A critical review of diagnostic methods for camel brucellosis Authors: Mammar Khames, Amaia Zúñiga-Ripa, Raquel Conde-Álvarez, Pilar M Muñoz, José María Blasco, Ignacio Moriyón Published: 2026 ABSTRACT Highlights•Most camel bacteriological studies are methodologically weak and of limited value.•No serological test has been validated for camels; standardization deficits are common.•The role of heavy-chain camel antibodies in these brucellosis tests is unknown.•Information in PCR-based studies is unreliable due to no standardization-validation.•Current prevalence data, risk analyses and reviews are flawed, hindering One Health actions.Keywords: Brucellosis, Camels, Diagnosis, Serology, Culture, PCR, RT-PCR Introduction Domestic camelids are multipurpose livestock comprising the dromedary (Camelus dromedarius), distributed mostly in Middle East and Africa, the two humped Bactrian camel (Camelus bactrianus), found mostly in northwestern China, southwestern Mongolia, Kazakhstan, Russia, Iran and other countries in Central Asia, as well as several species native to the Andean highlands. Camelids are vital to millions of households, especially for pastoralist communities in harsh environments where other livestock cannot thrive. They provide transport and essential products such as meat, milk, wool, hair and dung (Mohamed et al., 2024). Official figures for 2022 register approximately 37 million dromedaries and Bactrian camels (excluding feral populations) across 47 countries, many of which rank among the world’s poorest (World-Population-Review, 2025). Nevertheless, the figure likely underestimates true population sizes because many camel populations inhabit remote areas. A 2020 estimate suggested a global population exceeding 40 million, with projections reaching up to 60 million in coming decades (Faye, 2020). Camels are affected by several important bacterial zoonotic diseases (Hughes & Anderson, 2020; Mohammadpour et al., 2020), among which brucellosis has received special attention (Zhu et al., 2019). Caused by bacteria of the genus Brucella, this disease primarily manifests in domestic livestock as abortion and infertility, with bacterial shedding in aborted materials and milk serving as source of human brucellosis, a severe and debilitating illness. Brucellosis is globally distributed, and although prioritized in many endemic countries, its actual prevalence is largely unknown (Moreno et al., 2022). This uncertainty is attributable to the absence of pathognomonic clinical signs and the stealthy nature of the pathogen, which encumber awareness and suspicion of the bacteria involved and necessitate laboratory-based diagnosis (Laine et al., 2022; Moriyón et al., 2023; Vection et al., 2025; WOAH, 2023a). Even when diagnostic testing is implemented, WOAH recommendations and protocols are inconsistently applied (Vection et al., 2025), further compromising the reliability of reported prevalence data. Diagnostic methods for brucellosis are broadly classified as bacteriological (isolation and identification of Brucella spp.), immunological (detection of antibody or delayed-type hypersensitivity [DTH] responses), and DNA-based assays (detection of Brucella-specific DNA sequences). The WOAH provides guidelines for bacteriological and immunological methods, but not for DNA-based assays, emphasizing test standardization for reproducibility and quality control (WOAH, 2023a). While standardized protocols and WOAH or EU reference sera are available for common livestock species such as cattle, small ruminants, and swine, equivalent reference sera are lacking for camels and other affected species. Furthermore, the WOAH Manual stresses that serological tests must be validated for each host species and epidemiological context (WOAH, 2023a). However, as noted recently (Wernery et al., 2022) and despite earlier warning (Gwida et al., 2012; Sprague et al., 2012; Wernery, 2016), the standardization and validation of serological tests for camels have received minimal attention. Lack of standardization and validation are also extended problems of DNA-based diagnostic assays in animal brucellosis (Moreno et al., 2022), and the bacteriological methods also have methodological variables that influence their diagnostic performance. The validity of laboratory tests used for diagnosis of brucellosis in camels has been widely assumed rather than critically evaluated. This assumption is reflected in the uncritical acceptance of diagnostic sensitivity and specificity (DSe and DSp)1 estimates and reported seroprevalence values in reviews and meta-analyses (Abu-Basha et al., 2024; Adel, 2022; Akinyemi et al., 2022; Dadar et al., 2021, 2022; Dadar & Foster, 2025; Dagnaw et al., 2024; Gutema & Tesfaye, 2019; Hughes & Anderson, 2020; Menshawy et al., 2025; Mohammadpour et al., 2020; Mohammed et al., 2023; Refai, 2002; Sibhat et al., 2022; Tadesse, 2016; Waji & Neja, 2023; Wakene & Mamo, 2017; Wareth et al., 2014). Consequently, there is an urgent need for a critical appraisal of the laboratory tests applied to camels. The objective of the present review is therefore to identify key gaps and inconsistencies in current diagnostic approaches that must be addressed to enable effective brucellosis control in this important livestock. Literature search A systematic literature search was conducted in PubMed, CAB Direct, Scopus, Web of Science, and ScienceDirect for publications in English, French, or German. Search terms included combinations of “camel(s)” or “dromedary(ies)” with “brucellosis” or “Brucella,” adapted to the syntax requirements of each database. All retrieved records (n = 1168) were imported into Bookends (Sonny Software) for management. After removal of duplicates (n = 824) and exclusion of theses (n = 4), conference proceedings (n = 8), conference papers (n = 26), books (n = 7), and book chapters (n = 7), the titles and abstracts of the remaining 292 records were independently screened by MK, AZR, RCA, and IM for relevance to the objectives of this review. This process identified 22 review articles and 219 original research reports. An additional 10 original reports were retrieved through reference screening of included articles, resulting in a final total of 229 original research reports. Bacteriological diagnosis The isolation and identification of Brucella from an animal or its secretions is the only unequivocal proof of the disease. It is also an essential tool in epidemiological investigations, including molecular epidemiology, aspects in which it cannot be replaced by direct DNA-based diagnostic assays conducted in animal samples (Section 5). Indeed, bacteriology is laborious and hazardous, requires specialized knowhow and at least BSL2 infrastructure. Experience from domestic ruminants and swine shows that maximizing sensitivity requires the use of appropriate culture media or combinations of media, the collection and correct processing of a wide range of samples, the inoculation onto multiple plates of unusually large sample volumes (i.e., up to 0.5 mL or more per plate) starting from the lowest possible dilution (Moreno et al. 2022; Muñoz et al. 2022b), and incubation under conditions optimal for Brucella spp. Although culture, even optimized, may fail to detect the bacteria in some animals known to be infected by other evidence and if negative is not a reference of DSp, it remains irreplaceable as the gold standard for evaluating the relative DSe of other brucellosis diagnostic tests (Ducrotoy et al., 2018). Studies in non experimentally infected camels The literature review identified only 43 studies applying bacteriological methods to camels suspected of brucellosis. These investigations span approximately forty years and originate from just sixteen of the forty-seven camel-keeping countries (Table 1). Two studies claiming Brucella isolation employed culture media and incubation conditions incompatible with Brucella growth (Hirad et al., 2018; Mannivannan et al., 2022) and 7 failed to isolate Brucella (Adamu et al., 2007; Ahmed et al., 2017; Elshazly et al., 2020; Ghorbani et al., 2013; Hamdy & Amin, 2002; Salem et al., 1990; Wanjohi et al., 2012). Further limiting the number of camel-specific investigations, 3 studies were epidemiological analyses of human brucellosis cases (Bardenstein et al., 2021; Schulze zur Wiesch et al., 2010; Shimol et al., 2012) (Table 1). Among the remaining studies, one was a single-case report (Juhasz et al., 2019), and the others involved culture of milk, tissue, and/or abortion materials, generally with very low isolation rates (Table 1). Only 2 reports, each yielding a single isolate, concerned Bactrian camels (Bayasgalan et al., 2018; Kim et al., 2016) (an additional but inaccessible report describing B. abortus isolation in a Bactrian camel fetus in Russia is cited by Wernery (Wernery, 2016)). As discussed below, beyond this overall scarcity of bacteriological studies, most works reveal technical shortcomings that explain the low isolation rates, probable misidentifications and the above-mentioned failures to isolate Brucella. Table 1Bacteriological studies in naturally infected camels1,2.Table 1 dummy alt textReference [country]Samples (no.)Media and growth conditionsSpecies and biovar (no. strains)3Comments(Al Khalaf & el Khaladi, 1989) [Kuwait]Dromedary fetus stomach contents (5) and MRT+ milk (17).TSA or Brucella agar-serum-cycloheximide, bacitracin and polymyxin B; incubation with CO2.B. abortus biovar 1 (stomach content 2).Milk processing followed (Alton et al., 1988) but key variables (e.g. no of plates, inoculum size), etc.) were not described. Antibiotic supplement according to Alton et al. (1975). No milk samples yielded positive cultures. Stomach contents were inoculated into guinea pigs and brucellae were recovered at necropsy. Identification and biovar typing were performed in a WOAH reference laboratory.(Zowghi & Ebadi, 1989) [Iran]Lymph nodes (3500) from 300 dromedaries (serology not described).Serum dextrose agar -cycloheximide, bacitracin and polymyxin B;4 incubation with and without CO2.B. melitensis; biovars 1 (1) and 3 (2).No description of lymph node homogenization methods and amount seeded. Medium inhibiting some B. abortus and many B. melitensis strains (WOAH, 2023a). Conventional Identification and typing methods were used.(Nada et al., 1991) [Egypt]Retropharyngeal, supramammary, retropharyngeal and suprainguinal lymph nodes, spleen, liver and ovaries (902)TSA or Brucella agar-serum-cycloheximide, bacitracin;4 incubation with CO2.B. abortus biovar 7 (5 [1 from a seronegative camel]).No description of sample processing, amount seeded, and no of plates inoculated. Questionable implementation of classical biovar typing methods (the validity of B. abortus biovar 7 was suspended in 1986 (Corbel, 1988)).(Radwan et al., 1992) [Saudi Arabia]MRT+ milk from RBT+ dromedaries (120).BHI-dextrose agar; with Farrell’s antibiotic supplement; incubation with and without CO2.B. melitensis biovars 1 (26) and 2 (11).Insufficient amounts of inoculum (cotton swabs dipped in milk). Antibiotic supplement inhibitory for some B. abortus and many B. melitensis strains (WOAH, 2023a). Identification and biovar typing were performed in a WOAH reference laboratory.(Gameel et al., 1993) [Libya]Milk, vaginal swabs and fetuses (n.d.) of dromedaries from a farm with abortion cases.Farrell’s medium; incubation without CO2.B. melitensis biovar 1 (milk 5, fetuses 3; vaginal swab 1).No description of sample processing, amount seeded, and no of plates inoculated. Selective medium inhibiting some B. abortus and many B. melitensis strains; incubation without CO2 prevents isolation of B. abortus CO2-dependent biovars 1, 2, 3 and 4 (WOAH, 2023a).(Agab et al., 1994) [Sudan]Milk (n 22) and supramammary (2) and inguinal lymph nodes (5), testicles (5), vaginal swabs (3) and synovial fluid (1) from RBT+ dromedaries.Farrell’s medium; incubation with CO2.B. abortus biovar 3 (lymph nodes 2; vaginal swab 1)Insufficient amounts of inoculum (tissue homogenates, 0.1 mL on each of 3 plates; cotton swabs dipped in milk sediment). Medium inhibiting some B. abortus and many B. melitensis strains (WOAH, 2023a). Identification and biovar typing were performed in a WOAH reference laboratory.(Radwan et al., 1995) [Saudi Arabia]Milk from seropositive camels (120). Mammary tissues and lymph nodes from one additional seropositive camel.Farrell’s medium (incubation presumably with and without CO2).B. melitensis biovars 1, 2 and 3) (milk 41, necropsy 1)Insufficient amounts of inoculum (cotton swabs dipped in milk) (WOAH, 2023a). Unreliable serological screening: RBT performed in 60 °C heated serum rather than in normal serum (WOAH, 2023a) and BPAT and plate agglutination with serum dilutions used as “confirmatory tests” (see text).(Hegazy et al., 1998)[Egypt]Uteri and supramammary lymph nodes of seropositive dromedaries (RBT, SAT and Rivanol combined) (45).Brucella agar, Brucella medium base, and Blood agar base; incubation with no CO2.B. melitensis biovar 3 (n 4).Protocol for sample processing not described. Use of non-selective medium is a known cause of low isolation rates; incubation without CO2 prevents isolation of B. abortus CO2-dependent biovars 1, 2, 3 and 4 (WOAH, 2023a). Description of identification and biovar typing “according to physicochemical tests” calling the reported species and biovar identification into question.(Ramadan et al., 1998) [India]Carpal hygromas in 2 dromedaries.Farrell’s medium; incubation with and without CO2.B. melitensis (2, one from each animal).Medium inhibiting some B. abortus and many B. melitensis strains (WOAH, 2023a).(Khadjeh et al., 1999) [Iran]Popliteal, prescapular and supramammary Iymph nodes from seropositive (RBT-SAT and 2-ME) camels (5).Serum dextrose agar plates with antibiotics; incubation with CO2.B. melitensis biotype 1 (2) from Iymph nodes of two camels.No description of sample processing, amount seeded, and no of plates inoculated. Antibiotic supplement n.d. (presumably Farrells’s in (Alton et al., 1988) inhibiting some B. abortus and many B. melitensis strains (WOAH, 2023a)). Conventional Identification and typing methods were used.(El-Seedy et al., 2000) [Egypt]Uterus, placenta, fetal membranes, fetuses from RBT+ dromedaries (23).Serum Brucella agar without and with presumably Farrell’s antibiotics and liver infusion agar; incubation with CO2B. abortus biovar 1 (11) biovar 7 (6); and B. melitensis biovar 3 (3).Sample processing not described properly. Insufficient amounts of inoculum (samples streaked directly onto 2 plates of each medium). Medium inhibiting some B. abortus and many B. melitensis strains (WOAH, 2023a). Questionable implementation of classical typing methods (the validity of B. abortus biovar 7 was suspended in 1986 (Corbel, 1988)).(Hegazy et al., 2004) [Saudi Arabia]Mammary tissue and supramammary lymph nodes of RBT+ SAT+ dromedaries (7) with mastitisBrucella agar with (presumably) Farrell’s antibiotics; incubation without CO2.B. melitensis (5)Sample processing not described properly. Insufficient amounts of inoculum (samples streaked directly onto selective medium). Medium inhibiting some B. abortus and many B. melitensis strains; incubation without CO2 prevents isolation of B. abortus CO2-dependent biovars 1, 2, 3 and 4 (WOAH, 2023a). Identification and typing protocols are not described in the reference cited, calling the reported species identification into question.(Al-Majali et al., 2008) [Jordan]Stomach content; lungs, livers and spleens from aborted fetuses (7).TSA-cycloheximide, bacitracin, polymyxin B and nystatin; incubation with CO2.B. melitensis biovar 3 (4).Insufficient amounts of inoculum (100 μl of stomach content and swabs of lungs, livers and spleens). Medium inhibiting some B. abortus and many B. melitensis strains.(Dawood, 2008) [Jordan]Milk (n 26); aborted fetuses (stomach content, liver, spleen and lung) (n.d.) of RBT+, CFT+ dromedariesFarrell’s medium; incubation with CO2.B. melitensis biovar 3 (fetuses 2; fetuses; milk 2)Milk processing not described. Insufficient amounts of other inocula (swabs of liver, spleen and/or lungs directly on plates). Medium inhibiting some B. abortus and many B. melitensis strains (WOAH, 2023a).(Musa et al., 2008) [Darfur (Western Sudan)]Prescapular and supramammary lymph nodes from RBT+ dromedaries (6). Samples seeded were selected from the total by a Stamp-positive resultSerum dextrose agar; incubation without CO2B. melitensis biovar 3 (2;) and B. abortus biovar 6 (3).Methods insufficiently described. Selection of lymph node homogenates (processing and inoculum amount not described) by a positive Stamp staining severely compromises sensitivity. Use of non-selective medium is a known cause of low isolation rates; incubation without CO2 prevents isolation of B. abortus CO2-dependent biovars 1, 2, 3 and 4 (WOAH, 2023a). Some cultures were inoculated into guinea pigs and brucellae were recovered at necropsy. Identification and biovar typing were performed in a WOAH reference laboratory.(Omer et al., 2010) [SudanMilk (225), lymph nodes, testicular and udder tissues (n.d.) from RBT+, cELISA+ dromedaries. Samples seeded (n.d.) were selected from the total by a Stamp-positive result.Serum dextrose agar or tryptone agar; incubation with CO2 using candle jars.B. abortus biovar 6 (2)Tissue homogenates seeded on plates (processing and inocula not described; heavily contaminated samples were inoculated into guinea pigs). Selection of milk/tissue homogenates by a positive Stamp staining severely compromises sensitivity. Use of non-selective medium is a known cause of low isolation rates; plates discarded after 4 days of incubation further limit detection (WOAH, 2023a). Identification and biovar typing were performed in a WOAH reference laboratory.(Schulze zur Wiesch et al., 2010) [United Arabic Emirates]Milk samples from 3 camels.n.d.B. melitensis biovars 1 (2) and 2 (1)Epidemiological investigation of a laboratory-acquired infection. Culture medium and sample processing not described. The biovar 2 isolate was confirmed by MLVA-16 as the source.(Ibrahim et al., 2012) [Egypt]Milk (64) from RBT + dromedaries obtained in a 9-year time span.n.d.B. melitensis biovar 3 (19).Milk processing followed (Alton et al., 1988) but key variables (e.g. no of plates, inoculum size), etc.) were not described. Identification and typing methods were not reported.(Shimol et al., 2012) [Israel]Milk from a dromedaryn.d.B. melitensis biovar (1).Epidemiological investigation of an outbreak affecting 15 members of a family consuming milk from an infected camel (identified serologically after the outbreak). Sample processing not described. Culture medium n.d. Conventional Identification and typing methods were used.