Geriatric nutritional risk index at diagnosis predicts survival in pancreatic cancer

Introduction : Pancreatic cancer (PC) is one of the deadliest cancers with decreased overall survival due to cachexia and malnutrition. Early detection of nutritional risk is essential to improve treatment outcomes. Reliable and valid tools that are easily applicable in clinical settings will benefit to resolve this issue. The Geriatric Nutritional Risk Index (GNRI) utilizes commonly available clinical measures to assess nutritional risk, which may be useful in detecting risk as early as at diagnosis. Despite support for the GNRI in oncology, more research is needed to validate its use in PC. The purpose of this study was to examine the predictive value of the GNRI at diagnosis on survival in PC, controlling for demographics and cancer stage. Methods : 924 adult patients with primary PC, visited between January 1, 2012, and July 31, 2020 were included in the retrospective study using de-identified data from the UF Health electronic health record. GNRI calculated using the Lorenz formula for ideal weight (Wlo). Weight and serum albumin closest to diagnosis within 120 days used. GNRI = [1.489 x albumin (g/L)] + [41.7 x (weight/Wlo)]. Scores categorized into 3 groups: high (<92), low (92-98), no risk (>98). Survival: days between diagnosis and death with censoring. Covariates: age, race, sex, stage. Multiple imputation used for GNRI missing data (290 patients, 31.4%). Statistical analysis: descriptive statistics, survival analysis using accelerated failure time (AFT) models. Results : Results (Table 1) show GNRI at diagnosis, age, and stage predict survival, but not race or sex. Conclusion : Higher GNRI scores at diagnosis predicted longer survival in PC, after controlling for other influential variables. With patients at low and no risk having over twice the survival compared to those at high risk, the GNRI is effective in early detection for need of proactive nutritional interventions. Further research is needed to investigate the long-term effects of changes in GNRI score on survival. Table 1. Descriptive statistics and Survival analysis results Variables Total Sample(n=924) GNRI groups Univariate Results Multivariate Results High Risk (n=202) Low Risk (n=129) No Risk (n=303) p-value AF (95%CI) p-value AF (95%CI) GNRI groups High risk Ref. Ref. Ref. Ref. Low Risk 0.0019 2.26 (1.4-3.8) 0.0046 2.04 (1.25-3.34) No Risk <0.0001 3.24 (2.1-4.9) <0.0001 2.66 (1.77-4.00) Survival (days) Mean 1047 833 1091 1248 Median 654 278 717 968 Range 1-3164 5-3154 17-3128 8-3164 Age (years) Mean 71.7 72.7 69.7 71.0 0.0031 Per 1 year 0.0057 Per 1 year Median 73 73 72 72 0.98 (0.96-0.99) 0.98 (0.97-0.99) Range 27-101 27-94 34-93 30-98 Sex Male 501(54.2%) 105(52.0%) 65(50.4%) 166(54.8%) Ref. Ref. Female 423(45.8%) 97(48.0%) 64(49.6%) 137(45.2%) 0.2196 1.23 (0.89-1.70) Race White 741(80.2%) 168(83.2%) 103(79.8%) 251(82.8%) Ref. Ref. All Other 183(79.8%) 34(16.8%) 26(20.2%) 52(17.2%) 0.8755 1.03 (0.69-1.56 Stage Early: 1-2 480(51.9%) 91(45.0%) 64(49.6%) 182(60.1%) Ref. Ref. Ref. Ref. Late: 3-4 444(48.1%) 111(55.0%) 65(50.4%) 121(39.9%) <0.0001 0.27 (0.19-0.36) <0.0001 0.29 (0.22-0.40) Note: GNRI: Geriatric Nutritional Risk Index, AF: Acceleration Factor, CI: confidence interval