Introduction

Septic peritonitis is a common, life-threatening condition encountered in dogs and cats. Mortality rates have been reported to be as high as 68% with surgical intervention.^1,2 The recommended mainstay of treatment for septic peritonitis includes resuscitation and cardiovascular support, prompt treatment with broad-spectrum antimicrobials, and surgical debridement and source control.^3,4

Surgical treatment of septic peritonitis involves surgical debridement, correction of the septic source, and copious saline lavage of the abdomen, with or without employing either open or closed abdominal drainage techniques.^4–11 Despite being a mainstay of treatment, the efficacy of peritoneal lavage on clearance of bacteria from the abdominal cavity has not been proven, nor has it been shown to significantly impact survival in veterinary studies. A 2012 study by Swayne et al. evaluating the effect of peritoneal lavage on bacterial cultures in dogs with suspected septic peritonitis did not find a statistically significant number of dogs with a decrease in bacterial concentration postlavage, nor did they find a significant effect on survival for dogs with a change in pre- to postlavage cultures.¹² However, this study found positive prelavage cultures in only 14 of 33 cases (42%), which is lower than reported in other studies ranging from 82 to 100%.^{5–8,10,12,13}

Treatment of septic peritonitis often involves initiation of broad-spectrum antibiotic therapy prior to obtaining results of peritoneal bacterial culture and susceptibility. A previous study examining the effect of tympanic cavity evacuation and flushing on microbial isolates during total ear canal ablation with lateral bulla osteotomy in dogs showed that different susceptibility patterns of isolate pairs were detected in pre- and postlavage samples. Furthermore, results showed that bacteria remain in the tympanic cavity after evacuation and lavage and samples taken after tympanic cavity evacuation and lavage are likely to reveal novel isolates.¹⁴

In human medicine, appropriate empirical antimicrobial therapy has been established as a factor impacting survival for patients with sepsis.^15–18 The definition of empirical treatment is defined as that administered prior to microbiological documentation of infection.¹⁵ Appropriate antimicrobial therapy is typically defined as at least one of the drugs used in the empirical antimicrobial treatment administered within the first 24 hr of diagnosis as effective against the pathogen(s) isolated based on antimicrobial susceptibility results. Further, appropriate antimicrobial therapy assumes appropriate dose and route of administration.¹³ Despite being shown to significantly impact both short- and long-term outcomes in human septic patients, a 2015 study failed to show a significant effect on survival to discharge associated with appropriate empirical antimicrobial therapy in the treatment of canine septic peritonitis.¹³

The objective of this study was to evaluate and compare the differences in bacterial identity and susceptibility in samples obtained from the peritoneal surface on initial entry (prelavage) and after evacuation and lavage (postlavage) in dogs and cats who underwent exploratory laparotomy for a septic peritoneal effusion. Our second objective was to assess empirical broad-spectrum antimicrobial selection in comparison with bacterial culture results with regard to effect on short-term survival.

We hypothesize that (1) the postlavage peritoneal bacterial culture isolates and antimicrobial susceptibility patterns will be different compared with the initial peritoneal bacterial culture and susceptibility, (2) initial broad-spectrum antibiotic choice compared with the susceptibility of either culture will not affect clinical outcome, and (3) bacteria will remain in the peritoneal cavity after evacuation and lavage.

Materials and Methods

Data was collected from 35 patients (29 dogs, 6 cats) admitted to The Animal Medical Center between January 2011 and December 2015 for surgical treatment of septic peritonitis. Informed client consent was obtained prior to inclusion in the study, and the study protocol was approved by the Institutional Animal Care and Use Committee. Cats and dogs of any age, breed, gender, and body weight requiring exploratory laparotomy for treatment of a septic abdomen were included.

Animals without prelavage and postlavage bacterial (anaerobic and aerobic) culture and susceptibility results were excluded from the study. Animals who were not treated with empirical antimicrobial therapy or who had absence of evidence of septic peritonitis on abdominal explore were additionally excluded from the study.

Data collected for each case included age, breed, sex, weight, source of peritoneal contamination, and survival to discharge. Number, type, and susceptibility of microbial isolates obtained on prelavage and postlavage cultures were recorded.

