Are Fewer Damage Control Laparotomies Associated with a Decrease in the Rate and Severity of Surgical Site Infections?

Abstract

Background:

Damage control laparotomy (DCL) rates have declined and remained low. Given DCL’s association with surgical site infections (SSIs), we hypothesize SSI rates and severity, per Clavien-Dindo (CD) classification, have decreased concurrently.

Methods:

A retrospective study was performed of patients (≥16 years) who underwent trauma laparotomy from 01/2011–12/2020. Patients who survived ≤48 hours were excluded. Data were divided into period 1 (1/2011–12/2015, DCL rates declined) and period 2 (1/2016–12/2020, low DCL rates were sustained). SSIs were defined per the Centers for Disease Control and Prevention, and severity was graded by CD scale (I-V). SSIs were classified as minor (CD grade I-II) and major (CD grade III-V). Univariate and multivariable analyses were performed (P<0.05).

Results:

In total 1,975 patients met inclusion criteria. Between periods 1 and 2, there was an increase in Injury Severity Score (ISS); (19 [IQR 10, 29] vs 21 [IQR 12, 34], P=<0.001) and penetrating injuries (426 [44%] vs. 513 [51%], P=0.002), with a decrease in DCL (283 [29%] vs. 153 [15%], P=<0.001). A total of 355 (18%) patients developed a SSI, 18.5% (179) in period 1 and 17.5% (176) in period 2 (P=0.5). On multivariable analysis, after controlling for mechanism of injury, emergency room systolic blood pressure, wound classification, large bowel resection, ISS, splenectomy, and operating-room time at index laparotomy, DCL was associated with major SSIs.

Conclusions:

Despite decreases in DCL, SSI rates and severity after trauma laparotomy remain stable, potentially due to the increased injury severity and penetrating trauma. Identification of other potential modifiable risk factors is needed to decrease SSI incidence and severity.

Introduction

Damage control laparotomy (DCL) was first proposed in 1983 as a technique to manage acidotic, coagulopathic, and hypothermic trauma patients who would benefit from admission to the intensive care unit before definitive fascial closure.^1–2 However, as surgeons became more comfortable managing open abdomens, the use of DCL expanded to include over 30 different indications.³ This has led to the potential overutilization of DCLs,⁴ which have been associated with many complications, including surgical site infections (SSIs).⁵⁻⁶

SSIs are known to increase morbidity, mortality, and healthcare costs and have a significant impact on a patient’s quality of life.^7–9 There are two common ways to classify SSIs: by their location and the level of intervention needed to treat them.^10–12 By location, SSIs are classified into superficial, deep, and organ space.¹⁰ However, this classification does not gauge the impact an infection can have on a patient. The Clavien-Dindo (CD) classification allows for grading complications, such as SSIs, by their severity based on the progressive level of intervention required for treatment on a scale of I-V.^11–12 This classification focuses on the risk and invasiveness of the therapy used to correct a complication.

Extensive resources have been put in place to develop strategies and guidelines to decrease SSI rates¹³ and optimize patients before surgery by addressing modifiable risk factors.¹⁴ However, these options are limited in the trauma setting. Given the association of DCL with SSIs,^6,15 the primary aim of this study is to evaluate if an institutional reduction in the use of DCL over time was accompanied by a concomitant decrease in the rate and severity of SSIs. A secondary aim is to identify factors associated with major (≥CD IIIa) SSIs after emergency trauma laparotomy. We hypothesize that the rate and severity of SSIs have decreased with the reduction in the use of DCL and that DCL is strongly associated with major SSIs.

Materials and Methods

Study Design

A retrospective review of adult trauma patients (≥16 years old) at a level 1 trauma center who underwent emergency trauma laparotomy from January 2011 to December 2020 was performed. Patients who died within 48 hours of arrival were excluded from the study (a total of 189 patients died within the first 48 hours - 88 (8%) patients in period 1 and 101 (9%) patients in period 2 [p = 0.125]), since no SSIs occurred within this time frame and because death is a competing outcome with SSI.¹⁶ From November 2013 to October 2015, an institutional effort was carried out to decrease the rate of DCL where the indication was not widely agreed upon. The intervention included a multifaceted strategy to effect change, including auditing and providing feedback for every DCL case and having a physician champion. This project decreased the rate of DCL from 39% in the control period to 23% in the intervention period.¹⁷ Therefore, the data for this study was divided into period 1 (1/2011–12/2015), immediately before and after the implementation of the project to reduce DCL rate, and period 2 (1/2016–12/2020), when low DCL rates were sustained. The STROBE guidelines were used to ensure the study was reported correctly.¹⁸

