Lower Extremity Injuries and Revision Surgery Rates are Increased in Patients With Anorexia Nervosa and Bulimia Nervosa Compared With Controls Authors: Jackson G Woodrow, David G Deckey, Eugenia Lin, Jack M Haglin, Joshua S Bingham, Karan A Patel, Callie M Davies, Steven K Poon Published: 2026 ABSTRACT AbstractPurposeTo compare lower extremity soft tissue injury and revision surgery rates for patients with anorexia nervosa (AN) and bulimia nervosa (BN) compared with matched controls.MethodsThis retrospective cohort study used a deidentified claims database. Patients were included if they carried an International Classification of Diseases code for AN or BN. Respective matched cohorts for AN and BN were created using propensity score matching. Primary and revision surgeries for anterior cruciate ligament (ACL) reconstruction, posterior cruciate ligament reconstruction, knee collateral ligament reconstruction, meniscus repair or meniscectomy, and any ankle ligament reconstruction were identified by their CPT codes. Patients over the age of 40 and without a minimum 2‐year follow‐up after primary surgery were excluded. Differences in prevalence and revision rates were analyzed.ResultsA total of 56,146 patients with AN and 42,630 patients with BN were included in this study with equal numbers in respective matched cohorts. Patients with BN showed an increased risk of ACL injury (relative risk = 1.32, P = .019), medial collateral ligament injury (1.17, 0.04), calcaneofibular sprain (1.92, <0.001), deltoid sprain (1.41, 0.004), and unspecified ankle sprain (1.44, <0.001). Patients with AN showed an increased risk for calcaneofibular sprain (1.25, 0.007), deltoid sprain (1.28, 0.015), and unspecified ankle sprain (1.17, <0.001). Patients with BN were more likely to have revision ACL repair (1.89, 0.02), partial meniscus revision (3.44, <0.001) and any meniscus revision (4.74, <0.001) while patients with AN were more likely to have an ACL revision (2.81, <0.001), partial meniscus (4.37, <0.001), and any meniscus (4.52, <0.001).ConclusionsIn this study, patients with AN and BN had a significantly increased risk for various knee and ankle injuries and revision surgeries compared with matched controls. Additionally, patients with AN and BN who sustained these injuries were more likely to undergo a revision surgical procedure within the first 5 years of the primary procedure.Level of EvidenceLevel III, retrospective comparative study. Anorexia nervosa (AN) and bulimia nervosa (BN) are 2 common eating disorders that can affect anyone at any stage of their life, but are most commonly diagnosed in young women, with up to 30% of college‐aged women being affected. 1 AN is defined as a disorder of persistent behaviors causing restrictive energy or caloric intake relative to requirements and physiologic low body mass index (BMI), with additional psychological concerns including intense fear of weight gain or disturbance in one's body weight. 2 BN is classified as a separate disorder by which a person has persistent behaviors including recurrent episodes of binge eating (i.e., eating larger quantities of food than that which is larger than the average individual), with additional inappropriate compensatory behaviors to prevent weight gain including self‐induced vomiting, misuse of laxatives, diuretics, or other medications, fasting, or excessive exercise. 2 However, because of these compensatory behaviors, BN patients are diagnosed through a thorough medical history and not solely based on their BMI. 3 Because of the underlying total nutrient deficiency and endocrine disruption in AN and the inconsistent intake in BN, these diseases have substantial impact on the physical health of the patient. Up to 50% of adolescent girls and 70% of adolescent boys with AN had abnormal scores dual energy X‐ray absorptiometry, increasing the risk for developing osteoporosis, osteopenia, and pathologic fractures later in life. 4 AN has also been shown to decrease insulin‐like growth factor‐1 (IGF‐1) levels despite high growth hormone levels. Furthermore, the common compensatory mechanism of increased exercise to decrease net daily caloric intake can increase the risk of soft tissue injury in an already susceptible population. 