Researchers have found middle-age spread occurs in two distinct phases - casting doubt on the merits of using weight as a guide to health.
Weight gain is following by a thickening of girth
They found a thickening waistline in early middle age is accompanied by a rise in weight.
But although waists continue to expand with age, weight gain levelled off in later years as muscle turned to fat.
The study, by the Medical Research Council, appears in the Journal of Clinical Nutrition.
The findings challenge the wisdom of relying solely on Body Mass Index (BMI) to assess whether a person is a healthy weight.
BMI, now used widely by health workers, is calculated by dividing a person's weight in kilograms by the square of their height in metres.
But it can only measure weight gain, and not deposition of fatty tissue, which, the latest study shows, might not be accompanied by an increase in weight.
Geoff Der, from the MRC's Social and Public Health Sciences Unit in Glasgow, said: "As people get older it seems that their bodies change - they lose muscle and get fatter.
"This explains why middle-age spread might not be reflected on the bathroom scales."
He added that BMI was a good measure of lean body tissue, but an expanding waistline might be a more reliable measure of the amount of fatty tissue a person has gained.
"Although the people in the older middle age group in this study appeared to put on less weight than the younger people, their waist circumferences continued to grow over time.
"What appears to have been happening is that the increase in fat was being obscured by a loss of muscle mass."
The researchers carried out a nine-year study of 1,044 people aged either 39 or 59 in 1991.
The height, waist circumference and weight of each participant was measured in 1991, 1995 and 2000, and used to measure changes in body mass index over time.
Only one in five (20%) of the people maintained a stable weight as the study progressed.
Steady weight gain was measured in the younger group - more than 42% of study participants put on 10kg, 17% gained 5kg.
On average, both men and women in the younger group gained between 0.5kg and 1kg a year. This weight gain was fastest in their younger years.
Those in the older age group gained least weight in the second half of the study. Although their overall weight may not have changed, however, their waist circumference did.
The MRC study is the second in quick succession to question the use of BMI.
A University of Texas study found measuring the difference between the waist and hips was a more accurate way to identify people with the early signs of heart disease.
Dr Ian Campbell, medical director of the charity Weight Concern, said: "This study is another sharp reminder that there are more ways of measuring fatness than simple weight.
"As we get older its common for our lean tissue, muscle, to decrease as we become less active.
"Although it's common to keep increasing in total weight into our 60s, the full extent of that extra fat mass may well be obscured by muscle loss.
"A simple waist measurement can be enough to detect just how significant our increased risk of heart disease and type 2 diabetes has become.
"Everyone should regularly check their waistline, not just take a weekly step onto the bathroom scales. They may be giving false comfort."