Patient satisfaction and cosmetic outcome in a randomized study of cesarean skin closure.

OBJECTIVE
To evaluate patient satisfaction and patient and physician assessment of scar appearance after cesarean skin closure with suture versus staples.


METHODS
Women undergoing cesarean delivery (CD) at ≥23 weeks' gestation via low-transverse skin incisions at three hospitals in the CROSS Consortium were randomized to receive skin closure using subcuticular absorbable suture or nonabsorbable metal staples. The primary outcome of this substudy, patient satisfaction, was assessed by surveys at the postpartum visit using a 10-point Likert scale. Scar outcomes according to patients and trained observers were assessed at the primary research site using the Patient and Observer Scar Assessment Scale (POSAS). The POSAS is comprised of a patient-completed assessment including subjective data such as pain and itchiness, and an observer-completed assessment about cosmetic criteria.


RESULTS
Between June 2010 and August 2012, 746 women were randomized; 370 received suture and 376 received staples. Satisfaction data were available for 606 (81%). Complete patient scar assessment data were available for 577 (77%) and complete observer scar assessment data were available for 275 (57% of the 480 planned for evaluation at the primary research site). Demographic data for women in the two groups were similar. Satisfaction with the closure method was higher (superior) among women who received suture closure: median 10 (interquartile range 9, 10) versus 9 (interquartile ranges (IQR) 6, 10); p < .01. The suture group also had higher satisfaction with the scar's appearance at the postpartum visit: median nine (IQR 7, 10) versus 8 (IQR 6, 10); p = .02. Receiving one's preferred closure method was associated with higher patient satisfaction, and wound complications were associated with lower satisfaction. POSAS scores were superior (lower) in the suture group. Patient Scar Assessment Scale scores were median 15 (IQR 10, 25) for sutures versus 20 (IQR 11, 28) for staples; p < .01. Observer Scar Assessment Scale scores were median 12 (IQR 9, 15) for sutures versus 13 (IQR 9, 16) for staples; p = .01.


CONCLUSIONS
Satisfaction with the closure method, satisfaction with the scar's appearance, and patient and physician assessments of scar cosmesis were all superior in those closed with suture. These results further support the use of sutures for cesarean skin closure.


Introduction
Each year, almost one-third of women who deliver in the USA undergo cesarean delivery (CD) (32.0% in 2015). This makes CD as one of the most commonly performed surgeries in the country, and means approximately 1.2 million women received an abdominal scar from this surgery in 2017 [1]. In choosing a method of closure, surgeons must balance concerns for postoperative complications with patient satisfaction and cosmetic outcome [2]. Recent evidence suggests that suture is associated with lower risk of wound complications [3]. Yet, many obstetricians still use staples for closure [2].
No clear evidence exists to support either suture or staple closure with regard to patient satisfaction. Meta-analyses have found no significant difference between absorbable sutures and staples with regard to cosmesis and patient satisfaction, but previous reviews have found that many of the available studies are small in size and at risk of bias [3][4][5].
In an effort to generate more robust data for obstetricians choosing between suture and staples for the closure of low-transverse cesarean skin incisions, we designed a randomized clinical trial to assess the differences between these two closure methods with regard to wound complications, pain perception, patient satisfaction, and physician and patient assessment of wound cosmesis. Results regarding wound complications and pain perception are reported elsewhere [6,7]. Here, we report the results of our planned substudy of patient satisfaction, and of scar outcomes as assessed by patients and providers, after CD. Our objective was to determine whether CD closure with staples or with sutures was associated with higher patient satisfaction with the closure method, higher patient satisfaction with the scar appearance, or improved cosmetic outcome according to patients or physicians.

