The incidence of chronic diarrhea decreases with increasing serum calcium levels: a cross-sectional study based on NHANES 2005–2010

This cross-sectional study utilized continuous data from the National Health and Nutrition Examination (NHANES) study from 2005 to 2010. Our study included participants aged 20 years and older, who completed interviews and exams at designated mobile examination centers (MECs). Participants with gut health problems, missing serum calcium result and missing covariate data were excluded. NHANES is a health screening program that uses a multi-level probability model to select respondents for a representative sample, aiming to assess the health and nutritional status of American citizens [10]. Staff conducted extensive household interviews to collect demographic data and health histories.. At the MEC, a physical examination is performed and a blood sample is taken. The CDC Laboratory Science Division analyzed the serum samples obtained from the National Center for Environmental Health.

Participants provided informed consent to participate in the NHANES study, which was approved by the National Health Statistics Ethics Review Board. The initial study protocol is available on the National Institutes of Health Statistical Review Board website (https://​www.​cdc.​gov/​nchs/​nhanes/​irba98.​htm). The study was formally authorized by the Institutional Review Board. (Protocol No. 2005–06). Research data are freely available to researchers worldwide. As this is a secondary cross-sectional analysis, it did not require additional approval from the institutional ethics review committee.

Serum calcium was measured from blood samples collected at the MEC. The collected samples were stored at -30 °C, until they ware sent to the National Centre for Environmental Health for analysis. Detailed specimen collection and processing instructions from the NHANES Laboratory/Medical Technician Procedures Manual (LPM) were followed. In this study, serum calcium was analyzed as a continuous variable.

The Gut Health Questionnaire was completed in the MEC survey room using a computerized personal survey system. The Bristol stool scale, commonly used in research and clinical practice, was used to define chronic diarrhea [11]. Previous studies have demonstrate that individual stool results are strongly correlated with gastrointestinal transit time measurements of radiopaque markers [12].

Participants were shown a card with color pictures and descriptions of the seven Bristol stool types (types 1 to 7). Those who correlated their regular stool as being most similar to stool type 6 (fluffy stool with rough edges and a moist appearance) or type 7 (watery, non-solid stool) were classified as having chronic diarrhea [12, 13].

Age, sex, race, marital status, educational level, household income, smoking, alcohol abuse, albumin, serum iron, cholesterol, total protein, uric acid, serum vitamin D, sodium, potassium, and hemoglobin were analyzed in this study. Calcium intake, body mass index (BMI), diabetes, arthritis, renal failure, osteoporosis, thyroid disease, and physical activity were listed as possible confounding factors. Race and ethnicity were grouped into Mexican–American, Other Hispanic, Non-Hispanic White, Non-Hispanic Black and Other. Education level was divided into those without a high school diploma and those with at least a high school education. Marital status was classified as married, widowed/divorced and never married. Household income was divided into high earners and low earners, using a $20,000 threshold.

The Beckman Synchron LX20 analyzer was used to detect serum albumin (g/L), serum iron (μmol/L), serum cholesterol (mmol/L), serum phosphorus (mmol/L), total protein (g/L), uric acid (μmol/L), blood sodium (mmol/L), blood potassium (mmol/L). Analysis of 25-hydroxyvitamin D3 [25(OH)D3], 25-hydroxyvitamin D2 [25(OH)D2] and 3-epi serum samples by high performance liquid chromatography tandem mass spectrometry (UHPLC-MS/MS)- 25- Concentration of hydroxyvitamin D3 [3-epi-25(OH)D3]. Total vitamin D status was defined as the sum of 25(OH)D3 and 25(OH)D2. Hemoglobin (g/dL) was based on a Beckman Coulter counter calibration method using an automatic dilution mixer for sample processing and measured by a single-beam photometer.

Calcium intake was obtained from the MEC ​​food recall interview, and dietary information was collected 24 h before the interview. BMI was calculated from the measured height and weight, and expressed in kilograms per square meter. Weight was measured on a digital scale in pounds and converted to kilograms. Altitude was measured with an electronic altimeter, accurate to the millimeter.

