Evaluating consistency of anthropomorphic measurements in women with a history of gestational diabetes

Zuhra Musherraf, Western University of Health Sciences, College of Osteopathic Medicine of the Pacific, 309 E. Second Street, Pomona, CA 91766-1854.

E-mail address: zmusherraf@westernu.edu.

OBJECTIVE: Determination of waist circumference is now being recommended as a routine office measurement to identify women with metabolic syndrome and to monitor clinical response to thera- peutic interventions. As such, the accuracy of measurement is important. In this study, we tested accuracy of a series of anthropomorphic measurements by looking for internal consistencies between them and by comparing changes in circumference to changes in weight.

METHODS: We evaluated the fluctuations in anthropomorphic measurements performed by the same group of examiners on gestational diabetics being followed during the postpartum period.

RESULTS: One-hundred sixty-six patients were identified for the study who contributed a total of 502 changes in each of the variables over time. Comparisons of different waist measurements found significant changes in 35% of cases. As many as 40.4% of women had at least one episode of significant discordance when weight changed measurably in one direction but circumference changed in the opposite direction, and 41.6% of women had at least one measurement where a circumference changed significantly, but the weight did not. The frequency of these discordances was not different between obese versus nonobese women.

CONCLUSIONS: The inconsistencies found in this real-world experience suggest that the clinical significance of the anthropomorphic measurements may be limited and must be considered in the context of the patient’s weight and/or body mass index.

Body mass index (BMI) measurements have been found to provide useful estimates of obesity-related health risks at a population level, but may not be reliable for individual risk assessment.1,2 Anthropomorphic measurements (such as waist and hip circumference) and related measures (such as waist- to-hip ratios) have been found to be predictive of increased risks for chronic diseases (such as diabetes, hypertension, stroke, and myocardial infarction).3-5 Body fat distribution contributes to obesity-related disease risk independent of adi- posity.6 Waist measurements represent central obesity and are thought to be predictive of cardiovascular risk because they reflect the quantity of visceral fat.7-9 In particular, waist cir- cumference and related measurements have been reported to be more predictive for individual patients’ cardiovascular risk and/or all-cause mortality than BMI (this is true in the over- weight to mild obesity categories).3,10,11 Age and gender may also be important factors to consider when interpreting these measures. Waist-to-hip ratios (but not waist circumference) have been shown to be much more accurate than BMI in predicting risk for death in men and women over age 75.12 Waist-to-hip ratios, rather than BMI, have also been recom- mended for use in middle-aged persons.3,13

The National Cholesterol Education Program, the Group for the Studies of Insulin Resistance, the World Health Organization (WHO), and the American Association of Clinical Endocrinologists all consider an excessively large waist circumference (greater than or equal to 80 – 88 cm) or a waist-to-hip ratio greater than 0.85 in women as at least one defining criterion of metabolic syndrome.14,15 Some authors have suggested that measurement of waist circum- ference should become a routine vital sign, especially in patients with a normal BMI.16 Others have called for dif- ferent thresholds for defining abnormalities in waist and hip measurements for subjects from various geographic/ethnic backgrounds.16-19‌

Given the pivotal role that the measurements of waist and hip circumference have in predicting risk of disease and the significance that changes in those measurements may have in monitoring risks of those substantial disease states,20 it is important that these measurements be accurate and easy to obtain. Translation of earlier research reports into clinical practice may not render similarly reliable re- sults. In this study, we investigate the accuracy of routine measurements of waist and hip circumferences in a real- world clinical setting. Accuracy was assessed by examining the degree of internal consistency between changes in the two different measures of waist circumference and in the degree of internal consistency between changes in weight and changes in each of the 3 measured circumferences.

Women who have been diagnosed with gestational diabetes at the Harbor-UCLA Medical Center in Torrance, California are followed postpartum in a special clinic designed to encourage long-term healthy lifestyles and to monitor these high-risk women for development of glucose intolerance/diabetes and other markers of metabolic syndrome. This clinic serves an indigent, primarily Hispanic patient population.

