Flugsrud, Gunnar B.*; Nordsletten, Lars*; Espehaug, Birgitte†; Havelin, Leif I.‡; Meyer, Haakon E.‡§
Primary osteoarthritis of the hip is a major reason for pain and disability in the elderly. Because of this condition, 2 to 4 persons per thousand in the age group 60-80 years undergo a total hip replacement each year in the Nordic countries.1 A strong hereditary component has been demonstrated.2-5 The incidence increases with age and is higher in women.6 We have previously shown that a high body mass index (BMI), as well as physical activity at work, are important risk factors for later total hip replacement resulting from primary osteoarthritis.7 Other authors have shown previous hip injury,8,9 sports activity,10-13 and the presence of Heberden’s nodes9 to be associated with increased risk for primary hip osteoarthritis. Estrogen replacement therapy could be protective,14-17 whereas the effects of smoking habits and bone mineral density need further elucidation.
Weight reduction has proven useful in the conservative treatment of symptomatic knee osteoarthritis,18,19 and it has also been associated with reduced risk for later osteoarthritis of the knee.20,21 Little is known about the effect of middle-age weight reduction on the risk for later osteoarthritis of the hip.
The aim of this study was to investigate the association between adult weight change and the risk for later total hip replacement resulting from primary osteoarthritis. We also wanted to evaluate the association between BMI and the risk for later total hip replacement across a spectrum of ages to see whether the association changes with age.
During the years 1974-1978 and 1977-1983, the National Health Screening Service twice screened the population in 3 Norwegian counties for risk factors related to cardiovascular disease (Table 1). Participants answered a questionnaire concerning lifestyle factors, and screening nurses measured body height and weight in a standardized way. The number of people invited to both screenings was 47,797. For 40,739 of these, information on body height and weight was available from both screenings as well as from a nationwide compulsory tuberculosis screening completed from 1963-1975. Using the national personal identification code, we combined information from these screenings with data from the National Arthroplasty Register initiated in 1987. Reporting to the Register improved gradually, and since 1989 this register has captured more than 95% of all total hip replacements performed in Norway,22,23 including information on indications for surgery. We also used data from the Norwegian Registry of Vital Statistics on death and emigration during follow up. We excluded from the analysis all subjects in the screened cohort who died, emigrated, or underwent total hip replacement before the start of follow up. This left an eligible study group of 38,868 participants (81% of those invited to both cardiovascular screenings). The follow-up period in this investigation was from January 1, 1989, to June 16, 1998. The Norwegian Board of Health, the Data Inspectorate, and the Regional Committee on Ethics in Medical Research approved the study.
For each participant, weight change was assessed by linear regression of body weight over time. The slope (kg/year) of this straight line gave an estimate of each participant’s weight change.24,25 We also computed absolute weight change (in kilograms) as the difference between body weight recorded at the second cardiovascular screening and the tuberculosis screening. Relative weight change was computed as absolute weight change divided by body weight at the first examination.
We computed age-adjusted and multivariate adjusted relative risks using Cox proportional hazard regression. In the Cox analysis each subject’s survival time is computed either to an event (a first total hip replacement resulting from primary osteoarthritis) or to censoring (any incident other than an event, eg, death or end of registration, occurring after the start of follow up, that precludes the participant from later experiencing an event). Relative risks were estimated with incidence rate ratios. Multivariate analysis was stratified on sex and adjusted for the continuous variables age-at-screening and BMI (calculated as the mean from the 3 screenings). We also adjusted for height as a continuous variable and for the categorical variables of physical activity at work, physical activity during leisure, smoking, and marital status, as recorded at the second cardiovascular screening. Physical activity at work and in leisure were classified independently as sedentary, moderate, intermediate, or intensive.7 Smoking was classified as current, former, or never. Marital status was classified as unmarried, married, separated, divorced, or widowed. To evaluate the effect of body weight, a separate analysis was performed using weight in place of BMI.
To detect a possible change in the association between BMI and later total hip replacement with increasing age, we performed 3 separate Cox analyses, including BMI recorded at the first, second, or third screening. All 3 analyses were adjusted for age at screening, height, physical activity at work and in leisure, smoking, and marital status.
The study group included 18,705 men and 20,163 women (Table 2). The 3 screenings covered a mean time span of 13 years. From the first to the last screening, the mean body weight of both men and women increased 3.3% and mean BMI increased 3.7% (Table 3). Mean height decreased only very slightly (0.17%), so that body weight and BMI changed proportionately throughout the 3 screenings.
