The scoring system used to qualify the studies was useful to show that most of the investigations were not highly discrepant in terms of methodological quality. The analysis of the studies revealed that the main methodological flaw was the inaccurate exposure assessment, an issue already pointed out in other reviews focusing on caffeine consumption during human pregnancy.
Despite the biological plausibility that caffeine may increase the risk of fetal death, there is still no clear answer to the question: "Is there a true association between caffeine consumption during pregnancy and the risk of fetal death?". The small number of publications addressing the subject, methodological limitations in the reviewed studies, and overall risks only moderately elevated and marginally significant in most cases disallow stating with certainty that this association actually exists. In addition, due to the absence of studies showing negative results, the possibility of publication bias exists.
As mentioned by Signorello & McLaughlin 7 in their recent review of the effect of caffeine on abortion, an "optimal" observational investigation on the association between caffeine consumption and fetal death would also demand a huge cohort of women who agree to complete detailed questionnaires on consumption, to take samples for biochemical analysis of the different sources of caffeine consumed, and to allow drawing of repeated blood and urine samples to obtain accurate exposure information. This type of study would be difficult to plan and carry out, and also excessively expensive.
The present review highlighted the scarcity of publications on the subject and the need for well-designed future research to define the role of caffeine in fetal mortality. A well-designed investigation to contribute with convincing evidence on the association between caffeine consumption during pregnancy and fetal death would require a more in-depth approach to caffeine assessment, attempting to be sufficiently creative to suitably measure caffeine exposure without overly complicating the investigation.
The FDA has been advising women since 1980 to avoid caffeine or consume it only moderately during pregnancy. Although the available information linking caffeine to fetal death is incomplete, many health professional organizations 32 advise pregnant women to reduce caffeine intake. Although suspicion surpasses evidence at present, such caution appears to be prudent.
A. Matijasevich performed the relevant article searches, systemized the results, and wrote the draft and final version of the article. I. S. Santos oriented the methodological comments. F. C. Barros and I. S. Santos reviewed the draft and contributed to the review of the final version.