Bacteriology of Acute Appendicitis and Its Implication for the Use of Prophylactic Antibiotics

Patients and Methods

From January 1 to December 31, 2010, all patients with suspected acute appendicitis who underwent appendectomy were recruited prospectively. Incidental appendectomies and removals of normal appendices were excluded from the bacteriologic studies. The study was held at the National Yang-Ming University Hospital, located in north I-Lan County, where the agricultural population makes up 24% of all 220,374 residents. In this region, the hospital is the only place where 24-h emergency medical services and emergency surgery are provided. Institutional Review Board approval was obtained before acquisition of the patient health information.

Prophylactic antimicrobial drugs were given according to the national guidelines. For noncomplicated appendectomy, the Infectious Disease Society of the Republic of China and the Taiwan Surgical Association recommend cefoxitin, cefmetazole, or cefazolin + gentamicin + metronidazole. Alternative antibiotics are clindamycin + gentamicin, amoxicillin-clavulanic acid, or amoxicillin-sulbactam [20]. Thus, the operating surgeon can choose which antibiotic(s) to give because the guidelines offer a variety of options. The antibiotics usually were given before induction of anesthesia in the operating room. If the patients had fever and leukocytosis in the emergency department, antibiotics were started in the preoperative holding area. Postoperative antibiotics were not used except for patients who had a ruptured appendix or those with infectious complications.

The patients were operated on under the supervision or direct management of seven general surgeons. The patients received either open appendectomy via McBurney incision or a laparoscopic appendectomy according to the preference of each surgeon. The techniques utilized for appendectomy are described in brief. The open appendectomy was performed via McBurney incision in the lower-right quadrant. The appendix was divided at the base, and the stump was buried gently. In patients with a perforated appendix, the fluid in the peritoneal cavity was cleaned out by surgical gauze, and the wound was closed after washing with normal saline. Laparoscopic appendectomy was performed with a 12-mm trocar in the infra-umbilical position. Either the Hasson or the Veress needle technique was used to establish pneumoperitoneum according to the surgeon's preference. Two 5-mm ports were used, one in the left iliac fossa, and the other in the suprapubic position. The appendix base was tied and divided between two endoloops with laparoscopic scissors. The appendix was taken out from the 12-mm trocar without an endocatch bag. In the case of perforation, peritoneal lavage with normal saline was performed before closure. Laparotomy with a low midline incision was performed for diffuse peritonitis. The duty residents were recorded as the operator only if they performed the tasks from opening the peritoneal cavity to taking out the appendix; otherwise, the residents were recorded as assistants.

Immediately after removal, the appendices were incised open, and their luminal contents were swabbed. The swabs were placed in agar gel transport systems for separate anaerobic and anaerobic culture. The specimens were either sent to the microbiology laboratory directly or kept at 4°C until the next day if they were collected after office hours. Bacterial culture and identification were performed according to the laboratory protocol of National Yang-Ming University Hospital. Anaerobic culture and identification were performed using MGC Anaero Pack (Mitsubishi Gas Chemical Company, Tokyo, Japan) and the Remel RapID^™ ANA II system (Thermo Fisher Scientific, Lenexa, KS), respectively. The disc diffusion method was used for antibiotic susceptibility testing. All of the appendices were examined pathologically. Ruptured appendicitis was defined clinically during operation and confirmed by the pathologist.

The patients were observed for postoperative complications by chart review. Ileus was defined as repeat visits with self-limiting abdominal pain but without leukocytosis or fever. Intra-abdominal abscess was diagnosed by clinical symptoms and signs combined with imaging studies. Both sites with purulent discharge and those with local cellulitis were defined as infected. The recovery time for SSI was defined as the duration from the appendectomy to the date when all signs and symptoms had been resolved fully, as documented in the medical records. We did not call back our patients at a specific time. Instead, the patients' medical charts were reviewed in detail through June 30, 2011, when all appendectomies had been performed more than six months earlier. Any revisits because of postoperative complications were recorded.

