Investigators evaluated how screening for implausible self-reported intake alters apparent compliance with KDOQI nutrition targets in a Bangladeshi maintenance hemodialysis cohort. Intake plausibility used the EI:BMR ratio with an assumed PAL of 1.2; thresholds classified under-reporters (<0.8), acceptable reporters (0.8–1.3) and over-reporters (>1.3). Of 178 participants with complete dietary data, 65.2% were acceptable reporters, 32.0% under-reported, and 2.8% over-reported. In the unfiltered sample, 11.4% met the energy target (30–35 kcal/kg/day) and 33.0% met the protein target of ≥1.0 g/kg/day. Excluding implausible reports shifted adherence in opposite directions for intake versus restriction targets: energy adherence increased to 15.5% and protein adherence to 47.4%, whereas adherence to sodium and phosphorus limits fell (sodium: 73.6% to 65.5%; phosphorus: 88.2% to 84.5%). These opposing changes indicate that unscreened self-report can simultaneously overstate insufficiency and understate excess.
The team collected three non-consecutive 24-hour recalls per patient (one dialysis day, one nondialysis day, one weekend day), administered face-to-face during dialysis sessions by trained nutrition graduates using a multiple-pass method and standardized portion-estimation tools. Recipes and food items were entered into Nutritionist Pro™ using the Food Composition Table for Bangladesh. Energy and nutrient intakes were weight-adjusted and compared to KDOQI reference targets: energy 30–35 kcal/kg post-dialysis weight and protein 1.0–1.2 g/kg/day (≥1.2 g/kg/day used as the primary comparison endpoint for cross-study comparability). The energy intake–basal metabolic rate ratio was calculated using Harris–Benedict BMR estimates and post-dialysis body weight. The PAL of 1.2 reflected a sedentary maintenance dialysis population and matched comparable literature cut-offs cited by the authors.
The study enrolled 181 adults across four districts; mean age was 48 ± 13.8 years and sex distribution was 53% male, 47% female. Most patients (80.1%) dialysed twice weekly, with mean session duration 3.8 ± 0.2 hours. Hypertension was the leading reported cause of end-stage kidney disease (39.2%). Mean BMI (post-dialysis weight) was 22.8 ± 4.3 kg/m2 among the 178 analytic participants; 64.0% had normal BMI, 13.5% were underweight, and 22.5% were overweight or obese combined. Triceps skinfold thickness and mid-arm circumference differed by sex; TSF was higher among under-reporters. The authors report that over-reporters were numerically few (n = 5), rendering comparisons with that subgroup statistically unstable.
When analysed across the cohort, adherence to KDOQI energy and protein targets was low: only 20 of 176 patients met the energy target and 58 of 176 met the minimum protein target (≥1.0 g/kg/day). After excluding implausible reports, estimates of inadequate intake declined for energy and protein, indicating that underreporting inflated the apparent prevalence of insufficient intake. In contrast, restriction-target adherence (sodium and phosphorus) appeared higher in the unscreened dataset and fell after removing implausible reports, implying that underreporting masked dietary excess. The phosphorus-to-protein ratio exceeded 12 mg/g in nearly all participants (all but five), a finding with direct relevance to phosphorus management in dialysis patients.
Appetite was measured using the single-item Appetite and Diet Assessment Tool (ADAT) and grouped into good, fair, and poor categories. Poor appetite was common, affecting 42.0% of participants. The study team observed that poorer self-rated appetite did not correspond to lower measured energy or protein intake; instead, ADAT responses were associated with reporting plausibility (p = 0.024). This suggests appetite perception influenced how participants reported intake, rather than solely predicting actual lower intake values in this sample.
Descriptive statistics and group comparisons used one-way ANOVA and χ2 tests, with Fisher’s exact test applied where cell counts were small. A pre-specified sensitivity analysis combined under- and over-reporters into an ‘invalid reporter’ category to stabilize comparisons. The authors did not apply multiplicity correction across numerous nutrient comparisons and stated nominal p-values should be regarded as hypothesis-generating. Study design elements that limit inference include the cross-sectional, descriptive approach without a formal a priori power calculation for hypothesis testing; reliance on self-report even when plausibility-screened; and a very small over-reporter subgroup (n = 5) that limits subgroup stability. Missing biochemical data in some records and the use of a single PAL assumption (1.2) for all participants are further constraints. The paper does not report interventions, causality, or longer-term clinical outcomes.
Unanswered clinical questions the authors identify include whether plausibility-screened dietary assessment improves clinical decision-making in routine dialysis care, how best to integrate objective biomarkers or prospective monitoring in resource-limited settings, and what interventions most effectively correct the high phosphorus-to-protein ratios observed. These items remain to be addressed in future work.