SEARCH

SEARCH BY CITATION

Keywords:

  • hip circumference;
  • waist circumference;
  • cardiovascular disease;
  • diabetes;
  • mortality

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Objective: The waist circumference is widely viewed as a simple but effective measure for assessing obesity-related health risks, whereas measurement of the hip circumference is not currently prioritized. This study examines health risks associated specifically with hip circumference in a cohort of Swedish women, to determine whether information may be lost by excluding the hip circumference from health surveys.

Research Methods and Procedures: The subjects described in this report constitute a population-based sample of 38- to 60-year-old women who underwent anthropometric examinations in 1968. The 24-year incidence rates have been ascertained for myocardial infarction, combined cardiovascular diseases, and diabetes. All-cause, cardiovascular, and myocardial infarction mortality also were evaluated.

Results: Hip circumference was a significant independent inverse risk estimator for all endpoints studied. Using Cox regression with adjustment for age, smoking, body mass index, and waist circumference, the remaining variability associated with larger hips was associated with significantly fewer adverse health outcomes. The hip circumference became statistically informative after body mass index adjustment. The strongest protective associations were observed for cardiovascular disease and diabetes endpoints, although significant trends were also seen for total mortality. Considering hip and waist simultaneously, the strength of the inverse association for large hips generally exceeded the positive association for waist.

Discussion: Recent interest in the waist circumference as an effective screening tool has taken the focus off of the hip circumference. The present results suggest that collection of hip measurements should not be discontinued in assessment of obesity-related risk status and health promotion.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

In our quest for an informative, yet simple and convenient indicator of obesity and related health risks, waist circumference has been the focus of much recent public health attention (1, 2). Indeed, waist circumference has been proposed to be a better correlate of visceral fat and cardiovascular risk than the commonly used waist-to-hip circumference ratio (WHR) (1). A large waist has been shown to reflect both generalized obesity and centralized body fat distribution with high sensitivity and specificity, a result that implies that waist circumference could replace both body mass index (BMI) and WHR as a simple indicator of need for weight management (3). Despite this reasoning, it seems unlikely that the weight and height measurements will be excluded from public health screening efforts in the future. In contrast, there is anecdotal evidence from several ongoing studies that such is the fate of the hip circumference.

Before discarding the hip circumference in epidemiological surveys, it is important to consider whether any important information is likely to be lost by doing so. The purpose of this communication is to examine the health risks associated specifically with hip circumference using a population-based study of women in Gothenburg, Sweden (4). This study was among the first to reveal dramatically decreased longevity in women with a high WHR. It may be speculated that one of the reasons why the WHR has so strongly and consistently predicted mortality and morbidity endpoints is that the waist circumference indicates increased cardiovascular risk, possibly through increased visceral adiposity, whereas the hip is at the same time exerting a protective influence. Therefore, we now consider the waist and hip circumference as independent risk factors, using recent follow-up data from the Gothenburg Women's Health Study. Specifically, we focus on the independent predictive value of the hip circumference in relation to total mortality, diabetes, and selected cardiovascular endpoints, with the hypothesis that a large hip circumference may be a protective risk factor, if other anthropometric measures are held constant.

Research Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The subjects described in this report constitute a population-based cohort of women that has been followed since 1968. At that time, a randomly selected sample of 1462 women 38, 46, 50, 54, or 60 years of age (mean, 46.8 years; SD, 6.2) participated in a health examination study. This examination represented the baseline for a prospective follow-up study, with endpoints most recently established after 24 years. During the baseline examination, weight, height, waist circumference, and hip circumference were measured by one examiner in 1405 of the subjects, as described previously (4). Ascertainment of endpoints was accomplished by examining surviving subjects, sending questionnaires to nonparticipants, and conducting searches in the national mortality and local hospital registries. The outcome variables reported here are as follows: total mortality, mortality from myocardial infarction, and mortality from cardiovascular diseases, together with incidence of myocardial infarction, cardiovascular diseases, and diabetes mellitus.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The results shown in Table 1 suggest that the hip circumference is a significant and independent inverse risk estimator for all endpoints studied. Specifically, using proportional hazards regression with statistical adjustment for age, smoking status, BMI, and waist circumference, the remaining variability associated with larger hips was associated with significantly fewer adverse health outcomes. Most consistently, being in the highest quartile of hip circumference was associated with fewer cases of diabetes, myocardial infarction, and combined cardiovascular disease, compared with the lowest quartile. Hip circumference became statistically informative only after adjusting for BMI.

Table 1.  Hip circumference as a predictor of 24-year mortality and morbidity endpoints in 1405 Swedish women 38 to 60 years of age in the Gothenburg Women's Health Study
 Myocardial infarction mortality (n = 40)Myocardial infarction morbidity (n = 92)Cardiovascular disease mortality (n = 80)Cardiovascular disease morbidity (n = 177)Diabetes morbidity (n = 77)Total mortality (n = 257)
  • *

    Hip quartile cutpoints 94.5, 98.5, and 103.5 cm, multivariate models adjusted for age, smoking status, BMI, and waist circumference at baseline.

