A New Gold Standard? Impact of Broad-spectrum Penicillin-based Antibiotic Prophylaxis on Outcome After Pancreatoduodenectomy: Results of a Systematic Review and Meta-analysis (PROSPERO CRD42024559197)

Abstract

Objective:

This review evaluated whether broad-spectrum penicillin-based antibiotic prophylaxis (BS-AB) such as piperacillin-tazobactam might lead to better outcomes in pancreatoduodenectomy compared with standard care antibiotics, mainly cephalosporins (CE-AB).

Background:

Pancreatoduodenectomy is commonly associated with high postoperative infectious complications, contributing to increased morbidity, mortality, and health care costs.

Methods:

A systemic literature search (PubMed, EMBASE, Cochrane Library, and Web of Science) was conducted to identify suitable RCTs and non-RCTs. After inclusion, the data were analyzed using a random-effects model with the Mantel-Haenszel model or inverse variance to calculate odds ratio (OR) or mean difference (MD) with 95% confidence intervals (CI).

Results:

One RCT and 11 non-RCTs were included with 12,469 patients (35.3% BS-AB, 64.7% CE-AB). Surgical site infections (SSI) were significantly lower after BS-AB when compared with CE-AB, [OR 0.53; CI (0.32 to 0.86); P=0.01; I ²=79%] as well as the occurrence of postoperative pancreatic fistula (POPF) [OR 0.62; CI (0.47 to 0.81); P<0.01; I ²=0%], days of hospitalization [MD -2.02; CI (-4.08 to 0.03); P=0.05; I ²=98%] and mortality [OR 0.56; CI (0.34 to 0.95); P=0.03; I ²=0%]. Subgroup analyses of patients with preoperative biliary drainage demonstrated an even higher effect of BS-AB in reducing SSI [OR 0.45, CI (0.45 to 0.67); P=0.01; I ²=78%], POPF [OR 0.52; CI (0.36 to 0.75); P<0.01; I ²=0%] and mortality [OR 0.34; CI (0.15 to 0.76); P<0.01; I ²=0%].

Conclusion:

BS-AB significantly reduces the risk of infectious complications and surgical outcomes in pancreatoduodenectomy compared with CE-AB, particularly in patients with preoperative biliary drainage. These findings support the use of BS-AB as a new gold standard for patients undergoing pancreatoduodenectomy.

Pancreatoduodenectomy for benign or malignant lesions has substantially advanced over the past 2 decades due to improvements in surgical techniques and health care quality.^1–3 Nevertheless, it remains associated with a notable risk of postoperative morbidity and mortality affecting patient well-being and economic costs.^4–7 Although in-hospital mortality is under 4% in high-volume centers,⁸ postoperative infectious complications continue to remain a critical risk factor following pancreatoduodenectomy.^9,10

Perioperative antibiotic prophylaxis is generally recommended by the American Society of Health-System Pharmacists guidelines to reduce nosocomial infections and other complications in abdominal surgery.¹¹ In pancreatoduodenectomy, postoperative pancreatic fistula (POPF) occurs in 15% to 20% of patients after pancreatic head resections.^12–14 While the use of neoadjuvant therapy for advanced tumors may reduce the rate of POPF, it remains the primary root cause of mortality after pancreatic surgery.^15,16 Another major contributor to postoperative morbidity is surgical site infections (SSI), which occur in 6% to 17%.^17,18 SSIs are often linked to intra-abdominal infections from POPF due to inflammatory tissue reaction and bacterial growth.^19,20 In addition, biliary contamination after preoperative bile duct stenting is also associated with increased abdominal collections and the occurrence of resistant bacteria.^21–23 This is due to the upstream migration of bacteria from the intestinal tract to the biliary tree in stent-bearing patients, leading to prolonged antibiotic treatment and the development of antimicrobial-resistant microorganisms.^24,25

Antibiotic prophylaxis in elective abdominal surgery is the standard of care to reduce postoperative infectious complications, and pancreatoduodenectomy is no exception.^26,27 However, given the high rates of SSI and the rising prevalence of resistant bacteria, the identification of the optimal antibiotic regimen and its duration is still under investigation.^28,29 First-generation and second-generation cephalosporins, along with other commonly used antibiotics, have been routinely used in pancreatoduodenectomy and were recommended by ASHP therapeutic guidelines as standard care antibiotics.³⁰ However, these antimicrobials fail to cover common pathogens detected in bile fluid, such as Enterococcus or Klebsiella.^29,31 Recent studies indicate, therefore, that a broad-spectrum penicillin-based antibiotic treatment as superior regarding infectious complications, predominantly caused by Enterococcus and Enterobacter, in contrast to standard care antibiotics.³² In addition, some studies suggest that a prolonged duration of perioperative antibiotic treatment is associated with fewer infectious complications after pancreatic head resections, while others have shown no significant difference related to the duration of antibiotic treatment.^33,34 However, increasing prescription of antibiotics may add to the development of antimicrobial-resistant microorganisms and subsequent treatment resistance with implications on overall mortality and health care costs.^26,35

