Advertisement
British Journal of Anaesthesia
BJA

The association between obesity and disability in survivors of joint surgery: analysis of the health and retirement study

  • T.G. Gaulton
    Correspondence
    Corresponding author.
    Affiliations
    Department of Anesthesiology and Critical Care, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, USA

    Leonard Davis Institute for Health Economics, University of Pennsylvania, Philadelphia, PA, USA

    Center for Perioperative Outcomes Research and Transformation, University of Pennsylvania, Philadelphia, PA, USA
    Search for articles by this author
  • L.A. Fleisher
    Affiliations
    Department of Anesthesiology and Critical Care, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, USA

    Leonard Davis Institute for Health Economics, University of Pennsylvania, Philadelphia, PA, USA

    Center for Perioperative Outcomes Research and Transformation, University of Pennsylvania, Philadelphia, PA, USA
    Search for articles by this author
  • M.D. Neuman
    Affiliations
    Department of Anesthesiology and Critical Care, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, USA

    Leonard Davis Institute for Health Economics, University of Pennsylvania, Philadelphia, PA, USA

    Center for Perioperative Outcomes Research and Transformation, University of Pennsylvania, Philadelphia, PA, USA
    Search for articles by this author
Open ArchivePublished:November 20, 2017DOI:https://doi.org/10.1016/j.bja.2017.11.011

      Abstract

      Background

      Obesity is associated with osteoarthritis and the need for joint surgery. Obese patients who undergo joint surgery may have a higher risk of morbidity compared with normal or overweight patients but less is known about their risk of postoperative disability. The primary objective of our study was to determine the association between obesity and the development of new dependence in activities of daily living within 2 years after joint surgery.

      Methods

      We obtained data from the Health and Retirement Study, a longitudinal survey of older Americans. We included participants who indicated having joint surgery for arthritis. We defined obesity as a BMI ≥30 kg m-2. Our outcome was a new or increased dependence in one or more activities of daily living after surgery.

      Results

      We analysed data on 2519 respondents who underwent joint surgery for arthritis. Respondents had a median age of 69yr, 65.5% were female, 66.6% had joint replacement surgery and 45.3% were obese. The overall incidence of a new dependence within 2years was 22.1%. Obese respondents had a higher incidence of new dependence compared with non-obese respondents (25.4% vs 19.4%, P<0.001). In adjusted analysis, obese respondents had increased odds of developing dependence [odds ratio 1.35 (95% CI 1.09-1.68), P=0.007].

      Conclusions

      Obesity is associated with an increased risk of developing dependence in the 2 years after joint surgery. Our study findings identify a high-risk group that may benefit from targeted interventions and allocation of perioperative resources to optimize recovery and minimize longer-term disability.

      Key words

      • Obesity is associated with increased risk of postoperative complications, and this may exacerbate ongoing disability
      • Hip surgery had higher rates of longer-term disability when compared with knee surgery in obese patients
      Obesity is a worldwide epidemic with a prevalence exceeding one-quarter of the population in developed countries.
      • Ng M.
      • Fleming T.
      • Robinson M.
      • et al.
      Global, regional and national prevalence of overweight and obesity in children and adults 1980-2013: a systematic analysis.
      Patients with an obese BMI are more likely to undergo joint surgery because of a higher prevalence of arthritis compared with patients with a normal or overweight BMI.
      • Derman P.B.
      • Fabricant P.D.
      • David G.
      The role of overweight and obesity in relation to the more rapid growth of total knee arthroplasty volume compared with total hip arthroplasty volume.
      • 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.
      • Anandacoomarasamy A.
      • Caterson I.
      • Sambrook P.
      • Fransen M.
      • March L.
      The impact of obesity on the musculoskeletal system.
      Prior research has shown that obese patients have higher rates of postoperative complications after joint surgery, such as infection, poor wound healing and implant failure.
      • Rodriguez-Merchan E.C.
      The influence of obesity on the outcome of TKR: can the impact of obesity be justified from the viewpoint of the overall health care system?.
      • McElroy M.J.
      • Pivec R.
      • Issa K.
      • Harwin S.F.
      • Mont M.A.
      The effects of obesity and morbid obesity on outcomes in TKA.
      • Huddleston J.I.
      • Wang Y.
      • Uquillas C.
      • Herndon J.H.
      • Maloney W.J.
      Age and obesity are risk factors for adverse events after total hip arthroplasty.
      Obese patients also appear to have worse joint function and mobility.
      • Busato A.
      • Röder C.
      • Herren S.
      • Eggli S.
      Influence of high BMI on functional outcome after total hip arthroplasty.
      • Vincent H.K.
      • Horodyski M.
      • Gearen P.
      • et al.
      Obesity and long term functional outcomes following elective total hip replacement.
      Less is known about the effect of obesity on postoperative disability as defined by a dependence in an activity of daily living (ADL). Obese patients may be vulnerable to developing new or worsening disability after surgery for several reasons. Obesity is associated with increased dependence and restricted function at baseline and has become one of the biggest contributors to the burden of disease and lost quality-adjusted life years.
      • Jia H.
      • Lubetkin E.I.
      Trends in quality-adjusted life-years lost contributed by smoking and obesity.
      Relatedly, there is a high co-prevalence of conditions such as depression and chronic pain that are themselves detrimental to recovery. This elevated incidence of postoperative complications may further impair recovery and independence. Finally, postoperative services including physical therapy, home care and rehabilitation facilities that are fundamental parts of effective recovery from joint surgery may be underequipped and underprepared to deal with the increasing number of obese patients who are undergoing joint surgery.
      • Zhang N.
      • Li Y.
      • Temkin-Greener H.
      Prevalence of obesity in New York nursing homes: associations with facility characteristics.
      Understanding how obesity effects postoperative disability would allow for more informed surgical decision making and let obese patients better gauge how surgery would change their quality of life. This is especially important for caregivers of elderly obese patients where the decision to undergo surgery is difficult and information to guide decisions is scarce. As there is increasing focus on disability as a patient-centred outcome, identifying subgroups of patients that are at higher risk is an important component to improving the overall value of healthcare. Furthermore, it would facilitate appropriate allocation of postoperative resources in a healthcare system with increasing focus on cost. With the proliferation of bundle payments for joint surgery, the value of different forms of post-acute care has become an area of intense scrutiny. Therefore, the research would aim attention to a high-risk group that would require more research directed to understanding the influence of medical comorbidities on return to function and interventions to optimize care.
      The University of Michigan Health and Retirement Study (HRS) is a longitudinal survey of health and ageing conducted in over 20 000 adults in the USA since 1992. The HRS offers a unique opportunity to explore how obesity affects postoperative disability, particularly in an ageing surgical population. The primary aim of our study was to define the association between obesity and the development of ADL dependence after joint surgery.

