Evaluation of Data From 35 Dogs Pertaining to Dehiscence Following Intestinal Resection and Anastomosis
The objectives of this study were to evaluate blood and abdominal fluid lactate and glucose, fluid cytology, culture, and volume 24 and 48 hr following intestinal resection and anastomosis in dogs with and without closed-suction drains and to correlate findings with survival. Thirty-five client-owned dogs that underwent intestinal resection and anastomosis were prospectively enrolled in the study. Abdominal fluid was submitted for culture at surgery and again 24 hr postoperatively. Twenty-four and 48 hr postoperatively, blood and abdominal fluid glucose and lactate were measured and fluid was submitted for cytology. Abdominal fluid was collected either from a closed-suction drain or by abdominocentesis. Patients were followed either for 14 days or until death. Comparisons were made based on development of dehiscence and presence or absence of a drain. Patients with dehiscence were more likely to have positive cultures at 24 hr and to have had more bowel resected. Surviving patients without drains had significantly smaller differences in blood and fluid glucose and lactate both 24 and 48 hr postoperatively than surviving patients with drains. The significant differences identified between patients with and without drains suggests a need for further research into the effect of drains on abdominal fluid values.
Introduction
Intestinal resection and anastomosis (R&A) in dogs has a reported dehiscence rate of up to 16%, with mortality due to dehiscence and resultant septic peritonitis ranging from 20 to 80%.1–3 Preoperative peritonitis, hypoalbuminemia, intraoperative hypotension, and presence of a foreign body are factors that have been shown to increase the risk of dehiscence.1,2,4 One study found foreign bodies to be protective against dehiscence.3 Dehiscence most commonly occurs between 3 and 5 days, although occurrence has been reported 0–10 days postoperatively.1,2,5,6
During the immediate postoperative period, when patients are hospitalized and recovering from surgery, it would be ideal if there were an easy, inexpensive, and minimally invasive way to detect impending dehiscence or detect dehiscence as soon as it occurs so that appropriate treatment could be initiated sooner, decreasing morbidity and mortality. Unfortunately, in veterinary medicine, there is currently no way to detect impending dehiscence of an intestinal R&A.7–9 Available tests, including positive culture of abdominal fluid, only determine if dehiscence has already occurred, and those tests may have false negative or false positive results.9–11 Abdominal fluid culture also has the drawback of a delay of several days before results are available.12 If dehiscence is suspected, re-exploration of the abdomen is recommended; however, that carries a poor prognosis. In one study, two of three dogs and in another study two of two cats that were re-explored for dehiscence died perioperatively.2,10
Intestinal healing is negatively affected by local inflammation, infection, and poor blood supply, all of which may develop in the pre-, intra-, or postoperative period.6,13–15 Evaluation of abdominal fluid and blood work in the perioperative period should reflect inflammation, infection, and local ischemia. It is plausible that those factors, evaluated either individually or together, may be able to predict intestinal dehiscence, directing clinicians toward additional medical treatments or abdominal re-exploration to prevent dehiscence from occurring.7,16,17
The goals of this study were to evaluate blood and abdominal fluid glucose and lactate, fluid cytology, and volume at 24 and 48 hr postoperatively and fluid culture at 24 hr postoperatively from dogs that had undergone R&A with and without closed-suction abdominal drain placement. The study authors hypothesized that this evaluation would reveal significant differences between patients that would develop dehiscence and those that would not and that there would not be any significant differences in values between dogs with and without abdominal drains.
Materials and Methods
All cases of canine intestinal R&A performed between October 2010 and May 2012 were prospectively enrolled in the study. The hospital medical board approved the study, and informed verbal consent for enrollment in the study was obtained from the owners following surgery. Preoperative factors, including signalment, history, preoperative treatments, and preoperative blood work, were recorded. Age of patient and preoperative albumin, glucose, blood urea nitrogen, and creatinine were evaluated statistically. Intraoperative factors were recorded, including duration of anesthesia, location and length of bowel resected, reason for resection, if other procedures were performed, if intestinal perforation was present, if a culture was submitted, and if a closed-suction drain was placeda. Except for location of resection and reason for resection of these, all factors were evaluated statistically.