The aim of this systematic review and meta-analysis was to primarily gain evidence on broad-spectrum penicillin-based antibiotics versus standard care antibiotics, mainly cephalosporins (CE-AB), for surgical outcomes in pancreatoduodenectomy while evaluating their impact on postoperative complications, duration of hospitalization and mortality. In addition, the study aimed to investigate the effects in subgroups of stent-bearing patients and the influence of antibiotic treatment duration on these outcomes.

METHODS

This systematic review was conducted according to the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines³⁶ and as outlined in a predefined protocol (PROSPERO CRD42024559197). All stages of study selection, data abstraction, and quality assessment were carried out independently by 2 reviewers (S.W. and D.H.). Any disagreements were resolved by consulting a third reviewer (J.C.H. and T.H.).

Literature Search

The following electronic bibliographic databases were searched: MEDLINE (through PubMed), Web of Science, EMBASE, and Cochrane Central Register of Controlled Trials (CENTRAL). The search also included reference lists of relevant articles, as well as citing references in Web of Science. The search strategy was built using a combination of the following key search terms and synonyms as index or free text words and MeSH terms. We only included studies in English. No other restrictions were applied. The search was conducted until March 31, 2024. The search strategy was utilized as outlined in the PROSPERO protocol (CRD42024559197).

Study Selection

Both randomized controlled trials (RCT) and nonrandomized prospective and retrospective studies were used for the qualitative and quantitative synthesis of the systematic review. To determine the publications to be included, 2 researchers scanned the titles and abstracts of all studies retrieved according to the inclusion criteria. Full texts of all relevant articles were then further reviewed. Any disagreement was resolved by consensus or by consultation with a third reviewer.

Outcome Measures

As the primary outcome, surgical site infections (SSI) were assessed. According to the classification by the Center for Disease Control and Prevention, SSI were categorized into superficial, deep incisional, and organ/space SSI.³⁷ As secondary outcomes, the perioperative morbidity and mortality rate, as well as the incidence of clinically relevant POPF Grades B and C^38,39 and postpancreatectomy hemorrhage (PPH) were analyzed.

In addition, further nosocomial infectious complications such as incidence of bacteremia, pneumonia, and urinary tract infections, unplanned readmissions, reoperations, and the duration of hospitalization were assessed. Subgroup analyses for patients with preoperative biliary drainage (PBD) and duration of antibiotic therapy were performed.

Data Extraction

Data were extracted from the trials that met our final inclusion criteria using a standardized form. The form was piloted in the first 3 trials and revised accordingly. Two authors independently extracted data using the standardized form, and any queries/discrepancies were resolved by a third party. The data extraction sheet comprised the following predefined items: (i) study identifier (first author and year of publication); (ii) essential study data (country of study conduct, study design, mono-/multicenter, recruitment and follow-up period, treatment arms, number of subjects); (iii) baseline characteristics of study subjects (mean age, sex, type of disease, etc.); (iv) quality features. Finally, the outcome parameters described above were extracted for individual treatment groups as far as reported. The baseline comparability of the different treatment groups was evaluated.

Assessment of Risk of Bias

The methodological quality of included RCTs was assessed independently by 2 review authors (S.W. and D.H.) using the Cochrane Collaboration tool for assessing risk of bias 2.0.⁴⁰ The tool includes 5 standard domains of bias: bias due to confounding, bias due to deviations from intended interventions, bias due to missing outcome data, bias in measurement of the outcome and bias in selecting of the reported result. These domains were rated as high, low risk of bias, some concerns, or unclear. Finally, an overall risk of biased judgment was presented.

For non-RCTs, the ROBINS-I tool (risk of bias in nonrandomized studies - of Interventions)⁴¹ was used. This tool includes 7 standard domains of bias: bias due to confounding, bias due to selection of participants, bias in classification of interventions, bias due to deviations from intended interventions, bias due to missing outcome data, bias in measurement of the outcome, and bias in selecting of the reported result. These domains were rated as low, high, serious, or critical risk of bias. Also, in this device, an overall assessment of bias was conducted at the end. Furthermore, for each outcome the certainty of the evidence was rated to be low, moderate, serious or critical for each outcome using the GRADE system.⁴² This includes limitations in the design from the risk of bias assessment, indirectness of evidence, unexplained heterogeneity or inconsistency of results, imprecision of results, and publication bias (Supplemental Table 1, Supplemental Digital Content 1, http://links.lww.com/SLA/F439).

