To the Editor re: Reply to letters of Shapiro, Szarewski and Mansour:
Dr. Shapiro's suggestion that the rates of VTE in our studies are
egregiously low is unfounded. The incidence rates for users of
levonorgestrel pills in our study are consistent with those found in
earlier studies of idiopathic VTE in women of childbearing age (including
a large WHO hospital-based study which prospectively ascertained cases),
while the estimates for drospirenone are similar to those found for third-
generation oral contraceptives.1 2 The incidence rate quoted by Shapiro,
which came from a consensus statement arising from a workshop convened by
the manufacturers of the drospirenone pill,3 was based on a study which
included women with other proximate causes and other risk factors for VTE
where the rates are expected to be higher.4
While we were not able to validate VTE cases in the US database we
used very stringent case inclusion criteria, minimizing the chance of
including non-cases. Furthermore, any inclusion of non-cases would tend
to bias our result toward 1.0 (null finding), and would not explain the
increased risk of VTE in users of drospirenone OCs. In the GPRD study the
risk was highest among validated cases, suggesting that had we been able
to validate all cases the risk would have been higher than that reported.
We did not have difficulty finding controls. The smaller proportion of
controls using drospirenone pills, relative to cases, reflects the
association between drospirenone and VTE.
Confounding by family history of VTE is an unlikely explanation for
our results. Even if some preferential prescribing of drospirenone OCs
had occurred, the proportion of users is likely to be small because family
history is listed as a prescribing precaution. Hence family history would
need to carry an improbably high risk to explain a 2 to 3-fold excess risk
in users of drospirenone relative to levonorgestrel. BMI and smoking
status data were not available for the US study. However in the GPRD
study, adjustment for these factors had a negligible impact on the odds
ratios. The small number of women with missing BMI and smoking information
means that even extreme patterns of missing data would be insufficient to
explain our results. Nor would prescribing restrictions on drospirenone
OCs account for our results because conditions such as premenstrual
dysphoric disorder and severe acne are not risk factors for VTE
independent of BMI. Allowing for duration of use and calendar time did not
alter the odds ratios. We fail to see how the inclusion, rather than
exclusion, of non-idiopathic cases would help to 'account for unmeasured
Finally, the fundamental issue is the need to optimize the benefit to
risk profile of medications.5 Our studies contribute to emerging evidence
that drospirenone OCs carry a higher risk of VTE than levonorgestrel
pills, while systematic reviews have found no convincing evidence that
drospirenone confers benefits over and above those of other OCs.
1. World Health Organization Collaborative Study of Cardiovascular
Disease and Steroid Hormone Contraception. Effect of different
progestagens in low oestrogen oral contraceptives on venous thromboembolic
disease. Lancet. 1995;346:1582-8.
2. Jick H, Jick SS, Gurewich V, Wald Myers M, Vasilakis C. Risk of
idiopathic cardiovascular death and nonfatal venous thromboembolism in
women using oral contraceptives with differing progestagen components.
3. Reid RL, Westhoff C, Mansour D, de Vries C, Verhaeghe J, Boschitsch E,
et al. Oral contraceptives and venous thromboembolism consensus opinion
from an international workshop held in Berlin, Germany in December 2009.
Journal of Family Planning & Reproductive Health Care 2010;36:117-22.
4. Dinger JC, Heinemann LAJ, Kuhl-Habich D. The safety of a drospirenone-
containing oral contraceptive: final results from the European Active
Surveillance Study on oral contraceptives based on 142,475 women-years of
observation. Contraception 2007;75:344-54.
5. Dunn N. The risk of deep venous thrombosis with oral contraceptives
containing drospirenone. BMJ 2011;342:d2519 doi: 10.1136/bmj.d2519.
Competing interests: No competing interests