Retrospective Analysis of Post-Operative Antibiotics in Complicated Appendicitis

Patients and Methods

After Institutional Review Board approval, we identified all patients who underwent appendectomy for acute appendicitis from January 2012 to December 2016 at our institution. Through chart review, we selected the subset of patients who had complicated appendicitis. For the purposes of this study, complicated appendicitis was defined as an intra-operatively identified gangrenous or necrotic appendix without perforation or a non-iatrogenic perforated appendix, regardless of the presence of local or generalized contamination, periappendiceal abscess, or local or generalized peritonitis. The data collected included demographic and clinical characteristics at the time of presentation, intra-operative findings, post-operative antibiotic treatment, and complications and interventions related to the diagnosis of appendicitis within 30 days of appendectomy. For patients with a post-operative infectious complication, we included only the duration of antibiotics prior to identification of the infection.

Complicated appendicitis was stratified by whether perforation was identified intra-operatively. Post-operative antibiotic treatment was stratified as no antibiotics after appendectomy, administration of intravenous (IV) antibiotics alone, and administration of a combination of oral (PO) and IV antibiotics. Both predictive and outcome variables were analyzed using descriptive statistics for each subgroup. Continuous variables are reported as mean ± standard deviation (SD) and categorical variables as percentages. Student's t-test was used for continuous variables and χ² or Fisher exact tests for binary categorical variables as appropriate. Analysis of variance (ANOVA) with Bonferroni correction was used for multinomial continuous variables. Based on the significant predictors from the appropriate univariable analysis, multivariable logistic regression was used to calculate the odds of developing SSI with antibiotic treatment as the predictor variable. Two models were created for this purpose, one investigating the route of therapy and one the duration of therapy. Both of these were tested for goodness of fit using likelihood ratios. Statistical analysis was performed using STATA 14 (StataCorp 2015). A p value ≤0.05 was considered statistically significant.

Results

Of the 1,102 patients who underwent appendectomy for acute appendicitis during our study period, 188 (17.1%) were identified as having complicated appendicitis. The demographic, clinical, and outcome characteristics for these patients are summarized in Table 1. Of note, 45 patients (24%) were seen to have complicated appendicitis without perforation (i.e., a necrotic or gangrenous appendix) at the time of surgery, whereas 143 (76%) were deemed to have perforated appendicitis. Patients with a final diagnosis of perforated appendicitis as determined by the operating surgeon were more likely to have free fluid (p = 0.002) or perforation (p = 0.016) described in their pre-operative imaging reports than those without. As expected, these patients had a greater LOS (p = 0.006) and were significantly more likely to receive antibiotics post-operatively (p < 0.001). The incidence of 30-day post-operative complications was similar in the two groups. One urinary tract infection and two cases of Clostridioides difficile colitis occurred in the perforated appendicitis population. Thirty-one patients required a post-operative procedure (in most cases, drain placement for abscess evacuation). Two patients, both with perforated appendicitis, required re-operation (Table 1). No catastrophic events were identified.

Post-operative antibiotics were administered to 134 of the 188 patients (71.3%). Most patients who received post-operative antibiotics had an intra-operative finding of perforation (n = 120; 89.6%). Of the patients without perforation, only 15 (33.3%) received antibiotics post-operatively. Intra-operative determination of appendiceal perforation was significantly associated with administration of post-operative antibiotics (p < 0.001). The LOS was significantly longer for patients receiving IV antibiotics than for patients not receiving any antibiotics (p < 0.001). Otherwise, antibiotic administration did not differ according to demographic and clinical presentation characteristics or development of post-operative complications (Table 2).

On univariable analysis, a clinical diagnosis of diabetes (p = 0.04); decision to leave an intra-operative drain (p < 0.001); the presence of free fluid (p = 0.002), abscess (p = 0.039), or perforation (p = 0.012) on pre-operative imaging; and a greater LOS (p < 0.001) all were associated with development of SSIs. Patients who developed SSIs were more likely to have post-operative ileus, be re-admitted, or require an invasive procedure after their appendectomy to manage sequelae from complicated appendicitis (p < 0.001). The duration of antibiotic treatment pre-SSIs was significantly longer in patients who developed an SSI than in those who did not (p < 0.001) (Table 3).