Statistical Analysis

For all outcomes, a random-effect model was applied to account for methodological and clinical differences. Statistical heterogeneity among the effect estimates of the included trials was evaluated using the I ² statistic. We considered an I ² less than 25% to indicate low heterogeneity and an I ² >75% to indicate high heterogeneity.

Odds ratios (OR) and 95% CI were pooled for dichotomous outcomes using the Mantel–Haenszel (M-H) method. Continuous outcomes were pooled as mean difference and 95% CI with the inverse variance method. If trials only reported medians or ranges, the methods described by Hozo, 2005⁴³ were applied to calculate means and SD from the values reported.

Primary statistical analysis and meta-analysis were performed with program R.⁴⁴ Forest plots were used for the graphical presentation of (overall) effect estimates.

RESULTS

A total of 2813 references were identified through database searching, as shown in the PRISMA flow diagram (Fig. 1). No more references were identified by searching lists of retrieved studies. Fifteen full-text publications were finally assessed for eligibility, of which twelve were included for quantitative synthesis.^{18,32,33,45–56}

FIGURE 1.

Prisma flow diagram showing article selection process for review.

Study Characteristics

Study characteristics are presented in Table 1, along with the specific antibiotic regimens used. The analysis included 1 RCT and 11 non-RCTs from the United States (n=4), Japan (n=3), China (n=1), France (n=3), and Italy (n=1). Of these studies, 2 were prospective^46,48 and 10 were retrospective.^{18,33,45,47,49,50,53–56} Given the limited number of RCTs, both RCTs and non-RCTs were pooled for analysis. Whenever feasible, analyses were stratified according to the duration of antimicrobial therapy.

TABLE 1.

Study Characteristics

					Antibiosis		Duration of administration		Participants
References	Country	Study design	Diagnosis	Type of surgery	BS-AB	CE-AB	BS-AB	CE-AB	N (all)	BS-AB	CE-AB
Donald et al18	US	Non-RCT	NR	PD	Pip/Taz (if allergic: tigecycline)	Cefoxitin (62%), cefazolin + metronidazole (15%), clindamycin (8%)	PAP	PAP	140	106	34
Sourrouille et al33	France	Non-RCT	Malignant lesions	PD	Combination of gentamicin and Pip/Taz	Cefoxitin	Until POD 2, if bile culture negative; if positive: until POD 5	PAP	175	99	76
Kondo et al45	Japan	Non-RCT	Malignant lesions (95) and others (21)	PD	Cefmetazole + gentamicin (7), piperacillin (51)	Cefmetazole (55), cefazolin (2), sulbactam/cefoperazone (1)	POD 2	POD 2	116	58	58
Tanaka et. al49	Japan	Non-RCT	Pancreatic cancer (50%)	PD	VCM + Pip/Taz	Cefmetazole	POD 1	POD 1	72	32	40
Degrandi et al50	France	Non-RCT	Pancreatic cancer (94%)	PD	Pip/Taz	Cefazolin (if allergic: gentamycin +clindamycing)	POD 5	PAP	122	69	53
De Pastena et al46	Italy	Non-RCT	Peri-ampullary lesions	PD (86%)/ TP(14%)	Pip/Taz	Ampicillin-sulbactam	PAP	PAP	679	296	383
Kone et al247	US	Non-RCT	NR	PD	NSQIP broad-Abx	1st CE, 2nd-3rd CE	NR	NR	9136	3255	5881
Fromentin et al53	France	Non-RCT	Malignant lesions	PD	Pip/Taz	Cefoxitine / Cefamandole /Cefazoline	POD 5	PAP	308	81	227
Asukai et al54	Japan	Non-RCT	NR	PD	Pip/Taz	Cefazoline	POD 3; extended to POD 5 if preoperative infection	POD 3	349	153	196
D´Angelica et al48	US	RCT	NR	PD	Pip/Taz	Cefoxitin	PAP	PAP	778	378	400
Ocuin et al55	US	Non-RCT	Malignant lesions	PD	Pip/Taz; converted to oral amoxicillin/clavulanic acid at discharge	Cefoxitin	POD 10	PAP	87	34	53
Yang et al56	China	Non-RCT	PDAC (26.6%)/non-PDAC (73.4%)	PD	Pip/Taz	Ceftriaxone	POD 5	POD 3	507	267	240

Broad-Abx indicates broad-spectrum antibiotics according to NSQIP definition; BS-AB, broad-spectrum penicillin-based antibiotics; CE-AB, cephalosporin antibiotics; d, days; NR, not reported; OP, operation; PAP, perioperative antibiotic prophylaxis; PAP, perioperative antibiotic prophylaxis; PBD, preoperative biliary drainage; PD, pancreaticoduodenectomy; Pip/Taz, Piperacillin/Tazobactam; POD, postoperative day; RCT, randomized controlled trial; antibiotics are applied intravenously unless otherwise declared; TP, total pancreatectomy; VCM, vancomycin.