Surgery was performed by either one of three board-certified surgeons or one of three surgical residents. All intestinal R&As were closed with 4-0 polydioxanoneb in a simple interrupted pattern. The volume of abdominal fluid used to flush the abdomen and volume suctioned from the abdomen were not consistently recorded and were not analyzed in this study. No attempt was made to leave more or less fluid in the abdomen prior to closure. A closed-suction drain was placed and an aerobic culture takenc at the discretion of the primary surgeon. IV fluid rates, intra- and postoperative antibiotics, and duration of hospitalization varied and were determined by the primary clinician.
When abdominal drains were present, the reservoirs were either emptied and quantified q 2 hr in the immediate postoperative period or when they became full. Starting between 6 and 12 hr postoperatively, drain reservoirs were emptied and quantified approximately q 4 hr until removal, death, or euthanasia. If fluid production later increased, reservoirs were emptied more frequently. Quantification of fluid from drains was not evaluated in this study.
Approximately 24 hr postoperatively, an abdominal ultrasoundd,e was performed to look for fluid. The ultrasound was often performed just before scheduled emptying of the drain reservoir (when the suction was suspected to be lower) to try to increase the volume of retrievable fluid. The fluid volume was subjectively assessed and graded on a scale from 0 to 4. A grade of 0 was given when no fluid was seen in the abdomen. A grade of 1 was given when one small pocket of fluid was seen. A grade of 2 was given when more than one small pocket or one medium to large pocket of fluid was seen. A grade of 3 was given when several medium to large pockets of fluid were seen. A grade of 4 was given when a border of free fluid could be seen around multiple organs. Grades were assigned by one of three surgical residents. If a closed-suction abdominal drain was present, gloves were worn, the drain was clamped with a hemostat, the grenade removed, the tip of the drain swabbed with alcohol, and a minimum of 2 mL of fluid removed and discarded from the drain prior to collection of fluid for analysis (previous evaluation found drains at the authors’ hospital, which were frequently cut to a shorter length, held an average of 2 mL of fluid). Up to 6 mL of fluid was collected prior to replacing the grenade and removing the hemostat. If < 0.1 mL could be collected, this was recorded and no fluid was submitted. If a drain was not present, fluid was collected by ultrasound-guided abdominocentesis. After locating a fluid pocket, the skin was briefly scrubbed with alcohol prior to abdominocentesis. The needle was changed prior to dispensing fluid into culturettes. Abdominocentesis was performed by one of three surgical residents or one board-certified surgeon if a volume grade of 3 or 4 was assigned. Otherwise, abdominocentesis was performed by one of three board-certified internists.
Within 5 min of abdominal fluid collection, blood was drawn from a peripheral vein and both fluid and blood samples were analyzed within 10 min of blood collection for glucose and lactate levelsf. If abdominal fluid could not be obtained, blood was not collected.
Part of the remaining fluid was submitted for aerobicg and anaerobich culture and sensitivity. Lastly, four slides of unspun fluid and four slides of spun fluidi (3500 revolutions/min for 10 min for blood and 1600 revolutions/min for 10 min for fluid) were made and submitted, along with any remaining fluid, for cytologic evaluation by a single clinical pathology residentj. Fluid samples received by the laboratory were processed within 2 hr of acquisition according to standard laboratory procedure as follows. Leukocyte and red blood cell counts of each fluid were obtainedk and recorded. If the leukocyte counts were < 2 × 109 cells/L cytospin slides were prepared. If cells counts were 2–30 × 109 cells/L, sedimented slides were prepared by centrifugation of the samplel (1600 revolutions/min for 3 min). If cell counts were > 30 × 109 cells/L, direct preparations were made. If red blood cell counts were > 1 × 10, buffy coat slides were prepared.6 All slides (i.e., those premade at time of sample collection and those made in the laboratory) were routinely stained with Wright-Giemsa. All slides were microscopically evaluated and a 200 cell count differential performed and recorded. The presence of any microorganisms, neoplastic or atypical cells, and any degenerative changes were recorded. Samples were classified as pure transudate (< 2.5 × 109 cells/L and < 0.025 g/L protein), modified transudate (> 2.5 × 109 cells/L or protein > 0.025g/L) or exudate (> 5 × 109 cells/L).