5 , 6 The current literature has shown that both AN and BN increase the risks of developing orthopaedic bone and soft tissue injuries, but little is known about how patients recover from those initial injuries. 6 The purpose of this study was to compare lower extremity soft tissue injury and revision surgery rates for patients with AN and BN compared with matched controls. We hypothesized that patients with AN and BN are at an increased risk for both primary musculoskeletal injury and the need for revision surgeries compared with matched controls. METHODS This study used a matched cohort analysis through the PearlDiver M165 (PearlDiver, Colorado Springs, CO) Mariner database, a large deidentified national insurance claims database. The International Classification of Diseases, 9th and 10th Revision (ICD‐9 and ICD‐10), codes for AN (ICD‐9‐3071, ICD‐10‐F5000, ICD‐10‐F5001, and ICD‐10‐F5002) and BN (ICD‐9‐30751 and ICD‐10‐F502) were used to create cohorts with eating disorders. Two control populations, one of patients without AN and one of patients without BN, were then created and matched one‐to‐one based on age, Charlson Comorbidity Index (CCI), sex, and geographical region. While it is known that BMI plays a critical role in outcomes following a variety of orthopaedic procedures, we could not inherently control for this in our matching and final statistical analysis since patients with anorexia, by definition, will have a lower BMI than their matched controls. Matching based on CCI was chosen to avoid confounding by overall medical complexity, as eating disorder patients often have multiple medical comorbidities that could independently affect surgical outcomes. All groups were then filtered for patients less than 40 years old. Since no identifiable patient health information was queried, institutional review board approval was not required for this study. Ligamentous and meniscal injuries were then identified in the different groups through ICD codes for meniscus tear, anterior cruciate ligament (ACL) injuries, posterior cruciate ligament injuries, medial collateral ligament (MCL) injuries, lateral collateral ligament injuries, calcaneofibular sprain, ankle deltoid ligament injury, tibiofibular ligament injury, and unspecified ankle injury. All ligamentous injuries, including tears, were identified using ICD codes for “sprain” of the respective ligaments as there are no specific ICD codes designated for “tear” of such ligaments. Surgical interventions for ACL reconstruction, posterior cruciate ligament reconstruction, knee collateral ligament reconstruction, meniscus repair or meniscectomy, and ankle ligament reconstruction were identified by their respective Current Procedural Terminology (CPT) codes. Revision surgeries were identified using laterality filtering and a subsequent identical CPT code within 5 years of the initial date of surgery. Cases with multiligamentous or concomitant injuries were excluded to reduce variability in mechanism of injury. A full list of ICD and CPT codes used to query outcomes of interest is available in Tables S1 and S2. The prevalence of ligamentous and meniscal injuries, primary surgical intervention, and revision surgeries in the experimental and control cohorts was analyzed using χ2 analysis with a significance of P < .05. A risk ratio with a 95% confidence interval was also calculated for each injury, primary surgery, and revision surgery. All statistical analyses were performed with the PearlDiver integrated R statistical software (R Project for Statistical Computing, Vienna, Austria). RESULTS A total of 56,146 patients with AN and 42,630 patients with BN were included in this study with equal numbers in their respective control cohorts. Patients in the AN or BN and their respective control cohorts had similar distributions with regard to age, CCI scores, and sexes. A full description of patient demographics is presented in Table 1. TABLE 1Demographic Characteristics of Anorexia Nervosa and Bulimia Nervosa Matched CohortsParameterAnorexia NervosaAnorexia Nervosa Matched ControlBulimia NervosaBulimia Nervosa Matched ControlAge ± standard deviation21.2 ± 8.221.2 ± 8.224.1 ± 