Scheller NM, Pasternak B, Svanström H, Hviid A. Quadrivalent Human Papillomavirus Vaccine and the Risk of Venous Thromboembolism. JAMA. 2014;312(2):187-188. doi:10.1001/jama.2014.2198
A potential association between quadrivalent human papillomavirus (HPV) vaccination and venous thromboembolism (VTE) has been reported in 2 postlicensure safety studies.1,2 An analysis of the Vaccine Adverse Event Reporting System database found disproportionately high reporting of VTE following vaccination,1 and a study using the Vaccine Safety Datalink reported a doubling of VTE risk.2
Safety concerns can compromise immunization programs to the detriment of public health, and timely evaluations of such concerns are essential.3 In a self-controlled case-series (SCCS) study, we evaluated the potential link between quadrivalent HPV vaccination and VTE.
Linking individual-level information from Danish national demographic and health care registers, described in detail elsewhere,4 we constructed the source population of all Danish women aged 10 through 44 years from October 1, 2006, through July 31, 2013. The study was approved by the Danish Data Protection Agency. Ethics approval was not required.
Information on vaccination, use of oral contraceptives, use of anticoagulants, and the outcome of a first hospital diagnosis of VTE not related to pregnancy, surgery, or cancer was obtained from Danish registers.4 Because vaccinated and unvaccinated women may differ in ways that are difficult to measure and control for, we used the SCCS method in a population of women with a first diagnosis of VTE.5,6 By taking advantage of within-person comparisons, the SCCS design eliminates confounding by time-independent factors.
Cases were followed up from 10 years of age or October 1, 2006, until emigration, death, 45 years of age, or July 31, 2013. Incidence ratios were derived by conditional Poisson regression (SAS version 9.2, SAS Institute Inc), comparing incidence rates of VTE during predefined risk periods after each vaccine dose with all other observed periods in each individual (control periods). The main risk period was defined as 1 to 42 days from vaccination and was divided into shorter periods for additional analyses (1-14, 15-28, and 29-42 days).
Our first analysis included all VTE cases, both those exposed and unexposed to the quadrivalent HPV vaccine. In the SCCS method, unexposed cases do not contribute to estimates of the association between outcome and exposure but can be included in age adjustment. Second, we performed an analysis including only cases with exposure to the quadrivalent HPV vaccine to ensure that results did not vary.
Third, to increase the validity of VTE diagnosis, we conducted an analysis including only cases receiving anticoagulants at least 4 weeks after diagnosis of VTE. Fourth, we conducted an analysis stratified according to the age groups of 10-24 and 25-44 years because the incidence of VTE varies among women between the ages of 10 and 44 years. Fifth, because oral contraceptives are known to increase the risk of VTE, an analysis with adjustment for oral contraceptive use as a time-dependent variable was included.
A 2-sided 95% confidence interval that did not overlap 1.0 and P < .05 were considered statistically significant.
From a source population of 1 613 798 women, including 500 345 (31%) who received the quadrivalent HPV vaccine, we identified 4375 incident cases of VTE (Figure). Of these, 889 women (20%) were vaccinated during the study period. There was no association between the quadrivalent HPV vaccine and VTE during the 42 days following vaccination (crude incidence rates, 0.126 and 0.159 events/person-year for risk and control periods, respectively; incidence ratio, 0.77 [95% CI, 0.53-1.11]; Table).
Similarly, no association between the quadrivalent HPV vaccine and VTE was observed in subgroup analyses by age, including only anticoagulant-treated cases, only exposed cases, or when adjusting for oral contraceptive use (Table).
Our results, which were consistent after adjustment for oral contraceptive use and in girls and young women as well as mid-adult women, do not provide support for an increased risk of VTE following quadrivalent HPV vaccination. Two previous studies reported a potential association but one was based on reports from a passive surveillance system1 and the other included few vaccinated cases, many with known risk factors for VTE.2
Limitations of our study include the potential for residual confounding by unmeasured time-varying factors and invalid diagnoses of VTE; however, an analysis restricted to VTE cases with anticoagulant treatment after diagnosis was consistent with our other analyses.
Corresponding Author: Nikolai Madrid Scheller, MB, Department of Epidemiology Research, Statens Serum Institut, Artillerivej 5, 2300 Copenhagen S, Denmark (firstname.lastname@example.org).
Author Contributions: Mr Scheller had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: All authors.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Scheller.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Scheller, Svanström.
Study supervision: Pasternak, Svanström, Hviid.
Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.