RETRACTED: A meta-analysis examined the effect of topical vancomycin application in decreasing sternal wound infections post cardiac surgery
Abstract
To assess the impact of topical vancomycin (TV) application in decreasing sternal wound infections (SWIs) post cardiac surgery (CS), we lead a meta-analysis. Twenty-three thousand seven hundred and forty five participants had CS at the outset of the investigations, according to a thorough evaluation of the literature done up to November 2022; 8730 of them used TV, while 15 015 were controls. To assess the effectiveness of TV application in lowering SWIs following CS, odds ratios (OR) with 95% confidence intervals (CIs) were computed with dichotomous technique with a fixed- or random-effect model. The TV had significantly lower SWIs post CS (OR, 0.34; 95% CI, 0.20–0.57; P < .001), and deep SWIs post CS (OR, 0.26; 95% CI, 0.11–0.65; P = .004) compared with control as shown in Figures 2 and 3. Yet, there was no significant difference found amongst TV and control in superficial SWIs post CS (OR, 0.30; 95% CI, 0.07–1.30; P = .011). The TV had significantly lower SWIs, and deep SWIs post CS, and no significant difference was found in superficial SWIs post CS compared with control. The low number of included studies in this meta-analysis for superficial SWIs calls for precaution when analysing the outcomes.
1 INTRODUCTION
Despite being relatively uncommon, deep sternal wound infections (SWIs) are linked to higher rates of illness and death and lower long-term survival after heart surgery procedures.1 Although the need for strict intraoperative glycaemic control and intravenous antibiotic prophylaxis is without dispute, whether extra topically applied antibiotics could additionally decline the occurrence of deep SWI is still hotly contested. According to a latest expert consensus study on the prevention and management of SWIs, topical antibiotics must be administered to the cut margins of the sternum on opening and prior to closing all cardiac surgical operations requiring a sternotomy.2 This is the first time a Class I Endorsement with a Level of Evidence of B was assigned. The aforementioned assertion was principally motivated by several papers describing outstanding outcomes obtained with implantable gentamicin-collagen sponges3, 4 and topically applied topical vancomycin (TV) to the sternal margins.5, 6 In a multivariate analysis that took into account factors like body mass index >30 kg/m2, New York Heart Association Class IV, or Society of Thoracic Surgeons deep SWI Risk Index, a recent study showed that TV neither decreased the existence of deep SWI nor was its usage a negative predictor of deep SWI.7 We conducted a meta-analysis of studies evaluating the TV effect in lowering the occurrence of SWIs following cardiac surgery (CS) because the number of subjects included gave this study the best potential to inform decision-making.
2 METHOD
2.1 Research strategy
The current meta-analysis concentrated on the epidemiological statement and included studies that followed a predetermined study strategy.8
2.2 Sources of information
- The research was either retrospective, prospective, observational, or a RCT.
- The intended subjects consisted of subjects who had undergone CS.
- The intervention program's main component was TV.
- TV was compared with control.
- Those that failed to demonstrate how TV application reduced SWIs after CS.
- Research involving subjects who received care other than TV.
- The studies did not put many emphasis on the effects of comparison outcomes.
2.3 Identification
A protocol of search plans was developed in accordance with the PICOS principle9 and was defined as follows: P (population): those who have CS; I (exposure/intervention): SWIs were the outcome of the study's, which was conducted with no restrictions.10
We first prepared a full examination of the databases OVID, Embase, Cochrane Library, PubMed, and Google Scholar up to November 2022 for topical vancomycin, sternal wound infection, superficial sternal wound infection, and CS and deep sternal wound infection, as exposed in Table 1. All the encompassed papers were joint into an EndNote file, duplications were detached, the title and abstracts were reviewed, and any necessary changes were made to reject studies that failed to show a linking between TV and control in CS subjects.