      Methods

      Data were obtained from HRS, which is sponsored by the National Institute on Aging and is conducted by the University of Michigan. HRS is a longitudinal panel study that surveys approximately 20 000 people in the USA over the age of 50.
      • Juster F.T.
      • Suzman R.
      An overview of the health and retirement study.
      Respondents are interviewed in clusters approximately every 2 years on a wide array of topics including health status and disability. When a respondent is unable to be interviewed, often because of cognitive or medical problems, a proxy is contacted to answer questions for that respondent. We used the RAND HRS Data file, an easy to use longitudinal data set based on the HRS data (available at http://www.rand.org/labor/aging/dataprod/hrs-data.html). It was developed at RAND with funding from the National Institute on Aging and the Social Security Administration.

      RAND HRS Data, Version P. Produced by the RAND Center for the Study of Aging, with funding from the National Institute on Aging and the Social Security Administration. Santa Monica, CA. Available from: https://www.rand.org/labor/aging/dataprod/hrs-data.html [Accessed 11 November 2016]

      Respondents provided informed consent at the time of enrolment into HRS. Our analysis of public HRS data was deemed exempt from human subjects review by the University of Pennsylvania institutional review board.
      We obtained data from the 2004, 2006, 2008, 2010, 2012 and 2014 HRS interview clusters. We included subjects in our sample if they responded yes to the following question in HRS indicating receipt of joint surgery: 'in the last two years, have you had surgery or any joint replacement because of arthritis?' (Fig. 1). Respondents were excluded if they did not have a BMI reported in the interview cluster preceding the interview in which the patient reported having surgery. Our primary analysis excluded respondents who had an underweight BMI (<18.5 kg m-2) given that underweight patients have high rates of disability and their inclusion into the reference group might underestimate the effect of obesity;
      • Ferraro K.F.
      • Su Y.
      • Gretebeck R.J.
      • Black D.R.
      • Badylak S.F.
      Body mass index and disability in adulthood: a 20-year panel study.
      however, these respondents were included in secondary analyses to determine the effect of their inclusion on the primary outcome in a sensitivity analysis. Additionally, as the primary outcome is a new or increased dependence in an ADL, we excluded patients who had complete dependence on all ADLs at baseline and therefore were unable to have the primary outcome. If respondents indicated having surgery more than once, only the first episode of surgery was used in the analysis.
      Figure thumbnail gr1
      Fig 1Creation of analytical dataset. HRS, Health and Retirement Study.
      We restricted our study sample to HRS respondents who were alive at the time of the post-surgical interview and indicated that they had joint surgery. For respondents who died, HRS conducts an exit interview with a proxy to obtain information about the deceased respondent; however, exit interviews do not ask about joint surgery prior to death, preventing identification of decedents with recent joint surgery.

       Definition of obesity

      We defined obesity as a self-reported BMI of ≥ 30 kg m-2. Our reference group included respondents with a normal or overweight BMI, defined as a BMI ≥18.5 and <30 kg m-2. In a secondary analysis, we looked at baseline characteristics and the primary outcome across BMI categories. We defined BMI categories as follows: normal BMI (≥18.5 and <25 kg m-2), overweight BMI (≥25 and <30 kg m-2), obese BMI (≥ 0 and <40 kg m-2) and morbidly obese BMI (≥40 kg m-2).