Smoking status was categorized into two groups: participants who smoked 100 or more cigarettes during their lives were described as ≥ 100 cigarettes during their lifetime, those who smoke less than 100 cigarettes in their lifetime are considered to be non-smokers. Those who consume at least 12 alcoholic drinks annually were defined as drinkers [14]. Diabetes, arthritis, kidney disease, thyroid disease and osteoporosis were defined according to participant to the following questions: "Have doctors or other health professionals ever told you that you have diabetes, arthritis, kidney disease, thyroid disease or osteoporosis?" High intensity exercise was defined as at least 10 min of vigorous activity (exercises that resulted in profuse sweating and a significant increase in respiratory or heart rates) in the past 30 days. Moderate intensity exercise was defined as at least 10 min of physical activity (exercises that resulted in light sweating or a mild to moderate increase in respiratory or heart rates) during the past 30 days.

All analyses were performed using the R statistical package (http://​www.​R-project.​org, R Foundation) and, Free Statistical software, version 1.7.1 [15]. Demographic and clinical data are described using means ± standard deviations and frequencies (percentages). Data that followed a normal distribution were analyzed using the t-test, while non-parametric continuous variables were analyzed using the Kruskal–Wallis test. In multiple logistic regression, serum calcium was analyzed as both a continuous and categorical variable. Model 1 adjusted for age, gender, race, education, marital status, and household income, Model 2 adjusted for model 1 variables as well as including smoking status, drinking, high intensity exercise, moderate intensity activity, and body mass index, and Model 3 adjusted for model 2 as well as serum albumin, serum iron, serum cholesterol, serum phosphorus, serum total protein, uric acid, serum vitamin D, serum sodium, serum potassium, hemoglobin, and calcium intake. Model 4 adjusted for model 3 as well as diabetes, arthritis, renal failure, osteoporosis, thyroid disease, and depression. A sensitivity analysis was performed by excluding outliers with serum calcium outside the mean ± 2SD (0–2.6 mmol/L) or mean ± 3SD (0–2.7 mmol/L) ranges. P value < 0.05 was considered statistically significant. The dose–response association between serum calcium and risk of diarrhea was analyzed using a restricted cubic spline plot.

This study recruited 31,034 potential NHANES participants (2005–2010), of whom 17,132 adults (aged 20 years or older) completed interviews and were screened by the MEC for inclusion in our study. Participants who did not respond to the gut health and serum calcium questionnaires were excluded (n = 3,198). After eliminating participants with missing covariate data (n = 1592), 12,342 participants were included in the final analysis. Figure 1 shows a flowchart of the exclusion criteria. Table 1 shows the baseline characteristics of patients with diarrhea. In our study, diarrhea was more likely in participants who were male, non-Hispanic white, less educated, smoker, alcoholic, hypertensive, and those with little work or leisure time. Daily dietary intakes of serum albumin, serum calcium, serum iron, serum vitamin D, hemoglobin, uric acid, and calcium were significantly lower in participants with diarrhea compared with participants without diarrhea. Participants in the diarrhea group were characterized by their older age and high BMI.

A multivariate logistic regression analysis is presented in Table 2. In the unadjusted model, serum calcium was inversely related to diarrhea (OR 0.26; 95% CI 0.13–0.53). The outcome was similar after adjusting for age, gender, race, education, marital status, and household income (OR, 0.28; 95% CI, 0.14–0.58). After adjusting for other potentially confounding factors, including smoking status, drinking, physical activity, BMI, albumin, serum iron, cholesterol, phosphorus, total protein, uric acid, serum vitamin D, sodium, potassium, hemoglobin, calcium intake, diabetes, arthritis, kidney failure, osteoporosis, thyroid disease, depression, the negative association remained significant (all P < 0.001). This significant correlation also appeared in several adjusted models when serum calcium was analyzed as a categorical variable.

The dose-response relationships between serum calcium and diarrhea risk was examined using a restrictive cubic spline analysis. Figure 2 shows that a non-linear relationship between serum calcium and the risk of diarrhea does not hold. Result from the sensitivity analysis supported our findings.