All 12 studies involved patients undergoing pancreatoduodenectomy. One trial included patients undergoing total pancreatectomy in 14% of cases (n=95).⁴⁶ A total of 12,469 individuals received either penicillin-based BS-AB (n=4828) or CE-AB (n=7641). Sample sizes ranged from 72 to 9136 patients. Table 2 displays the baseline characteristics of the study populations (age, sex, ASA, risk factors).

TABLE 2.

Characteristics and Risk Factors of Participants

	Overall			Age (median)		Sex (M:F)		ASA <3/≥3		BMI (kg/m2) (median)		Diabetes (%)		Smoking (%)		Neoadjuvant therapy (%)
References	Participants	BS-AB	CE-AB	BS-AB	CE-AB	BS-AB	CE-AB	BS-AB	CE-AB	BS-AB	CE-AB	BS-AB	CE-AB	BS-AB	CE-AB	BS-AB	CE-AB
Donald et al18	140	106	34	63.3 (mean)	63.4 (mean)	49:57	12:22	44/62	12/22	25.7 (mean)	27.0 (mean)	31 (29.4)	6 (18.2)	9 (8.8)	4 (12.9)	NR	NR
Sourrouille et al33	175	99	76	61	64	58:41	34:42	87/12	69/7	24	23	15 (15.2)	12 (15.8)	NR	NR	NR	NR
Kondo et al45	116	58	58	69.2 (mean)	70.5 (mean)	35:23	30:28	52/6	53/5	21.7 (mean)	22.1 (mean)	24 (41.4)	19 (32.8)	NR	NR	NR	NR
Tanaka et al49	72	32	40	65.5	72	18:14	24:16	30/2	37/3	23.0 (mean)	21.5 (mean)	NR	NR	NR	NR	11 (34.4)	19 (47.5)
Degrandi et al50	122	69	53	68	66	42:27	27:26	NR	NR	24	24	11 (16)	11 (21)	28 (41)	22 (42)	19 (27)	15 (28)
De Pastena et al46	679	296	383	65 (mean)	66 (mean)	163:133	223:160	233/63	302/81	24 (mean)	24 (mean)	NR	NR	NR	NR	85 (28)	87 (23)
Kone et al47	9136	3255	5881	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR
Fromentin et al53	308	81	227	66	136	50:31	128:99	NR	NR	23	24	NR	NR	NR	NR	15 (19)	9 (4)
Asukai et al54	349	153	196	69	67	85:68	112:84	NR	NR	21.8	21.2	NR	NR	NR	NR	67 (43.8)	112 (57.4)
D´Angelica et al48	778	378	400	66.8	68.0	233:145	223:177	63/315	70/330	26.2 (mean)	26.5 (mean)	111 (29.4)	93 (23.3)	58 (15.3)	65 (16.3)	143 (37.8)	130 (32.5)
Ocuin et al55	87	34	53	64	63	13:21	28:25	NR	NR	NR	NR	7 (20.6)	12 (22.6)	NR	NR	9 (26.5)	8 (15.1)
Yang et al56	507	267	240	61.7 (mean)	61.7 (mean)	NR	NR	NR	NR	23.4 (mean)	23.4 (mean)	31 (14.4)	39 (20.3)	NR	NR	0 (0)	0 (0)

BMI indicates body mass index; BS-AB, broad-spectrum penicillin-based antibiotics; CE-AB, cephalosporin antibiotics; NR, not reported.

Critical Appraisal of Included Studies

Assessment of Risk of Bias

The comprehensive Risk of Bias analysis revealed that all included non-RCTs exhibited the greatest concerns in the domain of Bias due to missing data (ranging from serious to critical). However, most other domains were assessed to have a moderate risk of bias. The study by De Pastena⁴⁶ presented the highest overall risk, primarily due to numerous missing data points, issues with confounding, and deviations from intended interventions. In contrast, the studies by Ocuin⁵⁵ and Yang⁵⁶ were rated as having a moderate risk of bias, benefiting from robust participant selection and favorable performance across additional Cochrane criteria. The RCT by D’Angelica⁴⁸ demonstrated an overall low risk of bias, characterized by effective randomization, minimal missing data, and few deviations from intended interventions (Supplemental Table 2 and 3, Supplemental Digital Content 1, http://links.lww.com/SLA/F439).

Primary Outcomes

Surgical site infections (Overall). For overall SSI, 12 trials reported the incidence with a total of 12,469 patients.^{18,33,45–50,53–56} The number of SSI significantly differed [BS-AB vs. CE-AB: 972/4828 (20.1%) vs. 2022/7641 (26.5%)], favoring BS-AB over the CE-AB group [OR 0.46; 95% CI (0.33 to 0.63); P<0.01; I ²=78%; high level of certainty; Fig. 2].

