Capdevila O, Pallares R, Grau I, Tubau F, Liñares J, Ariza J, Gudiol F. Pneumococcal Peritonitis in Adult PatientsReport of 64 Cases With Special Reference to Emergence of Antibiotic Resistance. Arch Intern Med. 2001;161(14):1742-1748. doi:10.1001/archinte.161.14.1742
Few data are available regarding pneumococcal peritonitis. We studied the clinical characteristics of intra-abdominal infections caused by Streptococcus pneumoniae and its prognosis in relation to antibiotic resistance.
We reviewed all cases of culture-proved pneumococcal peritonitis. Patients with liver cirrhosis and primary pneumococcal peritonitis were compared with patients with Escherichia coli peritonitis.
Between January 1, 1979, and December 31, 1998, we identified 45 cases of primary pneumococcal peritonitis in patients with cirrhosis and 19 cases of secondary (or tertiary) pneumococcal peritonitis. Patients with cirrhosis and primary pneumococcal peritonitis vs those with primary E coli peritonitis had more frequent community-acquired infection, 73% vs 47%; pneumonia, 36% vs 2%; and bacteremia, 76% vs 33%; and higher attributable mortality (early mortality), 27% vs 9% (P<.05 for all). Secondary (or tertiary) pneumococcal peritonitis was associated with upper or lower gastrointestinal tract diseases; in most cases, the infection appeared after surgery. A hematogenous spread of S pneumoniae from a respiratory tract infection might be the most important origin of peritonitis; also, S pneumoniae might directly reach the gastrointestinal tract favored by endoscopic procedures or hypochlorhydria. There was an increased prevalence of penicillin and cephalosporin resistance up to 30.7% and 17.0%, respectively, although it was not associated with increased mortality rates.
Primary pneumococcal peritonitis in patients with cirrhosis more often spread hematogenously from the respiratory tract and was associated with early mortality. In secondary (and tertiary) pneumococcal peritonitis, a transient gastrointestinal tract colonization and inoculation during surgery might be the most important mechanisms. Current levels of resistance were not associated with increased mortality rates.
STREPTOCOCCUS PNEUMONIAE is a common pathogen that causes high morbidity and mortality around the world.1,2 It is the most common cause of community-acquired pneumonia and the second most common cause of purulent meningitis,3- 5 but intra-abdominal pneumococcal infections are rarely found.6- 9
Primary (or spontaneous) bacterial peritonitis occurs mainly in patients with liver cirrhosis and is usually caused by gram-negative bacilli.10- 12 It is thought that in most cases the enteric microorganism gains access to the peritoneal cavity without loss of integrity of the intestinal wall through a mechanism of bacterial translocation.12- 20 In addition, gram-negative bacteria can occasionally reach the peritoneal cavity by a hematogenous route from a distant primary focus (eg, a urinary tract infection).21
Primary pneumococcal peritonitis in patients with cirrhosis is usually associated with a respiratory tract infection such as pneumonia, and, in this case, the bloodstream (hematogenous spreading) is the most likely route of peritoneal fluid infection.6,7,22,23 In some patients, a respiratory tract focus is not clinically apparent, and, in these cases, the gastrointestinal tract has been hypothesized to be a source of pneumococci.7,22,24 However, pneumococci are soluble in bile salts,5 and, therefore, it is unlikely that S pneumoniae can grow in the normal gastrointestinal tract.
On the other hand, secondary intra-abdominal infection (secondary peritonitis) is due to the spread of gastrointestinal or genitourinary microorganisms into the peritoneal space from loss of integrity of the mucosal barrier. These are often polymicrobial infections and can take the form of generalized peritonitis or localized peritonitis (localized abscesses).14,15,22
To date, only anecdotal cases of adult patients with secondary pneumococcal peritonitis have been reported,25- 28 and some of them were associated with appendicitis.29,30 The mechanism of the pneumococcus that causes these infections is not clear because it is not found in the gastrointestinal tract.