      WHO. Physical status: the use and interpretation of anthropometry. Report of a WHO Expert Committee. WHO Technical Report Series 854. Geneva: World Health Organization, 1995. Available from: http://www.who.int/childgrowth/publications/physical_status/en/ [Accessed 15 January 2017]

       Definition of disability

      During each interview, HRS interviewers ask respondents or their proxy if they have any difficulty because of a physical, mental, emotional or memory problem with each of six ADLs. The ADLs assessed in HRS include bathing, dressing, walking, toileting, eating and getting in or out of bed. If a respondent had difficulty with an ADL, HRS then asked if the respondent required help with that ADL. We defined the primary outcome as an increase in the level of dependence in one or more ADLs at the post-surgical interview compared with the pre-surgical interview. Our primary outcome includes respondents who had no dependence in an ADL at baseline and then had difficulty or required help in that ADL after surgery as well as respondents who had difficulty with an ADL at baseline and then required help after surgery. We also assessed differences between obese and non-obese respondents for each individual ADL. In a secondary analysis, we further divided ADLs into two groups based on their relation to lower extremity joint function. The ADLs that might be more dependent on successful joint function were walking and getting in or out of bed. Eating, bathing, dressing and toileting were felt to be less reliant on lower extremity joint function and were grouped separately.
      HRS conducts interviews approximately every 2 years. As HRS does not contain the specific date of surgery, we were unable to determine the timing of ADL dependence following surgery. However, HRS does document the date of interview and we were therefore able to determine the time between the interview that the respondent indicated having surgery and the preceding interview to provide an approximate time frame of follow-up.

       Covariates

      We obtained baseline demographic and clinical covariates from the most recent interview wave before surgery. Covariates were defined based on whether the patient ever had the condition prior to surgery. Depression was measured on the 8-point Center for Epidemiological Studies-Depression (CES-D) scale. Total wealth was adjusted to 2011 values using the Consumer Price index inflation adjustment and numerical cut-offs were selected based on median values from the 2011 Census.

      U.S. Census Bureau. Distribution of Household Wealth in the U.S.: 2000 to 2011. Available from: https://www.census.gov/people/wealth/files/Wealth%20distribution%202000%20to %202011.pdf [Accessed 15 January 2017]

      Based on HRS items, we dichotomized type of surgery into joint replacement surgery and surgery without joint replacement. Furthermore, we categorized the location of joint surgery into hip, knee or other joint, which includes small joint, shoulder and spine surgery as examples. Given the possibility that obese respondents have different types of joint surgery compared with non-obese respondents, we reported the incidence of ADL dependence by type of joint surgery.

       Statistical analysis

      We calculated that we would need 1442 respondents to detect a 30% increase in the incidence of a new ADL dependence in obese respondents compared with non-obese respondents with a power of 80% and a two-tailed a of 0.05. We used prior work within HRS to estimate the incidence of ADL dependence in non-obese respondents.
      • Neuman M.D.
      • Werner R.M.
      Marital status and postoperative functional recovery.
      Continuous variables were compared using the Wilcoxon Rank Sum test and summarized as medians with inter-quartile ranges (IQR). Categorical variables were compared by the χ2 test and summarized as number of patients with proportions. We conducted bivariable analysis to determine the association between clinical and demographic variables and ADL dependence following surgery. We then performed a stratified analysis, as planned a priori, by age and type of joint surgery. We assumed an interaction was present when the test for homogeneity between the odds ratios (ORs) for different strata yielded a significant result. For multivariable analysis, we used logistic regression. We included the type of joint surgery and the location of the operated joint as independent covariates a priori into the multivariable model if there was no statistical interaction. We also included the interview year a priori to account for potential variations over time. Additional baseline covariates with P-values <0.20 on bivariate analyses were introduced sequentially into the multivariable model and retained if their inclusion resulted in a ≥15% change in the OR of the association between obesity and the development of ADL dependence or if they were independent predictors of the outcome based on a P-value of <0.10.

      Hosmer DW, Lemeshow S. Model building strategies and methods for logistic regression. In: Hosmer DW, Lemeshow S, editors. Applied Logistic Regression, 2nd ed. New York: John Wiley & Sons; 2000. p. 89–150

      Multivariable models were also developed for the association between obesity and different ADL categories. Finally, we performed additional univariable analyses to look at the change in respondent weight in pounds and the change in frequency of moderate or vigorous physical activity in the 2 years after joint surgery. The analysis of activity was limited to interview waves after 2004 because of a change in how HRS captured activity.
      For all calculations, a two-tailed P-value of <0.05 was considered significant. All statistical calculations were performed using STATA 12.0 (2011, StataCorp LP, College Station, TX, USA).