Multivariable logistic regression was utilized to assess both the route of antibiotic administration and the length of the post-operative course for an association with development of SSIs. Because we were concerned primarily about confounding by indication, we developed models controlling for variables significantly associated with SSI and antibiotic administration on univariable analysis, and we tested the goodness of fit of these models using likelihood ratio testing. Ultimately, we found that both models had the best goodness of fit when adjusted for free fluid on imaging and intra-operative drain placement. This analysis determined that patients receiving one day or three or more days of antibiotics had higher odds of developing an SSI than patients who did not receive any post-operative antibiotics. Patients receiving two days of antibiotics did not have a significantly higher risk of developing SSI than patients not receiving antibiotics. The IV route of antibiotic administration was significantly associated with higher odds of developing an SSI than was administration of a combination of IV and PO antibiotics (Tables 4 and 5).

Discussion

According to current U.S. guidelines, antimicrobial therapy for complicated intra-abdominal infections, including complicated appendicitis, should be limited to four to seven days after definitive source control [21]. However, it has been suggested that such guidelines have resulted in the administration of an arbitrary course of antibiotics post-operatively to patients with complicated appendicitis with the expectation of lower infectious complication rates but with no corresponding evidence to support such an intervention [6]. A 2005 Cochrane review of 45 studies covering 9,576 patients with appendicitis found that the most commonly suggested regimen for appropriate antibiotic coverage in all cases of appendicitis is a single pre-operative dose of any second-generation cephalosporin [15]. Unfortunately, the applicability of this meta-analysis to patients with complicated appendicitis, particularly those with perforated appendicitis, is questionable because of its reliance on studies that are outdated, focused on the pediatric population, or excluded cases of perforated appendicitis from the analysis [6]. In our review of the literature, we found only one randomized controlled trial, the STOP-IT trial, comparing the outcomes of short and long fixed-duration courses of antimicrobial therapy for intra-abdominal infection after adequate source control [22]. In this trial, a four-day fixed-duration antibiotic course was found to have infectious outcomes similar to those of longer courses of antibiotics. However, appendiceal sources of intra-abdominal infection comprised less than 15% of the patient population in this trial, making its applicability to the management of complicated appendicitis limited.

More recent data on antimicrobial therapy following appendectomy and adequate source control for complicated appendicitis have shown that longer antibiotic courses independently prolong the hospital stay without definitive improvement in the prevention of infectious complications [6,8,11,12,19]. In a retrospective study covering a five-year period at our institution, 17.1% of acute appendectomies were identified as having been performed for complicated appendicitis. Of these patients, 12.8% experienced post-operative ileus, 14.4% developed an SSI, and 13.8% required re-admission, reflecting a high rate of surgical complications in this population. Although surgeon-identified cases of appendiceal perforation, gangrene, necrosis, abscess, or phlegmon all were classified as complicated appendicitis, from the data we gathered, it is clear that surgeons managed perforated appendicitis differently than gangrenous or necrotic appendicitis. Patients with perforation were not only much more likely to receive post-operative antibiotics, they also received a longer course of antibiotics. However, the intra-operative anatomic findings did not appear to influence the route by which the surgeon administered antibiotics. More importantly, these differences in management decisions did not result in differences in infectious outcomes between the perforated and non-perforated cohorts. Thus, despite a lack of strong evidence to guide them, the surgeon's decisions were appropriate when it came to these differences in post-operative antimicrobial therapy decisions.

When cohorts were compared according to the development of SSIs, multiple variables were associated independently with the development of SSI, including a clinical diagnosis of diabetes; the presence of pre-operative imaging-identified free fluid, abscess, or perforation; and intra-operative drain placement. These variables, other than a clinical diagnosis of diabetes, are markers of clinical severity; one (imaging characteristics) a physiologic marker, although detected by computed tomography scan; the other (decision to leave a drain) a marker of surgeon judgement regarding severity based on intra-operative findings. Interestingly, none of these variables was found to influence surgeons' antimicrobial management: Only the surgeon's identification of a perforated versus non-perforated appendix was associated with the decision to initiate antibiotics post-operatively. These differences further suggest that clinical severity is not delineated adequately by the umbrella term “complicated appendicitis.” As such, when appendectomy is pursued to manage complicated appendicitis, our results strongly suggest a need to tailor post-operative antibiotic management on the basis of clinical and anatomic severity. As the debate over the optimal treatment approach regarding non-operative and operative complicated appendicitis continues, whether free fluid, abscess, or perforation detected on pre-operative imaging is a contraindication to appendectomy for complicated appendicitis remains unclear and warrants further exploration.

