Study Context
This cross-sectional study was undertaken as part of the project “Investigation of nutrition and diet of patients with pulmonary tuberculosis in poor areas in China” supported by the WHO Regional Office in the Western Pacific. It was conducted from November 2015 to April 2017 in two counties, that is, Linyun county located in Guangxi province and Lin county located in Shanxi Province. These two counties were national poverty counties and TB control programs were carried out well. TB case notification rates in 2016 were 63.16 per 100,000 in Lin county and 106.97 per 100,000 in Lingyun county, respectively.
Participants
Adult patients (age≥18 years) with active TB registered in Tuberculosis management information system from Nov 1st, 2015 through May11th, 2017 and signed consent form were recruited. Patients with extrapulmonary tuberculosis, and those aged 18 years and below or with severe complications were not eligible for the study. Pregnant or breast-feeding women, and those who refused to sign the consent form were also excluded. China Health and Nutrition Survey (CHNS) was conducted in 2015 among the general population in these two countries with the same component for nutrition assessment. The findings of the CHNS was taken as a comparison group for our study. Detailed information about CHNS 2015 was available on website https://www.cpc.unc.edu/projects/china/about/design/sample.
Sample Size
As BMI is commonly used in nutrition assessment, we applied the prevalence of BMI<18.5 which is also defined as malnutrition(13) ],[in the calculation of sample size. We assumed that the prevalence of malnutrition in the general population in poor areas and TB patients would be 6.7% (14)and 25.0%(2), respectively, and the required sample size was calculated to detect difference in the two proportions. The probability of a type I error was set at 0.05, the power of the study was estimated at 90% and the design effect was set at 1, determining a sample size of TB patients was 77 per study site. Considering participants’ refusal, we expanded the sample size to 150 per study site, and the final sample size of TB patients was 300 totally.
As for general population, after exclusion of subjects with missing data, the sample size was 496. A multistage, random cluster process was used to draw the sample of general population in CHNS 2015.
Socio-demographic and behavioral factors
Data for associated factors were taken from a questionnaire on personal information, including socio-demographic data (age, gender, education level, marital status, occupation and household income level) and behavioral data (alcohol consumption, smoking status and eating out-of-home). Age was categorized into 3 groups, 18-49 years, 50-64 years, 65 years and above, based on Chinese Dietary Reference Intakes (DRIs) 2013(12). Education level was categorized into 2 groups(15), that is, primary school and below, junior middle school and above. Household income level was evaluated by annual household income and was categorized into 3 groups, <20,000 yuan, 20,000-40,000 yuan, and ≥40,000 yuan(15). Alcohol consumption was defined as drinking wine, beer and Chinese spirit now or ever. A smoker was defined as smoking now or smoking previously but quitting now. Eating out-of-home was defined as eating at least one meal away from one’s own home or their residents’ home during the survey(16).
Assessment of Nutrient intake
Trained staff performed face-to-face interviews on each participant to obtain dietary intake data through a 2-day 24-hour dietary recall (24hdr) questionnaire, which was adapted from the method of 3-day 24hdr(17). Participants were instructed to record all food intake at home and away from home in the previous 2 days (one weekday and one weekend day). Consumptions of condiments were also recorded through a questionnaire. All investigations were completed after the patient registering in Tuberculosis management information system and before anti-tuberculosis treatment. Data for dietary intake of general population was obtained from a 3-day 24hdr questionnaire (two weekdays and one weekend day).
Total energy, four kinds of macronutrients and sixteen kinds of micronutrients were evaluated. Nutrient intakes of each patient were calculated based on Chinese Food Composition Tables (CFCT) 2004(18).
We also evaluated TB patients’ nutritional intake by applying Chinese dietary reference intakes (DRIs) 2013, with mean daily nutrient intakes compared with Recommended Nutrient Intakes (RNIs) and Adequate Intakes (AIs).
Statistical analysis
A Chi-square test and a Wilcoxon rank sum test were used to compare the characteristics of TB patients and the general population. Data of daily nutrient intakes were presented as means ± standard deviation (SD). A Wilcoxon rank sum test was conducted to compare differences in daily energy and nutrient intakes of TB patients and general population. As dietary recommendations are different for men and women, we compared TB patients’ and general populations’ daily nutrient intakes with DRIs by gender. Comparisons in proportion of participants whose nutrient intakes were below RNI/AI were analyzed with a Chi-square test and Fisher exact test. Since protein-calorie malnutrition (PCM) is the most common form of undernutrition in TB patients[11], we only examine factors related to insufficient energy and protein intakes using univariate and multivariate logistic regression models, and patients were classified into two groups: below RNI/AI and above RNI/AI.A P-value of less than 0.05 was considered statistically significant. All analyses were performed using SAS 9.4 (SAS Institute Inc, Cary, NC, USA).
Quality Control
The study was carried out on the basis of CHNS 2015, from which the investigating method and tool were used in our study. All on-site investigations were carried out by the county-level CDC and interviewers were trained with a standard protocol. Data was checked for completeness and accuracy on the day of investigation and sampled by provincial CDC for verification later. All data was double-entered into a database specially designed for this project.