The diagnosis of septic peritonitis for inclusion in the study was made on the basis of one or more of the following: preoperative peritoneal fluid cytology reported by the pathologist as consistent with septic peritonitis (septic exudate with intracellular bacteria), surgical findings consistent with septic peritonitis such as contamination from a ruptured viscus, or a positive bacterial culture of the peritoneal fluid.

All cases were treated with exploratory laparotomy by a board-certified veterinary surgeon or surgical resident. In all cases, abdominal explore was conducted via ventral midline celiotomy. The septic source was debrided or repaired, and the abdomen was copiously lavaged with sterile saline. The use of abdominal drainage techniques was at the discretion of the surgeon. All cases were treated with broad spectrum antimicrobials both perioperatively and postoperatively, and antibiotic choice was at the discretion of the surgeon or attending clinician when treatment was initiated before transfer to surgery. For cases in which a suspected diagnosis of septic peritonitis was made preoperatively, broad-spectrum antibiotics were initiated immediately, prior to transfer to surgery. If the diagnosis of suspected septic peritonitis was not made until the time of abdominal explore, the patient was treated with Cefazolin (22 mg/kg IV q 90 min) perioperatively, and further empirical antimicrobials were added at the surgeon’s discretion immediately following surgery. All patients received at least one dose of intravenous antibiotics prior to the start of surgery.

Aerobic and anaerobic bacterial culture samples were collected from the peritoneal surface of the abdominal body wall on initial approach to the abdomen (prelavage) and following copious saline lavage prior to abdominal closure (postlavage). Each culture was obtained in a standardized manner by swabbing the peritoneum with a commercial cotton-tipped swab in one swipe from the dorsal body wall to the ventral body wall. Care was taken to avoid contamination of the culture swab by skin or other surfaces. Once the sample was obtained, the cotton-tipped swab was immediately placed into a commercial Aimes gel culture collection and transport system^a and stored at room temperature until shipment to the laboratory.

The samples were submitted for aerobic and anaerobic bacterial culture and susceptibility testing. All samples were shipped overnight to an off-site laboratory. Samples obtained Sunday through Friday were plated the following morning. Samples obtained on Saturday were plated on Monday. All laboratory procedures followed guidelines provided by the Clinical Laboratory Standards Institute.

All samples were plated to Tryptic Soy Blood agar and MacConkey agar. Each sample was also given a thioglycollate enrichment broth to aid in rare bacteria recovery because the samples were collected from a sterile source. The blood plate was placed in 5% CO₂ at 35°C, and the MacConkey agar and thioglycollate enrichment broth were placed in a 35°C air incubator. The plates were examined for growth at 24, 48, and 72 hr. If bacterial identification was not made on the bench based on the colony morphology and biochemicals, the isolate was either sent to the Vitek^b or a BD Crystal may have been used. Every isolate was reported given that the cultures were taken from a presumably sterile source.

The susceptibilities were performed by the automated Vitek. In using the Vitek automated system, specially designed veterinary cards were used, which test the most applicable antibiotics for that particular organism and take into account the source and animal from which the culture was collected. If drugs that were not able to be placed on the card became available, interpretative criteria was provided based on making a Kirby-Bauer plate for the organism and drug.

The two bacterial culture and susceptibility results (isolates and antimicrobial susceptibility patterns) were compared for each patient with regard to bacterial isolates and bacterial susceptibility scores. Bacterial susceptibility scores were calculated for each culture as previously described.¹⁴ Briefly, the susceptibility score for each isolate from each peritoneal culture was calculated based on the number of drugs in the antibiogram to which the isolate was designated as susceptible as follows: 1 = susceptible to ≥80% of drugs in the antibiogram, 2 = susceptible to 50–79%, and 3 = susceptible to ≤50%. The susceptibility score for each culture was calculated as the sum of the susceptibility scores for the individual isolates.