There have been reported cases of primary and secondary pneumococcal peritonitis in prepubertal girls and in postpubertal healthy women, mainly post partum, after an abortion, after gynecologic procedures, or associated with intrauterine device use.25- 28,31 It is well known that S pneumoniae can colonize in the vagina and an ascending infection can occur.32,33 In children, colonization in the genitourinary tract with S pneumoniae occurs because of inadequate hygiene or orogenital sexual abuse.24,29,30 In these patients, when no apparent purulent foci in the genitourinary tract is found, such cases are usually called primary peritonitis.32
The main objectives of the present study were (1) to describe different types of intra-abdominal infections caused by S pneumoniae; (2) to determine the clinical characteristics, laboratory findings, and outcomes of patients with liver cirrhosis and primary peritonitis due to S pneumoniae and to compare them with those infected by Escherichia coli; and (3) to study antibiotic resistance in pneumococcal isolates and determine its clinical relevance.
The present study was carried out in Bellvitge Hospital, a 1000-bed teaching hospital for adult patients in Barcelona, Spain. It serves approximately 1 million people and admits more than 26 000 patients per year. This hospital does not have pediatrics and obstetrics departments.
We reviewed the clinical and microbiological data of all patients who had a diagnosis of diffused or localized peritonitis and a peritoneal sample that was positive for S pneumoniae. These patients were admitted to Bellvitge Hospital between January 1, 1979, and December 31, 1998, and most of them had been cared for and included in a protocol of invasive pneumococcal infections by one of us (R.P.).
Patients with liver cirrhosis and primary pneumococcal peritonitis (case patients) were compared with patients with cirrhosis and primary peritonitis due to E coli (control patients). For each case patient we selected a control patient according to the nearest date of positive culture. If a case or control patient had more than 1 episode of primary peritonitis, only the first episode was considered.
The diagnosis of liver cirrhosis was established using clinical, laboratory, and exploratory findings and did not require a liver biopsy with histologic confirmation. The stage of cirrhosis was determined by the criteria of Pugh et al.34,35
Primary peritonitis in a patient with cirrhosis was considered when clinical findings together with biochemical data of peritoneal inflammation and a positive ascitic fluid culture (for S pneumoniae [case patients] or E coli [control patients]) were present, and without any clinical or radiologic data suggesting a surgically treatable intra-abdominal focus.
Secondary (or tertiary) pneumococcal peritonitis was diagnosed in a patient with a localized or diffuse suppurative intra-abdominal process together with a positive culture for S pneumoniae and in whom an intra-abdominal, surgically treatable source was detected (or in whom the infection appeared after surgery).
We considered a hospital-acquired peritoneal infection (either primary or secondary peritonitis) when the episode occurred 48 hours after hospital admission (or appeared after surgery) and it was not in the incubated period. Community-acquired infection was considered when it was evident on hospital admission or within the first 48 hours.
Pneumonia was considered (definitive diagnosis) in a patient with signs or symptoms of a lower respiratory tract infection and a new pulmonary infiltrate on chest radiography, together with bacteremia or positive culture from a lower respiratory tract sample (eg, pleural fluid). We also considered pneumonia (presumptive diagnosis) in patients with clinical and radiographic findings compatible with pneumonia, with a positive sputum sample or no respiratory tract samples available for culture (all of our patients had concomitant positive ascitic fluid cultures for either S pneumoniae or E coli).
In all patients, demographic characteristics, clinical findings, and laboratory and medication data were obtained from hospital records or from previous data recorded in the protocol of invasive pneumococcal infections.
Previous hospitalization was defined as admission to any hospital during the previous 6 months.
We specifically investigated the performance of endoscopic procedures, peritoneovenous shunt implementation, or surgery within 1 month of hospital admission.
Previous antibiotics were considered when the patient received any antibiotic for prophylaxis or treatment for more than 48 hours during the previous 30 days of peritonitis.
Previous episodes of primary bacterial peritonitis were considered when they occurred within 1 year of the current hospital admission and by a different microorganism.
Septic shock was considered in a patient with a systolic blood pressure below 90 mm Hg and peripheral hypoperfusion together with clinical or bacteriological evidence of uncontrolled infection.
Antibiotic therapy was prescribed according to the attending physician's criteria and varied during the study period. The most common empirical antibiotic regimens were penicillin or ampicillin sodium plus aminoglycosides or aztreonam during the first years of the 1980s, and since 1986 almost all patients received a cephalosporin such as cefotaxime sodium or ceftriaxone sodium.
Mortality was considered when the patient died within 30 days of diagnosis of peritonitis. Attributable mortality (mortality probably related to infection) was considered when the patient died within 7 days of diagnosis and without another evident cause of death.
