Our proposed causal model between smoking, HPV antibodies, confounding factors, and subsequent HPV infection. We propose that smoking impairs the HPV-16 antibody response that in turn increases the risk of subsequent infection by that type (natural indirect effect). We also propose that smoking may have alternative mechanisms, represented here by the natural direct effect. Not depicted is the modeled interaction between smoking and HPV antibodies which could be a consequence of a more complex causal pathway involving smoking, sexual behavior, and immunity.