Statistical analyses were performed with STATA for Windows 10.0 (StataCorp, College Station, TX). The Student t-test and the Fisher exact test were used as appropriate for statistical analysis. In all statistical tests, p<0.05 was considered to be significant. All p values were two-sided.

Results

Between January 1 and December 31, 2010, 148 patients underwent appendectomy at our institution, including three interval appendectomies and 145 emergency appendectomies. Of the 145 patient who underwent emergency appendectomies, there were 106 (73%) who had pathologically proved acute non-perforated appendicitis, 12 (8%) had pathologically proved acute appendicitis with rupture, and 27 (19%) had a normal appendix. Of the 118 patients with pathology-proved acute appendicitis, the mean age was 38 years (range 5–82 years) and 71 (60%) were male. Of these 118 patients, 117 (99%) had a specimen sent for bacterial culture, the exception being a 59-year-old man who underwent laparoscopic appendectomy without postoperative complications. Therefore, the 117 patients with pathology-proved acute appendicitis were included in the bacteriologic analyses.

Bacteria were isolated from 115 of the 117 specimens sent for culture (98%). Of 115 patients with positive cultures, all had aerobic bacteria, and five had a mixture of aerobic and anaerobic bacteria. The most common aerobic organism was E. coli, which was present in 100 patients (85%) (Table 1). The other organisms were Klebsiella pneumoniae (30 cases; 26%), Streptococcus spp. (29 cases; 25%), Enterococcus spp. (21 cases; 18%), Pseudomonas aeruginosa (18 cases; 15%), five cases each of Proteus mirabilis and Klebsiella oxytoca, Bacteroides spp. in four cases (3%), Proteus vulgaris in three cases (3%), and one case each of Enterobacter spp., Moraxella spp., Comamonas spp. nonfermentative gram-negative bacilli, Porphyromonas asaccharolytica, and Pseudomonas alcaligenes. There were only five anaerobic isolates, which consisted of four Bacteroides spp. and one Porphyromonas asaccharolytica.

Four E. coli isolates (4%) were resistant to cefuroxime, and two of them were ESBL-producing (Table 2). All 30 K. pneumoniae isolates were sensitive to cefazolin except one that showed intermediate sensitivity. All 29 Streptococcus isolates were sensitive to cefazolin. All 21 Enterococcus isolates except one were sensitive to ampicillin. All 18 P. aeruginosa isolates were sensitive to amikacin, ceftazidime, and cefepime but most (7/8 tested) were resistant to cefuroxime (Table 2). All five anaerobic isolates were sensitive to cefoxitin except one B. fragilis strain, which showed intermediate sensitivity.

Four patients revisited the emergency department because of acute abdominal pain, and ileus was diagnosed in all of them. In total, 8 of 118 patients (7%) suffered from post-operative SSI. Seven of them were treated by debridement of the infected areas, combined with antibiotic therapy based on the initial culture results. One patient had cellulitis around the operative site without an underlying abscess, and this was treated with oral antibiotics. The patients with SSI had a mean recovery time of 28 days (range 15–49 days).

There were 107 patients receiving prophylactic antibiotics, eight receiving postoperative antibiotics, and two receiving no antibiotics at all. The 10 patients who did not receive antibiotics preoperatively had no postoperative complications and were excluded from the SSI statistical analyses.

There were 107 patients receiving various prophylactic antibiotic regimens, including cefoxitin in 26 cases, cefmetazole in six cases, cefazolin + gentamicin + metronidazole in 73 cases, amoxicillin/clavulanate in 1 case, and ceftriaxone in one case (for one 12-year-old girl with a rupture; Table 3). The SSI rates showed no significant differences among antibiotic regimens (p=1.000). Among the 107 patients with pathology-proved acute appendicitis who had received antibiotics pre-operatively, there was no statistical difference in age, gender, type of appendectomy surgery, or operating time between the patients with SSI and those without (Table 3). There also was no statistical difference in SSI rate among the various surgeons. Operations performed by resident surgeons did not show a higher infection rate. Patients with a ruptured appendix had a significantly higher SSI rate (3/12; 25%) than those without rupture (5/95; 5%; p=0.044; Table 3).