Mean age at start of follow-up was 55 years. During follow up, 799 participants received their first total hip replacement. The indication for surgery was primary osteoarthritis for 572 (235 men and 337 women). The 227 participants who had a total hip replacement for other indications were censored. The most common reasons other than primary osteoarthritis for total hip replacement were sequelae to a hip fracture (n = 95) and rheumatoid arthritis (n = 26). Participants were also censored during follow up as a result of emigration (n = 32) or death (n = 3107). The majority of participants were censored at the end of follow up (n = 34,930).
The rate of weight change ranged from −4.4 kg/year to 8.2 kg/year. The 25th, 50th, and 75th percentiles were −0.07 kg/year, 0.18 kg/year, and 0.45 kg/year, respectively, for men, and −0.11 kg/year, 0.14 kg/year, and 0.41 kg/year, respectively, for women. In both men and women, those with the highest weight gain had a higher mean BMI and mean body weight, and showed somewhat lower levels of physical activity (Tables 4 and 5). The proportion of exsmokers was highest among participants who gained most in weight. The incidence of total hip replacement showed no association with weight slope (Table 6). The age-adjusted and multivariate-adjusted relative risks for total hip replacement were also unaffected by weight slope.
We divided the cohort into quartiles according to BMI as recorded at the first screening, and performed the Cox analysis for each subgroup separately. There was no association between weight change and risk for later total hip replacement in any of the BMI categories. Including information on cardiovascular disease, diabetes mellitus, sick leave, or sick leave pension in the regression model did not alter the results (data not shown).
We tested for effect modification of weight change by each of the other variables used in the model. Only for weight change and age at screening in women did the interaction term reach statistical significance (P = 0.025). Stratifying the analysis on both sex and age at the second cardiovascular screening, none of the subgroups showed any association between high weight gain and risk for later total hip replacement. On the contrary, among women who were 50 years old or more at the second cardiovascular screening, those who were in the third and fourth quartiles of weight change showed a somewhat reduced risk for later total hip replacement compared with the lowest weight change quartile (relative risk [RR] = 0.60, 95% confidence interval (CI) = 0.38-0.95 for the third quartile, and RR = 0.51, CI = 0.32-0.82 for the fourth quartile).
In another analysis, we used weight change as a continuous variable in the multivariate analysis, and found a relative risk for men with a weight increase of 10 kg per 10 years of 0.93 (CI 0.69-1.25) compared with a man with no weight change. For women the corresponding relative risk was 0.85 per kg/year (CI 0.70-1.04).
In additional analyses, we compared the 5% of participants with the highest rate of weight gain (male >0.96 kg/year, female >0.94 kg/year) with the rest of the cohort and then the 5% of participants with highest rate of weight loss (male <-0.45 kg/year, female <-0.59 kg/year) with the other 95%. We found no differences in the risk for later total hip replacement (data not shown).
In a further analysis, we used the continuous variable absolute weight change rather than weight slope as an indicator of weight change. A man with 10-kg weight reduction from the first to the last screening had a relative risk for total hip replacement of 1.11 (CI 0.91-1.52) compared with a man with no weight change. For women the corresponding relative risk was 1.11 (CI 1.00-1.35). Using relative weight change as the indicator of weight change, there was no effect on the risk for total hip replacement (data not shown).
To detect any change, with increasing age, on the association between BMI and risk for later total hip replacement, we first found the relative risk for total hip replacement related to BMI at the tuberculosis screening (2.84 per 10 kg/m2 CI 1.79-4.41 for men, and 2.37 per 10 kg/m2 CI 1.97-3.11 for women). We then computed the relative risk for total hip replacement related to BMI as recorded at the next 2 screenings. There were no differences among these relative risks (Table 7).
Adult weight change between the ages of 34 and 47 years did not affect subsequent risk for total hip replacement. This was true irrespective of sex, initial BMI, and the magnitude of weight change. The investigation was strengthened by the population-based design and high response rates, and by the fact that we had screening information on most known confounders. However, information on hormone replacement therapy, previous hip injury, and bone mineral density was not available.
Comparisons between the National Arthroplasty Register and other official Norwegian registries indicate a high level of reporting. After excluding 1 hospital with poor reporting (not located within the counties of the present investigation), the completeness nationwide has been reported as 97%.22 During 1991-1995, 69% of the total hip replacements recorded in the Norwegian Arthroplasty Register were the result of primary osteoarthritis.6 This is in good agreement with the Swedish National Hip Arthroplasty Register, which during 1992-1996 reported that 69% of 48,007 total hip replacements were done for this indication.26 We have previously compared the Norwegian incidence of total hip replacement with an estimate of the actual demand for the same surgery in England and found similar values.7
Participants were 46-67 years old at the start of follow up and 56-77 at the end. Of all total hip replacements resulting from primary osteoarthritis recorded in the Norwegian Arthroplasty Register from the 3 counties, 3.8% were performed on patients younger than 55 years and 40.6% on patients aged 55-70 years (B. Espehaug, written personal communication, 1999). From 1991-1995, 13% of total hip replacements in the Norwegian Arthroplasty Register were the result of sequelae after fracture of the femoral neck, 8% the result of sequelae of hip dysplasia, and 4% the result of rheumatoid arthritis. This is very similar to the frequencies of indications recorded in the present investigation, suggesting that the total hip replacements that occurred during our follow up are representative of all total hip replacements that our study population would sustain during their lifetime.