At each visit, the patient’s weight is recorded, as are mea- surements of the hip and waist circumferences. All measure- ments in this study were obtained by two women’s health care nurse practitioners. Hip circumference was defined as the greatest lower truncal circumference, which includes the but- tocks. Waist circumference was measured at 2 (potentially different) sites. One measurement was defined as circumfer- ence at the site where the patient defined as her waist (the “true waist”). The second was the circumference that included the umbilicus (the “umbilical waist”). Each waist measurement was taken at the end of an exhalation with the patient having been told to relax her abdominal muscles.

This study was approved by the Institutional Review Board of the Los Angeles Biomedical Institute at Harbor- UCLA Medical Center to study these data on an exempt basis after the data were de-identified. All women who had at least 2 visits in which all the relevant measurements were recorded are included in the study. There were no exclusion criteria beyond inadequate data. Chi square tests were used to test for differences between dichotomous variables, and a Fisher t-test was used to test for differences in continuous variables. Statistical significance required a p < 0.05.

Results

After evaluating the medical records for completeness in all measurements for at least two visits, we identified 166 patients who contributed a total of 668 visits, which pro- vided information for 502 changes in each of the variables over time. The mean number of eligible return visits was four per patient, whereas the median was three. Baseline characteristics of the patient population are described in Table 1. The population was varied; baseline BMIs ranged from 19.5 to 45.6 kg/cm2.

Several analyses of the changes in waist and hip circum- ference measurements were performed to investigate their internal consistency as a way to validate accuracy. The first analysis of the reliability of these waist measurements was to compare the magnitude and direction of changes in the 2 different measurements of the waist circumference (the true and the umbilical circumferences). Internal consistency was established if a woman’s true waist circumference changed in the same direction as the umbilical waist circumference. Internal consistency was violated if the 2 measures changed significantly in opposite directions. Although others have reported precision in measurements to 0.1 cm12, we selected cm as a cut-off for accuracy because we believed that there would be general agreement that differences greater than that would be substantially wrong. Table 2 displays information about the 177 episodes in which significant discordance was found between the 2 measurements, that is, when one waist measurement changed by at least 2 cm (increased or decreased) and the other waist measurement changed by at least 2 cm in the opposite direction and produced a net difference of at least 4 cm. Interestingly, there was no difference in the frequency of discordant measurements between obese and nonobese women.

Next, internal consistency was tested by comparing changes in weight to changes in the circumference measure- ments. Discordance was first defined as the patient’s weight changed by more than two pounds in one direction (gain or loss), whereas the circumference measurement changed by more than 2 cm in the opposite direction (increase or de- crease). Again, these limits were set to allow for accuracy of scale (2 lb) and relatively large (2 cm) differences in cir- cumference measurement. Table 3 displays the frequency and degree of discordance found between weight change and each of the circumference measurements. The mean change in circumference was 5.84 cm when the women had an opposite change in weight of 5.8 lb. It can be seen that at least 30% of women had at least one episode of significant discordance between changes in weight and changes in each of the measurements of circumference. There was no sta- tistically significant difference in the frequency of such discordance between obese and nonobese women.

In a second analysis of internal consistency, discordance was identified when there was a significant change (>2 cm) in circumference but no significant change (<2 lb) in

weight. Table 4 displays the frequency and degree of these events. In the extreme, one woman had a 20.5-cm change recorded in her waist circumference that was not accompa- nied by any change in weight. At least one third of women had at least one episode of this discordance. Again, the frequency of this discordance did not vary between obese (BMI >30) and nonobese (BMI <30) subjects. There were also many episodes when the weight changed significantly but the curcumference measurements were stable (not shown).