(Hazem & Ismail, 2014) [Egypt]MRT+ milk (18)TSA-serum with no antibiotics; incubation with CO2B. melitensis (5)Culture methods insufficiently described and only milk sediment cultured. Use of non-selective medium is a known cause of very low isolation rates (WOAH, 2023a). Identification by AMOS PCR(Hosein et al., 2016) [Egypt]Lymph nodes and testicular tissues from males (n 128) and milk (26) of mRBT+, cELISA+ of apparently healthy dromedariesEwalt’s selective medium; incubation with CO2.Milk inoculated in guinea pigs.B. melitensis biovar 3 (2).Unreliable report: no of samples (128 plus 26) and seropositive animals (47 RBT+, 42 iELISA +) do not coincide. Protocol followed (Alton et al., 1988) but key variables (e.g. no of plates, inoculum size), etc.) were not described. No sample yielded Brucella on culture plates (the two strains were recovered from guinea pigs inoculated with milk).(Kim et al., 2016) [Mongolia]Bactrian camels (n.d.)n.d.B. abortus biovar 7 (1).No description of sample and culture methods. Strain included in the retrospective MLV characterization of a collection strains. Questionable implementation of classical typing methods (the validity of B. abortus biovar 7 was suspended in 1986 (Corbel, 1988)).(El-Sayed et al., 2017) [Egypt]Hygroma and mesenteric, retropharyngeal and supramammary lymph nodes of BPAT+ RBT+ and CFT+ dromedaries (n.d.) and fetal stomach contents of aborted animals (2).Farrell’s medium; incubation without CO2.B. melitensis biovar 3 (1).Protocol followed (Alton et al., 1988) but key variables (e.g. no of plates, inoculum size), etc.) were not described. Medium inhibiting some B. abortus and many B. melitensis strains; incubation without CO2 prevents isolation of B. abortus CO2-dependent biovars 1, 2, 3 and 4 (WOAH, 2023a).(Abd El Tawab et al., 2018) [Egypt]Tissues and lymph nodes from slaughtered seropositive dromedaries (n.d.)n.d.B. abortus (4 from lymph nodes and 1 from a testicle).Specific medium used n.d. Key aspects of culture methods (sample processing, no of plates, inoculum size) not described in the reference cited (Joint-FAO/WHO-Expert-Committee-on-Brucellosis., 1986). Tissues and lymph nodes not systematically obtained from every animal. Identification by AMOS PCR.(Bayasgalan et al., 2018) [Mongolia]Swabs (72) and/or milk (104) from bactrian camels with history of abortion, including 37 RBT+.Farrell’s and CITA media; with and without CO2B. abortus (1).Protocol for milk processing followed (Alton et al., 1988) but key variables (e.g. no of plates, inoculum size), etc.) were not described. Exceedingly low isolation rate (1/176 samples) despite high seroprevalence (also low for sheep: 0/75, goats:0/75 and cattle 3/119). Identification by Bruce-Ladder.(Foster et al., 2018) [Oman]Placenta, aborted fetuses, stomach, milk, vaginal swabs /discharge (n.d.) from aborted animals (n.d.)Sheep blood agar with Farrell’s selective supplement and 2.5% glucose; with or without CO2 for 7 days.B. melitensis (placenta 1, aborted materials 8, milk 4, vaginal swabs 2).Retrospective SNP and VTRN typing of isolates obtained in the 1997–2007 period. Sample processing, no of plates, etc. not described or referred. Medium inhibiting some B. abortus and many B. melitensis strains (WOAH, 2023a). Identification and typing by standard methods confirmed retrospectively by PCR according to Hinic et al. (2008).(Khan et al., 2018) [Saudi Arabia]Dromedary milk (80).Brucella agar plates incubated for 3 days at 30 °C.Brucella spp. (9).Not credible report. The medium and growth conditions (poorly described) would not support Brucella growth or isolation. Whole milk was seeded without processing; inoculum volume and no of plates were not reported. Identification relied solely on an unvalidated, insufficiently described “rep-PCR” (sic.) method.(Alamian & Dadar, 2019) [Iran]Milk (96) from dromedaries of unknown serological statusFarrell’s medium with CO2.B. abortus biovars 3 (1) and 5 (2).Protocol for milk processing followed (Alton et al., 1988) but key variables (e.g. no of plates, inoculum size), etc.) were not described. Medium inhibiting some B. abortus and many B. melitensis strains (WOAH, 2023a). Identification and typing by AMOS-PCR and Bruce-Ladder.(Juhasz et al., 2019) [Dubai]Blood, lymph nodes and other samples (see text), from an aborted RBT+ dromedary; several samples from the aborted fetus.Farrell’s, brain-heart and tryptic soy broth both with Farrell’s antibiotics, with CO2.B. melitensisCase study of a single camel. Thorough necropsy of an aborted dam and the fetus. Bacterial load determined (see text).(Dadar & Alamian, 2020) [Iran]Subscapular lymph nodes from serologically positive (10) and negative (30); RBT-SAT and 2-ME in series dromedaries.Farrell’s medium with CO2B. melitensis biovar 1 (2)Sample processing not described. Samples not representative of Brucella-colonized lymph nodes. Medium inhibiting some B. abortus and most B. melitensis strains (WOAH, 2023a). Identification and typing by AMOS-PCR and Bruce-Ladder. Self-contradictory report: culture, serology and a non-validated PCR do not correlate, and the 2 strains were isolated only from seronegative camels (see Table 2 and text).(Bardenstein et al., 2021) [Israel]Commercially sold milkn.d.B. melitensis (3).Epidemiological investigation of an outbreak affecting persons consuming the same brand of milk, 19 of whom were culture positive. Of these 19, at least 7 had the same WGS profile as the milk isolate. Culture methods not described.(Mannivannan et al., 2022) [Qatar]Serum from mRBT+, cELISA + dromedaries (20)TSA, Blood agar and MacConkey; Farrell’s medium with CO2.93% (sic.) of samples yielded B. melitensisNot credible report. Serum is not adequate for Brucella isolation and fifty percent of “Brucella” isolates grew on MacConkey (brucellae do not grow on this medium) (WOAH, 2023a). Identification based on Gram staining, bacterial morphology and VITEK-2 COMPACT GN, which is known to misclassify Ochrobactrum as Brucella (Mori et al., 2020).1These six studies failed to isolate brucellae and are not included in the Table: (Adamu et al., 2007; Ahmed et al., 2017; Elshazly et al., 2020; Ghorbani et al., 2013; Hamdy & Amin, 2002; Salem et al., 1990; Wanjohi et al., 2012).2Abbreviations: n.d., not done; MRT, milk ring test; SAT, serum agglutination in tube; 2-ME, 2-mercaptoethanol test; RBT, rose bengal test; CFT, complement fixation test; cELISA, competitive ELISA.3Biovar is indicated when reported. Bacteriological studies in naturally infected camels1,2. These six studies failed to isolate brucellae and are not included in the Table: (Adamu et al., 2007; Ahmed et al., 2017; Elshazly et al., 2020; Ghorbani et al., 2013; Hamdy & Amin, 2002; Salem et al., 1990; Wanjohi et al., 2012). Abbreviations: n.d., not done; MRT, milk ring test; SAT, serum agglutination in tube; 2-ME, 2-mercaptoethanol test; RBT, rose bengal test; CFT, complement fixation test; cELISA, competitive ELISA. Biovar is indicated when reported. A variety of camel samples have been used (Table 1). Although the frequent use of milk is explained by an easier access, it is well-known that shedding of brucellae in the milk of ruminants is intermittent, does not occur from all quarters and, as numbers of bacteria can be very low, an optimized protocol is required for successful isolation (Alton et al., 1988). There is evidence that at least some of these critical limiting factors occur in camels (see below). Also, 4 studies selected milk for culture by a positive Milk Ring Test (MRT) and one by a positive Stamp staining for Brucella, but MRT has no diagnostic value in camels (Section 4.13.) and Stamp staining of milk is very unsensitive (WOAH, 2023a). Five reports describe the lymph nodes investigated and, with one exception, they were no more than three (Table 1). However, it has been demonstrated in cattle and sheep that optimal sensitivity requires culturing a full set of necropsy samples (Blasco et al., 1994b; Ducrotoy et al., 2018), and this should be also the case of camels (Wernery et al., 2022). The exception to this scarcity of samples is the single-camel case study by Juhasz et al. (Juhasz et al., 2019) who systematically necropsied an aborted dam and isolated B. melitensis from the right submandibular, pharyngeal, retropharyngeal, prescapular, and tracheobronchial lymph nodes, as well as from lung and one udder quarter, but not from the left submandibular lymph node, other udder quarters, uterine horns, liver, blood or milk. It is also worth noting that of the three hygroma fluids tested from naturally infected dromedaries, two were culture-positive (Ramadan et al., 1998) and one, although not cultured, was Stamp-staining positive and RBT-positive (Musa et al., 2008), indicating the convenience of including carpal bursitis fluids in the above set of samples to increase the sensitivity of the culture. Juhasz et al. (Juhasz et al., 2019) also quantified the Brucella colony forming units (CFU) in the lungs pleural and abdominal fluids and liver of the aborted fetus as well as in the umbilical cord and placenta and found very high numbers (107 to 1010 CFU/g of tissue). While these figures could suggest that isolation of Brucella from these materials should be very productive, their use is not exempt of problems. Aside from the significant biohazard, if fetuses or placental tissues are not collected and refrigerated promptly, the fast-growing microbiota multiplies extensively. This results in heavily contaminated samples, practically impairing the isolation of the slow-growing organisms B. abortus and B. melitensis in culture. Instead, in domestic ruminants and pigs, vaginal swabs in an appropriate transport medium are the sample of choice because high numbers of brucellae are excreted in vaginal fluids for several weeks after abortion or premature delivery. Interestingly, of the 4 studies that used this approach (Table 1), three were successful (Agab et al., 1994; Foster et al., 2018; Gameel et al., 1993) and the one that failed, performed in Bactrian camels (Bayasgalan et al., 2018), is not very informative because the exceedingly poor isolation rates in camels and other species investigated simultaneously in areas of high seroprevalence (Table 1) raise concerns about the protocols used. Additional limitations in bacteriological implementation are summarized in Table 1. When reported, sample processing and inoculation protocols were frequently suboptimal, including insufficient inoculum size and limited numbers of culture plates (6 studies), absence of Brucella-selective media (5 studies), or incubation under a normal atmosphere (6 studies). The latter precludes isolation of B. abortus CO₂-dependent biovars (Table 1) (Alton et al., 1988; WOAH, 2023a). Farrell’s medium, or its antibiotic supplement in alternative basal media, was commonly used (16 studies), likely due to the commercial availability of the supplement. However, this medium inhibits some B. abortus strains and is strongly inhibitory to B. melitensis and B. suis (de Miguel et al., 2011; WOAH, 2023a), which combined with non-optimized samples and inoculum amounts further reduces the diagnostic efficacy. Table 1 also summarizes the reported species and biovars. Conventional identification requires an analysis of CO2 requirement, H2S production, sensitivity to thionine and basic fuchsine and agglutination with A and M monospecific sera, if possible complemented by phage typing (Alton et al., 1988). However, these methods require solid expertise, dye sensitivity in particular, and the monospecific sera and phages are not widely available. Thus, it is strongly recommended that presumptive isolates are sent to a reference laboratory for species and biovar typing always using well-maintained reference strains as controls (WOAH, 2023a). This was done in only 6 works (Table 1). To circumvent these difficulties, molecular methods applied to pure cultures were developed for identification and typing. They include AMOS-PCR for species (Bricker & Halling, 1994) Bruce-ladder for species and biovars (López-Goñi et al., 2011) and more recently Multiple Loci Variable Number of Tandem Repeats Analysis (MLVA) and Whole Genome Sequencing (WGS). AMOS-PCR and Bruce-ladder are simpler, and the latter is a WOAH-recommended method (WOAH, 2023a). According to classical and molecular typing, no B. abortus or B. melitensis biovar seems dominant in camels (Table 1) However, approximately half of the studies conducted after the introduction of AMOS-PCR and Bruce-ladder relied exclusively on classical biovar typing. Although the number of strains analyzed is limited and classical biovar typing is not always reliable, the available data provide no evidence for a camel-associated B. abortus biovar. This contrasts with the conclusions of a recent meta-analysis, which reported a predominance of B. abortus biovar 6 in camels (Dadar et al., 2022). One study reported the use of the automated VITEK-2 COMPACT GC system to identify isolates from dromedary sera (Mannivannan et al., 2022) (Table 1). However, serum is unsuitable for the isolation of Brucella spp., and the unusually high number of isolates obtained, together with their growth on MacConkey agar (not permissive for classical Brucella species), indicates major methodological flaws. Moreover, the VITEK-2 COMPACT GC system has been shown to misidentify Ochrobactrum spp. as Brucella (Mori et al., 2020), which likely explains these findings. Ochrobactrum spp. are predominantly environmental bacteria that have been erroneously included within the genus Brucella in several databases, leading to misclassification by automated identification systems (Moreno et al., 2023). These organisms grow rapidly at 30 °C, are Stamp-positive, and may cross-react with certain DNA probes, features that could also account for the results reported by Khan et al. (Khan et al., 2018), which are likewise incompatible with Brucella (Table 1). Study in experimentally infected camels Additional relevant evidence is provided by an experimental study in which 14 female dromedaries were exposed intranasally or intratracheally to a well-characterized B. melitensis strain Johnson et al. (2018). A total of 541 samples representing 43 organs and joint fluids from each animal were cultured on Farrell’s medium, brain–heart infusion (BHI) agar supplemented with Farrell’s antibiotics, and tryptic soy agar (TSA). The challenge strain was recovered from all 14 animals (48 samples), predominantly from submandibular, pharyngeal, prescapular, mammary, and lung-associated lymph nodes; however, bacterial distribution varied unpredictably among tissues and animals. The methodology warrants further consideration. The authors compared the performance of plates inoculated either directly by streaking cut tissue specimens or 0.1 mL of joint fluid or following centrifugation of tissue homogenates in 30 mL buffer, with 0.1 mL of the resulting pellet used for inoculation (concentration method). In addition, 1.0 mL of the pellet was inoculated into tryptic soy broth supplemented with Farrell’s antibiotics, incubated, and subsequently subcultured onto tryptic soy and BHI agar with Farrell’s antibiotics (enrichment–concentration method). Direct inoculation yielded poor results, while the enrichment–concentration method performed only slightly better to the concentration method. Although the authors favor the use of BHI as the basal medium because visual identification of Brucella suspected colonies may be easier, there were no differences in efficacy between the tryptic soy and the BHI agars in this and in a previous study (Wernery et al., 2015). Serological diagnosis in camels Characteristics of camel antibodies Camelid antibodies have peculiarities that make them exceptional among mammals (Hamers-Casterman et al., 1993). In addition to conventional IgM and IgG1 with identical light and heavy chains, Camelidae have IgG2 and IgG3 heavy chain antibodies (HCAbs) devoid of light chains. Thus, HCAbs cannot form the heterodimeric VL-VH antigen-binding pocket of IgM and IgG1 and bind the antigen only through VH. HCAbs have also shorter HCs lacking the CH1 domain and, while IgG2 has an unusually long hinge separating the CH2 and VH domains, IgG3 displays a very short hinge. HCAbs represent up to 75% of serum IgG (Hamers-Casterman et al., 1993) and, like IgG1, they are excreted in colostrum and milk where they are also in high amounts (Azwai et al., 1996; Yao et al., 2017). Camel IgA, which is present in mucosae and milk, also displays both conventional and HCAb structures (Conca et al., 2023). The structural characteristics of HCAbs are likely to influence antigen–antibody interactions in vitro. In brucellosis diagnosis, agglutination is the basis of Serum Agglutination Test (SAT), 2-Mercaptoethanol (2-ME), Rivanol, Rose Bengal test (RBT), Card Test, and Buffered Plate Agglutination Test (BPAT), and activation of guinea pig complement is essential for the Complement Fixation test (CFT) (Table 2). The reduced distance between adjacent VH domains in camel IgG3 (approximately half that of conventional heterodimeric IgG) is predicted to limit cross-linking capacity, thereby reducing agglutination and immunoprecipitation efficiency (Hamers-Casterman et al., 1993). As shown in humans, antibody-dependent complement activation is associated with C1q affinity for the Fc region and with hinge region length and flexibility (Frischauf et al., 2024). Accordingly, complement activation mediated by HCAbs may be impaired due to the absence of the CH1 domain and the short hinge region of IgG3. Consequently, it is premature to assume that the above-mentioned serological tests detect IgM, IgG1, IgG2, and IgA in camelids in the same manner as in ruminants, as assumed in a recent review (Dadar & Foster, 2025). Table 2Brucellosis serological tests used in camels.Table 2 dummy alt textTestAntigen (epitopes detected)Characteristics and standardizationSerum agglutination in tube (SAT), 2-Mercaptoethanol (2-ME) and Rivanol.S-Brucella cell suspension (O-PS)Quantitative tests carried out at neutral pH. In ruminants, SAT detects IgM and the agglutinating fraction of ruminant IgG (but not non-agglutinating IgG); there may be prozone-like effects. The antigenic suspension should be standardized using an official reference serum (equivalent of the WOAHISS, previously WSISSBaS, of bovine origin). In non vaccinated cattle, a titer ≥ 30 I.U./mL is considered positive (WOAH, 2023a).Under the test conditions, 2-mercaptoethanol and Rivanol [ethacridine lactate] preferentially inactivate IgM in ruminant sera. These tests do not show useful DSe/DSp in ruminants and are not currently recommended by WOAH (WOAH, 2023a).Buffered Brucella antigen tests (Rose Bengal test [RBT], Card test and buffered plate agglutination test [BPAT].S-Brucella cell suspension (O-PS).Qualitative tests carried out at pH 3.7. In domestic ruminants, IgM and IgG (acidity renders all IgG agglutinating) are detected, and there are no prozones. In cattle, the performance of RBT, Card test, and BPAT is practically identical. They require clean non-hemolyzed sera from well-coagulated blood (fibrine causes unspecific