The Japan Society for Surgical Infection: guidelines for the prevention, detection, and management of gastroenterological surgical site infection, 2018

The meta-analysis was conducted on seven observational studies^(5–11). The incidence of S. aureus in SSI was significantly higher in nasal carriers of S. aureus undergoing digestive surgery (odds ratio [OR] 9.0, 95% confidence interval [CI] 5.09–15.91) (Fig. 3-1). However, six of the seven studies included patients who were undergoing not only digestive surgery, but also other types of surgery^{(5, 7–11)}. Moreover, six of seven subjects were evaluated for the presence or absence of MRSA carriage only^(5–10).

Two meta-analyses were performed to evaluate the benefit of decolonization in patients with known nasal carriage of S. aureus and of universal decolonization in all preoperative patients. Two RCTs showed that decolonization in patients with nasal S. aureus carriage had significantly greater benefit than no treatment in reducing the S. aureus SSI rate (Fig. 3-2)^{(7, 12)}. There was no significant difference in mortality (Fig. 3-3). A prospective intervention cohort study also showed that the SSI rate in the decolonized group was not significantly higher than that of non-carriers (OR 0.8, 95% CI 0.19–3.44)⁽¹³⁾. All subjects in the decolonization group had received preoperative intranasal mupirocin ointment with or without a combination of chlorhexidine gluconate body wash. Two subsequent RCTs showed that universal decolonization did not reduce the SSI incidence significantly (Fig. 3-4)^{(7, 14)}. On the other hand, in a historical control study the incidence of SSI was significantly lower with universal decolonization than without it⁽¹⁵⁾. However, the spread of mupirocin resistance by universal decolonization is an important issue. For this reason, the committee members decided not to recommend universal decolonization (B, 4).

An analysis of six observational studies showed that the SSI rate in patients with preoperative malnutrition was higher than that in patients without malnutrition (OR 3.48, 95% CI 2.57–4.71, P < 0.00001) (Fig. 3-5)^(16–20). The reports used the prognostic nutrition indicators of serum albumin levels, prealbumin levels, and weight loss, to identify patients with malnutrition. A subsequent meta-analysis, consisting of two RCTs^{(21, 22)} and one observational study⁽¹⁷⁾ for SSI, revealed that nutrition improvement before surgery for patients with malnutrition reduced the incidence of SSI. The combined OR was 0.56 (95% CI 0.37–0.84) for the RCTs (Fig. 3-6) and 0.25 (95% CI 0.11–0.56) for the observational study. However, the methods and duration of appropriate nutrition interventions are unclear.

Three RCTs and one prospective assignment study showed that preoperative immunonutrition does not reduce the rate of SSI significantly in patients without malnutrition (RR 0.63, 95% CI 0.31–1.27) (Fig. 3-7). Moreover, preoperative immunonutrition did not result in significant improvement in terms of duration of hospital stay (Fig. 3-8)^{(23, 24)} [1, 4] and survival rates (Fig. 3-9)^{(22, 23, 25)}. However, several studies have added immunomodulatory nutrition to regular meals, so the study design may not have been appropriate.

An RCT⁽²⁶⁾ comparing SSI incidence in smokers with or without preoperative smoking cessation showed a tendency toward a decrease in the risk ratio (RR) of 0.48 (95% CI 0.2–1.16), but no significant difference was observed. In the meta-analysis comparing SSI incidence rates between preoperative smokers and nonsmokers, preoperative smoking was significantly associated with an SSI risk, with an OR of 1.79 (95% CI 1.59–1.94)^(27–57).

With respect to the incidence of SSI in relation to differences in alcohol consumption, the OR was 1.43 (95% CI 1.10–1.85) in seven observational studies (Fig. 3-11), and heavy drinking was a significant risk factor for SSI^{(39, 58–63)}. A small RCT of preoperative abstinence in drinkers did not show a significant effect on SSI reduction (RR 0.972; 95% CI 0.70–1.35)⁽⁶⁴⁾.

