Short-Course Antimicrobial Therapy Does Not Increase Treatment Failure Rate in Patients with Intra-Abdominal Infection Involving Fungal Organisms

Patients and Methods

The STOP-IT trial was an investigator-initiated, open-label, multi-center trial to investigate the optimal duration of Antimicrobial therapy after achieving source control in patients with complicated IAI. Over a five-year period, 518 patients were enrolled at 23 sites in the United States and Canada. Eligible patients were 16 years of age or older with complicated IAI who presented with fever (temperature ≥38.0°C), leukocytosis (peripheral white blood cell count >11,000/mcL), or gastrointestinal dysfunction caused by peritonitis that precluded intake of more than half their normal diet and who had undergone a source control intervention. Source control is critical to the management of IAI and was defined as a procedure to eliminate infectious foci, control factors that promote infection, and correct anatomic derangements to restore normal physiologic function. Patients with gastroduodenal perforations who underwent source control within 24 hours of symptom onset were excluded. As previously described, the use of antifungal therapy was left to the discretion of the clinician as long as it was in accordance with established guidelines, including either an azole or an echinocandin. The duration of antifungal therapy was determined by randomization group [11].

Data analysis was designed to test the hypothesis that fungal IAI, defined as an abdominal culture positive for a fungal organism, was associated with a higher rate of treatment failure, which was defined as a composite of recurrent IAI, surgical site infection (SSI), and death. Patients with available culture data were stratified on the basis of the presence of a fungal isolate and were analyzed in two ways: First, in a sub-group analysis of patients with fungal isolates, based on the original randomization to either a four-day course of Antimicrobial therapy (experimental group) or a traditional treatment duration based on resolution of symptoms (control group); and second, by comparison with those without a fungal isolate. Statistical significance was determined using the standard alpha value of <0.05.

Statistical analyses were performed using the Wilcoxon rank sum test for continuous variables and the χ² or Fisher exact test for categorical variables as appropriate. All statistical analysis was conducted using SAS software, version 9.3 (SAS Institute, Cary, NC).

Results

Four-hundred eleven patients (79% of the total patients seen during the study) had culture data available and were included for analysis. Of these, 58 (14%) had a positive fungal culture and were split between the control and experimental groups. The baseline characteristics of these patients in terms of demographics and co-morbidities were similar (Table 1). All fungal organisms isolated were yeast, with the most common listed in Table 1.

The control arm of the study contained 31 patients with fungal isolates, and the experimental arm contained 27. The only statistically significant difference between these groups was the number of days of Antimicrobial (antibiotic and antifungal) treatment (8 vs. 5; p = 0.046). There was no difference in days to enteral feeding, length of admission, or the composite outcome (Table 2).

Compared with patients without a fungal isolate, patients with fungal IAI were more likely to have malignant disease and coronary artery disease but were otherwise similar (Table 3). Fungal species were more frequently isolated from the duodenum and small intestine and were never isolated from the biliary system (Table 4). Fungi were rarely isolated from infections originating in the appendix. Patients with fungal IAI had longer hospital stays and more days to resumption of normal feeding, but there was no difference in the composite outcome (Table 5). There was a trend toward a longer duration of Antimicrobial therapy in patients with fungal IAI, but this did not reach statistical significance.

A total of 81 patients received antifungal therapy. Seventy-six (94%) were treated with an azole and 8 (10%) with an echinocandin, with three patients started on an azole and switched to an echinocandin. Patients who received antifungal therapy were more likely to have a hospital-acquired infection (24% vs. 11%; p = 0.001) and had higher Acute Physiology and Chronic Health Evaluation (APACHE) II scores (12 vs. 8; p < 0.0001). Of the 58 patients with a positive fungal culture, 37 received antifungal therapy. Considering only these 58 patients, those who were treated with an antifungal drug had trends toward higher APACHE II scores and more days to enteral feeding, both of which bordered on statistical significance, but otherwise had similar risk factors and outcomes (Table 6).

Discussion

In this post-hoc analysis of 58 patients with IAI involving fungal organisms who were enrolled in the STOP-IT trial, we identified no difference in clinical outcomes on the basis of the duration of Antimicrobial drugs after source control. Patients with fungal IAI had longer hospital stays and more days to enteral feeding but no difference in the mortality rate, likelihood of recurrent IAI, or SSI compared with those without a fungal component to their infection. This work suggests that the findings of the STOP-IT trial may be generalizable to fungal infections, which generally would be considered high risk. Hassinger et al. recently reported another analysis of the STOP-IT trial data and identified risk factors for treatment failure [12]. Evaluating patients with high-risk factors, they found no difference in treatment failure rates if they were assigned to the control arm or the short-course treatment arm. Our results agree with this study by demonstrating an equivalent rate of treatment failure with a shorter Antimicrobial course in another high-risk sub-group.

