Mortality after sternal reconstruction with pectoralis major flap vs omental flap for postsurgical mediastinitis: A systematic review and meta-analysis

Abstract

Background:

Deep sternal wound infections are rare but severe complications after median sternotomy and can be managed with sternal reconstruction. The use of pectoralis major flap (PMF) has traditionally been the first-line approach for flap reconstruction but the advantage in patients’ survival when compared to the omental flap (OF) transposition is still not clear. We performed a study-level meta-analysis evaluating the association of the type of flap on postoperative outcomes.

Methods:

A systematic saarch of the literature was performed to identify all studies comparing the postoperative outcomes of PMF versus OF for sternal reconstruction. The primary outcome was postoperative mortality. Secondary outcomes were the occurrence of sepsis, pneumonia, operative time, and length of stay. Binary outcomes were pooled using an inverse variance method and reported as odds ratio (OR) with corresponding 95% confidence interval (CI). Continuous outcomes were pooled using an inverse variance method and reported as standardized mean difference (SMD) with corresponding 95% CI.

Results:

Four studies with a total of 528 patients were included in the analysis. Overall, 443 patients had PMF reconstruction, and 85 patients had OF reconstruction. Baseline characteristics were similar in both groups. There were no statistically significant differences between PMF patients and OF patients in mortality (OR 0.6 [0.16; 2.17]; p = .09), sepsis (OR 1.1 [0.49; 2.47]; p = .43), pneumonia (OR 0.72 [0.18; 2.8]; p = .11), length of stay (SMD −0.59 [−2.03; 0.85]; p < .01), and operative time (SMD 0.08 [−1.21; 1.57]; p < .01).

Conclusion:

Our analysis found no association between the type of flap and postoperative mortality, the incidence of pneumonia, sepsis, operation time, and length of stay.

1 |. INTRODUCTION

Deep sternal wound infections (DSWIs) and mediastinitis are rare but severe complications after median sternotomy, with an incidence ranging from 0.2% to 8.0% of patients undergoing cardiac surgery procedures.^1–3 Despite this relative low incidence, DSWI and mediastinitis are challenging clinical problems with high mortality rate ranging from 7.3% to 21.6% as well as high morbidity and cost.^4,5 In a meta-analysis performed by our institution, DSWIs were associated with higher in-hospital mortality, long-term mortality, and major adverse cardiovascular events.⁶

Optimal management for DSWI is not clear. Surgical management includes closed irrigation and radical debridement, negative pressure wound dressing, pedicled muscle flaps (pectoralis muscle, latissimus dorsi, rectus abdominus), and omental flaps (OFs).⁷ The pectoralis major flap (PMF) represented the first-line approach in sternal reconstruction after mediastinitis but the advantage in patients’ survival when compared to the OF transposition is still not clear.⁸

We performed a systematic review and meta-analysis of the studies comparing postoperative outcomes after sternal reconstruction with PMF versus OF.

2 |. MATERIALS AND METHODS

This review follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines.⁹ Ethical approval was not required as no individual patients were involved in the study. A medical librarian (MD) performed a comprehensive search to identify studies comparing the postoperative outcomes of PMF versus OF for sternal reconstruction after postsurgical mediastinitis. On May 3, 2020, a systematic search of the literature was performed by querying Ovid MEDLINE, Ovid EMBASE, Cochrane Library databases, and CINAHL databases. The search included the following terms: pectoralis muscle flap, OF, DSWI, and mediastinitis. The complete search strategy for the Ovid MEDLINE database is provided in Supporting Information: Table S1.

After the exclusion of duplicate studies, available studies’ titles and abstracts were reviewed by two independent authors (G. C., T. A. Z). Any discrepancies were resolved by the senior authors (M. G.). Our inclusion criteria were studies written in English and comparing the clinical outcomes of patients undergoing sternal reconstruction with PMF versus OF for DSWI or mediastinitis. Case reports, conference presentations, editorials, animal studies, expert opinions, and studies not defining or reporting the outcomes of interest were excluded. Thirty-eight studies were selected for full-text review as the second round of screening. The reference list of the selected articles was also reviewed for relevant studies not captured by the original search. The quality of the included studies was assessed using the Newcastle–Ottawa Scale (NOS) for observational studies.

Two investigators (G. C., T. A. Z) independently performed data extraction, and the accuracy was verified by the senior author (M. G.). The variables included were study characteristics (publication year, institution, country, study period, study design, sample size), patient demographics (age, gender, smoking, diabetes, obesity, pulmonary disease, heart failure, peripheral vascular disease), reconstruction procedures (PMF, OF), postoperative outcomes (overall mortality, persistent sepsis, pneumonia, length of stay, operative time). The primary outcome was overall mortality.

