Practice Patterns for the Use of Antibiotic Agents in Damage Control Laparotomy and Its Impact on Outcomes

Patients and Methods

Patients with trauma undergoing laparotomy at the Virginia Commonwealth University (VCU) Medical Center, an urban Level 1 trauma center in Richmond, VA, over a one-year period were identified. Patients were grouped according to those who underwent DCL with temporary negative pressure therapy with an open abdomen and subsequent definitive abdominal closure or those whose abdominal fascia was primarily closed at the time of the initial operation. The primary closure laparotomy (PCL) group had the skin closed at initial operation except when obvious fecal contamination of the wound occurred. In those situations, both the PCL and DCL wounds were packed with gauze moistened with physiologic saline twice a day, followed by wound vacuum therapy once the wound was clean as determined by the attending trauma surgeon.

Demographic information collected included but was not limited to age, Injury Severity Score (ISS), Glascow Coma Score (GCS), Revised Trauma Score (RTS), hospital and intensive care unit (ICU) length of stay (LOS), number of days on ventilator, base deficit, and lactate on arrival. Deaths with LOS <5 days were excluded.

Patients undergoing DCL versus those with PCL were compared for antimicrobial agent use and its correlation with IAI (p < 0.05). The IAI were suspected based on clinical findings of two or more components of the systemic inflammatory response criteria and confirmed by cultures obtained either through percutaneous drainage with computed tomography guidance or at re-operation. Antimicrobial choice was based on Surgical Care Improvement Project (SCIP) guidelines allowing for coverage of gram-positive, gram-negative, and anaerobic micro-organisms. Antimicrobial choice was based on surgeon preference, and agents were continued peri-operatively depending on the amount of gross spillage and concern for development of intra-abdominal abscess. Patients with DCL had the abdomen closed between 48 and 72 hours after initial laparotomy and once they were hemodynamically stable. Outcomes for those patients who then went on to have an IAI develop were compared with those patients who did not have a diagnosis of IAI.

Statistical analyses consisted of chi-square/Fisher exact tests (categoric data) and Wilcoxon rank sum tests (continuous data). Logistic regression analyses were performed to identify predictors of abdominal infection. This study was approved by the VCU Institutional Review Board.

Results

There were 137 patients identified who underwent trauma laparotomies during the time frame of the study. Of these patients, 16 died within five days and were excluded from further analysis. The 121 patients who met inclusion criteria had an average age of 40.6 years ±18.9, an ISS of 21.4 ± 14.5, a GCS of 13.6 ± 3.6, and an RTS of 7.3 ± 1.4. In addition, their initial base deficit was −3.0 mg/dL ± 4.9, and their initial lactate was 3.8 mg/dL ± 2.4. For the overall study population, patients who received pre-operative antibiotic agents had an overall decrease in intra-abdominal infections (Table 1). Long-term use of antibiotic agents as associated with higher infection rates (p = 0.0004).

The 121 patients included 28 who underwent DCL and 93 who had a PCL. Table 2 compares patient demographics and outcomes between patients with PCL versus DCL. The DCL group was more severely injured (ISS: 29 DCL vs. 14 PC, p < 0.001) with more small and large bowel injuries (SLBI), although these were not statistically significant (53.6% DCL vs. 35.5% PCL, p = 0.12). The DCL group also had longer lengths of ICU and hospital LOS and a higher mortality rate.

Antibiotic use in the two groups is shown in Table 3. As demonstrated, antibiotic agents were different for both pre-operative use as well as post-operative administration with variation in the length of time antibiotic agents were used. The use of pre-operative antibiotics was much greater in the PCL group compared with the DCL group. There were 53.6% of patients with DCL who received antibiotic agents long term defined as greater than one postoperative day compared with 28% of PCL patients (p = 0.013).

Table 4 shows the relationship between use of antibiotic agents and the development of IAI in the PCL compared with the DCL group. There were no infections in the PCL group who did not receive pre-operative antibiotic agents. There was a 62.5% infection rate in DCL patients who did not receive pre-operative antibiotics, yet this was not statistically significant. Patients who received pre-operative antibiotic agents showed no difference in intra-abdominal infection rates between the two groups. Furthermore, prolonged antimicrobial use was significantly associated with higher IAI rates for both DCL and PCL.

Regression analyses demonstrated that neither ISS nor DCL were independent predictors of infection, but pre-operative antibiotic agent use was a negative predictor (odds ratio [OR] 0.20, 95% confidence interval [CI] 0.05–0.91, p = 0.037), while post-operative antibiotic agent use (OR 6.7, 95%CI 1.33–33.8, p = 0.044) and SLBI (OR 3.45, CI 1.03–11.5, p = 0.02) were positive predictors of infection with a receiver operating characteristic (ROC) of 0.81 (Table 5). When comparing outcomes data, those patients with IAI had a much higher hospital LOS compared with those in whom an IAI did not develop (28.7 ± 18.8 d vs. 12.7 ± 11.3 d). The ICU LOS was significantly longer in the IAI group (15.9 ± 13.6 d vs. 6.7 ± 9.1 d) compared with the non-IAI group.

