Joint Association of Physical Activity and Prognostic Nutritional Index on Survival in US Cancer Survivors: A Study Based on the <scp>NHANES</scp> Database

BACKGROUND: Nutritional-immune status and physical activity (PA) are key to prognosis in cancer survivors, yet their joint association with mortality remains unclear. Therefore, we examined the independent and combined effects of PA and the Prognostic Nutritional Index (PNI) on mortality outcomes. METHODS: We evaluated the independent and joint associations of PA and PNI with all-cause, cancer-specific, non-cancer, and cardiovascular mortality among 2420 US cancer survivors using data from the 2007-2016 National Health and Nutrition Examination Survey (NHANES). Prognostic Nutritional Index was used as an indicator of nutritional-immune status. We used multivariable Cox regression, extended Cox regression models, and restricted cubic spline analyses to evaluate the associations and dose-response relationships. Combined associations were examined using joint PNI-PA exposure models and Kaplan-Meier curves; multiplicative interaction was tested with likelihood ratio tests, and additive interaction was quantified using relative excess risk due to interaction (RERI), attributable proportion (AP), and synergy index (SI). Sensitivity analyses were performed using median and tertile PNI cut-off points. Meanwhile, subgroup analyses were conducted to investigate heterogeneity. RESULTS: Prognostic Nutritional Index was overall inversely associated with mortality risk (all-cause: HR = 0.762, cancer-specific: HR = 0.741, non-cancer: HR = 0.771, cardiovascular mortality: HR = 0.798; all p < 0.001), but this protective association was not constant and gradually attenuated over follow-up time. There was a graded inverse association between PA and mortality. Joint analyses revealed that the group with high PNI and sufficient PA was linked to an approximately 70% lower risk of all-cause mortality (p < 0.01). No multiplicative interaction was detected, whereas additive interaction was observed for all-cause mortality (RERI = 0.52; AP = 0.23; SI = 1.69). Sensitivity analyses supported the practical utility of PNI categorization, and subgroup analyses highlighted potentially heterogeneous associations across cancer types. CONCLUSION: Prognostic Nutritional Index and PA were independently and jointly associated with survival among cancer survivors, supporting integrated survivorship strategies targeting nutritional-immune status and physical activity.