The institutional laparotomy database was used to identify patients who underwent an emergency trauma laparotomy during the study periods stated above. Data collected from the medical records included basic demographic information, Injury Severity Scores (ISSs), mechanism of injury, emergency room vital signs, number of units of packed red blood cells given in the operating room, surgical procedures, SSIs, and hospital length of stay. Missing data was minimal, less than 1%, and was excluded from analysis.

Primary Outcome:

The primary outcome of the study was the rate of SSI, as defined by the Centers for Disease Control and Prevention: superficial, deep, and organ space (OS-SSI).¹⁰ Since it was difficult to retrospectively assess if SSIs were superficial, deep, or both, we grouped these two categories as incisional SSIs. SSIs were manually identified and classified by a general surgery resident (C.J.F.) and verified by another (S.M.U.). Patients were followed until discharge from the hospital, or readmission if applicable, and up to 30 days. In addition, SSI data was cross-referenced with the institution data reported to the National Healthcare Safety Network to ensure no SSIs were missed.

Clavien-Dindo Grade Assignment

CD grades were assigned to each SSI as previously described in the literature.¹¹ CD grade I refers to a SSI that involves opening the wound at the bedside for treatment. CD grade II indicates a SSI that was treated with antibiotics alone. CD grade IIIa refers to a SSI that requires a patient to undergo regional or local anesthesia, such as drain placement performed by interventional radiology, for treatment. CD grade V refers to a SSI that results in death.^11–12,19 Furthermore, the CD classification of SSIs were divided into minor and major, with CD grades I and II considered minor, and CD grade ≥IIIa considered major.^11–12

Statistical Analysis

Statistical analysis was performed using Chi-Square or Fisher’s Exact Test for categorical variables. T-test for normally distributed data and Mann-Whitney U test for nonparametric data were performed for continuous variables. Descriptive statistics were reported as medians with interquartile range (IQR). Multivariable logistic regression was used to control for known confounders, based on prior literature, and to assess SSI-related factors associated with major complications. A subgroup analysis of only DCL patients during the two study periods was prespecified. All statistical analyses were performed using Jamovi version 2.5.5 and R core team (2023).

The Institutional Review Board approved this study, IRB number HSC-GEN-13–0325. It complies with the Health Insurance Portability and Accountability Act. The Institutional Review Board waived informed consent for this retrospective cohort study.

Results

A total of 1,975 patients met the inclusion criteria, with 967 patients in period 1 and 1,008 patients in period 2. The median age of the study population was 33 (IQR 23, 46) years old, and 78% of the patients were male. Between periods 1 and 2, there was an increase in ISS (19 [IQR 10, 29] vs 21 [IQR 12, 34], P= <0.001) and penetrating mechanism of injury (426 [44%] vs. 513 [51%], P= 0.002), while there was a decrease in the use of DCL (283 [29%] vs. 153 [15%], P= <0.001) (Table 1). However, despite a decline in the use of DCL (Figure 1.A), the rates of SSIs remained stable between the two time periods (Table 1 and Figure 1.B), 18.5% (179) in period 1 and 17.5% (176) in period 2 (P= 0.543) as well as the severity per CD classification (Table 1). In addition, the length of hospital stay decreased between these two periods (10 [IQR 6, 21] vs. 9.5 [IQR 5, 20], P= 0.039), and mortality remained stable (5% first period vs. 7% second period, P= 0.109) (Table 1).

Table 1.

