Risk Factors for Surgical Site Infection in Neonates: A Systematic Review of the Literature and Meta-Analysis

Purpose: Surgical site infections (SSI) contribute to postoperative morbidity and mortality in children. Our aim was to evaluate the prevalence and identify risk factors for SSI in neonates. Methods: Using a defined strategy, three investigators searched articles on neonatal SSI published since 2000. Studies on neonates and/or patients admitted to neonatal intensive care unit following cervical/thoracic/abdominal surgery were included. Risk factors were identified from comparative studies. Meta-analysis was conducted according to PRISMA guidelines using RevMan 5.3. Data are (mean ± SD) prevalence. Results: Systematic review—of 885 abstracts screened, 48 studies (27,760 neonates) were included. The incidence of SSI was 5.6% (1,564 patients). SSI was more frequent in males (61.8%), premature babies (77.4%), and following gastrointestinal surgery (95.4%). Meta-analysis—10 comparative studies (16,442 neonates; 946 SSI 5.7%) showed that predictive factors for SSI development were gestational age, birth weight, age at surgery, length of surgical procedure, number of procedure per patient, length of preoperative hospital stay, and preoperative sepsis. Conversely, preoperative antibiotic use was not significantly associated with development of SSI. Conclusions: Younger neonates and those undergoing abdominal procedures are at higher risk for SSI. Given the lack of evidence-based literature, prospective studies may help determine the risk factors for SSI in neonates.


INTRODUCTION
Surgical site infections (SSI) are infections that occur postoperatively in the area of the body where the surgery took place. SSI can be superficial and involve the skin only, or more serious and involve other tissues, organs, or implanted material. SSI are among the most common hospital acquired diseases and are an important cause of morbidity and mortality in all patients, including neonates and infants (1,2).
Whilst the incidence and risk factors for SSI in adults and more recently in children have been defined and management guidelines have been established (3,4), yet little is known about SSI in neonates and infants.
The incidence of SSI is 2-5% in adult patients undergoing inpatient surgery (3). Risk factors associated with SSI included co-morbidities, advanced age, risk indices, patient frailty, and surgery complexity (5).
In children the rate of SSI ranged from 2.5 to 5.4% and dirty wounds, inpatient status, increased duration of surgery, or certain surgical disciplines (cardiovascular, general surgery, neurosurgery, and orthopedics) were associated with increased risk of developing an SSI (4).
Previous studies have shown that the incidence of SSI in neonates and infants can be as high as 17% (1,2). In this population of patients, several conditions have been reported to be associated with an increased risk of SSI, including admission to the neonatal intensive care unit (NICU), history of prematurity, low birth weight, mechanical ventilation, central venous access, associated co-morbidities, prolonged antibiotic administration, postsurgical hyperglycemia, and neutropenia (1,2,6,7).
In the present study, we aimed to establish the incidence of SSI in neonates and to identifyprognostic factors that may help stratify which neonates are at increased risk to develop this complication. A better understanding of the causes leading to SSI could reduce their incidence, help define guidelines, and eventually improve outcome.

METHODS
To investigate the incidence and risk factors of SSI in neonates, we conducted a systematic review of the literature and complemented it with a meta-analysis of comparative studies. Both the systematic review and the meta-analysis were drafted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (8). The protocol for this systematic review was registered on PROSPERO-international prospective register of systematic reviews (registration number: CRD42017069003) (9). Using a defined search strategy, three investigators (VDC, AB, and GL) independently searched scientific databases (PubMed, Medline, Cochrane Collaboration, Embase, Web of Science, Ovid) using a combination of keywords ( Table 1). MeSH headings and terms used were "neonate OR newborn" AND "surgery OR surgical" AND "wound infection OR surgical site infection" (Supplementary File 1).
All gray literature publications (i.e., reports, theses, conference proceedings, bibliographies, commercial documentations, and official documents not published commercially) were excluded. Only studies on neonates (<44 wks gestational age) and/or neonates admitted to the NICU following cervical/thoracic/abdominal surgery published since 2,000 were included. Case reports, case series with <10 patients, animal Abbreviations: SSI, surgical site infection; NICU, neonatal intensive care unit.

