Why does pregnancy change quit-medication decisions?

01

The pregnancy evidence has its own population

The 2020 Cochrane review searched to May 2019 and included 11 randomised trials with 2,412 pregnant participants. Nine studied nicotine replacement therapy (NRT), and two studied bupropion, alongside comparable behavioural support in the control groups. Those counts describe the evidence available at that search date, not all studies that exist today. They are also not evidence that two medicines are equally well studied.

[1]

02

Stopping smoking and a birth outcome are not one measure

The primary effectiveness outcome was abstinence in later pregnancy. The review's authors rated the evidence as low certainty and found that the NRT effect was not clear when potentially biased non-placebo trials were removed. Low adherence also limited interpretation. This is neither a guarantee of benefit nor proof that an intervention cannot help.

A finding of no clear difference in a measured birth outcome is not proof of no possible harm. An uncommon outcome may be estimated imprecisely, and measuring something at birth does not cover every later developmental outcome. These are limitations of the question and data, not a safety ranking between products.

[1]

03

A newer, larger study can still answer a narrower question

A 2025 four-country linked-record cohort examined major congenital malformations following first-trimester exposure to cessation medicines. Its exposure definition used medicine dispensing records, not direct observation of every dose taken. The analysis excluded stillbirths and did not study long-term functional or neurodevelopmental outcomes; the authors also described missing information about pregnancy terminations and smoking throughout the trimester. New data therefore add evidence for a defined question, not a certificate covering every pregnancy outcome.

[3]

04

Why guidance can change without certainty becoming complete

NICE's UK rationale describes evidence together with committee judgment and gaps needing further research. Its 2023 surveillance considered a newer trial but retained important uncertainty, particularly about longer-term outcomes. A recommendation may address how a service should work with incomplete evidence; it is not a statement that every relevant outcome has been settled. Read the jurisdiction, version and rationale alongside the recommendation, and let the local maternity team interpret actual product information.

[2]