Patient Characteristics

Variable	All N=1,975	Period 1 (1/2011–12/2015) N= 967	Period 2 (1/2016–12/2020) N=1,008	P value
Age, years (IQR)	33 (23, 46)	33 (24, 47)	33 (23, 46)	0.470
Sex				0.277
Male	1542 (78%)	745 (77%)	797 (79%)
Race/Ethnicity				<0.001
White	696 (35%)	389 (40%)	307 (30%)
Black	528 (27%)	220 (23%)	308 (31%)
Hispanic	539 (27%)	321 (33%)	218 (22%)
Other	211 (11%)	37 (4%)	174 (17%)
Mechanism of injury				0.002
Penetrating	939 (48%)	426 (44%)	513 (51%)
Pre-Hospital SBP, mmHg	119 (97, 138)	118 (98, 138)	120 (120, 140)	0.323
Pre-Hospital HR, beats per minute	99 (82, 118)	98 (82, 115)	100 (83, 120)	0.085
Pre-Hospital GCS	15 (13, 15)	15 (13, 15)	15 (14, 15)	0.258
ED SBP, mmHg	115 (94, 132)	115 (92, 132)	115 (95, 132)	0.398
ED HR, beats per minute	99 (83, 116)	99 (82, 115)	99 (83, 118)	0.468
ED GCS	15 (12, 15)	15 (12, 15)	15 (12, 15)	0.802
Injury severity score	19 (10, 30)	19 (10, 29)	21 (12, 34)	<0.001
OR RBC, units	1 (0, 4)	1 (0, 4)	1 (0, 4)	0.060
Damage control laparotomy	436 (22%)	283 (29%)	153 (15%)	<0.001
Small bowel resection	475 (24%)	196 (20%)	279 (28%)	<0.001
Large bowel resection	429 (22%)	177 (18%)	252 (25%)	<0.001
Splenectomy	420	200 (21%)	220 (22%)	0.535
Wound Class				<0.001
1	421 (21%)	196 (20%)	216 (21%)
2	527 (27%)	291 (30%)	236 (23%)
3	947 (48%)	477 (49%)	470 (47%)
4	88 (4%)	3 (0.3%)	85 (8%)
OR time of index laparotomy in hours (median, IQR)	2.6 (1.6, 3.3)	2.0 (1.4, 2.9)	2.5 (1.8, 3.5)	<0.001
Management of skin after fascial closure				0.426
Skin closed	1,467 (74%)	726 (75%)	741 (74%)
Outcomes
Any SSI	355 (18%)	18.5% (179)	17.5% (176)	0.543
All Incisional SSI (n, %)	151 (7.6%)	83 (8.6%)	68 (6.7%)	0.125
All OS-SSI (n, %)	287 (14.5%)	138 (14.3%)	149 (14.8%)	0.747
Incisional + OS-SSI	83 (4.2%)	42 (3.3%)	41 (4.1%)	0.760
CD classification				0.120
Grade I-II (n, %)	78 (4.0%)	47 (5%)	31 (3%)
Grade IIIa-V (n, %)	277 (14%)	132 (14%)	145 (14%)
Hospital LOS, days (median, IQR)	10 (5, 20)	10 (6, 21)	9.5 (5, 20)	0.039
Mortality	126 (6.4%)	53 (5%)	73 (7%)	0.109

Descriptive statistics for continuous variables are median, first, and third quartiles. Categorical variables are presented as frequency and percentage. IQR = Interquartile range; ED = emergency department; SBP = systolic blood pressure; HR = Heart rate; GCS = Glasgow coma scale; OR = operating room; RBC = red blood cells. SSI: Surgical site infection; Incisional SSI: Superficial and deep surgical site infection; OS-SSI = organ space surgical site infection; CD classification: Clavien-Dindo classification; LOS: Length of stay

Figure 1.

A. Percentage of Damage control laparotomy (DCL) per year. X-axis represents the year; Y-axis represents the percentage of DCLs among all trauma laparotomies performed in that year B. Percentage of Surgical Site Infections (SSIs) per year. X-axis represents the year; Y-axis represents the percentage of SSIs that developed among all trauma laparotomies performed in that year.

Damage Control Laparotomy versus Definitive Laparotomy Patients

A total of 436 (22%) patients underwent DCL. Patients who underwent DCL were more likely to have suffered blunt trauma and had more unstable vital signs at the scene and on arrival at the hospital compared to those patients who underwent definitive laparotomy. DCL patients were also more severely injured and had higher rates of small and large bowel resection compared to patients who had a definitive laparotomy (Table 2). In terms of abdominal closure of DCL patients, 88% received fascial closure, 7% died with an open abdomen, 2% had only their skin close (no fascial closure), 1% required a bridging mesh, and 2% had a split-thickness skin graft (Table 3). In addition, patients who underwent DCL had a shorter operating room time compared to those who underwent definitive laparotomy at their index operation, 2.0 (IQR 1.3, 2.9) vs. 2.4 (IQR 1.7, 2.9) hours, P=<0.001 (Table 2).