agglutinations).RBT and the Card test use an 8% suspension of B. abortus 19 or 1119–3cells. For cattle 30:30 μL serum:antigen and 4 min incubations at room temperature. For small ruminants: 75:25 μL (modified RBT [mRBT]) and 4 min incubations at room temperature is recommended. BPAT uses an 11% suspension of B. abortus 1119–3 cells, 80:30 μL serum:antigen and 8 min incubation at room temperature. The antigenic suspensions should be standardized using the same official reference serum as SAT (WOAH, 2023a).Complement fixation test (CFT).S-Brucella cell suspension (O-PS).Quantitative test that detects antibodies activating guinea pig complement. In bovines, and likely in small ruminants, CFT is an IgG1 test because IgM and IgG2 of this livestock do not activate guinea pig complement. Sera are heated (58 °C [bovine sera] or 63 °C [small ruminant sera] for 30 min) to inactivate the natural complement in domestic ruminant sera, but the optimal conditions for inactivation of complement in sera of other animals, including camels, has not been determined. The antigenic suspension should be standardized using the same official reference serum as SAT. In non-vaccinated cattle, a titer is ≥ 20 ICFTU/mL is considered positive (WOAH, 2023a). Not applicable in hemolyzed sera or sera showing anti-complementary activity.Indirect ELISAs (iELISAs).S-LPS (O-PS, core and lipid A)A family of quantitative immunosorbent assays that vary in adsorption protocols, antigen type and quality, antibody-detecting conjugates, enzyme-substrate and kinetics of color development. The animal species immunoglobulin and isotype detected (usually IgG) depends on conjugate specificity. Each assay should be standardized with reference sera of each animal species. For cattle iELISA, there are three WOAH reference sera, strong positive (WOAHELISASPISS), weak positive (WOAHELISAWPSS), and negative (WOAHELISANSS) (formerly OIEELISASPISS, OIEELISAwPISS and OIEELISANISS), which used together allow to standardize the operating or dynamic range of the assay (WOAH, 2023b); for sheep/goats there is also an international reference positive serum (ISaBmS) to standardize the iELISA analytical sensitivity threshold. There are no WOAH standard sera for pigs or camels.Competitive (or blocking) ELISA (cELISA)S-LPS (O-PS, core and lipid A)A family of quantitative immunosorbent assays. In cELISA an anti O-PS monoclonal antibody (MoAb) conjugate (varies by kit) competes with low-avidity antibodies in the serum tested. Each assay should be standardized using the same reference sera as for iELISA. cELISAs do not differentiate immunoglobulin classes and may be applicable across animal species.Fluorescence polarization assay (FPA).S-LPS hydrolytic polysaccharide (OPS-core).Fluid phase quantitative test carried out with fluorescein-labeled S-LPS hydrolytic polysaccharide (OPS-core). Antibody binding slows free rotation of the labelled molecule increasing polarization of a light beam. Requires specific reader. FPA does not differentiate immunoglobulin classes and may be applicable across animal species. Standardization uses the same sera as iELISA. Sera must be free of particulate matter, including lipids.Lateral flow immunochromatography (LFiC).S-LPS (O-PS, core and lipid A)A family of immunosorbent assays carried out on a chromatographic matrix. In camel studies, LFiC is named immunochromatographic test/assay (Eisa et al., 2014; Serhan et al., 2019), MSA Antigen Rapid Camel Brucella Ab Test (Almuzaini et al., 2025), and immunochromatographic assay (Abd El Tawab et al., 2018). Diluent flow directs the antibodies to the S-LPS line where those bound to the S-LPS are detected with an anti-immunoglobulin conjugate (usually colloidal gold). The animal species immunoglobulin and isotype detected depends on conjugate specificity. Brucellosis serological tests used in camels. Brucella antigens relevant in diagnostic tests All relevant brucellosis serological tests detect antibodies to the smooth (S) lipopolysaccharide (S-LPS) of the bacterial surface, and no assays based on alternative antigens achieves comparable DSe in at least humans, domestic ruminants and swine (Dieste-Pérez et al., 2015; Ducrotoy et al., 2016; Díaz & Moriyón, 1989). Within the S-LPS, the O-polysaccharide (O-PS) constitutes the immunodominant region, whereas the inner sections (core oligosaccharide and lipid A) carry epitopes of limited relevance if any in infections by S brucellae. The O-PS carries a highly repeated C epitope shared by all zoonotically significant S brucellae; consequently, antigens from B. melitensis and B. abortus are equally effective regardless of the infecting species (Ducrotoy et al., 2016), a finding also confirmed in camels (Azwai et al., 2001). Indeed, B. abortus antigens are normally used because well-characterized and relatively stable (see Section 4.3.) reference strains are available (WOAH, 2023a). Some gram-negative bacteria possessing O-PS structures related to those of the main S brucellae can cause false-positive serological reactions (FPSR) in S-LPS based assays (Ducrotoy et al., 2016). Although some camel studies assume that FPSR occur in RBT but not in CFT, iELISA, cELISA or LFiC (ICT) (Ahad et al., 2024; Edema et al., 2025; Esubalew & Robayo, 2017; Moti, 2022; Omer et al., 2010; Schelling et al., 2004; Serhan et al., 2019), no S-LPS based test can discriminate FPSR from true Brucella infections (Dieste-Pérez et al., 2015; Ducrotoy et al., 2018). As far as it is known, only Yersinia enterocolitica O:9 (whose O-PS is almost identical to that of S brucellae) is recognized as a significant source of FPSR but only under certain circumstances and mainly in cattle and swine. Although FPSR have been scarcely investigated in camels, five cases, including four isolations of Y. enterocolitica O:9 were reported in asymptomatic animals imported from China to Japan (Sunaga et al., 1983). Given that camels described as affected by brucellosis are often asymptomatic, FPSR in these animals deserves investigation. Test standardization and validation Antigenic stability is an issue whose importance cannot be overemphasized and one of the most important problems in the correct implementation of brucellosis serological tests. All S brucellae are prone to dissociate because some peculiarities of the O-PS coding genetic regions promote excisions (Mancilla et al., 2010, 2012). In vitro, these mutations cause the emergence of O-PS lacking rough (R) mutants (S-R dissociation) that overgrow the S bacteria and result in flawed diagnostic antigens (Ducrotoy et al., 2016). Consequently, precautions to minimize S-R dissociation during culture combined with standardization of the antigens and quality control using reference sera are essential. For this purpose, there are international WOAH standard sera of bovine and goat origin Table 2 kept at some WOAH brucellosis reference laboratories but, as they are in limited amounts, other reference centers should prepare their own secondary national standard sera (WOAH, 2023a). Standardization is necessary for reproducibility (and within laboratory repeatability) and does not imply validation (WOAH, 2023a) (see below). There is a paucity of information about the implementation of brucellosis serological tests standardization and quality control in the laboratories of camel keeping countries, an issue explored in Section 4.10. Validation is the process of determining the fitness of a standardized test for a specific purpose, most commonly diagnosis, and yields DSe and DSp values that vary by animal species and epidemiological context (WOAH, 2023c). However, few studies have assessed the DSe and DSp of serological tests in camels (discussed in Section 4.11). Consequently, most studies assume DSe and DSp values for RBT, CFT, iELISA, cELISA, and FPA based on otherwise unsubstantiated manufacturer claims for other species or extrapolations from studies conducted in small ruminants and cattle (Supplemental Tables S1, S2 and S3), an approach that is methodologically invalid (WOAH, 2023c). Brucellosis agglutination tests at neutral pH The agglutination tests at neutral pH used in camels are SAT, 2-ME and Rivanol and very rarely the USDA Plate Agglutination test (Radwan et al., 1983, 1992, 1995), SAT-phenol (Wernery & Wernery, 1990) or SAT-EDTA (Salisu et al., 2017), respectively described as reducing unspecific agglutinations in cattle (Ducrotoy et al., 2016) and (according to preliminary results) in camels (Wernery & Wernery, 1990). One study used 5% NaCl rather than saline, claiming that it reduces prozones in SAT and results in a slight increase in sensitivity but this claim is not supported by the use of properly characterized sera (Salem et al., 1990). SAT (Table 2) has been used in 59 of 177 studies since 1977 (Supplemental Table S1). However, the interpretation in terms of the immunoglobulin detected by this assay in camels is unclear and the DSe and DSp have not been established. Agglutination by camel IgG3 may be limited (see above) and it is unknown whether camel IgG1 and IgG2 include fractions unable to agglutinate at the neutral pH of SAT. This would not be unexpected because it occurs in bovine, small ruminant and human sera where non-agglutinating antibodies occur as a direct function of the time of infection limiting the DSe of SAT (Ducrotoy et al., 2016; Díaz & Moriyón, 1989). Moreover, the diagnostic titers 1:20 or 1:40 (IU/mL are seldom reported) used in 38 of the 59 studies are direct transpositions of the cut-off titers recommended for domestic ruminants (Table 2 and Table S1) and have not been validated in camels (WOAH, 2023a) (see also Section 4.10.). Many studies fail to specify the cut-off used or apply titers (e.g. 1:2, 1:5, 1:10, 1:100 or 1:160) without support from the literature of brucellosis in any animal species. These issues are compounded by the lack of descriptions of SAT antigen origin and/or standardization, or appropriate references, in a substantial proportion of studies (Table S1). The same concerns apply to 2-ME and Rivanol (efficient detection of all camel IgGs, diagnostic titers and standardization), tests aimed to eliminate IgM and to evaluate IgG responses in domestic ruminants. In addition, these tests were introduced for bovine brucellosis when it was believed that IgM was an important source of unspecific reactions and that only IgG indicated an active infection, two ideas proved erroneous (Ducrotoy et al., 2016). Furthermore, their use in camels assumes that the mercaptan in 2-ME depolymerizes camel IgM at the concentration and incubation conditions used in cattle while not affecting IgG1 or HCAbs, and that ethacridine lactate (the Rivanol reagent) selectively pricipitates camel IgM. However, there are no studies in camels proving this, and the parallelism cannot be taken for granted because of the IgG structural differences. In cattle, SAT shows comparatively low DSe (Ducrotoy et al., 2018), is of very limited usefulness and only under special circumstances (WOAH, 2023a) and its has been virtually discontinued. The 2-ME and Rivanol tests were found also of no practical value, and they have not been included in the WOAH (formerly OIE) Manual since 2004. Although their use in camels has also declined, these obsolete and not informative tests were still used in recent studies (Table S1). Brucellosis agglutination tests at acid pH Collectively termed buffered Brucella antigen tests, the RBT, Card test, and BPAT have been widely used in bovine and small ruminant brucellosis. Despite differences in formulation (Table 2), they show equivalent diagnostic performance in these animals. RBT is the most commonly used and, when sera are properly processed (fibrin may cause false-positive reactions), exhibits 96–100% DSe and 95–100% DSp in cattle not vaccinated with S19 and in the absence of FPSR, with values comparable to or slightly exceeding respectively those of iELISA and cELISAs (Ducrotoy et al., 2018). In small ruminants, optimization of DSe requires increasing the serum-to-antigen ratio (mRBT, Table 2), a change that does not compromise the very high DSp in sheep and goats not vaccinated with Rev.1 (Blasco et al., 1994a). The high DSe of RBT is explained by the ability of acidic buffer conditions to render the fraction of non-agglutinating IgG reactive, while maintaining detection of IgM; these principles also apply to BPAT (Ducrotoy et al., 2016). Frequently described as a screening test (but see Section 4.12), RBT (or mRBT) is the most widely used assay in camels (150 of 177 studies in Table S1). However, the effects of acidic pH on HCAbs and the presence of non-agglutinating IgG at pH 7 in camels remain unknown. Standardization is critical for RBT performance and cannot be assumed regardless of antigen source (Blasco et al., 1994a). The relevance of this issue is illustrated by a study of 14 camels experimentally challenged with B. melitensis and infection bacteriologically confirmed (Johnson et al., 2018), in which 540 serum samples collected over twelve months were tested using five commercial RBT antigens (Rose Bengal Antigen, [APHA Scientific], BENGATEST [Symbiontics Europe/Zoëtis], Pourquier Rose Bengal Ag [IDDEX], Rose Bengal Test APHA Scientific and Rose Bengal [Vircell]). Only one preparation (Rose Bengal [Vircell], marketed for human diagnosis) showed high sensitivity, highlighting marked variability among antigens (Soellner et al., 2018; Wernery et al., 2022). Although dose, route and strain of experimental infections are not representative of their field counterparts and thus the sera not suitable for test validation, they provide adjunct standards for comparison (Jacobson, 1998). Therefore, these findings strongly suggest inadequate standardization of at least four commercial RBT antigens used in camel studies (Supplemental Table S1). Consequently, RBT (and the mRBT used in 17 camel studies) remains to be validated in this species (WOAH, 2023a) (see also Section 4.10). The brucellosis complement fixation test (CFT) The brucellosis CFT uses guinea pig complement, a heterologous component that in bovines is activated by IgG1 but not by IgM or IgG2 (Ducrotoy et al., 2016). The capacity of camel IgM, IgG1, and HCAbs to activate guinea pig complement has not been investigated, which would be necessary for a proper interpretation of the CFT results and that may be compromised in HCAbs (see above). In addition, the CFT protocol includes heat inactivation of homologous complement in the test sera (Table 2), but the effects of the prescribed temperature and duration on camel sera are unknown. As with the serum agglutination test, CFT has not been validated in camels in terms of DSe and DSp (WOAH, 2023a) (see also Section 4.10), and diagnostic titers are extrapolated from ruminant studies. Of the 177 studies reviewed, 57 employed CFT (Table S1), but only 7 reported results in international units (ICFTU); of these, 5 applied the 20 ICFTU/mL cut-off recommended for cattle (WOAH, 2023a). When described (18 studies did not report titers), diagnostic titers most often ranged from 1:4 to 1:8 (23 studies) but cut-offs of 1:10, 1:16, 1:20, or 1:40 were also used (Table S1). As with SAT, information on antigen origin/standardization is often lacking. In many camel studies, CFT was used as a “confirmatory” test following RBT, an erroneous strategy discussed below (Section 4.12.). Brucellosis iELISA and cELISA Under the name of iELISA, there is a family of assays that differ in antigen (S-LPS of variable purity or S-LPS hydrolytic polysaccharide in some kits [Table S2]) and conjugate (protein A, protein G, anti-IgG monoclonal antibody, anti-IgG polyclonal antibody of either H+L or Fc specificity), enzyme-substrate and kinetics of color development. Consequently, in addition to validation for host species and epidemiological situation, (WOAH, 2023a, c), each ELISA kit needs to be standardized with appropriate sera (a minimum of two, optimally three; Table 2) to provide a useful dynamic range. Although the information specified by manufacturers and/or authors is often incomplete, these variables are reflected in camel studies (Supplemental Table S2). Conjugate specificity and standardization are critical as they affect the background and top reactivity of the dynamic range, and hence the DSe/DSp upon validation. Most camel studies published after 2015 that included iELISA used the ID Screen Brucellosis Serum Indirect ELISA Multispecies (IDVet, France) (23 of 33 studies; Tables S1 and S2). The manufacturer provides this kit for the detection of antibodies against B. abortus, B. melitensis, and B. suis in bovine, ovine, caprine, and porcine serum, and it is not directly applicable to sera from other livestock species. Although this specific information could not be found on the webpage of the manufacturer (it merely gives a list of publications where the kit was used), two works indicate that the conjugate is protein G-based (Benfodil et al., 2022; Khan et al., 2020), which is known to bind IgG of ruminants and swine efficiently. However, in affinity chromatography protein G does not bind the HCAb IgG2 and camel IgG1 and HCAb IgG3 elute under different conditions suggesting a different kind of interaction with protein G (De Genst et al., 2006; Hamers-Casterman et al., 1993). These differences may explain why in a screening of sera from 160 animal species, a protein G-conjugate did not bind dromedary and Bactrian camel sera in the range of concentrations tested (Stöbel et al., 2002). Also, in the above-mentioned study in experimentally exposed camels the ID Vet iELISA was less sensitive than the RBT (Soellner et al., 2018; Wernery et al., 2022). It seems therefore that protein G-conjugates have some limitations for camel serological studies and that, at least, they would need to be adjusted to conditions different from those set for domestic ruminants and swine. Indeed, the ID Screen kit is calibrated to detect the bovine OIEELISASPISS (Table 2), and the manufacturer positive and negative control sera are also of bovine origin (Table S2). Similar considerations apply to the in-house iELISA of Shome et al. (Shome et al., 2013) (protein G-based developed for equines), to the iELISA of the NIVEDI and possibly to the bovine IDEXX Brucellosis Serum iELISA respectively used in some studies (Tables S1 and S2). The IDEXX kit was also reported not to have a high degree of sensitivity in the sera of experimentally infected camels (Soellner et al., 2018; Wernery et al., 2022). The PrioCHECK® Brucella Ab 2.0 indirect ELISA kit uses an anti-ruminant IgG1 monoclonal antibody (MoAb) conjugate (Supplemental Table S2). Although the information is not on the respective manufacturer web pages, other iELISA brands used in camel studies have also conjugates of ruminant specificity because they are described as meeting EU requirements and WOAH standards set for bovines and/or are marketed for diagnosis in ruminants (Supplemental Table S2). Thus, the use