When the antibiotic sensitivities of the bacterial isolates were compared with the preoperative antibiotics, there were 74 sensitive cases, 12 resistant cases, and 21 indeterminate cases (intermediate sensitivity or sensitivity test for specific antibiotics not available; Table 4). The SSI rates were higher for resistant and indeterminate groups (p=0.014; Table 4). When the P. aeruginosa isolates were addressed, there was no isolate sensitive to the preoperative antibiotic given. In contrast, there were six resistant isolates with two of them associated with SSI, and 11 indeterminate isolates with three of them associated with SSI (Table 4).

There was no statistical difference in the presence of E. coli, K. pneumonia, Streptococcus, Enterococcus, Proteus, anaerobic organisms, or number of bacterial isolates for SSI (Table 5). The four patients with E. coli resistant to second-generation cephalosporin or producing ESBL did not have a SSI. However, P. aeruginosa was isolated frequently among the patients with an SSI (5/8 in the SSI group vs. 12/99 in the non-SSI group; p=0.002).

Discussion

In a national registered study in Finland, the incidence of appendicitis was highest in patients ages 15–24 years [1]. In our study, the mean age was 38 years with a wide range, from 5 to 82 years. As in other rural regions in Taiwan, most of the young people in the studied region emigrated to nearby Taipei City. The higher age in our study may reflect the aging characteristics of our population more than the true epidemiology of this disease.

In our study, there were 106 patients who had pathology-proved acute nonperforated appendicitis, and 12 patients who had pathology-proved acute appendicitis with rupture. Our negative appendectomy rate (19%, 27/145) may be somewhat higher than in previous reports, where the rate ranged from 6% to 19% [21–23]. However, it is noteworthy that historically, the negative appendectomy rate for patients with acute appendicitis typically exceeded 20% [24]. Negative appendectomy and perforation rates usually correlate inversely. Accordingly, our proportion of perforated appendicitis was lower.

In 1984, Lau et al. reported a detailed bacteriologic study of 161 patients who had been operated on for appendicitis [13]. E. coli was isolated from the lumen of the appendix in 98 (60.9%) of these cases. Furthermore, there were 23.6% Klebsiella spp. and 8.7% P. aeruginosa strains isolated [13]. However, this early report covered only a period of six months and did not involve a pathologic review. In 1990, Bennion et al. recovered 223 anaerobic and 82 aerobic bacteria from appendiceal tissue in 30 patients [12]. They found E. coli strains were isolated in 83.3% of these cases and P. aeruginosa in 23.3% [12]. However, these results were obtained from patients older than 12 years who had gangrenous or perforated appendicitis. In our study, all appendicitis specimens were proved pathologically and include 12 specimens (10%) that were from ruptured appendices. Among the 117 specimens, E. coli was isolated in 100 cases (85%), K. pneumonia in 30 (26%), Streptococcus spp. in 19 (25%), Enterococcus spp. in 21 (18%), and P. aeruginosa in 18 (15%). Our results represent a full year of study of the bacteriology of acute appendicitis across all age groups in a limited rural region.

The distribution of bacteria obtained from the appendiceal lumen in early appendicitis may be different from that of peritonitis secondary to appendiceal rupture. Although the cultural yields obtained from intraperitoneal fluid (pus) or abscess contents correlated directly with intra-abdominal sepsis [13], the cultural yields of appendiceal luminal abscess might be associated with SSI because of direct contact with the incision during operation [25]. Although specific pathogen recovery via site culture was valuable, surgical site culture had two limitations in our study. First, when infection was found, pus culture results required a few more days, and most of our surgeons did not repeat site culture but instead treated empirically or according to the initial culture results. Second, although perioperative antibiotics had been used widely in appendicitis, the pus culture from the surgical site often yielded no bacterial growth despite frank infection [26].

