Objective
We sought to perform a metaanalysis to synthesize randomized clinical trials of cesarean skin closure by subcuticular absorbable suture vs metal staples for the outcomes of wound complications, pain perception, patient satisfaction, cosmesis, and operating time.
Study Design
A systematic search was performed using MEDLINE, Cochrane Databases, and ClinicalTrials.gov registries. We included randomized trials comparing absorbable suture vs metal staples for cesarean skin closure. Data were abstracted regarding wound complications, patient pain perception, patient satisfaction, cosmesis as assessed by the physician and patient, and operating time.
Results
Twelve randomized trials with data for the primary outcome on 3112 women were identified. Women whose incisions were closed with suture were significantly less likely to have wound complications than those closed with staples (risk ratio, 0.49; 95% confidence interval [CI], 0.28–0.87). This difference remained significant even when wound complications were stratified by obesity. The decrease in wound complications was largely due to the lower incidence of wound separations in those closed with suture (risk ratio, 0.29; 95% CI, 0.20–0.43), as there were no significant differences in infection, hematoma, seroma, or readmission. There were also no significant differences in pain perception, patient satisfaction, and cosmetic assessments between the groups. Operating time was approximately 7 minutes longer in those closed with suture (95% CI, 3.10–11.31).
Conclusion
For patients undergoing cesarean, closure of the transverse skin incision with suture significantly decreases wound morbidity, specifically wound separation, without significant differences in pain, patient satisfaction, or cosmesis. Suture placement does take 7 minutes longer than staples.
Cesarean is one of the most common surgeries performed worldwide and rates are increasing despite efforts to the contrary. Many of the surgical steps have been individually assessed previously, eg, prophylactic antibiotic administration, development of the bladder flap, techniques for expansion of the uterine incision and removal of the placenta, closure of the uterine and fascial incisions, closure of the subcutaneous space when ≥2 cm, and closure of the skin incision. Our previous systematic review as well as Dahlke et al concluded that the data remain conflicting with regards to whether it is better to close the cesarean skin incision with suture or with staples.
Though optimizing the individual steps of a cesarean is important with respect to providing the best possible care for patients, the skin incision is the visible reminder to a patient about her cesarean. Despite the potential complexity of the surgery, the occurrence of a wound complication may be the aspect that the patient most clearly recalls. In addition, wound complications may result in prolonged hospital stay, readmission, increased time away from work, decreased infant bonding time, and increased health care expenditures.
Choice of closure continues to vary among clinicians, most commonly between absorbable subcuticular suture and nonabsorbable metal staples. Additional evidence, including a trial recently published by 2 authors of this metaanalysis, has emerged regarding the optimal closure of the cesarean skin incision. As such our goal was to examine the pertinent randomized clinical trials (RCTs) to evaluate the incidence of wound complications, pain perception, patient satisfaction, cosmesis, and operating time when the cesarean skin incision was closed with suture vs with staples.
Materials and Methods
Sources
The principles embodied in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement were used in compiling this metaanalysis. MEDLINE via Ovid and PubMed searches were performed in July 2014 to encompass the past 50 years of trials; additionally the Cochrane Databases were searched and clinical trials were identified using Ovid and Clinical Key and by searching ClinicalTrials.gov. Ovid MEDLINE was searched using Medical Subject Headings (MeSH) without restrictions for text words or word variations for the following search terms: “cesarean,” “caesarean,” “suture,” “suture techniques,” “stitches,” “staple,” “surgical stapling,” “skin,” “closure,” “wound complications,” “wound infection,” “wound healing,” “wound closure,” “cicatrix,” “scar,” “randomized controlled trial,” and “randomized clinical trial.” “Humans” was set as the only limit. A second search was conducted in PubMed to identify nonindexed citations using the search terms “caesarean,” “cesarean,” “skin,” “wound complications,” “wound healing,” “wound infection,” “surgical wound infection,” “staple,” “suture,” “stitches,” and “trial.” There were no language restrictions applied. Searches were performed by a health sciences library specialist, a reference librarian, and the primary author (A.D.M.). This search was conducted as a brand new search and not part of the original Cochrane review or RCT by this author. As this is a metaanalysis, it was considered exempt from institutional review board approval.
Study selection
All identified abstracts were independently reviewed by the primary author (A.D.M.) and an additional author (V.B. or M.S.) and full articles were retrieved and reviewed for trials considered for inclusion. The primary author and additional author independently reviewed the manuscripts to assess for inclusion or exclusion criteria for this metaanalysis. We included only RCTs comparing subcuticular absorbable suture with nonabsorbable metal staples for cesarean skin closure. We chose to exclude RCTs that compared absorbable staples, nonabsorbable suture, or stapling devices ( Figure 1 ) . We excluded ongoing trials; studies that assessed surgical techniques for cesarean, but not the skin incision; and studies that did not provide applicable data for inclusion in the metaanalysis (eg, abstracts that did not report sample size).