of all these iELISAs in camels presupposes that camel IgG1 and HCAbs are efficiently detected by the anti-ruminant IgG conjugate under conditions optimized for ruminants. However, several MoAbs to bovine IgG1, IgG2 and light chain failed to react with camel sera in an iELISA study while significant reactivity was observed with MoAbs to IgM and IgA (Henning & Nielsen, 1992). Also, MoAbs to bovine IgG1, IgG2, IgM and IgA failed to react with camel IgG1 and HCAbs (Azwai et al., 1993). Although these results do not exclude cross-reactivity in epitopes not detected by those MoAbs, they show that such cross-reactivity is limited to an unknown extent. Consequently, it cannot be assumed that those iELISAs detect camel IgG1 and HCAbs as efficiently as ruminant IgG, or that the specificity is properly adjusted. An early study employed an in-house iELISA with a conjugate specific for camel immunoglobulins, but it contains misconceptions regarding the DSe/DSp of RBT and thus employs inadequately characterized sera (Azwai et al., 2001) (Table S1). Only two more recent works (Eckstein et al., 2022; Hanon & Al-Sary, 2019) used an iELISA designed to detect camel IgG. However, the commercial kit (EUROIMMUN Anti-Brucella ELISA Camel IgG) is no longer listed on the manufacturer’s website, and critical information supporting its application, such as conjugate specificity (IgG1 and/or heavy-chain antibodies) and assay standardization and validation data are not publicly available. Furthermore, the kit showed low sensitivity in camel sera in the above-mentioned experimental study (Soellner et al., 2018; Wernery et al., 2022). Some of the limitations of iELISAs are overcome by cELISAs, which do not require an anti-immunoglobulin conjugate because they measure binding of an O-PS specific MoAb-enzyme conjugate to O-PS in competition with antibodies that result from infection. Nevertheless, cELISAs differ in avidity and fine epitope specificity of the MoAb, and the host antibody avidity (functional affinity) and titers vary depending on the antigenic stimulus and time of evolution of the infection. Therefore, cELISAs risk to exclude antibodies that may be significant in diagnosis or, conversely, loose DSp if DSe is optimized. This intrinsic limitation explains why commercially available cELISAs (including those used in camels) adjusted to the maximum J (Youden) index show inadequate DSp as screening tests in cattle or swine (Muñoz et al., 2012; Praud et al., 2016) and why the claim that cELISAs can exclude S19 vaccinal and FPSR antibodies is unfounded, unless DSe is reduced to an unacceptably low level in order to increase DSp by ROC analysis (Ducrotoy et al., 2018). These characteristics of cELISAs underscore the need to tailor each assay to the specific features of the host antibody response. Among the cELISAs used in camel studies, reliable information is available for four assays (Supplemental Table S3). These assays differ in protocol (one- or two-step competition) and in the MoAbs employed and therefore require distinct optimization of the MoAb-to-serum concentration balance. This makes the use of appropriate reference standard sera critical. However, standard sera for cELISAs are of bovine origin (Table 2), and because ruminant and camel immunoglobulin characteristics differ substantially, the assay dynamic range implicit in the standardization cannot be assumed to be equivalent and should be specifically validated for camels. Moreover, assay cut-offs should also be established through ROC analysis to determine DSe and DSp. Evidence of these limitations is apparent in camel studies (Table S1). Zolzaya et al. (Zolzaya et al., 2014) investigated brucellosis seroprevalence in Bactrian camels by testing RBT-positive sera with two cELISA kits. Of 118 RBT-positive samples, 111 and 11 were classified as positive using COMPELISA VLA and SVANOVIR cELISAs, respectively, when performed according to manufacturers’ instructions. Regardless of whether RBT is considered an appropriate reference test (see above), this discrepancy is notable. Kit non-equivalence was also reported in experimental camel studies (Soellner et al., 2018; Wernery et al., 2022). The authors compared the three commercially available kits (Table S3) and, although individual results were not provided, only the Ingenasa cELISA (currently Ingezim Brucella Compact 2.0, Gold Standard Diagnostics) showed a high degree of sensitivity. Manufacturer-recommended serum dilutions may also introduce error. Wernery et al. (Wernery et al., 2021) reported that serum from a camel strongly positive by RBT, SAT, and CFT tested negative in the Ingenasa cELISA at the recommended 1:160 dilution but became positive at 1:320. Similar prozone-like effects (“hook effects”) have been described in human ELISAs and are hypothesized to result from monovalent binding, affinity-related effects, and/or steric interference from residual complement (Jacobs et al., 2015). To our knowledge, such effects have not been reported in brucellosis immunosorbent assays using ruminant or porcine sera. Whether this phenomenon occurs in camels, potentially related to their unique immunoglobulin features, warrants further investigation. To summarize, since all studies employing iELISAs or cELISAs followed the instructions of the manufacturers (Tables S1, S2 and S3) and this implies questionable standardization and no validation (WOAH, 2023a) (Section 4.10.) the validity of the large number of studies using these assays (Table S1) should be evaluated bearing in mind these limitations. Immunochromatographic tests Immunochromatographic assays are rapid tests that admit vertical and lateral formats (LFiC). Although the latter is the most common, LFICs are not uniform with regards to S-LPS quality, immunoconjugate and detection system (Table 2). To date, five different LFiC assays have been evaluated across 7 studies, all using camel sera of unknown brucellosis status (Table S1). Notably, no reference sera are available for LFiC in any animal species. Two studies using abattoir sera evaluated the brucellosis prototype LFiC of ruminant specificity described by Abdoel et al. (Abdoel et al., 2008)). These studies reported either identical number of positives by RBT and LFiC (14/210) (Eisa et al., 2014) or fewer positives by LFiC than by RBT (1 vs. 4/180, respectively) (Madu et al., 2016). Another LFiC employed a protein G-based kit (Genomix Molecular Diagnostics, India), which shares the limitations described above for protein G-based iELISAs, and for which the authors claimed %DSe/DSp 80/90 without supporting evidence (Jain et al., 2021). Two studies evaluated a LFiC based on a MoAb against camel IgG (MSA Antigen Rapid Camel Brucella Ab Test). Relative to RBT, this kit yielded slightly different results (57 and 51 positives, respectively, in 625 sera (Almuzaini et al., 2025)) or identical results (98 positives in 353 sera (Khan et al., 2016); Section 4.11.). Serhan et al. (Serhan et al., 2019) assessed a Rapid Camel Brucella Ab kit (possibly protein A-based like similar kits of BIONOTE [Korea]) in 487 sera from camels suspected of brucellosis (Section 4.11). The number of positives differed markedly between RBT and LFiC (36 vs. 59, respectively), and from the manufacturer’s claimed performance (%DSe/DSp 100/95 relative to RBT in 106 sera of camels of unknown status: https://en.bionote.co.kr/en/product/rapid/view.html?idx=41&srh_cate=6&curpage=1#list03). Finally, one study evaluated a fourth different LFiC kit marketed for cattle brucellosis (Quicking Biotech Co Ltd. [China]; https://www.quickingbio.com/bovine-rapid-test/w81085-bovine-brucella-ab-test.html?utm_source=chatgpt.com) and reported broadly concordant results with BPAT, mRBT, SAT, and Rivanol tests (Abd El Tawab et al., 2018) (Table S1). In conclusion, whereas LFiC is an attractive approach, the number of studies is insufficient, results are contradictory and none has used properly defined sera that could grant validation. Brucellosis fluorescence polarization assay (FPA) The characteristics of brucellosis FPA are summarized in Table 2. Antibody binding reduces the free rotational movement of the labeled O-PS molecule, thereby increasing light polarization. However, the reduction in free rotation caused by conventional antibodies and HCAbs is expected to differ because of differences in molecular weight, an aspect that has not yet been investigated. The test is sensitive to temperature, which may partially explain the poor performance reported under hot climatic conditions (Falzon et al., 2019). In addition to proper standardization, FPA requires validation for different animal species and epidemiological settings. However, neither standard sera nor validation studies are available for camels, a limitation reflected in the studies summarized in Table S1. Gwida et al. (Gwida et al., 2011) used an FPA (Diachemix, currently marketed by Ellie LLC [USA]) using a 20 mP cut-off that the maker claims distinguishes B. abortus infected and vaccinated cattle without independent support (https://ellielab.com/brucella-fpa/). The assay was compared with RBT, SAT, CFT, cELISA, and RT-PCR using the latent class analysis (LCA) discussed in Section 4.11, according to which the positive/negative % correlation of FPA was 99/67 in relation to the latent classes. Although this LCA is based on uncertain assumptions, FPA was the serological test with the lowest negative correlation (Section 4.11.), strongly suggesting a high proportion of false-positive results and inadequate validity of the cattle-derived cut-off in camels. Khan et al. (Khan et al., 2016) evaluated the same FPA kit using the manufacturer’s cut-off in 343 camel sera, with RBT as the relative standard, and found lower DSe/DSp values than those obtained with CFT or the MSA LFiC assay (Section 4.11). Finally, Serhan et al. (Serhan et al., 2019) applied the same FPA kit using a 95 mP cut-off reportedly derived from ROC analysis, although no methodological details were provided for this ROC analysis (control positive and negative sera numbers, animal species and criteria for defining them as positive and negative). The results were subsequently used in a Bayesian analysis that yielded high estimated DSe/DSp values (95.0/99.9), but the study raises substantial methodological concerns (Section 4.11) Overall, there is no evidence that FPA can replace CFT, as stated without supporting information in a recent camel brucellosis review (Dadar & Foster, 2025). The limited number of available studies is constrained by inadequate control definitions for estimating DSe/DSp. Moreover, these studies do not indicate that FPA offers an improvement over simpler diagnostic methods for animal brucellosis. Finally, the high cost and infrastructure requirements of FPA, including control of laboratory temperature, severely limit its feasibility in regions where camels are commonly bred. Serological test consistency The relevance of the standardization and cut-off issues discussed for each of the tests can be probed by comparing the relative results across different studies or, when individual data are reported, within the same study. Relative to the test used most often (i.e., RBT) for the sake of comparisons, the result of this comparison across different studies is as follows (see also Supplemental Table S4). Among 30 studies employing both RBT and SAT, 13 reported higher, 9 about the same and 8 less RBT-positive results compared with SAT. This implies that in different studies the DSe of SAT relative to RBT would be respectively lower, similar or higher or, conversely, that the SAT DSp would be higher, similar or lower than that of RBT. Indeed, SAT consistently shows less DSe than RBT in domestic ruminants due to its inability to detect non-agglutinating antibodies (Ducrotoy et al., 2016; WOAH, 2023a). Comparable inconsistencies were observed when RBT was compared with other assays. For CFT, 23, 5, and 9 studies reported higher, similar, or lower RBT-positive results, respectively; corresponding figures were 11, 1, and 6 for iELISA, and 7, 4, and 10 for cELISA. Such discrepancies have been never observed in well-conducted studies in domestic ruminants. For example, in cattle, RBT shows slightly higher DSe and DSp than CFT and cELISA and performance comparable to iELISA when evaluated using gold-standard sera (Ducrotoy et al., 2018). Discrepancies are also evident at the individual-animal level. Abdelnassir et al. (Abdelnassir et al., 2020) reported 124 and 68 positive sera by RBT and cELISA, respectively. Although an apparent inference would be that all cELISA-negative sera should be RBT-negative and that cELISA-positive sera should be RBT-positive, 23.6% ELISA-negative samples were RBT-positive and 13.2% of ELISA-positive samples were RB- negative. This large disagreement is totally unexpected given the known basis of these tests and their performance in cattle (Ducrotoy et al., 2016, 2018). Similar inconsistencies were obtained by Khan et al. (Khan et al., 2020). In 381 investigated sera, 60, 18, 13 and 5 were positive exclusively by CFT, c-ELISA, iELISA and RBT, respectively, (see Venn diagram in Fig. 2 of Khan et al. study). Logically, the authors concluded that (sic) “these findings call for validation and standardization of tested kits for camel brucellosis”. Lack of validation and standardization likewise provides the most plausible explanation for the above-summarized results of Abdelnassir et al. (Abdelnassir et al., 2020) and for the inconsistent and contradictory relative performance of tests across camel studies (Table S4). DSe/DSp estimations Frequentist analysis is a strategy to assess the DSe/DSp by comparison with gold standards or, alternatively, relative standards (i.e., another diagnostic test), and the former are the best choice (Jacobson, 1996). In brucellosis, valid gold standards are limited to samples from unequivocally infected animals (i.e., culture-positive; Section 3 and (Ducrotoy et al., 2018)) or from truly brucellosis-free animals. Relative standards are acceptable only if previously validated and if the error in their DSe/DSp estimates is corrected for and not carried over in the comparison (Jacobson, 1996), conditions unmet by any serological study in camels. No camel study reports DSe using sera from culture-positive animals and only Khan et al. (Khan et al., 2016) included sera from unquestionably brucellosis-free camels, albeit in a limited number. These authors examined 10 sera from brucellosis-free Australia and 343 from endemic countries (United Arab Emirates, Qatar and Sudan). Although the Australian sera were RBT-negative, suggesting of a high specificity, the sample size is largely insufficient for robust estimation (Jacobson, 1996) and, moreover, these sera were not used as a negative gold standard. Instead, DSe/DSp estimates for other tests were calculated relative to the total RBT-positive (n 98) and RBT-negative (n 255) sera. Therefore, the assessment used populations of unknown status yielding results (c.a. CFT 99.0/98.5, LFiC [MSA kit in the publication] 98.0/99.0 and FPA, 84.7/97.7) that necessarily depend on the DSe and DSp of RBT itself, which respectively remain to be determined and confirmed. Other frequentist studies used CFT or iELISA rather than RBT as the relative standard but included no control sera, and reported very low numbers of positive samples, precluding reliable estimation (8 CFT-positive and 11 RBT-positive in 1550 sera (Gessese et al., 2014), 39 iELISA-positive and 49 RBT-positive in 1245 sera (Ghanem et al., 2009)). Since neither CFT nor the iELISA had been validated for camels, their selection as reference standards has no sound basis. Finally, one study failed to specify any reference standard invalidating the DSe/DSp figures reported (Eisa et al., 2014), while another used pooled RBT results from humans and camels as a reference, a flawed approach (Rabah et al., 2020). To circumvent the need for reference standards, LCA has been applied. Gwida et al. (Gwida et al., 2011) examined 895 serum samples from two groups of camels imported in the United Arab Emirates in 2008 and 2009 using RBT (Pourquier Institute), SAT and CFT (standardized according to WOAH), the SVANOVIR Brucella-Ab cELISA and FPA (Ellie LLC [then Diachemix] USA) plus RT-PCR (Section 5). The serological test protocols and cut-offs were those indicated by manufacturers for bovine brucellosis and the raw positive results in the 895 samples were RBT 633, SAT 632, CFT 639, cELISA 616, FPA, 710 and RT-PCR 759. Since classical LCA could not be applied to the two subpopulations included because of the lack of prevalence data for the 2009 group, the authors divided the 895 sera in “not infected” and “infected”, the latter defined by a positive result in RT-PCR and/or two positive serological tests, which were regarded as presumed true status of infection setting two latent classes in the study. However, this classification relies on unvalidated tests with unknown DSe/DSp, including RT-PCR, such that defining infection by agreement among tests introduces unknown errors. Therefore, as acknowledged by the authors, this LCA cannot answer the question as to which test or group of tests produces wrong results: RBT, SAT and CFT for which the agreement was good or cELISA or FPA or RT-PCR, and it only allows estimation of correlations under the assumed latent class model (Sections 4.8 and 5.2.1.). Bayesian LCA was subsequently applied in several studies. Sherna et al. (Serhan et al., 2019) evaluated four tests in 487 sera from dromedaries suspected of brucellosis: RBT (Synbionics, France), cELISA (APHA, Table S3), FPA and an LFiC kit (Section 4.8.). The positive reactors were RBT 36, cELISA 51, FPA 50 and LFiC 59, and the %DSe/DSp (and 95% confidence limits) assessed using Markov Chain Monte Carlo with Gibbs sampling were (c.a.): RBT 68.9/99.9 (56.5–80.7/99.6–100), cELISA 94.9/99.8 (88.2–99.4/99.4–99.9), FPA 95.0/99.9 (88.2–99.5/99.6/99.6–99.9) and LFiC 98.7/99.0 (94.3–98.7/98.3–99.0). Using a similar Bayesian methodology, Khalafalla et al. (Khalafalla et al., 2020), evaluated the RBT (but from Zoetis, France), the APHA cELISA and the skin test with Brucellergene OCB in 68 camels from farms with a history of brucellosis but unknown individual status. The number of reactors were: RBT 36, cELISA 36, and skin test 29, and %DSe/DSp (and 95% confidence limits) resulting from the Bayesian analysis: RBT 93.3/97.8 (83.4–93.3/92.4–97.8), cELISA 94.8/98.5 (84.8–94.8–93.9–98.5) and skin test 70.7/98.8 (56.2–70.7–94.9–98.8). In both studies, cut-offs were those proposed for cattle for APHA cELISA (Section 4.7.) and of unknown origin for FPA (95 mP, Section 4.9.). The raw results and DSe/DSp estimates of RBT and cELISA in each of these 2 studies are remarkably different. Moreover, the low DSe of RBT obtained by Sherhan et al. (Serhan et al., 2019) is difficult to explain, as noted by the authors. They attribute the RBT