Variability in the Preoperative Rheumatoid Arthritis Medication Management in Hand Surgery Authors: Luna Toma, Ethan Cossu, Leah Demetri, Brandon E Earp, Dafang Zhang, Philip E Blazar Published: 2026 ABSTRACT AbstractPurposePreoperative management of biological disease modifying antirheumatic drugs (bDMARDs) in patients with rheumatoid arthritis (RA) undergoing hand surgery is a subject of ongoing debate. While consensus supports continuation of corticosteroids and conventional synthetic DMARDs, no definitive guidelines exist for bDMARDs. This study aimed to characterize the degree of perioperative medication management variability among patients with RA undergoing hand surgery and to examine whether continuation of bDMARD therapy is associated with an increased risk of postoperative surgical site infection.MethodsA retrospective chart review was conducted at a single academic institution of patients with RA who underwent hand surgery between June 2015 and May 2020. Perioperative medication continuation rates were calculated for corticosteroids, conventional synthetic DMARDs, and bDMARDs.ResultsOf 171 patients included, 157 (92%) were on pharmacological therapy at the time of surgery. Continuation rates were 82% for corticosteroids, 73% for conventional synthetic DMARDs, and 70% for bDMARDs. Six patients (3.5%) developed a surgical site infection within 30 days.ConclusionsPerioperative bDMARD management in patients with RA undergoing hand surgery is variable and lacks standardization. Prospective, multicenter studies are needed to establish evidence-based guidelines for this population.Study typeRetrospective cohort study.Type of study/level of evidenceTherapeutic III.Key words: Immunosuppression, Perioperative risk, Postoperative infection, Rheumatoid arthritis surgery timing, Steroids The treatment of rheumatoid arthritis (RA) has advanced over the decades with the advent of disease modifying antirheumatic drugs (DMARDs), thereby reducing the need for surgical intervention in the hand and wrist.1,2 Current pharmacological management includes nonsteroidal anti-inflammatory drugs, corticosteroids, and DMARDs.3 However, because these medications modulate immune function, clinicians have raised concerns about their preoperative use and the potential for postoperative complications, making perioperative medication management an important clinical consideration. For corticosteroids and conventional synthetic DMARDs, more than two decades of evidence, including data specific to hand surgery, has yielded a consistent consensus that continuation throughout the perioperative period is recommended.4, 5, 6, 7, 8, 9, 10, 11, 12 The management of biological DMARDs (bDMARDs) and the targeted synthetic DMARD tofacitinib, however, remains a subject of ongoing debate that includes balancing the risk of disease flares and accelerated joint deterioration upon drug discontinuation against the potential for surgical site infection (SSI) or wound dehiscence if therapy is maintained.13, 14, 15 This risk evaluation is further complicated by the elevated infection susceptibility observed in patients with RA.16 Current literature reflects this unresolved tension: several studies advocate for preoperative discontinuation of bDMARDs citing elevated infection rates,6,9,12,17,18 while others have found an increase in disease flares when the medication was discontinued.8,10,11,13,14,19, 20, 21, 22, 23 Consequently, no definitive guidelines have emerged. The present study pursues two objectives. First, we characterize the degree of real-world perioperative management variability among patients with RA undergoing surgery at our institution. Second, we examine whether continuation of bDMARD therapy in this cohort is associated with an increased risk of postoperative infection. Materials and Methods This study was approved by the Institutional Review Board and conducted as a retrospective chart review at a single academic institution. The initial database query identified patients seen between June 1, 2015, and May 15, 2020, using three parameters: International Classification of Diseases-10 diagnosis codes for RA, a panel of relevant hand surgery Current Procedural Terminology codes, and a list of medications being actively used in RA treatment at the time of surgical booking, which together generated a confidence score for the likelihood of a true RA diagnosis. This query yielded 3,635 patients, who were subsequently cross-matched against an internal research database from our rheumatology department of patients likely to have a true RA diagnosis. After crossmatching with the rheumatology database, only patients with a RA likelihood of at least 0.99 were retained, yielding 328 patients.24 Of these 328 patients initially identified, 171 were excluded for the following reasons: no active RA diagnosis (n = 79), osteoarthritis (n = 24), non-RA inflammatory arthritis (n = 20), no pharmacological RA treatment (n = 14), psoriatic arthritis (n = 8), fibromyalgia (n = 7), polymyalgia rheumatica (n = 7), systemic lupus erythematosus (n = 4), scleroderma (n = 2), other inflammatory conditions (n = 2), and no documented surgical procedure (n = 4). A final cohort of 157 patients was included in the analysis. Baseline demographic and clinical characteristics were collected (Table 1). For the first study aim, medication data were extracted for