Evaluation of indices of obesity in men: descriptive study

BMJ 1998; 316 doi: (Published 09 May 1998) Cite this as: BMJ 1998;316:1428
  1. Derrick Pounder, professor,
  2. David Carson, research student,
  3. Michael Davison, research student,
  4. Yoshiyuki Orihara, research fellow
  1. Department of Forensic Medicine, Dundee Royal Infirmary, Dundee DD1 9ND
  1. Correspondence to: Professor Pounder
  • Accepted 31 October 1997

The intra-abdominal accumulation of fat is more common in men; it is an independent risk factor for diabetes mellitus and cardiovascular disease.1 The assessment of obesity usually focuses on abdominal fatness, but the most appropriate clinical measurement of obesity has been disputed. 2 3 We evaluated several measures of obesity by comparing them with observed intra-abdominal and subcutaneous fat during postmortem examination.

Methods and results

A series of 100 men was studied. Cases of prolonged hospitalisation, severe trauma, chronic wasting disease, and postmortem decomposition were excluded from analysis. Measurements were made using techniques suitable for supine cadavers. Body weight was measured to within 1 kg, body length to within 1 cm, hip circumference at the iliac crests to within 1 cm, and waist circumference at the umbilicus to within 1 cm. Uniform cores of subcutaneous fat were obtained from over the biceps, anterior thigh, anterior chest, and anterior abdomen. These samples were weighed to within 0.1 g. The greater omentum was excised as was the pararenal fat; these were weighed to within 1 g.

The mean age at the time of death was 52.8 years (median 55.5, range 17 to 89, interquartile range 37 to 67). Mean body weight was 74 kg (median 73, range 39 to 142, interquartile range 65 to 83). In nine cases the body mass index (weight (kg)/height (m2)) was <20; in 50 it was 20 to 24; in 32 it was 25 to 29; in 8 it was 30 to 39; and in 1 it was ≥40. Waist circumference was <94 cm in 69 cases; 94 to 101 cm in 19, and ≥102 cm in 12.

The strongest correlation between the weight of the subcutaneous fat taken from the four sites was between the fat at the abdomen and that at the chest (0.60); the weakest correlation was between the fat at the abdomen and that at the upper arm (0.36). The correlation between the weight of the intra-abdominal fat and the weight of the subcutaneous fat was 0.57 for the anterior chest, 0.38 for the abdomen, 0.45 for the upper arm, and 0.28 for the thigh. When 82 cases with body mass index between 20 and 30 were analysed, the correlation was 0.49 for the anterior chest, 0.23 for the abdomen, 0.39 for the upper arm, and 0.15 for the thigh.

When the weight of intra-abdominal fat was used as the outcome r2 predictive value was 40% for body weight, 37% for body mass index, 61% for waist circumference, 47% for hip circumference, 43% for waist to hip ratio, and 54% for waist to height ratio. For the 82 cases with body mass index between 20 and 30 the r2 predictive value was 32% for body weight, 31% for body mass index, 58% for waist circumference, 40% for hip circumference, 38% for waist to hip ratio, and 51% for waist to height ratio. Waist circumference was strongly correlated with body weight (r=0.90), body mass index (r=0.90), and hip circumference (r=0.95). For the 82 cases with body mass index between 20 and 30 r was 0.81 for body weight, 0.80 for body mass index, and 0.91 for hip circumference. When non-obese cases were grouped by waist circumference or body mass index a trend of increasing intra-abdominal fat mass was seen (table).

Mean, median, and range of weight (g) of intra-abdominal fat from the greater omentum and pararenal area in 59 non-obese subjects by body mass index (kg/m2) and waist circumference (cm) at the umbilicus

View this table:


Body mass index is the traditional measure of obesity; it is an index of weight relative to stature. The waist to hip circumference was the first index adopted as a measure of the abdominal location of body fat.1 The results of studies that used computed tomography have suggested that waist circumference correlates better with abdominal visceral fat mass.2 Advocates of using waist circumference to measure obesity argue that it is a straightforward measurement that relates to both body weight and the distribution of fat.3

Indices of obesity can be evaluated during post- mortem examination by comparing them with the amount of body fat observed. Measurements cannot be made in the same manner as during clinical practice, so values observed at necropsy may not correspond exactly with clinical indices. Within these limitations this study confirms that waist circumference correlates more strongly with intra-abdominal fatness than any other suggested index, and that this is true across the range of body mass index values.

When the cases were grouped together (table) a significant increase in intra-abdominal fat mass occurred before the point at which guidelines in the United Kingdom recommend implementation of a weight loss programme—that is, waist measurement of 94 cm and body mass index >25. The appropriateness of using a body mass index of 25 to determine the point at which weight reduction should occur has been questioned,4 and it has also been suggested that the ideal body mass index should be about 22.5 Above 22 some men accumulate large amounts of intra-abdominal fat and the health risks associated with this fat.


Contributors: DP initiated the study, designed the protocol, helped analyse data, wrote the paper, and is guarantor for the study. MD helped design the protocol, review the literature, collect and analyse the data. DC and YO collected and analysed data.

Funding: The Crown Office funds the department of forensic medicine.

Conflict of interest: None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract