Martin · Journal of clinical oncology : official journal of the American Society of Clinical Oncology 2015 · Prospective cohort study · n=11,123

Diagnostic criteria for the classification of cancer-associated weight loss.

Cited 799 times in the scientific literature.

Level 3 - non-randomized controlled study

Prospective observational cohort study for prognostic classification with validation cohort.

PubMed 25422490 · doi:10.1200/JCO.2014.56.1894 · record verified 2026-08-27

What was done

A population-based derivation cohort of 8,160 Canadian and European patients with cancer was followed prospectively until death. Baseline body mass index (BMI) and percent weight loss (%WL) were recorded and analyzed using multivariable models adjusting for age, sex, cancer site, stage, and performance status. Based on survival associations across five BMI categories and five %WL categories, a 5x5 prognostic matrix (grades 0–4) was established and evaluated in an independent validation cohort of 2,963 patients.

What was found

Mean overall %WL was -9.7% ± 8.4% and mean BMI was 24.4 ± 5.1 kg/m². Both %WL and BMI independently predicted survival (P < .01). Median overall survival progressively declined with worsening grade (P < .001): grade 0 (weight-stable, BMI ≥ 25.0 kg/m²), 20.9 months (95% CI, 17.9 to 23.9); grade 1, 14.6 months (95% CI, 12.9 to 16.2); grade 2, 10.8 months (95% CI, 9.7 to 11.9); grade 3, 7.6 months (95% CI, 7.0 to 8.2); and grade 4, 4.3 months (95% CI, 4.1 to 4.6). Significant survival discrimination by grade was confirmed across specific cancer types, stages, age groups, performance status levels, and in the 2,963-patient validation sample.

Why it matters

This system provides an empirically validated, standardized tool to stratify mortality risk in cancer patients by accounting for both percentage weight loss and baseline BMI.

Limits

The abstract does not define the time duration over which weight loss occurred. It also does not provide details on specific cancer types, systemic cancer treatments received, body composition metrics such as sarcopenia, or non-cancer causes of weight loss.

Cited by