results to supposed limitations of the RBT (sic. “…false unspecific agglutination reactions, low sensitivity particularly in chronic cases, and relatively low specificity of the RBT”). However, such limitations are disproved by the experience in domestic ruminants (Ducrotoy et al., 2018) and the two references provided by the authors are misrepresented: one does not show unspecificity or low sensitivity of RBT in camels (Khan et al., 2016) and the other is a study in humans, not in camels (Díaz et al., 2011). Moreover, the latter shows the excellent specificity and sensitivity of RBT in acute and long evolution (“chronic”) cases. Therefore, the discrepancy between the results of Sherhan et al. and Khalafalla et al. point out to issues in test standardization and/or validation (Sections 4.5., 4.7. and 4.10) and/or in the Bayesian modeling assumptions. Concerning the latter, the descriptions in these works are succinct, and it is not possible to evaluate the underlying assumptions, that can be highly variable and are largely unreliable in Bayesian analyses of brucellosis serological tests (Ducrotoy et al., 2018). Other Bayesian analyses raise similar concerns. Elsohaby et al. (Elsohaby et al., 2022) investigated 921 sera using BPAT, RBT and iELISA (IDEXX, Table S2) and obtained, respectively, 84, 73 and 73 positive results but the two subsequent analyses are invalid. The authors carried out a first analysis using vague priors for BPAT and for RBT and informative DSe/DSp priors for iELISA taken from Gwida et al. (Gwida et al., 2011) (see above). However, Gwida et al. used a cELISA, not an iELISA, and reported data calculated on the above-discussed unprecise definition of “uninfected” and “infected animals” that, as underlined above, are not true DSe or DSp estimates. Hence, the iELISA informative priors (supposed DSe/DSp, and derived percentiles and distribution) applied are invalid interpretations of the original work. Elsohaby et al. (Elsohaby et al., 2022) also used a second approach with only vague priors for the three tests. However, an assumption (in both analyses) was the conditional dependence of BPAT and RBT and the conditional independence of the iELISA (Elsohaby et al., 2022). To support this choice, the authors defined BPAT and RBT as “antigen tests” and the iELISA as an “antibody test”, when in fact these three tests measure antibodies (Ducrotoy et al., 2016) and are thus likely to be conditional dependent. In Bayesian analyses of brucellosis serological tests, conditional independence, which cannot be relaxed because it requires additional and not always verifiable assumptions, is difficult to establish: the antigen and the analytes are the same in all tests and, even though the IgM and IgG subclasses detected vary depending on the test, these immunoglobulins both overlap and their relative proportion shifts over the course of infection (Ducrotoy et al., 2016). Therefore, the same tests can be conditionally dependent or independent depending upon the time lapse after infection and the antigenic stimulus. Accordingly, these parameters are highly heterogeneous at individual level in naturally infected populations, thus preventing the application of sound criteria in Bayesian analysis in brucellosis serology (Ducrotoy et al., 2018). Finally, Meletis et al. (Meletis et al., 2023) conducted a Bayesian analysis of RBT and SAT in camels based on only 2 RBT-positive and no SAT-positive sera (n = 31). The DSe/DSp estimates and 95% probability intervals (respectively 65–98 and 97–99 for RBT and 75–99 and 97–99 for SAT) derived from such sparse data are obviously invalid. Moreover, the authors employed informative priors taken from posterior estimates obtained from another Bayesian analysis of brucellosis serological tests conducted in sheep and goats (Anisur Rahman et al., 2013). These priors are not transferable to camels and are further affected by the above-outlined methodological problems of Bayesian analysis of brucellosis serological tests. Confirmatory strategies A total of 77 studies reported the use of a “confirmatory” test following an RBT (or mRBT)-positive result (Table S1), reflecting a widespread misconception regarding the appropriate use of brucellosis serological tests. It is true that in ruminant eradication programs that combine S19 or Rev.1 vaccination with test-and-slaughter, culling of vaccinated seropositive animals is minimized by applying the simpler RBT as a screening test followed by the technically demanding CFT on RBT-positive animals (Blasco et al., 2021). However, it should not be overlooked that in such programs repeated testing is essential, and that CFT (a quantitative assay) meets two key criteria: (i) CFT titers decline more rapidly than RBT-detected antibodies in vaccinated animals that become protected; and (ii) under repeated testing, persistent or increasing titers are indicative of infection. Used in this way, CFT is a complementary (not a “confirmatory”) test that allows to assess the persistence of infected animals in a vaccinated herd or flock. However, in the absence of vaccination, CFT exhibits slightly lower DSe and no higher DSp than RBT, such that confirming RBT-positive results with CFT in unvaccinated animals yields false-negative results (Ducrotoy et al., 2016, 2018). Unfortunately, misinterpretation of the RBT–CFT strategy has led to the erroneous assumption that RBT is highly sensitive but poorly specific in unvaccinated animals (Ducrotoy et al., 2016, 2018). In camel studies, this misconception is further compounded by the wide range of supposed “confirmatory” tests. Among the 74 studies that employed such “confirmatory” tests after RBT, 32 used CFT, 15 iELISA, 13 SAT, 11 cELISA, 5 2-ME, 4 Rivanol, 1 LFiC, and 1 the USDA plate agglutination test, either individually or in combination. This heterogeneity bears no parallel with the RBT–CFT serial testing strategy applied in vaccinated cattle and sheep and is incompatible with numerous camel studies reporting fewer RBT reactors than positive results obtained with the purported “confirmatory” tests (Tables S1 and S4). Milk antibody tests in camels The MRT and iELISAs are assays used in ruminants to detect anti-S-LPS antibodies in milk (Ducrotoy et al., 2016). Fifteen studies (Tables 1, 3, and 4; Table S1) employed the MRT, using either raw camel milk (14 studies) or camel milk supplemented with MRT-negative bovine milk (see below). However, the application of MRT to camel relies on doubtful assumptions. In cow milk, IgA and IgM and to a lesser extent IgG adsorb to fat globules via their Fc regions promoting clumping of the stained S brucellae that then accumulate in the cream layer to form a characteristic ring (Ducrotoy et al., 2016). Although camel milk also contains IgA and IgG, an important fraction of both is made of HCAbs, a major difference with respect to ruminants. In addition, MRT performance is also related to physicochemical properties specific of bovine milk, as shown by the lack of usefulness in sheep and goat milk (Alton et al., 1988; WOAH, 2023a). While camel has a fat globule size distribution like that of cow milk, it exhibits poor creaming capacity, attributed to the absence of a fat agglutinin (Farah & Rüegg, 1991). This poor creaming likely explains the failure of MRT ring formation in camel milk artificially supplemented with homologous immunoglobulins (Bastawrows et al., 2000) and the findings of Van Straten et al. (Van Straten et al., 1997), who reported MRT-negative results in milk from three seropositive camels unless bovine milk was added. Similar observations were made by Musa and Shigidi (Musa & Shigidi, 2001) and implemented in one subsequent study (Omer et al., 2010), but systematic studies are lacking. All these differences strongly question the usefulness of the MRT in camel milk. The uncertainties are not dissipated by the only wok reporting DSe/DSp estimates. Almashhadany et al. (Almashhadany et al., 2023) investigated 250 milk samples from animals of unknown serological status, reporting 21 culture-positive, 29 MRT-positive, and 26 iELISA-positive with estimated %DSe/DSp of 80.8/96.6 for MRT and 80.7/99.1 for iELISA relative to culture results. However, selective culture media were not used, a precaution essential in milk (WOAH, 2023a). Moreover, milk processing and protocols for identification and typing are not described, making the reported data unreliable and thus unsuitable for DSe/DSp estimates (Table 1). In addition, the authors calculated the DSp with respect to the number of culture negative samples, which is not a valid negative gold standard control (Section 3). Several studies applying iELISA in camel milk relied on in-house protocols using uncharacterized antigens and lacked control sera (Almashhadany et al., 2023; El-Naggar et al., 2006; Van Straten et al., 1997); consequently, their findings are preliminary and of limited, if any, interpretative value. Commercial iELISAs have been used in only 3 studies. The bovine milk BRUCELISA M kit (APHA Scientific UK) yielded negative results in 75 milk samples (Tanimoun et al., 2021), and the PrioCHECK® Brucella Ab 2.0 indirect ELISA kit was negative when applied to 3 individual camel milk plasmas and 11 milk pools (Wainaina et al., 2020). However, the absence of appropriate controls precludes confirmation that these samples were truly negative, and these kits share the conjugate specificity and validation limitations discussed above for serum iELISAs. Finally, Khan et al. (Khan et al., 2018) reported 12 positive results among 18 camel milk samples tested using a Sigma-Aldrich iELISA kit (for which no validation information is currently available; Table S2), with milk diluted 1:10 in phosphate buffer. However, the authors fail to justify the dilution factor used, the iELISA has the intrinsic limitations in camel samples as those of other iELISAs and the study used flawed controls (RBT in milk diluted 10 times in phosphate buffer, and culture [Table 1]). Overall, the validity of iELISAs for detecting anti-Brucella antibodies in camel milk remains to be investigate based on well-characterized samples and validated assays. PCR studies in camels In camel brucellosis, both conventional PCR and RT-PCR have been used in DNA extracted from a variety of samples to complement and/or compare the results with other diagnostic methods. Conventional PCR studies These studies are summarized in Table 3, where can be seen that they are methodologically heterogeneous. Sample types included milk, milk cream or centrifugation pellets (but not both, as recommended for culture), blood, blood clots, serum, diverse tissues, hygroma fluid, vaginal swabs, and seminal plasma. However, only 9 of the 30 studies reported the sample preparation protocol and the amount of material processed (Alamian & Dadar, 2019; Bahari et al., 2021; Getachew et al., 2025; Hamdy & Amin, 2002; Hazem et al., 2023; Hazem & Ismail, 2014; Ibrahim et al., 2012; Khan et al., 2018; Saxena et al., 2021), two critical issues as they affect reproducibility and sensitivity. DNA extraction methods varied widely. Twenty-one studies used commercial extraction kits from at least thirteen different manufacturers; 2 studies employed an in-house detergent–proteinase K-phenol protocol (Leal-Klevezas et al., 1995), 2 studies applied protocols designed for DNA extraction from isolated Brucella colonies (Abdalla et al., 2021; Ibrahim et al., 2012) originally described by Unver et al. and Wilson et al. (Unver et al., 2006; Wilson et al., 1990), 1 study applied protocols developed for Listeria monocytogenes grown in culture (Dehkordi et al., 2012) and 1 study did not describe the extraction method (Hosein et al., 2016). Primers were similarly diverse. Three studies used commercial PCR kits (SACACE (Al-Busadah et al., 2017; Khalafalla et al., 2017) and AccuPower Brucella PCR Kit (Hanon & Al-Sary, 2019)) despite manufacturer specifications indicating their intended use for identification of bacterial isolates rather than direct diagnosis from clinical samples (https://sacace.com/?id=0&sid=1&q=brucella+kit; https://eng.bioneer.com/20-k-2906.html; https://eng.bioneer.com/20-k-2906.html?utm_source=chatgpt.com). Six studies applied AMOS-PCR, 2 in combination with Bruce-ladder; 1 used Bruce-ladder alone; and 2 targeted the bp26 gene directly from extracted samples, although these assays were originally developed for species identification and typing of Brucella in pure cultures (Bricker & Halling, 1994; García-Yoldi et al., 2006; López-Goñi et al., 2008). Other studies employed genus-specific primers targeting the bcsp31 (31-kDa protein) gene (10 studies), the 16S rRNA gene (1 study), sequence RM57 (1 study), or IS711 in combination with primers from a conserved region of B. abortus NC_007624.1 (Table 3). Noteworthy, 3 studies reported simultaneous detection B. abortus and B. melitensis in 1 (Rabah et al., 2020), 3 (Getachew et al., 2025) and 4 (Khamesipour et al., 2014) camels, a finding of questionable validity (see below). Table 3Conventional PCR in dromedaries1,2.Table 3 dummy alt textReference [Country]Samples (no.) and protocolControlsResults3Comments3(Hamdy & Amin, 2002) [Egypt]Milk from camels (12) from farms (52 cows, 21 sheep, 18 goats, 12 camels), from farms with a history of brucellosis.DNA extraction: 400 μL of the cream of centrifuged (20 mL) milk; Triton X100-SDS-proteinase K/phenol (Leal-Klevezas et al., 1995).Primers and detection: AMOS-PCR (Bricker & Halling, 1994); PAGE.DSe controls.Culture: cream and sediment on Farrell’s medium.Serology: positive in at least one test (RBT, SAT or MRT).DSp controls: dairy cows (50) from Brucella-free herd.Culture: all -.Serology +: 1 (SAT+, RBT +, MRT +).PCR +: 1 (RBT+ SAT+ MRT +).DSe/DSp unknow. No DSe controls: Farrell inhibitory for some B. abortus and many B. melitensis strains and all samples culture negative. PCR conducted only in camels positive in at least one test; non-camel-validated serological tests are not valid relative DSe/DSp standards; other DSp controls in the study are not valid (milk from seronegative cows rather than from brucellosis-free camels).(Alshaikh et al., 2007) [Saudi Arabia]Serum from seropositive camels (26).DNA extraction: centrifuged serum (amount not described) digested with proteinase K, ethanol precipitated, dissolved in Tris-EDTA, and DNA extracted with Tissue extraction kit (QiAgen, Germany).Primers and detection: three Brucella genus (B4/B5 from BCSP31 [31Kd protein]); F4/R2 of 16S rRNA (Romero et al., 1995) and IS711 (Bricker & Halling, 1994); B. abortus and B. melitensis species-specific of AMOS-PCR (Bricker & Halling, 1994); PAGE.DSe controls.Culture: n.d.Serology: positive in at least two tests (RBT, SAT, cELISA, CFT).DSp controls: n.d.PCR+:B4/B5 26.F4/R2 18.AMOS-PCR B. abortus 8 (no B. melitensis).DSe/DSp unknown. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls: PCR conducted only in camels positive in at least two tests (RBT, SAT, cELISA, CFT); no samples from brucellosis-free camels.(Dehkordi et al., 2012) [Iran]Abomasal contents of aborted fetuses (618)DNA extraction: Amount not described; method developed for Listeria monocytogenes in pure culture; DNA purified using Fermentas kit (GmbH, [Thermo Fisher Scientific] Germany).Primers and methodology: primers targeting omp2 as in (Leal-Klevezas et al., 1995).No DSe or DSp controls.PCR + 201.Non-validated DNA extraction methodology.DSe/DSp unknown: no controls of any kind.(Ibrahim et al., 2012)Milk from RBT+ camels (64) and seronegative cows (50).DNA extraction: milk (500 μl) mixed with (100) NaCl- EDTA-Tris–HCl (pH 7.6), at 80 °C for 15 min, digested with SDS-proteinase K (50 °C, 3 h), NaCl-CTBA precipitated, extracted according to Wilson et al. (1990); pellet resuspended in 25 μl of sterile distilled water and 1 μg added to the PCR mixture.Primers and detection: P1 6633 TCG GTT GCC AAT and P2 6634 CTT GCC TTT CAG GTC from BCSP31 (31Kd protein); PAGE.DSe controls.Culture: Table 1.Serology: RBT and iELISA in serum, MRT in milk.DSp controls: 50 milk samples from serologically negative cows.Culture 19/64.Serology: MRT, 22/64; iELISA 22/64.PCR + 18/64DSp unknown. No DSp controls: PCR in milk from seronegative cows, rather than milk from brucellosis-free camels.(Ghorbani et al., 2013) [Iran]Blood from apparently healthy camels (310).DNA extraction: blood (amount not described); ZistFannavari (Iran) DNA Purification kit modified by authors for blood samples.Primers and detection: genus-specific nested-PCR with primers from 16S rRNA described by the authors; PAGE.DSe controls.Culture: (Table 1)Serology: mRBT, RBT, SAT and 2-ME.DSp controls: n.d.Culture: all -.Serology: mRBT+ 39, RBT+ 27; SAT, 7, 2-ME+ 6.PCR+: 6/39 seropositive; 3/(no.?) seronegative.DSe/DSp unknown. No DSe/DSp controls: all samples culture negative; non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls (study reports PCR negatives in 61 controls not described in Mat. Meth. and PCR positives in seronegative samples).(Shome et al., 2013) [India]Sera (78) from camels with a history of abortion.DNA extraction: serum (amount not described); DNeasy blood and tissue kit protocol (QiAgen, USA).Primers and detection: B4/B5 from BCSP31 (31Kd protein) gene (Baily et al., 1992); PAGE.DSe controls.Culture: n.d.Serology: RBT, SAT and iELISA.DSp controls: n.d.Serology:RBT+ 7, SAT 2, iELISA 2.PCR+ 1/(only 42 tested).DSe/DSp unknown. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls (no samples from brucellosis-free camels).(Hazem & Ismail, 2014) [Egypt]MRT+ milk (18) from camels w/wo. reproductive disorders (156).DNA extraction: 400 μL of the cream of centrifuged (20 mL) milk; Triton X100-SDS-proteinase K/phenol (Leal-Klevezas et al., 1995).Primers and detection: IS711 specific primers (Leal-Klevezas et al., 1995); PAGE.DSe controls.Culture: (Table 1).Serology: serum BPAT, Card Test, SAT, 2-ME (1:25), Rivanol; milk MRT.DSp controls: n.d.Culture: 5/18 MRT+.Serology (156 sera): BPAT+11, SAT+ 9, 2-ME 7+, Rivanol+ 7; MRT+ 18.PCR+ 6/18 MRT+DSe/DSp unknown. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls: PCR conducted only in MRT (not valid in camels; see text) + milk (no samples from brucellosis-free camels).(Khamesipour et al., 2014) [Iran]Blood (123) and lymph node samples (123) from camels (123).DNA extraction: amount of sample not described; CinnaGen DNA extraction kit™ (Cinnagen, Tehran, Iran.Primers and detection:Brucella genus primers as in (Khamesipour et al., 2013); B. abortus and B. melitensis specific primers as in (Mirnejad et al., 2012).DSe controls.Culture: n.d.Serology: n.d.DSp controls: n.d.PCR + in 123 blood and 123 lymph nodes: Brucella 14 and16; B. abortus 5 and 4; B. melitensis 3 and 2; both spp. in the same sample 2 and 2; undetermined (“other”) Brucella spp. 6 and 10.Brucella genus primers not validated for ASp.DSe/DSp unknown. No DSe or DSp controls of any kind.Purported co-infections by two Brucella species in 4 camels. Detection of “other Brucella” not valid since species-specific primers were used.(Ullah et al., 2015) [Pakistan]Camels (387) collected randomly that were RBT+.DNA extraction: serum (amount not described); QIAmp DNA Mini Kit (Qiagen).Primers and detection: B4/B5 from BCSP31 (31Kd protein) gene (Baily et al., 1992); PAGE.DSe controls.Culture: n.d.Serology: RBT.DSp controls: n.d.Serology:RBT+ 2.PCR+ 0/2 RBT+.DSe/DSp unknown. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls: PCR conducted only in seropositive camels (no samples from brucellosis-free camels).(Hosein et al., 2016) [Egypt]Culture + milk (2) from apparently healthy camels (1126).DNA extraction: Not described.Primers and detection: genus specific bp26 primers (García-Yoldi et al., 2006).DSe controls.Culture: (Table 1).Serology: mRBT, cELISA.DSp controls: n.d.Culture +: 2 B melitensis.Serology (1126 sera): mRBT+ 47, cELISA+ 42.PCR+ 2/2 culture+ milk.Incomplete protocol description.DSe/DSp unknown. Insufficient no of culture + DSe controls. No DSp controls: PCR performed only in culture-positive milk (no samples from brucellosis-free camels).(Al-Busadah et al., 2017) [Saudi Arabia]Semen from fertile (40) and infertile (94) dromedaries.DNA Extraction: seminal plasma (amount not described); DNA/RNA Extraction Kit (Vivantis Technologies).Primers and detection: PCR kit (SACACE). PAGEDSe controls.Culture: n.d.Serology: n.d.DSp controls: n.d.PCR+: Brucella in 2/44 infertile camels; (fertile, n.d.).Unknown DSe/DSp. Non valid DSe/DSp controls: fertility/infertility cannot be considered as controls (infertility has multiple causes).(Khalafalla et al., 2017) [Saudi Arabia]Uterine swabs (182).DNA extraction: uterine swabs (processing not described); DNA/RNA extraction kit (Vivantis Technologies).Primers and detection: PCR kits (SACACE); PAGE.DSe controls.Culture: n.d.Serology: iELISA.DSp controls: n.d.Serology: iELISA+ 15.PCR+ 27.DSe/DSp unknow. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls (no samples from brucellosis-free camels).(Shahzad et al., 2017) [Pakistan]RBPT+ cELISA + from a serosurvey (761).DNA extraction: serum (amount not described); DNA extraction kit (Favorgen, FABGK 001).Primers and detection: B4/B5 from BCSP31 (31Kd protein) gene (Baily et al., 1992); PAGE.DSe controls.Culture: n.d.Serology: RBT; cELISA.DSp controls: n.d.Serology (761 sera): RBT+ 26; (RBT - cELISA+: 18.PCR+ 2/18 (RBT+, cELISA+).DSe/DSp unknow. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls: PCR performed only in seropositive sera (no samples from brucellosis-free camels).(Khan et al., 2018) [Egypt]Milk (80).DNA extraction: pellet of 10 mL of milk treated with lysozyme, proteinase K-SDS (2 h, 55 °C) and phenol-chloroform-isoamyl alcohol conventional extraction as in (Romero & López-Goñi, 1999).Primers and detection: 16rRNA F4/R2; Omp2 JPF/JPR, IS711 B1/B2; PAGE.DSe controls.Culture: (Table 1)Milk: RBT; serum iELISA in samples arbitrarily diluted 1:10 in water.DSp controls: n.d.Culture +: 9.PCR (16 rRNA and/or IS711 primers) + 9/9 culture+.Non-dependable results. Identification methodology inappropriate (Table 1); RBT is not a milk test; serum iELISA not validated for milk and reported by authors to detect IgM in milk; no DSp controls (relative standards or samples from brucellosis-free camels).(Alamian & Dadar, 2019) [Iran]Milk (96) from apparently healthy, clinically asymptomatic dromedaries.DNA extraction: pellet from centrifuged milk (1.5 ml); High Pure PCR template Kit (Roche).Primers and detection: AMOS-PCR (Bricker & Halling, 1994) and Bruce-ladder (López-Goñi et al., 2011); PAGE.DSe controls.Culture: (Table 1).Serology: n.d.DSp controls: n.d.Culture: 3 B abortus strains.PCR Bruce-ladder +: 14 B abortus.DSe/DSp unknown. Incomplete DSe controls: Farrell’s inhibitory for some B. abortus and many B. melitensis strains. No DSp controls (relative standards or samples from brucellosis-free camels).(Hanon & Al-Sary, 2019) [Iraq]Blood (237; 209 female and 28 male camels).DNA extraction: blood (amount not described); DNA extraction kit KAPA Express Extract KK7100Primers and detection: PCR PreMixAccuPower® Bioneer and PAGE.DSe controls.Culture: n.d.Serology: RBT, Camel iELISA.DSp controls: n.d.Serology: RBT+ 51, camel iELISA+ 39.PCR+: 7 B abortus; 5 B melitensis.DSe/DSp unknow. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls (no samples from brucellosis-free camels).(Dadar & Alamian, 2020) [Iran]Blood (2854) from apparently healthy dromedaries.DNA extraction: lymph nodes (sample processing not described); Exgene Cell SV kit (Gene All).Primers and detection: AMOS-PCR (Bricker & Halling, 1994) and Bruce-ladder (López-Goñi et al., 2011); PAGE.DSe controls.Culture: (Table 1).Serology: RBT, SAT, and 2-ME (≥ 1/80)DSp controls: n.d.RBT+10, PCR+ 7/10 (all culture -)RBT- 30, PCR+ 2 (both culture +).DSe/DSp unknow. No DSe/DSp controls: serology, culture and PCR do not correlate; non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls (no samples from brucellosis-free camels).(Elshazly et al., 2020) ElshazlyGenitalia of adult nonpregnant she-camels 500).DNA extraction: genital organs (sample processing not described); DNeasy Blood &Tissue Kit (Qiagen).Primers and detection: B4/B5 from BCSP31 (31Kd protein) gene (Baily et al., 1992); PAGE.DSe controls.Culture: n.d.Serology: n.d.DSp controls: n.d.No PCR + samplesNon-dependable results. No PCR + sample and no DSe or DSp controls (culture, relative standards or samples from brucellosis-free camels).(Rabah et al., 2020) [Egypt]Randomly selected camels (100) that were RBT+.DNA extraction: blood (amount not described) and QIAampe DNA mini kit or Roche PCR Template Preparation kit.Primers and detection: B4/B5 from BCSP31 (31Kd protein) gene (Baily et al., 1992); primers targeting BMEI1162 (B. melitensis) (Mutnal et al., 2007) and alkB (B. abortus) (Song KyungYoung et al., 2019); PAGEDSe controls.Culture: n.d.Serology: RBT, BPAT, CFT.DSp controls: n.d.Serology: 10 RBT+ 10, BPAT+ 10, CFT 9.PCR+: 9 (7 B abortus,1 B melitensis, 1 both B. abortus and B. melitensis)/10 RBT+DSe/DSp unknown. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls: PCR conducted only in seropositive camels and samples from brucellosis-free camels not tested. Purported co-infections by two Brucella species in 1 camel.(Shabbir et al., 2020) [Pakistan]Pools of 4–5 sera from 74 herds.DNA extraction: sera (amount not described) DNA (FavorPrep Genomic DNA Extraction Kit, Favorgen).Primers and detection: B4/B5 from BCSP31 (31Kd protein) gene (Baily et al., 1992) plus spp. identification using AMOS-PCR (Bricker & Halling, 1994); PAGE.PCR performed in pools of sera made irrespective of the positive or negative result in an iELISA.Herd prevalence.iELISA + 31/74.RT-PCR 19/74 (B. abortus).DSe/DSp unknown. PCR done with pools of sera and dilution factor effect not tested in ASe experiments. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls (no samples from brucellosis-free camels).(Abdalla et al., 2021) [Sudan]Whole blood (n.d.).DNA extraction: method designed for Brucella in pure culture (Unver et al., 2006).Primers and detection: IS711 primers for B. melitensis and B. abortus as in (Hinic et al., 2008).DSe controls.Culture: n.d.Serology: RBT, SAT.DSp controls: n.d.RBT+19/500SAT 17/500No PCR+ (n.d.)DNA extraction methodology inappropriate. No of samples tested by PCR unclear (only seropositive?) but all negative.DSe/DSp unknown. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls (no samples from brucellosis-free camels).(Bahari et al., 2021) [Iran]Blood, brain, liver and portal lymph nodes from apparently healthy camels (100).DNA extraction: 300 μl blood (WizPrep™ DNA Extraction Kit) and 10–20 mg of tissues (WizPrep™ gDNA Mini Kit).Primers and detection: B4/B5 from BCSP31 (31Kd protein) gene (Baily et al., 1992) plus identification using an abortus-melitensis-canis-suis multiplex PCR (Imaoka et al., 2007) and sequencing of PCR products.DSe controls.Culture: n.d.Serology: n.d.DSp controls: n.d.PCR+: B. abortus in blood (2 camels) and in liver and lymph node (1 camel).Unknown DSe/DSp. No DSe or DSp controls (culture, relative standards or samples from brucellosis-free camels).(Saxena et al., 2021) [India]Whole blood of RBT+ (15) camels (177).DNA extraction: 100 μl blood (Qaigen Genomic DNA isolation kit).Primers and detection: Bruce-ladder (López-Goñi et al., 2011); PAGE.DSe controls.Serology: RBT.DSp controls: n.d.RBT+: 15.PCR+: 0/15 (and 0/10 RBT- tested).DSe/DSp unknown. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls: PCR conducted only in seropositive camels (no samples from brucellosis-free camels).(Hazem et al., 2023) [Egypt]Milk from dairy camels (270) with history of reproductive disorders, including abortion.DNA extraction: 400 μL of the cream of centrifuged milk (20 mL); Triton X100-SDS-proteinase K/phenol (Leal-Klevezas et al., 1995).Primers and detection: AMOS-PCR (Bricker & Halling, 1994); PAGE.DSe controls.Culture: (Table 1).Serology: mRBT, CFT.DSp controls: n.d.Culture+: B. melitensis (2); B. abortus (1).Serology: RBT+ 40, CFT+ 32.PCR+: 22 (no spp. data)Methodology unclear: no spp. identification by AMOS-PCR.DSe/DSp unknown. No appropriate DSe controls: Farrell inhibitory for some B. abortus and many B. melitensis strains; non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls (no samples from brucellosis-free camels).(Meena et al., 2023) [India]Samples (200) collected randomly from farmed camels.DNA extraction: blood clots; DNASure Blood Mini Kit (Genetix).Primers and detection: genus-specific from IS711 adopted from Amoupour et al. (2019) and from a conserved region of B. abortus NC_007624.1 (this work). PAGEDSe controls.Culture: n.d.Serology: RBT, iELISADSp controls: n.d.Serology (in 200 sera):RBT+ 15; iELISA+ 18; RBT+ iELISA+ 11/200.PCR+ (in 200 blood clots) 5.DSe/DSp unknown. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls (no samples from brucellosis-free camels).(Selmi et al., 2024) [Tunisia]Blood from iELISA+ (10) apparently healthy camels (470)DNA extraction: blood (amount not described) Wizard® Genomic DNA purification (Promega).Primers and detection: B4/B5 from BCSP31 (31Kd protein) gene (Baily et al., 1992); PAGE.DSe controls.Culture: n.d.Serology: iELISA.DSp controls: n.d.Serology: iELISA+ 10.PCR +: 0/10.DSe/DSp unknown. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls: PCR conducted only in seropositive camels (no samples from brucellosis-free camels).(Xingxing et al., 2024) [China]Blood from RBT+ (14) camels (320.DNA extraction: blood (amount not described); blood DNA extraction kits (not specified).Primers and detection: Brucella RM57 primers described by the authors; PAGEDSe controls.Culture: n.d.Serology: RBTDSp controls: n.d.Serology:RBT+ 14.PCR+:13/14 RBT+DSe/DSp unknown. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls: PCR conducted only in seropositive camels (no samples from brucellosis-free camels).(Almuzaini et al., 2025) [Saudi Arabia]Supramammary lymph nodes and uterine tissues from seropositive camels (10).DNA extraction: sample processing not described; gSync™ DNA Extraction Kit (Geneaid).Primers and detection: genus specific primers (García-Yoldi et al., 2006); PAGE.DSe controls.Culture: n.d.Serology: RBT, CFT, iELISA, LFiC.DSp controls: n.d.Serology:PCR+: 4/10 seropositives (test n.d.).Unclear report: it does not describe culture but shows Bruceladder of an isolate; it does not report which samples were PCR positive and if the camels were seropositive. DSe/DSp unknown. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls: PCR conducted only in seropositive camels (no samples from brucellosis-free camels).(Getachew et al., 2025) [Ethiopia]Blood from camels (160) that were RBT+ and iELISA+ (7).DNA extraction: blood (200 µL) clots (QIAGEN DNA extraction kit).Primers and detection: AMOS-PCR (Bricker & Halling, 1994); PAGE.DSe controls.Culture: n.d.Serology: RBT, iELISA.DSp controls: n.d.Serology: 7 RBT+, iELISA +.AMOS-PCR+: 5 simultaneously + for B. abortus and B. melitensis (see also Table 4)/7 seropositivesDSe/DSp unknown. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls: PCR conducted only in seropositive camels and samples from brucellosis-free camels not tested. Purported co-infections by two Brucella species in 3 camels.(Hussein et al., 2025) [Egypt]Milk samples (6) from RBT +, iELISA + camels.DNA extraction: sample processing and amount extracted not described; DNeasy Blood and Tissue Kit (Qiagen).Primers and detection: AMOS-PCR (Bricker & Halling, 1994); PAGE.DSe controls.Culture: n.d.Serology: RBT, iELISA.DSp controls: n.d.PCR + 5/6 milk samples (2 B abortus; 3 B melitensis).DSe/DSp unknown. Non-camel-validated serological tests are not valid DSe/DSp relative standards. No other DSp controls: PCR conducted only in seropositive camels (no samples from brucellosis-free camels).1Abbreviations in this column: DSe, diagnostic sensitivity; DSe, diagnostic specificity; n.d.: not done.2Unless otherwise indicated the iELISA kits were intended for ruminants or for ruminants and pigs. For a discussion of serological tests in these works, see Section 3.3. Brucellosis serological tests used in camels and Table S1.3Positive or negative results are indicated with the signs + or -, respectively. Conventional PCR in dromedaries1,2. Abbreviations in this column: DSe, diagnostic sensitivity; DSe, diagnostic specificity; n.d.: not done. Unless otherwise indicated the iELISA kits were intended for ruminants or for ruminants and pigs. For a discussion of serological tests in these works, see Section 3.3. Brucellosis serological tests used in camels and Table S1. Positive or negative results are indicated with the signs + or -, respectively. PCR protocols developed for identification or typing are typically optimized to amplify purified DNA from bacterial isolates rather than from animal samples, in which host DNA, inhibitory substances, and tissue structure may interfere with DNA extraction and amplification. Consequently, assessment of analytical sensitivity (ASe) using samples from brucellosis-free animals spiked with viable bacteria is informative. However, this approach was applied in only the 2 studies discussed below. In a study investigating milk from camels and other livestock, ASe was evaluated for AMOS-PCR using cow milk rather than camel milk (Hamdy & Amin, 2002) (Table 3). The reported ASe was 100 CFU/mL for PCR and 100–1000 CFU/mL for the milk culture control. The second study determined ASe in camel milk spiked with B. melitensis using primers targeting 16S rRNA, omp2 or IS711 (Khan et al., 2018) (Table 1, Table 3). Although the authors reported an ASe of 16 CFU/mL for 16S rRNA and IS711 assays, direct examination of the figure published indicate a more realistic value of 160 CFU/mL, and the validity of the study is compromised by deficiencies in bacteriological methodology (Table 1). The ASe values reported in these 2 studies are markedly higher than the sensitivity achievable by well-performed bacteriological methods in milk from naturally infected cows (1 CFU/mL) (Muñoz et al., 2022b). Regarding analytical specificity (ASp), all studies listed in Table 3 assumed sufficient specificity to avoid cross-identification with phylogenetically related bacteria. While AMOS-PCR does not amplify DNA from Agrobacterium spp., Rhizobium spp., Ochrobactrum anthropi, undetermined Ochrobactrum spp., or Rhodospirillum rubrum ((Bricker & Halling, 1994), the bcsp31 gene primers B4/B5 used in 9 studies (Table 3) amplify homologous sequences in certain Ochrobactrum strains. (Casañas et al., 2001). Cross-detection was not examined for other, less commonly used primers (Khamesipour et al., 2014; Mirnejad et al., 2012). In animal brucellosis, DSe should ideally be assessed using bacteriological culture as the gold standard or, alternatively, a validated reference test with known DSe and DSp, such as a properly validated serological assay (Section 4.11) (Ducrotoy et al., 2018; Jacobson, 1998). Notably, only 3 of the 29 studies listed in Table 3 performed bacteriological culture. Evaluation of DSp requires samples from animals originating from brucellosis-free areas or, at minimum, from confirmed brucellosis-free herds. Samples from merely seronegative animals are inadequate, particularly when obtained from the same epidemiological unit as positive animals. Likewise, negative bacteriological results are not informative for DSp estimation (3, 4.11). When these criteria are applied, none of the studies in Table 3 provides data useful to estimate the DSe or, even more importantly, the DSp of the PCR assays used. Five studies did not include DSe/DSp controls (Al-Busadah et al., 2017; Bahari et al., 2021; Elshazly et al., 2020; Khamesipour et al., 2014; Shabbir et al., 2020). In addition, the lack of DSp negative controls in 21 studies and the use of non-validated serological tests preclude comparison between the diagnostic performance of the latter and the PCR protocol applied. The use of pools of sera from animals of unknown status (Shabbir et al., 2020) further prevents drawing conclusions, as the dilution effect on potentially PCR positive samples is undefined. Also, the results reported are often contradictory. For example, In 10 studies, the number of PCR-positive samples was lower than the number of RBT-positive samples (Almuzaini et al., 2025; Getachew et al., 2025; Ghorbani et al., 2013; Hanon & Al-Sary, 2019; Hazem et al., 2023; Hazem & Ismail, 2014; Saxena et al., 2021; Shahzad et al., 2017; Shome et al., 2013; Ullah et al., 2015), whereas 2 studies reported similar numbers of PCR- and RBT-positive results (Rabah et al., 2020; Xingxing et al., 2024) (Table 3). Moreover, one study detected 3 AMOS-PCR-negative results in the milk of 10 RBT-positive camels that were culture-negative in lymph node samples; conversely, 2 AMOS-PCR-positive results were reported in milk from 2 RBT-negative camels that were culture-positive (Dadar & Alamian, 2020) (Table 3). Considering that in domestic ruminants RBT-positive culture-negative animals are very exceptional, these results need to be confirmed. RT-PCR This methodology may be applied qualitatively (or semi-quantitatively when Ct values are reported) to detect the presence or absence of the target sequence, or quantitatively when calibrated against known copy numbers. Strictly, the former corresponds to real-time PCR (RT-PCR), whereas the latter should be designated quantitative PCR (qPCR) (Bustin et al., 2009). Although some camel studies describe the method used as qPCR and the original in vitro protocol from which it was adopted may be qPCR (Ameni et al., 2024, 2025; Getachew et al., 2025; Wainaina et al., 2020) (Table 4), none presents the calibration plot, number of copies estimated and the corresponding DSe/DSp adjustments, typically determined by ROC analysis in qPCR studies. Therefore, only RT-PCR has been applied in camel brucellosis. Table 4RT-PCR studies1,2.Table 4 dummy alt textReference [Country]Samples (no.) and protocolCt cut-off and controlsResultsComments(Gwida et al., 2011) [United Arab Emirates]Sera from apparently healthy camels (895) imported from Sudan.DNA extraction: serum (200 μl) High pure PCR Template preparation Kit (Roche Diagnostics).Primers and methodology: BCSP31 (31 Kd protein) gene primer for Brucella, and IS711 insertions primers for B. abortus and B. melitensis and probes according to Probert et al. (2004); TaqMan Real-Time PCR with TaqMan Universal PCR Master Mix.Positive Ct ≤ 40.DSe controls.