Advertisement
Research Article|Articles in Press

Bariatric surgery, osteoarthritis and arthroplasty of the hip and knee in Swedish Obese Subjects – up to 31 years follow-up of a controlled intervention study

Open AccessPublished:February 05, 2023DOI:https://doi.org/10.1016/j.joca.2022.11.015

      Summary

      Objective

      To study the long-term effect of obesity and bariatric surgery on incidences of osteoarthritis and arthroplasty of hip and knee.

      Design

      Hazard ratios (HR) and incidence rates (IR) of osteoarthritis and arthroplasty of hip and knee were studied in the prospective, controlled, non-randomized Swedish Obese Subjects (SOS) study (bariatric surgery group, n = 2007; matched controls given usual obesity care, n = 2040) and the SOS reference cohort (n = 1135, general population). Osteoarthritis diagnosis and arthroplasty for osteoarthritis were captured from the National Swedish Patient Register. Median follow-up time was 21.2 (IQR 16.4–24.8), 22.9 (IQR 19.1–25.7), and 20.1 years (IQR 18.7–20.9) for the control group, surgery group and reference cohort, respectively.

      Results

      The surgery group displayed lower incidence of hip osteoarthritis (IR 5.3, 95% CI 4.7–6.1) compared to controls (IR 6.6, 95% CI 5.9–7.5, adjHR 0.83, 95% CI 0.69–1.00) but similar incidence of hip arthroplasty. Similar incidence of knee osteoarthritis was observed in the surgery group and controls, but knee arthroplasty was more common in the surgery group (IR 7.4, 95% CI 6.6–8.2 and 5.6, 95% CI 4.9–6.4, adjHR 1.45, 95% CI 1.22–1.74). The reference cohort displayed lower incidences of osteoarthritis and arthroplasty of hip and knee compared with the surgery group and controls.

      Conclusion

      Bariatric surgery did not normalize the increased risk of knee and hip osteoarthritis in patients with obesity but was associated with an increased incidence of knee arthroplasty compared to the control group. With the limitations inherent to the present data, additional studies are needed to confirm these results.

      Trial registration

      clinicaltrials.gov Identifier: NCT01479452.

      Keywords

      Introduction

      Obesity is a major risk factor for osteoarthritis of the hip and knee
      • Bijlsma J.W.
      • Berenbaum F.
      • Lafeber F.P.
      Osteoarthritis: an update with relevance for clinical practice.
      • Anderson J.J.
      • Felson D.T.
      Factors associated with osteoarthritis of the knee in the first national Health and Nutrition Examination Survey (HANES I). Evidence for an association with overweight, race, and physical demands of work.
      • Pottie P.
      • Presle N.
      • Terlain B.
      • Netter P.
      • Mainard D.
      • Berenbaum F.
      Obesity and osteoarthritis: more complex than predicted!.
      • Reyes C.
      • Leyland K.M.
      • Peat G.
      • Cooper C.
      • Arden N.K.
      • Prieto-Alhambra D.
      Association between overweight and obesity and risk of clinically diagnosed knee, hip, and hand osteoarthritis: a population-based cohort study.
      . Osteoarthritis and arthroplasty are expected to increase because of the growing obesity pandemic
      • Lohmander L.S.
      • Gerhardsson de Verdier M.
      • Rollof J.
      • Nilsson P.M.
      • Engstrom G.
      Incidence of severe knee and hip osteoarthritis in relation to different measures of body mass: a population-based prospective cohort study.
      ,
      • Flugsrud G.B.
      • Nordsletten L.
      • Espehaug B.
      • Havelin L.I.
      • Engeland A.
      • Meyer H.E.
      The impact of body mass index on later total hip arthroplasty for primary osteoarthritis: a cohort study in 1.2 million persons.
      . Bariatric surgery is an effective obesity treatment resulting in long-term weight reduction
      • Carlsson L.M.
      • Peltonen M.
      • Ahlin S.
      • Anveden A.
      • Bouchard C.
      • Carlsson B.
      • et al.
      Bariatric surgery and prevention of type 2 diabetes in Swedish obese subjects.
      , but whether bariatric surgery is associated with a reduced risk of hip and knee osteoarthritis and a reduction of the need for arthroplasty is not clear.
      Uncontrolled studies indicate that bariatric surgery can improve symptoms of knee osteoarthritis, reduce markers of cartilage turnover, and possibly reduce the need for arthroplasty
      • Hacken B.
      • Rogers A.
      • Chinchilli V.
      • Silvis M.
      • Mosher T.
      • Black K.
      Improvement in knee osteoarthritis pain and function following bariatric surgery: 5-year follow-up.
      • Edwards C.
      • Rogers A.
      • Lynch S.
      • Pylawka T.
      • Silvis M.
      • Chinchilli V.
      • et al.
      The effects of bariatric surgery weight loss on knee pain in patients with osteoarthritis of the knee.
      • Richette P.
      • Poitou C.
      • Garnero P.
      • Vicaut E.
      • Bouillot J.L.
      • Lacorte J.M.
      • et al.
      Benefits of massive weight loss on symptoms, systemic inflammation and cartilage turnover in obese patients with knee osteoarthritis.
      • Hooper M.M.
      • Stellato T.A.
      • Hallowell P.T.
      • Seitz B.A.
      • Moskowitz R.W.
      Musculoskeletal findings in obese subjects before and after weight loss following bariatric surgery.
      • Abu-Abeid S.
      • Wishnitzer N.
      • Szold A.
      • Liebergall M.
      • Manor O.
      The influence of surgically-induced weight loss on the knee joint.
      • Fonseca Mora M.C.
      • Milla Matute C.A.
      • Ferri F.
      • Lo Menzo E.
      • Szmostein S.
      • Rosenthal R.J.
      Reduction of invasive interventions in severely obese with osteoarthritis after bariatric surgery.
      • Korenkov M.
      • Shah S.
      • Sauerland S.
      • Duenschede F.
      • Junginger T.
      Impact of laparoscopic adjustable gastric banding on obesity co-morbidities in the medium- and long-term.
      . However, controlled studies with longer follow-up are lacking and little is known about the effects of bariatric surgery on hip osteoarthritis. Studies on bariatric surgery and arthroplasty have mostly focused on possible beneficial effects of bariatric surgery in regard to complication rates and functional outcomes after arthroplasty, with divergent results
      • Werner B.C.
      • Kurkis G.M.
      • Gwathmey F.W.
      • Browne J.A.
      Bariatric surgery prior to total knee arthroplasty is associated with fewer postoperative complications.
      • Martin J.R.
      • Watts C.D.
      • Taunton M.J.
      Bariatric surgery does not improve outcomes in patients undergoing primary total knee arthroplasty.
      • Watts C.D.
      • Martin J.R.
      • Houdek M.T.
      • Abdel M.P.
      • Lewallen D.G.
      • Taunton M.J.
      Prior bariatric surgery may decrease the rate of re-operation and revision following total hip arthroplasty.
      • Severson E.P.
      • Singh J.A.
      • Browne J.A.
      • Trousdale R.T.
      • Sarr M.G.
      • Lewallen D.G.
      Total knee arthroplasty in morbidly obese patients treated with bariatric surgery: a comparative study.
      • Ighani Arani P.
      • Wretenberg P.
      • Ottosson J.
      • Robertsson O.
      • Annette W.D.
      Bariatric surgery prior to total knee arthroplasty is not associated with lower risk of revision: a register-based study of 441 patients.
      .
      We here report on the long-term risk of osteoarthritis and the need for hip and knee arthroplasty after bariatric surgery in the Swedish Obese Subjects (SOS) study, a prospective, controlled intervention trial, and in the SOS reference cohort, a reference cohort from the general population. We hypothesized that treatment with bariatric surgery would be associated with less osteoarthritis and less need for arthroplasty based on the large weight loss induced by bariatric surgery.