all RA-related pharmacological agents. For each agent, the date of the last preoperative dose and the date of surgery were recorded, and the interval between them was calculated. Perioperative medication management was classified as follows: Corticosteroids were considered discontinued if the patient had not taken a dose for more than 1 day before surgery. For DMARDs, discontinuation was defined as the omission of at least one full dosing interval before the procedure. For each pharmacological class, the proportion of patients in whom therapy was continued versus discontinued was calculated. The primary outcome for the second study aim was the occurrence of an SSI within 30 days of the procedure. Table 1Patient Demographics and Surgical Characteristics∗Patient Demographics and Surgical CharacteristicsTotalPatients171Age (mean ± SD)62 ± 11Female sex141 (82)White ethnicity143 (84)Current or past smoker82 (48)Diabetes mellitus23 (13)Performed procedureSurgery on right side96 (56)Carpal Tunnel Release69 (40)Trigger Finger Release35 (20)∗All values are in n (%), unless otherwise indicated. Patient Demographics and Surgical Characteristics∗ All values are in n (%), unless otherwise indicated. The data sets generated during and/or analyzed during the current study are available from the corresponding author on request. Results Patient demographics and surgical characteristics are presented in the Table 1. Of the 157 patients included in the study, DMARDs were the most commonly prescribed medication class (59%), followed by nonsteroidal anti-inflammatory drugs (40%) and systemic corticosteroids (33%). Among DMARD users, the majority were on sDMARDs (67%), with methotrexate and hydroxychloroquine being the most frequently used agents. Of bDMARDs (33%), adalimumab (Humira) was the most commonly prescribed agent, accounting for half of all bDMARDs (Table 2). The most commonly performed procedures were carpal tunnel release (n = 69; 40%), followed by trigger finger release (n = 35; 20%). Surgery was performed on the right hand in 56% cases (n = 96). Table 2Pharmacological Rheumatoid Arthritis Treatment at Time of Surgery∗Medicationn (%)Patients on pharmacological treatment157 (92)Medication class (% is that of the patients on pharmacological treatment)DMARDs92 (59)Nonsteroidal anti-inflammatory drugs68 (40)Systemic corticosteroids56 (33)Nonbiological DMARDs (n = 62, 67% of DMARDs, % is that of the patients on DMARDs)Methotrexate31 (29)Hydroxychloroquine31 (29)Leflunomide8 (13)Tofacitinib (Xeljanz)2 (3)Sulfasalazine1 (2)Biological DMARDs (n = 30, 33% of DMARDs, % is that of the patients on DMARDs)Adalimumab (Humira)15 (50)Etanercept (Enbrel)5 (17)Tocilizumab (Actemra)5 (17)Infliximab (Remicade)2 (7)Certolizumab1 (3)Rituximab (Rituxan)1 (3)Golimumab (Simponi)1 (3)∗Some patients received more than one medication within and across drug classes. Pharmacological Rheumatoid Arthritis Treatment at Time of Surgery∗ Some patients received more than one medication within and across drug classes. Of the 157 patients, 36 patients (23%) discontinued their medication regimen preoperativley. Among patients on corticosteroids, 46 of 56 patients (82%) continued their medication without interruption. Among sDMARD users, 45 of 62 patients (73%) also continued their medication. Among bDMARD users, 21 of 30 patients (70%) continued their medication. Within this group, management varied by agent: Nine of 15 patients (60%) on adalimumab (Humira), four of five (80%) on both etanercept (Enbrel) and tocilizumab (Actemra), and one of two (50%) on infliximab (Remicade) continued the medication before surgery. Patients on certolizumab, rituximab, or golimumab (n = 1 each) continued their medication before surgery. Among sDMARDs, 19 patients on methotrexate (61%), 6 on leflunomide (75%), and 29 on hydroxychloroquine (94%) continued their medication. Of the two patients on tofacitinib (Xeljanz), one (50%) continued their medication. Six patients (3.8%) developed an SSI within 30 days of the procedure. All six were women and white, five (83%) were active smokers, and one (17%) had a diagnosis of diabetes mellitus. At the time of surgical booking, three were taking nonsteroidal anti-inflammatory drugs, and three were on a bDMARD, one of whom was also concurrently taking methotrexate. No patients in the infection group were taking systemic corticosteroids. Only one patient of those taking bDMARDs (33%) in this group withheld medication (adalimumab) before surgery. In comparison, in the cohort of patients without infection (151; 29 on bDMARD), 8 (28%) patients withheld bDMARD medication. No statistically significant difference was observed in continuation rates between the infection group and the no-infection group (P = .45). Discussion This retrospective cohort study examined perioperative medication management among patients with RA undergoing hand surgery, characterizing real-world management variability and evaluating whether bDMARD continuation is associated with increased postoperative infection risk. Our findings reflect the broader uncertainty in the literature and offer institution-level insight into how clinicians navigate this unresolved clinical question. The high rates of medication continuation observed across all drug classes suggest that the default clinical posture leans toward maintaining pharmacological therapy through the