      Results

      Over the full study period, 3748 HRS subjects responded yes to having joint surgery for arthritis, and 2519 respondents met our inclusion criteria (Fig. 1). 1140 (45.3%) respondents were obese and 1379 (54.7%) respondents were not obese. Obese respondents were more likely to have had joint replacement surgery and knee surgery than non-obese respondents (Table 1). Obese respondents were also younger, had lower median total adjusted wealth, fewer total years of education, and a higher incidence of self-rated poor health, diabetes and hypertension. They were also more likely to have difficulty or require help with an ADL prior to surgery (Table 2). Non-obese respondents were more likely to be white, older, reside in a nursing home and require a proxy to complete the interview. The median time between the pre-surgical to the post-surgical interview was 730 (IQR 669-792) days and there was no difference in the timing between interviews by BMI category (P=0.33).
      Table 1Baseline characteristics of 2519 survivors of joint surgery. CES-D, Center for Epidemiology Studies-Depression
      Characteristics n (%) or median (IQR)Obese 1140 (45.3)Non-obese 1379 (54.7)P-value
      Age (yr)67 (61-73)71 (64-78)<0.001
      Female756 (66.3)894 (64.8)0.44
      RaceWhite899 (78.7)1199 (87.0)<0.001
      Black193 (16.9)119 (8.6)
      Years of education<12259 (22.7)241 (17.5)0.001
      =12411 (36.1)484 (35.1)
      >12470 (41.2)654 (47.4)
      Total adjusted wealth (2011 dollars)114871 (3037-343613)206960 (65 077-484 330)<0.001
      Married787 (69.0)946 (68.7)0.84
      Smoker (current or former)120 (10.5)163 (11.8)0.31
      Residence in nursing home22 (1.93)50 (3.63)0.01
      Self-rated healthPoor144 (12.6)121 (8.77)<0.001
      Fair320 (28.1)272 (19.7)
      Good414 (36.3)488 (35.4)
      Very good226 (19.8)408 (29.6)
      Excellent36 (3.16)90 (6.53)
      Hypertension819 (71.8)739 (53.6)<0.001
      Heart disease262 (23.0)318 (23.1)0.96
      Diabetes308 (27.0)158 (11.5)<0.001
      Psychiatric disorder241 (21.1)251 (18.2)0.06
      Cancer152 (13.3)218 (15.8)0.08
      Stroke93 (8.16)100 (7.25)0.40
      Chronic lung disease125 (11.0)128 (9.28)0.16
      Memory impairment23 (2.02)38 (2.76)0.23
      Proxy respondent37 (3.25)78 (5.66)0.004
      Pain that limits activities587 (51.5)550 (39.9)<0.001
      CES-D score1 (0-3)1 (0-3)<0.001
      Joint replacement783 (68.7)895 (64.9)0.05
      Type of surgeryKnee753 (66.1)627 (45.5)<0.001
      Hip153 (13.4)321 (23.3)
      Other234 (20.5)431 (31.3)
      Table 2Baseline ADL dependence in 2519 survivors of joint surgery. Number of respondents (%)
      ObeseNon-obeseP-value
      Any baseline ADL dependence*568 (49.8)526 (38.1)<0.001
      EatingDifficulty only50 (4.39)48 (3.48)0.38
      Needs help11 (0.96)18 (1.31)
      BathingDifficulty only126 (11.1)103 (7.47)0.001
      Needs help63 (5.53)56 (4.06)
      ToiletingDifficulty only224 (19.7)215 (15.6)0.006
      Needs help28 (2.46)22 (1.60)
      WalkingDifficulty only201 (17.6)155 (11.2)<0.001
      Needs help51 (4.47)59 (4.28)
      Getting in or out of bedDifficulty only207 (18.2)200 (14.5)<0.001
      Needs help77 (6.75)49 (3.55)
      DressingDifficulty only227 (19.9)202 (14.7)<0.001
      Needs help141 (12.4)86 (6.24)
      Within our full sample, 557 of the 2519 respondents (22.1%) developed a new or increased ADL dependence following joint surgery for arthritis. In comparison, the incidence of new or increased ADL dependence in respondents who did not indicate having joint surgery over a similar 2 year period was 13.2%. Following surgery, 289 of the 1140 obese patients (25.4%) had an increased incidence of new or increased ADL dependence compared with 268 of the 1379 non-obese patients (19.4%) on univariable analysis (P <0.001). Using multivariable analysis (Table 3), an obese BMI was associated with a statistically significant increased risk of developing an ADL dependence following joint surgery compared with a non-obese BMI [OR 1.35 (95% CI: 1.09-1.68), P=0.007].
