Some Contraceptives Carry Higher Venous Thromboembolism Risk Than Others

Among hormonal contraceptive users, risk of venous thromboembolism (VTE) varied by method and dose, with combined pills and injections having the highest risk, a Danish nationwide cohort study found.

The standardized VTE rates per 10,000 person-years for the following hormonal contraceptives were:

• Combined pills: 10.0 (95% CI 9.2-10.9)
• Vaginal rings: 8.0 (95% CI 4.6-12.8)
• Patches: 8.1 (95% CI 1.5-25.1)
• Progestin-only pills: 3.6 (95% CI 2.8-4.7)
• Injections: 11.9 (95% CI 4.4-25.6)
• Intrauterine devices (IUDs): 2.1 (95% CI 1.7 -2.6)
• Implants: 3.4 (95% CI 1.7-6.3)
• No contraceptive use: 2.0 (95% CI 1.9-2.1)

When exclusively considering confirmed VTE cases, these associations persisted, reported Harman Gailan Hassan Yonis, MD, of Aalborg University Hospital in Denmark, and colleagues in a JAMA research letter.

“We found that combined hormonal contraceptives, particularly those containing third-generation progestins, were associated with the highest risk,” Yonis told MedPage Today. “In contrast, intrauterine devices did not show an increased risk compared to women who did not use hormonal contraception.”

When it came to VTE excess per 10,000 person-years, authors found it varied by combined pill formulation. For 20-μg estrogen pills with levonorgestrel it was 3.0 (95% CI -1.8 to 7.7) and 14.2 (95% CI 9.2-19.3) for combined pills containing third-generation progestins like desogestrel.

Compared with nonuse, VTE incidence rate ratios across contraceptive methods were:

• Combined pills: 4.6 (95% CI 4.2-5.0)
• Vaginal rings: 4.5 (95% CI 3.1-6.5)
• Patches: 5.0 (95% CI 2.1-12.0)
• Progestin-only pills: 1.8 (95% CI 1.4-2.3)
• Injections: 5.7 (95% CI 3.5-9.3)
• IUDs: 1.0 (95% CI 0.8-1.1)
• Implants: 2.4 (95% CI 1.4-4.0)

VTE, which is a significant cause of maternal mortality, typically forms in the legs and can travel to the lungs. Yonis noted that the study was motivated by a need for information about VTE risk associated with newer formulations of hormonal contraception.

The risk differences between methods are “crucial for personalized contraceptive counseling and helping women make informed choices based on their individual risk factors,” said Yonis.

Sarah Averbach, MD, of UC San Diego Health, who was not involved in the study, said that previous large observational studies have demonstrated that “the relative risk of VTE is elevated among combined hormonal contraceptive users” though the “absolute attributable risk is still low given how rare VTE is.” She noted that it was unfortunate that researchers were unable to adjust for body mass index, smoking, and family history all together.

Progestin-only contraceptives have not been thought to be linked with VTE, though more recent research has suggested that depot medroxyprogesterone acetate, a high dose and potent injectable progestin, may have increased VTE risk, she said.

“This study provides further evidence that injectable progestins, unlikely any other progestin-only contraceptives, may be associated with an elevated risk of VTE,” Averbach told MedPage Today. Like past data, this study can help guide physicians in counseling patients with conditions that put them at higher risk for VTE who may want to consider lower risk options.

Averbach also pointed out that VTE risk skyrockets in pregnancy, thus “by preventing pregnancy, contraception actually prevents VTE.”

This nationwide cohort study of nearly 1.4 million women between the ages of 15 and 49 used Danish national registers and followed patients from Jan. 1, 2011, or their 15th birthday, until July 1, 2021, emigration, death, or an exclusionary event. Hormonal contraception use was determined through prescription records. Women with a history of thrombosis, cancer, thrombophilia, liver or kidney disease, infertility treatment, hormone therapy, oophorectomy, hysterectomy, polycystic ovary syndrome, or endometriosis were excluded. Overall, 2,691 VTEs occurred.

Authors noted some limitations, including that residual confounding may have remained and external generalizability may have been limited because of the homogeneity of the population.

Please enable JavaScript to view the

comments powered by Disqus.