Statistical Analysis
Data were analyzed using GraphPad Prism 6.2 Macintosh Version (10). Data were compared using Fisher's exact test and are expressed as mean ± SD. When median and range were reported, mean ± SD were estimated, as previously reported (11). The meta-analysis was conducted with RevMan 5.3 (12), using the fixed-effects model to produce risk ratio (RR) for categorical variables and mean differences (MD) for continuous variables, along with 95% confidence intervals (CI). We produced I 2 values to assess homogeneity and quantify the dispersion of effect sizes.

Quality Assessment
Risk of bias for individual studies was assessed in duplicate (VDC and GL) using the methodological index for non-randomized studies (MINORS) (13). Differences between the two reviewers were resolved through consensus and discussion with another author (AZ). The total score for this 12-item instrument ranges 0-24 points with a validated "gold standard" cut-off of 19.8.
Two authors (FM and AZ) independently evaluated the present systematic reviews and meta-analysis using A Measurement Toll to Assess Systematic Reviews (AMSTAR) (14). The PRISMA checklist of our study was then completed (8).

DISCUSSION
The overall rate of SSI in adult and pediatric patients is ∼ 2-5% and SSI are associated with increased morbidity, mortality, healthcare costs, and length of hospital stay (3,62). Risk factors for SSI have been identified in predominantly adult populations and include advanced age, hyperglycemia, malnutrition, co-morbidities, risk indices, patient frailty, prior infections, complexity of surgery, increased operative time, and increased blood loss during surgery (5,45). With regard to children, certain surgical disciplines (cardiovascular, general surgery, neurosurgery, and orthopedics) were associated with increased risk of developing an SSI (4). SSI development substantially increases the clinical and economic burden of surgery, because of prolonged hospitalization of the patient, diagnostic tests, and treatment. Moreover, SSI negatively impact on patient physical and mental health as well as on loss of earnings during recovery (63,64). Our study shows that neonates undergoing abdominal surgery are at high risk of SSI and that age at surgery and length of procedure are the main predictors of SSI in those admitted to NICU. Similarly, male gender and gestational age may be associated to the development of SSI, but the present data are not conclusive.
Although there is an abundance of literature on SSI in adults, there is a lack of studies having examined risk factors for SSI in neonates undergoing surgical interventions.
In the present systematic review of the literature on more than 27,000 patients, we found an incidence of SSI of 5.5%. Interestingly, the overall rate of SSI in the neonatal age group in our review is comparable to the rates reported in the older pediatric age group (51,65,66). This finding suggests that neonates may be less prone to SSI than it might be expected based        on their alleged fragility, as they represent a special population that is thought to be at higher risk for infection due to their immature immune systems (6,67,68).
According to our study, premature infants represent a significant proportion of infants who require surgical interventions (51), and this is confirmed by the overall prevalence of SSI in this cohort (77%). When examining the type of surgical intervention, the vast majority of infants underwent gastrointestinal surgery (96%). In particular, abdominal surgery was significantly associated with an increased risk of SSI (23,54). The most common type of surgical intervention described in the included articles was laparotomy for congenital abdominal wall defects, necrotizing enterocolitis, or congenital bowel obstruction. This surgical procedure is known to compromise  2  2  2  2  2  2  2  2  2  2   Total score  13  16  13  14  14  14  14  16  14  16 the integrity of the gastrointestinal tract and to potentially result in bacterial translocation. Surgical wounds following a neonatal laparotomy are classified, at best, as clean-contaminated wounds, which justify the highest prevalence of SSI in this sub-group (22). The Canadian Pediatric Surgery Network recently reported an overall 15% incidence of SSI in infants with gastroschisis who underwent immediate (<6 h after birth) or delayed closure (55). For this reason, to reduce time of visceral exposure, the authors have proposed the gastroschisis sutureless closure, as it is also associated with a reduced risk of SSI (54).
To define the risk factors that are to be considered by the surgeon to estimate the risk of SSI at the time of neonatal surgery, we selected comparative studies that analyzed neonates with or without SSI. Interestingly, we observed that neonates who developed SSI had a younger GA compared to those who did not. This was a validation of the results from the systematic review that showed a high prevalence of prematurity among patients who developed SSI. Likewise, also a lower birth weight was associated with an increased risk of SSI in the included studies, even if the SSI group had an older age at the time of their procedures compared to the group who neonates who did not develop SSI. Moreover, this group also had a longer preoperative length of stay and it likely required a greater number of invasive diagnostic and therapeutic procedures per patient, as already reported by Garcia and Lejus (24,34). In fact, this study confirmed a significant difference in number of surgical procedures between neonates with SSI and those without. This figure may be related to the severity of illness, with sicker patients being more likely to require additional procedures, although one third of SSI did occur after a single or first procedure. A long preoperative admission is a risk factor for SSI, as it is proportional to the severity of the underlying clinical conditions, the need for invasive devices and treatments (included prolonged antibiotics), and it promotes nosocomial flora colonization. Furthermore, neonates admitted to NICU are highly susceptible to nosocomial infections. As an expected consequence, the presence of a systemic infection significantly influenced the SSI incidence in our study.
Our present study also highlighted that the length of surgery is another risk factor for SSI. Length of surgery was not identified by all studies as a risk factor for SSI, and our meta-analysis confirmed the findings of Clements and coworkers who found a longer operative time in patients developing SSI (22). Prolonged visceral exposure may negatively impact on surgical outcome as consequence of skin contamination. Furthermore, the longer the operative time the deeper the surgical stress response, as the invasiveness of surgery and the length of procedure significantly correlated to oxidative stress activation and cortisol response (69)(70)(71). The latter may have an impact on postoperative outcome, such as the development of infectious complications, including SSI.
This study did not show any difference in preoperative antibiotic administration between neonates who developed SSI and those who did not. Whilst standardized preoperative antibiotic protocols in the adult population have shown to reduce the rate of SSI, a consensus is lacking among pediatric surgeons regarding preoperative antibiotic prophylaxis especially in neonates (1,4). The different definitions of antibiotic prophylaxis and antibiotic regimen used in the analyzed studies may have led to our findings. The isolation of skin flora from a large number of wound cultures suggests that standardization of preoperative prophylaxis could potentially have an impact on the rate of SSI as has previously been demonstrated in pediatric patients (72,73).