      Table 3Primary analysis of the association between obesity and new or increased ADL dependence after joint surgery
      Univariable analysisMultivariable analysis
      Odds ratio (95% CI)P-valueOdds ratioP-value (95% CI)
      Obesity1.41 (1.17-1.70)<0.0011.35 (1.09-1.68)0.007
      Age1.01 (1.00-1.02)0.0271.02 (1.01-1.03)0.001
      History of diabetes1.80 (1.44-2.25)<0.0011.29 (1.01-1.66)0.05
      History of psychiatric disorder2.03 (1.63-2.52)<0.0011.31 (1.01-1.70)0.04
      Baseline ADL dependence2.51 (2.07-3.04)<0.0011.29 (1.08-1.56)0.006
      Self-rated healthExcellent1 (Reference)
      Very good0.93 (0.50-1.71)0.820.94 (0.50-1.75)0.84
      Good2.25 (1.26-4.01)0.0061.76 (0.97-3.19)0.06
      Fair3.22 (1.80-5.78)<0.0011.92 (1.05-3.53)0.04
      Poor5.59 (3.05-10.3)<0.0012.17 (1.13-4.17)0.02
      Joint replacement surgery1.08 (0.89-1.33)0.421.01 (0.73-1.41)0.94
      Operative jointHip1 (Reference)
      Knee0.70 (0.55-0.90)0.0050.69 (0.52-0.90)0.007
      Other0.78 (0.59-1.02)0.070.74 (0.52-1.05)0.10
      Total adjusted wealth (2011 dollars)<02.16 (1.54-3.04)<0.0011.67 (1.15-2.42)0.007
      0 to < 68 8281.38 (1.06-1.80)0.021.10 (0.83-1.46)0.50
      68828 to < 205 9851 (Reference)
      >205 9850.72 (0.56-0.91)0.0070.85 (0.65-1.10)0.22
      Interview year0.96 (0.91-1.01)0.110.94 (0.88-0.99)0.03
      The incidence of new dependence (whether having difficulty or needing help) for each ADL is shown in Table 4. For both obese and non-obese respondents, the ADLs that had the highest incidence of dependence were dressing and walking, although there were no statistically significant differences between the two groups for an individual ADL.
      Table 4New or increased ADL dependence in 2519 survivors of joint surgery. Number of respondents (%)
      ObeseNon-obeseP-value
      New or increased ADL dependence*289 (25.4)268 (19.4)<0.001
      EatingNew difficulty16 (1.40)26 (1.89)0.35
      New help19 (1.67)29 (2.10)0.43
      New difficulty or help35 (3.07)55 (3.99)0.22
      BathingNew difficulty33 (2.89)23 (1.67)0.04
      New help44 (3.86)48 (3.48)0.61
      New difficulty or help77 (6.75)71 (5.15)0.09
      ToiletingNew difficulty65 (5.70)62 (4.50)0.62
      New help22 (1.93)23 (1.67)0.17
      New difficulty or Help87 (7.63)85 (6.16)0.15
      WalkingNew difficulty52 (4.56)43 (3.12)0.06
      New help38 (3.33)41 (2.97)0.61
      New difficulty or help90 (7.89)84 (6.09)0.08
      Getting in or out of bedNew difficulty31 (2.72)35 (2.54)0.78
      New help33 (2.89)42 (3.05)0.82
      New difficulty or help64 (5.61)77 (5.58)0.97
      DressingNew difficulty44 (3.86)40 (2.90)0.18
      New help66 (5.79)70 (5.08)0.43
      New difficulty or help110 (9.65)110 (7.98)0.14
      In our secondary analysis, obesity was associated with increased odds of developing new or increased dependence in lower extremity ADLs (walking and getting in or out of bed) compared with having a non-obese BMI [OR 1.38 (95% CI: 1.03–1.84), P=0.03]. The overall incidence of dependence in the lower extremity ADLs was 10.4%. Obesity was also associated with increased odds of developing a new or increased dependence in non-lower extremity ADLs (bathing, eating, dressing and toileting) [OR 1.26 (95% CI: 1.00-1.57), P=0.046]. The overall incidence of dependence in these ADLs was 17.7%.
      To account for potential factors in the casual pathway between obesity and ADL dependence, we assessed changes in respondent weight and the frequency of moderate or vigorous exercise in the 2 years after joint surgery compared with presurgical values. Obese respondents had a median weight loss of 0.23 (IQR -5.44 to +2.72) pounds compared with non-obese respondents who on average had no change in weight (IQR -2.72 to +2.27). Furthermore, we found that most obese and nonobese respondents did not change their frequency of moderate or vigorous exercise after surgery (77.6% us 75.3%, P=0.17).

