Prognostic nutritional index, Naples prognostic score and Osaka prognostic score as predictors of acute respiratory distress syndrome in non-diabetes mellitus patients with sepsis: a loss of predictive efficacy of NPS and OPS in diabetes mellitus cohorts.

BACKGROUND

Acute respiratory distress syndrome (ARDS) induced by sepsis is associated with uncontrolled immunoinflammatory responses and nutritional status. The purpose of this study was to explore the value of several prognostic scoring indices (Prognostic nutritional index (PNI), Naples prognostic score (NPS), and Osaka prognostic score (OPS)) in predicting ARDS risk among sepsis patients.

METHODS

2089 sepsis patients were retrospectively enrolled, and divided into diabetes mellitus and non-diabetes mellitus groups based on the presence of comorbid diabetes mellitus. Baseline clinical data and laboratory test indicators at admission were collected, and the PNI, NPS, and OPS were calculated. The differences in PNI, NPS, and OPS between ARDS and non-ARDS patients were compared. Logistic regression analysis was performed to explore the associations of PNI, NPS, and OPS with ARDS in septic patients.

RESULTS

1772 patients did not develop ARDS, whereas 317 cases with ARDS. PNI level in ARDS patients was lower than that in patients without ARDS (p < 0.001). ARDS patients had notably higher proportions of NPS scores of 3 or 4 points and OPS scores of 2 or 3 points compared with non-ARDS patients (p < 0.001). The proportion of diabetes mellitus was significantly lower in the ARDS group than non-ARDS group (p = 0.030). In non-diabetes mellitus patients, logistic regression analysis revealed that low PNI (odds ratio [OR]: 3.764, 95% confidence interval [CI]: 2.549-5.557, p < 0.001), NPS score 3-4 (OR: 2.537, 95% CI: 1.302-4.944, p = 0.006), and OPS score 2-3 (OR: 3.189, 95% CI: 1.326-7.670, p = 0.010) were independently associated with ARDS. In patients with diabetes mellitus, low PNI (OR: 2.037, 95% CI: 1.256-3.306, p = 0.004) was independently associated with ARDS, however, neither NPS nor OPS yielded statistically significant results.

CONCLUSIONS

PNI, NPS, and OPS were predictive indicators for ARDS risk in sepsis patients without diabetes mellitus; however, NPS and OPS lack corresponding predictive value in diabetes mellitus cohorts.