Body weight and BMI are risk factors for osteoarthritis in the knee and hip. The association is considered to be stronger for knee osteoarthritis.27,28 However, the association we previously reported between BMI and hip osteoarthritis21 is very similar to that described between BMI and knee osteoarthritis. In the Framingham osteoarthritis study weight loss over the previous 10 years was associated with reduced risk for symptomatic knee osteoarthritis in women approximately 60 years of age with baseline BMI 25 kg/m2 or more, but weight gain did not increase the risk.20 This effect might have been the result of reduced joint pathology, but might also be explained by less weight provocation of already pathologic joints. In another Framingham study (64% women), weight gain was associated with increased risk for radiographic knee osteoarthritis.21 In that report the screening period covered 4 years, ending 4 years before index radiographs taken at a mean age of 71 years.
At the last screening 95% of the participants were aged 54 years or less, which is younger than the usual age of onset for primary osteoarthritis of the hip.28 There was no difference between operated and nonoperated participants with regard to change in physical activity from the first to the second cardiovascular screening. This supports the assumption that symptoms of osteoarthritis were negligible at all 3 screenings and that hip pain did not influence weight change from the first to the last screening.
For the majority of the study group body weight and BMI remained fairly constant throughout the screening period (and possibly even up to the age of hip replacement). However, even in the subgroups with extreme weight gain or weight loss we could not demonstrate any change in risk for later total hip replacement resulting from osteoarthritis. In the study population 3524 participants (with 51 total hip replacements) had a decrease in BMI from first to third screening of at least 2 kg/m2, and 1126 participants (with 24 total hip replacements) had a decrease in BMI of at least 3 kg/m2. There were 10,107 participants (145 total hip replacements) with a BMI increase of at least 2 kg/m2, and 5244 participants (78 total hip replacements) with a BMI increase of at least 3 kg/m2. The power to detect a 67% risk reduction (or a 3-fold increase) among the most extreme of these groups was greater than 95%.
If one could determine at what age a high BMI is most detrimental to the hip joint, this information would help to focus a preventive strategy. Case-control studies of male and female recipients of total hip replacements indicated that recalled BMI at 40 years was more predictive of later hip osteoarthritis than recalled BMI at 20, 30, and 50 years of age.29,30 These reports indicated that BMI impact peaked at age 40 years, but the differences were very small. We found that the association between BMI and later total hip replacement resulting from primary osteoarthritis remained unchanged during the fourth and fifth decades of life.
In this study we found no association between adult weight change and the risk for later total hip replacement resulting from primary osteoarthritis. Future research should focus on weight reduction during the first 3 decades of life or after the age of 55 years to define a preventive strategy against disabling osteoarthritis of the hip.
1.Ingvarsson T, Hagglund G, Jónsson H Jr, et al. Incidence of total hip replacement for primary osteoarthrosis in Iceland 1982-1996. Acta Orthop Scand. 1999;70:229–233.
2.Ingvarsson T, Stefánsson SE, Gulcher JR, et al. A large Icelandic family with early osteoarthritis of the hip associated with a susceptibility locus on chromosome 16p. Arthritis Rheum. 2001;44:2548–2555.
3.Antoniades L, MacGregor AJ, Matson M, et al. A cotwin control study of the relationship between hip osteoarthritis and bone mineral density. Arthritis Rheum. 2000;43:1450–1455.
4.Lanyon P, Muir K, Doherty S, et al. Assessment of a genetic contribution to osteoarthritis of the hip: sibling study. BMJ. 2000;321:1179–1183.
5.Lindberg H. Prevalence of primary coxarthrosis in siblings of patients with primary coxarthrosis. Clin Orthop. 1986;203:273–275.
6.Havelin LI, Furnes O, Espehaug B. [Report June 2002 Norwegian Arthroplasty Register]. Bergen, Norway: Orthopaedic Dept, Haukeland University Hospital; 2002.
7.Flugsrud GB, Nordsletten L, Espehaug B, et al. Risk factors for total hip replacement due to primary osteoarthritis: a cohort study in 50,034 persons. Arthritis Rheum. 2002;46:675–682.