Discussion

Obesity is epidemic in the United States; estimates are that for the first time in four centuries, life expectancy of the next generation may be less than for the current one pri- marily because of obesity and obesity-related diseases. Met- abolic syndromes have been defined to identify individuals at high risk for cardiovascular disease. Visceral adiposity, as determined by magnetic resonance imaging/computed to- mography studies, is most closely linked to dyslipidemia, insulin resistance, and other cardiovascular risk factors.20

Because direct measurement of visceral fat is not possi- ble, surrogate measures have been developed to identify patients who require medical interventions and to monitor the progress of those treatments. Weight, BMI, waist cir- cumference, hip circumference, and waist-to-hip ratios have all been studied. Anthropomorphic measurements have been found to be more sensitive than weight or BMI. Many studies have confirmed a high correlation between waist circumference and viseral obesity.7-9,21-24 Furthermore, Janssen et al. found that within each BMI category, those patients with larger waist circumferences had higher risks for virtually every obesity-related comorbidity when compared with nor- mal-waisted patients.25

Hu has demonstrated that the waist circumference mea- surement is also simpler to measure than the waist-to-hip ratio, that it is associated with fewer measurement errors, and that waist circumference is the most practical and sim- ple to interpret.20 Different cutoffs have been suggested for different ethnicities to reflect differences in muscle mass and bone density.17

However, our study found that the accuracy of measure- ments of circumference can be questionable in real-world prac- tice. Testing available data for internal consistency, there were high rates of changes that did not make sense. In particular, we found that 57.7% of women had a significant change in ab- dominal girth (>2 cm) in one direction (increase or decrease) measured by “true waist”, when the other measure of girth “umbilical waist” had a significant change in the opposite direction. Because the “true waist” might be considered an unreliable gauge, we also compared measurements of circum- ference to a more objective measurement (weight). We again found that at least 30% of women had at least one time when her circumference changed significantly (>2 cm) in one direc- tion (increase or decrease), whereas her weight changed sig- nificantly (>2 lb) in the opposite direction. We also found changes in circumference when there was no accompanying change in weight; significant changes in weight were also seen with no changes in circumference. Virtually every woman who had multiple visits had at least one episode of inconsistent changes in some variable. Interestingly, obesity did not affect the answers; the frequency of inconsistent changes was no different between women with BMI >30 compared with nonobese women. If we had used lower thresholds for discor- dance (<1 cm or <1 lb), as has been done in other studies, the results would have been even significant.

It is interesting to note that, as important as these mea- surements appear to be, there are no standardized protocols for determining any of the anthropomorphic measurements. WHO has provided guidance, but those recommendations are not universally followed. National Health and Nutrition Examination Survey criteria specified that the waist circum- ference is to be measured with a steel measuring tape to the nearest 0.1 cm at the high point of the iliac crest at minimal respiration.21 However, these criteria are often, but not always, used.22 For example, for hip circumference, Sön- mez et al. measured the circumference in the standing po- sition from the plane of both major trochanters.2 On the other hand, Zhang et al. measured hip circumference at the level of maximum width of the buttocks with the subject in a standing position.16 Noble determined hip circumference at the widest point over the greater trochanter.23 Similarly, Sönmez et al. measured the waist circumference in the standing position at a point midway between the lowest rib and the iliac crest.2 Zhang measured waist circumference cm above the umbilicus.16 The Department of Defense advises the clinician to “feel to locate the upper hip/bone and top of the right iliac crest and locate the midpoint. The tape is placed in a horizontal plane around the abdomen at the level of that landmark. Measurements are made to the nearest half inch.24 Others record the waist circumference at the level of the umbilicus.5,23

Although we recognize the theoretical importance of waist circumference, our study highlights the need to con- sider several real-world sources for potential error. For example, distinct problems arise by measuring waist cir- cumference at only one site. As women gain weight, the abdominal panniculus often falls. This can create a situation in which the upper abdomen may become more slender. At the same time as the umbilicus droops toward the mons, the lower abdomen may increase greatly in girth. To reflect the variability of anatomical locations of both visceral and sub- cutaneous fat, this study defined 2 different measures of waist circumference. Some of the discordance seen between changes in true waist and umbilical waist measurements may be explained by this “settling” or “drooping” effect.

