Characteristics of participants.

Objectives
This study investigates the relationship between smoking and total BMD and examines the mediating role of inflammatory markers in this relationship.
Method
In total, 22,022 participants were included in this study, based on data from the National Health and Nutrition Examination Survey for the periods 2001–2006 and 2011–2018. Weighted linear regression models and restricted cubic splines(RCS) were leveraged to examine the linear or nonlinear relationship between serum cotinine levels and total BMD. Additionally, mediation analysis was leveraged to appraise the potential mediating effects of inflammatory markers, such as lymphocytes, monocytes, neutrophils, and platelets, in the relationship between cotinine and total BMD.
Results
After fully adjusting for all covariates, an increase of one unit in cotinine corresponded to a 0.00022 g/cm2 decrease in total BMD (Beta = −0.00022, 95% CI: −0.0003 ~ −0.0000, P = 0.0069). The RCS analysis indicated an “n-shaped” relationship between cotinine and total BMD (P-nonlinear = 0.0069). According to the mediation analysis, monocytes and neutrophils acted as mediators in the relationship between cotinine and total BMD, with mediation effects accounting for 19.8% and 19.6%, respectively.
Conclusions
Smoking serves as a risk factor for reduced BMD, and the impact on BMD is partially mediated by inflammatory markers such as monocytes and neutrophils. Platelets can moderate the effect of cotinine on total BMD to some extent.