The administration of infliximab (IFX) has not been identified as a risk factor for SSI in patients with inflammatory bowel disease (OR 0.94; 95% CI 0.62–1.41)^(65–78). In an observational study by Ahn et al., the onset of postoperative infection, with or without long-term steroid administration, was examined in patients with inflammatory bowel disease. The steroid administration group had a significantly higher incidence of postoperative infectious complications (OR 5.83; 95% CI 1.063–32.021)⁽⁷⁹⁾. In a report by Miki et al.⁽⁸⁰⁾, postoperative infectious complications were examined by dividing the preoperative steroid total dose of 12 g as the boundary into two groups of high dose and low dose. The low-dose group had significantly lower postoperative infectious complications than the high-dose group (OR 3.40; 95% CI 1.172–9.862). Based on these results, long-term or high-dose steroids are regarded as a risk factor for postoperative SSI incidence (C).

Ten RCTs that analyzed the presence or absence of MBP in colorectal surgery and the incidence of SSI^(81–90) found no difference between the RR of 1.02 (95% CI 0.82–1.28). In a comparison between OAMBP and MBP in colorectal surgery, there was a RR of 0.61 (95% CI 0.46–0.82) based on the analysis of the 10 RCTs^(901–100); OAMBP was found to reduce the rate of SSI significantly. Two large multiple-case studies compared OAMBP and no MBP in colorectal surgery.^{(101, 102)} The OR was 0.42 (95% CI 0.35–0.50), which was significant, and SSI reduction effect was also significantly higher for patients who received OAMBP.

The meta-analysis^(103–112) showed that cleansing the skin with chlorhexidine did not reduce the occurrence of SSI (RR 0.94; 95% CI 0.85–1.05).

In a meta-analysis comparing clipper hair removal, depilation and shaving^(113–119), the RR of 0.54 (95% CI 0.38–0.78) was significant and the incidence of SSI was significantly lower after clipper hair removal. In the comparison of depilatory cream and shaving, the SSI was lower after the depilatory cream, not significantly^(120–124) (RR 0.52; 95% CI 0.24–1.11). In the comparison between no hair removal and shaving^{(113, 121, 125–128)}, the RR was 0.58 (95% CI 0.34–0.98), which was significant, and the incidence of SSI was low in the absence of hair removal.

Figure 4-1 shows the results of a meta-analysis of 13 RCTs evaluating prophylactic antibiotic treatment for laparoscopic cholecystectomy^(129–141). Prophylactic antibiotic treatment resulted in a significant reduction in SSI incidence, and given that population sizes were sufficiently large, at 2000 patients in each RCT, the information was given a level of evidence of A. Similarly, Fig. 4-2 shows the results of a meta-analysis of the effects of prophylactic antibiotic treatment for inguinal hernia surgery. The results of these analyses demonstrate that prophylactic antibiotic treatment is effective, even for laparoscopic cholecystectomy and radical inguinal hernia surgery, both of which involve a low degree of surgical site contamination. Therefore, one can infer that prophylactic antibiotic treatment is effective in the field of gastrointestinal surgery with a higher degree of contamination and incidence of SSIs. Hence, prophylactic antibiotic treatment is deemed beneficial in gastrointestinal surgery.

Three cohort studies compared SSI incidence after administration within 1 h before the surgical incision versus more than 1 h before the surgical incision^(142–144). Overall, 3606 patients received prophylactic antibiotic treatment within 1 h before the surgical incision and 3386 received prophylactic antibiotic treatment more than 1 h before the surgical incision. Analysis revealed an OR of 0.91 (95% CI 0.71–1.15) and no significant difference between the groups. However, theoretically, attaining an appropriate blood concentration of prophylactic antibiotics by the time of surgical incision is preferred, and hence the conventional administration method is recommended.

Although no high-quality study has shown that the SSI incidence is reduced by intraoperative re-administration of prophylactic antibiotics, from the viewpoint of PK and PD, it is logical that an appropriate blood concentration of antibiotics should be maintained during surgery, so intraoperative re-administration is preferred. Nevertheless, the appropriate timing of re-administration and how to adjust timing when severe bleeding occurs have not been established.