       ADL dependence across BMI categories

      We assessed baseline characteristics and ADL dependence across BMI categories (Supplementary Table S1). Morbidly obese respondents were younger, had lower median total adjusted wealth, were more likely to rate their health as fair or poor, and had higher incidences of diabetes, hypertension, pain and baseline ADL dependence. In looking at the type of joint surgery, there was no difference in the rates of joint replacement surgery between respondents with an overweight, obese or morbidly obese BMI, although these rates were higher than in normal BMI respondents. Morbidly obese respondents were more likely to have knee surgery while normal and overweight BMI respondents were more likely to have hip surgery. For the primary outcome, morbidly obese respondents had a higher incidence of a new or increased ADL dependence compared with other BMI categories after joint surgery on univariate analysis (P-value for trend <0.001).
      Relatedly, we assessed BMI as a continuous variable to account for the potential upward bias of the primary outcome that can occur when categorizing self-reported BMI.
      • Preston S.H.
      • Fishman E.
      • Stokes A.
      Effects of categorization and self-report bias on estimates of the association between obesity and mortality.
      In multivariable analysis, a 1kgm-2 increase in BMI was associated a 3% increased odds of developing a new or increased ADL dependence (95% CI: 2-5%, P<0.001).

       ADL dependence by type of joint surgery

      We looked at differences in ADL dependence by the type of joint surgery. Respondents who had joint replacement surgery had a similar incidence of ADL dependence compared with respondents who had surgery for arthritis that did not involve joint replacement (22.6% us 21.2%, respectively, P=0.42). Additionally, we found that respondents who had hip surgery had higher incidences of ADL dependence compared with respondents who had knee surgery (26.8% us 20.5%, respectively, P=0.02).

       Sensitivity analysis

      Two hundred and twenty-six respondents did not have an available pre-surgical BMI. Respondents without an available BMI had a median age of 54 yr, 42% were female and 59.7% were white. Fifty per cent underwent knee surgery and 41.2% had joint replacement surgery. Baseline ADL dependence was also not available in respondents without a reported BMI and therefore the incidence of new or increased ADL dependence after surgery could not be measured.
      Twenty respondents had an underweight BMI and were excluded from the primary analysis. Consistent with prior literature showing a positive correlation between being underweight and disability, we found that the incidence of new or increased ADL dependence in the underweight respondents was 40% following joint surgery. Underweight respondents were more likely to have joint replacement surgery (65%) and hip surgery (50%). We did not find that the inclusion of underweight respondents into the primary analysis significantly changed the risk of obesity on developing an ADL dependence after joint surgery [OR 1.33 (95% CI: 1.07-1.65), P=0.01].