Similar to the findings in prior studies, fungi frequently were isolated from colonic and small-bowel sources. Compared with the rest of the study population, patients with fungal isolates were more likely to have infections from small bowel and duodenal sources but not from gastric sources. Prior studies suggest that fungi are rarely encountered in IAI of appendiceal origin [3]. Our data indicate that fungi can complicate IAI of appendiceal origin, but support the prior findings that this is rare. It is interesting that fungi were never recovered from the biliary system.

In this cohort, there were 106 patients for whom no culture data were available, and they were not included in this analysis. These patients were more likely to have IAI of appendiceal origin and had lower APACHE II scores but otherwise were demographically and clinically similar to the patients analyzed. One common reason these patients did not have cultures was that fungal cultures are not common in routine appendiceal IAI. Furthermore, given that fungi are uncommon in infections of appendiceal origin, we do not expect that the failure to capture them in this analysis would significantly bias our findings, because if there are patients who would benefit from a longer course of antifungal therapy, it is most likely to be those with more severe infections or more critical illness.

We observed longer stays and more days to resumption of enteral feeding among patients with positive fungal cultures. A possible explanation for this finding is that patients having fungal cultures were more likely to have malignant disease and coronary artery disease, with trends toward older age and higher APACHE II scores (see Table 3). These differences suggest that patients with fungal isolates had more co-morbidities and greater physiologic disturbances. Additionally, these patients were more likely to have infections originating from the duodenum and small bowel, and having a proximal GI infectious source may lead to slower tolerance for enteral intake. In contrast to several other studies, our data do not demonstrate more deaths in patients with fungal IAI compared with patients without a fungal component [3,7,8]. One potential explanation for this result is that many other studies focused on patients in intensive care units (ICUs), whereas our study included both ICU and non-ICU patients. It could be that the higher mortality rate associated with fungal isolates is accounted for primarily in patients with more severe physiological disturbance and who are more likely to have nosocomial infection. The small sample and the inclusion of lower-acuity, non-critically ill patients likely limit our ability to detect a difference between groups. The range of Candida species of yeast observed in this study is similar to what has been reported elsewhere for fungal IAI [4,5].

One challenge in this study is that we cannot be sure that all fungal isolates detected were clinically relevant, as invasive candidiasis is a difficult diagnosis to arrive at with certainty [13–15]. Although yeast commonly are isolated in gastrointestinal perforations, they may be non-pathologic. Current guidelines recommend against initiating antifungal therapy in a number of clinical settings [16]. Still, some studies have shown that the presence of a fungal isolate is a predictor of a worse outcome, although it remains unknown if the yeast themselves were the causal agent or if they were a marker of risk for morbidity and death [3,4,8]. It is likely that some of the patients identified in this study had fungal isolates that were non-pathologic and did not play a role in their outcomes. Presumably, in the 21 patients with fungal isolates who did not receive antifungal therapy, the presence of fungi was believed not to be significant clinically. In some patients, such as those with gastroduodenal perforations who receive early source control, guidelines recommend limiting Antimicrobial therapy to 24 hours, so antifungal therapy for four days or less may be appropriate in certain groups. Notably, however, patients with gastroduodenal perforations who underwent early source control were excluded from the STOP-IT trial, so our data likely do not reflect over-treatment of this patient population. Because only eight patients received an echinocandin and three of these also received an azole, we did not analyze outcomes on the basis of antifungal treatment regimen.

The use of prospective multi-center data to evaluate an important clinical entity is a strength of this study. However, as a post-hoc analysis of a randomized trial not specifically designed to evaluate fungal IAI, our results are inherently limited. Furthermore, not all patients adhered to protocol, and, although the difference was not statistically significant, this could bias toward a finding of equivalence between the intervention and control groups. This may be attributable to the way in which the duration of treatment was counted. Not all patients were started on antifungal therapy empirically and may have begun a four-day course of antifungal therapy only once culture data became available, after already receiving one or more days of anti-bacterial therapy, for a total of more than four days of therapy overall. Another limitation of this study is its small sample, which may limit its ability to detect smaller but clinically significant differences between the control and intervention groups. However, although the sample is small, it is similar to the size of the populations in several other studies of fungal IAI [3,5,8]. It also is possible that some patients had false-negative fungal cultures, which would reduce our ability to detect a difference between the two groups.

Conclusions

In this study, we demonstrated no difference in the rate of treatment failure among patients with IAI involving fungal isolates who were treated with a shorter course of antifungal therapy after obtaining adequate source control in comparison those treated for a time based on sign and symptom resolution. We further demonstrate that these patients had longer stays than those without a fungal isolate but did not have a higher rate of treatment failure. These findings suggest that shorter antifungal treatment duration may spare patients the adverse side effects and cost of prolonged treatment without compromising treatment efficacy when fungal organisms are present in IAI.