      Subjects and methods

      The SOS study

      The SOS study is a prospective controlled intervention study including patients with obesity, one group treated with bariatric surgery, and a matched control group of patients receiving usual obesity care at their primary health care centre. The design of the SOS study has been described
      • Ahlin S.
      • Peltonen M.
      • Sjoholm K.
      • Anveden A.
      • Jacobson P.
      • Andersson-Assarsson J.C.
      • et al.
      Fracture risk after three bariatric surgery procedures in Swedish obese subjects: up to 26 years follow-up of a controlled intervention study.
      . The primary endpoint of the SOS study was overall mortality; osteoarthritis and arthroplasty were not predefined end points. Patients were recruited via mass media campaigns and recruitment via surgical departments and primary health care centres between September 1, 1987 and January 31, 2001. The study was conducted at 25 surgical departments and 480 primary health care centres. The inclusion criteria were 37–60 years of age and a Body Mass Index (BMI) >34 kg/m2 for men and >38 kg/m2 for women. Exclusion criteria were previous bariatric surgery, gastric ulcer during the past 6 months, previous surgery for gastric or duodenal ulcer, ongoing malignancy, active malignancy during the past 5 years, bulimic eating pattern, myocardial infarction during the past 6 months, drug or alcohol abuse, psychiatric or cooperation problems or other contraindicating conditions such as anti-inflammatory or glucocorticoid treatment. Identical inclusion- and exclusion criteria were used for the surgery and matched control groups.
      The recruitment of study participants is described in Supplementary Fig. S1. Initially, 8966 of 11,453 study applicants fulfilled the criteria for age, weight and height and were asked to return a study questionnaire. A matching examination with assessment of anthropometry, electrocardiogram (ECG), blood pressure, blood samples, was offered to the 7593 study applicants who returned the study questionnaire, and 6905 completed the examination. After matching examinations, 5335 were eligible for participation. The surgery group was created from the 2010 eligible applicants who chose to undergo bariatric surgery. The control group (n = 2037) was created from eligible applicants who chose not to undergo bariatric surgery and received usual care. The control group was matched to the surgery group using 18 variables (sex, age, height, weight, waist and hip circumferences, systolic blood pressure, serum levels of cholesterol and triglycerides, diabetes status, menopausal status, smoking, four psychosocial variables, two personality traits). Matching was performed on a group level to keep the mean values of the matching variables similar in the control group and the surgery group
      • Pocock S.J.
      • Simon R.
      Sequential treatment assignment with balancing for prognostic factors in the controlled clinical trial.
      . For three individuals who initially choose to undergo surgery, the surgical intervention was not performed and for the per protocol analysis used in this current study, these three individuals are included in the control group (n = 2040). The patients in the per protocol surgery group (n = 2007) were treated with either gastric banding (n = 376), vertical banded gastroplasty (n = 1365) or gastric bypass (n = 266). The surgeon decided type of surgical procedure. A baseline examination was performed before study start and participants followed-up after 0.5, 1, 2, 3, 4, 6, 8, 10, 15 and 20 years
      • Ahlin S.
      • Peltonen M.
      • Sjoholm K.
      • Anveden A.
      • Jacobson P.
      • Andersson-Assarsson J.C.
      • et al.
      Fracture risk after three bariatric surgery procedures in Swedish obese subjects: up to 26 years follow-up of a controlled intervention study.
      .

      The SOS reference cohort

      The SOS reference cohort was created as a reference cohort to the SOS study
      • Sjostrom L.
      Review of the key results from the Swedish Obese Subjects (SOS) trial – a prospective controlled intervention study of bariatric surgery.
      ,
      • Larsson I.
      • Berteus Forslund H.
      • Lindroos A.K.
      • Lissner L.
      • Naslund I.
      • Peltonen M.
      • et al.
      Body composition in the SOS (Swedish Obese Subjects) reference study.
      . A computer-based procedure was used to randomly select participants between 37 and 60 years of age from the Swedish population registry, who received a written invitation to participate. Participants were recruited between August 1994 and December 1999, a period when the majority of the patients of the SOS study were recruited. In total, 1135 participants (524 men and 611 women) were included and underwent a baseline examination and assessment of anthropometry, blood biochemistry and filled in questionnaires similar to those in the SOS study.

      Ethics and patient consent

      Seven regional ethics review boards approved the studies. All patients gave informed consent.

      Patient and public involvement

      Patients or the public were not involved in study design, participant recruitment or conduct of the study.

      Osteoarthritis and arthroplasty in hip and knee

      ICD8, ICD9 and ICD10 codes for osteoarthritis diagnosis and for arthroplasty of the hip and knee were used to identify events that required hospitalization or specialist outpatient treatment in the Swedish National Patient Register (Supplementary Table S1). Hence, this study captures events of osteoarthritis and arthroplasty requiring orthopaedic specialist care. The Swedish National Patient register has 99% coverage for the inpatient part of the register and 78% coverage for specialist outpatient part of the register with a high positive predictive value (85–95%) for overall diagnosis and a high proportion (about 93%) identification of all cases of knee and hip arthroplasty compared to national quality registers
      • Ludvigsson J.F.
      • Andersson E.
      • Ekbom A.
      • Feychting M.
      • Kim J.L.
      • Reuterwall C.
      • et al.
      External review and validation of the Swedish national inpatient register.
      . Only hip arthroplasty accompanied by hip or general osteoarthritis diagnosis and knee arthroplasty accompanied by knee or general osteoarthritis diagnosis were included. No national registers for primary health care are available in Sweden, so osteoarthritis diagnosed and managed in primary health care was not included.

      Risk factors of osteoarthritis at baseline

      Information on age, sex, physical activity, postmenopausal state, and pain in knee or hip were retrieved from questionnaires at baseline. The questions used to define hip and knee pain were formulated as: Have you had pain in the hip/knee that periodically restricted your working capacity during the past 12 months? (Yes/No). BMI was calculated by dividing the measured weight in kilogrammes with the measured height in metres squared. For estimation of BMI at 20 years of age, information on self-reported weight at 20 years of age from the matching questionnaire was used with height at baseline. Hip and knee arthroplasty before baseline were identified by searching the national patient register for procedure codes for hip and knee arthroplasty.

      Follow-up

      The present report includes data from the National patient register until December 31, 2018 for the SOS study and until December 31, 2016 for the SOS reference cohort. Median time of follow-up was 21.2 years (interquartile range (IQR) 16.4–24.8) for the control group, 22.9 (IQR 19.1–25.7) for the surgery group, and 20.1 years (IQR 18.7–20.9) for the SOS reference cohort.

