Prevalence of hyperhomocysteinemia in South Asian patients with heart disease: a systematic review and meta-analysis Authors: Mysha Nowrin Alabbi, Rifah Nanziba, Naila Masfiqua Malek, Nafisa Ahmed, Harisul Hoque, Mahbubul H Siddiqee, Badhan Bhattacharjee, Nadia S Deen Published: 2026 ABSTRACT AbstractBackgroundHyperhomocysteinemia, characterized by an elevated concentration of serum homocysteine, is a well-recognized risk factor for cardiovascular pathophysiology. South Asians are known to have a high burden of cardiovascular disease, and hyperhomocysteinemia is commonly observed in this population. The present systematic review and meta-analysis aimed to assess the pooled prevalence of hyperhomocysteinemia and estimate the serum homocysteine concentrations among South Asians with heart disease.MethodsFollowing PRISMA 2020 reporting guidelines, systematic searches were conducted to identify studies focusing on the South Asians, heart disease and homocysteine. Random-effects meta-analyses were performed using MetaXL to determine the pooled prevalence of hyperhomocysteinemia among patients with heart disease, and standardized mean difference in serum homocysteine concentration between patients with and without heart disease. Meta-regression analyses were performed to identify the source of data heterogeneity. Additionally, GRADE tool was used to assess the certainty of evidence for the study outcomes.ResultsA total of 47 articles comprising 6094 heart disease patients from India, Pakistan, Bangladesh and Afghanistan met the inclusion criteria. The overall pooled prevalence of hyperhomocysteinemia was 61% (95% CI: 49–72%) with a significant heterogeneity (I2 = 98%; p < 0.01). The weighted mean serum homocysteine concentration was 20.98 µmol/L (SD: 4.57). Additionally, the standardized mean difference in serum homocysteine concentration between patients with and without heart disease was 2.22 (95% CI: 1.63–2.82). Subgroup analyses of pooled prevalence and standardized mean difference based on different covariates demonstrated high degree of heterogeneity (I2 = 85–98%; p < 0.01) and the subsequent meta-regression analyses indicated that study precision, as indicated by the standard error and the number of patients, substantially contributed to the observed heterogeneity. Finally, GRADE assessment rated certainty of evidence for prevalence as low and for SMD as moderate.ConclusionDespite variability and limited country representation, this first meta-analysis of South Asian population shows that the prevalence of hyperhomocysteinemia is high among heart disease patients in this region, with homocysteine concentrations significantly exceeding the diagnostic threshold of ≥ 15 µmol/L. These findings stress the need for homocysteine-lowering interventional studies to confirm hyperhomocysteinemia as a potential risk factor of heart disease in this region.Trial registrationProspero registration number: CRD42023424762.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12872-026-06074-3.Keywords: Homocysteine, Hyperhomocysteinemia, South Asia, Prevalence, Heart disease Introduction Homocysteine (Hcy) is a sulfur-containing nonproteinogenic amino acid derived from the metabolism of methionine. It plays a crucial role in DNA methylation, antioxidant, such as glutathione synthesis, and redox homeostasis [1–4]. The metabolism of Hcy depends on several enzymes, including cystathionine beta-synthase and methylenetetrahydrofolate reductase (MTFR), as well as some essential cofactors such as vitamins B6 and B12 and folic acid. Mutations affecting these enzymes or deficiencies in these cofactors can impair Hcy metabolism, leading to elevated serum Hcy concentrations- a condition known as hyperhomocysteinemia (HHcy) [1, 2, 5]. HHcy has been implicated in the pathophysiology of numerous diseases [2, 3, 6]. Accumulating studies link HHcy with cardiovascular diseases, such as heart diseases including coronary artery disease, myocardial infarction, congenital heart disease, acute coronary syndrome, and ischemic heart disease [7–11]. HHcy has been found to promote endothelial dysfunction, decrease glutathione peroxidase activity, induce atherosclerosis, and trigger inflammation in the heart tissue [2, 3, 12]. South Asians (SA) representing a quarter of the global population include people from Afghanistan, Bangladesh, Bhutan, India, the Maldives, Nepal, Pakistan, and Sri Lanka [13, 14]. Many people of this region lack access to food and proper treatment. A greater incidence of heart disease has been described in SA than in people of other ethnicities [14–16]. SA people have also been reported to suffer from HHcy [17]. Although the mechanisms behind this association is not understood, inadequacy of nutritious food, and genetic background might predispose SA people to HHcy [18]. In recent years, an increasing number of studies have shown that SA patients with heart disease have increased serum Hcy concentrations compared with healthy individuals [19–24]. However, the single-centered nature and small sample size of most of these studies limit their generalizability. To address this limitation, the present systematic review and meta-analysis aimed to estimate the pooled prevalence of HHcy in SA patients with heart disease using a defined cutoff value for HHcy. The standardized mean difference in serum Hcy concentration in SA patients with and without heart disease was also analyzed. Additionally, the variation in HHcy prevalence based on various covariates was analyzed. Methods The latest recommendations and guidelines proposed by the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) [25] were followed for this study (Fig. 1). The study protocol was registered in PROSPERO (CRD42023424762). PRISMA chart detailing the search results and number of eligible studies Data sourcing and search strategy A preliminary search was conducted to select the eligibility