Figures 4-3 and 4-4 show forest plots of the SSI incidence after single dosing and repeated dosing of prophylactic antibiotics for gastrectomy and colectomy. Overall, the analysis for gastrectomy revealed a risk ratio (RR) of 0.97 (95% CI 0.55–1.68) and showed that SSI incidence was not increased significantly, even in the preoperative/intraoperative treatment group^(144–147). However, since the population size was not sufficiently large, this was given a level of evidence of B. Similarly, even in cases of elective colectomy for colorectal cancer, the SSI incidence did not increase significantly with single-dose treatment versus repeated-dose treatment. However, because there were only two RCT reports, the population size was small, and because the control group and intervention group differed greatly between the reports, the information was given a level of evidence of C.

A similar description was found in some American⁽¹⁴⁸⁾ and European^{(149, 150)} guidelines and no significant difference was found. There were three interventional RCTs with SSI incidence as the outcome, with the same conclusion^(151–153). Three observational studies with SSI incidence as outcome and three RCTs with a colony-forming unit (CFU) as an outcome were also confirmed^{(151, 154–158)}. The superiority of either hygiene method has not been confirmed in any meta-analysis.

Rubbing with alcohol-based hand antiseptic is often performed instead of the conventional hand washing (scrubbing) with running water and soap for surgical hand hygiene. These methods have been compared in some guidelines^(148–150), but it has not been confirmed which method is superior. Intervention studies targeting SSI incidence as an outcome have not shown any superiority of either method. Moreover, observational studies using SSI incidence and RCTs with CFU as an outcome have not confirmed the superiority of either method.

A study from Japan identified some reports that the cost of the rubbing method is cheaper than the scrub method⁽¹⁵⁹⁾, but the number of cases was low and the evidence level is low.

The most widely used agents for surgical skin preparation were povidone iodine or alcohol-based chlorhexidine gluconate solutions. The World Health Organization’s (WHO) 2016 Guidelines for the Prevention of Surgical Site Infections include recommendations for surgical site skin preparation⁽¹⁵⁰⁾. Alcohol-based antiseptic solutions based on chlorhexidine gluconate for surgical site skin preparation were recommended. It is important to identify which skin antiseptics for surgical site skin preparation are most effective for preventing SSI.

The meta-analysis was conducted with three RCTs^(160–162). These three trials compared the efficacy of aqueous-based povidone iodine with alcohol-based chlorhexidine gluconate solutions for preventing SSI. The meta-analysis indicated that alcohol-based antiseptic solutions based on chlorhexidine gluconate are more effective than aqueous-based povidone iodine in reducing the risk of SSI in patients undergoing gastrointestinal surgery (Fig. 5-1). However, several factors require attention when applying this recommendation in clinical practice. The concentration of antiseptic agent varied among the studies, with the iodophor compound ranging from 5–10% and chlorhexidine gluconate ranging from 0.5–2.5%. Furthermore, we cannot use chlorhexidine gluconate solutions with a concentration over 1% in Japan. In future studies, we need to examine the appropriate concentration of chlorhexidine gluconate solutions for clinical practice in Japan. There are eight reports of a fire in the operating room associated with the use of alcohol-based antiseptics solutions in Japan. Measures should be taken to prevent such incidents before the use of alcohol-based antiseptic solution for surgical site skin preparation. We also need to check if patients have allergies to components of the antiseptics. Aqueous-based chlorhexidine gluconate should be used when a patient has alcohol intolerance.

Olanexidine has been introduced for surgical site skin preparation. Olanexidine has antimicrobial activity against a wide range of Gram-positive and Gram-negative bacteria. It also has bactericidal activity against Vancomycin-resistant enterococci, Pseudomonas aeruginosa, Serratia spp., and Cepacia spp. A clinical trial to elucidate whether these various concentrations antiseptics and Olanexidine are more effective than alcohol-based antiseptics should be performed.