      Discussion

      In our analysis of HRS, obesity was associated with an increased risk of developing a new or increased dependence in an ADL within approximately 2 years after joint surgery compared with normal or overweight individuals. To our knowledge, this is the first study to look at the effect of obesity on ADL dependence in joint surgery.
      Obese patients are becoming more likely to have joint surgery given their high rates of arthritis.
      • Derman P.B.
      • Fabricant P.D.
      • David G.
      The role of overweight and obesity in relation to the more rapid growth of total knee arthroplasty volume compared with total hip arthroplasty volume.
      • 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.
      • Anandacoomarasamy A.
      • Caterson I.
      • Sambrook P.
      • Fransen M.
      • March L.
      The impact of obesity on the musculoskeletal system.
      Initial research in joint replacement surgery showed that obese patients had higher rates of complications such as infection, poor wound healing, joint failure and need for surgical revision. Further studies revealed that obese patients had higher pain scores and reduced mobility postoperatively compared with normal weight counterparts.
      • Busato A.
      • Röder C.
      • Herren S.
      • Eggli S.
      Influence of high BMI on functional outcome after total hip arthroplasty.
      • Vincent H.K.
      • Horodyski M.
      • Gearen P.
      • et al.
      Obesity and long term functional outcomes following elective total hip replacement.
      However, these studies have used composite outcome scores such as the Oxford Hip and Knee Scores, which aggregate assessments of pain, mobility and function and may be difficult for patients to interpret and use to make informed surgical decisions. Prior studies on joint surgery have not directly addressed ADL dependence, an important and understandable patient-centred outcome.
      We found that obese respondents had a higher risk of ADL dependence even after adjustment for important predictors of postoperative disability, such as diabetes, socioeconomic status, baseline ADL dependence, respondent age and type of joint surgery. Furthermore, although obese respondents were more likely to have knee surgery and joint replacement surgery, we did not find that the type of joint surgery altered the association between obesity and dependence. In fact, respondents who had knee surgery that lower incidences of ADL dependence compared with respondents who had hip surgery.
      Several mechanisms may account for the observed difference in risk. Obesity is associated with chronic inflammation, which may negatively modulate the response to surgery and increase postoperative complications.
      • Kern P.A.
      • Ranganathan S.
      • Li C.
      • Wood L.
      • Ranganathan G.
      Adipose tissue tumor necrosis factor and interleukin-6 expression in human obesity and insulin resistance.
      Obese patients undergoing joint surgery have higher rates of infection, compromised wound healing and implant failure. These complications can delay recovery and may be a potential mechanism that warrants further investigation. There may also be differences in the quality and availability of postoperative care between obese and non-obese patients, which are especially crucial for return of function following joint surgery. Acute care facilities may be underprepared to deal with the needs of obese patients given the requirement for additional equipment and supplies.
      • Zhang N.
      • Li Y.
      • Temkin-Greener H.
      Prevalence of obesity in New York nursing homes: associations with facility characteristics.
      When we looked at the primary outcome in greater depth, we found that the risk of dependence for respondents with an obese BMI was higher for lower extremity ADLs compared with non-lower extremity ADLs. Our findings are consistent with prior research, which showed that post-surgical joint function for obese patients may be worse than their normal or overweight counterparts.
      • Busato A.
      • Röder C.
      • Herren S.
      • Eggli S.
      Influence of high BMI on functional outcome after total hip arthroplasty.
      • Vincent H.K.
      • Horodyski M.
      • Gearen P.
      • et al.
      Obesity and long term functional outcomes following elective total hip replacement.
      Lower extremity joint surgery (hip and knees) also represented the majority (73.6%) of our study sample. Nonetheless, we still found that, in the non-lower extremity ADLs of bathing, eating, dressing and toileting, respondents with an obese BMI also had an increased risk of dependence. It is also worth noting that the overall incidence of a new or increased dependence in these ADLs was 17.7%, which suggests that there may be substantial vulnerability after surgery, even outside of factors that affect joint function. This was especially true for respondents that required a proxy to complete the interview as they had a 43.5% incidence of new or increased dependence in non-lower extremity ADLs after joint surgery. Respondents that required a proxy had a higher baseline prevalence of memory impairment compared with respondents who did not require assistance to complete the interview (13.0% vs 1.91%). Patients with evidence of cognitive impairment at the time of surgery are an extremely susceptible population and more work is needed to determine how cognition affects postoperative recovery and independence.
      Additionally, we found that the proportion of respondents with any ADL dependence, not just new or increased, was lower after surgery than at baseline, which suggests an overall improvement in disability for patients who have joint surgery for arthritis. The improvement was seen for both obese and non-obese respondents and needs to be considered when making recommendations for the surgical treatment of arthritis. Notably, despite this improvement in ADL dependence, respondents did not have a clinically significant change in weight or a change in the frequency of moderate or vigorous exercise in the 2 years after surgery when compared with presurgical values.
      Lastly, we found that morbidly obese respondents had a higher incidence of developing an ADL dependence even compared with obese respondents. We did not perform multivariable analysis given the small number of morbidly obese respondents. Prior research has shown that morbidly obese patients have an elevated risk of several adverse post-surgical outcomes.
      • McElroy M.J.
      • Pivec R.
      • Issa K.
      • Harwin S.F.
      • Mont M.A.
      The effects of obesity and morbid obesity on outcomes in TKA.
      They may be particularly vulnerable following joint surgery where mobility can be limited and influence functional recovery. This has important implications with regard to establishing appropriate recommendations for how to manage morbidly obese patients with arthritis and whether surgery should be delayed in favour of non-surgical management or more aggressive weight loss reduction through bariatric surgery.
      • 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.
      Our study has several strengths. Data were used from HRS, which is a large, representative survey of adults in the USA and results are generalizable across a range of socioeconomic status, race, geography and family structure. We also used ADL dependence as our primary outcome, an important patient-centred endpoint.
      There are several limitations to the study. HRS contains self-reported data, which may introduce misclassification of the exposure, outcome and/or covariates. However, prior studies using HRS have shown good correlation between self-reported BMI and physical measurements from in-person interviews.
      • Suemoto C.K.
      • Gilsanz P.
      • Mayeda E.R.
      • Glymour M.M.
      Body mass index and cognitive function: the potential for reverse causation.
      Moreover, while HRS does not contain information of the date of surgery, we found no differences in the interview dates between obese and non-obese respondents. Our study sample includes different types of joint surgery. While obese respondents were more likely to have joint replacement surgery and knee surgery, we did not find that the type of surgery alerted the association between obesity and ADL dependence. Furthermore, we felt that the inclusion of different types of joint surgery would be more pragmatic and make our results have stronger external validity. While HRS does not capture information on joint surgery prior to a respondent's death, prior studies have not shown a mortality difference between obese and nonobese patients after joint surgery.
      • Huddleston J.I.
      • Wang Y.
      • Uquillas C.
      • Herndon J.H.
      • Maloney W.J.
      Age and obesity are risk factors for adverse events after total hip arthroplasty.
      Finally, as we are limited to the questions asked in the HRS survey, there is the potential for confounding from unmeasured variables in addition to constraints on the generalizability of our results to other surgeries.

       Summary

      In this retrospective analysis of a large, population-based survey, we found that obesity was associated with an increased risk of developing an ADL dependence following joint surgery. Further study is required to understand the mechanisms of poor functional recovery in obese patients in order to design targeted interventions, effectively allocate resources and better inform patients about postoperative disability.

      Authors' contributions

      Conception and study design, data analysis, interpretation of data, writing up the first draft of the paper, revising the manuscript critically for important intellectual content, final approval of the version to be published: T.G.G.
      Conception and study design, data analysis, interpretation of data, revising the manuscript critically for important intellectual content, final approval of the version to be published: M.D.N.
      Conception and study design, revising the manuscript critically for important intellectual content, final approval of the version to be published: L.A.F.