At approximately 48 hr postsurgically, the above procedures were repeated, with the exception of the culture and sensitivity unless the patient had died, been euthanized, or discharged prior to the 48 hr collection time. Patients were followed until death, euthanasia, or the 2 wk recheck examination, and were categorized as either having had an intestinal dehiscence or not having had a dehiscence. If a patient died or was euthanized, the reason and the number of days from surgery were recorded. If the patient was reexamined prior to 14 days postoperatively or did not return for evaluation, their owners were contacted by phone after 14 days.
For analysis of the preoperative, intraoperative, and postoperative factors previously listed, patients were divided into the following group comparisons: (1) patients that developed dehiscence (D) and those that did not (ND), (2) patients that had a drain placed and did (DCS) or did not (NDCS) develop dehiscence, (3) patients that did not develop dehiscence and did (NDCS) or did not (NDNCS) have a drain placed, (4) patients that did not have a drain placed and did (DNCS) or did not (NDNCS) develop dehiscence, and (5) patients that developed dehiscence that did (DCS) and did not (DNCS) have a drain placed.
The Shapiro-Wilk test was used to evaluate normality for all continuous variables (i.e., age, preoperative serum albumin, glucose, creatinine, blood urea nitrogen, duration of anesthesia, length of bowel resected, subjective fluid volume, cytologic white blood cell count, paired serum, and abdominal fluid glucose and lactate). Because nearly all continuous variables were not normally distributed, the median (range) was used to describe these variables. The Wilcoxon rank sum test or the Student t test (depending on if the data were not normally or were normally distributed, respectively) was used to compare continuous variables between groups. Similarly, the paired t test or Wilcoxon signed rank test was used to compare continuous variables between time periods. The Fisher's exact test (if the expected value in any cell was < 5) or χ2 test was used to compare categorical variables (i.e., if other gastrointestinal surgery was performed, if intestinal perforation was found at the time of surgery, if cultures were positive or negative, if a closed suction drain was placed, if bacteria were seen cytologically, and if dehiscence occurred or not) between groups. A P < .05 was considered significant for all comparisons. All analyses were performed using a statistical software programm.
Results
Thirty-five cases of canine intestinal R&A were included in this study. Mean age was 6.5 yr (range, 6 mo to 12 yr). Sex of the patients included 14 castrated males, 18 spayed females, and 3 intact females. Patients were hospitalized a median of 2 days (range, 1–5 days). Reasons for intestinal R&A included neoplastic and non-neoplastic masses (n = 16), foreign bodies (n = 11), intussusception (n = 3), leaking previous surgery site (n = 3), self-trauma (n = 1), and idiopathic ischemic bowel (n = 1). Additional procedures performed included liver biopsy (n = 10), gastrotomy (n = 7), splenectomy (n = 6), lymph node biopsy (n = 6), enterotomy (n = 1), gastric biopsy (n = 1), cholecystectomy (n = 1), adrenalectomy (n = 1), ovariohysterectomy (n = 1), liver lobectomy (n = 1), and omental biopsy (n = 1).
Regions of intestine resected in dogs developing dehiscence included the mid-jejunum (n = 3), distal jejunum to transverse colon (n = 1), and distal duodenum and proximal jejunum (n = 1). Regions of intestine resected in dogs that did not develop dehiscence included proximal duodenum (n = 1), mid-duodenum (n = 2), distal duodenum (n = 1), distal duodenum to proximal jejunum (n = 1), distal duodenum to mid-jejunum (n = 1), proximal jejunum (n = 2), mid-jejunum (n = 12), mid- and distal jejunum (n = 2), distal jejunum (n = 5), distal jejunum and ileum (n = 1), mid-jejunum to ileum (n = 1), and distal jejunum to ascending colon (n = 1).
