Patient Generated Subjective Global Assessment as a prognosis tool in women with gynecologic cancer

Abstract

The aim of this study was to assess the nutritional status (NS) of women hospitalized for gynecologic tumors and relate it to such outcomes as hospital length of stay and 1-y mortality. Methods: We assessed 146 women diagnosed with gynecologic tumors who were admitted to a referral oncologic hospital in November 2012. Data collected included medical history, duration and reason for admission, and cases of death within 1 y. Results: NS was assessed using Patient-Generated Subjective Global Assessment (PG-SGA). The receiver operating characteristic curve was used to define the best cutoff point for discriminating individuals who did or did not die. We used proportional hazards regression to assess associations between malnutrition and 1-y mortality. According to the PG-SGA, 62.4% of the women were classified as being at nutritional risk or having moderate or severe malnutrition. Sorting patients by stage of cancer, there was no statistical difference in NS classification according to the different cancer sites. The median hospital stay, in days, was statistically lower in patients classified as well nourished. Individuals with a score above the cutoff point of 10 were 30.7 times more likely (95% confidence interval, 11.8–79.4) to die. There was a 52.1% rate of mortality within 1 y. Patients classed as having some degree of malnutrition had a significantly lower median survival rate. A diagnosis of cervical cancer and severe malnourishment increases the likelihood of death. Conclusions: Our findings suggest that the PG-SGA can be considered not just as an indicator of nutritional risk, but also as a major predictor of prognosis and mortality in this population.