Biological samples (eg, blood and peritoneal fluid) were studied at the Microbiology Laboratory. Strains of S pneumoniae were identified using standard methods. Ascitic fluid cultures were performed by the method of bedside inoculation of blood culture bottles with ascites. Antibiotic susceptibility to penicillin was initially determined with a 1-µg oxacillin sodium disk using the Kirby-Bauer disk diffusion method, and strains that showed a zone of inhibition of less than 20 mm were considered nonsusceptible to penicillin. Minimal inhibitory concentrations (the lowest concentration that inhibits pneumococcal growth) of isolated strains from 1979 to 1992 were determined using the agar dilution method in Mueller-Hinton agar supplemented with 5% sheep blood and containing antimicrobials, as described previously.36 Minimal inhibitory concentrations of isolated strains from 1993 to 1998 were determined using the microdilution method in Mueller-Hinton broth supplemented with 5% lysed horse blood, as recommended in the 1992 criteria of the National Committee for Clinical Laboratory Standards.37 The following 7 antimicrobial agents were tested: penicillin G, ceftriaxone/cefotaxime, erythromycin, clindamycin, sulfamethoxazole and trimethoprim, tetracycline, and chloramphenicol. The results of susceptibility tests were evaluated according to the 1998 criteria of the National Committee for Clinical Laboratory Standards.
Statistical analysis was carried out using a statistical software package (SPSS for Windows, version 9.0; SPSS Inc, Chicago, Ill). Data are given as mean ± SD. Continuous variables were compared using the unpaired t test, and categorical variables were compared using χ2 or Fisher exact tests when appropriate. Statistical significance was established at P<.05 (2-tailed). When many variables were analyzed, and to adjust for multiple comparisons, the Bonferroni adjustment was used (calculated by dividing α = .05 by κ variables).38
At Bellvitge Hospital, during the 20-year study period, we identified 64 patients with intra-abdominal infections caused by S pneumoniae. During this period, the 64 intra-abdominal pneumococcal isolates represented 4.3% of a total of 1476 S pneumoniae strains isolated from clinical specimens (all sterile fluid samples; sputum samples were not included). Comparing 1979 to 1988 with 1989 to 1998, the percentage of pneumococcal isolates from abdominal samples vs the total number of sterile fluid samples was 3.6% (22/604) and 4.8% (42/872), respectively (P = .27).
Of 64 patients with intra-abdominal pneumococcal infections, 45 with liver cirrhosis had primary pneumococcal peritonitis and the remaining 19 had secondary (or tertiary) pneumococcal peritonitis.
We analyzed all patients admitted to Bellvitge Hospital between January 1, 1989, and December 31, 1998, with a diagnosis of chronic liver disease (n = 7535); approximately 90% of them had liver cirrhosis (an estimated 6871 patients). Primary bacterial peritonitis (of any etiology) occurred in 513 patients (8% of all patients with liver cirrhosis). Primary pneumococcal peritonitis (30 cases during 1989-1998) represented 5.8% of all primary bacterial peritonitis and 0.4% of all hospital admissions for liver cirrhosis.
Forty-five case patients with primary pneumococcal peritonitis were compared with 45 control patients with primary peritonitis due to E coli (Table 1).
The statistically significant differences between the 2 groups were that patients with pneumococcal peritonitis were more frequently smokers, had associated pneumonia, more often had bacteremia, less frequently had a nosocomial-acquired infection, and had lower ascitic fluid white blood cell counts (Table 1).
Although not statistically significant, patients with pneumococcal peritonitis had more associated comorbidities, were more frequently alcohol abusers, received previous antacid treatment, and had a previous endoscopy. The etiology and stage of liver cirrhosis were similar in both groups; most patients had an advanced stage of the disease according to Child-Pugh scores (Table 1).
Mortality in patients with primary pneumococcal peritonitis and E coli peritonitis at 30 days was 47% (21 patients) and 36% (16 patients), respectively (P = .28) (Table 1). Patients with primary pneumococcal peritonitis had higher attributable mortality than patients with E coli peritonitis: 27% (12 of 45 cases) vs 9% (4 of 45 controls) (P = .05).
