Paperonium

HIV and pregnancy · Lancet HIV · 2021

Summary: an HIV drug’s brain-and-spine defect risk is at most 0.10%. Its own numbers give 0.99%.

The summary: 0 cases in 384at most 0.10%The paper’s stated formula,same countsat most 0.99%Earlier studies of this drug5 rates, 0.30% to 1.04%Other HIV drugs, earlier study0.10%0.10%0.99%0%0.2%0.4%0.6%0.8%1%
The summary:0 cases in 384at most 0.10%The paper’sstated formula,same countsat most 0.99%Earlier studies ofthis drug5 rates,0.30% to 1.04%Other HIV drugs,earlier study0.10%0.10%0.99%0%0.5%1%1.5%

Defect rate, % of pregnancies: what 0 cases in 384 allows, and earlier studies

The paper says

In Brazil, none of 384 pregnancies in which the mother was taking the HIV drug dolutegravir at conception showed a brain or spine defect. Zero cases can’t prove zero risk, so the summary gives a range for the true rate: 0% to 0.10%.

Its own data show

The paper names the standard formula it used. For no cases in 384 it gives 3.84 ÷ (384 + 3.84) = 0.99%, nearly ten times the summary’s 0.10%. The same formula matches every other such range in the paper, like 0% to 0.36% for no cases in 1,068 pregnancies on another HIV drug.

A range’s top end is the highest rate the data can’t rule out, and zero cases prove little: even if 0.4% of pregnancies were affected, 384 would show none about 1 time in 5. As printed, the study’s main result seems to rule out all five rates earlier studies found for this drug; corrected, it includes four.

“The summary” means the paper’s abstract: the short summary printed at its top, which is the part most readers see.

Read the paper →Full technical write-up →

Dolutegravir and pregnancy outcomes among women on ART in Brazil: A national cohort study · Lancet HIV (2021) · doi:10.1016/S2352-3018(20)30268-X