      Statistical analysis

      Mean values with standard deviations (SD) or numbers and percentages were used to report baseline characteristics. Time to first event (first event of osteoarthritis of the knee, osteoarthritis of the hip, arthroplasty of the knee, arthroplasty of the hip) was analysed with Kaplan–Meier estimates of cumulative incidence. Differences in cumulative incidences between the surgery group, the control group and the reference cohort were analysed with log-rank tests in unadjusted analysis. Cox proportional hazard regression was used in both unadjusted and adjusted analyses. The analyses were adjusted for age, sex, and proportion of women in menopause at baseline. Participants with previous history of arthroplasty of the knee or hip were excluded in the knee or hip arthroplasty analyses, respectively. Participants with a diagnosis of osteoarthritis at baseline were not excluded from the analysis of osteoarthritis. Instead, work-restricting pain in hip or knee was used as a proxy for prevalent osteoarthritis and sensitivity analyses including and excluding participants with pain from the body area of interest were performed. Hazard ratio (HR) or adjusted HR (adjHR) with 95% confidence intervals (CI) were used to present data. Additional sensitivity analyses with competing risk regression models suggested by Fine and Gray
      • Fine J.P.
      • Gray R.J.
      A proportional hazards model for the subdistribution of a competing risk.
      were conducted to account for differences in overall mortality between the study groups
      • Carlsson L.M.S.
      • Sjoholm K.
      • Jacobson P.
      • Andersson-Assarsson J.C.
      • Svensson P.A.
      • Taube M.
      • et al.
      Life expectancy after bariatric surgery in the Swedish obese subjects study.
      . The relative risk estimates from these models are presented as subhazard ratios with corresponding CI.
      A per-protocol approach was used. All participants were included in their study group until any bariatric surgery was performed in the control group or the participants in the surgery group removed their bariatric procedure, after which the participants were censored from the analysis. Stata statistical package 12.1 (Stata-corp 2011, Stata Statistical Software: Release 12, College Station TX, USA; StataCorp LP) was used for statistical analyses.

      Results

      Patient characteristics

      The SOS reference cohort had a higher proportion of men (46.2%) and a lower BMI at baseline (25.2 (SD 3.8)) compared to the SOS surgery group (29.2% men; BMI 42.4 (SD 4.5)) and the SOS control group (29.1% men; BMI 40.1 (SD 4.7)) (Table I). A larger proportion of participants reported pain in the knee (38.9%) and hip (24.5%) at baseline in the surgery group compared to the control group (knee 25.7%; hip 18.5%) and SOS reference cohort (knee 19.2%; hip 11.6%).
      Table IBaseline characteristics of study participants in the SOS study and the SOS reference cohort
      SOS surgery groupn

       = 2007
      SOS control groupn

       = 2040
      SOS reference cohortn

       = 1135
      Age (SD)47.2 (5.9)48.7 (6.3)49.5 (7.0)
      Males587 (29.2 %)593 (29.1 %)524 (46.2 %)
      BMI (SD)42.4 (4.5)40.1 (4.7)25.2 (3.8)
      Proportion of BMI-categories at 20 years of age
       normal (<25)725 (36.3 %)869 (42.8 %)1039 (93.8 %)
       Overweight (25–29.99)744 (37.3 %)721 (35.5 %)62 (5.6 %)
       Obesity (≥30)528 (26.4 %)439 (21.6 %)7 (0.6 %)
      Proportion with diabetes344 (17.2 %)263 (12.9 %)38 (3.4 %)
      Proportion with hypertension1571 (78.4 %)1301 (63.8 %)312 (27.6 %)
      Proportion of women in menopause433 (30.5 %)529 (36.6 %)172 (28.2 %)
      Daily smoking518 (25.8 %)422 (20.8 %)235 (20.7 %)
      Proportion physically active in their leisure time1071 (53.3 %)1351 (66.2 %)1012 (89.2 %)
      Pain in knee or hip at baseline
       Hip492 (24.5 %)378 (18.5 %)132 (11.6 %)
       Knee781 (38.9 %)524 (25.7 %)218 (19.2 %)
      Previous arthroplasty
       Hip11 (0.5 %)8 (0.4 %)1 (0.1 %)
       Knee5 (0.2 %)2 (0.1 %)2 (0.2 %)
      Data are presented as mean value and standard deviations in parenthesis or number and percent. The body-mass index is the weight in kilogrammes divided by the square of the height in metres. For estimation of BMI at 20 years of age, information on self-reported weight at 20 years of age from the matching questionnaire was used, with height at baseline. Information on hip and knee pain was assessed from baseline questionnaires. Diabetes was defined as a fasting blood glucose >6.1 mmol/l or use of diabetes medication. Hypertension was defined as a diastolic blood pressure >90 mm Hg or systolic blood pressure >140 mm Hg or use of hypertension medication. Physically active was defined as at least 4 h of physical activity (e.g., walking, going by bicycle) per week.
      Following bariatric surgery, the surgery group obtained long-term weight reduction (Fig. 1). In contrast, the control group did not obtain any major weight loss (less than 3 kg) during the entire follow-up period. At 1 year after bariatric surgery, the mean BMI for the surgery group was 31.8 (95% CI 31.5–32.0), while that of the control group was 39.9 (95% CI 39.7–40.1). No data on weight changes during the follow-up were available for the reference cohort.
      Fig. 1
      Fig. 1BMI with 95 % confidence intervals during follow-up in the surgery group and control group in the SOS study and BMI at study start in the SOS reference cohort.

      Osteoarthritis and arthroplasty of the hip

      For total number of events, incidence rates (IR) and unadjusted hazard ratios, see Table II, Table III. The surgery group had a lower risk of hip osteoarthritis diagnosis during follow-up, compared to the control group (adjHR 0.83; 95% CI 0.69–1.00; Fig. 2(A), Table II). However, the incidence of hip arthroplasty for the surgery group was similar to that of the control group (adjHR 0.90; 95% CI 0.72–1.11; Fig. 2(B), Table II).
      Table IINumber of events, incidence rates and Cox hazard models of osteoarthritis and arthroplasty of the hip for the SOS study
      Surgery groupControl groupP-value
      Osteoarthritis of the hip
      Number of participants20072040
      Events, n217256
      Events per 1000 p-y (95 % CI)5.3 (4.7–6.1)6.6 (5.9–7.5)0.006
      Cox proportional HR (95 % CI)0.77 (0.65–0.93)1.00.006
      Cox proportional adj HR (95 % CI)0.83 (0.69–1.00)1.00.046
      Arthroplasty of the hip
      Number of participants19962032
      Events, n162184
      Events per 1000 p-y (95 % CI)4.0 (3.4–4.6)4.7 (4.1–5.5)0.064
      Cox proportional HR (95 % CI)0.82 (0.66–1.01)1.00.064
      Cox proportional adj HR (95 % CI)0.90 (0.72–1.11)1.00.309
      Adjusted analyses are adjusted for age, sex, and proportion of women in menopause at baseline. Participants with previous history of arthroplasty of the hip were excluded in the hip arthroplasty analyses. P-y = person years.
      Table IIINumber of events, incidence rates and Cox hazard models of osteoarthritis and arthroplasty of hip in the SOS surgery and control groups compared to the SOS reference cohort
      Reference cohortSurgery groupP-valueControl groupP-value
      Osteoarthritis of the hip
      Number of participants113520072040
      Events, n80217256
      Events per 1000 p-y (95 % CI)3.7 (3.0–4.6)5.3 (4.7–6.1)6.6 (5.9–7.5)
      Cox proportional HR (95 % CI)1.01.33 (1.03–1.73)0.0291.72 (1.34–2.22)<0.001
      Cox proportional adj HR (95 % CI)1.01.46 (1.12–1.91)0.0051.75 (1.35–2.26)<0.001
      Arthroplasty of the hip
      Number of participants113419962032
      Events, n55162184
      Events per 1000 p-y (95 % CI)2.5 (1.9–3.3)4.0 (3.4–4.6)4.7 (4.1–5.5)
      Cox proportional HR (95 % CI)1.01.55 (1.14–2.12)0.0051.90 (1.41–2.57)<0.001
      Cox proportional adj HR (95 % CI)1.01.78 (1.30–2.44)<0.0011.98 (1.45–2.69)<0.001
      Participants treated with arthroplasty of the hip before baseline were excluded from the hip arthroplasty analysis. The SOS reference cohort is used as reference in the Cox Proportional Hazard regression analysis. The adjusted analysis is adjusted for age, sex and proportion of women in menopause at baseline. P-y = person years.
      Fig. 2
      Fig. 2Kaplan–Meier estimates of cumulative incidence of osteoarthritis and arthroplasty of the hip in the SOS intervention study and SOS reference cohort. The upper panels illustrate osteoarthritis of the hip (A), and arthroplasty of the hip (B). The lower panels illustrate arthroplasty of the hip in participants without hip pain at baseline (C), and arthroplasty of the hip in participants with hip pain at baseline (D).
      Compared to the reference cohort, both the surgery and control groups showed an increased risk of being diagnosed with hip osteoarthritis (adjHR 1.46; 95% CI 1.12–1.91 and adjHR 1.75; 95% CI 1.35–2.26, respectively) and of undergoing hip arthroplasty (adjHR 1.78; 95% CI 1.30–2.44 and adjHR 1.98, 95% CI 1.45–2.69, respectively) (Fig. 2(A) and (B), Table III). Results from subgroup analyses including patients with and without hip pain at baseline in general confirmed results from main analyses (Fig. 2(C) and (D), Supplementary Tables S2 and S3). In addition, results from the competing risk regression models were in agreement with the Cox regression models (Supplementary Tables S4 and S5).

