Fig. 4

Levels of salivary immune markers following Prenatal Total Oral Rehabilitation (PTOR). A Eotaxin levels in saliva B ITAC Eotaxin levels in saliva, and C MCP1 Eotaxin levels in saliva. To assess the PTOR-related longitudinal association between immune markers and possibly related factors, we first fitted linear mixed effects models with random intercepts for participants of salivary cytokine levels (pg/ml) of immune markers over time on demographic and clinic factors. The model allows us to compare the cytokine levels of immune markers at follow-up visits to those at baseline by setting the time points as categorical. Other variables included were race, education, vaginal yeast infection, work status, whether brushing teeth twice daily, gestational age (> = 20 weeks), ICDAS level, BOP, pain score, and the number of decayed teeth. We then assessed the association between the microbial species and the immune markers using Lasso-penalized linear mixed effects models with random intercepts using saliva and plaque samples, respectively. The results are shown in tables on the right side for each immune marker. The microbiome data in both samples, together with the same set of covariates as in the previous models, were used as candidate variables. The tuning parameters were chosen using Jones’ BIC criterion for longitudinal data