  • Wald χ2 and p levels (below) indicating independent contributions to predictive models. Hip and waist circumference as continuous (BMI residual) values with further multivariate adjustment for BMI, age, smoking status, and other circumference. BMI adjusted for waist and hip residuals, age, and smoking status.

  • Coefficients negative for all endpoints.

  • §

    All coefficients positive.

Quartiles*      
Hip Q11 (reference)1 (reference)1 (reference)1 (reference)1 (reference)1 (reference)
Hip Q20.41 (0.15–1.2)0.59 (0.31–1.13)0.61 (0.32–1.2)0.76 (0.48–1.22)0.41 (0.17–1.01)0.69 (0.48–1.00)
Hip Q30.47 (0.18–1.20)0.56 (0.3–1.06)0.44 (0.22–0.88)0.76 (0.48–1.22)0.57 (0.27–1.23)0.55 (0.37–0.82)
Hip Q40.18 (0.05–0.67)0.34 (0.14–0.79)0.30 (0.12–0.78)0.43 (0.23–0.82)0.31 (0.12–0.8)0.59 (0.35–0.99)
Continuous effects      
Hip residuals10.517.117.610.310.15.9
 p = 0.001p = 0.0001p = 0.0001p = 0.001p = 0.001p = 0.02
Waist§ residuals2.81.89.93.43.824.7
 p = 0.1p = 0.2p = 0.001p = 0.07p = 0.05p = 0.0001
BMI§ (kg/m2)5.212.31.87.517.62.0
 p = 0.02p = 0.0005p = 0.2p = 0.006p = 0.0001p = 0.2

On the lower portion of Table 1, we have directly compared the predictive values for hip circumference, waist circumference, and BMI values in their continuous forms. To reduce colinearity among waist, hip, and BMI, the residual method was used to obtain waist and hip circumference indicators that reflect deviation from expected values at a given BMI. These effects were all considered simultaneously in one multivariate model; the Wald χ2 test statistic and its associated p level are reported here. Consistently, the regression coefficients were negative for the hip indicator and positive for both waist and BMI. Compared with waist, hip circumference was a statistically stronger predictor for all endpoints, except for total mortality. Compared with waist and hip residuals, BMI was the strongest predictor of diabetes mellitus.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

This analysis underscores the importance of continuing to measure hip circumference in studies of health and longevity among women. Recent interest in waist circumference as an effective screening tool has taken the focus off of the hip. Yet, Seidell et al. (5) reported that men and women with narrower than expected hips had a 2- to 3-fold excess risk of being diabetics, after adjustment for waist circumference. This supports the earlier observations of Hartz et al. (6) that women with wide hips were less likely to report (prevalent) hypertension, diabetes, and gallbladder disease, after adjustment for relative weight and waist circumference. The results reported here are consistent with these previous cross-sectional findings, and the prospective design of the present study strengthens the conclusion by demonstrating that the hips were narrow many years before the conditions emerged and that similar associations are present for fatal cardiovascular disease endpoints.

Future research will be required to elucidate the physiological significance of these findings. For instance, narrow hips may reflect small pelvic size, gluteal muscle, and/or fat masses, which, in turn, may relate to a number of factors, including early life exposures, lipoprotein lipase activity, or circulating estrogen levels. In addition, it is plausible that the small hip girth and the health endpoints studied here have common genetic origins. In the meantime, it seems premature to conclude that hip circumference measurements are unnecessary in the assessment of obesity-related risk status and health promotion. The present results suggest that despite a possible clash between a perfect body and a healthy body, narrow hips seem to be a great health disadvantage to women.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

This work was funded by grants from the Swedish Council for Planning and Coordination of Research and the Swedish Medical Research Council.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • 1
    World Health Organization (1998) Obesity, Preventing and Managing the Global Epidemic: Report of a WHO Consultation on Obesity. World Health Organization: Geneva, Switzerland.
  • 2
    National Heart Lung and Blood Institute (1998) Clinical guidelines on the identification, evaluation and treatment of overweight and obesity in adults—the evidence report. Obes Res. 6 (Suppl 2).
  • 3
    Lean, M. E. J., Han, T. S., Morrison, C. E. (1995) Waist circumference as a measure for indicating need for weight management. Br Med J. 311: 158161.
  • 4
    Bengtsson, C., Björkelund, C., Lapidus, L., Lissner, L. (1993) Associations of serum lipid concentrations and obesity with mortality in women: 20-year follow-up of participants in the prospective population study in Gothenburg, Sweden. Br Med J. 307: 13851388.
  • 5
    Seidell, J. C., Han, T. S., Feskens, E. J. M., Lean, M. E. J. (1997) Narrow hips and broad waist circumferences independently contribute to increased risk of non-insulin-dependent diabetes mellitus. J Intern Med. 242: 401406.
  • 6
    Hartz, A. J., Rupley, D. C., Rimm, A. A. (1984) The association of girth measurements with disease in 32,856 women. Am J Epidemiol. 119: 7180.