Discussion

Numerous guidelines have been developed that address the role of antimicrobial agents in abdominal surgical procedures, but it is uncertain whether these are applicable in the setting of DCL. Most guidelines have focused on minimizing surgical site infection (SSI), yet a large portion of patients undergoing DCL have hollow viscous organ injury with gross contamination or significant infection precluding closure of the skin. These surgically created wounds heal instead by secondary intention to minimize infection rates and post-operative wound complications.

The most widely recognized of these protocols is the SCIP, which has significant implications for medical reimbursement and quality control [11]. The Eastern Association for the Surgery of Trauma (EAST) 2012 update in practice management guidelines addressing the role of antibiotic agents in the management of penetrating abdominal trauma specifically addressed the role of antimicrobial agents in DCL, but was also unable to provide evidenced-based recommendations [12]. One of the most comprehensive sets of guidelines surrounding antimicrobial prophylaxis in surgical procedures was a joint venture between the American Society of Health-System Pharmacists (ASHP), the Infectious Diseases Society of America (IDSA), the Surgical Infection Society (SIS), and the Society for Healthcare Epidemiology of America (SHEA) [13], yet these guidelines did not address patients undergoing DCL.

Similarly, our knowledge of the key infectious complications surrounding DCL remains largely undocumented. The largest multi-institutional study to date followed the natural course of patients enrolled over a two-year period and showed that abdominal abscess was associated with a failure to achieve definitive primary closure [10]. Goussous et al. [14] proposed that the etiology underlying DCL (i.e., intra-abdominal hemorrhage vs. intra-abdominal sepsis) would influence rates of infectious complications; however, this did not hold true. Lastly, Vogel et al. [15] studied infectious complications manifesting themselves outside of the abdomen in patients with trauma undergoing DCL and demonstrated an inability to achieve primary abdominal closure was associated with ventilator-associated pneumonia, blood stream infections, and surgical site infections (SSIs). The above studies focused predominately on achievement of definitive fascial closure as the primary study outcome rather than infectious complications despite their being closely related.

Thus, the current literature fails to specifically address whether our goals to minimize infectious complications of DCL should focus on SSI, intra-abdominal abscess, or non-abdominal sources, and what, if any, antimicrobial prophylaxis is warranted. Surgical site infection is often not an issue those patients undergoing DCL whose wounds are closed by secondary intention, as in our institution. One study is of particular interest because it addressed the specific issue of SSIs in patients with trauma according to SCIP non-compliance after the implementation of SCIP protocol hospital-wide for all surgical procedures. The rate of SSI was increased with trauma patients undergoing damage control operation and with enteric injuries [16].

In this study, we sought to document practice patterns associated with the use of antimicrobial agents within a Level-1 trauma center in the setting of both DCL and PCL for patients with trauma. We demonstrated greater consistency with pre-operative antibiotic administration in the primarily closed patients. It is possible that the patients undergoing DCL were hemodynamically unstable at the time of incision, and that focus was shifted away from the administration of pre-operative antibiotic agents. These patients may have also received pre-operative antibiotic agents by anesthesiologist but may not have been documented appropriately because of the emergent nature of the situation.

A standardized pre-operative checklist for trauma laparotomy specifically addressing pre-operative antibiotic agents may improve compliance. There was considerable variation in the administration of prolonged post-operative antibiotic agents. The majority of patients whose abdomens were primarily closed did not receive prolonged antimicrobial agents, consistent with current guidelines. This is in contrast to more than 50% of patients undergoing DCL who were administered prolonged antimicrobial agents in excess of one day. Whether the decision to extend antibiotic duration was because of the mere presence of an open abdomen and an assumed risk, a concern for intra-abdominal infection or gross spillage, or the a lack of awareness of current guidelines as applicable to the standard laparotomy remains unknown.

Intra-abdominal infection rates, in this study, appeared to be influenced by both the presence and absence of pre-operative antibiotic agents and the duration of post-operative antibiotic administration. Patients with DCL who received no pre-operative antimicrobial agents had higher IAI rates. Similarly, patients with DCL who received post-operative antimicrobial agents experienced higher IAI rates that increased with prolonged duration. These findings have significant clinical impact because the patients in whom these infections developed had much longer ICU and hospital LOS.

This study has many limitations, but it is the first study to address these issues of prophylactic antibiotic agents in DCL after the revised EAST guidelines. This study is limited, because we were not able to account for variations in the types of antimicrobial agent administered. We have since established a protocol for antibiotic agents in the open abdomen with plans for future analyses of infection rates. All pre-operative antibiotic agents, however, were consistent with SCIP guidelines. This study is also a single-center retrospective review with small numbers in some analyses. Last, we did not address the issue of SSI or other hospital-associated infections because these data were not consistently available from the outpatient setting. Further study on a multi-institutional level is warranted.

This study suggests that in the setting of DCL, pre-operative antimicrobial agents should be administered. Prolonged antimicrobial agents are not warranted in the absence of documented IAI at the time of laparotomy. Defining appropriate guidelines for this patient population is of paramount importance to minimize IAI and worse outcomes.