SSIs often develop after gastrointestinal surgical procedures; these wounds are mostly classified as clean-contaminated wounds. Various strategies, such as single-use drapes, are used to reduce wound contamination during a surgical procedure; however, their effectiveness in preventing wound infection is limited. For this reason, adhesive drapes are now used widely to reduce wound contamination.

We identified three RCTs^(164–166) and one observational study⁽¹⁶⁷⁾. The observational study found that SSI was significantly less likely when surgery was performed with adhesive drapes. The meta-analysis was conducted with three RCTs^(164–166). Results from the meta-analysis indicated that surgery with adhesive drapes did not reduce the risk of SSI compared with surgery without adhesive drape (Fig. 5-2).

Despite using various strategies to minimize surgical wound contamination during surgery, incisional SSIs often occur after gastrointestinal surgical procedures. These wounds are usually classified as clean-contaminated wounds. Currently, surgical wound protectors, which comprise a non-adhesive plastic sheath attached to a single or double rubber ring that secures the sheath to the wound edges, are available to reduce wound edge contamination.

A meta-analysis was conducted with eight RCTs^(168–175) comparing the use of a surgical wound protector with conventional wound protection for preventing SSI. The meta-analysis indicated that wound protector devices are significantly more effective than conventional wound protection for reducing the risk of SSI in patients undergoing gastrointestinal surgery (Fig. 5-3). One trial performed an economic evaluation of wound protector devices⁽¹⁷⁶⁾ and found no evidence to suggest that wound-edge protection devices are a cost-effective device to reduce SSI. There were several limitations in this analysis. First, the structural design of wound protector devices varied among the studies. Second, this analysis was unable to evaluate the efficacy in preventing SSI between the single-ring and the double-ring wound protector devices. Third, all studies targeted open gastrointestinal surgery. It is unknown whether the use of wound protector devices in laparoscopic surgeries can reduce the risk of SSI.

We cannot provide any recommendations for the use of double-gloving (D), changing gloves (C), or repeated hand washing (D) during digestive surgery for preventing SSI due to the lack of supporting evidence (D). However, double-gloving can reduce the exposure risk since the incidence of glove perforation is significantly lower for the inner gloves (A).

The WHO stated that a recommendation could not be formulated because of the lack of evidence⁽¹⁵⁰⁾. However, it stated that most surgeons prefer to double-glove because bacterial contamination of the surgical field may occur in the event of glove perforation and most surgeons prefer to wear double gloves for their own protection against injury or infections.

There were no RCTs or observational studies concerning double-gloving for digestive surgery and the incidence of SSI. On the other hand, five RCTs were identified concerning glove perforations^(177–181). In the meta-analysis of these five RCTs, the incidence of perforation in single-gloving was significantly higher than that in the inner gloves used in double-gloving (Fig. 5-5). We could not find any RCTs in digestive surgery concerning changing gloves during surgery. Although we found an observational study with low quality evidence for changing gloves, there were no significant effects⁽¹⁸²⁾. Moreover, we could not find any RCTs or OBSs in digestive surgery concerning repeated hand washing for preventing SSI. Thus, we could not formulate any recommendations for using double-gloving, changing gloves, or repeated hand washing during digestive surgery to prevent SSI. However, the incidence of glove perforation was significantly lower with double-gloving than with single-gloving.

The limitations of this analysis were as follows: (1) There was no cost analysis. (2) There was no evidence for whether glove perforation can actually lead to occupational infections. (3) Using double-gloving may reduce operability, although no evidence could be provided.

We could not address this CQ fully as there are no RCTs on the correlation between changing instruments and preventing SSI. Further studies are needed.

A total of 10 RCTs and 5 observational studies were identified^(183−197). We analyzed the incidence of SSI and the duration of hospitalization.