criteria of the studies. Afterwards, the articles indexed in PubMed, SCOPUS and Google Scholar were systematically searched based on the eligibility criteria. The initial searches were conducted between June 15, 2023, and December 31, 2023 and the latest searches were performed in October–November 2025. To identify relevant study articles, the search strategy was designed to retrieve all articles published between January 2000 and October 2025, with a focus on the average Hcy level and prevalence of HHcy among SA patients with heart disease. Various combinations of Boolean operators’ keywords were used in the search process, i.e., ‘homocysteine,’ ‘hyperhomocysteinemia,’ ‘homocystinuria,’ ‘Hcy,’ ‘HHcy,’ ‘incidence,’ ‘prevalence,’ ‘level of homocysteine’ ‘heart patients,’ ‘coronary artery disease,’ ‘myocardial infarction,’ ‘South Asia,’ ‘Bangladesh,’ ‘India,’ ‘Pakistan,’ ‘Nepal,’ ‘Sri Lanka,’ ‘Bhutan,’ ‘Maldives,’ and ‘Afghanistan.’ To cover all the relevant articles for this study, the position of the keywords was periodically changed in the search bar. Additionally, ‘similar articles’ suggested by the repositories (PubMed, SCOPUS, and Google Scholar) were explored. To maximize the search efficiency, a forward and backward search was also conducted by manually checking the citing and reference lists of the individual articles. The search strings used across three different databases used for the literature search are available in Table S1. Finally, the WHO database was examined by using similar search terms as an additional source for country-specific information regarding Hcy and HHcy. Eligibility criteria HHcy is usually categorized into three classes: mild (15–30 µmol/L), moderate (31–100 µmol/L) and severe (> 100 µmol/L) [26–29]. Therefore, in this study, HHcy is defined if the serum (or plasma) concentration of Hcy is equal to or above 15 µmol/L and only studies that followed this definition were included. Additionally, the term ‘patients with heart disease’ was used in this study to address patients who are suffering from a critical medical condition associated with the heart, which includes coronary artery disease, myocardial infarction, congenital heart disease, acute coronary syndrome, and ischemic heart disease. Articles were included in this systematic review if they met the following inclusion criteria: (I) peer-reviewed original articles reporting either average Hcy levels or the prevalence of HHcy among patients with heart disease; (II) participants aged 18 years or above and who lived in SA countries, including Afghanistan, Bangladesh, Bhutan, India, the Maldives, Nepal, Pakistan and Sri Lanka. (III) Articles published between January 2000 and October 2025, and (IV) all peer-reviewed studies published in the English language if they matched the other eligibility criteria. On the other hand, the following exclusion criteria were applied in this study: (I) studies focused on countries located outside of South Asia; (II) studies that used less than 15 µmol/L as the cutoff value for HHcy; (III) non-peer-reviewed publications, letters to the editor, short communications, review articles, meta-analyses, etc.; (IV) articles published in languages other than English; and (V) studies that satisfied the inclusion criteria, but full-text articles could not be obtained from the authors after requesting them. Mendeley reference manager (version 2.62.0) was used to manage references and remove duplications. After eliminating duplicates, titles and abstracts were screened, and full-text screening was implemented for the relevant articles. Data extraction Data extraction was initially performed independently by four coauthors (MNA, NMM, RN and NA) between January 7, 2024, and January 21, 2024. Afterward, another coauthor (BB) rechecked the overall procedure to finalize the number of eligible articles. Any discrepancies between these four coauthors during data extraction were resolved through discussion with the project supervisor (NSD). An Excel sheet was used to develop a summary table that contained information on all the selected study articles (Table 1), including the reference (name of the first author and year of publication), study period, study location, number of study participants, age of the study participants, method of measuring the concentration of Hcy, study design, study setting, average concentration of Hcy, and number of people affected with HHcy. Table 1Summary of the selected studiesReferencesStudy locationStudy periodNumber of patients with heart diseaseAge rangeTypes of heart diseaseStudy designMethod of measurementHHcy among patients with heart diseaseAverage concentration of Hcy (µmol/L)MeanSDSaluddin et al., 2005 [30]Karachi, PakistanJune 2001- December 200163 (M-45, F-18)Mean age 47 yearsMICCFPIANot reported19.432.46Naiksatam et al., 2022 [31]Gujarat, IndiaAugust 2018-August 2020150 (M-109, F-4155–64 yearsCADCSEIA41Not reportedNot reportedLaghari et al., 2009 [32]Sindh, Pakistan2.5 years (Did not specify the timeline)110 (M-71, F-39)40–65 yearsMICCHPLCNot reported23.142.40Bhagwat et al., 2009 [33]Maharashtra, IndiaOctober 2001—January 20045020–50 yearsIHDCCHPLCNot reported31.3918.35Abraham et al., 2006 [34]Punjab, IndiaNot reported70 (M-57, F-13)Mean age 57 yearsACSCCSPIANot reported22.8113.91Angeline et al., 2005 [35]Tamil Nadu, IndiaNot reported60 (M-49, F-11)Mean age 47 yearsMICCEIANot reported24.596.14Bhargava et al., 2014 [36]New Delhi, India1 year (Did not specify the timeline)43Not reportedCADNot reportedCLIA22Not reportedNot reportedSingh et al., 2008 [37]Punjab, IndiaNot reported75 (M-75, F-0)30–60 yearsCADCCEIANot reported18.922.11Shenoy et al., 2014 [38]Karnataka, IndiaAugust 2012—November 201251 (M-42, F-9)31–65 yearsCADCCELISANot reported23.356.08Kaur et al., 2016 [39]Chandigarh, IndiaJuly 2006—December 2009350 (M-301, F-49)20 years or aboveMINot reportedEIA15120.0510.45Khare et al., 