8.924.1 ± 8.9Body Mass Index16.6 ± 2.323.6 ± 4.123.7 ± 3.924.3 ± 3.8Female n, (%)48,581 (86.5%)48,581 (86.5%)38,120 (89.4%)38,120 (89.4%)Male n, (%)7565 (13.5%)7565 (13.5%)4510 (10.6%)4510 (10.6%)Charlson Comorbidity Index ± standard deviation0.36 ± 0.710.12 ± 0.410.39 ± 0.780.14 ± 0.47 Demographic Characteristics of Anorexia Nervosa and Bulimia Nervosa Matched Cohorts Patients with AN showed an increased risk for calcaneofibular sprain (relative risk [RR] 1.25, 95% CI 1.06‐1.46, P = .007), deltoid sprain (RR 1.28, 95% CI 1.05‐1.55, P = .015), and unspecified ankle sprain (RR 1.17, 95% CI 1.09‐1.25, P < .001) (Figure 1A), and an increased risk of a revision ACL reconstruction (RR 2.81, 95% CI 1.63‐4.85, P < .001), revision partial meniscectomy (RR 4.37, 95% CI 2.53‐7.57, P < .001), and any revision meniscus procedure (RR 4.52, 95% CI 2.85‐7.17, P < .001) (Figure 1C). A full analysis of injuries, primary surgeries, and revision surgeries can be found in Tables 2 and 3. Patients with BN showed an increased risk of sustaining an ACL tear (RR = 1.32, 95% CI 1.11‐1.57, P = .019), MCL injury (RR 1.17, 95% CI 1.01‐1.34, P = .04), calcaneofibular sprain (RR 1.92, 95% CI 1.59‐2.32, P < .001), deltoid sprain (RR1.41, 95% CI 1.12‐1.76, P = .004), and unspecified ankle sprain (RR 1.44, 95% CI 1.30‐1.60, P < .001) compared with their matched cohort (Figure 1B). These patients had an increased risk for a revision ACL reconstruction (RR 1.89, 95% CI 1.105‐3.229, P = .02), revision partial meniscectomy (RR 3.44, 95% CI 2.06‐5.76, P < .001), and any revision meniscus procedure (RR 4.74, 95% CI 2.63‐8.53, P < .001) (Figure 1D). Relative risk for injuries and revision surgeries. (A) Relative risk for injuries in patients with AN compared with matched controls. (B) Relative risk for injuries in patients with BN compared with matched controls. (C) Relative risk for revision surgeries in patients with AN compared with matched controls. (D) Relative risk for revision surgeries in patients with BN compared with matched controls. Boldface indicates statistical significance of P < .05.. (AN, anorexia nervosa; BN, bulimia nervosa.) TABLE 2Odds Ratio of Primary Injuries in Patients With Anorexia and Bulimia Compared With Matched CohortsInjuryOdds Ratio (95% Confidence Interval)AnorexiaBulimiaACL sprain0.89 (0.77‐1.04)1.32 (1.11‐1.58)∗ PCL sprain0.47 (0.26‐0.87)∗∗ 1.06 (0.55‐2.01)MCL sprain0.95 (0.83‐1.08)1.17 (1.01‐1.36)∗ LCL sprain1.04 (0.83‐1.30)1.20 (0.93‐1.56)Knee cruciate ligament sprain0.80 (0.72‐0.89)∗∗ 1.30 (1.15‐1.47)∗∗ Calcaneofibular sprain1.25 (1.06‐1.46)∗ 1.92 (1.59‐2.32)∗∗∗ Deltoid sprain1.28 (1.05‐1.55)∗ 1.41 (1.12‐1.76)∗∗ Meniscus tear0.78 (0.72‐0.84)∗∗∗ 1.05 (0.97‐1.15)Tibiofibular ligament sprain1.02 (0.86‐1.20)1.22 (1.00‐1.48)∗ Unspecified ankle sprain1.16 (1.06‐1.27)∗ 1.44 (1.30‐1.60)∗∗∗ Any ankle sprain1.17 (1.09‐1.25)∗∗∗ 1.46 (1.34‐1.59)∗∗∗ ACL, anterior cruciate ligament; LCL, lateral collateral ligament; MCL, medial collateral ligament; PCL, posterior cruciate ligament.∗ P < .05.∗∗ P < .01.∗∗∗ P < .001. Odds Ratio of Primary Injuries in Patients With Anorexia and Bulimia Compared With Matched Cohorts ACL, anterior cruciate ligament; LCL, lateral collateral ligament; MCL, medial collateral ligament; PCL, posterior cruciate ligament. P < .05. P < .01. P < .001. TABLE 3Odds Ratio of Revision Surgeries in Patients With Anorexia and Bulimia Compared With Matched CohortsInjuryOdds Ratio (95% Confidence Interval)AnorexiaBulimiaAny meniscus revision4.52 (2.85‐7.17)∗∗∗ 4.47 (2.63‐8.53)∗∗∗ Partial meniscectomy revision4.37 (2.53‐7.57)∗∗∗ 3.44 (2.06‐5.76)∗∗∗ Revision partial meniscectomy4.38 (2.53‐7.57)∗∗∗ 3.44 (2.06‐5.76)∗∗∗ Any ankle revision ligament repair1.19 (0.87‐1.63)1.74 (0.94‐3.22)Revision ACL reconstruction2.81 (1.63‐4.85)∗∗∗ 1.89 (1.11‐3.23)∗ ACL, anterior cruciate ligament.∗ P < .05.