The results of a meta-analysis of all 10 RCTs showed that antimicrobial-coated sutures used in digestive surgery were effective for preventing SSI (RR 0.68; 95% CI 0.48–0.95; p = 0.03) (Fig. 5-5). However, in the analyses distinct from suture material, we could find significant efficacy only for poly-filament suture (RR 0.45; 95% CI 0.26–0.77; p = 0.004) in six RCTs, but not even for these in four RCTs (RR 0.79; 95% CI 0.54–1.17; p = 0.24). Analysis of five RCTs showed that the duration of hospitalization did not decrease significantly when antimicrobial-coated sutures were used (risk differences [RD] − 0.5; 95% CI − 16.68–6.69; p = 0.40). Thus, we recommend the use of antimicrobial-coated sutures to prevent SSI in digestive surgery. However, studies on cost benefit, efficacy in a pediatric population, or adverse events are lacking. All analyzed results of this systematic review can be found in the report by Uchino et al.⁽¹⁹⁸⁾.

We could find only one RCT published since 2000, that examines the efficacy of wound irrigation in appendectomies⁽¹⁹⁹⁾. In a systematic review and meta-analysis reported in 2015, wound irrigation was found to be significantly more effective in preventing SSI⁽²⁰⁰⁾. However, this meta-analysis included 15 appendectomies among 34 RCTs and 33 of the 34 studies were conducted before 2000. It is difficult to apply these results to recent cases of abdominal surgery, or even laparoscopic surgery. Therefore, recent evidence is insufficient on the relationship between wound irrigation and SSI.

We found two RCTs on high-pressure wound irrigation^{(199, 201)} in appendectomy and hepatic surgery, and three observational studies^(202−204). Although a significant effect was found in the meta-analysis for RCTs (Fig. 5-6) and OBSs (Fig. 5-7), evidence levels were low in both RCTs due to unclear definitions of outcome, randomization, concealment, and assessment of outcomes. Moreover, several methods of high-pressure irrigation were used in these studies, including pulsation and syringe with or without a thin needle.

Wound irrigation with disinfectant or antibiotic solution has not been practiced in digestive surgery since 2000.

There were two RCTs concerning the use of wound irrigation with electrolyzed acidic aqueous solution in preventing SSI^(205,206); however, no significant efficacy was observed for preventing SSI (RR 0.42; 95% CI 0.09–2.03; p = 0.19). Conversely, using electrolyzed acidic aqueous solution could be harmful since wound dehiscence and herniation were increased in one RCT (RR 2.28; 95% CI 1.03–5.04; p = 0.04)⁽⁸⁾. Based on these results, we recommend wound irrigation, especially with high pressure, for preventing SSI. However, a recommendation for wound irrigation with disinfectant, antibiotics, or electrolyzed acidic aqueous solution could not be formulated due to the lack of evidence in digestive surgery.

Figures 5-8 and 5-9 show the results of a meta-analysis of three RCTs^(207−209) and two observational studies^(210,211), respectively. Peritoneal lavage appeared to be harmful in one RCT on elective hepatic surgery, although there was no significant difference (RR 2.31; 95% CI 0.99–5.36)⁽¹⁾. The other two RCTs were conducted on restricted to open appendectomy^(208,209). Two observational studies were also conducted on restricted to open appendectomy^(210,211). Although the RCTs did not find significant efficacy, analysis of the observational studies found that it was significantly harmful. However, we could not apply these results to universal digestive surgery for either emergency surgery with dirty/infected wounds or elective surgery with clean-contaminated wounds. Therefore, we cannot comment on the efficacy of peritoneal lavage for preventing SSI due to the lack of evidence.

The reported RCTs^(212–214) were considered with few observational studies and high evidence. Two observational studies were also considered^{(215, 216)}. There was no significant difference in mortality rates after total gastrectomy and distal gastrectomy with versus without drain placement (Figs. 5-10 and 5-11, respectively). We recommend that drain placement after gastrectomy be judged according to the clinical outcome and skills of the individual institution (B, 3).

Results from a meta-analysis of 13 RCTs had a high evidence level because of the homogenous population and the inclusion of more than 500 cases^(217–229). There is no disadvantage for patients without drain placement because the mortality rate was not different with or without a drain. Moreover, the SSI rate did not differ among the groups, although that in the no-drain group tended to be lower than that in the drain group (Fig. 5-12). On the other hand, most investigators recommend no drain after laparoscopic cholecystectomy because of some benefits, such as shortening the operation time (Fig. 5-13). Therefore, committee members recommend that drainage after laparoscopic cholecystectomy is unnecessary (A, 4).