2005 [40]Mumbai, India2.5 years (Did not specify the timeline)120 (M-110, F-10)21 −39 yearsMICCEIA6323.4916.11Palazhy et al., 2015 [41]Kerala, India16 months (Did not specify the timeline)151 (M-151, F-0)35–70 yearsMI & CADCCELISANot reported16.68.60Mukhopadhyay et al., 2018 [42]Northern IndiaApril 2014—October 201540 (M-37, F-3)27- 75 YearsACSCCPMNot reported23.8116.73Tantray et al., 2019 [43]Hyderabad, IndiaNot reported100 (M-69, F-31)Mean age 44–47 yearsCADCCNot reportedNot reported19.71.80Shah et al., 2018 [23]Peshawar, PakistanOctober 2016—September 2017128 (M-86, F-42)25–40 yearsCADCCFPIA12844.514.01Iqbal et al., 2005 [44]Karachi, PakistanJanuary 2001- June 2001224 (M-169, F-55)30–70 YearsMICCFPIANot reported188.36Prajapati et al., 2015 [45]Gujarat, IndiaJanuary 2008—December 2012200 (M-185, F-15)22–80 yearsACSCSCLMIANot reported25.0215.14Munshi et al., 2019 [46]Mumbai, IndiaNot reported120 (M-91, F-29)25–70 yearsCADCCNot reportedNot reported15.1Not reportedBhargava et al., 2012 [47]New Delhi, IndiaNot reported3518 years or aboveCADCCCLIANot reported27.174.16Mahalle et al., 2013 [48]IndiaNot reported300 (M-216, F-84)25–92 YearsCADCSELISA28635.7516.17Jaswal et al., 2008 [49]India2 years100 (M-90, F-10)25–59 yearsACS &CADCCNot reported7634.14Not reportedKumar et al., 2011 [50]Aligarh, IndiaNot reported30 (M-30, F-0)18–44 yearsMINot reportedEIA2526.96.20Shukla et al., 2019 [51]June 2012—December 2014787 (M-736, F-51)18–39 yearsMINot reportedNot reported189Not reportedNot reportedGujarat, IndiaVishwajeet et al., 2018 [52]Northern IndiaJuly 2014—December 2015101 (M-97, F-4)18–40 yearsMICCEIANot reported22.4825.48Karim et al., 2015 [53]Dhaka, BangladeshJuly 2010—June 2011100 (M-78, F-22)18 years or aboveMICSNot reportedNot reported15.494.02Kumar et al., 2009 [54]New Delhi, IndiaSeptember 2004—March 2007368 (M-339, F-29)Median age 48–54 yearsCADCCHPLCNot reported14.9Not reportedKhan et al., 2000 [55]Karachi, PakistanNot reported31 (M-29, F-2)18–59 yearsMICCFPIANot reported14.971.13Iqbal et al., 2005 [56]Sindh, PakistanJanuary-June 2001, August–September 2003397 (M-309, F-88)30–74 yearsMICCFPIANot reported22.981.51Chaudhary et al., 2017 [57]Northern IndiaNot reported60 (M-52, F-8)18–40 yearsCADCSNot reported44Not reportedNot reportedBhandari et al., 2017 [58]Kolkata2013–20145022–35MICSNot reported21Not reportedNot reportedAlam et al., 2012 [59]Dhaka, BangladeshOctober 2009- September 201017621–75MICSFPIA12020.214.30Alom et al., 2016 [60]Dhaka, BangladeshJuly 2011- December 201160 (M-52, F-8)21–80ACSCCFPIA3219.516.30Muzaffar et al., 2019 [61]Lahore, PakistanNovember 2014- October 201510525–60IHDCSFPIA8322.39.22Bansal et al., 2021 [62]Amritsar, IndiaNot reported4030–80 yearsCADCCELISA2915.724.48Kumar et al., 2016 [63]North IndiaNot reported100(M-91, F-9)23–59 yearsACS &CADCCELISA7634.14Not reportedIjaz et al., 2015 [64]Lahore, PakistanOctober 2009—June 20103020–45 yearsMICCCLMIA2118.15.30Shanoli et al., 2017 [65]West Bengal, IndiaNot reported116 (M-94, F- 22)35–64 yearsCADCCNot reported5816.8810.84Muzaffar et al., 2021 [66]Lahore, PakistanNot reported105 (M-79, F-26)25–60 yearsCADCCFPIA8322.339.22Kabir et al., 2009 [67]Dhaka, BangladeshJanuary 2003—December 200350 (M-40, F-10)Median age 52–53 yearsIHDCCNot reported5018.262.32Arumalla et al., 2011 [68]Andhra Pradesh, IndiaNot reported5018–40 yearsMICSFPIA3319.696.94Begum et al., 2022 [69]Dhaka, BangladeshMarch 2021—February 20225218 years or aboveCADCSCLMIA3016.986.44Sainani et al., 2009 [70]IndiaNot reported6530–60 yearsCADCCEIA2221.252.20Afzal et al., 2011 [71]Lahore, PakistanSeptember 2009—August 201030 (M-20, F-10)18–80 yearsMICCEIA512.993.51Almas et al., 2020 [72]PakistanJuly 2017—January 201812035–80 yearsACSCSNot reported69Not reportedNot reportedSadhu et al., 2023 [73]Andhra Pradesh, IndiaJanuary 2022- December 202235 (M-29, F-6)18–45 yearsMICCNot reported1215.61.80Ullah et al., 2025 [74]AfghanistanMarch 2022- December 2023Pakistan 160 (M-102, F-58)25–80 yearsCADCSImmune-electrochemical assayNot reported21.67.30Afghanistan 91 (M-49, F-42)21.07.7Siripuram et al., 2025 [75]Telangana, IndiaJanuary 2021- December 2023245 (M-145, F-100)30–70 yearsCAD and MICSAutomated analyzerNot reported19.64.50Abbreviations: FPIA Fluorescence polarization immunoassay, EIA Enzyme immune assay, ELISA Enzyme-linked immune immunosorbent assay, HPLC High-performance liquid chromatography, CLIA Chemiluminescent immunoassay, CLMIA Chemiluminescent microparticle immunoassay, PM Photometric method, SPIA Solid-phase immune assay, CAD Coronary artery disease, MI Myocardial infarction, ACS Acute coronary syndrome, IHD Ischemic heart disease, CC Case control, CS Cross-sectional, Hcy Homocysteine, HHcy Hyperhomocysteinemia, M Male, F Female Summary of the selected studies Abbreviations: FPIA Fluorescence polarization immunoassay, EIA Enzyme immune assay, ELISA Enzyme-linked immune immunosorbent assay, HPLC High-performance liquid chromatography, CLIA Chemiluminescent immunoassay, CLMIA Chemiluminescent microparticle immunoassay, PM Photometric method, SPIA Solid-phase immune assay, CAD Coronary artery disease, MI Myocardial infarction, ACS Acute coronary syndrome, IHD Ischemic heart disease, CC Case control, CS Cross-sectional, Hcy Homocysteine, HHcy Hyperhomocysteinemia, M Male, F Female Evaluation of the quality of the selected articles and GRADE assessment Three of the coauthors (MNA, NMM, and RN) independently assessed the methodological quality of the selected study articles following the 9-item tool explicitly developed by the Joanna Briggs Institute [76, 77]. Afterward, two authors (NSD and BB) rechecked the procedure of the methodological quality assessment, and any discrepancies concerning the outcome of the quality assessment were resolved through discussion. On the basis of the critical appraisal tool scoring system, if the answer was “Yes”, then 1 was implied, and if the answer was “No” or “Unclear”, then 0 was implied. Total scores ranging from 0–5 were considered high risk of bias, scores ranging from 6–7 were considered moderate risk of bias, and scores ranging from 8–9 were considered low risk of bias (Table S2). The risk of bias was assessed for all the selected articles included in this study. Any disagreements during the assessment were resolved through discussion with other group members. Additionally, the certainty of evidence of the study outcomes (1. prevalence of HHcy among patients with heart disease, 2. standardized mean differences (SMD) between healthy individuals and patients with heart disease was assessed using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) framework. This involved evaluating risk of bias, inconsistency, indirectness, imprecision, and publication bias, with ratings of high, moderate, low, or very low assigned to each outcome [78, 79]. Statistical analyses The average concentration of Hcy and the prevalence of HHcy were considered summary measurements. According to the literature, the serum concentration of Hcy mainly depends on the activity of certain enzymes and vitamins, and their effects might vary depending on ethnicity [1, 2, 5]. Since the individual studies included in the final summary table (Table 1) did not report the effects of all of these variables, a random effects model was used to obtain the pooled prevalence of HHcy and pooled standardized mean differences, with 95% confidence intervals (CIs). Additionally, leave-one-out sensitivity analysis was performed to evaluate the influence of each study on the pooled estimate and heterogeneity. Heterogeneity was assessed via Cochran’s Q test and I2 statistics. Substantial heterogeneity was indicated with an I2 of more than 75% [80]. The 95% prediction intervals were calculated by using Tau2 statistics to reflect the expected range of effects in future comparable populations [10, 81]. Publication bias was also assessed graphically using a funnel plot and statistically through Egger’s test for the selected studies included in the meta-analysis [82, 83]. Additionally, Meta-regression analysis was also performed to understand the source of heterogeneity. Meta-analysis, leave-one-out sensitivity analysis, and subgroup analysis was performed using MetaXL (Microsoft Excel version 16). Meta-regression and Egger’s test was performed by Stata version 17.0 using the metareg, metabias, and metan commands. Results A total of 7,668 articles were retrieved from different databases following the initial search. After applying the eligibility criteria, 47 articles involving 6094 patients with heart disease were finally identified as eligible for this systematic review [23, 30–75] (Fig. 1). The study period for data collection in individual studies ranged from 6 months to 2.5 years; 16 out of 47 studies did not specify their study period [34, 35, 37, 42, 46–48, 50, 55, 57, 62, 63, 65, 66, 68, 70] (Table 1). Study characteristics As assessed by the eligibility criteria, more than half of the selected articles were published on patients from India (30 out of 47), followed by Pakistan (11 out of 47) and Bangladesh (5 out of 47), and contained data from 4062, 1,503 and 438 patients with heart disease, respectively. We found only one study from Afghanistan representing 91 patients with heart disease. Intriguingly, no study has been conducted in other SA countries, including Nepal, Bhutan, the Maldives, and Sri Lanka. The age range of the study participants ranged from 18–92 years. Only one study did not provide information about the age of the participants [36]. The sex distribution revealed a predominance of males, constituting more than 70% (4324 out of 6094) of the study population. Approximately one-fourth (12 out of 47) of the selected studies did not mention the gender of the study participants [33, 36, 47, 58, 59, 61, 62, 64, 68–70, 72]. Myocardial infarction was the most frequently reported (19 studies, comprising 2,784 patients) heart disease, followed by coronary artery disease (17 studies; comprising 1,973 patients), acute coronary syndrome (5 studies; comprising 490 patients), and ischemic heart disease (3 studies; comprising 253 patients). Only 4 studies reported patients who had multiple heart conditions [41, 49, 63, 75]. The study setting was hospital-based for almost all of the selected articles, except one, which was community-based [33]. The study design was case‒control for more than half of the selected articles (30 out of 47), and the rest were mostly cross-sectional studies. A few studies (4 out of 47) did not mention anything specific about the study design [36, 39, 50, 51]. Various analytical methods, such as fluorescence polarization immunoassay (FPIA), enzyme immunoassay (EIA), enzyme-linked immune immunosorbent assay (ELISA), high-performance liquid chromatography (HPLC), and chemiluminescent immunoassay (CLIA), have been used to determine the concentration of Hcy in the study population. FPIA (10 out of 47) and EIA (9 out of 47) were the methods most frequently used for Hcy measurement (Table 1). Nearly one-fourth of the selected studies (11 out of 47) did not mention the method of Hcy measurement [43, 46, 49, 51, 53, 57, 58, 65, 67, 72, 73]. Prevalence of HHcy More than half of the selected articles (26 out of 47) mentioned the number of participants with HHcy in their studies. This subset of the study population consisted of 3302 patients with heart disease. The overall pooled prevalence of HHcy was 61% (95% CI: 49–72%), with a significant level of heterogeneity (I2 = 98%, p < 0.01) (Fig. 2). Moreover, leave-one-out sensitivity analyses demonstrated small variations in the pooled prevalence values, ranging between 58 and 62% (95% CI: 48–74%) with a persistently high heterogeneity (I2 = 97–98%) (Table S7). The corresponding 95% prediction interval of the pooled prevalence was estimated between 0% −100%, indicating