The empirical antimicrobial selection was evaluated and compared with the antibiotic susceptibility results of both the initial and postlavage peritoneal culture and susceptibility. Empirical antimicrobial selection was considered appropriate if all isolates from the peritoneal culture were susceptible to the chosen drugs. For cases with multiple empirical antimicrobials administered, empirical antimicrobial selection was considered appropriate if all isolates from the peritoneal culture were susceptible to at least one of the chosen drugs. Empirical antimicrobial selection was considered inappropriate if at least one isolate was resistant to all the antimicrobials administered. The clinical outcome, defined as survival to discharge, was evaluated with regard to antibiotic choice and bacterial culture and susceptibility results.

Data was assessed to be either categorical, ordinal, or nonparametric discrete numerical data. Confidence intervals (95%) for proportions in categorical data were calculated based on a binomial distribution. Distributions in categorical data were compared by the χ² test of association or by Fisher exact test. Numerical data was compared between categories by either the Wilcoxon rank sum test for independent data or the Wilcoxon sign rank test for paired (pre-versus post-) data. P values ≤.05 were considered statistically significant.

Results

In total, 35 patients met the inclusion criteria. Of these, 29 dogs and 6 cats were included. There were 10 spayed female dogs, 3 intact female dogs, 13 castrated male dogs, and 3 intact male dogs. There were three spayed female cats, two castrated male cats, and one intact male cat. The median age for dogs was 8.1 yr (range 0.8–16.3 yr) and for cats was 11.0 yr (range 3.1–13.0 yr). Median body weight of dogs was 7.2 kg (range 2.0–48.0 kg) and of cats was 4.2 kg (range 2.8–4.8 kg).

Dog breeds included in the study were 5 mixed breeds, 3 golden retrievers, 3 Labrador retrievers, 2 Cavalier King Charles spaniels, 2 Yorkshire terriers, and 1 each of 14 additional breeds (Appenzeller, Coton de Tulear, Dogue de Bordeaux, French bulldog, Havanese, Jack Russell terrier, Maltese, miniature dachshund, miniature pinscher, miniature schnauzer, Pekingese, Pomeranian, shih tzu, and toy poodle). Feline breeds included three domestic shorthairs, one Maine coon, one Persian, and one Sphynx.

The source of abdominal contamination was the gastrointestinal tract in 24 cases (68.6%), the urogenital tract in 7 cases (20%), and the hepatobiliary system in 4 cases (11.4%).

Twenty-five patients (71%) were treated with antibiotics prior to transfer to surgery. Of these patients, 10 were started on antibiotics the day of surgery by the receiving service when the diagnosis of septic peritonitis was made or suspected. The remaining 15 patients were currently being treated with previously prescribed antibiotics at the time of diagnosis of septic peritonitis and subsequent abdominal exploratory surgery.

Microbiological growth occurred in 74.3% (26/35) of samples obtained prelavage. Organism growth occurred in 20/29 (69.0%) of dogs and 6/6 (100%) of cats. There was no significant difference in percentage of prelavage cultures with positive growth between dogs and cats (P = .304). Microbiological growth occurred in 74.3% (26/35) of samples obtained postlavage. Overall, microbial growth occurred in at least one culture (either pre- or postlavage) in 88.6% of patients.

Organisms isolated from each case are shown in Table 1. Overall, prelavage culture results revealed no growth in 9 cases (25.7%), Escherichia coli in 13 cases (37.1%), Enterococcus sp. in 7 cases (20.0%), Staphylococcus sp. in 5 cases (14.3%), Enterobacter sp. in 2 cases (5.7%), Streptococcus sp. in 2 cases (5.7%), Proteus sp. in 1 case (2.9%), Klebsiella sp. in 1 case (2.9%), Candida sp. in 1 case (2.9%), and Yeast sp. in 1 case (2.9%). Postlavage culture results revealed no growth in 9 cases (25.7%), E coli in 11 cases (31.4%), Enterococcus sp. in 9 cases (25.7%), Staphylococcus sp. in 3 cases (8.6%), Enterobacter sp. in 3 cases (8.6%), Streptococcus sp. in 3 cases (8.6%), Proteus sp. in 3 cases (8.6%), Klebsiella sp. in 1 case (2.9%), Candida sp. in 2 cases (5.7%), and Clostridium perfringens in 1 case (2.9%). There was no significant difference in bacterial isolates pre- versus postlavage (P = .984).