Study outcomes
The primary outcome was wound complications. This was defined as a composite of wound infection, separation, hematoma, seroma, or readmission secondary to a wound concern. These were defined as per the individual trials. Secondary outcomes were patient pain perception at discharge, patient satisfaction at 2 months postoperatively, cosmesis as assessed by the physician and patient at 2 months postoperatively, and operative time. Patient pain perception and satisfaction were collated among studies if a 10-point scale was used, typically the visual analog scale in which higher scores represent higher patient satisfaction and more pain. Cosmesis was assessed if studies used the validated Physician Observer Scar Assessment Scale (OSAS) for assessment; this scale has both subjective (Physician Scar Assessment Scale [PSAS]) and objective (OSAS) components as scored by the patient and physician, respectively. Patient assessment of cosmesis (PSAS) was scored from 6-60 and physician assessment of cosmesis (OSAS) was scored from 5-50: lower scores are considered superior.
When applicable, attempts were made to contact authors to obtain more detail on outcomes not clearly reported in the manuscript. Additionally we analyzed the primary outcome stratified by body mass index (<30 kg/m 2 [nonobese] vs ≥30 kg/m 2 [obese]).
Risk of bias assessment
Each of the individual manuscripts was independently reviewed by the primary author and 1 additional author, both of whom assigned a low, high, or unclear risk of bias for all studies for the following 6 categories: (1) random sequence generation, (2) allocation concealment, (3) blinding of outcome assessment, (4) incomplete outcome data, (5) selective reporting, and (6) other bias (any bias that did not fit into categories 1-5). We investigated publication bias using a funnel plot that we assessed visually.
Statistical analysis
Data from included studies were entered into software (Review Manager [RevMan], version 5.2; The Cochrane Collaboration, Copenhagen, Denmark). Summary risk ratios (RRs) with 95% confidence intervals (CIs) were calculated for categorical variables ( Figures 2-5 and Appendix : Supplemental Figures 1-3 ); the weighted mean differences (MD) with 95% CI were computed for continuous variables ( Supplemental Figures 4-8 ). The χ 2 test for heterogeneity was used to assess for statistical heterogeneity along with the I 2 , which reflects the magnitude of heterogeneity. If it was reasonable to assume that the studies were similar with respect to the trials’ methods and sample and χ 2 test for heterogeneity P value was ≥ .10, we applied fixed effects analyses. However, if substantial heterogeneity was present (ie, χ 2 P value < .10), random effects analyses were applied. Tau 2 , χ 2 , and I 2 statistics are reported in all figures for reference, as descriptors of heterogeneity. The P value for overall effect is also presented and is considered significant if < .05. We did not perform any additional analyses (eg, sensitivity or subgroup analysis) secondary to the desire to include as many data as possible for the primary outcome.
Results
After removal of duplicates, ongoing trials, non-RCTs, and RCTs that did not compare only absorbable suture vs nonabsorbable staples for cesarean skin closure, 12 studies were included in our analysis ( Figure 1 ) ; data on 3112 women were available for the primary outcome. Of those 12 articles, 6 had not been included in previous systematic reviews. Two of the trials include cesareans performed via vertical skin incisions. One study limited inclusion to those undergoing emergency surgery. One study compared suture vs staples by using each method on one half of the incision and randomizing the laterality of each method. One trial included uterine and wound infections together in their reported infection rate. Two of the included studies randomized women to closure vs nonclosure of the subcutaneous tissue as well as closure of the skin incision with suture or staples. A number of authors provided additional information (including unpublished data), for the purpose of this metaanalysis.
Figures 2-5 and Supplemental Figures 1-3 and 7 present metaanalysis results based on RR. The overall metaanalysis RR of 0.49 delineates that women whose cesarean skin incision was closed with suture were 51% less likely to have a wound complication than those closed with staples (76/1571 [4.84%] vs 201/1541 [13.04%]; 95% CI, 0.28–0.87; P < .01) ( Figure 2 ). For Figures 4 and 5 and Supplemental Figures 1-3 and 7 , χ 2 P values indicate that there was not substantial heterogeneity among studies, so fixed effects metaanalyses were applied.
There were no significant differences in wound infections between those closed with suture and those closed with staples (46/1571 [2.92%] vs 98/1541 [6.36%]; RR, 0.80; 95% CI, 0.36–1.75; P = .58) ( Figure 3 ).