2.1 |. Statistical analysis

Binary outcomes (mortality, sepsis, pneumonia) were extracted as the number of events in each group and were pooled using an inverse variance method and reported as odds ratio (OR) with corresponding 95% confidence interval (CI). Continuity correction was applied for cell with zero event.

Continuous outcomes (operative time, length of stay) were extracted as the mean and standard deviation for each group and were pooled using an inverse variance method and reported as standardized mean difference (SMD) with corresponding 95% CI.

For both types of outcomes common (also known as fixed) and random effects estimates were reported.

Statistical heterogeneity and consistency were assessed with I², which described the percentage of the variability in effect estimates that is due to heterogeneity rather than sampling error (chance). Low, moderate, and high heterogeneity were defined as I² less than 25%, 25%–50%, and more than 50%, respectively. The presence of a small-study effect was assessed visually by inspection of the funnel plot and, when possible, quantitatively by means of Egger’s test.

A univariate meta-regression was performed by regressing the meta-analytic estimates against the publication year, age, proportion of women, and diabetes.

In all analyses, the PMF was the reference group.

Statistical analyses were performed in R (version 4.0.3; R Project; R Foundation for Statistical Computing) using the packages: meta, dmetar. A p value < .05 was assumed as statistically significant.

3 |. RESULTS

Overall, 936 studies were identified in the primary search and 4 were included in the final analysis.^10–13 The PRISMA flow diagram outlining the study selection process and the quality of included studies according to the NOS are provided in Supporting Information: Figure S1 and Table S2, respectively. All studies were of high quality (NOS score 8–9). They were all observational studies published between 1998 and 2021 in four different countries (Mexico, United States, Germany, and Canada). Study characteristics are reported in Table 1. All studies included comparative retrospective cohorts, except for Lopez-Monjardin et al.¹⁰ who compared retrospective PMF to prospectively collected data for OF.

TABLE 1.

Characteristics of the included studies

Author	Publication year	Country	Study period	Study design	Sample size	PMF	OF
Lopez-Monjardin	1998	Mexico	January 1987 to December 1996	Retrospective versus prospective	33	21	12
Milano	1999	United States	1988–1998	Retrospective	59	38	21
Tewarie	2019	Germany	July 2010 to July 2016	Retrospective	39	20	19
Marzouk	2021	Canada	January 2000 to December 2020	Retrospective	397	364	33

Abbreviations: OF, omental flap; PMF, pectoralis major flap.

A total of 528 patients were included in the analysis. The number of patients in each study ranged from 33 to 397 with a median sample size of 49 patients (interquartile range: 36–228). Overall, 443 patients had PMF reconstruction, and 85 patients had OF reconstruction.

The mean age range was 62.0–70.1 in the PMF group and 60.0–70.3 in the OF group, 141 (31.8%) female patients were included in the PMF group and 31 (36.5%) in the OF group, males were 302 (68.2%) in the PMF group and 54 (63.5%) in the OF group. The prevalence of diabetes ranged from 39.0% to 57.0% in patients treated with PMF and from 33.0% to 51.5% in patients treated with OF. The prevalence of obesity ranged from 14.3% to 52.2% in patients treated with PMF and from 25.0% to 66.7% in patients treated with OF (Supporting Information: Table S3). All the included studies mentioned the heart surgery procedure patients underwent before developing mediastinitis, with the only exception for Milano et al.¹¹ About 86.8% of the cardiac procedures were CABG, 8.5% were isolated valve procedures, and 4.7% were other cardiac procedures.

OF was harvested using the opening surgical technique in all studies except for Marzouk et al.¹³ in which OF was laparoscopically harvested. Different techniques were used for OF coverage and PMF harvesting. Lopez-Monjardin et al.¹⁰ described that 5/12 OFs were covered with full-thickness meshed skin, 5/12 OFs were covered with pectoralis major musculocutaneous advancement flap, and 2/12 OFs were covered with skin flaps; PMFs were transposed on the dominant thoracoacromial neurovascular pedicle in all their cases. Milano et al.¹¹ used free OFs in all the cases and PMFs were either unilateral or bilateral. Tewarie et al.¹² covered all the OFs with pectoral advancement flaps and used bilateral PMF in all cases. In Marzouk et al.¹³ OFs were covered with PMF or split-thickness skin grafts with or without direct skin closure and PMFs were dissected without deltopectoral separation.

There were no statistically significant difference between PMF patients and OF patients in mortality (OR 0.6 [0.16; 2.17]; p = .09) (Figure 1), sepsis (OR 1.1 [0.49; 2.47]; p = .43) (Figure 2), pneumonia (OR 0.72 [0.18; 2.8]; p = .11) (Figure 3), length of stay (SMD −0.59 [−2.03; 0.85]; p < .01) (Figure 4), and operative time (SMD 0.08 [−1.21; 1.57]; p < .01) (Figure 5) (Table 2).