In addition, there is a possibility of error in measurement. There is a suggestion of this potential in the way data have been reported in earlier studies. Price et al. performed du- plicate measurements of waist and hip circumference and reported discarding instances when there was greater than 3% difference between the duplicate measurements. They averaged remaining data points and reported waist circum- ference to the nearest millimeter.12

Beyond the definitional and mechanical factors such as respiratory excursion and positional changes that may in- troduce measurement errors, patients themselves may also contribute some to the uneven findings. In their embarrass- ment, there is a tendency for the patients to “suck it in” or contract their abdominal muscles whenever they are being measured. In reproductive-aged women, the menstrual cycle may also affect abdominal circumference measurements. During the premenstrual phase, the patient’s bloating and constipation may add to the total circumference, as may food or air in the gastrointestinal tract or an enlarged uterus. Although Janssen et al. and Zhang reported recording their measurements to the nearest millimeter,16,25 our data sug- gest that at least changes less than 2 cm are too unreliable to be considered clinically significant.

Even though all the measurements were made by trained professionals, it is clear that very unusual patterns of weight change and anthropomorphic measurement changes were observed. Therefore, just as BMI may be an indirect and imperfect measurement of adiposity and has several limita- tions,20 our study finds that measurements of waist and hip circumference also are indirect and (occasionally) imperfect measures of visceral obesity. Clearly precision may not be needed to establish the initial diagnosis of visceral obesity, but accuracy and reliability of these measurements may be very important in monitoring a patient’s response to treat- ment. Bosy-Westphal et al. have concluded that waist cir- cumference and hip-to-waist ratios are not able to capture all the information that is needed to assess obesity-related metabolic risk, especially in the context of lean body mass.1

Although changes in these anthropomorphic measures are important in assessing risk status and determining the effect of therapeutic interventions,6 our study suggests that a more uniform and evidence-based method of anthropometric measurements would be helpful but may not be sufficient. Instead, the totality of weight and BMI changes, as well as changes in waist and hip measurements, should be consid- ered in the context of the patient’s body habitus before judging the success or failure of a given intervention.‌

References

Bosy-Westphal A, Geisler C, Onur S, Korth O, Selberg O, Schrezen- meir J: Value of body fat mass vs anthropometric obesity indices in the assessment of metabolic risk factors. Int J Obes (Lond) 30:475-483, 2006

Sönmez K, Akçakoyun M, Akçay A, Demir D, Duran NE, Gençbay M, et al: Which method should be used to determine the obesity, in patients with coronary artery disease? (body mass index, waist cir- cumference or waist-hip ratio). Int J Obes Relat Metab Disord 27:341- 346, 2003

Yusuf S, Hawken S, Ounpuu S, Bautista L, Franzosi MG, Commerford P, et al: INTERHEART Study Investigators. Obesity and the risk of myocardial infarction in 27,000 participants from 52 countries: a case-control study. Lancet 366(9497):1640-1649, 2005

Vazquez G, Duval S, Jacobs DR Jr, Silventoinen K: Comparison of body mass index, waist circumference, and waist/hip ratio in predict- ing incident diabetes: a meta-analysis. Epidemiol Rev 29:115-128, 2007

Gupta R, Rastogi P, Sarna M, Gupta VP, Sharma SK, Kothari K: Body-mass index, waist-size, waist-hip ratio and cardiovascular risk factors in urban subjects. J Assoc Phys India 55:621-627, 2007

Eckel RH, Grundy SM, Zimmet PZ: The metabolic syndrome. Lancet 365:1415-1428, 2005

Ferland M, Després JP, Tremblay A, Pinault S, Nadeau A, Moorjani S, et al: Assessment of adipose tissue distribution by computed axial tomography in obese women: association with body density and an- thropometric measurements. Br J Nutr 61:139-148, 1989

