Subcutaneous closure after caesarean delivery: evidence update
G. Vilchez,E. Kontopoulos
DOI: https://doi.org/10.1111/1471-0528.14637
2017-06-01
Abstract:According to the National Center for Health Statistics of the Center for Disease Control, there were 1 284 551 caesarean deliveries in the USA in 2014. As such, it is the most common abdominal surgery performed in the USA, with a 5% incidence of wound complications. Among variables affecting wound complication rates is the suturing of the subcutaneous tissue, proponents of closing this layer argue for the associated decrease in dead space, which could potentially decrease the incidence of seromas, infection and wound breakdown. Detractors argue that both the additional operating time and the suture material could add to the risk of infection. A prior Cochrane meta-analysis [Anderson et al. Cochrane Database Syst Rev 2004; (4):CD004663], which included seven studies, concluded that closure of the subcutaneous layer was associated with an overall decrease in any type of wound complication [RR (95%) = 0.68 (0.52–0.88], wound seroma, and haematoma [RR(95% CI) = 0.52 (0.33–0.82)] but showed no difference in the rate of wound infection compared with non-closure of this layer. In this issue of BJOG, Pergialiotis et al. present a meta-analysis on the effect of subcutaneous closure in wound complication rates after caesarean delivery (Pergialiotis et al, BJOG 2017; 124:1018–25). The study includes six of seven articles identified in the prior Cochrane review and adds an article published in 1994 [Bohman et al. Infect Dis Obstet Gynecol 1994;1:259-64], which used unblinded randomisation. This article adds a total of 2079 new patients to 1617 patients included in the Anderson review (Table 2). The addition of these patients confirmed the previous findings that wound complication was significantly less in the closure group than in the non-closure group [OR (95% CI) = 0.66 (0.47–0.93)]. Of interest, although the risk reduction was lower (0.66 versus 0.68), the CI was wider (0.47–0.93 versus 0.52– 0.88) and, similarly, the risk of seroma or haematoma was significantly decreased [RR (95% CI) = 0.5 (0.33– 0.84], as in the Anderson review [(RR (95% CI) = 0.52 (0.33–0.82)]. Thus, the addition of newer series and the quasi-randomised trial of Bohman et al. did not change the original conclusions of Anderson’s review, while not narrowing the CI. Nonetheless, the effect is in the same direction in both, which provides support for the clinical recommendation. Is this the final word on the subject? Clearly not. Although the metaanalyses show decrease in seroma, haematoma or ‘any type of wound complication’, it is apparent from the literature that the definitions of such complications are neither objective nor standardised. The combined outcome of ‘any type of wound complication’ does not provide clear objective guidelines to the practitioner. Definitions for partial versus total wound breakdown, seroma and haematoma would be required to quantify the true effect of the surgical intervention. Furthermore, a multiple logistic regression model could exclude possible confounders (gestational age, indication for procedure, operating time, maternal co-morbidities, cervical dilation, suture material, surgical technique—continued versus interrupted stitches, maternal body mass index, additional surgical procedures—drainage, irrigation etc.) in order to reflect accurately the true contribution of the surgical step in reducing the incidence of wound complications. Given how often the procedure is performed, this metaanalysis should serve as a catalyst for the design of comprehensive randomised clinical trials addressing the true role of this surgical step in reducing the rate of wound complications after caesarean sections.