      Osteoarthritis and arthroplasty of the knee

      A similar risk of being diagnosed with knee osteoarthritis was observed in the surgery and control groups (adjHR 1.08; 95% CI 0.95–1.22; Fig. 3(A), Table IV). A higher incidence of knee arthroplasty was observed in the surgery group compared to the control group (adjHR 1.45; 95% CI 1.22–1.74; Fig. 3(B), Table IV). Both the surgery group and the control group displayed a higher risk of knee osteoarthritis (adjHR 3.28; 95% CI 2.64–4.08 and adjHR 3.03; 95% CI 2.44–3.77, respectively) and of knee arthroplasty (adjHR 4.44; 95% CI 3.15–6.29 and adjHR 3.06; 95% CI 2.17–4.33, respectively), compared to the SOS reference cohort (Fig. 3(A) and (B), Table V). For total number of events, IR and unadjusted hazard ratios including subgroup analysis with patients with and without knee pain at baseline, see Fig. 3(C) and (D), Table IV, Table V and Supplementary Tables S6 and S7. The subgroup analyses of osteoarthritis and arthroplasty in participants with and without knee pain at baseline in general confirmed results of main analyses. The additional sensitivity analyses with competing risk regression models were also in line with the main Cox regression models analyses (Supplementary Tables S8 and S9).
      Fig. 3
      Fig. 3Kaplan–Meier estimates of cumulative incidence of osteoarthritis and arthroplasty of the knee in the SOS intervention study and SOS reference cohort. The upper panels illustrate osteoarthritis of the knee (A), and arthroplasty of the knee (B). The lower panels illustrate arthroplasty of the knee in participants without knee pain at baseline (C), and arthroplasty of the knee in participants with knee pain at baseline (D).
      Table IVNumber of events, incidence rates and Cox hazard models of osteoarthritis and arthroplasty of the knee for the SOS study
      Surgery groupControl groupP-value
      Osteoarthritis of the knee
      Number of participants20072040
      Events, n553514
      Events per 1000 p-y (95 % CI)14.8 (13.6–16.1)14.2 (13.0–15.5)0.788
      Cox Proportional HR (95 % CI)1.02 (0.90–1.15)1.00.788
      Cox proportional adjHR (95% CI)1.08 (0.95–1.22)1.00.227
      Arthroplasty of the knee
      Number of participants20022038
      Events, n294218
      Events per 1000 p-y (95 % CI)7.4 (6.6–8.2)5.6 (4.9–6.4)0.006
      Cox proportional HR (95 % CI)1.28 (1.07–1.52)1.00.006
      Cox proportional adj HR (95 % CI)1.45 (1.22–1.74)1.0<0.001
      Adjusted analyses are adjusted for age, sex, and proportion of women in menopause at baseline. Participants with previous history of arthroplasty of the knee were excluded in the knee arthroplasty analysis. P-y = person years.
      Table VNumber of events, incidence rates and Coz hazard models of osteoarthritis and arthroplasty of the knee in the SOS surgery and control groups compared to the SOS reference cohort
      Reference cohortSurgery groupP-valueControl groupP-value
      Osteoarthritis of the knee
      Number of participants113520072040
      Events, n101553514
      Events per 1000 p-y (95 % CI)4.8 (3.9–5.8)14.8 (13.6–16.1)14.2 (13.0–15.5)
      Cox proportional HR (95 % CI)1.03.10 (2.51–3.85)<0.0013.05 (2.46–3.79)<0.001
      Cox proportional adj HR (95 % CI)1.03.28 (2.64–4.08)<0.0013.03 (2.44–3.77)<0.001
      Arthroplasty of the knee
      Number of participants113320022038
      Events, n40294218
      Events per 1000 p-y (95 % CI)1.8 (1.3–2.5)7.4 (6.6–8.2)5.6 (4.9–6.4)
      Cox proportional HR (95 % CI)1.03.85 (2.76–5.36)<0.0013.01 (2.15–4.23)<0.001
      Cox proportional adj HR (95 % CI)1.04.44 (3.15–6.26)<0.0013.06 (2.17–4.33)<0.001
      Participants treated with arthroplasty of the knee before baseline were excluded from the knee arthroplasty analysis. The SOS reference cohort is used as reference in the Cox Proportional Hazard regression analysis. The adjusted analysis is adjusted for age, sex and proportion of women in menopause at baseline. P-y = person years.