Rankinen T, Kim SY, Pérusse L, Després JP, Bouchard C: The pre- diction of abdominal visceral fat level from body composition and anthropometry: ROC analysis. Int J Obes Relat Metab Disord 23:801- 809, 1999

Onat A, Avci GS, Barlan MM, Uyarel H, Uzunlar B, Sansoy V: Measures of abdominal obesity assessed for visceral adiposity and relation to coronary risk. Int J Obes Relat Metab Disord 28:1018-1025, 2004

Winter Y, Rohrmann S, Linseisen J, Lanczik O, Ringleb PA, Hebe- brand J, et al: Contribution of obesity and abdominal fat mass to risk

of stroke and transient ischemic attacks. Stroke 2008 [Epub ahead of print]

Hu FB, Willett WC, Li T, Stampfer MJ, Colditz GA, Manson JE: Adiposity as compared with physical activity in predicting mortality among women. N Engl J Med 351:2694-2703, 2004

Price GM, Uauy R, Breeze E, Bulpitt CJ, Fletcher AE: Weight, shape, and mortality risk in older persons: elevated waist-hip ratio, not high body mass index, is associated with a greater risk of death. Am J Clin Nutr 84:449-460, 2006

Kragelund C, Omland T: A farewell to body-mass index? Lancet 366:1589-1591, 2005

Lorenzo C, Williams K, Hunt KJ, Haffner SM: The National Choles- terol Education Program–Adult Treatment Panel III, International Di- abetes Federation, and World Health Organization definitions of the metabolic syndrome as predictors of incident cardiovascular disease and diabetes. Diabetes Care 30:8-13, 2007

Alberti KG, Zimmet P, Shaw J: IDF Epidemiology Task Force Con- sensus Group. The metabolic syndrome—a new worldwide definition. Lancet 366:1059-1062, 2005

Zhang X, Shu XO, Yang G, Li H, Cai H, Gao YT, et al: Abdominal adiposity and mortality in Chinese women. Arch Intern Med 167:886- 892, 2007

Deurenberg P, Deurenberg-Yap M, Guricci S: Asians are different from Caucasians and from each other in their body mass index/body fat per cent relationship. Obes Rev 3:141-146, 2002

Molarius A, Seidell JC, Sans S, Tuomilehto J, Kuulasmaa K: Varying sensitivity of waist action levels to identify subjects with overweight or obesity in 19 populations of the WHO MONICA Project. J Clin Epidemiol 52:1213-1224, 1999

Misra A, Wasir JS, Vikram NK: Waist circumference criteria for the diagnosis of abdominal obesity are not applicable uniformly to all populations and ethnic groups. Nutrition 21:969-976, 2005

Hu FB: Obesity and mortality: watch your waist, not just your weight. Arch Intern Med 167:875-876, 2007

Center for Disease Control and Prevention: The Third National Health and Nutrition Examination Survey (NHANES III 1988-94) reference manuals and reports. Bethesda, MD: Nathional Center for Health Statistics; 1996 (CD-ROM)

Center for Disease Control and Prevention: The Third National Health and Nutrition Examination Survey (NHANES III 1988-94) reference manuals and reports. Bethesda, MD: Nathional Center for Health Statistics; 1996 (CD-ROM)

Noble RE: Waist-to-hip ratio versus BMI as predictors of cardiac risk in obese adult women. West J Med 174:240-241, 2001

Heinrich KM, Jitnarin N, Suminski RR, Berkel L, Hunter CM, Alvarez L, et al: Obesity classification in military personnel: a comparison of body fat, waist circumference, and body mass index measurements. Mil Med 173:67-73, 2008

Janssen I, Katzmarzyk PT, Ross R: Body mass index, waist circum- ference, and health risk: evidence in support of current National Institutes of Health guidelines. Arch Intern Med 162:2074-2079, 2002