Advertisement

Postsurgical barrier strategies to avoid the recurrence of intrauterine adhesion formation after hysteroscopic adhesiolysis: a comment

Published:February 16, 2022DOI:https://doi.org/10.1016/j.ajog.2022.02.016
      To the Editors:
      In this network meta-analysis (NMA), Vitale et al
      • Vitale S.G.
      • Riemma G.
      • Carugno J.
      • et al.
      Postsurgical barrier strategies to avoid the recurrence of intrauterine adhesion formation after hysteroscopic adhesiolysis: a network meta-analysis of randomized controlled trials.
      compared antiadhesive strategies for women undergoing hysteroscopic adhesiolysis followed by mechanical prevention of intrauterine adhesions. The authors used the surface under the cumulative ranking (SUCRA) method to rank the interventions. Based on the SUCRA scores, the study concluded a copper intrauterine device together with an intrauterine balloon (46.4%), hyaluronic acid gel (79.8%), hyaluronic acid gel plus intrauterine device (49.9%), and dried amnion graft (53.8%) ranked the highest for preventing adhesions recurrence, improving fecundity, postsurgical adhesion severity, and menstrual pattern improvement, respectively.
      • Vitale S.G.
      • Riemma G.
      • Carugno J.
      • et al.
      Postsurgical barrier strategies to avoid the recurrence of intrauterine adhesion formation after hysteroscopic adhesiolysis: a network meta-analysis of randomized controlled trials.
      However, when considering the limitation of SUCRA, this conclusion might be inappropriate.
      It could be very misleading to conclude the effectiveness or harmfulness of treatments by only relying on the SUCRA score but ignoring the certainty of the evidence, as the SUCRA approach only focuses on point estimates of effect. This approach ignores the possibility that chance can explain the differences between SUCRA scores (precision of estimates), the magnitude of the absolute difference between rankings, and, most importantly, the certainty of the evidence.
      • Rücker G.
      • Schwarzer G.
      Ranking treatments in frequentist network meta-analysis works without resampling methods.
      For example, in 1 NMA of pain treatments for non–low back musculoskeletal injuries, Busse et al
      • Busse J.W.
      • Sadeghirad B.
      • Oparin Y.
      • et al.
      Management of acute pain from non-low back, musculoskeletal injuries : a systematic review and network meta-analysis of randomized trials.
      reported both the SUCRA and certainty of the evidence. Fentanyl ranked the highest effect for pain relief (<2 hours after treatment) but proved to be low or very low certainty evidence. This means that it is very unsure that the high effect seen for fentanyl is true.
      Unfortunately, the authors did not consider the certainty of the evidence for network estimates. Considering the wide confidence intervals for the network estimates in adhesions recurrence (the primary outcome), the certainty of evidence might be low. If the authors used the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) contextualized framework to rank the treatments,
      • Brignardello-Petersen R.
      • Izcovich A.
      • Rochwerg B.
      • et al.
      GRADE approach to drawing conclusions from a network meta-analysis using a partially contextualised framework.
      which avoids the limitations of SUCRA, they might get different but more reliable rankings.
      In conclusion, when ranking the effectiveness and/or harm of treatments in an NMA, we suggest that the authors should not only rely on the SUCRA scores of treatments but also consider the certainty of evidence to avoid making misleading conclusions.

      References

        • Vitale S.G.
        • Riemma G.
        • Carugno J.
        • et al.
        Postsurgical barrier strategies to avoid the recurrence of intrauterine adhesion formation after hysteroscopic adhesiolysis: a network meta-analysis of randomized controlled trials.
        Am J Obstet Gynecol. 2022; 226: 487-498.e8
        • Rücker G.
        • Schwarzer G.
        Ranking treatments in frequentist network meta-analysis works without resampling methods.
        BMC Med Res Methodol. 2015; 15: 58
        • Busse J.W.
        • Sadeghirad B.
        • Oparin Y.
        • et al.
        Management of acute pain from non-low back, musculoskeletal injuries : a systematic review and network meta-analysis of randomized trials.
        Ann Intern Med. 2020; 173: 730-738
        • Brignardello-Petersen R.
        • Izcovich A.
        • Rochwerg B.
        • et al.
        GRADE approach to drawing conclusions from a network meta-analysis using a partially contextualised framework.
        BMJ. 2020; 371: m3907

      Linked Article