Bacteroides fragilis was reported as the leading bacterium isolated from the lumen of the appendix, accounting for 88.8% of the isolates in the study of Lau et al. [13]. In another study, using optimized sampling, transport, and culture techniques, Bennion et al. reported that B. fragilis was isolated in 73.3% of specimens tested [12]. Our study identified a mean of 7.4 anaerobes in each specimen, much greater than the mean of 0.9% anaerobes suggested as normal in the literature [12]. In our study of 117 specimens, the anaerobic isolates consisted only of four B. fragilis and one Porphyromonas spp. Although our recovery rate for anaerobic bacteria was low, there were methodological differences. In the study of Bennion et al., bacteria were isolated from the appendiceal tissue in the area of necrosis, and 18 of the 30 specimens (60%) were perforated. In our study, the bacteria were isolated by swabs from the appendiceal lumen, and only 12 of 117 of the appendices (10%) were ruptured. In the past, marked increases in the counts of anaerobic bacteria have been noted as one moves from normal toward gangrenous appendices [27]. Our results could be in line with the report of Gladman et al. from the United Kingdom, who showed a low isolation rate of anaerobes (40/361; 11%) in acute appendicitis by swab culture [25]. Finally, considering that most of the antibiotics given were active against the anaerobes, the use of preoperative antibiotic prophylaxis and preoperative antibiotic consumption in some debilitated patients might suppress the growth of anaerobes [28]. However, probably, our collection of specimens into anaerobic tubes was delayed, and the time from sample collection to plating was too long. In our study, 79 (68%) of 117 patients were operated on during off-hours, and the mean time from the collection of the specimen to plating was 17 h. The specimens were stored at 4°C until the next working day if they were not sent to the microbiology laboratory immediately. Although the specimen was stored at low temperature, the detection rate might become lower. Bennion et al. suggested that oxygen diffuses better at colder temperatures [12], and our tubes might not have been absolutely gas-tight.

In early reports, anaerobic organisms were regarded as more important than fecal aerobic organisms in the pathogenesis of SSI after appendectomy, and therefore, it was believed that clindamycin produced a significant reduction in the overall rate of infection (33% vs. 17%) [29]. Later, in a prospective multicenter study of Bauer et al. in 1989, 1,735 patients undergoing appendectomy received either 2 g of cefoxitin as prophylaxis or no prophylaxis [7]. Cefoxitin is a cephamycin that should be active against the clinically important gram-negative bacteria and B. fragilis, but not against Pseudomonas. Overall, the patients receiving prophylaxis showed a significant reduction in the incidence of SSI compared with the control group (2.5% vs. 8.3%, respectively) [7]. In our study, the SSI rate was 3/12, respectively (25%) for ruptured appendicitis and 5/95 (5.3%) for nonperforated appendicitis. However, the study reported by Bauer et al. excluded patients with a perforated appendix or appendiceal abscesses but included patients with a normal appendix. Thus, their SSI rate under prophylactic antibiotics can be stratified into 3.1% (18/579) for non-perforated appendicitis. Thus, the site infection rate (5.3% for nonperforated appendicitis) in our study was slightly higher than that in the prospective trial by Bauer et al.

The SSI was found 4 d to 18 d (mean 9 d) after appendectomy. Although nosocomial infection was defined as an infection absent at admission that became evident 48 h or more after admission [30], our bacterial swabs were collected on admission. In addition, the antibiotic sensitivity tests in our study demonstrated P. aeruginosa 100% sensitive to both ceftazidime and ciprofloxacin, which is compatible with community-acquired rather than nosocomial P. aeruginosa [30].