FIGURE 1.

Forest plot of studies comparing the effect of pectoral major flap and omental flap on the primary outcome of mortality. CI, confidence interval; OR, odds ratio.

FIGURE 2.

Forest plot of studies comparing the effect of pectoral major flap and omental flap on sepsis. CI, confidence interval; OR, odds ratio.

FIGURE 3.

Forest plot of studies comparing the effect of pectoral major flap and omental flap on pneumonia. CI, confidence interval; OR, odds ratio.

FIGURE 4.

Forest plot of studies comparing the effect of pectoral major flap and omental flap on length of stay. CI, confidence interval; SD, standardized difference.

FIGURE 5.

Forest plot of studies comparing the effect of pectoral major flap and omental flap on operative time. CI, confidence interval; SD, standardized difference.

TABLE 2.

Summary of the main outcomes

Outcome	No. of studies	OR/SMD (95%), p value	Heterogeneity (I2, p)
Mortality	4	0.6 [0.16; 2.17], p = .09	54%; p = .09
Sepsis	4	1.1 [0.49; 2.47], p = .43	0%; p = .43
Pneumonia	2	0.72 [0.18; 2.8], p = .11	61%; p = .11
Length of stay	3	−0.59 [−2.03; 0.85], p < .01	94%; p < .01
Operative time	2	0.08 [−1.21; 1.57], p <.01	92%; p < .01

Abbreviations: OR, odds ratio; SMD, standardized mean difference.

Meta-regression did not identify any association between the tested variables and the OR for the primary outcome (Table 3).

TABLE 3.

Results of the meta-regression for the primary outcome

Variable	Estimate	95% CI	p Value
Age	−0.2235	−0.6747 to 0.2277	.33
Female	−4.7908	−24.0320 to 14.4504	.63
Diabetes	−9.7495	−32.8078 to 13.3088	.41
Study year	−0.0938	−0.1675 to −0.0201	.01

Abbreviation: CI, confidence interval.

4 |. DISCUSSION

Herein we have reported the results of a meta-analysis of fur studies including 528 patients undergoing sternal reconstruction with PMF versus OF after DSWI or mediastinitis. Our primary finding was that there is no significant difference in postoperative mortality between patients treated with PMF versus patients treated with OF.

Although DSWI and mediastinitis represent serious complications occurring after cardiac surgery, the best surgical treatment still needs to be identified. PMF has been considered the gold standard in sternal reconstruction while OF has been relegated as a secondary option, particularly for patients with lower third sternum involvement or patients with extensive loss of chest wall soft tissue and insufficient skin for closure.¹⁴ Interestingly, OF contains immunological and proangiogenic factors that justified the good postoperative outcomes in reconstructive general and plastic surgery, but its application to cardiothoracic surgery appeared to be limited.¹⁵ On the other hand, a prior analysis conducted by Ghazi et al. showed that OF reconstruction is related to an increased mortality, especially when used as a salvage procedure. However, this association may not be directly related to the OF but rather to the complexity of the clinical situation leading up to its use.¹⁶ Indeed, in the present analysis, OF was transposed as a savage procedure, when the clinical status of the patient was already deteriorated.

A prior review and meta-analysis of six studies performed by Wingerden et al.¹⁷ in 2011, showed that the difference in postoperative mortality between OF and PMF was not statistically significant (relative risk 1.29, 95% CI 0.58–2.88). The number of studies included in this analysis was larger than ours as the inclusion criteria were different. The authors decided to include all the papers related to the treatment of DSWI that reported data about mortality for different types of flap and then, they performed the meta-analysis comparing PMF versus OF. Instead, we decided to include only the studies which primary purpose was to compare clinical outcomes between the two types of flaps. Despite the difference in selection criteria and timeline, both meta-analyses showed that the primary outcome of mortality was not associated to the flap choice.

This study must be interpreted in the setting of the following limitations: comparisons between the two flaps may be underpowered to detect meaningful clinical differences because of the limited number of available studies on this topic. Also, there is heterogeneity in the surgical technique, as the OF was harvested endoscopically in 38.8% of the patients (all of them included in the analysis conducted by Marzouk et al.), while 61.2% were treated with open harvesting technique.

5 |. CONCLUSIONS

In conclusion, our analysis demonstrated no association between the type of flap and postoperative mortality, the incidence of pneumonia, sepsis, operation time, and length of stay. Therefore, well-designed and adequately powered studies are required to determine the impact of different reconstruction strategies on short- and long-term outcomes in patients with DSWI and mediastinitis.