Date Published: February 15, 2018
Publisher: Public Library of Science
Author(s): Ji-Sook Kong, Yeon-Kyung Lee, Mi Kyung Kim, Mi-Kyeong Choi, Young-Ran Heo, Taisun Hyun, Sun Mee Kim, Eun-Soon Lyu, Se-Young Oh, Hae-Ryun Park, Moo-Yong Rhee, Hee-Kyong Ro, Mi Kyung Song, Tatsuo Shimosawa.
This study was conducted to develop an equation for estimation of 24-h urinary-sodium excretion that can serve as an alternative to 24-h dietary recall and 24-h urine collection for normotensive Korean adults. In total, data on 640 healthy Korean adults aged 19 to 69 years from 4 regions of the country were collected as a training set. In order to externally validate the equation developed from that training set, 200 subjects were recruited independently as a validation set. Due to heterogeneity by gender, we constructed a gender-specific equation for estimation of 24-h urinary-sodium excretion by using a multivariable linear regression model and assessed the performance of the developed equation in validation set. The best model consisted of age, body weight, dietary behavior (‘eating salty food’, ‘Kimchi consumption’, ‘Korean soup or stew consumption’, ‘soy sauce or red pepper paste consumption’), and smoking status in men, and age, body weight, dietary behavior (‘salt preference’, ‘eating salty food’, ‘checking sodium content for processed foods’, ‘nut consumption’), and smoking status in women, respectively. When this model was tested in the external validation set, the mean bias between the measured and estimated 24-h urinary-sodium excretion from Bland-Altman plots was -1.92 (95% CI: -113, 110) mmol/d for men and -1.51 (95% CI: -90.6, 87.6) mmol/d for women. The cut-points of sodium intake calculated based on the equations were ≥4,000 mg/d for men and ≥3,500 mg/d for women, with 89.8 and 76.6% sensitivity and 29.3 and 64.2% specificity, respectively. In this study, a habitual 24-hour urinary-sodium-excretion-estimation model of normotensive Korean adults based on anthropometric and lifestyle factors was developed and showed feasibility for an asymptomatic population.
Sodium intake has been reported to be a known risk factor for hypertension in many, animal as well as epidemiological studies [1–4]. Excess sodium intake additionally is a known risk factor for cardiovascular diseases and stroke [5, 6], renal disease , stomach cancer , and osteoporosis [9, 10]. Although the WHO recommends a sodium intake of <2,000 mg/d (<5 g/d salt) to prevent diet-related chronic diseases , the average level among Koreans exceeds that level by approximately two-fold . Sodium intake is considered a national problem, and the Korean government has implemented policies to reduce it. The gold standard for estimation of sodium excretion or intake has been considered to be the 24-h urine collection method , which, not coincidentally, is the most valid and reliable. However, 24-h urine sampling imposes a high burden on subjects, and is also time- and cost-intensive in a large population. In this study, we developed alternative easy, simple, low-cost and gender-specific equations for estimation of 24-hUNa. The equations utilize a simple questionnaire that includes 7 variables for men (age, body weight, four dietary variables [‘eating salty food’, ‘Kimchi consumption’, ‘Korean soup or stew consumption’, ‘soy sauce or red pepper paste consumption’] and smoking status) and 7 variables for women (age, body weight, four dietary variables [‘salt preference’, ‘eating salty food’, ‘checking sodium content for processed foods’, ‘nut consumption’] and smoking status). Each equation’s validity was evaluated with an independent dataset. We established an easy, low-cost, and gender-specific method for estimation of 24-hUNa using daily-habits questionnaire that accounts for age, body weight, dietary behavior, and smoking. This study enrolled healthy volunteers with normal blood pressure and no other diseases. Our habitual 24-hour urinary-sodium-excretion-estimation model of normotensive Korean adults based on anthropometric and lifestyle factors showed good performance for an asymptomatic population. This suggests that it is a feasible potential model for monitoring of sodium intake in a Korean population. Further study of the prediction model using simple questionnaires yielding the most accurate estimation of 24-h urinary-sodium excretion at both the population-mean and individual levels would be recommended. Source: http://doi.org/10.1371/journal.pone.0192588