∗∗∗ P < .001. Odds Ratio of Revision Surgeries in Patients With Anorexia and Bulimia Compared With Matched Cohorts ACL, anterior cruciate ligament. P < .05. P < .001. DISCUSSION Our study showed that patients with AN and BN were associated with increased risk of knee meniscal and ligamentous injuries and ankle ligamentous injuries compared with a matched cohort. Additionally, both cohorts were associated with experiencing higher rates of revision knee surgeries. Compared with their respective matched cohorts, patients with AN had an increased risk for calcaneofibular sprain, deltoid sprain, other ankle sprains, while patients with BN had an increased risk of sustaining ACL tears, MCL injuries, calcaneofibular sprains, deltoid sprains, and other ankle sprains. Patients with AN experienced a 3‐fold higher risk for ACL revision and 4.5‐fold higher risk for revision meniscal surgery compared with their matched cohort, and patients with BN had an increased risk for revision ACL surgery and revision meniscal surgery. This study correlates eating disorders with specific revision rates in orthopaedic surgeries to better inform patients about potential re‐injury risk and revision surgery risk compared with a population without AN or BN. The cause for these increased risks is likely multifactorial. IGF‐1, a key growth factor, plays an important role in both the formation of bone during adolescence and the maintenance of bone and connective tissue strength through adulthood. 7 In individuals with AN, chronic malnutrition has been shown to decrease IGF‐1 levels, resulting in decreased bone mineral density (BMD), which elevates the risk of injuries. 8 In animal models, IGF‐1 has also been shown to accelerate the healing process following tendon and ligament injuries, and its deficiency in AN could be responsible for the increased need for revision surgeries. 9 , 10 In BN, although BMD is usually preserved, the compensatory behaviors of purging and laxative abuse can lead to electrolyte, micronutrient, and macronutrient imbalances. Patients with BN have shown decreased intake of vitamins C and D and protein with excessive intake of sodium, fats, and carbohydrates. 11 These deficiencies, particularly vitamin C and protein, can impair collagen synthesis and tissue healing, potentially explaining the increased injury and revision surgery rates observed. 11 Furthermore, excessive exercise as a compensatory mechanism may predispose patients to overuse injuries and poor healing in an already vulnerable population. 12 , 13 , 14 Based on the current understanding of AN's effect on ligament quality, our study found that patients with AN were not at an increased risk of injuring their ACL (OR 0.89). 15 However, for those who did suffer an ACL injury, they were almost 3 times more likely to require a revision ACL surgery compared with a matched control cohort. The lack of statistically significant difference in initial ACL injuries is further supported by the Schmerler et al. database study which found the relative risk of sustaining an ACL injury in a patient with AN was 0.88 compared with a matched cohort (P = .855), with our patient population having an odds ratio of 0.989. 6 Additionally, the study by Schmerler et al. found the relative risk of requiring an ACL primary reconstruction to be similar at 2.83 (P = .037) compared with 2.81 in our study. 6 Other studies report subsequent revision rates for ACL reconstructions as high as 8%, though how this rate varies across demographics remains unclear. 16 However, multiple other risk factors for requiring revision surgery have been identified, including younger age, high activity level, graft type selection, tunnel positioning, and concomitant meniscal injuries. 16 , 17 , 18 , 19 These rates are predominantly derived from athletic populations, as ACL injuries most commonly occur in sports participation, particularly in cutting and pivoting activates. 20 However, the specific impact of underlying medical conditions, such as eating disorders, on revision risk remains poorly understood in these populations. Similarly, our finding of increased revision rates following meniscectomy in both AN and BN patients warrants additional consideration, as meniscectomy is typically considered a procedure with minimal reliance on tissue healing. 