– Culture: n.d.– Serology: RBT, SAT, CFT, cELISA and FPA.DSp controls: n.d.RT-PCR +: 759/895 (all B. abortus), 534 seropositive and 118 seronegative in all tests.DSe/DSp unknown (serological tests not validated; no DSp controls). According to a LCA, RT-PCR has high DSe but low DSp (see text).(Dehkordi et al., 2012) [Iran]Abomasal contents of aborted fetuses (618).DNA extraction: Amount not described; method designed for Listeria monocytogenes in pure culture; DNA purified using Fermentas kit (GmbH, currently Thermo Fisher Scientific] Germany).Primers and methodology: Primers targeting BMEII0466 and BruAb2_0168 for species identification in Foodproof Brucella kit (Hygiena, USA) with internal and positive controls.Positive Ct, n.d.No DSe or DSp controls.RT-PCR in 618 samples: B. melitensis 51; B. abortus 143; both B. melitensis and B. abortus 25.Non-validated DNA extraction methodology.DSe/DSp unknown: no controls of any kind.Purported co-infections by two Brucella species.(Khamesipour et al., 2015) [Iran]Blood (135) and lymph nodes (135) from apparently healthy camels.DNA extraction: blood and subscapular lymph nodes (amount and processing not reported); DNA extraction kit DNP, CinnaGen.Primers and methodology. IS711 primers for Brucella spp., B. melitensis and B. abortus as in (Hinic et al., 2008). TaqMan Real-Time PCR with TaqMan Universal PCR Master Mix. No amplicon contamination and amplification failure controls described.Positive Ct ≤ 35 (but distilled water instead of template DNA as negative control).DSe controls.– Culture: n.d.– Serology: n.d.DSp controls: n.d.RT-PCR+: 22/135.Brucella spp.18 blood (3 B abortus and 15 unknown) and 4 lymph nodes (all B. melitensis).15 positive bloods were not Brucella (sic. “unknown”)Unclear report: 15 “unknown” positives (no template controls).(Fatima et al., 2016) [Pakistan].Sera from randomly selected camels (200).DNA extraction. Serum (amount not reported) DNeasy Blood and Tissue Kit (QIAGEN, Germany).Primers and methodology: BCSP31 (31 Kd protein) gene primer for Brucella, and IS711 insertions primers for B. abortus and B. melitensis and probes according to Probert et al. (2004); TaqMan Real-Time PCR with TaqMan Universal PCR Master Mix. No amplicon contamination and amplification failure controls described.Positive Ct ≤ 35.DSe controls.– Culture: n.d.– Serology: RBT, cELISA.DSp controls: n.d.Seropositive: 10 (RBT+10, cELISA+:4).RT-PCR+: 3 (both Brucella and B. abortus primers); all RBT+, cELISA+.RT-PCR- all seronegative.DSe/DSp unknown (serological tests not validated; no DSp controls). PCR less sensitive than serology or serology less specific?(Juhasz et al., 2019) [Dubai]Single animal case study: Blood, 11 lymph nodes (see Culture), liver uterus, udder and mammary secretions of an aborted dromedary.DNA extraction: Tissues/organs (n 20) (20 mg) incubated in ATL buffer (Qiagen) and lysate processed in a MagNa Pure station (Roche) with MagNa Pure DNA extraction kit.Primers and methodology: BCSP31 (31 Kd protein) gene primer for Brucella, and IS711 insertions primers for B. abortus and B. melitensis and probes according to Probert et al. (2004); TaqMan RT PCR with primers in AmpliTaq Gold buffer and TaqMan probes (Qiagen) (Probert et al., 2004). No amplicon contamination and amplification failure controls described.Positive Ct ≤ 40.DSe controls.– Culture: (Table 1 and text).– Serology: RBTDSp controls: n.d.Animal RBT+.Culture+: 9/20 samples.PCR+: 1/20 samples(No RT-PCR +, culture -samples)Case study strongly suggests culture more sensitive than the RT-PCR protocol used (see text).(Khan et al., 2020) [Egypt]Sera (381).DNA extraction: sera (amount not reported) by QIAcube machine (QIAGEN) using the QIAamp DNA Mini Kit (QIAGEN).Primers and methodology (BCSP31 (31 Kd protein) gene primer for Brucella, and IS711 insertions primers for B. abortus and B. melitensis and probes according to Probert et al. (2004); B. suis primers as in (Hänsel et al., 2015). TaqMan RT PCR with TaqMan Universal PCR Master Mix. No amplicon contamination and amplification failure controls described.Positive Ct ≤ 38.DSe controls.– Culture: n.d.– Serology: RBT, CFT, iELISA, cELISA.DSp controls: Escherichia coli, O. intermedium and S. aureus DNA; serum/a from non-infected animals.Serology: RBT+ 59; iELISA+ 87, cELISA+ 77, CFT+: 118+ (only 4 animals + in all tests).RT-PCR+: 32 (25 B abortus, 5 B suis and 2 B melitensis);DSe/DSp unknown (no DSe controls [wide disagreement among serological tests, and tests not validated]; no DSp controls: n° and origin of serum/sera from non infected animals not described). PCR less sensitive than serology or serology less specific?(Wainaina et al., 2020) [Egypt]Pools from farms (11) or individual (3) milk from camelsDNA extraction: sediment (200 μL) from milk (1.5 mL); DNeasy Blood and Tissue Kit (Qiagen).Primers and methodology: BCSP31 (31 Kd protein) gene primer for Brucella and probe according to Probert et al. (2004). TaqMan RT PCR with primers in AmpliTaq Gold buffer and TaqMan probes (Qiagen) (Probert et al., 2004) (RT-PCR; named qPCR in the report).Positive Ct ≤ 39DSe controls.– Culture: n.d.– Serology: n.d.DSp controls: n.d.RT-PCR+: 0/11 and 0/3RT-PCR described as qPCR.Invalid report for DSe/DSp assessment: no controls DSe/DSp and use of pooled samples.(Manivannan et al., 2021b) [Qatar]Serum (45), placental tissue (40), blood (25), placental fluid (13), milk (50) uterine swab (30) samples (no. of camels, n.d.).DNA extraction: Magnapure compact (Roche) and Qiagen extraction kit for blood; Qiagen extraction kit for tissues. Amounts and sample processing not reported.Primers and methodology: BCSP31 (31 Kd protein) gene primer for Brucella, and IS711 insertions primers for B. abortus and B. melitensis and probes according to Probert et al. (2004); TaqMan RT PCR with TaqMan Universal PCR Master Mix.Positive Ct ≤ 39DSe controls.– Culture: n.d.– Serology: tests applied to both serum and blood (see Table S1).DSp controls: n.d.Serological data not interpretable (obtained applying tests to both serum) samples.RT-PCR+: blood and serum 20/70; 14/40 tissue samples; 6/13 placental fluids; 5/50 milk; uterine swabs 5/30.Two different kits used for DNA extraction from blood.Invalid report for DSe/DSp assessment: no DSe/DSp controls and serological tests invalid in blood samples.(Manivannan et al., 2021a) [Qatar]A total of 150 tissue samples (no. of camels n.d.): lung (15), liver (21), placenta (36), spleen (25), kidney (20) and placental fluid (33).DNA extraction: tissue (25–30 mg) QIAamp-DNA mini kit (QIAGEN).Primers and methodology: BCSP31 (31 Kd protein) gene primer for Brucella, and IS711 insertions primers for B. abortus and B. melitensis and probes according to Probert et al. (2004); TaqMan RT PCR with TaqMan Universal PCR Master Mix.Positive Ct ≤ 39DSe controls.– Culture: n.d.– Serology: n.d.DSp controls: n.d.RT-PCR+: 77/150 (lung 9/15, liver 15/2, placenta 16/36, spleen 18/25, kidney 5/20, placental fluid 14/33). Brucella species not reported.Unclear report: no Brucella spp. despite specific primers; Figure 1 shows Bruceladder results (no description of strains, DNA etc.). No of camels not reported. Unknown DSe/DSp (no controls).(Muturi et al., 2021) [Kenya]Sera from seropositive camels (32).DNA extraction: serum (200μl) QIAamp DNA Blood Mini Kit.Primers and methodology: IS711 primers and probes for Brucella genus, B. abortus, B. melitensis and B. suis in Supplemental Material 1 of the publication. No amplicon contamination and amplification failure controls described.Positive Ct ≤ 39DSe controls.– Culture: n.d.– Serology: iELISA.DSp controls: n.d.iELISA +: 32PCR +: Brucella 16; B. melitensis 5; species not identified for 11..DSe/DSp unknown (assessment not possible: PCR conducted only on seropositive camels; iELISA not validated).(Njeru et al., 2021) [Kenya]Sera from seropositive animals (camels, sheep, goats and cattle) (Table S1).DNA extraction: serum (200μl) QIAamp DNA Blood Mini Kit.Primers and methodology: BCSP31 (31 Kd protein gene primer for Brucella, and alkB gene BMEI1162 IS711 insertions primers for B. abortus and B. melitensis and probes according to Probert et al. (2004) and PerfeCTa qPCR masterMix.Positive Ct ≤38DSe controls.– Culture: n.d.– Serology: RBT, mRBT, iELISADSp controls: n.d.(RBT-mRBT-iELISA) + camels 10/50 (60/532 for camels, sheep, goats and cattle together).RT-PCR+ 49/60 (all B. abortus); (individual data for each animal species not reported).DSe/DSp unknown (assessment not possible: PCR conducted only on animals selected by a positive result in at least one test; serological tests not validated; RT-PCR data for sheep, cattle, goats and camels pooled; no DSp controls).(Taha & Abbas, 2021)MRT+ milk (15) from lactating camels (15) RBT+, mRBT+, cELISA + camels in farms with a history of brucellosis.DNA extraction: not described.Primers and methodology: IS711 primers as in (Bounaadja et al., 2009) and TaqMan RT PCR (incomplete description). No amplicon contamination and amplification failure controls described.Positive Ct ≤ 39DSe controls.– Culture: n.d.– Serology: RBT and mRBT, cELISA. Milk: MRTRBT+, mRBT+, cELISA +: 15MRT+. 9/15RT-PCR+ 11/15.Unclear report and results: methodology not properly described; serological tests not validated; MRT not reliable in camel milk. Authors assume RT-PCR is the “gold standard”.(Mohammed Babiker et al., 2022) [Saudi Arabia]Milk (“representative” [sic] samples) and vaginal swabs (n.d.) from seropositive animals taken during a control test and slaughter and Rev 1 vaccination program.DNA extraction: sample processing and amount not reported. MagNA pure Compact Nucleic Acid isolation Kit.Primers and methodology: authors describe the LightMix kit Brucella melitensis and LightCycler® 480 (Roche Diagnostics) as identifying B. melitensis but according to the manufacturer, the kit (TaqMan mix) uses primers for Brucella genus IS711. No amplicon contamination control described; amplification failure control possibly included in kit.Positive Ct (not reported).DSe controls.– Culture: n.d.– Serology: cELISA.DSp controls: n.d.RT-PCR positive in some but not all milk samples from cELISA positive camels; negative in vaginal swabsUnclear report. Kit does not detect B. abortus.Data cannot be evaluated because of the extensive introduction of imported camel in the study herds, RT-PCR run only in “representative” (sic.) milk and insufficient data on herd management. No DSe controls.(Wakjira et al., 2022) [Ethiopia]Camels (248) clinically healthyDNA extraction: blood clots (amount not described) from seropositive animals QIAMP DNA Mini Kit (QIAGEN).Primers and methodology: BCSP31 (31 Kd protein gene primer for Brucella, and IS711 insertions primers for B. abortus and B. melitensis and probes according to Probert et al. (2004). TaqMan RT PCR (no other details given). No amplicon contamination and amplification failure controls described.Positive Ct ≤ 45DSe controls.– Culture: n.d.– Serology: iELISA.DSp controls: n.d.iELISA+: 21/201RT-PCR+: 0/21DSe/DSp unknown (assessment not possible: PCR conducted only on seropositive camels; iELISA not validated; no DSp controls). Unusually permissive Ct. PCR less sensitive than serology or serology less specific?(Ameni et al., 2024) [Abu Dhabi]Sera from 356 camels.DNA extraction: sera (amount not reported); NucleoSpin Tissue kit (Macherey-Nagel GmbH & Co.KG).Primers and methodology. BCSP31 (31 Kd protein) gene primers for Brucella and probe according to Probert et al. (2004). FIREPol EvaGreen qPCR Supermixt used qualitatively (RT-PCR; named qPCR in the report). No amplicon contamination control described.Positive Ct ≤ 40.DSe controls.– Culture: n.d.– Serology: cELISA and iELISA.DSp controls: n.d.iELISA+ and cELISA+ 51/356.RT-PCR+ 13/51.RT-PCR described as qPCR. DSe/DSp unknown (assessment not possible: PCR conducted only seropositive camels; serological tests not validated). PCR less sensitive than serology or serology less specific?(Getachew et al., 2025) [Ethiopia]Blood from RBT+ and iELISA+ camels (160).DNA extraction: blood (200 µL) clots (QIAGEN DNA extraction kit).Primers and methodology. IS711 primers for Brucella spp. B. melitensis and B. abortus and probes as in (Hinic et al., 2008). TaqMan RT PCR with TaqMan Universal PCR Master mix, no AmpErase UNG (also, AMOS-PCR and PAGE; Table 2) (named qPCR in the report). No amplicon contamination control described.Positive Ct ≤ 38.DSe controls.– Culture: n.d.– Serology: RBT, iELISA.DSp controls: n.d.Seropositive: RBT+7; ELISA+:6.RT-PCR+: 5 (all RBT+).RT-PCR described as qPCR. DSe/DSp unknown (assessment not possible: PCR conducted only seropositive camels; serological tests not validated; no DSp controls). PCR less sensitive than serology or serology less specific?(Ameni et al., 2025) [Abu Dhabi]Sera from 356 camels (and abattoir workers).DNA extraction: sera (amount not reported); NucleoSpin Tissue kit (Macherey-Nagel GmbH & Co.KG).Primers and methodology: IS711 insertions primers for B. abortus and B. melitensis and probes according to Probert et al. (2004) and B. ovis primers from aroA according to Edao et al. (2023) (see text for discussion). RT-PCR (named qPCR in the report) conducted using a CFX96Real-TimePCRsystem (Bio-Rad, USA) with SYBRGreen to monitor amplification. No amplicon contamination control described.Positive Ct ≤ 40.DSe controls.– Culture: n.d.– Serology: cELISA.DSp controls: n.d.cELISA +: 30.RT-PCR +: 20/30; B. ovis 10; B. abortus 6; B. melitensis 2; B. ovis-B. abortus co-infection 2; B abortus-B. melitensis co-infection 1.RT-PCR described as qPCR. DSe/DSp unknown (assessment not possible: PCR conducted only seropositive camels; serological tests not validated; no DSp controls). PCR less sensitive than serology orserology less specific? Authors purportedly describe B. ovis as the most frequently detected species in camels and purported B. ovis infections in humans. Also, purported co-infections by several species in camels (and in humans).1Abbreviations: DSe, diagnostic sensitivity; DSe, diagnostic specificity; n.d.: not done.2All iELISA kits were intended for ruminants or for ruminants and pigs. For a discussion of serological tests in these works, see Section 3.3. Brucellosis serological tests used in camels and Table S1. RT-PCR studies1,2. Abbreviations: DSe, diagnostic sensitivity; DSe, diagnostic specificity; n.d.: not done. All iELISA kits were intended for ruminants or for ruminants and pigs. For a discussion of serological tests in these works, see Section 3.3. Brucellosis serological tests used in camels and Table S1. The 17 studies using this methodology are summarized in Table 4. One study does not describe the DNA extraction and purification procedure, and 7 do not report the sample volume processed. In the remaining studies, the extracted material was limited (e.g., 200 μL of serum; clots from 200 μL of blood; 200 μL of sediment from 1.5 mL of milk; or 20–30 mg of tissue), and at least six different commercial kits were used. Although primer selection and protocols were more homogeneous than in conventional PCR studies, variability was also significant. Eight studies used the Brucella genus plus B. abortus and B. melitensis primers described by Probert et al. (Probert et al., 2004) (in one case only the genus specific primers; in another complemented with a fourth set of B. suis primers); 2 used the Brucella genus plus B. abortus and B. melitensis primers of Hinic et al. (Hinic et al., 2008); and 1 used the B4/B5 primers targeting the Brucella bcsp31 gene (Baily et al., 1992). TaqMan-based RT-PCR was the most frequently employed methodology (10 studies), usually, but not always, with TaqMan probes. Of the 17 studies, 12 do not report the use of positive Brucella DNA and non-template controls, precautions respectively necessary to rule out amplification failure and amplicon contamination. Concerning ASp, the original protocol of Probert et al. included ASp assessments against Agrobacterium spp., O. anthropi and five additional Ochrobactrum strains (Probert et al., 2004) and that of Hinic et al. evaluated cross-detection of Agrobacterium spp. (Hinic et al., 2008). As noted above, B4/B5 primers cross-detect some Ochrobactrum strains (Casañas et al., 2001). However, no study evaluated the ASe using spiked samples or provides data suitable for estimating the DSe/DSp of the RT-PCR protocol used (Table 4). One study analyzed pools of milk (Wainaina et al., 2020); 4 contain inconsistencies that limit interpretability (Manivannan et al., 2021a; Mohammed Babiker et al., 2022; Taha & Abbas, 2021; Wainaina et al., 2020); and 3 lacked any DSe/DSp controls (Khamesipour et al., 2015; Manivannan et al., 2021a; Wainaina et al., 2020). In 7 studies, the lack of DSp controls and/or the use of non-validated serological tests precludes comparison of the diagnostic performance of the serological tests and the RT-PCR protocol. Notably, 2 studies reported B. melitensis-B. abortus and B. ovis-B. abortus co-infections in the same animal (Ameni et al., 2025; Dehkordi et al., 2012). One of these studies further reported B. ovis-B. abortus more frequently than B. abortus or B. melitensis (Ameni et al., 2025), and also B. ovis detection in exposed workers, despite it is well-known non-zoonotic nature. The validity of these studies is discussed below. Finally, the study performed by Juhasz et al. (Juhasz et al., 2019) (Section 3) compared RT-PCR and culture in 19 organs/tissues and blood from a single infected camel. Remarkably, RT-PCR was positive in only 1 (an udder quarter) of the 9 culture-positive samples, and no culture-negative sample was RT-PCR-positive. Consistent with these findings, among the 541 samples collected from 43 organs and joint fluids in a study of 11 experimentally infected camels (Section 3.2), only the tarsal joint fluid of a single animal tested positive by RT-PCR (Johnson et al., 2018). In contrast, culture was positive in this RT-PCR-positive sample and in other 47 samples. Discussion This review highlights significant methodological limitations in the bacteriological, serological and DNA-based brucellosis diagnostic tests applied to camels that is necessary to address in future studies. Concerning bacteriology, deficiencies in sample number and type, processing methods, inoculum volume, culture media, growth conditions, and strain typing collectively contribute to the low isolation rates reported in most investigations. Evidence from the case study by Juhasz et al. (Juhasz et al., 2019) and experimental infections described by Johnson