Women whose incisions were closed with subcuticular suture were 71% less likely (28/1110 [2.52%] vs 100/1064 [9.40%]; RR, 0.29; 95% CI, 0.20–0.43; P < .01) ( Figure 4 ) to have a wound separation as compared to women whose incisions were closed with staples. In the study in which half of the incision was closed with suture and the other half with staples, it was noted that 3 separations occurred in the middle of the cesarean scar; these were not included in this forest plot as it is unclear whether these complications should be attributed to suture or staple closure.
There were no significant differences in hematoma, (10/629 [1.59%] vs 7/648 [1.08%]; RR, 1.45; 95% CI, 0.59–3.54; P = .42) ( Supplemental Figure 1 ), seroma (6/497 [1.21%] vs 6/511 [1.17%]; RR, 1.02; 95% CI, 0.35–3.00; P = .97) ( Supplemental Figure 2 ), and readmission (4/589 [0.68%] vs 4/573 [0.70%]; RR, 0.97; 95% CI, 0.24–3.90; P = .97) ( Supplemental Figure 3 ) between those closed with suture and those closed with staples, respectively.
Supplemental Figures 4-8 present metaanalysis results based on MD. For Supplemental Figures 4-6 and 8 , random effects metaanalyses were applied as the χ 2 P values indicate that heterogeneity may be present among the included studies. For Supplemental Figure 7 , fixed effects metaanalyses were applied.
There were no significant differences between groups with respect to pain perception at discharge (MD, –0.13; 95% CI, –0.47 to 0.21; P = .46) ( Supplemental Figure 4 ), patient satisfaction at 2 months (MD, 0.27; 95% CI, –0.46 to 1.00; P = .47) ( Supplemental Figure 5 ), OSAS at 2 months (MD, 0.68; 95% CI, –2.59 to 3.95; P = .68) ( Supplemental Figure 6 ), or PSAS at 2 months (MD, –1.61; 95% CI, –3.23 to 0.01; P = .05) ( Supplemental Figure 7 ). Though the differences were not statistically significant, the MD for PSAS ( Supplemental Figure 7 ) scores was 1.61 lower for those closed with suture as compared to those closed with staples; for PSAS, lower scores are considered superior.
The only outcome that favored staples was operating time. All the included studies consistently demonstrated longer operative time for suture closure. The overall metaanalysis MD indicates that closure with suture was significantly longer by 7.20 minutes than closure with staples (95% CI, 3.10–11.31; P < .01) ( Supplemental Figure 8 ).
Authors of 5 studies provided their wound complication results stratified by obesity, ie, obese vs not obese. Obese patients were 49% less likely to have wound complications if their incisions were closed with suture as compared to staples (34/508 [6.69%] vs 67/522 [12.84%]; RR, 0.51; 95% CI, 0.34–0.75) ( Figure 5 ). Nonobese patients were 62% less likely to have a wound complication if their incisions were closed with suture as compared to staples (18/422 [4.27%] vs 48/427 [11.24%]; RR, 0.38; 95% CI, 0.23–0.64) ( Figure 5 ).
Table 1 depicts the characteristics of the included studies. Table 2 describes which of the specific wound complications were reported by the individual trials. Figure 6 shows the risk of bias assessments for the included studies. In Figure 6 , “low risk” (indicated with a green circle with a ‘+’) indicates: computer-generated or shuffle-cards (randomization), sequentially numbered opaque envelopes (allocation), outcome assessor blinded (blinding of outcome assessment), accounted for lost to follow-up and this was <30% of those randomized (incomplete outcome data), prespecified outcomes stated (selective outcome reporting), no other potential sources of bias identified (other bias). Unclear risk of bias is indicated by a yellow circle with a ‘?’; high risk of bias is indicated by a red circle with a ‘-’ sign. Regarding incomplete primary outcome data, “unclear” indicates that the number of dropouts were stated, but an intention-to-treat analysis not performed for 2 to 5 patients who did not receive assigned closure. Given that the number that did not receive assigned closure was so low, we decided that this would result in an unclear risk of bias, rather than a high risk of bias for not performing an intention-to-treat analysis. One study had >30% lost to follow-up for their 4-month primary outcome so was listed as high risk of attrition bias. Two studies had >30% lost to follow-up at 6 months and 1 year, but had low attrition at 4 and 6 week follow-up, respectively; so we determined that other bias was high, but that attrition bias was low. Other bias was listed as unclear for 1 study that performed an interim analysis secondary to patients with staples reporting severe discomfort with removal and as high for 1 study that stopped early due to the results of an unplanned interim analysis.