FIGURE 2.

Forest plot comparing overall surgical site infections with broad-spectrum antibiotics (BS-AB) versus standard care antibiotics (CE-AB). BS-AB indicates broad-spectrum penicillin-based antibiotics; CE-AB, cephalosporin antibiotics.

Five trials reported the overall SSI incidence following perioperative antibiotic prophylaxis (PAP) with a total of 10,849 patients.^18,45–48 The number of SSI significantly differed [BS-AB vs. CE-AB: 830/4093 (20.3%) vs. 1700/6756 (25.2%)], favoring BS-AB over the CE-AB group [OR 0.53; 95% CI (0.32 to 0.86); P=0.01; I ²=79%; high level of certainty, Supplemental Figure 1, Supplemental Digital Content 1, http://links.lww.com/SLA/F439].

In the prolonged antibiotic therapy (PAT) group, 7 studies^{33,49,50,53–56} reported overall SSI in a total of 1620 patients, again significantly favoring BS-AB with a higher level of certainty in prolonged treatment [BS-AB vs. CE-AB: 142/735 (19.3%) vs. 322/885 (36.4%); OR 0.41; 95% CI (0.28 to 0.60); P<0.01; I ²=52%; high level of certainty, Supplemental Figure 2, Supplemental Digital Content 1, http://links.lww.com/SLA/F439].

Superficial surgical site infections. Based on 11 trial involving 12,161 patients,^{18,33,45–50,54–56} the incidence of superficial SSI was significantly lower with BS-AB compared with CE-AB [BS-AB vs. CE-AB: 265/4747 (5.6%) vs. 612/7414 (8.3%); OR 0.44; 95% CI (0.28 to 0.68); P<0.01; I ²=68%; high level of certainty; Suppl. Figure 3, Supplemental Digital Content 1, http://links.lww.com/SLA/F439].

Five trials reported the incidence of superficial SSI following PAP, with a total of 10,849 patients included.^18,45–48 The number of superficial SSI significantly differed [BS-AB vs. CE-AB: 235/4093 (5.7%) vs. 548/6756 (8.1%)] favoring the BS-AB group over the CE-AB group [OR 0.38; 95% CI (0.20 to 0.74); P<0.01; I ²=82%; high level of certainty, Supplemental Figure 4, Supplemental Digital Content 1, http://links.lww.com/SLA/F439].

In the PAT group, 6 studies reported superficial SSI with a total of 1312 patients,^{33,49,50,54–56} again significantly favoring BS-AB in prolonged treatment [BS-AB vs. CE-AB: 30/654 (4.6%) vs. 64/658 (9.7%); OR 0.51; 95% CI (0.26 to 0.97); P=0.04; I ²=45%; high level of certainty, Supplemental Figure 5, Supplemental Digital Content 1, http://links.lww.com/SLA/F439].

Deep Incisional Surgical Site Infections

For deep SSI, five trials reported the incidence, with a total of 1994 patients included.^{45,46,48,49,54} The number of deep SSI was without significant difference [BS-AB vs. CE-AB: 6/917 (0.7%) vs. 20/1077 (1.9%); OR 0.44; 95% CI (0.13 to 1.47); P=0.18; I ²=32%; high level of certainty]. Due to the limited number of patients per subgroup, there was no stratification according to the duration of antibiotic treatment (Supplemental Fig. 6, Supplemental Digital Content 1, http://links.lww.com/SLA/F439).

Organ/Space Surgical Site Infections

For organ/space SSI, 11 trials reported the incidence, with a total of 12,161 patients included.^{18,33,45–50,54–56} The number of organ/space SSI significantly differed favoring BS-AB over CE-AB [BS-AB vs. CE-AB: 696/4747 (14.7%) vs. 1422/7414 (19.2%); OR 0.54 95% CI (0.39 to 0.75); P=<0.01; I ²=69%; high level of certainty, Figure 3].

FIGURE 3.

Forest plot comparing organ/space surgical site infections with broad-spectrum antibiotics (BS-AB) versus standard care antibiotics (CE-AB). BS-AB indicates broad-spectrum penicillin-based antibiotics; CE-AB, cephalosporin antibiotics.

Five trials reported the incidence following PAP, with a total of 10,849 patients included.^18,45–48 The number of organ/space SSI significantly differed [BS-AB vs. CE-AB: 606/4093 (14.8%) vs. 1.247/6756 (18.5%), favoring the BS-AB group over the CE-AB group [OR 0.69; 95% CI (0.49 to 0.97); P=0.03; I ²=68%; high level of certainty, (Supplemental Fig. 7, Supplemental Digital Content 1, http://links.lww.com/SLA/F439)].

