The association between dietary inflammatory index and latent tuberculosis infection: Confounding role of country of birth.

BACKGROUND

Mounting evidence suggests that diet-induced inflammation may modulate the risk of Latent Tuberculosis Infection (LTBI). However, the reported association between the Dietary Inflammatory Index (DII) and LTBI remains inconsistent, and potential confounding factors such as country of birth have not been fully elucidated in previous studies.

OBJECTIVE

This study aimed to evaluate the confounding effect of country of birth on the apparent association between DII and LTBI, using data from the National Health and Nutrition Examination Survey (NHANES). A key hypothesis was tested that the observed relationship between DII and LTBI could be explained by the stratification of nativity status.

METHODS

A cross-sectional analysis was conducted on 3892 eligible participants from NHANES with no missing values for core variables. Comprehensive variables were collected, including core indicators DII,LTBI status, and active tuberculosis status defined in accordance with the diagnostic criteria specified in the NHANES operational guidelines), demographic characteristics (country of birth, age, gender, ethnicity, household size), socioeconomic factors (educational attainment, Poverty Income Ratio [PIR], marital status), and health-related covariates (Body Mass Index [BMI], smoking habits, alcohol intake, Hypertension [HPT], Diabetes Mellitus [DM]). Logistic regression models, smooth curve fitting, and stratified analyses by country of birth (US-born vs. non-US-born) were applied to quantify the association between DII and LTBI, with sequential adjustment for potential confounders.

RESULTS

The DII values of the participants ranged from -4.63 (highly anti-inflammatory diet) to 5.47 (highly pro-inflammatory diet), with a mean of 1.40±1.91, indicating a slight pro-inflammatory dietary tendency overall. LTBI prevalence varied by TB classification, and TB-positive groups showed significantly higher active tuberculosis rates than the TB-negative group. In the initial multivariate logistic regression model, adjusted for demographic, socioeconomic, and health-related factors (excluding country of birth), a significant inverse association was observed between DII and LTBI (OR = 0.93, 95 % CI 0.88-0.98, p = 0.011). However, this association was attenuated to non-significance after further adjustment for country of birth (OR = 0.96, 95 % CI 0.90-1.01). Stratified analyses confirmed the lack of a significant association between DII and LTBI in both US-born (OR = 0.95, 95 % CI 0.88-1.04) and non-US-born (OR = 0.96, 95 % CI 0.88-1.03) subgroups, with all interaction p-values > 0.05. Additionally, the initial linear dose-response relationship between DII and LTBI was no longer maintained after controlling for country of birth. The authors acknowledge that the country of birth variable is a relatively crude indicator, and residual confounding may persist in the current analysis.

CONCLUSIONS

In this nationally representative sample, the initially observed inverse association between DII and LTBI was not sustained after adjusting for country of birth, indicating that nativity status is a critical confounder of the apparent DII-LTBI relationship. The findings highlight that the previously reported protective effect of an anti-inflammatory diet against LTBI may be attributable to the confounding influence of country of birth rather than a direct causal effect. Future epidemiological studies investigating diet-infection relationships should rigorously assess and control for country of birth as a potential confounding factor, and explore more refined indicators to reduce residual confounding.