Study design and interventions
This prospective, multicenter, randomized clinical trial was conducted at Thomas Jefferson University Hospital in Philadelphia, PA (the primary research site), Lankenau Medical Center in Wynnewood, PA, and Yale-New Haven Hospital in New Haven, CT, USA. This group comprised the CROSS (Cesarean Randomized controlled trial Of Suture versus Staples) Consortium. The trial received approval from the Institutional Review Boards at each institution and was registered on ClinicalTrials.gov (Identifier: NCT01211600). No changes to the methods were made after trial commencement.
We included women undergoing primary or repeat CD at 23 weeks' gestational age via a low-transverse skin incision. Informed consent was obtained from all included patients by physicians during their prenatal visits or on presentation to labor and delivery. We excluded women with a predetermined set of comorbidities that would put them at an increased risk for wound complications: poorly controlled diabetes, current CD through a vertical skin incision, chronic steroid use, active lupus flare, human immunodeficiency virus or acquired immunodeficiency syndrome, current treatment for cancer, a history of radiation to the abdomen or pelvis, current treatment with immunosuppressant medications, or hypersensitivity to either closure material. At Yale, age less than 18 years was also an exclusion criterion following the institutional review board's request.
Before the surgery, each patient received intravenous antibiotics and abdominal preparation with either chlorhexidine or Betadine solution. Subcutaneous tissue was closed with plain or chromic gut interrupted suture if the depth of this layer was 2 cm. Patients were randomized to the closure of the skin incision with subcuticular absorbable suture or nonabsorbable metal staples which were removed between 4 and 10 days after the surgery. If randomized to suture closure, either 4-0 Vicryl (polyglactin 910) or 4-0 Monocryl (poliglecaprone 25) was used at the surgeon's discretion. Steri strips were placed at the time of suture placement or staple removal, and patients were advised to remove them in 7 days.
Randomization was in a computer-generated 1:1 ratio with random-sized blocks stratified within each center by current BMI (<30 versus 30) and primary versus repeat CD. A sequentially numbered, opaque envelope was opened by the circulating nurse at the time that fascial closure was completed for each enrolled patient. Further details regarding randomization and surgical techniques have been reported in the original manuscript describing the risk of wound complications [6].

Patient satisfaction
Our primary outcome was patient satisfaction with the closure method. This outcome was assessed at the postpartum visit using a paper survey 10-point Likert scale with the question, "How satisfied are you with the method used to close your skin incision?" Higher scores were superior. Patients also rated their satisfaction with the scar appearance on 10-point Likert scales by answering the question, "How satisfied are you with the appearance of your incision?" Patients who did not return for postpartum visits were contacted by telephone to complete the survey.

Scar assessments by patient and physician
Surgical scars at the postpartum visit were evaluated using the Patient and Observer Scar Assessment Scale (POSAS), a validated scar assessment tool previously used for the assessment of CD incisions [8][9][10][11][12][13]. The POSAS is composed of a patient-completed Patient Scar Assessment Scale (PSAS) and an objective component called the Observer Scar Assessment Scale (OSAS). The PSAS evaluates six items: pain, itchiness, color, stiffness, thickness, and irregularity. The OSAS evaluates five items: vascularity, pigmentation, thickness, relief, and pliability. Each item is scored on a 10-point scale comparing the patient's skin to normal skin, in which one represents normal skin. The score is summed separately for PSAS (range 6-60) and OSAS (range 5-50), and often totaled to give the POSAS score (range 11-110). Lower scores indicate closer resemblance to normal skin, and are superior.
Subjects at all sites were eligible for PSAS assessment. For logistic reasons and in order to reduce interobserver variability, we limited OSAS measurements to the primary research site, Thomas Jefferson University Hospital. Seven obstetric evaluators were trained (by JF) to complete OSAS assessments. We convened periodic retraining meetings in which cesarean scar photographs and their appropriate scoring were reviewed. All evaluators were blinded to the method of closure. If a patient did not return for her postpartum visit, multiple attempts were made to reschedule the appointment for OSAS completion.

