Design: Systematic review and meta-analysis of 18 RCTs and 10 other studies
Population: Users of combined oral contraceptives, continuous or extended-cycle versus 21/7 or 24/4
Primary outcome: Unintended pregnancy over one year
Effect: OR 0.79 (95% CI 0.41 to 1.52); Pearl Index difference −0.20 (−1.31 to 0.92)

A systematic review in BMJ Sexual & Reproductive Health pooled 18 randomised trials and 10 other studies comparing continuous or extended-cycle combined oral contraceptives with standard 21/7 or 24/4 regimens.

Contraceptive effectiveness was similar: the odds of an unintended pregnancy over a year did not differ meaningfully, and Pearl Index values were close. Bleeding was the real difference. Some trials found fewer total bleeding and spotting days with extended use, but unscheduled bleeding was more common. The pooled odds of a venous thromboembolism were imprecise, with an estimated absolute difference of under one extra case per 1000 users. Satisfaction was high with both approaches.

For counselling, this supports offering a shortened or skipped hormone-free interval to women who want fewer withdrawal bleeds, painful periods or menstrual migraine, as long as they expect some breakthrough spotting, particularly early on. The authors note higher discontinuation and missing data, which weakens the acceptability findings more than the effectiveness ones.

Offer continuous or extended use of a combined pill to women who want fewer withdrawal bleeds; effectiveness is similar to 21/7 or 24/4.
Warn that unscheduled spotting is more common with extended use, especially in the first months, so it is not mistaken for failure.
Apply the usual eligibility checks for combined hormonal contraception; extending the cycle does not change who can take it.
Consider extended use for women whose symptoms cluster in the pill-free week, such as headaches or dysmenorrhoea.

Why it matters

The monthly withdrawal bleed is a convention rather than a safety requirement, and this review supports saying so to patients.

The statistics, in plain English

An odds ratio of 0.79 with a confidence interval from 0.41 to 1.52 crosses 1.0, so the review cannot say extended use is better or worse at preventing pregnancy — only that no difference was found. The VTE estimate (OR 1.05, 0.16 to 6.79) is so wide it is uninformative about relative risk; the absolute estimate of about 0.7 extra cases per 1000 users is the more useful figure.