The meta-analysis was conducted with six RCTs^(230–235). Mortality rates were not significantly different between the drain and no-drain groups (Fig. 5-14). The SSI rates tended to be higher in the drain group (Fig. 5-15) and ascitic fluid leakage was significantly reduced in the no-drain group (Fig. 5-16). Therefore, as drain placement after hepatectomy without biliary reconstruction tends to increase SSI rates, ascitic leakage, extend the duration of hospitalization, no drain is recommended with high evidence levels (A, 4).

The meta-analysis was conducted with three RCTs^(236−238). The Van Buren’s (2014) RCT study⁽²³⁷⁾ for a pancreatoduodenectomy was discontinued by the study’s Medical Safety Commission (Fig. 5-17)⁽²³⁹⁾. Therefore, it is suggested that no drainage may increase the mortality rate, and drain placement is better from the viewpoint of medical safety (B, 2b). Concerning the timing of drain removal, the group that had early drain removal after surgery within certain removal criteria had fewer intraabdominal infections (Fig. 5-18), and the duration of hospitalization was also shortened significantly (Fig. 5-19). Early removal is recommended by the committee, although this does not have a high evidence level and the number of cases is limited.

It is important to note that all evidence was collected for open surgery using a Penrose drain. Although the evidence level was moderate, most appendectomies are now performed under laparoscopy and the clinical background of the studies and medical reality differs. With this limited evidence, a meta-analysis was conducted with seven RCTs^(239−245). Mortality rates (Fig. 5-20) were lower in the no-drain group. Therefore, drain placement after appendectomy is unnecessary, unless patients need abdominal lavage.

Drain placement does not affect the clinical outcome of suture failure, abscess formation, or mortality, after colon surgery or rectal surgery. Although there was no significant difference in rectal surgery, the incidence of mortality (Fig. 5-21) tended to be lower in the drain group^(246–249). In colon surgery, we did not identify any merits of drain placement or of no drain placement^(250–253). The committee recommend that drain placement is unnecessary (Recommendation 4). There was no significant difference in rectal surgery, and the merits of drain placement are unclear, although drain placement tends to reduce serious complications (Recommendation 3).

The meta-analysis was conducted with seven RCTs^(254–260), which showed that the incidence of superficial incision wounds in SSI was significantly reduced when subcutaneous drains were placed (Fig. 5-22). With respect to the high evidence level of the studies included in the meta-analysis, subcutaneous draining appears to reduce the incidence of SSI after gastrointestinal tract surgery (Evidence level, B). However, there is no clear recommendation because the subcutaneous drains in clinical practice are not widespread, and the indication in clinical practice is inconclusive.

Meta-analysis of six RCTs for superficial SSI^(261−266) revealed a significantly lower incidence of SSI when absorbable sutures were used than when non-absorbable sutures were used (Fig. 5-23). The meta-analysis from appendectomy alone showed that the incidence of superficial SSI was significantly lower for absorbable sutures than non-absorbable sutures. Therefore, absorbable sutures are more clinically valuable than non-absorbable sutures for superficial SSI and wound dehiscence in primary closure using subcutaneous sutures^(261−266).

A meta-analysis of four RCTs for superficial SSI and wound dehiscence for subcutaneous suturing^{(262, 263,267, 268)} revealed that the incidence of wound dehiscence was significantly lower for continuous sutures than for interrupted sutures (Fig. 5-24). A meta-analysis of eight RCTs for superficial SSI and hernia formation for fascia closure^(269–276) revealed that the SSI rate and hernia formation were not different between continuous and interrupted sutures. Because there was less wound dehiscence associated with continuous sutures, interrupted sutures are strongly recommended for subcutaneous suturing.