that the future studies are likely to produce considerable variability. The overall pooled prevalence of HHcy among SA patients with heart disease. Each horizontal line of the forest plot represents an individual study, and the box plot shows the prevalence for that study. The diamond at the bottom represents the overall prevalence when all the individual studies are combined together and averaged. The horizontal points of the diamonds represent the limit of the 95% confidence interval We segregated the 26 articles that mentioned the number of patients with HHcy on the basis of their study location to determine the country-specific prevalence. More than half of the selected articles (16 out of 26) were published on the Indian population, followed by the Pakistani population (6 out of 26) and the Bangladeshi population (4 out of 26). The studies conducted in India comprised 2396 patients with heart disease. The pooled prevalence of HHcy among patients from India was 57% (95% CI: 42–72%), with a high degree of heterogeneity (I2 = 98%, p < 0.01) (Fig. 3a). The overall pooled prevalence of HHcy among 518 Pakistani patients included in this study was 73% (95% CI: 45–94%), with a high degree of heterogeneity (I2 = 97%, p < 0.01) (Fig. 3b). Additionally, the total study population from Bangladesh included 388 patients with heart disease and the overall pooled prevalence of HHcy in this population was 58% (95% CI: 48–67%), with a high degree of heterogeneity (I2 = 97%, p < 0.01) (Fig. 3c). Pooled prevalence of HHcy among Indian (a), Pakistani (b) and Bangladeshi (c) patients with heart disease. Each horizontal line of the forest plot represents an individual study, and the box plot shows the prevalence for that study. The diamond at the bottom represents the overall prevalence when all the individual studies are combined and averaged. The horizontal points of the diamonds represent the limit of a 95% confidence interval The prevalence of HHcy was also analyzed based on study-design (Figure S1), method of Hcy measurement (Figure S2), type of heart diseases (Figure S3), risk of bias (Figure S4), and age of study participants (Figure S5). In terms of age, only 1 subgroup (45 years and below) consisting of 6 studies qualified to be included in the calculation of the prevalence (Figure S5), since it was not possible to subgroup study participants based on different age groups due to a large range of age used in most studies (Table 1). Importantly, all the subgroup analyses (Figure S1 to S5) demonstrated high level of heterogeneity. The results of these analyses are presented in Table 2. Table 2Summary of subgroup analysis for the prevalence of HHcyCategoryStudies includedSubgroup in each category (n)*Prevalence (95% CI)I2p-valueStudy design26 out of 26Case–control (13)67% (50%−82%)96% < 0.01Cross-sectional (9)65% (45%−82%)97% < 0.01Design Not reported (4)49% (30%−68%)96% < 0.01Type of heart disease20 out of 26CAD (10)68% (45%−88%)98% < 0.01MI (10)50% (36%−63%)96% < 0.01Method20 out of 26EIA (6)42% (29%−55%)91% < 0.01FPIA (6)79% (59%−93%)96% < 0.01Method Not reported (8)60% (37%−81%)98% < 0.01Excluding method not reported (18)65% (51%−78%)97% < 0.01Risk of Bias26 out of 26Low (4)68% (26%−100%)99% < 0.01High (7)62% (51%−73%)85% < 0.01Moderate (15)58% (41%−74%)98% < 0.01Age8 out of 2645 years and below (8)64% (32% −91%)99% < 0.01*n = number of studies included in each subgroup; subgroups containing less than 4 studies were not included in the analyses Summary of subgroup analysis for the prevalence of HHcy *n = number of studies included in each subgroup; subgroups containing less than 4 studies were not included in the analyses Measurement of serum Hcy More than three-fourths of the selected articles (37 out 47) reported the mean concentration of plasma Hcy among patients with heart disease. This subset of the study population comprised 4246 patients: 2284 (54%) from India, 1383 (33%) from Pakistan, 488 (12%) from Bangladesh and 91 (1%) from Afghanistan. The weighted mean concentration of Hcy was found to be 20.98 µmol/L (SD 4.57) considering all 37 of these articles. Country-wise analysis revealed that Pakistani patients had the highest weighted mean concentration of Hcy (21.78 µmol/L; SD 7.34), followed by India (19.78 µmol/L; SD 2.2) and Bangladesh (17.64 µmol/L; SD 2.35) (Figure S6). To compare the concentration of serum Hcy between patients with heart disease (cases) and healthy individuals (controls), the standardized mean difference (SMD) was calculated. Among the thirty case–control studies included in this systematic review, twenty-five studies comprising 1990 patients with heart disease (cases) and 1466 healthy individuals (controls) were qualified for the analyses for SMD (Fig. 4). The remaining 5 case‒control studies were not included, as either SD was not reported in these studies or serum mean level of Hcy was not properly segregated between case and control (Table S3). The overall SMD between cases and controls was 2.22 (95% CI: 1.62–2.82), with substantial heterogeneity (I2 = 98%, p < 0.01) (Fig. 4). The leave-one-out sensitivity analyses showed small differences in the pooled SMD values, ranging from 2.04 and 2.32 (95% CI: 1.47–2.93), with a persistently high heterogeneity (I2 = 97–98%) (Table S8). Moreover, the 95% prediction interval was approximately −0.93 to 5.37, indicating the likelihood of generation of considerable variability in the pooled SMD in future studies. Forest plot representing the overall SMD between patients with heart disease and normal healthy individuals. Each horizontal line of the forest plot represents an individual study, and the box plot shows the prevalence for that study. The diamond at the bottom represents the overall prevalence when all the individual studies are combined together and averaged. The horizontal points of the diamond represent the limit of the 95% confidence interval