Validating Appetite Assessment Tools Among Patients Receiving Hemodialysis.

OBJECTIVE

To test the performance of appetite assessment tools among patients receiving hemodialysis (HD).

DESIGN

Cross-sectional.

SUBJECTS

Two hundred twenty-one patients receiving HD enrolled in seven dialysis facilities in Northern California.

INTERVENTION

We assessed 5 appetite assessment tools (self-assessment of appetite, subjective assessment of appetite, visual analog scale [VAS], Functional Assessment of Anorexia/Cachexia Therapy [FAACT] score, and the Anorexia Questionnaire [AQ]).

MAIN OUTCOME MEASURES

Reported food intake, normalized protein catabolic rate, and change in body weight were used as criterion measures, and we assessed associations among the appetite tools and biomarkers associated with nutrition and inflammation. Patients were asked to report their appetite and the percentage of food eaten (from 0% to 100%) during the last meal compared to usual intake.

RESULTS

Fifty-eight (26%) patients reported food intake ≤ 50% (defined as poor appetite). The prevalence of anorexia was 12% by self-assessment of appetite, 6% by subjective assessment of appetite, 24% by VAS, 17% by FAACT score, and 12% by AQ. All the tools were significantly associated with food intake ≤ 50% (P < .001), except self-assessment of appetite. The FAACT score and the VAS had the strongest association with food intake ≤ 50% (C-statistic 0.80 and 0.76). Patients with food intake ≤ 50% reported weight loss more frequently than patients without low intake (36% vs 22%) and weight gain less frequently (19% vs 35%; P = .03). Normalized protein catabolic rate was lower among anorexic patients based on the VAS (1.1 ± 0.3 vs 1.2 ± 0.3, P = .03). Ln interleukin-6 correlated inversely with food intake (P = .03), but neither interleukin-6 nor C-reactive protein correlated with any of the appetite tools. Furthermore, only the self-assessment of appetite was significantly associated with serum albumin (P = .02), prealbumin (P = .02) and adiponectin concentrations (P = .03).

CONCLUSIONS

Alternative appetite assessment tools yielded widely different estimates of the prevalence of anorexia in HD. When considering self-reported food intake as the criterion standard for anorexia, the FAACT score and VAS discriminated patients reasonably well.