A meta-analysis of two RCTs on the rate of SSI rate and wound complications after skin closure using subcutaneous sutures versus a skin stapler^(277,278) revealed that the incidence of each clinical indicator for subcutaneous sutures tended to be less than that for skin stapling. The RCTs consisted of more than 1000 subjects and had high statistical power. Subcutaneous sutures resulted in less wound thickness than a skin stapler. Patient satisfaction after subcutaneous sutures were used ranked excellent and obtained 54% (268/511), whereas patient satisfaction after a skin stapler was used obtained only 42.7% (211/494: P = 0.002).

The number of patients in each study was small, so the evidence level was C. A meta-analysis of six RCTs that compared bioadhesives and subcutaneous sutures^(279–284) revealed that the incidence of SSI and wound dehiscence was similar for both methods. The cost of using bioadhesives could be higher than that of subcutaneous sutures and it can also cause chemical burns. Therefore, the clinical application of these materials must be carried out carefully.

Figure 6-1 shows the results of a meta-analysis of 29 RCTs^(285−313) on early recovery after surgery (ERAS)/Fast Track Surgery (FTS) for digestive surgery. Implementation of an ERP like ERAS/FTS significantly reduced the incidence of SSI after digestive surgery. The length of hospital stay (standardized mean difference [SMD]: − 1.05 day; 95% CI − 1.41, − 0.75) and overall postoperative complications (RR 0.76; 95% CI 0.63, 0.93) were significantly lower in the ERAS/FTS groups. Another meta-analysis of 27 RCTs also demonstrated that ERAS/FTS for abdominal or pelvic surgery had a similar effect on SSI prevention (RR 0.77; 95% CI 0.58, 0.98)⁽³¹⁴⁾ [30]. Therefore, we recommend ERP after digestive surgery to decrease the risk of SSI (A, 2a). However, it remains unclear which program components are optimal to prevent SSI after various types of digestive surgery because several components constituting ERP were different in every type of digestive surgery reported.

Meta-analysis of six RCTs^(315–320) on preoperative CHO loading did not show any effect on SSI prevention after digestive surgery (Fig. 6-2). Another meta-analysis of 8 RCTs ahowed no significant difference in the incidence of postoperative complications (RR 0.85; 95% CI 0.66, 1.08) between preoperative CHO loading and a placebo^(315–322). A meta-analysis of preoperative CHO loading for elective surgery, including cardiovascular and hip joint surgery, also showed no effect on preventing surgical complications (RR 0.88; 95% CI 0.50, 1.55)⁽³²³⁾. Therefore, we do not recommend the preoperative administration of CHO to prevent SSI (A, 3).

The American College of Surgeons (ACS)/Surgical Infection Society (SIS) guideline recommends a target blood glucose level to prevent SSI of 110–150 mg/dL, and less than 180 mg/dL for cardiovascular surgery⁽³²⁴⁾. The CDC guideline recommends a target blood glucose level of less than 200 mg/dL⁽³²⁵⁾ and the WHO global guideline suggests 110–150 mg/dL or less than 150 mg/dL without a definitive recommendation⁽¹⁵⁰⁾. To address the optimal blood glucose target level for SSI prevention in digestive surgery, four RCTs^(326−329) and three observational studies^(330–332) were identified. In a meta-analysis of the four RCTs, the summary estimate showed a significant benefit of intensive glucose control compared with conventional control for reducing the incidence of SSI in patients with and those without diabetes (Fig. 6-3). The intensive group’s target blood glucose levels were 80–110 mg/dL in the four RCTs, resulting in a high incidence of hypoglycemic events (RR 7.11, 95% CI 2.15, 23.55). The target levels of blood glucose were different in each observational study: 80–140 mg/dL, less than 125 mg/dL, and less than 180 mg/dL, respectively. The incidences of SSI in each glucose control group were significantly lower than those in the reference groups. A recent meta-analysis reported that the effect was similar in studies with a target blood glucose level of less than 110 mg/dL (RR 0.50; 95% CI 0.35, 0.73) and an upper limit target level of 110–150 mg/dL (RR 0.43; 95% CI 0.29, 0.63)⁽³³³⁾. Considering this evidence, we recommend blood glucose target levels of less than 150 mg/dL to reduce the incidence of SSIs in digestive surgery patients with and those without diabetes mellitus; however, the available evidence is low quality and hypoglycemic events should be avoided in intensive glucose control (B, 2b).

