Estimating Differences in Risk of Chronic Kidney Disease Based on Water Intake in a National Sample

Abstract

<b><i>Background:</i></b> In agricultural communities in Central and South America, Egypt, India, and Sri Lanka, an unexplained form of chronic kidney disease affects agricultural workers. Termed chronic kidney disease of unknown origin (CKDu), it disproportionately affects young men in their 30s–40s and is unrelated to the traditional risk factors of diabetes, hypertension, and obesity [1–3]. Recent investigations suggest that agricultural work in the USA carries similar risks, as reduced kidney function has been found among those working in US agriculture [4–5]. However, researchers are yet to determine the etiology of the disease [6–8]. Central to the hypotheses of CKDu is the reduced blood flow to the kidneys due to inadequate hydration during periods of intense physical labor. <b><i>Objectives:</i></b> The primary aim of the current investigation was to identify if a relationship between hydration and kidney function exists among the general population by using the data from the National Health and Nutrition Examination Survey (NHANES). We hypothesize that reduced hydration will be associated with reduced kidney function. <b><i>Methods:</i></b> Data were retrieved from the NHANES dataset from 3 sample years 2005/2006, 2007/2008, and 2011/2012. Data were merged across all 3 periods with survey weights adjusted for combining across multiple years. Participants were excluded if they had missing data for hydration or kidney function, or if they were &#x3c;19 year. Kidney function was categorized low risk, moderate risk, or high risk for impaired function based on estimated glomerular filtration rate and albumin creatinine ratio according to the National Kidney Foundation [9]. Hydration was classified based on total water intake (TWI) extracted from plain water intake and water from food. Participants were labeled as high if they met or exceeded sex-specific water recommendations, 3.7 and 2.7 L/day for men and women, respectively; otherwise they were labeled as low. A survey-weighted proportional odds logistic regression model was fitted to assess the association between water intake and kidney function, while controlling for other demographic, socio-economic, behavioral, and socio-economic risk factors [10–12]. <b><i>Results:</i></b> Of the 13,056 participants initially sampled, 10,651 participants are included in the analysis after cleaning and including survey weights. 9,125 (85.67%) of participants were in the low-risk group, 1,128 (10.59%) were classified as medium-risk, while the remaining 398 (3.74%) were high risk (Fig. 1). Adjusting for survey weights, results suggest that the estimated rate of high-risk kidney function was 5% more for low water drinkers compared to high water drinkers (Fig. 2). There is strong evidence of a difference in CKD risk categories based on TWI (χ²(1) = 13.1, <i>p</i> value &#x3c;0.0001) from a survey-weighted proportional odds logistic regression model, but only moderate evidence of a difference when controlled for sodium/potassium ratio, education, age, gender, ethnicity, income, BMI, blood pressure, diabetes, smoking, and alcohol consumption (χ²(1) = 3.3, <i>p</i> value = 0.067). <b><i>Conclusions:</i></b> Not meeting recommended daily TWI was associated increased presentation of high-risk kidney function. Even though the NHANES data are not focused on areas where chronic kidney disease is prevalent, results from this are an indication that hydration does play a role in kidney function.