Correlation of the Dietary Protein Intake between Those Estimated from a Short Protein Food-Recall Questionnaire and from 24-Hour Urinary Urea-Nitrogen Excretion in Stages 3-4 Chronic Kidney Disease Patients

INTRODUCTION: High protein intake may accelerate progression of chronic kidney disease (CKD). Estimation of dietary protein intake (DPI) is indispensable for management of CKD, but to achieve optimum DPI is quite challenging in routine clinical practice. We recently studied a beneficial effect of utilizing integrated care on the management of CKD at the rural community level. In that study, we created a short protein food-recall questionnaire (S-PFRQ) as a working tool to estimate DPI of the CKD patients during home visit by community health personnel. Herein, we reported the initial evaluation of the reliability of S-PFRQ from our previous study. OBJECTIVE: We compared the amount of DPI obtained from S-PFRQ with that obtained from protein-equivalent of total nitrogen appearance (PNA). METHODS: In the previous ESCORT-2 study, 914 patients with CKD stage 3 or 4, who were living in the rural area of Thailand, were prospectively followed while receiving integrated care for 36 consecutive months. During home visits by community nurses from subdistrict health centers, dietary food recall was made, recorded in S-PFRQ, and DPI was obtained. Among these, sixty patients were randomly selected, and 24-h urine was collected for urinary urea-N and estimation of PNA. A correlation was made between DPI obtained from S-PFRQ and PNA. RESULTS: The DPIs derived from S-PFRQ and PNA were 28.8 ± 14.8 and 39.26 ± 17.79 g/day, respectively. The mean difference and 95% CI between the 2 methods was −10.43 (−7.1 to −13.8) g/day, respectively (P < 0.001). Interclass correlation between these 2 methods was 0.24, P = 0.007. The difference between the 2 methods remained constant across different amounts of DPI. CONCLUSION: The DPI estimated from S-PFRQ significantly correlated to that from PNA. However, the S-PFRQ method yielded a DPI value which was about 10 g of protein or 25% less than the PNA method. Despite this amount of difference, this S-PFRQ is user-friendly and could be used during field work as an easy and simple tool for DPI estimation in resource-limiting condition.