Five dogs either died or were euthanized within 14 days of surgery, all due to dehiscence (median, 3 days; range, 1–5 days). The remaining dogs were alive at the time of the follow-up examination (22 dogs) or when the follow-up phone call was made (8 dogs) at least 14 days postoperatively. Four of the five dogs were euthanized, one on days 1, 2, 4, and 5 postoperatively. Three of those patients had intestinal dehiscence noted during necropsy, and one had copious abdominal effusion and fluid cytology consistent with dehiscence (i.e., degenerative neutrophils with intracellular bacteria). One dog died 3 days postoperatively, which was < 24 hr after abdominal re-exploration. At the time of death, that patient had cytology consistent with intestinal dehiscence (i.e., numerous intracellular bacteria and intestinal debris). Three of the five patients that developed dehiscence had an R&A performed for masses (one sarcoma, one chronic hematoma, one lymphoplasmacytic enteritis) and two for foreign bodies. Rate of survival to 14 days postoperatively was 85.7%, and the rate of dehiscence was 14.3%.
Fluid was collected twice in 22 patients, once in 7 patients (because 5 of those patients had either too little or no fluid to be collected on the second day, one was discharged < 48 hr postsurgically, and one patient was euthanized < 48 hr postsurgically), and 6 patients did not have any abdominal fluid collected (because 5 patients had either too little or no fluid to be collected on either day and 1 patient was discharged < 24 hr postsurgically). Complications associated with abdominocentesis were only encountered in one patient because bowel contents were aspirated at 24 hr. A second attempt at aspiration was successful and collection at 48 hr occurred without complication. That patient was doing well when the owner was contacted by phone 21 days postoperatively.
When patients were divided into groups for comparisons, there were 5 dogs in the D group, 30 in the ND group, 10 in the NDCS group, 4 in the DCS group, 1 in the DNCS group, and 20 in the NDNCS group. Only one patient was included in the DNCS group; therefore, no comparisons were made between either the DNCS and DCS groups or the DNCS and NDNCS groups. The presence of only one patient in the DNCS group also prevented calculation of power for comparison between the DNCS group and the other groups. Assuming the ratio of animals would be consistent between groups and assuming a power of 0.8 and an α of 0.05, the number of patients needed to find a significant difference between the DNCS group and the other two groups was calculated for differences in serum and fluid glucose and lactate at 24 and 48 hr. It would be necessary to have 40 patients in the DNCS group when comparing that group to the DCS group for differences in serum and fluid lactate and glucose at 24 hr. Forty-eight and 42 patients, respectively, would be needed when comparing differences in serum and fluid lactate and glucose at 48 hr. It would be necessary to have 233 and 178 patients in the DNCS group when being compared with the NDNCS group for differences in serum and fluid lactate and glucose at 24 hr, respectively, and 75 and 3719 patients when comparing those groups for differences in serum and fluid lactate and glucose at 48 hr, respectively.
Power was calculated for the comparisons of differences in blood and fluid glucose and lactate between the other groups. Power was highest when comparing the NDCS and NDNCS groups (0.44 and 0.75 for differences in blood and fluid glucose at 24 and 48 hr, respectively, and 0.66 and 0.66 for differences in blood and fluid lactate at 24 and 48 hr, respectively). That comparison would have required the fewest patients to reach a power of 0.8, with the NDCS group requiring between 12 and 36 patients and the NDNCS group requiring between 10 and 22 patients. Power was lower when comparing the D and ND groups (0.22 and 0.1 for differences in blood and fluid glucose at 24 and 48 hr, respectively and 0.11 and 0.56 for differences in blood and fluid lactate at 24 and 48 hr, respectively). That latter comparison would have required between 32 and 367 patients in the D group and 8 and 81 patients in the ND group to reach a power of 0.8.
Preoperative blood work was evaluated for differences in total WBC count, albumin, blood urea nitrogen, creatinine, and glucose. There were no significant differences between any of the groups.
Age of patient at time of surgery, length of bowel resected, duration of anesthesia, if bowel perforation was discovered at the time of surgery, and if other gastrointestinal surgical procedures were performed were evaluated for differences between groups. The age at surgery was significantly older in the DCS group (median, 112.5 mo; range, 84–134 mo) than the NDCS group (median, 48 mo; range, 12–96 mo; P = .011). The length of bowel resected was significantly longer in the D group (median, 8 inches; range, 1–28 inches) than the ND group (median, 24 inches; range, 8–30 inches; P = .015). Duration of anesthesia was not significantly different between groups. Surviving patients that had a drain placed (the NDCS group) were significantly more likely to have had a bowel perforation that was discovered at the time of surgery than the NDNCS group (P < .001). There was no significant difference in performance of additional gastrointestinal procedures between groups.