We did not identify any RCTs or observational studies investigating the relationship between oral hygiene and the incidence of SSI in digestive surgery in English language journals. Only one observational study found that preoperative dental care significantly reduced severe pneumonia after esophagectomy (Fig. 6-4)⁽³³⁴⁾. The evidence level of this retrospective study was low. However, preoperative oral care appears to have become adopted widely in the field of cancer surgery. We decided not to provide a recommendation about preoperative oral care and SSI prevention.

A meta-analysis was conducted with two RCTs^(335,336) (Fig. 6-5). Intraoperative warming and maintaining normothermia significantly reduces the risk of SSI after surgery compared with non-warming care. However, the evidence level is moderate (B) because the two RCTs contained small sample sizes and there have been no recent large-scale studies on digestive surgery. Intraoperative hypothermia causes not only SSI but also postoperative shivering, delayed emergence from anesthesia, and abnormal coagulation. Thus, patients should be warmed and normothermia with a core temperature > 36 °C maintained during surgery using methods such as forced-air warming, warming blankets, and a fluid warmer.

The meta-analysis was conducted on 10 RCTs on digestive surgery^(337−346). High perioperative oxygen concentrations (F_IO₂ of 0.8) do not reduce the risk of SSI significantly after digestive surgery, compared with the usual standard of care (F_IO₂ of 0.3–0.35) (Fig. 6-6). However, when meta-analysis was conducted on four RCTs on high-concentration oxygen administered during and 2 h or more after colorectal surgery^{(337, 339,340,346)}, high F_IO₂ during and 2–6 h after surgery reduced the risk of SSI after colorectal surgery significantly (Fig. 6-7).

The indication for high perioperative F_IO₂ should be evaluated carefully, especially for patients with respiratory diseases such as chronic obstructive pulmonary disease and interstitial pneumonia for whom high oxygen concentrations may exacerbate respiratory failure. In the 10 RCTs on digestive surgery included in this meta-analysis, no harmful injury was reported in the control group, or after the administration of high-concentration oxygen for about 3 h during surgery and up to 6 h after surgery. On the other hand, the effect of high F_IO₂ during long operations has not been verified, and its safety is unknown. The beneficial effect of perioperative high F_IO₂ on SSI prevention is limited, and its safety is unclear. Therefore, the committee members do not recommend it (recommendation level 3). Further research is needed.

A meta-analysis was conducted with seven RCTs on digestive surgery (Fig. 6-8)^(347–353). Early postoperative oral and enteral feeding does not prevent SSI. The meta-analysis did not show that early postoperative oral and enteral feeding was useful for SSI prevention. However, it is an established element of the ERAS protocol and its usefulness for shortening the length of hospital stay and other advantages is well documented. Therefore, the committee members recommend it.

The results of a meta-analysis of eight RCT reports on the combination of hydrocolloid material and silver-containing wound protection material^(355−362) showed that the material for each control varied. Thus, it was considered evidence level B. In the meta-analysis, the use of any protective material reduced the incidence of wound infection significantly (Fig. 7-1). Despite detection bias or technique-related inconsistencies, wound infection is reduced by protective dressings (B, 2b).

NPWT for primary incisional wounds had a significantly lower incidence of SSI than gauze dressing from the meta-analysis of four RCTs^(363−366) (Fig. 7-2). Seroma formation also tended to be lower with NPWT although the difference was not significant (Fig. 7-3). Thus, NPWT may reduce the incidence of SSI in primary incisional wounds. This recommendation has been made based on consideration that the indication of the disease and the type of surgical procedures varied. The negative pressure and the period were also unstable. Furthermore, Japanese public insurance does not yet cover NPWT for primary incisional wounds.