Table 2.

Admissions, demographics, and injuries between patients who underwent definitive laparotomy compared to those who underwent damage control laparotomy.

Variables	Definitive laparotomy N= 1,539	DCL N=436	P value
Age, years (IQR)	32 (IQR 23 –46)	36 (IQR 24–49)	0.004
Sex	1,204 (78%)	338 (78%)	0.752
Male
Race/Ethnicity			0.463
White	530 (35%)	166 (38%)
Black	422 (27%)	106 (24%)
Hispanic	420 (27%)	119 (27%)
Other	166 (11%)	45 (10%)
Mechanism of injury			<0.001
Penetrating	779 (51%)	160 (37%)
Pre-Hospital SBP, mmHg	121 (IQR 102 –140)	102 (IQR 82–129)	<0.001
Pre-Hospital HR, beats per minute	96 (IQR 82 – 113)	109 (IQR 90 –126)	<0.001
Pre-Hospital GCS	15 (IQR 14–15)	14 (IQR 6–15)	<0.001
ED SBP, mmHg	120 (IQR 101 –136)	91 (IQR 76–114)	<0.001
ED HR, beats per minute	96 (IQR 81 –112)	111 (IQR 91 – 130)	<0.001
ED GCS	15 (IQR 14 – 15)	14 (IQR 3 – 15)	<0.001
Injury severity score	17 (IQR 10 – 27)	29 (IQR 21 – 38)	<0.001
OR RBC, units	0 (IQR 0 – 2)	6 (IQR 2 – 13)	<0.001
Small bowel resection	333 (22%)	142 (33%)	<0.001
Large bowel resection	283 (18%)	146 (34%)	<0.001
Splenectomy	308 (20%)	112 (26%)	0.011
Wound Class			<0.001
1	358 (23%)	54 (12%)
2	349 (23%)	178 (41%)
3	759 (49%)	188 (43%)
4	73 (5%)	15 (3%)
OR time of index laparotomy in hours (median, IQR)	2.4 (1.7, 2.9)	2.0 (1.3, 2.9)	<0.001
Management of skin after fascial closure			<0.001
Skin closed	1,287 (84%)	180 (41%)
Outcomes
Any SSI	180 (12%)	175 (40%)	<0.001
All Incisional SSI (n, %)	74 (5%)	77 (18%)	<0.001
All OS-SSI (n, %)	140 (9%)	147 (34%)	<0.001
Incisional + OS-SSI	42 (4%)	41 (4%)	0.760
CD classification			0.376
Grade I-II (n, %)	43 (24%)	35 (20%)
Grade IIIa-V (n, %)	137 (76%)	140 (80%)
Hospital LOS, days (median, IQR)	8 (IQR 5 – 16)	22 (IQR 12 – 34)	<0.001
Mortality	55 (4%)	71 (16%)	<0.001

Descriptive statistics for continuous variables are median, first, and third quartiles. Categorical variables are presented as frequency and percentage. ED = emergency department; SBP = systolic blood pressure; HR = Heart rate; GCS = Glasgow coma scale; OR = operating room; RBC = red blood cells. SSI: Surgical site infection; Incisional SSI: Superficial and deep surgical site infection; OS-SSI = organ space surgical site infection; CD classification: Clavien-Dindo classification; LOS: Length of stay

Table 3.

Fascial closure types of patients who underwent damage control laparotomy (DCL).

DCL Patients	All (N=436)	Period 1 (1/2011–12/2015) N= 283	Period 2 (1/2016–12/2020) N= 153	P value
Fascial closure type				0.004
Die with an open abdomen	29 (7%)	10 (4%)	19 (12%)
Only skin close	11 (2%)	9 (3%)	2 (1%)
Split thickness skin graft	7 (2%)	6 (2%)	1 (0.7%)
Bridging mesh	5 (1%)	4 (1%)	1 (0.7%)
Fascial closure	384 (88%)	254 (90%)	130 (85%)

Even though there were no differences in the rates of SSI between the two periods, patients who underwent DCL had higher rates of all SSIs and higher rates of OS-SSIs compared to patients who underwent definitive laparotomy (Table 2). However, there were no significant differences in minor CD (24% vs 20%) and major CD (76% vs. 80%) SSI complications between these two groups (Table 2). In addition, DCL patients who developed a SSI had a higher number of re-explorations (mean of 2.9 vs 2.4, P= 0.007) and a longer interval of time between index operation and abdominal closure (mean 2.2 vs. 1.9 days P= 0.028) compared to DCL patients without a SSI, and similar trends were seen with OS-SSIs.