Statistical analyses
The sample size was determined for the primary outcome of the CROSS consortiumwound complicationsto a goal of 850 women in total. As a result of decreasing enrollment and the relocation of the Consortium's principal investigator (DM), the study was stopped early after enrolling 746 patients. No data were analyzed until after this decision was made.
Statistical analysis was performed using SAS 9.4 (SAS Inst, Cary, NC). Analyses were performed according to the as-treated principle. Satisfaction scores and POSAS scar assessments were analyzed using medians with interquartile ranges (IQR). The Wilcoxon test was used to determine associations between satisfaction scores and wound complications, stratified by the randomization group. Various covariates were investigated for their impact on satisfaction using an univariate logistic regression, with satisfaction scores dichotomized as high (8) or low (<8). Since the odds ratios from these variables are of different types and thus not directly comparable, a c-statistic was calculated for the significant covariates to determine which predicted method and cosmetic satisfaction more strongly. POSAS scar assessments were analyzed using the Wilcoxon rank sum test. A p-value of <.05 was considered significant.
Since the interval between surgery and the postpartum visit varied, satisfaction scores and POSAS outcomes were not always assessed at the same time postoperatively. The Poisson regression was used to assess any impact of time-to-response on satisfaction scores or POSAS outcomes. Additionally, we evaluated whether time influenced responses in the suture group differently than in the staples group. No significant effect of time was found in either analysis, and thus we did not control for time-to-response in the analyses. A nonparametric Wilcoxon test was then used to compare these variables between the two closure groups. This process was repeated for a sample restricted to postpartum surveys completed at less than 100 days postpartum, and no appreciable difference was found between this time-restricted sample and the full sample of subjects. Thus, surveys completed at all time points were included.

Results
The enrollment was conducted from June 2010 to August 2012. At all study locations, 746 women were randomized to cesarean skin closure with suture or staples. Of the 370 subjects allocated to receive a suture, two received staples. One of these subjects had inadequate anesthesia so staples were chosen for speed of closure, and the other subject's reason was unknown. All subjects allocated to staples received staples (n ¼ 376).
Background data among all subjects were similar between groups (Supplementary Table 1). Satisfaction data were collected on 606 (81%) of the 746 total subjects, 299 in the suture group and 307 in the staples group (Supplementary Figure 1). There were no significant differences between the 606 included subjects and those lost to follow up with regard to all variables listed in Supplementary Table 1, with the exception of marital status: more subjects in the included group were married (43.6 versus 37.1%, p < .01). Satisfaction results were assessed at a median of 47 days postoperatively (76% of results were assessed within 8 weeks postoperatively and 91% within 12 postoperatively). The PSAS data were available for 577 subjects out of 746 (77%) and OSAS data were available on 275 (57% of the 480 planned for evaluation at the primary research site). The PSAS data were assessed at a median of 46 days postoperatively (80% of results were assessed within 8 weeks postoperatively and 93% within 12 weeks postoperatively). The OSAS data were assessed at a median of 46 days postoperatively (70% of results were assessed within 8 weeks postoperatively and 81% within 12 weeks postoperatively). Scar lengths were similar between the groups: mean 13.8 cm for suture closure and 13.9 cm for staple closure, t-test p-value .61. The median width was 3 mm for both groups.

Patient satisfaction
Satisfaction with the closure method was higher in women who received suture than in those who received staples: median score 10 (IQR 9, 10) versus nine (IQR 6, 10), p < .01 (Table 1). Similarly, satisfaction with the scar appearance was higher among patients closed with suture. These differences in both types of satisfaction were observed among both obese and nonobese patients.
We assessed whether wound complications confounded patient satisfaction, by comparing satisfaction scores among patients with and without a wound complication, stratified by skin closure method (Supplementary Table 2). Wound complications were associated with lower satisfaction with the closure method, in both suture (p < .01) and staple groups (p ¼ .03).
In the univariate analysis, three independent variables predicted high satisfaction (defined as score 8) (Supplementary Table 3). These predictors of high satisfaction were: absence of a wound complication, increasing PSAS score, and preference concordance (defined as receiving the patient's preferred method of closure). Women with a wound complication were only one-third as likely to have high satisfaction as women without a wound complication (OR 0.32 for satisfaction with closure method, and or 0.29 for satisfaction with scar appearance). c-Statistics were calculated to determine which of these significant odds ratios more strongly predicted satisfaction: preference concordance was the strongest predictor of satisfaction with the closure method (0.81). The two types of satisfaction were positively correlated, regardless of whether a wound complication was present (Spearman's rho 0.69, p < .01).