Pseudomonas aeruginosa is a major pathogen during appendicitis and peritonitis and has a frequency of 10.3% among children [31]. With gangrenous and perforated appendicitis, P. aeruginosa has been reported to be the third leading aerobic bacterium (7/30; 23%) after E. coli and Streptococcus [12]. Yellin et al. reported a randomized, double-blinded comparison of ampicillin-sulbactam and clindamycin with gentamicin in 1985 [32]. They included 107 patients with gangrenous or perforated appendicitis but excluded those with early appendicitis. Via the culture of peritoneal fluid adjacent to the appendix, Pseudomonas was recovered in 19 patients (18%), but 17 of them had perforated appendicitis. In our study, P. aeruginosa was isolated by appendiceal luminal swabs in 18 cases (15%), and only three of them had perforated appendicitis. These discrepant findings may be ascribable to the ways cultures were taken (peritoneal fluid vs. luminal swabs) and the clinical stages (perforated vs. early appendicitis). To agree with the report of Yellin et al., we suggest that P. aeruginosa deserves further attention.

Although the recommended second-generation cephalosporins have been proved to reduce the overall infection rate with appendicitis, these drugs are not effective against Pseudomonas [7]. Pseudomonas aeruginosa has a resistance rate of 8% to 48% to gentamicin and 92% to the third-generation cephalosporin ceftizoxime [33–35]. Although no significant difference was found when comparing cefoxitin with other third-generation cephalosporins in prophylactic use, such studies used only third-generation cephalosporins, which have low-to-moderate effectiveness against P. aeruginosa [14,15]. At the same time, two studies showing that post-operative antibiotics were ineffective at reducing the rates of SSI did not use antibiotics that target Pseudomonas [36,37]. In our study, when the P. aeruginosa isolates from the patients receiving preoperative antibiotics were addressed, there were no isolates sensitive to the preoperative antibiotics, six resistant isolates with two associated with infections, and 11 indeterminate isolates with three associated with infections. The 11 P. aeruginosa isolates with indeterminate sensitivities came from patients given preoperative gentamicin, because our microbiology laboratory did not test gentamicin sensitivity routinely for P. aeruginosa. Accordingly, the result with such a high infection rate in the indeterminate group might be attributable to gentamicin failure against P. aeruginosa. This would imply an adverse clinical result under the current prophylactic guidelines for acute appendicitis. The clinical significance of P. aeruginosa infection needs to be evaluated further and the choice of prophylactic antibiotics revisited in the light of the findings.

Several limitations of this study need to be acknowledged. First, the cases were recruited from seven surgeons with different degrees of experience. In addition, 19/107 appendectomies (18%) were performed by resident surgeons. However, there was no statistical difference in the complication rates among attending and resident surgeons. Second, whereas some surgeons preferred laparoscopic appendectomy, others preferred a traditional open appendectomy. Laparoscopic surgery has been reported to have a lower SSI rate, from 0.6% to 2% [38–40] in uncomplicated appendicitis to 18% in complicated appendicitis [4]. Our SSI rate (3/43; 7%) for laparoscopic appendectomy was higher than reported elsewhere. We did not routinely use an extraction bag during laparoscopic appendectomy, even though such use has been reported to be associated with a lower SSI rate [40]. In addition, our national health insurance did not pay for laparoscopic appendectomy until 2007, which meant surgeons might lack experience and could have resulted in a higher infection rate during the study year. Third, prophylactic antibiotics were not given to 10/117 patients (9%) because of a lack of adherence to guidelines. In addition, the selection of prophylactic antibiotics was not randomized among various guideline regimens, and selection bias cannot be excluded.

In conclusion, the organism involved most commonly in acute appendicitis was E. coli, which was followed by K. pneumoniae, Streptococcus spp., Enterococcus spp., and P. aeruginosa. Isolation of P. aeruginosa was associated with SSI despite antibiotic prophylaxis. When an updated recommendation is made, coverage of P. aeruginosa is strongly recommended, and a further randomized trial is warranted.