| Database | Search strategy |
|---|---|
| Pubmed | #1 ‘topical vancomycin’[MeSH Terms] OR ‘sternal wound infection’[All Fields] #2 ‘superficial sternal wound infection’[All Fields] OR ‘deep sternal wound infection’ OR ‘cardiac surgery’ [All Fields] #3 #1 AND #2 |
| Embase | ‘topical vancomycin’/exp OR ‘sternal wound infection’/exp #2 ‘superficial sternal wound infection’/exp OR ‘deep sternal wound infection’ OR ‘cardiac surgery ‘ #3 #1 AND #2 |
| Cochrane library | (topical vancomycin):ti,ab,kw (sternal wound infection):ti,ab,kw (Word variations have been searched) #2 (superficial sternal wound infection): ti,ab,kw OR (deep sternal wound infection): ti,ab,kw OR (cardiac surgery): ti,ab,kw (Word variations have been searched) #3 #1 AND #2 |
2.4 Screening
The first author's last name, the length of the research, its publication year, and nation and region where the studies were lead were the bases used to condense the data. Other bases included population, participants, demographic information, therapeutic and clinical characteristics, categories, quantitative and qualitative appraisal methods, and the information source.11 Two authors independently inspected the research's methodology to determine the prospect of studies bias. The methodological quality was appraised using the ‘risk of bias instrument’ from the Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0.12 According to the appraisal criteria, which were ordered as low, unclear, or high, each study was given a rating and a risk of bias score.
2.5 Eligibility
The main discovery was the effect of TV application on lowering SWIs following CS, and a summary was created by comparing TV to control.
2.6 Inclusion
For subcategory and sensitivity analyses, comparisons amongst TV and control were done, but only papers that reported and evaluated the power of the TV compared with control were encompassed.
2.7 Statistic appraisal
In the current meta-analysis, the odds ratio (OR) was calculated with a 95% confidence interval with dichotomous or continuous methods with a fixed-or random-effect model. I2 index was determined which had a range of 0 to 100 (%).13 Indicators of no, low, moderate, and high heterogeneity were represented by I2 index readings of 0%, 25%, 50%, and 75%, respectively. When I2 was less than 50%, the fixed-effect model was designated, and when I2 was greater than 50%, the random-effect model was designated. By stratifying the first appraisal founded on the formerly declared consequence categories, a subcategory analysis was done. For the existing meta-analysis, statistical significance for differences amongst subcategories was definite as a P-value ≥.05. Publication bias was measured qualitatively and quantitatively using funnel plots of the logarithm of ORs against their standard errors and the Egger regression test (publication bias was considered existing if P ≥ .05).9 Two-tailed tests were used to compute all P-values. Reviewer Manager Version 5.3 was used to present the statistical analyses and graphs.
3 RESULTS
The results of this study are shown in Table 2. Twenty-three thousand seven hundred and forty five participants with CS were enrolled in the chosen studies; 8730 of them used TV, while 15 015 served as controls. From a total of 1054 applicable studies that were appraised, 10 papers between 1989 and 2022 that met the inclusion criteria and were encompassed in the meta-analysis were selected.7, 14-22
| Study | Country | Total | Topical vancomycin | Control |
|---|---|---|---|---|
| Vander Salm, 198914 | USA | 416 | 223 | 193 |
| Lazar, 201415 | USA | 3265 | 1075 | 2190 |
| Hamman, 201416 | USA | 1866 | 548 | 1318 |
| Lazar, 20187 | USA | 14 492 | 4997 | 9495 |
| Pervaiz, 201917 | Pakistan | 276 | 138 | 138 |
| Jadhao, 202018 | India | 862 | 430 | 432 |
| Mahmood, 202119 | Pakistan | 180 | 90 | 90 |
| Donovan, 202220 | Germany | 1251 | 662 | 589 |
| Shah, 202221 | India | 100 | 50 | 50 |
| Servito, 202222 | USA | 1037 | 517 | 520 |
| Total | 23 745 | 8730 | 15 015 |
The TV had significantly lower SWIs post CS (OR, 0.34; 95% CI, 0.20–0.57; P < .001) with moderate heterogeneity (I2 = 62%), and deep SWIs post CS (OR, 0.26; 95% CI, 0.11–0.65; P = .004) with moderate heterogeneity (I2 = 65%) compared with control as shown in Figures 2 and 3. However, there was no significant difference found between TV and control in superficial SWIs post CS (OR, 0.30; 95% CI, 0.07–1.30; P = .011) with moderate heterogeneity (I2 = 68%) as shown in Figure 4.
