Phase Angle Is a Stronger Predictor of Hospital Outcome than Subjective Global Assessment-Results from the Prospective Dessau Hospital Malnutrition Study

Abstract

This prospective cohort study of 16,943 consecutive patients compared phase angle (PhA, foot-to-hand at 50 kHz) and subjective global assessment (SGA) to predict outcomes length of hospital stay (LOS) and in-hospital mortality in patients at risk of malnutrition (NRS-2002 ≥ 3). In 1505 patients, the independent effects on LOS were determined by competing risk analysis and on mortality by logistic regression. In model I, including influence factors age, sex, BMI, and diagnoses, malnourished (SGA B and C) patients had a lower chance for a regular discharge (HR 0.74; 95%CI 0.69−0.79) and an increased risk of mortality (OR 2.87; 95%CI 1.38−5.94). The association of SGA and outcomes regular discharge and mortality was completely abrogated when PhA was added (model II). Low PhA reduced the chance of a regular discharge by 53% in patients with a PhA ≤ 3° (HR 0.47; 95%CI 0.39−0.56) as compared to PhA > 5°. Mortality was reduced by 56% for each 1° of PhA (OR 0.44; 95%CI 0.32−0.61). Even when CRP was added in model III, PhA ≤ 3° was associated with a 41% lower chance for a regular discharge (HR 0.59; 95%CI 0.48−0.72). In patients at risk of malnutrition, the objective measure PhA was a stronger predictor of LOS and mortality than SGA.

Keywords: bioimpedance analysis; inflammatory status; medical patients; nutritional status; screening; surgical patients.

Figure 1

Flow diagram of study population for cohort 0 (all patients screened) and study cohort (consenting patients screened NRS ≥ 3). The multivariate analysis was performed on the study cohort.

Figure 2

Cumulative incidence of discharge (A) and death (B) over 50 days after hospital admission according to SGA categories A (n = 584), B (n = 783), C (n = 131). SGA: subjective global assessment.

Figure 3

Cumulative incidence of discharge (A) and death (B) over 50 days after hospital admission according to PhA categories ≤ 3° (n = 327), 3–4° (n = 482), 4–5° (n = 424), > 5° (n = 272).

Figure 4

Receiver—operating—characteristic (ROC) curve for the prediction of in-hospital death by numerical PhA. AUC = area under the curve.

Figure 5

Cumulative incidence of death (A) and discharge (B) over 50 days after hospital admission according to PhA < 5th percentile (n = 777) or ≥ 5th percentile (n = 565) as classified according to age, sex, BMI-class, and population-specific reference values [26]. Areas in grey shade indicate 95%CI.