      Discussion

      We here report that bariatric surgery was associated with a reduced incidence of hip osteoarthritis compared to a usual care control group with obesity. However, this did not result in a reduced incidence of hip arthroplasty. We found a similar incidence of knee osteoarthritis in the bariatric surgery group and the control group, but unexpectedly, treatment with bariatric surgery was associated with higher incidence of knee arthroplasty. However, compared to the general population in the reference cohort, osteoarthritis and arthroplasty of the hip and knee were more common among patients with obesity treated with bariatric surgery or usual care.
      Previous results from the SOS study showed that among participants who reported pain of the hip and knee at baseline, a higher proportion had recovered from pain after 2 years in the surgery group compared to the control group
      • Peltonen M.
      • Lindroos A.K.
      • Torgerson J.S.
      Musculoskeletal pain in the obese: a comparison with a general population and long-term changes after conventional and surgical obesity treatment.
      . We now show that with a median follow-up of about 20 years, bariatric surgery is associated with a lower incidence of hip osteoarthritis diagnosis. In contrast, for knee osteoarthritis diagnosis we found a similar incidence in patients with obesity treated with bariatric surgery and usual care. Our finding of a lower incidence of hip osteoarthritis diagnosis among bariatric surgery patients is consistent with a recently published propensity score matched cohort study on bariatric surgery patients with follow-up of maximum 14 years
      • Burkard T.
      • Holmberg D.
      • Wretenberg P.
      • Thorell A.
      • Hügle T.
      • Burden A.M.
      The associations between bariatric surgery and hip or knee arthroplasty, and hip or knee osteoarthritis: propensity score-matched cohort studies.
      . The same study also observed lower incidence of knee osteoarthritis diagnosis among bariatric surgery patients
      • Burkard T.
      • Holmberg D.
      • Wretenberg P.
      • Thorell A.
      • Hügle T.
      • Burden A.M.
      The associations between bariatric surgery and hip or knee arthroplasty, and hip or knee osteoarthritis: propensity score-matched cohort studies.
      , contrasting with our present finding with similar incidence of knee osteoarthritis in the surgery group and the matched control group. Shorter term, smaller and/or uncontrolled studies have suggested that both a modest reduction in weight
      • Felson D.T.
      • Zhang Y.
      • Anthony J.M.
      • Naimark A.
      • Anderson J.J.
      Weight loss reduces the risk for symptomatic knee osteoarthritis in women. The Framingham Study.
      • Atukorala I.
      • Makovey J.
      • Lawler L.
      • Messier S.P.
      • Bennell K.
      • Hunter D.J.
      Is there a dose-response relationship between weight loss and symptom improvement in persons with knee osteoarthritis?.
      • Gersing A.S.
      • Schwaiger B.J.
      • Nevitt M.C.
      • Zarnowski J.
      • Joseph G.B.
      • Feuerriegel G.
      • et al.
      Weight loss regimen in obese and overweight individuals is associated with reduced cartilage degeneration: 96-month data from the osteoarthritis initiative.
      • Jin X.
      • Gibson A.A.
      • Gale J.
      • Schneuer F.
      • Ding D.
      • March L.
      • et al.
      Does weight loss reduce the incidence of total knee and hip replacement for osteoarthritis? – A prospective cohort study among middle-aged and older adults with overweight or obesity.
      and bariatric surgery
      • Hacken B.
      • Rogers A.
      • Chinchilli V.
      • Silvis M.
      • Mosher T.
      • Black K.
      Improvement in knee osteoarthritis pain and function following bariatric surgery: 5-year follow-up.
      • Edwards C.
      • Rogers A.
      • Lynch S.
      • Pylawka T.
      • Silvis M.
      • Chinchilli V.
      • et al.
      The effects of bariatric surgery weight loss on knee pain in patients with osteoarthritis of the knee.
      • Richette P.
      • Poitou C.
      • Garnero P.
      • Vicaut E.
      • Bouillot J.L.
      • Lacorte J.M.
      • et al.
      Benefits of massive weight loss on symptoms, systemic inflammation and cartilage turnover in obese patients with knee osteoarthritis.
      • Hooper M.M.
      • Stellato T.A.
      • Hallowell P.T.
      • Seitz B.A.
      • Moskowitz R.W.
      Musculoskeletal findings in obese subjects before and after weight loss following bariatric surgery.
      • Abu-Abeid S.
      • Wishnitzer N.
      • Szold A.
      • Liebergall M.
      • Manor O.
      The influence of surgically-induced weight loss on the knee joint.
      • Fonseca Mora M.C.
      • Milla Matute C.A.
      • Ferri F.
      • Lo Menzo E.
      • Szmostein S.
      • Rosenthal R.J.
      Reduction of invasive interventions in severely obese with osteoarthritis after bariatric surgery.
      • Korenkov M.
      • Shah S.
      • Sauerland S.
      • Duenschede F.
      • Junginger T.
      Impact of laparoscopic adjustable gastric banding on obesity co-morbidities in the medium- and long-term.
      ,
      • Stefanik J.J.
      • Felson D.T.
      • Apovian C.M.
      • Niu J.
      • Margaret Clancy M.
      • LaValley M.P.
      • et al.
      Changes in pain sensitization after bariatric surgery.
      results in decreased incidence of osteoarthritis of the knee and symptom improvement. These differences in outcomes may be related to the longer follow-up of the present study, a lack of properly BMI-matched control groups in other studies, and that we have monitored only specialist care osteoarthritis diagnoses.
      We observed similar incidences of hip arthroplasty in the matched control group (IR 4.7/1000 person years) and the surgery group (IR 4.0/1000 person years), while knee arthroplasty was more common in the bariatric surgery group (IR 7.4/1000 person years) than in the control group (IR 5.6/1000 person years). In a clinical perspective, these IR correspond to in average 1.8 additional knee arthroplasty among 1000 bariatric surgery patients per year compared with a group of 1000 individuals with conventional obesity treatment. In contrast, a recent cohort study reported increased incidences for both hip and knee arthroplasty in the bariatric surgery group
      • Burkard T.
      • Holmberg D.
      • Wretenberg P.
      • Thorell A.
      • Hügle T.
      • Burden A.M.
      The associations between bariatric surgery and hip or knee arthroplasty, and hip or knee osteoarthritis: propensity score-matched cohort studies.
      .
      Based on the documented association of BMI with risk for osteoarthritis and arthroplasty of the hip and knee, reports of improvement of osteoarthritis symptoms, and decrease in radiographic worsening of joint structure following weight loss
      • Bijlsma J.W.
      • Berenbaum F.
      • Lafeber F.P.
      Osteoarthritis: an update with relevance for clinical practice.
      • Anderson J.J.
      • Felson D.T.
      Factors associated with osteoarthritis of the knee in the first national Health and Nutrition Examination Survey (HANES I). Evidence for an association with overweight, race, and physical demands of work.
      • Pottie P.
      • Presle N.
      • Terlain B.
      • Netter P.
      • Mainard D.
      • Berenbaum F.
      Obesity and osteoarthritis: more complex than predicted!.
      • Reyes C.
      • Leyland K.M.
      • Peat G.
      • Cooper C.
      • Arden N.K.
      • Prieto-Alhambra D.
      Association between overweight and obesity and risk of clinically diagnosed knee, hip, and hand osteoarthritis: a population-based cohort study.
      ,
      • Hacken B.
      • Rogers A.
      • Chinchilli V.
      • Silvis M.
      • Mosher T.
      • Black K.
      Improvement in knee osteoarthritis pain and function following bariatric surgery: 5-year follow-up.
      • Edwards C.
      • Rogers A.
      • Lynch S.
      • Pylawka T.
      • Silvis M.
      • Chinchilli V.
      • et al.
      The effects of bariatric surgery weight loss on knee pain in patients with osteoarthritis of the knee.
      • Richette P.
      • Poitou C.
      • Garnero P.
      • Vicaut E.
      • Bouillot J.L.
      • Lacorte J.M.
      • et al.
      Benefits of massive weight loss on symptoms, systemic inflammation and cartilage turnover in obese patients with knee osteoarthritis.
      • Hooper M.M.
      • Stellato T.A.
      • Hallowell P.T.
      • Seitz B.A.
      • Moskowitz R.W.
      Musculoskeletal findings in obese subjects before and after weight loss following bariatric surgery.
      • Abu-Abeid S.
      • Wishnitzer N.
      • Szold A.
      • Liebergall M.
      • Manor O.
      The influence of surgically-induced weight loss on the knee joint.
      • Fonseca Mora M.C.
      • Milla Matute C.A.
      • Ferri F.
      • Lo Menzo E.
      • Szmostein S.
      • Rosenthal R.J.
      Reduction of invasive interventions in severely obese with osteoarthritis after bariatric surgery.
      • Korenkov M.
      • Shah S.
      • Sauerland S.
      • Duenschede F.
      • Junginger T.
      Impact of laparoscopic adjustable gastric banding on obesity co-morbidities in the medium- and long-term.