Next, to find the country-specific mean difference in the concentration of Hcy, the case‒control studies were distributed according to the study location. Among the 25 case‒control studies included in these analyses, 15 were conducted in India, 8 were in Pakistan, and 2 were in Bangladesh. Owing to the limited number of studies, data from study participants in Bangladesh were not included in this analysis. The study population of India included 1109 patients with heart disease and 933 healthy individuals. The weighted mean concentrations of Hcy in the cases and controls were 19.45 µmol/L (SD 2.41) and 10.99 µmol/L (SD 5.32), respectively. The overall SMD was 2.25 (95% CI: 1.44–3.05), with significant heterogeneity (I2 = 98%, p < 0.01) (Fig. 5a). Next, a total of 721 patients with heart disease and 423 healthy individuals were included in the study population of Pakistan. The weighted mean concentrations of serum Hcy in the cases and controls were 18.801 µmol/L (SD 2.50) and 10.77 µmol/L (SD 5.45), respectively. The overall SMD was 2.24 (95% CI: 1.11–3.36), with high heterogeneity (I2 = 98%, p < 0.01) (Fig. 5b). The overall SMD between Indian (a) and Pakistani (b) patients with heart disease and normal healthy individuals. Each horizontal line of the forest plot represents an individual study and the box plotted as prevalence for that study. The diamond at the bottom represents overall prevalence when all the individual studies are combined together and averaged. The horizontal points of the diamond represent the limit of 95% confidence interval Additionally, SMD was analyzed based on method of Hcy measurement, type of heart diseases, risk of bias, and age of study participants (Figure S7). Again, only 5 studies qualified to be included in the calculation of SMD based on the participants’ age (45 years and below) (Figure S7h). All the subgroup analyses (Figure S7) demonstrated high level of heterogeneity. The results of these analyses are presented in Table 3. Table 3Summary of subgroup analysis for SMD of Hcy concentration between heart patients (cases) and healthy individuals (controls)CategoryStudies includedSubgroup in each category (n)*SMD (95% CI)I2p-valueType of heart disease19 out of 25CAD (9)3.05 (1.85—4.25)98% < 0.01MI (10)2.09 (1.14—3.03)98% < 0.01Method16 out of 25EIA (6)1.56 (0.28—2.84)98% < 0.01FPIA (6)2.22 (1.14—3.31)98% < 0.01Method not mentioned (4)3.71 (0.88—6.54)99% < 0.01Excluding method not reported (21)1.96 (1.37–2.55)98% < 0.01Risk of Bias22 out of 25Low (4)2.25 (0.97–3.53)99% < 0.01Moderate (17)1.95 (1.26—2.64)98% < 0.01Age5 out of 2645 years and below (5)1.78 (0.53 −3.04)97% < 0.01*n = number of studies included in each subgroup; subgroups containing less than 4 studies were not included Summary of subgroup analysis for SMD of Hcy concentration between heart patients (cases) and healthy individuals (controls) *n = number of studies included in each subgroup; subgroups containing less than 4 studies were not included Source of heterogeneity The data of prevalence and SMD demonstrated a high heterogeneity across all the variables tested (study location, number of patients, study design, method of Hcy measurement, type of heart disease, study quality, mean age and percentage of male participants). To identify the sources of the heterogeneity, next we performed meta-regression analyses using several models with available covariates. The analyses were conducted considering two main outcomes: prevalence of HHcy among SA patients with heart diseases (26 studies), and SMD between healthy individuals and patients (25 studies). For the prevalence, among all the models, the model incorporating only standard error and patient number was the most significant (p = 0.0078), explaining 31.62% of variance (Table S4). In contrast, all models that tested the SMD were significant (p < 0.05), explaining more than 50% of variance in most cases. (Table S5). Across all the models, the model including standard error and patient number explained the largest proportion of variance (Adj R2 = 58.53%), with a strong overall model fit (p = 0.0001). Altogether, these findings indicate that study precision and sample size are the key contributors to heterogeneity. Quality assessment, publication bias, and GRADE assessment The risk of bias score was calculated for each study (Table S3) following the method described by the Joanna Briggs Institute in 2020 [76, 84]. Among the 47 studies selected, 11 had a low risk of bias, 27 had a moderate risk of bias, and the rest had a high risk of bias. Additionally, funnel plots (Figure S8) showing asymmetry and suggesting publication bias for the prevalence of HHcy and SMD. Furthermore, Egger’s test (p < 0.05) confirmed the presence of small-study effects and potential publication bias, indicating that smaller studies may overestimate both prevalence and SMD (Table S6). GRADE assessment rated the certainty of evidence as low (prevalence) to moderate (SMD) (Table 4). Table 4Summary of result using GRADE toolOutcomePooled Effect (95% CI)Heterogeneity (I2)Study DesignNumber of patients with heart disease (n)Certainty of Evidence (GRADE)Prevalence (26 studies)61% (49%–72%)98%Observational3,252⨁⨁◯◯ Low*SMD (25 studies)2.22 (1.63–2.81)98%Observational1,990⨁⨁⨁◯ Moderate***Evidence downgraded for inconsistency (I2 = 95.87%), publication bias (Egger’s p = 0.0000), and study design. No clear evidence of indirectness; whereas imprecision is unclear because of wide CI and variable study sizes** Evidence downgraded for inconsistency (I2 = 94.88%) and publication bias (Egger’s p = 0.0044) and study design. Upgraded for meta- regression analysis where covariates, particularly standard error, showing a strong explanatory factor of heterogeneity (Adj R2 > 54%) which indicate it could be an important source heterogeneity. No clear evidence of indirectness; whereas imprecision is unclear because of variable patients’ number with large CI Summary of result using GRADE tool *Evidence downgraded for inconsistency (I2 = 95.87%), publication bias (Egger’s p = 0.0000), and study design. No clear evidence of indirectness; whereas imprecision is unclear because of wide CI and variable study sizes ** Evidence downgraded for inconsistency (I2 = 94.88%) and publication bias (Egger’s p = 0.0044) and study design. Upgraded for meta- regression analysis where covariates, particularly standard error, showing a strong explanatory factor of heterogeneity (Adj R2 > 54%) which indicate it could be an important source heterogeneity. No clear evidence of indirectness; whereas imprecision is unclear because of variable patients’ number with large CI Discussion To the best of our knowledge, this is the first systematic review and meta-analysis that assessed the extent of HHcy in SA patients with heart disease. The overall prevalence of HHcy in SA patients with heart disease was observed 61% (95% CI: 49–72) (Fig. 2). These data are comparable to those of a study conducted on Chinese patients with coronary artery disease, in which the prevalence of HHcy was 62.08% [85]. Other studies in Chinese population have shown that 31–52% of patients with hypertension exhibited HHcy [86, 87]. In contrast, lower rates of prevalence have been reported in other ethnic groups; for instance, 45.3% of Romanians with cardiovascular metabolic syndrome [88] and 26% of Nigerian hypertensive subjects [89] have been documented to have HHcy. Collectively, these findings indicate that HHcy may be more common in SA patients than in patients from many other regions of the world, supporting the need for HHcy screening as part of cardiovascular risk assessment in this region. Although the exact mechanisms that explain the region-specific association of HHcy with cardiovascular disease is not well understood, several factors have been implicated. For instance, deficiencies in vitamins B6 and B12, and mutations in enzymes, involved in Hcy metabolism may play a crucial role. In a recent study in Nepal, homozygous mutation in gene encoding methylenetetrahydrofolate reductase (MTFR) was found in hypertensive individuals with an average homocysteine level of 19.4 µmol/L [90]. Similarly, HHcy was observed to be negatively associated with vitamin B12 and folate in Indian patients with hypertension [91]. Further interventional and molecular studies are warranted to validate the association of these factors with HHcy in SA population. Meanwhile, nutritional intervention such as dietary supplementations of vitamin B6, vitamin B12 and folate, and screening for polymorphisms in metabolic enzymes could be considered as preventive strategies in this region. The present study revealed that in SA patients, the weighted mean concentration of serum Hcy is 21.02 µmol/L (SD 4.76), which markedly exceeds the optimal threshold (< 15 µmol/L). A hospital-based study conducted approximately a decade ago in the UK reported that the average Hcy concentration among patients of SA ethnicity with heart failure was 14.4 µmol/L [92]. Moreover, evidence from a recent meta-analysis indicated elevated homocysteine concentrations among Indian and Pakistani patients with coronary artery disease with mean values ranging from 14.7 to 44.5 µmol/L—higher than those observed in other regional cohorts. [10]. This apparent increase in the serum Hcy concentration in SA patients might be associated with other comorbidities, genetic factors, and nutritional and metabolic conditions, since nutritional determinants, such as folate and B-vitamin status, genetic polymorphisms (particularly MTHFR C677T), regional differences in dietary habits, and methodological inconsistencies in homocysteine assays have all been reported to affect circulating homocysteine levels [93, 94]. Country-wise analyses revealed that Pakistani patients had the highest prevalence of HHcy (73%; 95% CI: 45–94%). In comparison, the prevalence rates of HHcy in Indian (57%; 95% CI: 42–72%) and Bangladeshi (58%; 95% CI: 48–67%) patients were relatively similar (Fig. 3a-3b). Similarly, the overall weighted mean concentration of Hcy was highest in Pakistani patients (21.78 µmol/L; SD: 7.34), followed by Indian (19.78 µmol/L; SD: 2.20) and Bangladeshi (17.64 µmol/L; SD: 2.35) patients (Figure S6). Many factors may contribute to this observed geographical disparity in HHcy. One factor could be a deficiency of vitamin B12, folic acid and other essential cofactors and enzymes that are involved in Hcy metabolism in Pakistani patients compared with Indian and Bangladeshi patients. Studies have shown that over 50% of Pakistani women of reproductive age are deficient in folic acid and vitamin B12 [95]. In comparison, 8% and 48% of Indian women [96] and nearly 9% and 23% of Bangladeshi women of reproductive age are folate- and vitamin B12-deficient, respectively [97]. In addition to the nutritional determinants of HHcy, the genetic background of patients may contribute to HHcy. For example, a common C677T mutation in the gene encoding the enzyme methyl-tetrahydrofolate reductase (MTHFR), which plays a crucial role in Hcy metabolism, has been found to lead to increased risk of HHcy [98]. Multiple reports have revealed the presence of the MTHFR C677T gene polymorphism among Indian, Pakistani and Bangladeshi populations [24, 99, 100]. While this systematic review and meta-analysis address a critical health concern among SA population, the study has certain limitations which should be considered when interpreting the findings. Substantial heterogeneity in the overall prevalence of HHcy and SMD in the concentrations of Hcy between cases and controls was observed (I2 = 98%, p < 0.01). A leave-one-out sensitivity analyses demonstrated that exclusion of any single study did not significantly alter the pooled prevalence or SMD estimates, indicating that the overall results were not driven