perioperative period. Corticosteroids and sDMARDs were continued at rates of 82% and 73%, respectively, consistent with, though not fully adherent to, the clear recommendation for uninterrupted perioperative use of these agents.4, 5, 6, 7, 8, 9, 10, 11, 12 The minority that withheld medication may represent individual clinical exceptions, such as concerns about wound healing in specific surgical contexts, or simply gaps in protocol adherence. For bDMARDs, the continuation rate of 70% shows that in the absence of definitive guidelines, clinicians were more inclined to continue these agents before surgery. The small sample sizes within each drug subgroup precludes meaningful comparative conclusions around the variation in continuation, ranging from 50% for infliximab to 80% for etanercept and tocilizumab. The clinical concern, whether bDMARD continuation confers an increased risk of postoperative SSI, was not supported by the data in this cohort. Six patients (3.8%) developed an SSI within 30 days of surgery, and no statistically significant difference in bDMARD continuation rates was identified between the infection and noninfection groups. This finding is concordant with literature challenging the assumption that perioperative bDMARD continuation independently elevates infection risk in hand surgery, and supports the position that discontinuation may not be necessary to achieve acceptable surgical outcomes.11,20,21 Nevertheless, with only six infection events in total, our sample size is too small to draw reliable conclusions about the presence or absence of a true association. A notable methodological strength of this study was the use of a validated rheumatology database that assigned a probabilistic confidence score for the likelihood of a true RA diagnosis. By crossmatching the initial patient cohort against this database and retaining only those with a matching probability of at least .99, we were able to minimize diagnostic misclassification. However, this may have introduced selection bias by excluding patients with a legitimate RA diagnosis who were not captured in the rheumatology database. Berthold et al20 conducted an observational single-center study of 1,596 elective orthopedic and hand surgery procedures performed on 892 patients with inflammatory rheumatic disease at Lund University Hospital between 2003 and 2009. Their study design offered a comparison between two sequential institutional practice eras: group A (2003–2005), in which tumor necrosis factor inhibitors were discontinued before surgery, and group B (2006–2009), in which tumor necrosis factor inhibitors were continued.20 The SSI rates were 3.0% in group A and 5.3% in group B, though this difference did not reach statistical significance.20 Our SSI rate is similar to that observed across both of their groups. Notably, the higher raw infection rate in group B was driven largely by foot surgery, the only subgroup in which the difference reached significance, while hand surgery rates remained comparable between eras.20 The findings of the present study are supported by the recent work of Cathey et al,21 who examined perioperative bDMARD management across 7,929 patients with RA undergoing elective hand surgery using the TriNetX Research Network. Cathey et al21 report continuation rates only for bDMARDs and targeted synthetic DMARDs, precluding comparison for corticosteroids and conventional synthetic DMARDs. Among bDMARD users, their continuation rate of 34% stands in contrast to the 70% observed in our cohort. This difference likely reflects both institutional and temporal factors, as their data set spans 2006−2025. With respect to outcomes, Cathey et al21 found that estimated risks of SSIs and wound morbidity did not notably differ between patients who continued versus discontinued bDMARDs perioperatively, a finding consistent with our own. However, their study additionally demonstrated that bDMARD continuation was associated with a 50% lower estimated risk of 30-day postoperative disease flares.21 The two studies are complementary: the scale of the TriNetX data set confers substantial statistical power and broad generalizability, while our study offers individual-level granularity. Several limitations of this study must be acknowledged. The retrospective, single-institution design limits generalizability. Confounding variables including disease severity, perioperative antibiotic use, type of surgical procedure, operative time and surgeon-specific factors were not controlled in this investigation. The absence of data on postoperative disease flares represents a notable gap, as flare rates are central to the risk-benefit calculus of perioperative bDMARD management. Additionally, the classification of perioperative discontinuation based on dosing interval omission may not fully capture the pharmacodynamic variability across agents with differing half-lives and mechanisms. Conclusions The findings of this study reinforce the conclusion that perioperative bDMARD management in patients with RA undergoing hand surgery remains highly variable and lacks standardization. Prospective, multicenter studies are needed to establish the evidence necessary for definitive guideline development in this population. Conflicts of Interest No benefits in any form have been received or will be received related directly to this article. References