In the PAT group, 6 studies^{33,49,50,54–56} reported organ/space SSI in a total of 1312 patients, again significantly favoring BS-AB with a higher level of certainty in prolonged treatment [BS-AB vs. CE-AB: 90/654 (13.8%) vs. 175/658 (26.6%); OR 0.44; 95% CI (0.30 to 0.64); P<0.01; I ²=37%; high level of certainty, Supplemental Fig. 8, Supplemental Digital Content 1, http://links.lww.com/SLA/F439].

Secondary Outcomes

Mortality. Eight trials^{33,46,48–50,53–55} reported the mortality rate with a total of 2570 patients included. The overall mortality rate was 1.8% (21 out of 1142) in the BS-AB group compared with 3.9% (56 out of 1428) in the CE-AB group. The mortality significantly differed, favoring the BS-AB group over the CE-AB group [OR 0.56; CI (0.34 to 0.95); P=0.03; I ²=0%; high level of certainty, Supplemental Fig. 9, Supplemental Digital Content 1, http://links.lww.com/SLA/F439].

Postpancreatectomy specific complications. Ten trials reported the incidence of clinically relevant POPF with a total of 3046 patients included.^{33,45,46,48–50,53–56} The POPF rate significantly favored BS-AB over CE-AB [BS-AB vs. CE-AB: 217/1428 (15.2%) vs. 340/1618 (21%); OR 0.71; CI (0.59 to 0.87); P<0.01; I ²=26%; high level of certainty, Figure 4].

FIGURE 4.

Forest plot comparing postoperative pancreatic fistula with broad-spectrum antibiotics (BS-AB) versus standard care antibiotics (CE-AB). BS-AB indicates broad-spectrum penicillin-based antibiotics; CE-AB, cephalosporin antibiotics.

The POPF rate significantly differed, favoring the BS-AB group over the CE-AB group, even after PAP, which was reported by 3 trials,^45,46,48 with 1573 included patients [OR 0.62; CI (0.47 to 0.81); P<0.01; I ²=0%; moderate level of certainty], but not following PAT, as assessed by 7 trials,^{33,49,50,53–56} including 1473 patients [OR 0.82; CI (0.63 to 1.08); P=0.16; I ²=40%; high level of certainty, Supplemental Fig. 10, 11, Supplemental Digital Content 1, http://links.lww.com/SLA/F439].

PPH was analyzed by 6 trials^{33,50,53–56} with a total of 1401 patients included in the PAT group. The PPH rate did not differ significantly between the groups [BS-AB vs. CE-AB: 53/664 (8.0%) vs. 74/737 (10.0%); OR 1.01; CI (0.68–1.51); P=0.96; I ²=0%; high level of certainty, Supplemental Fig. 12, Supplemental Digital Content 1, http://links.lww.com/SLA/F439].

Nosocomial infections. Bacteremia was reported in 4 trials^33,48,50,56 with a total of 1582 patients included. The incidence of bacteremia showed a significant reduction by using BS-AB [OR 0.36; CI (0.17 to 0.78); P<0.01; I ²=64%; high level of certainty, Supplemental Fig. 13, Supplemental Digital Content 1, http://links.lww.com/SLA/F439]. The incidence of pneumonia was also significantly reduced in the BS-AB group [OR 0.45; 95% CI (0.27 to 0.78); P < 0.01; I ²=39%; moderate level of certainty], as documented by 5 studies^{33,46,48,50,56} with a total of 2261 patients (Supplemental Fig. 14, Supplemental Digital Content 1, http://links.lww.com/SLA/F439). The incidence of urinary tract infections also significantly decreased with BS-AB as shown in 5 studies^{33,46,48,50,56} including 2261 patients [OR 0.39; 95% CI (0.19 to 0.82); P=0.01; I ²=31%; high level of certainty, Supplemental Fig. 15, Supplemental Digital Content 1, http://links.lww.com/SLA/F439].

Readmissions. Three trials^46,48,55 involving a total of 1544 patients recorded the rate of readmissions. The analysis showed no significant difference in readmission rates between the antibiotic prophylaxis regimens [OR 0.92; CI (0.68 to 1.24); P=0.56; I ²=0%; very low level of certainty, Supplemental Fig. 16, Supplemental Digital Content 1, http://links.lww.com/SLA/F439].

Unplanned reoperation. Five trials^{46,48,50,53,55} involving a total of 1974 patients reported the incidence of unplanned reoperations. The use of BS-AB significantly reduced the rate of unplanned reoperations [OR 0.61; CI (0.38 to 0.96); P=0.03; I ²=14%; high level of certainty, Supplemental Fig. 17, Supplemental Digital Content 1, http://links.lww.com/SLA/F439].