TABLE 1 Source of Abdominal Contamination, Pre- and Postperitoneal Lavage Culture Results, Pre- and Postperitoneal Lavage Susceptibility Scores, and Empirical Antimicrobials Administered

View Fullscreen

Bacterial susceptibility scores for each case are shown in Table 1. The median bacterial susceptibility score was 2 (range 0–8) for prelavage culture results and 2 (range 0–6) for postlavage culture results. There was no significant reduction in bacterial susceptibility score following peritoneal lavage (P = .797).

The overall survival rate was 74.3%. Twenty-three (79.3%) of twenty-nine dogs and three (50.0%) of six cats survived and were discharged from the hospital. There was no significant difference in survival rates between species (P = .162).

The survival rate for patients with positive prelavage microbial growth was 69.2% and for patients with no microbial growth on prelavage culture was 88.9%. There was no significant difference in survival based on having a positive versus negative prelavage culture (P = .391). The survival rate for patients with positive postlavage microbial growth was 73.1% and for patients with no microbial growth on postlavage culture was 77.8%. There was no significant difference in survival based on having a positive versus negative postlavage culture (P = 1.000).

The empirical antibiotics chosen for each case and whether empirical antimicrobial selection was appropriate based on pre- and postperitoneal lavage culture and susceptibility results for each case are shown in Table 1. The overall percentages of bacterial isolates susceptible to the most commonly chosen empirical antibiotics are depicted in Figure 1. Of the antimicrobials commonly used, those most frequently effective against the cultured isolates were Cefotaxime, Ceftazidime, and Imipenem. Based on prelavage culture results, empirical antimicrobial selection was appropriate in 14 of 26 cases (53.8%) for which antimicrobial susceptibility results were available. Of patients with appropriate empirical antimicrobial selection based on prelavage cultures, 11 of 14 cases survived to discharge (78.6%) compared with 7 of 12 patients (58.3%) with inappropriate antimicrobial selection. Survival rates for patients with appropriate versus inappropriate antimicrobial selection based on prelavage susceptibility results were not significantly different (P = .401). Based on postlavage culture results, empirical antimicrobial selection was appropriate in 16 of 26 (61.5%) cases for which antimicrobial susceptibility results were available. Of patients with appropriate empirical antimicrobial selection based on postlavage cultures, 13 of 16 cases survived to discharge (81.3%) compared with 6 of 10 patients (60.0%) with inappropriate antimicrobial selection. Survival rates for patients with appropriate versus inappropriate antimicrobial selection based on postlavage susceptibility results were not significantly different (P = .369).

View Figure

• Download as Powerpoint
• Download Figure

Discussion

Findings of this study showed no significant difference in microbial isolates or susceptibility score pre- versus postperitoneal lavage. Similarly, in the study by Swayne et al., the number of dogs who had a decrease in concentration of bacteria cultured from pre- to postlavage samples was not significant.¹² However, in contrast to this study, our study found a higher proportion of cases with positive microbial growth in at least one culture in 88.6% of patients. This finding is more consistent with percentages reported in other studies.^5-8,10,13 A potential reason for a higher percentage of positive culture results seen in our study could be attributed to the more stringent inclusion criteria used. In addition to the criteria used in our study, cases were included in the Swayne et al. study if they had sonographic evidence of free abdominal fluid in the presence of a foreign body or a mass or following previous intestinal surgery, or if abdominal fluid cytology showed suppurative inflammation, including cases without evidence of intracellular bacteria.¹² Other potential causes may include differences in sample handling, study populations, and in preoperative treatment protocols between different institutions.