      ,
      • Joseph G.B.
      • McCulloch C.E.
      • Nevitt M.C.
      • Lynch J.
      • Lane N.E.
      • Link T.M.
      Effects of weight change on knee and hip radiographic measurements and pain over 4 years: data from the osteoarthritis initiative.
      , we expected a decrease in the incidence of osteoarthritis and arthroplasty following the marked weight-loss induced by bariatric surgery. Several factors may have contributed to our somewhat counterintuitive results. First, the apparent failure of the surgery-induced weight-loss to reduce the incidence of arthroplasty, compared to the control group with usual care, may at least in part be the result of a biased restriction of offering in particular knee arthroplasty to the patients in the control group with chronic severe obesity. The marked weight loss induced in the bariatric surgery group may have diminished the same biased restriction
      • Berend K.R.
      • Lombardi Jr., A.V.
      • Mallory T.H.
      • Adams J.B.
      • Groseth K.L.
      Early failure of minimally invasive unicompartmental knee arthroplasty is associated with obesity.
      • Dewan A.
      • Bertolusso R.
      • Karastinos A.
      • Conditt M.
      • Noble P.C.
      • Parsley B.S.
      Implant durability and knee function after total knee arthroplasty in the morbidly obese patient.
      Workgroup of the American Association of H, Knee Surgeons Evidence Based C
      Obesity and total joint arthroplasty: a literature based review.
      • Ricciardi B.F.
      • Giori N.J.
      • Fehring T.K.
      Clinical faceoff: should orthopaedic surgeons have strict BMI cutoffs for performing primary TKA and THA?.
      . Second, almost two out of three patients in the bariatric surgery group had overweight or obesity already at age 20
      • Kristensson F.M.
      • Andersson-Assarsson J.C.
      • Svensson P.A.
      • Carlsson B.
      • Peltonen M.
      • Carlsson L.M.S.
      Effects of bariatric surgery in early- and adult-onset obesity in the prospective controlled Swedish obese subjects study.
      . At study baseline and a mean age of about 48, their average BMI exceeded 40. Many thus had a history of overweight or obesity for much of their adult life and could have developed severe osteoarthritis already at the time of their bariatric surgery
      • Flugsrud G.B.
      • Nordsletten L.
      • Espehaug B.
      • Havelin L.I.
      • Engeland A.
      • Meyer H.E.
      The impact of body mass index on later total hip arthroplasty for primary osteoarthritis: a cohort study in 1.2 million persons.
      ,
      • Flugsrud G.B.
      • Nordsletten L.
      • Espehaug B.
      • Havelin L.I.
      • Meyer H.E.
      Weight change and the risk of total hip replacement.
      . This may have compromised their knee or hip joint beyond the ‘point-of-no-return’ and prevented a possible ‘rescue’ from arthroplasty by surgery-induced weight loss. Third, even with a mean weight-loss of 30 kg at 1 year after bariatric surgery, the mean BMI of this group remained greater than 30 for the duration of the almost 23-year follow-up, leaving them at a continued increased risk of osteoarthritis and arthroplasty. The association between BMI and osteoarthritis, and BMI and arthroplasty, is stronger for the knee than for the hip, perhaps related to differences in joint anatomy and loading
      • Lohmander L.S.
      • Gerhardsson de Verdier M.
      • Rollof J.
      • Nilsson P.M.
      • Engstrom G.
      Incidence of severe knee and hip osteoarthritis in relation to different measures of body mass: a population-based prospective cohort study.
      . This may at least in part explain the differences in results between hip and knee osteoarthritis in the bariatric surgery group observed in the present study. Finally, endocrine factors may contribute to the apparent failure of bariatric surgery to reduce knee arthroplasty in particular. Oestrogen appears to be more protective against knee osteoarthritis than hip osteoarthritis
      • Hunter D.J.
      • Bierma-Zeinstra S.
      Osteoarthritis.
      . Weight loss after bariatric surgery leads to a decrease in oestrogen levels in both females and males due to reduction of the aromatization of androgens to oestrogen in adipose tissue
      • Nelson L.R.
      • Bulun S.E.
      Estrogen production and action.
      ,
      • Escobar-Morreale H.F.
      • Santacruz E.
      • Luque-Ramirez M.
      • Botella Carretero J.I.
      Prevalence of ‘obesity-associated gonadal dysfunction’ in severely obese men and women and its resolution after bariatric surgery: a systematic review and meta-analysis.
      . Hence, bariatric surgery may reduce an important protective factor against knee osteoarthritis.
      Our study has strengths and limitations. The SOS study is a prospective, matched controlled study which together with the SOS reference cohort provides a unique opportunity to study the long-term effects of bariatric surgery on osteoarthritis and arthroplasty compared both to a control group with sustained obesity, and with a reference cohort from the general population. The present report with a follow-up of up to 31 years includes data on osteoarthritis diagnosis and arthroplasty of the hip and knee obtained from reliable national health registers with a very high coverage of specialist outpatient and inpatient care. An important study limitation was that the ethics committees did not approve randomization due to high post-operative mortality risk when the study started in 1987. Matching was performed on group level and based on 18 different variables. In spite of this, several sources of bias, e.g., patient treatment preference and willingness to undergo surgical treatment, likely remained. Hence, it is possible that the surgery group participants were more prone to undergo arthroplasty than the control group which may have contributed to the somewhat unexpected results with higher IR of knee arthroplasty. Another example is the higher proportion of participants reporting pain of the knee and hip in the bariatric surgery group compared to the control group. However, subgroup analyses of osteoarthritis and arthroplasty in participants with and without pain at baseline in general confirmed the results from the main analyses. Due to the lack of national registers covering primary health care in Sweden, study participants with osteoarthritis diagnosed and treated only in primary care could not be reliably identified and were not included in the current analyses. Use of national registry data involves the risk of missing data and misclassifications, while the validity of our methodology is supported by the results of comparisons of the study groups with obesity with the reference cohort from the general population, confirming the well-known risk increases with obesity
      • Anderson J.J.
      • Felson D.T.
      Factors associated with osteoarthritis of the knee in the first national Health and Nutrition Examination Survey (HANES I). Evidence for an association with overweight, race, and physical demands of work.
      • Pottie P.
      • Presle N.
      • Terlain B.
      • Netter P.
      • Mainard D.
      • Berenbaum F.
      Obesity and osteoarthritis: more complex than predicted!.
      • Reyes C.
      • Leyland K.M.
      • Peat G.
      • Cooper C.
      • Arden N.K.
      • Prieto-Alhambra D.
      Association between overweight and obesity and risk of clinically diagnosed knee, hip, and hand osteoarthritis: a population-based cohort study.
      ,
      • Bourne R.
      • Mukhi S.
      • Zhu N.
      • Keresteci M.
      • Marin M.
      Role of obesity on the risk for total hip or knee arthroplasty.
      ,
      • Lim Y.Z.
      • Wang Y.
      • Cicuttini F.M.
      • Giles G.G.
      • Graves S.
      • Wluka A.E.
      • et al.
      Obesity defined by body mass index and waist circumference and risk of total knee arthroplasty for osteoarthritis: a prospective cohort study.
      . It is possible that the incidence of osteoarthritis was underestimated in the SOS reference cohort due to less contact with specialist care because of fewer comorbidities and less need for health care in general, whereas the SOS study participants may be more likely to be diagnosed with osteoarthritis because of higher need for specialist health care due to comorbidities. Finally, the mean age of the participants was close to 50 years at inclusion. With a median follow-up of 20–23 years, their mean age was around age 70 at the end of follow-up. In Sweden, the mean age for arthroplasty of the hip and knee is about 68–69 years, suggesting that results regarding arthroplasty may still change with continued follow-up
      • W-Dahl A
      • Kärrholm J.
      • Rogmark C.
      • Nauclér E.
      • Nåtman J.
      • Bülow E.
      • et al.
      The Swedish Arthroplasty Register Annual Report 2021.
      .
      In conclusion, the overall results of our long-term controlled study in patients with obesity suggest that bariatric surgery and weight-loss at a mean age of about 50 years was not associated with a normalization of the increased risk of knee and hip osteoarthritis and arthroplasty related to severe obesity.

