Comment on “Clinical manifestation, outcomes in pregnant women with COVID-19 and the possibility of vertical transmission: a systematic review of the current data”

To the Editor,

We read with great interest the study by Han and colleagues published in this Journal recently (volume 48, issue 9) [1]. In a word, the authors aimed to assess perinatal outcomes of COVID-19 infections during pregnancy and the possibility of vertical transmission. We want to congratulate the authors for establishing an informative systematic review and shed some light on this infection in a particular group. Certainly, the findings of Han et al. are an extremely commendable achievement and add to the literature on SARS-CoV-2 infections during pregnancy. While agreeing with the conclusions made by Han and colleagues, we believe that some concerns should be discussed in this important study.

First, the authors should present a full electronic search strategy for at least one database as online supplemental material, including any limits used, such that it could be repeated [2].

Second, the authors should assess the risk of bias in individual studies according to the PRISMA recommendations, which are mandatory for any systematic review [2].

Third, a total of 36 studies involving 1,103 pregnant women with COVID-19 meeting the study criteria were included in this systematic review. However, the authors did not exclude studies suspected of containing duplicate reporting. In other words, it is essential to note some pregnant women or neonates with COVID-19 may have been included in multiple publications, as admission dates overlap for reports from the same hospital.

The cases from Qiancheng et al. [3] should be considered a duplicate as a larger retrospective case series from Chen et al. [4] reported by the same hospital with the periods of recruitment overlapped. Of note, Qiancheng et al. clarified that 11 cases were from the same hospital included in Chen et al., which reported 118 pregnant women with COVID-19 in Wuhan. The two cases reported by Cooke et al. [5] should be considered a duplicate for similar reasons as a larger retrospective cohort study by Knight et al. [6], which reported 427 pregnant women with COVID-19 in the UK. Furthermore, we note the authors included a study reported by Yan et al. that was pooled from a national registry including 25 hospitals in China with study dates overlap (between 20 January 2020 and 24 March 24, 2020).

Isolated case reports and repeat case series from the same hospital should be excluded to avoid duplicate data from large retrospective studies as duplication can give a potentially biased picture. Studies suspected of including duplicate reporting can be identified based on hospital location, recruitment periods, maternal and neonatal characteristics. Although duplicate reporting has small numbers in this systematic review and a reanalysis are not likely to change the results, we humbly suggest that the authors extract the name of hospitals and recruitment periods. When a hospital had published their cases more than once, if the recruitment periods overlapped, only the most informative study with the bigger sample size was included to minimize the possibility of double counting [7].