8.Gelber AC, Hochberg MC, Mead LA, et al. Joint injury in young adults and risk for subsequent knee and hip osteoarthritis. Ann Intern Med. 2000;133:321–328.
9.Cooper C, Inskip H, Croft P, et al. Individual risk factors for hip osteoarthritis:obesity, hip injury, and physical activity. Am J Epidemiol. 1998;147:516–522.
10.Spector TD, Harris PA, Hart DJ, et al. Risk of osteoarthritis associated with long-term weight-bearing sports:a radiologic survey of the hips and knees in female ex-athletes and population controls. Arthritis Rheum. 1996;39:988–995.
11.Vingård E, Sandmark H, Alfredsson L. Musculoskeletal disorders in former athletes. A cohort study in 114 track and field champions. Acta Orthop Scand. 1995;66:289–291.
12.Lindberg H, Roos H, Gärdsell P. Prevalence of coxarthrosis in former soccer players. 286 players compared with matched controls. Acta Orthop Scand. 1993;64:165–167.
13.Vingård E, Alfredsson L, Goldie I, et al. Sports and osteoarthrosis of the hip. An epidemiologic study. Am J Sports Med. 1993;21:195–200.
14.Barrett-Connor E. Postmenopausal estrogen therapy and selected (less-often-considered) disease outcomes. Menopause. 1999;6:14–20.
15.Sandmark H, Hogstedt C, Lewold S, et al. Osteoarthrosis of the knee in men and women in association with overweight, smoking, and hormone therapy. Ann Rheum Dis. 1999;58:151–155.
16.Vingård E, Alfredsson L, Malchau H. Osteoarthrosis of the hip in women and its relation to physical load at work and in the home. Ann Rheum Dis. 1997;56:293–298.
17.Nevitt MC, Cummings SR, Lane NE, et al. Association of estrogen replacement therapy with the risk of osteoarthritis of the hip in elderly white women. Study of Osteoporotic Fractures Research Group. Arch Intern Med. 1996;156:2073–2080.
18.Huang MH, Chen CH, Chen TW, et al. The effects of weight reduction on the rehabilitation of patients with knee osteoarthritis and obesity. Arthritis Care Res. 2000;13:398–405.
19.Messier SP, Loeser RF, Mitchell MN, et al. Exercise and weight loss in obese older adults with knee osteoarthritis:a preliminary study. J Am Geriatr Soc. 2000;48:1062–1072.
20.Felson DT, Zhang Y, Anthony JM, et al. Weight loss reduces the risk for symptomatic knee osteoarthritis in women. The Framingham Study. Ann Intern Med. 1992;116:535–539.
21.Felson DT, Zhang Y, Hannan MT, et al. Risk factors for incident radiographic knee osteoarthritis in the elderly:the Framingham Study. Arthritis Rheum. 1997;40:728–733.
22.Havelin LI, Engesæter LB, Espehaug B, et al. The Norwegian Arthroplasty Register:11 years and 73, 000 arthroplasties. Acta Orthop Scand. 2000;71:337–353.
23.Havelin LI, Espehaug B, Vollset SE, et al. The Norwegian arthroplasty register. A survey of 17,444 hip replacements 1987-1990. Acta Orthop Scand. 1993;64:245–251.
24.Meyer HE, Tverdal A, Selmer R. Weight variability, weight change and the incidence of hip fracture:a prospective study of 39,000 middle-aged Norwegians. Osteoporos Int. 1998;8:373–378.
25.Matthews JN, Altman DG, Campbell MJ, et al. Analysis of serial measurements in medical research. BMJ. 1990;300:230–235.
26.Herberts P, Malchau H. [Annual Report 2001]. Gothenburg, Sweden: Orthopedic Dept, Sahgrenska University Hospital; 2002.
27.Felson DT, Lawrence RC, Dieppe PA, et al. Osteoarthritis:new insights. Part 1: the disease and its risk factors. Ann Intern Med. 2000;133:635–646.
28.Cooper C. The epidemiology of osteoarthritis. Rheumatology. In: Kippel JH, Dieppe PA, 1994; 731–734 Gower Medical Publications. London
29.Vingård E. Overweight predisposes to coxarthrosis. Body-mass index studied in 239 males with hip arthroplasty. Acta Orthop Scand. 1991;62:106–109.
30.Vingård E, Alfredsson L, Malchau H. Lifestyle factors and hip arthrosis. A case referent study of body mass index, smoking and hormone therapy in 503 Swedish women. Acta Orthop Scand. 1997;68:216–220.
© 2003 Lippincott Williams & Wilkins, Inc.