| Variable | Mackeen et al, 2014 | Sharma et al, 2014 | Figueroa et al, 2013 | Huppelschoten et al, 2013 | Aabakke et al, 2013 | de Graaf et al, 2012 | Chunder et al, 2012 | Cromi et al, 2010 | Basha et al, 2010 | Rousseau et al, 2009 | Gaertner et al, 2008 | Frishman et al, 1997 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| n | 746 | 112 | 398 | 145 | 59 | 133 | 734 | 180 | 430 | 101 | 100 | 66 |
| Primary outcome | Wound complications | Cosmesis and pain preference after emergency cesarean | Wound complications | Cosmesis | Patient preference | Cosmesis | Wound complications | Cosmesis | Wound complications and satisfaction | Postoperative pain | Cosmesis | Cosmesis |
| Nonabsorbable metal staples | 376 | 59 | 198 | 77 | 59 | 68 | 373 | 45 | 206 | 49 | 51 | 34 |
| Subcuticular absorbable suture | 370 | 53 | 200 | 68 | 59 | 64 | 361 | 90 | 224 | 52 | 49 | 32 |
| Suture material | 4–0 poliglecaprone/polyglactin | 3–0 poliglecaprone | 4–0 poliglecaprone | 3–0 poliglecaprone | 3–0 polyglactin | 3–0 poliglecaprone | Polyglycolic acid | 3–0 glyconate or polyglycolic acid | 4–0 poliglecaprone | 4–0 poliglecaprone | 3–0 polyglactin | 4–0 polyglycolic acid |
| Skin incision type | Low transverse | Pfannenstiel | Pfannenstiel and vertical | Pfannenstiel | Joel-Cohen | Pfannenstiel | Pfannenstiel | Pfannenstiel | Pfannenstiel and vertical | Pfannenstiel | Pfannenstiel | Pfannenstiel |
| Timing of antibiotic prophylaxis | Preoperatively | Preoperatively | After cord clamp | Not reported | After cord clamp | After cord clamp | After cord clamp | Not reported | Preoperatively | Not reported | After cord clamp | Not reported |
| Subcutaneous closure | If ≥2 cm | Not closed | If >2 cm | 50/50 | If >2.5 cm | 50/50 | Not stated | If ≥2 cm | If >2 cm | 100% | 50/50 | None |
| Staple removal day | 4–10 | 6 | 3–4 for Pfannenstiel; 7–10 for vertical | 4 | 3 | 7 | 10 | 7 | 3–4 | 3 | 6 | At discharge |
| Postoperative follow-up for wound complications | 4–8 wk | Unclear, but it appears this was performed at day 3 and at 6 wk | 4–6 wk | 6 wk | 3 mo | Unclear at what point these were assessed | Day 1, 3, and 10 | 2 mo | 2–4 wk | 6 wk | 4 mo | 6 wk |
| Postoperative pain assessment | VAS | VAS | 0–10 scale | NRS | 0–10 scale | VAS | Not assessed | Not assessed | Not assessed | 0–10 scale | Assessed based on author’s survey tool | 10-cm Pain scale |
| Incision assessment | POSAS | POSAS | Stony Brook SES | POSAS and NRS | VAS and Singer SES | POSAS | Not specifically stated | POSAS and VSS | Telephone interview regarding wound complications | Singer SES | Cosmesis based on author’s assessment scale | Subjective |
| Patient satisfaction assessment | Subjective | Not assessed | Subjective | Not assessed | Patient preference was assessed | Not assessed | Not assessed | VAS | Telephone interview | General satisfaction scale | Assessed based on author’s survey tool | Not assessed |
| Study site | Thomas Jefferson University, Lankenau Medical Center, Yale University (United States) | Prasad Government Medical College and Hospital, Kangra (India) | University of Alabama (United States) | Jeroen Bosch Hospital (The Netherlands) | University of Copenhagen (Denmark) | Spaarne Hospital and Academic Medical Center in Amsterdam (The Netherlands) | Addington Hospital, University of Kwa-Zulu-Natal (South Africa) | University of Insubria (Italy) | Lehigh Valley Health Network (United States) | Centre Hospitalier Universitaire de Québec (Canada) | University Hospital of Zurich (Switzerland) | Women and Infant’s Hospital (United States) |
| Study length | 2010 through 2012 | July through September 2013 | 2009 through 2010 | 2007 through 2009 | 2010 through 2011 | 2007 through 2008 | 2009 through 2010 | 2006 through 2008 | 2008 through 2009 | 2007 through 2008 | Started 2003 | 1995 through 1996 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