by individual studies. However, 95% prediction intervals, highlight considerable between-study variability, limiting the direct generalizability of the pooled estimates across diverse populations and clinical settings. An extended meta-regression analysis by utilizing several models, identified study precision, associated with standard error and sample size, consistently explained a modest proportion of heterogeneity for both prevalence and SMD outcomes. In contrast, study-level characteristics such as study location, study design, assay method, study quality, type of heart disease, mean age, and sex distribution did not substantially reduce residual heterogeneity or improve model fit. Notably, additional sensitivity analyses were conducted by excluding studies that did not report homocysteine measurement methods: the pooled estimates remained comparable to the primary analyses, and the heterogeneity remained high (Table 2), indicating that measurement method alone is unlikely to be the primary driver of the observed variability. These findings suggest that imprecision in individual study estimates, potentially arising from small sample sizes or measurement inaccuracy, rather than any additional covariates, may contribute to between-study variability. Such patterns align with established methodological principles describing how study precision and sample size influences heterogeneity in meta-analyses [101]. Additionally, cautions should be exercised before generalizing the findings of this study since several methodological and study-specific limitations further constrain the interpretations. First, although the study aimed to represent the entire SA populations, eligible studies were not available from Nepal, Bhutan, Sri Lanka, or the Maldives and this may not have reflected the full view of the region. Second, only studies that used cutoff values of HHcy ≥ 15 µmol/L were included and therefore other studies that used different threshold of HHcy may have excluded from the analyses. Third, detailed age- and sex-stratified data were inconsistently reported, restricting subgroup analyses. Fourth, key biological determinants of homocysteine metabolism—such as nutritional status (folate, vitamin B₆ and B₁₂ intake), dietary patterns, and genetic polymorphisms (e.g., MTHFR variants), were rarely reported and therefore could not be incorporated into the analyses. These unmeasured factors may have contributed substantially to the observed residual heterogeneity. Fifth, the use of different methods for measuring serum Hcy concentrations across different studies could introduce assay bias to our findings. Sixth, some included studies lacked essential particulars such as study period, standard deviations, or demographic details, whereas others reported unusually high values of standard deviation; these factors may have affected the pooled estimates. Seventh, most included studies reported only a single outcome measure (either prevalence of hyperhomocysteinemia or mean serum homocysteine levels), and only a very small subset reported both outcomes within the same study. In addition, the distribution of studies across countries was highly uneven, with the majority originating from a single country and only a few studies from others, which limit multivariate or hierarchical meta-analytical approaches. However, these limitations are common in most of the similar meta-analyses and were therefore unavoidable. Finally, the findings indicate that while the pooled estimates are robust, substantial unexplained heterogeneity remains, limiting direct extrapolation of the data to all South Asian populations. Future studies should prioritize standardized reporting of demographic, nutritional, genetic, and methodological variables and consider multivariate or hierarchical meta-analytic approaches to better account for correlated sources of variability and country-level effects. Conclusion This study reports a high prevalence of HHcy among SA patients with heart disease, with six out of ten individuals exhibiting this condition. The average concentration of Hcy among patients with heart disease exceeds the normal reference range. Pakistani patients presented the highest concentration of Hcy, followed by Indian and Bangladeshi patients. While the high degree of heterogeneity observed in the pooled prevalence and SMD data may limit the generalizability of the findings of this study, this limitation suggests some important recommendations for future studies. For instance, cautions must follow to reduce the imprecision in individual study; complete sets of data of different covariates must be used; factors, including age, sex, lifestyle, dietary habits, and genetic polymorphisms present in the study population should ideally be included in analyses- all these measures will increase the reliability of the pooled estimates. Additionally, as part of the comprehensive strategies used to manage cardiac health in this region, screening of HHcy among at-risk population, and preventive and intervention-based measures addressing HHcy should be performed. Supplementary Information Supplementary Material 1. Supplementary Material 2. Acknowledgements Not applicable. Authors’ contributions Conceptualization and Supervision – N.S.D.; Methodology – M.N.A., N.M.M., R.N. and B.B.; Review & Editing Review & Editing – N.S.D., B.B., H.H., N.A. and M.H.S.; Original Draft – M.N.A., N.M.M. R.N. and, B.B; Data Extraction and Analysis—M.N.A., N.M.M., R.N. and B.B; Revision: B.B., N.M.M., and N.S.D. Funding Not applicable. Data availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Badhan Bhattacharjee, Email: badhan.bhattacharjee@g.bracu.ac.bd. Nadia S. Deen, Email: nadia.sultana@bracu.ac.bd References Associated Data Supplementary Materials Supplementary Material 1. Supplementary Material 2. Data Availability Statement The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.