Our study also has several limitations. A major limitation
is that the diagnosis of HPV infection was presumptive and
based on the presence of an abnormal pap test. Studies
do however show a strong association (97–99%) between
cervical dysplasia and the presence of HPV infection [15, 16].
Another limitation of our study is potential misclassification
as women who had negative cytology and no STIs may have
had HPV infection. However, this misclassification would
be in the direction of the null hypothesis and should not
have appreciably impacted our results. Finally, there may be
unknown confounder that exists and were not controlled for
in the multivariable analysis.