et al. (Johnson et al., 2018) demonstrates that a correct methodology parallels that used in ruminants to optimize Brucella detection (Ducrotoy et al., 2018). Maximizing the chances of Brucella isolation from necropsy samples requires systematic examination of multiple lymph nodes, including bilateral submandibular and all mammary nodes. Similarly important, the inoculum per plate taken from tissue homogenates in the minimal amount possible of buffer must be abundant and the number of plates increased. Direct smearing of portions of tissues on culture plates is ineffective and constitutes a waste of both laboratory resources and valuable samples. The same precautions (volume, inoculum size and number of plates) apply to milk culture. Finally, the above-discussed challenges associated with performing bacteriology investigations with fetal tissues compared with the advantages of culturing numerous vaginal smears obtained from aborted camels should be taken into consideration. The choice of selective media is also relevant. It is worth noting that the antibiotics in Farrell’s medium inhibit many B. melitensis and B. suis and some B. abortus strains. The antibiotics in CITA medium are less inhibitory but even so a combination of Farrell, CITA and modified Thayer-Martin media is recommended for optimal diagnostic performance (de Miguel et al., 2011). These studies also require systematic implementation of existing (AMOS-PCR and Bruce-ladder, and WGS whenever possible) DNA-based methods for molecular typing, a complement necessary to overcome the above-mentioned imprecisions about Brucella biovar distribution in camels in current reviews (Dadar et al., 2022). PCR-based assays have simplified the characterization of brucellae obtained in pure culture. Many of these methods have been also applied directly on DNA extracted from field samples. However, the large diversity of protocols hinders standardization, and diagnostic studies with appropriate control samples in domestic livestock or wildlife are largely lacking (Moreno et al., 2022). Camel studies are no exception, and no available evidence identifies a protocol with optimal DSe and DSp. This limitation affects not only the selection and standardization of amplification and detection methods but also sample processing and DNA extraction procedures. Blood, blood clots and serum have been used in camel studies, yet none has compared the information provided by these three related samples or compared them with culture, the gold standard. Studies in human brucellosis, however, demonstrate that blood requires specific processing to prevent PCR inhibition by hemoglobin and that blood clots are unsuitable for PCR or RT-PCR (Yagupsky et al., 2019). Similarly, although Brucella can be cultured from either the cream and fractions of milk or from both several camel PCR studies use only one fraction, reducing the probability of detection. Commercial DNA extraction and purification kits or platforms are logically preferred, and over 90% of the studies use such kits from different makers or formats. However, comparative studies in vitro show that they are neither equivalent nor perform equally in all types of samples (Preis et al., 2019; Queipo-Ortuño et al., 2007; Tomaso et al., 2010). Notably, although examined in saline extracts of formalin-fixed tissues, a comparative study in samples of two serologically positive camels found significant differences among five commercial extraction kits (Tomaso et al., 2010). Further studies on this topic are necessary. Detection efficiency depends on both the quantity of extracted material and the Brucella DNA load. When reported, sample volumes include 200–400 μL of blood, 20–30 mg of tissue (200 mg in the above-described experimental infection), and 200–400 μL of milk cream or sediment from 1.5 mL of milk (Table 3, Table 4). For pure DNA, ASe data suggest detection limits corresponding to approximately 37–38 genome copies ((Probert et al., 2004), 10 copies (Hinic et al., 2008), and ∼25 copies for B4/B5 primers targeting BCSP31 (Queipo-Ortuño et al., 2008), assuming ∼4 fg DNA per genome (Gwida et al., 2011; Tomaso et al., 2010). Under idealized assumptions of complete inhibitor removal, 100% extraction efficiency, and equivalence between DNA copies and viable bacteria, these limits imply bacterial concentrations of ∼100 organisms/mL in blood or milk and >1000 CFU/g in tissue for the Probert protocol. As indicated above, no true qPCR has been applied in camels so that the number of DNA copies detected was not assessed in any study. For RT-PCR, an indirect estimate is the Ct number at which the signal becomes higher than the control. Most Ct cut offs are below 38–40 (Table 4), which controls (when correct and included) and experience show is likely to be appropriate. Yet, only 3 studies provided Ct plots (Ameni et al., 2025; Manivannan et al., 2021b; Muturi et al., 2021) and one reported the Cq/Ct (Khan et al., 2020). In these reports, most positive signals were close to the Ct threshold, indicating low effectiveness of the complete process (i.e., sample size, extraction, detection, etc.). In contrast, well-conducted bacteriology can detect 1 CFU in tissue homogenates or 1 CFU/mL in cow milk (Muñoz et al., 2022a, b). This disparity in sensitivity likely contributes to the findings from the only 2 rigorous camel studies, which reported considerably lower sensitivity of RT-PCR compared with bacteriology (Johnson et al., 2018; Juhasz et al., 2019). Regarding specificity (i.e., cross-detection), in vitro evaluations of ASp exists for AMOS-PCR and the protocols of Probert et al. and although more limited for Hinic et al. (Bricker & Halling, 1994; Hinic et al., 2008; Probert et al., 2004). Among the Brucella -related bacteria examined, Agrobacterium and Rhizobium species are rhizosphere dwellers unlikely to contaminate animal samples, but Ochrobactrum spp. are ubiquitous and have been isolated from domestic animals or their environment (Choi et al., 2020; Yamamoto et al., 2010). Although cross-amplification of some Ochrobactrum strains was excluded for several protocols (Bricker & Halling, 1994; Hinic et al., 2008; Probert et al., 2004; Taha & Abbas, 2021), the B4/B5 primers targeting the bcsp31 gene used in camel studies amplify certain Ochrobactrum strains (Casañas et al., 2001), and other camel PCR or RT-PCR protocols were not evaluated. Moreover, additional members of the Ochrobactrum–Brucella group have been identified since those ASe were performed. They include O. soli (Choi et al., 2020) Falsochrobactrum ovis (Kampfer et al., 2013), Pseudochrobactrum algeriensis (Loperena-Barber et al., 2022) and P. saccharolyticum (Pereira et al., 2023), all isolated from ruminants, indicating the need for expanded specificity testing of DNA amplification methods in brucellosis. Besides, given the enlarging taxonomy of the Brucellaceae, unidentified related bacteria cannot be excluded, stressing the importance of assessing direct diagnosis PCR protocols using controls from animals rigorously proved to be Brucella-free. Amplicon contamination represents an additional source of false-positive results requiring strict controls at levels. The detection by PCR or RT-PCR of animals simultaneously positive for B. abortus and B. melitensis in 5 studies (Ameni et al., 2025; Dehkordi et al., 2012; Getachew et al., 2025; Khamesipour et al., 2014; Rabah et al., 2020) is reminiscent of PCR and RT-PCR studies reporting B. abortus and B. melitensis co-infections in African cattle (Mengele et al., 2024; Mitterran et al., 2020) and warrants critical scrutiny. Given the extensive cross-immunity among Brucella species and the absence of documented mixed infections in large-scale field studies where multiple host species coexist (De Massis et al., 2019; Taleski et al., 2002; Verger et al., 1989), such findings ought to be evaluated critically. Similar concern raises the report that camels are more often infected by B. ovis than by B. abortus or B. melitensis, or co-infected by B. ovis and B. abortus, and of abattoir workers infected by B. ovis (Ameni et al., 2025). Since B. ovis is primarily sexually transmitted among sheep and non-zoonotic, such results represent a laboratory artifact, which could also explain the simultaneous detection of B. melitensis and B. abortus DNA in some samples. Significantly, none of these studies reported amplicon contamination controls, true negative control samples or confirmatory bacteriology. Therefore, robust exclusion of contamination and corroboration by bacteriology are required to substantiate such claims. Although most PCR or RT-PCR studies included serological tests and some are presented as a comparative study (Ghorbani et al., 2013; Gwida et al., 2011) (Section 4), none included proper DSp controls. Consequently, the assertion that RT-PCR is a gold standard (Khan et al., 2020; Taha & Abbas, 2021) or a confirmatory test (Xingxing et al., 2024) has no bases. One LCA study comparing RT-PCR, SAT, RBT, cELISA, and FPA (Gwida et al., 2011) reported markedly lower agreement of RT-PCR with the latent classes (see below) compared with conventional serological tests. This result can be interpreted to mean a substantial proportion of false-positive RT-PCR results (as suggested by the raw data [Table 4]) or a low performance of the classical serological tests. Therefore, when the results of PCR and serological tests are in conflict in camels, as they did most often, the discrepancy cannot be attributed confidently to DSe/DSp deficiencies in either method, and interpretations on terms of prevalence and further elaborations of the data are not granted. Regarding serological studies, most if not all employed diagnostic tests that were neither validated nor verified against a local gold-standard serum panel comprising culture-confirmed positive and truly brucellosis-free camels. Furthermore, the cross-study comparison of serological assays conducted here (including SAT, BPAT, RBT, TAT, and ELISA) demonstrates substantial inconsistencies rather than the agreement described in a recent review on camel brucellosis (Dadar & Foster, 2025). These discrepancies indicate first that WOAH standardization protocols are not consistently implemented. This problem is compounded by the use of immunoconjugates not validated for camel immunoglobulins, reliance on manufacturer-recommended cut-offs established for other species (rather than cut-offs derived from camel-specific validation studies), application of ruminant-based thresholds in classical tests such as SAT and CFT, and the lack of test harmonization. An additional difficulty is that sample sizes are frequently calculated using seroprevalence estimates derived from previous studies conducted using non validated tests either in the same area or, more commonly, in different regions. However, these underlying studies are themselves unreliable for the reasons discussed above. Collectively, these limitations compromise the accuracy of most original studies, including seroprevalence estimates and risk factor assessments (respectively 157 and 73 of 176 studies in Table S1). The implications of these deficiencies are considerable. First, a substantial proportion of the published studies are unlikely to accurately represent the true epidemiological situation of camel brucellosis, thereby calling into question the validity of subsequent analyses, including comparisons across different geographic regions, meta-analyses, and systematic reviews (Abu-Basha et al., 2024; Adel, 2022; Akinyemi et al., 2022; Dadar et al., 2021, 2022; Dadar & Foster, 2025; Dagnaw et al., 2024; Gutema & Tesfaye, 2019; Hughes & Anderson, 2020; Menshawy et al., 2025; Mohammadpour et al., 2020; Mohammed et al., 2023; Refai, 2002; Sibhat et al., 2022; Tadesse, 2016; Waji & Neja, 2023; Wakene & Mamo, 2017; Wareth et al., 2014). Second, and more importantly, the widespread use of inadequately validated diagnostic tools may adversely affect public health policies, trade decisions, control of animal movements and One Health interventions related to brucellosis in camels and other livestock species in contact with these animals. There is a clear need for studies characterizing the response kinetics and profiles of IgM, IgG1 and HCAbs IgG2 and IG3 during Brucella infection in camels to determine their diagnostic relevance, rather than assuming equivalences with the IgM, IgG1, IgG2 and IgA responses ruminants, as it has been done recently (Dadar & Foster, 2025). Such investigations should be complemented by analyses of how specific immunoglobulin classes and subclasses are detected by assays based on different principles, including agglutination at neutral and acidic pH, guinea pig complement activation, immunoconjugate specificity and cross-reactivity, and potential competition effects among isotypes. Although technically demanding, these studies are essential for advancing the serological diagnosis of camel brucellosis. Useful precedents exist in earlier camel antibody research (Azwai et al., 1993, 1995, 1996) and future work can also build on established knowledge of protein A and protein G reactivity in relation to HCAbs (Hamers-Casterman et al., 1993) and on the extensive basic information on ruminant serology (Ducrotoy et al., 2016). Like in DNA-based methods, after solving standardization issues, validation is the immediate target identified in this and some previous works (Ahmed et al., 2024; Gwida et al., 2012; Khan et al., 2020; Sprague et al., 2012; Wakjira et al., 2022; Wernery et al., 2022; Wernery, 2016). Overall, neither frequentist nor Bayesian approaches applied to camel brucellosis serology have provided reliable DSe/DSp estimates, primarily due to the absence of valid reference standards, insufficient validation and standardization of tests, and questionable modeling assumptions. Bayesian analyses have not been performed according to the standards of this methodology and they yield contradictory estimations. Moreover, because of the complexities of brucellosis tests and antibody responses (Ducrotoy et al., 2018), this approach is unlikely to circumvent the issue of the validity of priors and the reliability of the dependent or independent conditionality assumptions. Obviously, frequentist analyses require obtaining samples that could serve as DSe/DSp gold standards. Well-performed bacteriological studies in camels are essential because even relative standards require previous validation (Jacobson, 1998). Thus, a collection of well-characterized sera from culture-positive and brucellosis free camels would be a great asset in this direction. In this regard, validation of the RBT seems the first target for defining a relative standard. As it happens in domestic ruminants (Ducrotoy et al., 2016), the few studies carried out with a properly standardized antigen strongly suggest that the RBT it is not outperformed by other WOAH classical or sophisticated tests (Gwida et al., 2011; Khan et al., 2016) and, importantly, its purported lack of DSp is a misconception. In addition, its simplicity, rapidity, easy standardization and quality control and low cost are desirable properties in most areas where camels are bred. Indeed, the RBT should also be investigated with regards to its ability to detect the different antibodies of the camel repertoire. Certainly, all these studies require sustained international cooperation but above all the involvement of well-equipped and trained reference laboratories in camel keeping countries. In conclusion, although logistical and environmental constraints inherent to camel husbandry may hinder optimal sampling and culture procedures, rigorously designed bacteriological studies remain essential, as they constitute the reference (gold standard) for validating DSe of serological and DNA amplification diagnostic tests. Regarding the latter, systematic evaluations encompassing sample type, processing, DNA extraction, amplification protocols, and appropriate positive and negative controls are lacking for PCR and RT-PCR in camels, and the most solid available evidence indicates low DSe relative to bacteriology. Strict exclusion of DNA contamination and complementary bacteriological investigation are essential in future studies. Similarly, systematic implementation of standardization and repeatability controls for serological tests, as well as access to collections of well-characterized sera from culture-positive and Brucella free camels for validation are necessary. Statement for studies in humans/animals This review article does not include experiments carried out by the authors in animals or humans. CRediT authorship contribution statement Mammar Khames: Methodology, Conceptualization. Amaia Zúñiga-Ripa: Methodology, Conceptualization. Raquel Conde-Álvarez: Methodology, Conceptualization. Pilar M. Muñoz: Methodology. José María Blasco: Writing – review & editing, Methodology. Ignacio Moriyón: Writing – review & editing, Writing – original draft, Methodology, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements The authors gratefully acknowledge the expert assistance of Lucía Zubasti-Madoz (University of Navarra Library) in retrieving the pdf publications used in this review. This review was carried out in the frame of projects PID2023-146797OB-C31 and PID2023-146797OB-C32 financed by MCIN/AEI/10.1303910.13039/501100011033. The work of P.M.M. (CITA) and R. C-Á (UNAV) was also supported by REPRODIVAC project under Grant Agreement No 10106081 through the HORIZON—CL6-2021-FARM2FORK-01–06 Call. Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.vas.2026.100735. Diagnostic specificity (DSp) is defined as the proportion of true negatives correctly identified by the test, i.e., the number of true negative results divided by the total number of true negatives tested. Diagnostic sensitivity (DSe) is defined as the proportion of true positives correctly identified, i.e., the number of true positive results divided by the total number of true positives tested. Values equal or close to 0.5 (or 50%, when expressed as a percentage) indicate a diagnostic performance no better than random classification. These parameters should not be confused with analytical sensitivity (ASe) and analytical specificity (ASp), which refer, respectively, to the minimum detectable amount of a given analyte and the ability of the assay to distinguish the target analyte from potential interfering substances. ASe and ASp are typically expressed in appropriate measurement units (e.g., colony-forming units, antibody dilution, nanograms, femtograms or genome copies). Contributor Information Mammar Khames, Email: drmammar@hotmail.fr. Amaia Zúñiga-Ripa, Email: azuniga@unav.es. Raquel Conde-Álvarez, Email: rconde@unav.es. Pilar M. Muñoz, Email: pmmunnoz@cita-aragon.es. José María Blasco, Email: jblasco@unizar.es. Ignacio Moriyón, Email: imoriyon@unav.es. Supplementary materials References Associated Data Supplementary Materials