      Author contributions

      Dr. Ahlin had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
      Concept and design: Ahlin, Lohmander, Carlsson, Peltonen.
      Acquisition, analysis, or interpretation of data: All authors.
      Drafting the manuscript: Ahlin, Lohmander.
      Critical revision of the manuscript for important intellectual content: All authors.
      Statistical analysis: Peltonen.
      Obtained funding: Ahlin, Sjöholm, Svensson, Carlsson.
      Administrative, technical, or material support: All Authors.
      Supervision: Ahlin, Lohmander, Carlsson.
      Approval of the final version of the manuscript: All authors.

      Declaration of competing interest

      No other disclosures were reported.

      Funding and role of the funding source

      This study was funded from grants from the Health & Medical Care Committee of the Region Västra Götaland. (grant no. VGFOUREG-931560 and VGFOUREG-941125), the Swedish Research Council (2020-01303, 2021-01496), the Swedish government under the LUA/ALF agreement concerning research and education of doctors (ALFGBG-966076, ALFGBG-965046, ALFGBG-965955), the Magnus Bergvall Foundation (grant no 2019-03489), Johan and Lisa Grönberg foundation (2014-00030 and 2015-00064), the Swedish diabetes foundation (DIA2019-417) and Wilhelm and Martina Lundgren foundation (2015-0509). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

      Availability of data and materials

      The data is subject to legal restrictions according to national legislation. Confidentiality regarding personal information in studies is regulated in the Public Access to Information and Secrecy Act (SFS 2009:400). There is a possibility to apply to get access to public documents that an authority holds. In this case, the University of Gothenburg is the specific authority that holds the documents. A request to get access to public documents can be rejected or granted with reservations. If the authority refuses to disclose the documents the applicant is entitled to get a written decision that can be appealed to the administrative court of appeal. Contact person, data inquiries from fellow researchers: Jan Borén, Professor, Prefect; Head of Institute of Medicine, the Sahlgrenska Academy, University of Gothenburg [email protected]

      Acknowledgements

      We would like to thank all the co-workers in the SOS study team and all the staff members at the 25 surgical departments and 480 primary health care centres involved in the SOS study.

      Appendix A. Supplementary data

      The following is the Supplementary data to this article.