Median subjectively assessed fluid volume scores were higher at 24 and 48 hr in the D than ND group, but that difference was not significant (Table 1). When comparing scores of abdominal effusion at 24 to 48 hr in the NDCS and NDNCS groups, the NDCS patients trended toward having significantly less volume decrease over time (P = .06).
Denotes a comparison between groups for which P = .075.
Denotes a comparison between groups for which P = .06.
Serum lactate and glucose and abdominal fluid lactate and glucose were evaluated independently and compared with each other at 24 and 48 hr postoperatively between three groups (Tables 2A, 2B). Changes in glucose and lactate over time were also compared for serum and abdominal fluid individually and compared with each other. There were no significant differences between the D and ND groups. Change in serum lactate from 24 to 48 hr and difference in blood to fluid lactate at 48 hr trended toward significance when comparing the D and ND groups (P = .074 and P = .074).
Denotes a comparison between groups for which P = .029.
Denotes a comparison between groups for which P = .011.
Denotes a comparison between groups for which P = .013.
Denotes a comparison between groups for which P = .016.
There was a significant difference in serum and fluid lactate values at 24 hr (P = .013) and 48 hr (P = .016) in the NDCS and NDNCS groups with patients with drains having significantly greater differences in lactate. There was also a significant difference in serum and fluid glucose values at 24 hr (P = .029) and 48 hr (P = .011) in the NDCS and NDNCS groups with patients with drains having significantly greater differences in glucose.
Fifteen patients had cultures of abdominal fluid taken intraoperatively. Nine of the 15 cultures were positive. Four of the cultures grew two bacteria and five cultures grew one. Bacteria cultured included Enterococcus spp. (seven dogs), Escherichia coli (three dogs), Pasteurella spp. (one dog), and Kluyvera cryocrescens (one dog). Six of the bacteria cultured were resistant to six or more antibiotics. Three of the five patients that developed dehiscence had positive cultures at surgery. The cultures of two of those patients grew Escherichia coli and Enterococcus spp. and the other patient’s culture grew Kluyvera cryocrescens. One culture of bile and choleliths was also taken and grew highly resistant Enterobacter sakazakii and Enterococcus spp. There were no significant differences in cultures taken at the time of surgery between any groups.
Twenty-nine patients had cultures of abdominal fluid taken at 24 hr postoperatively. Of those 29 cultures, 16 were positive. Of those cultures, 12 grew one bacterium, three grew two, and one grew four. Bacteria cultured included Staphylococcus pseudintermedius (four dogs), Escherichia coli (four dogs), Staphylococcus schleiferi (three dogs), Enterococcus spp. (two dogs), Pseudomonas aeruginosa (two dogs), Morganella morganii (one dog), Clostridium spp. (one dog), Enterobacter cloacae (one dog), nonpathogenic Diptheroid spp. (one dog), Micrococcus spp. (one dog), nonpathogenic Bacillus spp. (one dog), and Enterobacter sakazakii (one dog). Seven of the bacteria cultured were resistant to six or more antibiotics. All five of the patients that developed intestinal dehiscence had positive cultures at 24 hr postoperatively. The cultures of three of those patients grew one bacteria each (Escherichia coli in two cases and Enterobacter species in one case), one culture grew two bacteria (Morganella morganii and Clostridium spp.), and one culture grew four bacteria (Escherichia coli, Pseudomonas aeruginosa, Enterococcus spp., and coagulase-positive Staphylococcus pseudintermedius). There were significantly more positive cultures in the D group than the ND group at 24 hr (P = .044).
Fourteen patients had cultures taken at both times, and six of those patients were positive at both times (three of those developed dehiscence). It approached significance for the D group to have positive cultures at both time points compared with the D group (P = .055).
Only two patients grew the same bacteria in both cultures (Escherichia coli in one patient that developed dehiscence, Enterobacter sp. in a patient that did not). Four patients grew different bacteria, two of which developed dehiscence (Escherichia coli and Enterococcus spp. at surgery, Morganella morganii and Clostridium spp. at 24 hr; and Kluyvera cryocrescens at surgery, Enterobacter spp. at 24 hr), and two of which did not (Enterococcus spp. at surgery, coagulase-positive Staphylococcus pseudintermedius at 24 hr; and Escherichia coli at surgery and coagulase-positive Staphylococcus pseudintermedius at 24 hr).