Duration of hospitalization. Based on reports from 6 trials^{33,46,48–50,55} involving a total of 1913 patients, the duration of hospitalization was significantly reduced by choice of antibiotic prophylaxis favoring BS-AB over CE-AB [MD −2.02; CI (−4.08 to 0.03); P=0.05; I ²=98%; very low level of certainty, Supplemental Fig. 18, Supplemental Digital Content 1, http://links.lww.com/SLA/F439].

Further Subgroup Analyses

Preoperative biliary drainage. A total of 6151 patients, accounting for 33.4% of the entire population, underwent PBD. This subgroup was divided between 2427 patients in the intervention group and 3724 in the control group. Within the group of patients having received a PBD, BS-AB significantly reduced the incidence of all nearly analyzed endpoints compared with CE-AB at a high level of certainty.

Specifically, BS-AB was associated with a significant reduction in overall SSI^{47,48,50,53,56} in a total of 6151 patients [OR 0.45; 95% CI (0.30 to 0.67); P<0.01; I ²=78%; high level of certainty, Fig. 5], the incidence of organ/space SSI^47,48,50,56 [OR 0.39; CI (0.33 to 0.47); P<0.01; I ²=0%; high level of certainty, Supplemental Fig. 19, Supplemental Digital Content 1, http://links.lww.com/SLA/F439], the incidence of POPF^48,50,53,56 [OR 0.52; CI (0.36 to 0.75); P<0.01; I ²=0%; moderate level of certainty, Supplemental Fig. 20, Supplemental Digital Content 1, http://links.lww.com/SLA/F439], unplanned reoperations^48,50,53 [OR 0.24; CI (0.10 to 0.55); P<0.01; I ²=48%; moderate level of certainty, Supplemental Fig. 21, Supplemental Digital Content 1, http://links.lww.com/SLA/F439], and mortality^48,50,53 [OR 0.34; CI (0.15 to 0.76); P<0.01; I ²=0%; high level of certainty, Supplemental Fig. 22, Supplemental Digital Content 1, http://links.lww.com/SLA/F439]. For superficial SSI, no significant difference was found between BS-AB and CE-AB^47,48,50,56 [OR 0.75; CI (0.29 to 1.92); P=0.54; I ²=84%; very low level of certainty] in a total of 6005 patients (Supplemental Fig. 23, Supplemental Digital Content 1, http://links.lww.com/SLA/F439).

FIGURE 5.

Forest plot comparing overall surgical site infections in patients with a biliary stent with broad-spectrum antibiotics (BS-AB) versus standard care antibiotics (CE-AB). BS-AB indicates broad-spectrum penicillin-based antibiotics; CE-AB, cephalosporin antibiotics.

DISCUSSION

This meta-analysis demonstrates that penicillin-based BS-AB significantly reduces SSIs, postoperative infectious complications, pancreas-specific complications such as clinically relevant POPF, unplanned operations, and mortality compared with standard care antibiotics (CE-AB), especially in patients with PBD. The broader antibacterial spectrum of BS-AB appears to address key pathogens more effectively, reducing infection-related complications and the incidence of POPF. These findings suggest that BS-AB provides a distinct advantage in preventing complications associated with pancreatoduodenectomy, particularly in high-risk groups such as PBD patients, where the microbial burden is higher and traditional CE-AB may be insufficient.

Hospital-acquired infections, particularly deep and organ/space SSI, are a major contributor to the extended length of hospital stay, increased mortality, and higher health care costs, especially in pancreatic head surgery.^57–59 Addressing these complications requires a comprehensive approach, including advancements in surgical techniques, wound protectors, and strategies such as antiseptic wound irrigation.^60,61 Among these, the choice of antibiotic prophylaxis plays a pivotal role. BS-AB offers distinct advantages over cephalosporin-based regimens by targeting a wider array of pathogens, particularly those commonly present after pancreatoduodenectomy.²⁴ For instance, BS-AB regimens like piperacillin-tazobactam provide superior coverage against Gram-negative organisms, anaerobes, and certain Gram-positive organisms like Enterococcus and Enterobacter, effectively addressing gaps left by CE-AB.^20,32 Strikingly, in our study, SSI were significantly lower in the BS-AB group, alongside reductions in other infectious complications, such as urinary tract infections, which collectively translate into shorter hospital stays and lower rates of unplanned reoperations. In addition, this meta-analysis demonstrates that BS-AB is less frequently associated with severe conditions like hospital-acquired pneumonia and bacteriaemia. Although these organisms play a minor role in such infections, the findings highlight a broader effect of BS-AB over CE-AB in reducing overall postoperative infection risks.

