[Increased chance of dying among nursing home patients with lower body weight]

To investigate a possible relationship between changes in weight and mortality among nursing home patients of 65 years and older, after adjustment for age, sex and main diagnosis.

Descriptive longitudinal study.

Dutch nursing home in Delft, the Netherlands.

For three years after admission the body weight of patients newly admitted from April 1st 1987 to April 1st 1989 was measured every three months under the same conditions. It was recorded if a patient died or was discharged in this three-year period. The effects of body weight at admission, sex, age and diagnostic category (psychogeriatric, neurological, orthopaedic, oncological, other) on mortality were analysed using the univariate Cox regression analysis. The effect of the time-dependent covariate body weight on mortality in the nursing home was analysed with a combined logistic regression analysis per 3-month interval.

The mean weight of the females increased from 59.7 kg (n = 194) at the time of admission to 63.7 kg (n = 57) three years later and that of the males from 64.7 kg (n = 70) to 70.3 (n = 11) due to the low body weight of those who died. During the first 3 months after admission mortality was high (14%) and only diagnosis related. In subsequent 3-month periods there was a significant association between mortality and weight, weight change, age and sex: the risk of mortality was higher for those with a lower body weight, weight loss, higher age and male sex.

Nursing home patients with a low body weight or loss of weight after the first 3 months after admission, have an increased risk of death.