Prognostic value of various nutritional risk markers in patients hospitalized for the treatment of genitourinary cancer: A retrospective study

• Published in: Clinical nutrition ESPEN Vol. 55; pp. 44 - 50
• Main Authors: Takagi, Kimiaki, Takahashi, Hiroshi, Miura, Tomomi, Yamagiwa, Kasumi, Kawase, Kota, Muramatsu-Maekawa, Yuka, Yamaha, Masayoshi, Nakane, Keita, Koie, Takuya, Minoshima, Kenichi
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01-06-2023
• Subjects: Aged; Genitourinary cancer; Humans; Inpatients; Mortality; Nutritional assessment; Nutritional Status; Predictability; Prognosis; Retrospective Studies; Urogenital Neoplasms - diagnosis; Urogenital Neoplasms - therapy; MNA; NRI; Mortality; CI; Genitourinary cancer; HR; aHR; SGA; MNA-SF; IDI; Nutritional assessment; CONUT; Predictability; vs; GNRI; BMI
• ISSN: 2405-4577; 2405-4577
• DOI: 10.1016/j.clnesp.2023.03.002

Because malnutrition adversely affects the prognosis of patients with cancer, accurate nutritional status assessment is important. Therefore, this study aimed to verify the prognostic value of various nutritional assessment tools and compare their predictability.
We retrospectively enrolled 200 patients hospitalized for genitourinary cancer between April 2018 and December 2021. Four nutritional risk markers, namely, Subjective Global Assessment (SGA) score, Mini-Nutritional Assessment-Short Form (MNA-SF) score, Controlling Nutritional Status (CONUT) score, and Geriatric Nutritional Risk Index (GNRI), were measured at admission. The endpoint was all-cause mortality.
SGA, MNA-SF, CONUT, and GNRI values were all independent predictors of all-cause mortality (hazard ratio [HR] = 7.72, 95% confidence interval [CI]: 1.75–34.1, P = 0.007; HR = 0.83, 95% CI: 0.75–0.93, P = 0.001; HR = 1.29, 95% CI: 1.16–1.43, P < 0.001; and HR = 0.95, 95% CI: 0.93–0.98, P < 0.001, respectively) even after adjustment for age, sex, cancer stage, and surgery or medication. However, in the model discrimination analysis, the net reclassification improvement of the CONUT model (vs. SGA: 0.420, P = 0.006 and vs. MNA-SF: 0.57, P < 0.001) and GNRI model (vs. SGA: 0.59, P < 0.001 and vs. MNA-SF: 0.671, P < 0.001) were significantly improved compared to the SGA and MNA-SF models, respectively. The combination of CONUT and GNRI models also had the highest predictability (C-index = 0.892).
Objective nutritional assessment tools were superior to subjective nutritional tools in predicting all-cause mortality in inpatients with genitourinary cancer. Measurement of both the CONUT score and GNRI might contribute to a more accurate prediction.