      Declaration of interest

      None declared.

      Funding

      This work was supported by a grant from the National Institutes of Health (grant number K08 AG043548) to M.D.N.

      Supplementary material

      The following is the supplementary data related to this article:

      References

        • Ng M.
        • Fleming T.
        • Robinson M.
        • et al.
        Global, regional and national prevalence of overweight and obesity in children and adults 1980-2013: a systematic analysis.
        Lancet. 2014; 384: 766
        • Derman P.B.
        • Fabricant P.D.
        • David G.
        The role of overweight and obesity in relation to the more rapid growth of total knee arthroplasty volume compared with total hip arthroplasty volume.
        J Bone Joint Surg Am. 2014; 96: 922-928
        • 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 Rheum. 2006; 54: 802-807
        • Anandacoomarasamy A.
        • Caterson I.
        • Sambrook P.
        • Fransen M.
        • March L.
        The impact of obesity on the musculoskeletal system.
        Int J Obes. 2007; 32: 211-222
        • Rodriguez-Merchan E.C.
        The influence of obesity on the outcome of TKR: can the impact of obesity be justified from the viewpoint of the overall health care system?.
        HSSJ. 2014; 10: 167
        • McElroy M.J.
        • Pivec R.
        • Issa K.
        • Harwin S.F.
        • Mont M.A.
        The effects of obesity and morbid obesity on outcomes in TKA.
        J Knee Surg. 2013; 26: 083-088
        • Huddleston J.I.
        • Wang Y.
        • Uquillas C.
        • Herndon J.H.
        • Maloney W.J.
        Age and obesity are risk factors for adverse events after total hip arthroplasty.
        Clin Orthop Relat Res. 2012; 470: 490-496
        • Busato A.
        • Röder C.
        • Herren S.
        • Eggli S.
        Influence of high BMI on functional outcome after total hip arthroplasty.
        Obes Surg. 2008; 18: 595-600
        • Vincent H.K.
        • Horodyski M.
        • Gearen P.
        • et al.
        Obesity and long term functional outcomes following elective total hip replacement.
        J Orthop Surg Res. 2012; 7: 16
        • Jia H.
        • Lubetkin E.I.
        Trends in quality-adjusted life-years lost contributed by smoking and obesity.
        Am J Prev Med. 2010; 38: 138-144
        • Zhang N.
        • Li Y.
        • Temkin-Greener H.
        Prevalence of obesity in New York nursing homes: associations with facility characteristics.
        Gerontologist. 2013; 53: 567-581
        • Juster F.T.
        • Suzman R.
        An overview of the health and retirement study.
        J Hum Resour. 1995; 30: S7
      1. RAND HRS Data, Version P. Produced by the RAND Center for the Study of Aging, with funding from the National Institute on Aging and the Social Security Administration. Santa Monica, CA. Available from: https://www.rand.org/labor/aging/dataprod/hrs-data.html [Accessed 11 November 2016]

        • Ferraro K.F.
        • Su Y.
        • Gretebeck R.J.
        • Black D.R.
        • Badylak S.F.
        Body mass index and disability in adulthood: a 20-year panel study.
        Am J Public Health. 2002; 92: 834-840
      2. WHO. Physical status: the use and interpretation of anthropometry. Report of a WHO Expert Committee. WHO Technical Report Series 854. Geneva: World Health Organization, 1995. Available from: http://www.who.int/childgrowth/publications/physical_status/en/ [Accessed 15 January 2017]

      3. U.S. Census Bureau. Distribution of Household Wealth in the U.S.: 2000 to 2011. Available from: https://www.census.gov/people/wealth/files/Wealth%20distribution%202000%20to %202011.pdf [Accessed 15 January 2017]

        • Neuman M.D.
        • Werner R.M.
        Marital status and postoperative functional recovery.
        JAMA Surgery. 2016; 151: 194
      4. Hosmer DW, Lemeshow S. Model building strategies and methods for logistic regression. In: Hosmer DW, Lemeshow S, editors. Applied Logistic Regression, 2nd ed. New York: John Wiley & Sons; 2000. p. 89–150

        • Preston S.H.
        • Fishman E.
        • Stokes A.
        Effects of categorization and self-report bias on estimates of the association between obesity and mortality.
        Ann Epidemiol. 2015; 25: 907
        • Kern P.A.
        • Ranganathan S.
        • Li C.
        • Wood L.
        • Ranganathan G.
        Adipose tissue tumor necrosis factor and interleukin-6 expression in human obesity and insulin resistance.
        Am J Physiol Endocrinol Metabol. 2001; 280: E745-E751
        • 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 Joint J. 2016; 98-B: 1180-1184
        • Suemoto C.K.
        • Gilsanz P.
        • Mayeda E.R.
        • Glymour M.M.
        Body mass index and cognitive function: the potential for reverse causation.
        Int J Obes. 2015; 39: 1383