Four of five patients that developed dehiscence had positive cultures at one or both sampling times for Escherichia coli, Enterobacter spp., or Enterococcus spp. Patients were significantly more likely to develop dehiscence if they cultured enteric bacteria (Escherichia coli, Enterococcus spp., Enterobacter spp.) at any time point (P = .026)
Total cell count for abdominal fluid at 24 and 48 hr was compared between groups and no significant differences were found (Table 3). Patients in the D group (mean, 30,700 cells; range, 4200–779,000 cells) had higher cell counts than the ND group at 24 hr (mean, 14,600 cells; range, 600–77,000 cells) and 48 hr. In the D group, mean was 36,200 cells and range was 32,800–39,600 cells, and in the ND group, mean was 8500 cells and the range was 1400–70,800 cells).
Abdominal fluid was neutrophilic in all patients in which it was collected. The percentage of neutrophils in the abdominal fluid at 24 and 48 hr was compared between groups and no significant differences were found (Table 3).
Bacteria were seen cytologically in only four patients, all at 24 hr. Two of those patients developed intestinal dehiscence and two did not. Degenerative neutrophils were seen in two of these four patients, both at 24 hr, one of which developed dehiscence and one of which did not. One of the two patients that developed dehiscence had a positive culture at surgery and at 24 hr. The other patient was not cultured at surgery, but had a positive culture at 24 hr, as well as intracellular bacteria and degenerative neutrophils seen cytologically at 24 hr. The two patients that did not develop dehiscence had negative cultures at surgery. One patient had intracellular bacteria seen on cytology but that was not cultured at 24 hr. The other patient had extracellular bacteria and degenerative neutrophils seen cytologically at 24 hr and grew a nonpathogenic Bacillus spp. at 24 hr.
Discussion
Although the results of this study did not show significant differences in abdominal fluid to blood glucose or lactate values, abdominal fluid volume, or cytology in patients that did and did not develop dehiscence, the number of patients was small and the power was low. Many of the comparisons trended toward significance, suggesting that a study with a larger number of patients might find significant differences that could be used to create cut-off values to guide the decision to surgically re-explore the abdomen. Based on the authors’ analyses, it may be necessary to have several hundred patients enrolled in a study to find significant differences between the D and ND groups’ blood and fluid glucose and lactate. This study was short in duration because it was based around the mean historical hospitalization time of about 48 hr for patients following intestinal R&A. A study that extended past 48 hr postoperatively may also be more likely to find differences between patients that do and do not develop dehiscence considering that dehiscence occurred a mean of 3 days postoperatively in this study and 3–5 days in other reports.1,2,5,6
Length of bowel resected is infrequently reported. The results of this study suggest that resection of longer lengths of bowel may be associated with an increased risk of dehiscence, which differs from one previous study that found no association between length resected and survival.2 Although this study found no association between age and dehiscence when comparing the D and ND groups, there was a significant difference between the DCS and NDCS groups. Again, that finding may be due to the small number of patients in this study. Age has not been associated with an increased risk of dehiscence in previous studies.2,4 The NDCS group of patients was more likely to have perforations of the gastrointestinal tract than the NDNCS group of patients. That finding makes sense because patients with perforation and gross abdominal contamination would have been more likely to have drains placed.
Blood to abdominal fluid glucose and lactate cut-off values have been successfully determined for diagnosis of preoperative cases of septic peritonitis.12,14,18 Poor blood flow to the intestines results in increased anaerobic metabolism and resultant increased abdominal fluid lactate and decreased abdominal fluid glucose concentrations. Lactate is also increased, and glucose decreased secondary to bacterial infection.13,18 There is a strong correlation between blood and serum lactate and glucose values and preoperative septic peritonitis.11,14,18 A difference in serum lactate and abdominal fluid lactate of < −2.0 mmol/L is 100% sensitive and 63% specific for septic peritonitis, abdominal fluid lactate > 2.5 mmol/L is 91% sensitive and 100% specific for septic peritonitis, and a difference in serum glucose and abdominal fluid glucose > 1.1 mmol/L is 100% sensitive and 100% specific for septic peritonitis in dogs.11,18 It was not possible to determine cut-off values for blood to abdominal fluid glucose and lactate for predicting postoperative dehiscence and need for abdominal re-exploration in the patients included in this study. The difference in blood to fluid lactate at 48 hr and the change in blood to fluid lactate from 24 to 48 hr trended toward significance, warranting further evaluation in future studies.
