Prolonged Therapy Is Not Associated with Delayed Identification of Recurrent Intra-Abdominal Infection

Patients and Methods

The University of Kansas Medical Center Institutional Review Board granted approval for this study. We analyzed all adult patients with a colon-related complicated IAI between 2016 and 2020 at our institution. Only patients who underwent a source control procedure during their initial operation were included. The presence of a recurrent IAI was recorded. Demographics, comorbidities, total duration of antibiotic therapy, and in-hospital mortality were recorded.

The cohort was divided initially into two groups based on whether they sustained a recurrent IAI or no recurrent IAI during their admission or within 30 days of initial operation. The initial duration of antibiotic use after the source control procedure was identified for all patients. Those receiving 4 ± 1 days of antibiotic therapy were considered to have received standard course therapy whereas those receiving >5 days were considered prolonged course. The primary outcomes were the presence of and the days to identification of a recurrent IAI. We considered the diagnosis of a secondary infection to be delayed if it was identified more than 10 days post-operatively. We defined the recurrence latency period as the number of days between the cessation of the initial antibiotic treatment and the diagnosis of the recurrent infection.

Standard univariate techniques were used. Continuous variables were analyzed using Wilcoxon rank sum, χ ² and Fisher exact test were used where appropriate. Statistical analysis was defined as p value <0.05. Statistical analysis was performed using SAS 9.4 (SAS Institute, Cary, NC).

Results

A total of 76 patients met inclusion criteria of which 48.7% were perforated diverticulitis, 18.4% were non-diverticulitis sigmoid perforations, 17.1% were non-diverticulitis colon perforations in the transverse or descending colon, 7.9% were cecal or right colon perforations, 7.9% were anastomotic leaks. Overall 17 (22.4%) of these patients had a recurrent IAI. Demographics and comorbidities are provided in Table 1. Patients with recurrent infections were slightly younger than patients without recurrent infections, had lower rates of pre-operative anticoagulation, and were less likely to have their initial surgery performed by an acute care surgeon. There was no difference in the days of antibiotic agents after source control between the recurrent infection and no recurrent infection groups (p = 0.6). Furthermore, there was no significant difference in the proportion of patients receiving the standard course of antibiotic agents between the two groups. There was a difference in the total days of antibiotic use between the two groups, with the recurrent infection group averaging 10 more days of antibiotic use than the non-recurrence group. The 10 most common antibiotic agents for the initial infection are provided in Table 2.

Of the 17 patients with a recurrent infection (Table 3), eight (41.7%) had a prolonged course of therapy for their initial infection. However, there were no differences in median days to recurrence (9 vs. 11.5; p = 0.29), or the proportion of patients with a delayed (>10 days) diagnosis of recurrence (44.4% vs. 75.0%; p = 0.33) when stratified by standard versus prolonged course. Those who were defined as having a standard course of therapy had fewer days of antibiotic agents after source conrol than those receiving a prolonged course. The recurrence latency period was five days for those who received a standard course and zero days for who received a prolonged course but this difference was not statistically significant. Patients with recurrent infections had longer median total treatment durations than those without a recurrent infection (7 vs. 17; p < 0.0001). The mortality rate for those with recurrent infection was 0% versus 15.3% (9 patients) for those without a recurrent infection although this did not reach statistical significance. Differences between survivors and non-survivors are listed in Table 4. In general, non-survivors were older, more likely to be immunosuppressed, or have a damage control operation. Non-survivors were also had higher Sequential Organ Failure Assessment (SOFA) and Acute Physiology and Chronic Health Evaluation (APACHE) II scores than survivors.

Discussion

Recurrent IAIs remain an important issue in the treatment of the patient with a complicated IAI. Ideal treatment would reduce the rate of recurrence, while not obscuring the diagnosis when recurrences do occur. As such, the time to recurrence should be an important outcome variable when evaluating antibiotic durations. In this retrospective study, we did not identify an association between the duration of treatment for the initial complicated IAI and the time to diagnosis of a recurrent infection.

Studies of short course antibiotic durations for a variety of common infections have become more frequent in recent years. These studies define clinical failure in a variety of ways, including recurrence, but the time to clinical failure or recurrence appears to be frequently ignored in the current literature.^5–9 The STOP-IT trial identified an almost five-day mean delay in the time to identification of a recurrent intraabdominal infection (mean 15.1 vs. 10.8 days; p < 0.001). A study by Smith et al. ¹⁰ is one of the few other than STOP-IT to evaluate the time to treatment failure. In their study treatment failure was defined as a composite of recurrent complicated IAI, secondary infection, or all-cause in-hospital mortality. Whereas the composite treatment failure outcome was indeed delayed by a median of two days (6 vs. 8 days; p = 0.001) in the group that received a prolonged course, this finding appears to be driven largely by a delay in death (median 6 vs. 18 days; p < 0.001). There was no statistical difference in the median days until diagnosis of a recurrent infection (7 vs. 8 days; p = 0.102). ¹⁰ Our recurrence rate was 22.4%, which is higher than that reported in STOP-IT (overall 14.7%) but lower than reported by Smith et al. ¹⁰ (overall 40.8%). ⁴ However, given the overall lower sample size, our pool of recurrent patients is low. We suspect that we are underpowered to detected clinically significant differences in the time to identification of recurrent infections or in the recurrence latency period. These limitations highlight the difficulty with studying recurrent infections.

The STOP-IT trial did not find any differences between short- and long-course antibiotic durations with regards to their primary composite outcome (secondary infection, surgical site infection, or death). Furthermore, all subsequent subgroup analyses of the STOP-IT trial failed to identify a group that benefitted from prolonged initial therapy.^11-18 This study also found no relation between the initial duration of antibiotic agents and delayed identification of secondary IAI.

Despite our failure to identify an association between prolonged antibiotic use and delayed identification of recurrent IAI, there are benefits of using a shorter course. The use of any antibiotic agents after IAI has been found to alter the rectal flora and antibiotic resistance in a variety of bacterial species. ¹⁹ Although a course of antibiotic agents after source control for IAI is often unavoidable and necessary, antibiotic stewardship remains an important principle of treatment to avoid the selection of highly resistant organisms. Furthermore, the unnecessary prolonged use of any pharmaceutical has the potential to increase healthcare costs and place further burden on the patient.

This study did have several limitations that must be considered. First, we are underpowered, not only to detect differences in our primary outcome, but also to control for potential confounding variables. Second, as a single-center retrospective analysis, this study is subject to the limitations inherent to that study design and may not be broadly generalizable. Third, all non-survivors in the study were determined to not have a secondary IAI because they did not survive long enough in the hospital to sustain one. This could potentially result in survivorship bias.

Conclusions

In concordance with the STOP-IT trial, we found that antibiotic course duration is not associated with the rate of secondary IAIs. However, we failed to identify an association between the duration of antibiotic agents and time to identification of a secondary IAI. These findings further support the adoption of a shorter antibiotic course in accordance with Surgical Infection Society guidelines. More work is needed to determine if the time to diagnosis of recurrent IAI may be influenced by prolonged initial therapy. We further support increased reporting of the time to event outcomes in studies of this type.