With regard to the duration of antibiotic administration, this study found that an antibiotic course of three days or more compared with a shorter course did not lead to an improvement in post-operative complications. This finding corroborates several other investigations that found prolonged post-operative antibiotics had no association with differences in surgical complications [6,8,11,12,19]. A 2014 study by Kimbrell et al. of 52 patients with operatively managed complicated appendicitis found no significant difference in abscess development between patients receiving >24 hours of post-operative antibiotics and those who did not [6]. In contrast, after adjusting for post-operative drain placement and free fluid on pre-operative imaging, our investigation demonstrated that 24 hours of post-operative antibiotic therapy and >72 hours of post-operative antibiotic therapy were both associated with higher odds of developing an SSI compared with 48 hours of antibiotic therapy. However, this discrepancy in findings becomes less important when one considers the limitations of the small sample of both the study by Kimbrell et al. and the current investigation and the broader definition of SSI used in this study. It is likely that in both investigations, the number of patients receiving 24 hours or less of antibiotics was disproportionately small, indicating that surgeons tend to prescribe a more protracted course of antibiotics in patients with anatomically severe disease.

Although post-operative antibiotic administration was not found to reduce the rate of complications, it was associated with longer hospital stays. Perforation also was found to be independently associated with longer stays. These findings are consistent with prior investigations that likewise found that prolonged post-operative antibiotic courses among patients with operative complicated appendicitis increased the hospital LOS without providing an appreciable clinical benefit [19]. However, given that perforation was the only variable associated with antibiotic administration, attributing a longer LOS to antibiotic administration alone is confounded by a clinical susceptibility bias. It is likely that the severity of disease attributed to perforated appendicitis influenced providers to favor a more cautious post-operative management course in this group of patients, resulting in decisions that prolonged hospitals stays independent of the need for extended antibiotic therapy.

In an attempt to standardize disease severity and guide therapy in the management of acute appendicitis, the American Association for the Surgery of Trauma (AAST) created an anatomically based, time-independent approach to assessing a patient's disease in which a higher grade represents a greater severity of disease. In this classification of acute appendicitis, perforated appendicitis is of worse severity than complicated appendicitis where the appendiceal wall remains intact, and as such, is associated with a higher risk of open procedures, a higher risk of complications, and longer LOS [23]. Our study supports this method of disease classification, as patients with complicated non-perforated appendicitis have similar post-operative outcomes and a shorter LOS than those with perforated appendicitis despite being less likely to receive post-operative antibiotics. However, this calls into question whether the current umbrella classification of operative complicated appendicitis including gangrenous, necrotic, and perforated appendices in the same group, is appropriate. Further research comparing these higher-grade disease entities independently may assist not only in validating the current AAST grading scale but also allow study of each of them as potentially different processes with different management requirements. In particular, whether non-perforated but gangrenous or necrotic appendicitis warrants its own unique designation outside of complicated appendicitis and thus its own post-operative management paradigm is of great interest moving forward.

Given that the data represented in this study are from a small sample at a single institution without a standardized protocol, our findings concerning surgeons' post-operative antimicrobial decisions are difficult to generalize. Nonetheless, our primary concern when devising the analytic plan for this study was addressing confounding adequately by indication, which we felt would be driven mainly by clinical severity at presentation and intra-operative determinations of severity, both of which could affect surgeons' decisions about the prescription of antibiotics post-operatively. Although we sought to adjust for clinical and anatomic severity, we could adjust only for the variables captured in our dataset, which makes our model imperfect when controlling for confounding by indication.

Complicated appendicitis is both common and morbid, yet its classification and management are not uniformly agreed on. Although non-operative management currently is favored over immediate appendectomy [5], a definitive consensus on the best treatment approach does not exist; and a significant number of cases are managed with urgent surgical intervention without strong data to drive their post-operative management. The results of this study highlight the clinical inadequacy of grouping operative appendicitis with perforation, gangrene, necrosis, abscess, or phlegmon under the same umbrella of complicated appendicitis in regard to post-operative SSI risk and appropriate post-operative antimicrobial management decisions. In view of these results, rather than manage operative complicated appendicitis uniformly, it would be reasonable to adopt a standardized protocol where patients with perforated appendicitis, as determined by the surgeon intra-operatively, or appendicitis where the surgeon decides to leave a drain, receive no more than three days of antibiotics post-operatively. However, given the limitations inherent in a retrospective study of this size and design, an adequately powered, randomized trial to clarify the optimal duration and route of antibiotic therapy in patients with complicated appendicitis, ideally stratified by perforation versus gangrene or necrosis, would improve patient care after surgery for this common condition.