      References

        • Bijlsma J.W.
        • Berenbaum F.
        • Lafeber F.P.
        Osteoarthritis: an update with relevance for clinical practice.
        Lancet. 2011; 377: 2115-2126
        • Anderson J.J.
        • Felson D.T.
        Factors associated with osteoarthritis of the knee in the first national Health and Nutrition Examination Survey (HANES I). Evidence for an association with overweight, race, and physical demands of work.
        Am J Epidemiol. 1988; 128: 179-189
        • Pottie P.
        • Presle N.
        • Terlain B.
        • Netter P.
        • Mainard D.
        • Berenbaum F.
        Obesity and osteoarthritis: more complex than predicted!.
        Ann Rheum Dis. 2006; 65: 1403-1405
        • Reyes C.
        • Leyland K.M.
        • Peat G.
        • Cooper C.
        • Arden N.K.
        • Prieto-Alhambra D.
        Association between overweight and obesity and risk of clinically diagnosed knee, hip, and hand osteoarthritis: a population-based cohort study.
        Arthritis Rheumatol. 2016; 68: 1869-1875
        • Lohmander L.S.
        • Gerhardsson de Verdier M.
        • Rollof J.
        • Nilsson P.M.
        • Engstrom G.
        Incidence of severe knee and hip osteoarthritis in relation to different measures of body mass: a population-based prospective cohort study.
        Ann Rheum Dis. 2009; 68: 490-496
        • Flugsrud G.B.
        • Nordsletten L.
        • Espehaug B.
        • Havelin L.I.
        • Engeland A.
        • Meyer H.E.
        The impact of body mass index on later total hip arthroplasty for primary osteoarthritis: a cohort study in 1.2 million persons.
        Arthritis Rheumatol. 2006; 54: 802-807
        • Carlsson L.M.
        • Peltonen M.
        • Ahlin S.
        • Anveden A.
        • Bouchard C.
        • Carlsson B.
        • et al.
        Bariatric surgery and prevention of type 2 diabetes in Swedish obese subjects.
        N Engl J Med. 2012; 367: 695-704
        • Hacken B.
        • Rogers A.
        • Chinchilli V.
        • Silvis M.
        • Mosher T.
        • Black K.
        Improvement in knee osteoarthritis pain and function following bariatric surgery: 5-year follow-up.
        Surg Obes Relat Dis. 2019; 15: 979-984
        • Edwards C.
        • Rogers A.
        • Lynch S.
        • Pylawka T.
        • Silvis M.
        • Chinchilli V.
        • et al.
        The effects of bariatric surgery weight loss on knee pain in patients with osteoarthritis of the knee.
        Arthritis. 2012; 2012504189
        • Richette P.
        • Poitou C.
        • Garnero P.
        • Vicaut E.
        • Bouillot J.L.
        • Lacorte J.M.
        • et al.
        Benefits of massive weight loss on symptoms, systemic inflammation and cartilage turnover in obese patients with knee osteoarthritis.
        Ann Rheum Dis. 2011; 70: 139-144
        • Hooper M.M.
        • Stellato T.A.
        • Hallowell P.T.
        • Seitz B.A.
        • Moskowitz R.W.
        Musculoskeletal findings in obese subjects before and after weight loss following bariatric surgery.
        Int J Obes. 2007; 31: 114-120
        • Abu-Abeid S.
        • Wishnitzer N.
        • Szold A.
        • Liebergall M.
        • Manor O.
        The influence of surgically-induced weight loss on the knee joint.
        Obes Surg. 2005; 15: 1437-1442
        • Fonseca Mora M.C.
        • Milla Matute C.A.
        • Ferri F.
        • Lo Menzo E.
        • Szmostein S.
        • Rosenthal R.J.
        Reduction of invasive interventions in severely obese with osteoarthritis after bariatric surgery.
        Surg Endosc. 2020; 34: 3606-3613
        • Korenkov M.
        • Shah S.
        • Sauerland S.
        • Duenschede F.
        • Junginger T.
        Impact of laparoscopic adjustable gastric banding on obesity co-morbidities in the medium- and long-term.
        Obes Surg. 2007; 17: 679-683
        • Werner B.C.
        • Kurkis G.M.
        • Gwathmey F.W.
        • Browne J.A.
        Bariatric surgery prior to total knee arthroplasty is associated with fewer postoperative complications.
        J Arthroplast. 2015; 30: 81-85
        • Martin J.R.
        • Watts C.D.
        • Taunton M.J.
        Bariatric surgery does not improve outcomes in patients undergoing primary total knee arthroplasty.
        Bone Jt Lett J. 2015; 97-B: 1501-1505
        • Watts C.D.
        • Martin J.R.
        • Houdek M.T.
        • Abdel M.P.
        • Lewallen D.G.
        • Taunton M.J.
        Prior bariatric surgery may decrease the rate of re-operation and revision following total hip arthroplasty.
        Bone Jt Lett J. 2016; 98-B: 1180-1184
        • Severson E.P.
        • Singh J.A.
        • Browne J.A.
        • Trousdale R.T.
        • Sarr M.G.
        • Lewallen D.G.
        Total knee arthroplasty in morbidly obese patients treated with bariatric surgery: a comparative study.
        J Arthroplast. 2012; 27: 1696-1700
        • Ighani Arani P.
        • Wretenberg P.
        • Ottosson J.
        • Robertsson O.
        • Annette W.D.
        Bariatric surgery prior to total knee arthroplasty is not associated with lower risk of revision: a register-based study of 441 patients.
        Acta Orthop. 2021; 92: 97-101
        • Ahlin S.
        • Peltonen M.
        • Sjoholm K.
        • Anveden A.
        • Jacobson P.
        • Andersson-Assarsson J.C.
        • et al.
        Fracture risk after three bariatric surgery procedures in Swedish obese subjects: up to 26 years follow-up of a controlled intervention study.
        J Intern Med. 2020; 287: 546-557
        • Pocock S.J.
        • Simon R.
        Sequential treatment assignment with balancing for prognostic factors in the controlled clinical trial.
        Biometrics. 1975; 31: 103-115
        • Sjostrom L.
        Review of the key results from the Swedish Obese Subjects (SOS) trial – a prospective controlled intervention study of bariatric surgery.
        J Intern Med. 2013; 273: 219-234
        • Larsson I.
        • Berteus Forslund H.
        • Lindroos A.K.
        • Lissner L.
        • Naslund I.
        • Peltonen M.
        • et al.
        Body composition in the SOS (Swedish Obese Subjects) reference study.
        Int J Obes Relat Metab Disord. 2004; 28: 1317-1324
        • Ludvigsson J.F.
        • Andersson E.
        • Ekbom A.
        • Feychting M.
        • Kim J.L.
        • Reuterwall C.
        • et al.
        External review and validation of the Swedish national inpatient register.
        BMC Public Health. 2011; 11: 450
        • Fine J.P.
        • Gray R.J.
        A proportional hazards model for the subdistribution of a competing risk.
        J Am Stat Assoc. 1999; 94: 496-509
        • Carlsson L.M.S.
        • Sjoholm K.
        • Jacobson P.
        • Andersson-Assarsson J.C.
        • Svensson P.A.
        • Taube M.
        • et al.
        Life expectancy after bariatric surgery in the Swedish obese subjects study.
        N Engl J Med. 2020; 383: 1535-1543
        • Peltonen M.
        • Lindroos A.K.
        • Torgerson J.S.
        Musculoskeletal pain in the obese: a comparison with a general population and long-term changes after conventional and surgical obesity treatment.
        Pain. 2003; 104: 549-557
        • Burkard T.
        • Holmberg D.
        • Wretenberg P.
        • Thorell A.
        • Hügle T.
        • Burden A.M.
        The associations between bariatric surgery and hip or knee arthroplasty, and hip or knee osteoarthritis: propensity score-matched cohort studies.
        Osteoarthr Cartil Open. 2022; 4
        • Felson D.T.
        • Zhang Y.
        • Anthony J.M.
        • Naimark A.
        • Anderson J.J.
        Weight loss reduces the risk for symptomatic knee osteoarthritis in women. The Framingham Study.
        Ann Intern Med. 1992; 116: 535-539
        • Atukorala I.
        • Makovey J.
        • Lawler L.
        • Messier S.P.
        • Bennell K.
        • Hunter D.J.
        Is there a dose-response relationship between weight loss and symptom improvement in persons with knee osteoarthritis?.
        Arthritis Care Res. 2016; 68: 1106-1114
        • Gersing A.S.
        • Schwaiger B.J.
        • Nevitt M.C.
        • Zarnowski J.
        • Joseph G.B.
        • Feuerriegel G.
        • et al.
        Weight loss regimen in obese and overweight individuals is associated with reduced cartilage degeneration: 96-month data from the osteoarthritis initiative.
        Osteoarthr Cartil. 2019; 27: 863-870
        • Jin X.
        • Gibson A.A.
        • Gale J.
        • Schneuer F.
        • Ding D.
        • March L.
        • et al.
        Does weight loss reduce the incidence of total knee and hip replacement for osteoarthritis? – A prospective cohort study among middle-aged and older adults with overweight or obesity.
        Int J Obes. 2021; 45: 1696-1704
        • Stefanik J.J.
        • Felson D.T.
        • Apovian C.M.
        • Niu J.
        • Margaret Clancy M.
        • LaValley M.P.
        • et al.
        Changes in pain sensitization after bariatric surgery.
        Arthritis Care Res. 2018; 70: 1525-1528
        • Joseph G.B.
        • McCulloch C.E.
        • Nevitt M.C.
        • Lynch J.
        • Lane N.E.
        • Link T.M.
        Effects of weight change on knee and hip radiographic measurements and pain over 4 years: data from the osteoarthritis initiative.
        Arthritis Care Res. 2022;
        • Berend K.R.
        • Lombardi Jr., A.V.
        • Mallory T.H.
        • Adams J.B.
        • Groseth K.L.
        Early failure of minimally invasive unicompartmental knee arthroplasty is associated with obesity.
        Clin Orthop Relat Res. 2005; 440: 60-66
        • Dewan A.
        • Bertolusso R.
        • Karastinos A.
        • Conditt M.
        • Noble P.C.
        • Parsley B.S.
        Implant durability and knee function after total knee arthroplasty in the morbidly obese patient.
        J Arthroplast. 2009; 24 (94 e1–3): 89-94
        • Workgroup of the American Association of H, Knee Surgeons Evidence Based C
        Obesity and total joint arthroplasty: a literature based review.
        J Arthroplast. 2013; 28: 714-721
        • Ricciardi B.F.
        • Giori N.J.
        • Fehring T.K.
        Clinical faceoff: should orthopaedic surgeons have strict BMI cutoffs for performing primary TKA and THA?.
        Clin Orthop Relat Res. 2019; 477: 2629-2634
        • Kristensson F.M.
        • Andersson-Assarsson J.C.
        • Svensson P.A.
        • Carlsson B.
        • Peltonen M.
        • Carlsson L.M.S.
        Effects of bariatric surgery in early- and adult-onset obesity in the prospective controlled Swedish obese subjects study.
        Diabetes Care. 2020; 43: 860-866
        • Flugsrud G.B.
        • Nordsletten L.
        • Espehaug B.
        • Havelin L.I.
        • Meyer H.E.
        Weight change and the risk of total hip replacement.
        Epidemiology. 2003; 14: 578-584
        • Hunter D.J.
        • Bierma-Zeinstra S.
        Osteoarthritis.
        Lancet. 2019; 393: 1745-1759
        • Nelson L.R.
        • Bulun S.E.
        Estrogen production and action.
        J Am Acad Dermatol. 2001; 45: S116-S124
        • Escobar-Morreale H.F.
        • Santacruz E.
        • Luque-Ramirez M.
        • Botella Carretero J.I.
        Prevalence of ‘obesity-associated gonadal dysfunction’ in severely obese men and women and its resolution after bariatric surgery: a systematic review and meta-analysis.
        Hum Reprod Update. 2017; 23: 390-408
        • Bourne R.
        • Mukhi S.
        • Zhu N.
        • Keresteci M.
        • Marin M.
        Role of obesity on the risk for total hip or knee arthroplasty.
        Clin Orthop Relat Res. 2007; 465: 185-188
        • Lim Y.Z.
        • Wang Y.
        • Cicuttini F.M.
        • Giles G.G.
        • Graves S.
        • Wluka A.E.
        • et al.
        Obesity defined by body mass index and waist circumference and risk of total knee arthroplasty for osteoarthritis: a prospective cohort study.
        PLoS One. 2021; 16e0245002
        • W-Dahl A
        • Kärrholm J.
        • Rogmark C.
        • Nauclér E.
        • Nåtman J.
        • Bülow E.
        • et al.
        The Swedish Arthroplasty Register Annual Report 2021.
        2021