Fluid volume subjectively evaluated at 24 and 48 hr postoperatively found that patients that developed dehiscence had more fluid than patients that did not and that patient with drains had more fluid than patients without drains, although those differences were not significant. There are several limitations to the fluid volume evaluation in this study. Lavage fluid used in surgery was not standardized and abdominal drains were present and regularly emptied in many of the patients, which could have significantly affected fluid volume. Most of the patients that developed dehiscence had drains placed; however, it is not possible to know how much of the fluid production was related to dehiscence compared with inflammation secondary to the presence of a drain. In addition, a validated technique for evaluating fluid volume ultrasonographically was not used, and multiple individuals assessed fluid volume in included patients. To the authors’ knowledge, there are no papers evaluating fluid volume ultrasonographically postoperatively after R&A, only after trauma.19
Past studies have suggested that cytologic evaluation of abdominal fluid, via either abdominocentesis or diagnostic peritoneal lavage, can be used to gauge the intra-abdominal environment pre- and postoperatively. This study found no significant difference in WBC numbers between any of the groups although the D group did have a higher median WBC count than the ND group, which was consistent with previous studies. Abdominal fluid cytology in a normal dog collected by diagnostic peritoneal lavage should reportedly have a WBC count < 0.5 × 109 cells/L with nondegenerate neutrophils predominating.7,13,20 WBC count has been shown to increase in the postoperatively period after celiotomy alone (15.1 × 109 cells/L 1 day postoperatively), intestinal R&A (18.0 × 109 cells/Lin the first 3 days postoperatively) and to be highest after experimental R&A dehiscence (31.7 × 109 cells/L 1 day postoperatively and 686.0 × 109 cells/L 2 days postoperatively).12,13 Unfortunately, those studies were primarily experimental, and in most cases had collected fluid via diagnostic peritoneal lavage, making it difficult to compare those results to this study.12,13,20 One human study suggested using a cut off WBC count > 5.0 × 109 cells/L along with the appearance and smell of the fluid and gram stain results to surgically re-explore the abdomen, with some success.7
The authors of the current study found that significantly more patients that developed dehiscence had a positive culture at 24 hr. That finding correlates with preoperative sepsis being a risk factor for dehiscence, which was shown in previous papers.1,2,4 It also suggests that culture results could help predict dehiscence as early as 24 hr postoperatively. Unfortunately, the clinical utility of a positive culture is limited due to delay in results. It should also be noted that, in this study, patients developing dehiscence did so within 5 days of surgery. It is unclear if patients with dehiscence over 5 days postoperatively may be less likely to have positive culture results at 24 hr.
This study found no correlation between visualization of bacteria on cytology of abdominal fluid and dehiscence. This may have been due to the small number of cases of dehiscence and to the limited duration of the stud. Three of the patients developed dehiscence after 48 hr, and bacteria may have been visualized in those patients if the study duration had been increased. Two of the four cases with bacteria seen on cytology at 24 hr did not progress to intestinal dehiscence. Thus, positive cytology alone cannot be used as criteria to surgically re-explore the abdomen. Bacteria in the abdominal cavity postoperatively can originate from the intestinal tract during surgery, from the intestinal tract postoperatively through incisional dehiscence, or from other sources, such as from either an incisional infection or abdominal drain.12,21 False-positive and false-negative results have been shown postoperatively in other studies.12,19,22 There was a high prevalence of gram-positive Staphylococcus spp. cultured at 24 hr compared with predominately gram-negative bacteria, such as Escherichia coli and Enterobacter spp. and other enteric bacteria such as Enterococcus spp. at the time of surgery. That finding suggests that either contamination of the abdomen perioperatively or of the culture sample at the time of collection may have occurred. It is interesting to note that even though cultures rarely grew the same bacteria at both sample times, patients that cultured enteric bacteria at any point were more likely to develop dehiscence.
