Analysis of data from two large controlled trials of the bivalent human papilloma virus (HPV) vaccine (Cervarix™) found no evidence that the vaccine was associated with increased risk of miscarriage.

HPV vaccine is likely to be given to large numbers of young women of reproductive age, therefore any effect of the vaccine on pregnancy could reduce the overall health benefits of vaccination. This paper, prepared largely independently of the manufacturer, examines data on miscarriage rates in two large trials of the bivalent HPV vaccine (Cervarix, GSK). The two trials, termed PATRICIA and CVT, involved women aged between 15 (PATRICIA) or 18 (CVT) and 25 who received three doses of the HPV vaccine at baseline, 1, and 6 months, or hepatitis A vaccine as control. They had a pregnancy test before each dose, and were asked to use contraception from one month before the first vaccination to two months after the last. As part of follow-up, information on pregnancies and their outcomes was collected throughout the study period. Primary outcome for this analysis was miscarriage and other pregnancy outcomes.

In total, 26,130 women were enrolled into the two studies (13,075 to HPV vaccine and 13,055 to hepatitis A vaccine). There were 3,599 pregnancies and 373 miscarriages eligible for analysis, and the rates were 11.5% and 10.1% in the HPV and hepatitis A vaccine arms respectively – both within the range of international rates. In the primary analysis, there was no statistically significant difference between the two values (one-sided P=0.16). A number of sophisticated statistical analyses were also carried out to reduce the effects of not knowing the exact date of conception: these too indicated no significant difference between the groups. In a range of subset analyses, there was no indication of any statistically significant difference between the groups for any subset.

The authors conclude that overall, there was no evidence for any association between the bivalent HPV vaccine and miscarriage. They note that there was a variation in the differences for pregnancies conceived closest to vaccination: this may be either a real small effect or statistical noise. As a result, the possibility of a small increase in risk around this point cannot be excluded completely and further analysis of trial data is needed to examine this further.