Background

Although advanced pancreatic ductal adenocarcinoma (PDAC) is characterized by progressive weight loss and nutritional deterioration, the effect of cancer cachexia and body mass index (BMI) at diagnosis on survival remains unclear.

Methods

We retrospectively evaluated a prospectively collected internal cancer registry of PDAC cases from 2006-2014 at the Kaiser Permanente Medical Center. Cancer cachexia was defined as weight loss greater than 5% over the 6 months prior to diagnosis. Multivariate cox proportional hazards regression was used to assess the influence of cachexia on survival. To evaluate effect measure modification of this relationship, we performed additional analyses stratified by race, BMI class, stage, receipt of surgery and receipt of chemotherapy. We tested for heterogeneity by fitting models with an interaction term for cachexia and the modifying variable.

Results

Of the 977 patients, 611 (63%) were identified with cachexia. Cachexia in PDAC patients was prevalent across all stages of disease and BMI classes. Patients with cachexia had lower survival (median 4.3 months, IQR 1.7-10.0) compared to those without cachexia (median 5.2 months, IQR 2.0-10.6), log-rank P=0.03. In this analysis BMI at diagnosis was not associated with survival. In the multivariate Cox regression, cachexia was independently associated with decreased overall survival (HR 1.24, CI: 1.06-1.45, P=0.01). However, the effect of cachexia on survival outcomes was modified by receipt of chemotherapy. Cachectic patients who did not receive chemotherapy had a 40% increase in risk of death compared to non-cachectic patients (HR 1.40, CI: 1.12-1.75), while those receiving chemotherapy were unaffected by cachexia (HR 1.04, CI: 0.82-1.32, P_interaction=0.01).

Conclusions

In the largest cohort of pancreatic cancer patients examined to date, cachexia and not obesity is independently associated with worse survival in PDA and its effect is negated by systemic chemotherapy.