Results of this study showed no significant difference in survival rates for cases with positive versus negative pre- or postlavage cultures. Further, survival was not significantly affected by appropriate empirical antibiotics based on either pre- or postlavage culture results. These findings are consistent with those of Dickinson et al. but contradictory to the benefit of appropriate empirical antibiotics in treatment of sepsis in people.^13,15-18 It is possible that the lack of significant difference in survival rates for animals treated with appropriate versus inappropriate empirical antibiotics is due to sample size in our study. Based on the observed difference in survival rates of patients receiving appropriate empirical antimicrobials (81%) and patients receiving inappropriate empirical antimicrobials (60%) based on postlavage culture and susceptibility results, 73 patients in each group (146 patients total) would be required to have 80% power to detect a significant difference in these rates with an alpha error of 0.05. Likewise, differences in survival rates for cases with appropriate empirical antibiotics (78%) versus inappropriate empirical antibiotics (58%) based on prelavage culture and susceptibility results may have been statistically significant with larger case numbers. Further studies with larger study populations should be performed to evaluate the effect of appropriate empirical antibiotics on survival rates for canine and feline patients with septic peritonitis. Additional possible explanations for the differences seen between the veterinary studies in comparison with the human literature include species differences and that all cases of sepsis in our study were due to surgical sources. Human medical literature has shown that the beneficial effect of appropriate empirical antimicrobial therapy is greater for nonsurgical sepsis cases than for cases of surgical sepsis.¹⁵

The most common bacterial isolates both pre- and postlavage were E coli followed by Enterococcus sp. Given that nearly 70% of cases in this study had the gastrointestinal tract as their septic source, this correlates with the expected enteric bacterial population. Likewise, E coli and Enterococcus sp. were the first and second most common bacteria isolated in two other recent studies on canine septic peritonitis.^12,13 The empirical antibiotics chosen in this study were selected as they are typically effective against the expected bacterial populations. However, with growing concerns for antimicrobial resistance, culture and susceptibility results will likely continue to become increasingly important in the treatment of septic peritonitis. Therefore, we recommend performing bacterial culture and susceptibility testing in all cases of septic peritonitis so that antimicrobial therapy can be tailored based on these results in cases that are not responding to initial empirical treatments.

Empirical antibiotics were defined as being inappropriate if at least one bacterial isolate was resistant to all antibiotics used in the initial treatment plan, which is consistent with the definition used in previous studies.¹³ However, in addition to antimicrobial susceptibility found with in vitro laboratory testing, several additional confounding variables exist that could influence clinical outcome independently from susceptibility results. One such example would be bacterial virulence factors, such as endotoxin, which can be associated with a greater degree of patient morbidity regardless of antimicrobial susceptibility. Additionally, host immunocompetence should be considered, as it is likely that in the majority of immunocompetent animals, the susceptibility pattern will not affect clinical outcome as long as the septic source is effectively treated with surgical source control and thorough abdominal lavage. In cases in which multiple antimicrobials were used, drug synergism can also lead to increased efficacy of the drugs in vivo. The effects of such factors on clinical outcome were not evaluated in this study. In addition, for certain types of bacteria, it has been shown that in vitro susceptibility results do not consistently predict in vivo efficacy (e.g., when considering fluoroquinolones for treatment of enterococcal infections). This additional information must also be taken into account when choosing appropriate antimicrobials based on culture and susceptibility results.

Five cases in this study had negative prelavage but positive postlavage culture results. A plausible explanation for this finding was that these cases had a walled-off, localized source of peritonitis prior to surgery. Following surgical repair of the septic source and peritoneal lavage, the bacteria were likely spread throughout the abdomen, leading to positive microbial growth on postlavage cultures. Another possible explanation is inappropriate sample handling of the prelavage sample, leading to a false-negative result.