LIMITATIONS OF THE STUDY
Our study has some limitations. The first relates to the relative small number of studies available for the meta-analysis, with only two prospective cohort study and one national prospective database included. Nonetheless, the population of neonates reported in the studies was not small, with more than 27,000 neonates included.
The second limitation is that we could not analyze important variables, such as the use of adequate antibiotic prophylaxis, as they were not reported in the few studies selected in our analysis. Finally, the third limitation is the relative heterogeneity of the patient population: although we tried to limit the study to neonates, we included both patients with a post-conceptional age below 44 weeks and infants admitted in the NICU following abdominal, cervical, and thoracic surgery.
As a consequence, in our meta-analysis, none of the studies reached the gold standard cut-off on MINORS of 19.8 out of 24 (Table 4). However, when independently assessed by two authors using AMSTAR, the present systematic reviews and metaanalysis received a relevant score (Supplementary File 2) and the PRISMA checklist was completed (Supplementary File 3).

CONCLUSION
In conclusion, SSI is a significant complication in neonates admitted to NICU that can negatively impact their outcome by prolonging hospital stay and increasing the risk for further complications, such as potentially fatal sepsis. Younger and smaller neonates at birth, those requiring longer or multiple operative procedures, and those with prolonged preoperative hospital stay and preoperative sepsis are at higher risk for SSI. These patients require special attention with close monitoring during their post-operative course. Given the lack of evidence in the literature, well-designed prospective studies on large cohorts of neonates may help setting up specific guidelines for the prevention and treatment of SSI in this particular population.

DATA AVAILABILITY
All datasets generated for this study are included in the manuscript and/or the Supplementary Files.

AUTHOR CONTRIBUTIONS
VC, AB, GL, FM, and AZ: conception and design, analysis and interpretation, drafting, final approval. VC, AB, and GL: data acquisition. VC, GL, FM, and AZ: quality assessment. GL, FM, and AZ: revision.