To date, veterinary medicine has focused on evaluation of different abdominal wall closure techniques to decrease risk of dehiscence or detect dehiscence early. Closure techniques include closed suction drains, open abdominal drainage, sump Penrose drainage, planned abdominal re-exploration, and vacuum-assisted closure, which has had some success in human medicine.10,23–26 So far, results showing one closure method as superior to another has been lacking. 10,23,25,27 It is unclear how the presence or absence of a drain affects the intra-abdominal environment. In the current study, there were significant differences in blood to fluid glucose and lactate between patients that survived with and without a Jackson-Pratt drain, with patients without drains having lower differences, or “more normal” values, at both 24 and 48 hr. Those results are similar to a recent study that looked at the effect of Jackson-Pratt drains on glucose and lactate postceliotomy, finding that after the first 2–4 days values became consistent with a septic abdomen.12 The differences in fluid parameters between patients with and without drains may be due to either a change in the abdominal environment or a change within the drain tubing. The study authors attempted to eliminate the chance of the differences being due to a change within the drain tubing by discarding a minimum of 2 mL of fluid prior to analysis. The study authors suspect that the presence of an abdominal drain alters the environment within the abdomen resulting in values consistent with sepsis. It is also possible that there was a selection bias, with more significant abdominal pathology leading to the decision to place a drain, regardless of ultimate survival or death.
A recent study evaluated vacuum assisted closure after surgery on dogs with septic abdomens. Although there were only three surviving patients, it is interesting to note that their blood to abdominal fluid glucose and lactate differences improved to within normal values within 2 days postoperatively.25 The values were most similar to those of the patients in the NDNCS group, all of which had “normal” differences between their serum and fluid lactate and glucose at 24 and 48 hr. It may be that, for an unknown reason, vacuum-assisted closure with continuous suction may be less likely to cause changes consistent with a septic abdomen than Jackson-Pratt closed-suction drains. Alternately, vacuum-assisted closure may be more successful at removing fluid, resulting in either fluid spending less time sitting in the abdomen or drain where it could be developing falsely elevated lactate or decreased glucose.
A test to determine what patients would develop dehiscence or had a high probability of dehiscence would be useful. The authors did not find a single parameter that could reliably predict those cases that would develop dehiscence. Some future areas of research that have shown promise in human medicine include evaluation of intra-abdominal pressure and the use of microdialysis catheters at the resection site to measure glucose, lactate, lactate/pyruvate ratio, and inflammatory cytokines.16,17,28,29 Work in this area has shown that lactate/pyruvate ratios during the first 2 days postoperatively may be able to predict future dehiscence and that positioning of the catheter near the R&A site yields more reliable results.
The primary shortcoming of this study is the small number of cases, which may have precluded differences in blood to abdominal fluid glucose and lactate from reaching significance and precluded comparison between certain groups, such as patients that died that did and did not have drains. Having > 2 days of evaluation may have improved results. The evaluation of fluid volume was also subjective and difficult to standardize.
Further experimental or prospective clinical studies with larger numbers of cases should be undertaken to determine if cut-off values for blood to abdominal fluid glucose and lactate and abdominal fluid volume exist, and to evaluate abdominal fluid lactate/pyruvate levels and cytokine levels. A study to evaluate the impact of closed-suction drains on abdominal fluid values after intestinal surgery would also be useful to resolve the uncertainty over the trustworthiness of values obtained from a closed-suction drain.
Conclusion
This study did not show a difference in any values between patients that did or did not have intestinal dehiscence, with the exception that patients with intestinal dehiscence were significantly more likely to have a positive culture of abdominal fluid at 24 hr postoperatively and to have had a longer section of bowel resected. There was, however, a significant difference in blood to abdominal fluid glucose and lactate in patients with and without drains, suggesting further research is needed regarding the impact of closed-suction drains on postoperative abdominal fluid values.
Contributor Notes
E. Mouat's updated credentials since article acceptance are VMD, DACVS.
E. Mouat's present affiliation is Puget Sound Veterinary Referral Center, Tacoma, WA.