Based on our study, of the antimicrobials commonly used, those most frequently effective against isolates cultured from septic peritonitis cases were Cefotaxime, Ceftazidime, and Imipenem, as shown in Figure 1. Given these findings, thought may be given to choosing one of these antibiotics empirically for treatment of septic peritonitis in the perioperative period, while culture and susceptibility results are pending. However, our study also showed no difference in survival rate for cases in which empirical antimicrobial selection was inappropriate. We suspect that this is because the source of contamination was removed with surgery, and any residual peritoneal bacteria are at low enough numbers to be controlled by the patient’s natural immune defenses. However, it is possible that the lack of significant difference in survival with appropriate versus inappropriate empirical antibiotics is due to the small sample size and low power for data analysis, and therefore this could be found to significantly impact outcome if a study was performed with larger case numbers. The antibiotics found to be most frequently effective against the isolates cultured are typically reserved for cases with documented antimicrobial resistance to the more commonplace and lower-generation antimicrobials and are of antimicrobial classes that are critically important in human medicine. If choosing one of these antibiotics for empirical treatment, thought must be given to the possible implications of inducing bacterial resistance to these stronger and newer-generation antibiotics. The practitioner must be responsible in their decision to use these higher-generation antimicrobials, and this choice should only be made in light of clinical decline despite treatment with lower-generation antimicrobials or based on documented antimicrobial resistance determined by bacterial culture and susceptibility testing.

Limitations of this study include a small sample size, lack of standardization of surgical technique among surgeons, and lack of standardization of antimicrobial protocols. Additionally, bacterial load pre- and postlavage could not be calculated, as the sampling technique did not allow for bacterial quantification. It is possible that the lack of significant difference between pre- and postlavage culture results and survival rates for cases with positive versus negative cultures as well as appropriate versus inappropriate empirical antimicrobial therapy is due to low statistical power because of the small sample size. However, cases were collected over a long period (5 yr) at a busy referral hospital with a high caseload. Therefore, to achieve higher case numbers in future studies, thought should be given to a multi-institutional study design. Strict inclusion criteria were also chosen for our study, which may have limited the number of cases eligible for enrollment.

Antimicrobial protocols were not standardized in this study. For patients in whom a diagnosis of suspected septic peritonitis was made preoperatively, broad-spectrum antibiotics were initiated immediately, prior to transfer to surgery. However, in other cases, the diagnosis of suspected septic peritonitis was not made until the time of abdominal explore. These cases were treated with Cefazolin perioperatively, and additional empirical antimicrobials were added at the surgeon’s discretion immediately postoperatively. Additionally, some cases included in this study had been previously started on empirical antimicrobials at the time of prior surgery or based on other clinical parameters suggestive of infection but without a diagnosis of suspected septic peritonitis. Whereas all cases received at least one dose of intravenous antimicrobials prior to the initial skin incision, there was significant variation in preoperative antimicrobial administration. This variation could have affected both culture and susceptibility results as well as clinical outcome. No conclusions could be made on the effect of timing of initiation of empirical antibiotics in relation to surgery on culture results or clinical outcome. Further studies should be performed to evaluate this effect.

An additional limitation in our study was that the amount of lavage solution used was not standardized among surgeons, and the definition of copious lavage was at each individual’s discretion. The amount of saline used was not routinely recorded and therefore conclusions regarding the effect of volume of lavage on peritoneal culture results could not be made. Further studies would be needed to evaluate this effect.

Although there was no significant difference between pre- and postlavage peritoneal culture and susceptibility results when looking at the study population as a whole, approximately half the cases had at least one microbial type that was isolated from only one of the two cultures. Therefore, clinical decisions regarding antibiotic protocol could change depending on if culture samples were obtained pre- versus postperitoneal lavage. Based on these findings, we recommend obtaining cultures both pre- and postperitoneal lavage, to ensure that all organisms are identified. Clinically, if this is not financially feasible, combining pre- and postlavage cultures may increase the chances of identifying all microbes, rather than eliminating one of the two cultures, and therefore could be considered as an alternative to submitting two separate cultures. The effects of combining cultures on culture and susceptibility results was not evaluated in this study, however, and future studies should be performed to evaluate the effects of this practice.

Conclusion

In conclusion, our study found no significant difference between pre- and postlavage peritoneal cultures in isolate type or susceptibility score in cases treated surgically for septic peritonitis. Further, no significant difference in survival rate was seen for cases with appropriate versus inappropriate antimicrobial selection based on either pre- or postlavage culture and susceptibility results. Additional studies are needed to assess if clinical outcome is affected by only obtaining cultures pre- or postlavage, or by treating with empirical antibiotics not based on culture results for cases of septic peritonitis.