Although the majority of the encompassed RCTs did not include specific variables like age, gender, and ethnicity into stratified representations to inspect their impact on the comparison consequences, visual inspection of the funnel plot and quantitative extents finished using the Egger regression test both exposed no sign of publication bias (P = .88).
4 DISCUSSION
At the start of the investigations, 23 745 participants had CS at the start of the examination; 8730 of them used TV, while 15 015 were control.7, 14-22 The TV had significantly lower SWIs, and deep SWIs post CS compared with control. However, there was no significant difference found between TV and control in superficial SWIs post CS. The low number of designated studies in the meta-analysis for superficial SWIs calls for precaution when analysing the results.
Lazar and colleagues found that TV did not lessen the occurrence of deep SWI post CS in their largest study to date.7 This specific study was identified as the source of significant heterogeneity in the analysis of SWIs in the current meta-analysis for the following reasons, in addition to unbalanced baseline subject characteristics: The study covers 12 years with undeniable changes in perioperative antibiotics, glycaemic control, and sternal closure methods; non-reporting of superficial SWIs; exclusion by design of those undergoing heart transplantation who have infective endocarditis, given that they are most likely to benefit from television; exclusion of nine subjects who had deep SWI at three months following surgery; and non-reporting of sternal. Only 25% of subjects (2321/9413 from 2003 to 2010) received TV, which had a 1.29% rate of deep SWI (121 deep SWIs/9413 subjects) during the first seven years. However, in the second period (2011–2015), 53% of subjects (2676/5079 subjects) received TV, which had a nearly 5-fold lower rate (0.29% [15 deep SWIs/5079]). The outcomes must have been strongly influenced by factors other than TV for this to have happened. In contrast, TV placed on the sternal edges in combination with perioperative antibiotics and firm glycaemic control removed superficial SWIs, deep SWIs, and any SWIs in consecutive unselected CS subjects. It is important to emphasise the simplicity and applicability of TV. It is made by combining 2 g of powder of vancomycin with 2 mL of very little saline to create a wax that resembles bone wax and is then applied to the sternal halves' cut margins either right away following sternotomy5 or just prior to sternal approximation, the sternal wires are in position.15
This meta-analysis verified the TV in controlling of deep SWI.23-30 To show these potential correlations and link the effects of TV with the consequences investigated, more research is still needed. This was also proposed in other comparable meta-analysis studies that produced similarly inspiring consequences.31 Meanwhile, our meta-analysis study was incapable to define if different gender, ages, and ethnicities are linked to the consequences, well-conducted RCTs are essential to evaluate these features and the mixture of different ages, genders, ethnicities, and other alternatives of subjects.
In conclusion, the TV had significantly lower SWIs, and deep SWIs post CS compared with control. However, there was no significant difference found between TV and control in superficial SWIs post CS.
5 LIMITATIONS
There may have been bias in this study's selection because so many of the publications that were discovered were not encompassed in the meta-analysis. The eliminated papers, however, did not adhere to the standards of our meta-inclusion analysis. Moreover, we were incapable to ascertain whether or not age, gender, or ethnicity had a control on the consequences. The drive of the study was to measure how TV application affected the post-CS SWIs. Because of a lack of specificity, the study's use of data from other studies might have produced bias. Individuals’ nutritional status, gender, and age were factors that could induce bias. Not published studies and lost data might, unfortunately, cause a bias in the evaluated effect.
6 CONCLUSIONS
The TV had significantly lower SWIs, and deep SWIs post CS compared with control. However, there was no significant difference found between TV and control in superficial SWIs post CS. Analysing the findings requires caution because of the small number of studies that were included in the meta-analysis for superficial SWIs.
Open Research
DATA AVAILABILITY STATEMENT
The datasets analyzed during the current study are available from the corresponding author on reasonable request.