21 The elevated revision rates may reflect additional factors outside of tissue quality that can also be affected by nutritional deficiencies. The present study also highlighted similarities and differences between patients with AN and BN, further emphasizing the need for the patient's individual condition to be considered during the treatment process. Patients with BN were more likely to have injuries to their ACL and MCL compared with matched controls, while this trend was not seen in patients with AN. One explanation for this is that health care providers are less likely to pursue a workup for an eating disorder when the patient has a normal weight or BMI, as is the case with BN. 22 Furthermore, orthopaedic injuries may be the first obvious external manifestation of an eating disorder and a driving factor for seeking initial medical attention. 23 These overall findings suggest the potential need for preoperative evaluations tailored to patients with eating disorders, potentially including pre‐surgical nutrition or rehabilitation plans. Given the multifactorial nature of eating disorders and their various presentations and manifestations, a multidisciplinary approach is necessary for effective management and prevention of subsequent injuries. Collaboration between orthopaedic surgeons, primary care physicians, mental health professionals, and dieticians is often recommended. 14 , 24 , 25 For athletes aiming to return to sport safely, the involvement of a primary care sports medicine physician can be invaluable in creating a tailored, gradual return‐to‐play plan with respect to their level of relative energy deficiency in sport. While orthopaedic surgeons are equipped to manage the acute injuries, dieticians are best suited to address nutritional deficiencies, and cognitive behavior therapy with specialized therapists has proven most effective for the long‐term management of disordered eating. 26 Furthermore, providers from all disciplines and specialties should be aware of signs indicating a potentially undiagnosed AN or BN, which could prevent potential subsequent complications. An integrative care model not only helps prevent further injuries but addresses root causes as primary prevention. Limitations Being a retrospective database study, there are inherent limitations. Although PearlDiver allowed for access to large amounts of patient data, this study lacked insight into the nature surrounding the initial injuries, surgery, and possible revision surgeries. It was not known if the revision surgeries were because of traumatic or atraumatic failure, incomplete post‐operative rehabilitation plan, activity level, participation in sports, or other external factors. Technical failures, including malpositioning of ACL tunnels, hyperlaxity, increased tibial slope, and specific type of meniscal tear (i.e., root tear, bucket‐handle tear, radial split) can all impact likelihood of revision knee surgeries and could not be adjusted for using our database model. 16 , 17 , 19 , 27 Because of the format of extracted data, a subgroup analysis for the effect of sex on outcomes was unable to be performed. Concomitant procedures performed at the time of ACL reconstruction, such as meniscal repair, could also impact outcomes. Additionally, we were not able to identify or control for the surgical approach, graft selection, or baseline level of activity from the data available. Importantly, this study used a revision surgery as a proxy for graft failure. In many instances, patients who fail an initial surgery may opt to forgo further surgical intervention. 27 The analysis also could not control for rehabilitation protocol differences or variations in activity levels, which may influence revision surgery rates. CONCLUSIONS In this study, patients with AN and BN had a significantly increased risk for various knee and ankle injuries and revision surgeries compared with matched controls. Additionally, patients with AN and BN who sustained these injuries were more likely to undergo a revision surgical procedure within the first 5 years of the primary procedure. SUPPORTING INFORMATION Additional supporting information can be found online in the Supporting Information section. DISCLOSURES The authors (J.G.W., D.G.D., E.L., J.M.H., J.S.B., K.A.P., C.M.D., S.K.P.) declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this article. Supporting information Supplementary Material REFERENCES Associated Data Supplementary Materials Supplementary Material