On multivariable analysis, after controlling for mechanism of injury, emergency room systolic blood pressure, wound classification, large bowel resection, ISS, splenectomy, and operating-room time at index laparotomy, DCL was associated with major SSIs (CD grade ≥ IIIa) (Table 4). When performing independent regression analysis for OS-SSI and incisional SSI, the association between DCL and both types of infection remained clinically and statistically significant (Supplemental Tables 1–2). When including only patients who had their skin closed after a laparotomy, a total of 1,473 patients (75% of the total population) of whom 186 underwent DCL, DCL was associated with an even higher odds ratio of developing an incisional SSI (Supplemental table 3).

Table 4.

Predictors of Major Surgical Site infections (or Clavien-Dindo Classification grade ≥ IIIa)

Variable	Odds ratio	95% Confidence Interval	P value
Damage Control laparotomy	4.3	3.0 – 6.1	<0.001
Operative time at index laparotomy in hours	1.1	1.1–1.3	0.004
ED Systolic blood pressure	0.99	0.98 – 0.99	0.036
Large bowel resection	3.9	2.8 – 5.3	<0.001
Penetrating Injury (Ref. Blunt)	2.2	1.5 – 3.4	<0.001
Wound Class (Ref. wound class 1)
2	3.0	1.6 – 5.7	<0.001
3	2.7	1.4 – 5.3	0.005
4	4.3	1.8 – 10.2	<0.001
Injury severity score	1.0	0.99–1.0	0.486
Splenectomy	2.4	1.6– 3.5	<0.001

Discussion

The rates and severity of SSIs after trauma laparotomy in our patient population have remained stable over the recent decade despite the decrease in DCL rates. Since the initial project reduced the rate of DCLs from 39% to 23%, our DCL rate has been further reduced to approximately 15%. Although the institutional DCL rate decreased and remained low over time, other risk factors for SSI were more common in period 2, including injury severity and a penetrating mechanism of injury. Nonetheless, our overall rate of SSIs after trauma laparotomy is similar to or lower than those reported in other studies – 7.6% for superficial and deep incisional SSIs and 14.5% for OS-SSIs.^20–22

After the initial reduction in DCL use from 39% (2011–2013) to 23% (2013–2015), there were no significant differences in the rates of superficial SSIs (8% versus 7%, P= 0.544) or in OS-SSIs (12% versus 16%, P= 0.148) in the pre- versus post-intervention period.¹⁷ However, this initial study, like other observational cohort studies, was limited by the inability to adjust for unknown confounders. In a subsequent pilot randomized trial comparing DCL to definitive laparotomy in selected patients in whom there was equipoise about the closure strategy, there was a clinically, but not statistically, significant reduction in superficial SSIs (19% to 6%, P= 0.349) and in OS-SSIs (38% to 28% P= 0.496), in the DCL group.²³ It is unclear if this difference was due to a higher than average rate of SSIs in the definitive laparotomy group, which had a higher ISS, 34 [20, 43] vs. 29 [22, 41], and a higher number of deaths (33% versus 0%).²³ Such chance imbalances can occur in small randomized trials. Our current study, which is larger and spans a longer time period than the initial study, suggests that while DCL is associated with a higher rate of SSIs, there are other contributing factors. Thus, although DCLs have been associated with both incisional and OS-SSIs,^16,24 reducing DCLs may not be sufficient to prevent SSIs.