Beyond non–pancreas-specific complications, this meta-analysis also highlights the significant impact of BS-AB on pancreas-specific outcomes, particularly with the incidence of clinically relevant POPF. A previous study from our center demonstrated that infections caused by E. coli are strongly associated with the development of severe POPF after pancreatoduodenectomy, emphasizing the role of bacterial contamination in the progression of pancreatic fluid collections to clinically significant complications.¹⁹ Moreover, recent findings indicate that Enterococcus faecalis, through its collagen-degrading enzymes and ability to activate tissue matrix metalloproteinase 9, contributes to anastomotic leakage in colorectal surgery, which may also play a role in the occurrence of POPF.⁶² Strikingly, Ellis et al demonstrated that cefoxitin-resistant bacteria, including Enterobacter and Enterococcus, were associated with a higher incidence of POPF when patients were treated with cefoxitin rather than penicillin-based antibiotics (piperacillin-tazobactam).³² Notably, in this meta-analysis, the use of BS-AB, even as single-shot or prolonged prophylaxis, was associated with a significant reduction in the incidence of POPF (OR: 0.71). This suggests that the broader antibacterial coverage of penicillin-based antibiotics, including efficacy against both Gram-positive organisms like Enterococcus and Gram-negative organisms such as Enterobacter, E. coli and others frequently isolated from bile cultures, plays a pivotal role in diminishing these risks early on. While the incidence of PPH did not differ significantly between BS-AB and CE-AB regimens, the prevention of clinically relevant POPF through improved bacterial control is likely to mitigate downstream complications, including bleeding events. These findings underscore the importance of selecting an antibiotic regimen that addresses both the general and pancreas-specific infectious risks in pancreatoduodenectomy, ultimately improving patient safety and postoperative recovery.

Several studies have shown that PBD is associated with an increased risk of severe infectious complications due to bacterial contamination of bile fluid.^63,64 This contamination often leads to resistant bacteria, particularly in patients with prolonged stent placement or repeated endoscopic interventions, which potentially increase postoperative infections and contribute to worse outcomes.⁶⁵ Gram-negative pathogens and Enterococcus are routinely isolated from bile cultures in PBD patients, posing a significant task for perioperative infection control.²³ In this meta-analysis, BS-AB demonstrated a significant reduction in overall SSI, organ/space infections, and pancreas-specific complications such as POPF in stent-bearing patients compared with CE-AB. The superior efficacy of BS-AB in this subgroup can be attributed to their broader coverage of resistant bacteria, which is critical in covering the higher infection risk associated with PBD. These findings are in line with prior studies indicating that the selective pressure utilized by biliary stents increases the likelihood of resistant pathogens and limits the efficacy of traditional prophylactic antibiotics, such as cephalosporins.^21,24 By effectively addressing these resistant bacteria, BS-AB reduces the infectious burden and prevents secondary complications, such as organ/space SSI and bacteremia, which are particularly common in the high-risk group of PBD patients.

Despite the meta-analysis not identifying a clear trend regarding the optimal duration of antibiotic therapy in individuals with or without PBD, there is a lack of robust data from randomized controlled trials supporting extended durations. However, recent studies suggest that extending perioperative antibiotic therapy up to 5 days can reduce postoperative complications, including reoperations and intensive care stay.^33,66 On the other hand, studies like the STOP-IT Trial⁶⁷ indicate that prolonged antibiotic therapy does not provide additional benefits over shorter durations for intra-abdominal infections, though its applicability to elective pancreatoduodenectomy may be limited. Therefore, identifying specific high-risk patients who could benefit from prolonged therapy remains crucial in future studies.

Limitations of this meta-analysis include the heterogeneity of the included studies, variability in the antibiotic prophylaxis regimens, and the duration of antibiotic therapy, which affects the validity of the results. In addition, the retrospective nature of most studies, with only 1 randomized controlled trial, should be noted. Nevertheless, current data support the recommendation of broad-spectrum prophylaxis, such as piperacillin/tazobactam, to reduce postoperative complications and mortality following pancreatoduodenectomy. This highlights the ongoing challenge of balancing antibiotic efficacy with the risks of antibiotic resistance, particularly in high-risk patients.

CONCLUSIONS

Broad-spectrum penicillin-based antibiotic prophylaxis, compared with standard care antibiotics, primarily cephalosporins (CE-AB), significantly reduces postoperative infectious complications, clinically relevant POPF, unplanned reoperations, and mortality in pancreatoduodenectomy. With the increasing number of patients undergoing neoadjuvant treatments and preoperative biliary drainage, BS-AB prophylaxis has also proven to be an effective approach, reducing the incidence of pancreas-specific complications and morbidity in these emerging groups. Further prospective randomized trials are essential to improve these recommendations and better define optimal treatment strategies.

Supplementary Material

sla-283-418-s001.docx ^{(5MB, docx)}