Scar assessments by patient and physician
Of the 746 patients at all study locations, PSAS cosmetic outcome data were available for 577 subjects (77%). Of the 480 patients randomized at the primary site, OSAS cosmetic outcome data were available for 275 (57%) (Supplementary Figure 1).
The OSAS scores favored suture closure: median 12 (IQR 9, 15) versus 13 (IQR 9, 16), p ¼ .01 (Table 2). When stratified by BMI, this difference persisted among nonobese patients only. Of the individual OSAS components, only pigmentation and thickness were different between groups. The Kruskal-Wallis test was performed to determine whether observers differed in their total score, and there was evidence of a difference between observers (p < .01). In order to compare the OSAS scores between suture and staple groups while adjusting for the observer, a stratified Wilcoxon test was performed which showed little evidence of a difference between the two groups (p ¼ .25).

Conclusions
We found that patients were more satisfied with suture closure than staple closure. Suture closure also resulted in superior patient-assessed cosmesis, but the cosmetic superiority of sutures according to trained observers became nonsignificant when accounting for interobserver variability. Higher satisfaction was found among those who received their preferred method of closure, even though most patients had no preoperative preference for a particular closure method.
Several prior studies, including smaller randomized trials, have not found differences in satisfaction between closure methods [4,[14][15][16]. One larger randomized trial conducted by Figueroa et al. [16], involving 350 patients who received either subcuticular suture or staples, also found no difference in satisfaction scores. Our observation that higher satisfaction occurs with suture closure differs from previous Table 1. Patient satisfaction according to method of cesarean skin closure.
studies. Given the overall high satisfaction rates in our study, however, the clinical significance of this new finding is less certain. Perhaps unsurprisingly, wound complications seem to be associated with patient satisfaction. The occurrence of a wound complication reduced the chance by two-thirds of having high satisfaction in our study. Similarly, another randomized trial with 435 subjects showed higher satisfaction with suture closure compared to staple closure, but that difference became nonsignificant when the investigators controlled for wound complications [17].
Like our trial, previous studies on cosmesis have shown little difference between suture and staples. An unblinded trial with subjective, nonvalidated scar assessments of 50 patients by Frishman et al. [18] reported sutures were cosmetically superior. Although a recent trial using POSAS found that staples were cosmetically superior [9], most studies assessing cosmetic outcome with validated scales have reported no differences between suture versus staples closure [10,[13][14][15][16]. A meta-analysis found that sutures led to a better cosmetic outcome at 6-12 months, but not at 6-8 weeks, postoperatively [19]. In this regard, our study is consistent with prior work. Regardless, cosmetic outcome is likely to be only one of several factors that contributes to satisfaction with the closure method, so the clinical significance of a difference in cosmesis is uncertain.
The strengths of this study include the randomized, large sample size, and the use of a previously validated instrument to assess cosmetic outcomes. By limiting our exclusion criteria only to those attributes that are known to significantly impair wound healing, the generalizability of our results is strengthened in comparison to other, more restrictive studies [14].
Limitations of this study include the low follow-up rate. We believe that the stress and time pressures of caring for a new infant and potentially returning to work were likely responsible, but it is possible that follow-up rates were differentially affected by satisfaction with cesarean closure method, potentially confounding the primary outcome. Additionally, despite our efforts, we did not avoid interobserver variability in our OSAS measurements. This variability seems to explain the difference in OSAS scores between the suture and staple groups.
Cosmetic outcome is certainly important to patients and providers alike, but it is just one of many considerations that affect patient satisfaction. When randomized to suture or staple closure, patients were more satisfied with sutures, showing evidence that sutures directly improve patient satisfaction. Since we observed that a wound complication is a predictor of poor satisfaction, and we know from prior studies that sutures have fewer complications, our results also imply that sutures improve patient satisfaction indirectly by reducing wound complications [3,6]. For all of these reasons, subcuticular suture closure should be the method of choice for skin closure of low-transverse cesarean skin incisions.