Strategies exist to reduce the risk of incisional SSIs. Leaving the skin open has been suggested as a strategy to prevent incisional SSIs,^25–26 but this strategy is more resource intensive and results in a less cosmetic outcome. Furthermore, in a retrospective single-center study, He et al. did not identify an increased rate of superficial SSIs with primary skin closure.²⁷ Similarly, our study did not show any differences in incisional SSI rates between those patients who had their skin left open versus closed after DCL (16% vs 19%, P= 0.403). In addition, incisional negative pressure wound therapy (iNPWT), in addition to primary skin closure, has been promoted to decrease superficial SSI rates.^28–29 In a meta-analysis of non-randomized studies, Lakhani et al. reported a decrease in superficial and deep SSIs after emergency laparotomy with iNPWT.³⁰ Alternatively, an EAST multi-center study found a lower rate of superficial SSIs with skin closure with wicks versus skin closure with iNPWT (9.8% versus 16.2%, P= 0.361) but the difference was not statistically significant.³¹ Future research should focus on identifying strategies to prevent incisional SSIs after DCL.

Unlike for incisional SSIs, there are limited interventions that have been proven to reduce OS-SSIs after emergency surgery. Large bowel injury and resection have been identified in this and prior studies to be a risk factor for OS-SSIs.²⁴ The impact of the timing of colonic anastomosis on OS-SSI rates has been evaluated, with several studies suggesting that delayed anastomosis does not increase the risk of OS-SSI, particularly if it is performed in selected patients and on the first takeback.^32–36 Other studies recommend diversion with a stoma, such as for destructive colon injuries.³⁵ Direct peritoneal resuscitation has been suggested as an adjunct to facilitate fascial closure and to decrease the risk of intra-abdominal infectious complications,³⁷ although studies have conflicting results.³⁸ A recent study suggested an increase in fungal intra-abdominal infections after direct peritoneal resuscitation.³⁹ Lastly, the use of broader spectrum antibiotics such as ertapenem may reduce OS-SSIs;⁴⁰ however, the impact of broad-spectrum antibiotics on bacterial resistance patterns needs to be further studied.

Of the predictors of major SSIs that we identified in this study, operative time at the index case, DCL, and time to fascial closure in patients who underwent DCL were the only potentially modifiable risk factors to decrease SSIs. The association of increased operative time with increased risk of SSI across surgical specialties was previously reported in a systematic review,⁴¹ and in a retrospective study of trauma laparotomies.²⁴ Therefore, the decrease in DCL rates might have also been counterbalanced by the increase in operative time experienced by definitive laparotomy patients. In addition, adjuncts after DCL, such as hypertonic saline infusion, have been described to reduce time to fascial closure, but the data is conflicting. While observational studies suggest a benefit,^42–43 a recent randomized trial did not demonstrate an advantage.⁴⁴ A study by Loftus et al. reported improved time to closure with a standardized protocol that used judicious fluid replacement and early return to the operating room in addition to hypertonic saline.⁴⁵ Our institution already utilizes hypertonic saline infusion.

The limitations of this study are those associated with retrospective, single-center cohort studies, such as limited data accuracy and lack of generalizability. Although prior studies have reported an association between DCL and SSI similar to that noted in ours, ^{46, 16} we cannot conclude that there is causation, especially since patients undergoing DCL often have multiple risk factors for SSI development. Although the only randomized trial comparing DCL to definitive laparotomy suggested a causal relationship,²³ larger trials would be needed to definitively answer the question. Unfortunately, lack of equipoise may prevent such a trial from being performed. The rate of SSIs, particularly incisional SSIs, may have been underestimated since minor interventions, such as removing a few staples, may not always be well documented in the medical records. Also, patients may present to other hospitals for treatment post-discharge. It is possible that changes in patient characteristics and injury patterns, as well as evolving clinical management practices over the study period, may have served as temporal confounders.^47–48 For example, increased use of prehospital blood over the study period may have resulted in unexpected survivors who may be more injured.⁴⁹ The median ISS was increased in the post-intervention period, although there was no significant difference in pre-hospital or arrival vital signs in our study.

Conclusion

Despite decreases in DCL use, the rates and severity of SSIs after trauma laparotomy in our patient population have not decreased but have remained stable over the recent decade. The decreased use of DCLs may have been offset by the increase in penetrating injuries and injury severity experienced by our patients and possibly by an increase in operative time experienced by definitive laparotomy patients. Therefore, there needs to be a focus on other potentially modifiable risk factors, such as shorter time to fascial closure after DCL, to impact the frequency and severity of SSIs after trauma laparotomy.

Data Availability Statement

Data will be available upon request and in compliance with institutional policy.