During the study, we found a significant decrease in the mortality rate in both groups. Thus, from 1979-1988 to 1989-1998, mortality decreased in patients with pneumococcal peritonitis from 73% (11 of 15 patients) to 33% (10 of 30 patients) (P = .01) and in patients with E coli primary peritonitis from 64% (9 of 14 patients) to 32% (10 of 31 patients) (P = .04).
We studied 19 patients (age, 59.7 ± 18.6 years; 14 [74%] were men) with diffuse or focal pneumococcal peritonitis in whom an apparent source of the infection was detected; they were divided into upper and lower abdominal infections.
The 12 patients with upper abdominal pneumococcal infections had associated gastroduodenal or pancreaticobiliary tract disease.
Of 7 patients with gastroduodenal ulcer (perforated or bleeding), in 1 case S pneumoniae was isolated in pure culture at the time of initial surgery and in the remaining 6 cases the microorganism was iso lated 4 to 31 days after surgery (4 had polymicrobial infection).
Of the other 5 patients, 3 were admitted to the hospital because of an exacerbation of chronic pancreatitis, and during hospitalization a pancreatic pseudocyst or abscess was diagnosed; S pneumoniae was isolated from a sample obtained from scanning guided puncture 12 to 46 days after hospitalization (none of these patients underwent previous surgery). The other 2 patients were admitted to the hospital because of biliary tract neoplasm (n = 1) or pancreatic neoplasm (n = 1). The patient with a diagnosis of pancreatic neoplasm underwent surgery on hospital admission, and S pneumoniae grew in pure culture from an exudate obtained from a perineoplasm abscess. The patient with biliary tract neoplasm was admitted to the hospital for elective surgery; 11 days later, an intra-abdominal abscess was detected in which S pneumoniae grew in pure culture.
Of 7 patients with lower pneumococcal abdominal infections, 6 were admitted to the hospital for elective surgery (5 with intestinal neoplasm and 1 with Gardner syndrome). The intra-abdominal infection appeared 5 to 20 days after surgery (in 4 of the 6 cases, S pneumoniae was isolated in pure culture). The last patient is a postpartum woman who was admitted to the hospital because of abdominal pain and fever and was diagnosed as having an intra-abdominal abscess in which S pneumoniae grew in pure culture.
Most patients with secondary pneumococcal peritonitis had some predisposing factors for pneumococcal infection or colonization. Of 19 patients, alcohol abuse was present in 14 (74%), liver cirrhosis in 2 (11%), chronic bronchitis or chronic obstructive pulmonary disease in 7 (37%), and clinical and radiologic findings suggestive of a lower respiratory tract infection coinciding with or preceding the intra-abdominal pneumococcal infection in 5 (26%). Only 4 patients (21%) underwent previous endoscopy, and 8 (42%) had previous antacid treatment.
All but one patient required drainage (surgically or by scanning guide puncture); this was a patient with cirrhosis who died (see the following paragraph). All but 3 patients received antibiotic therapy; these 3 patients recovered with only surgical drainage.
Two (11%) of 19 patients died, both within 3 days of diagnoses of pneumococcal infection. One of them, a 73-year-old man with liver cirrhosis in an advanced stage, had a polymicrobial infection after surgery for a perforated gastric ulcer; a culture from an abdominal fluid puncture grew S pneumoniae and Pseudomonas aeruginosa (resistant to initial therapy with ceftriaxone and clindamycin); no surgical drainage could be performed. The second patient was a 78-year-old man with chronic obstructive pulmonary disease and gastrointestinal tract bleeding secondary to duodenal ulcer who developed a polymicrobial infection after surgery in which grew S pneumoniae and Morganella morganii; he died despite surgical drainage and antibiotic therapy (amoxicillin and clavulanate potassium plus tobramycin sulfate).
Overall, in the abdominal pneumococcal isolates (n = 64) we found an increase in penicillin resistance during the study: 14% (2 of 14 strains) between 1979 and 1985, 33% (5 of 15 strains) between 1986 and 1991, and 40% (14 of 35 strains) between 1992 and 1998 (P = .05). The mortality rate did not differ significantly between patients infected with penicillin-resistant and penicillin-susceptible strains (Table 2). Only 4 patients had resistance to cefotaxime/ceftriaxone, and one of them died.
Table 3 shows our experience with antibiotic resistance in 1095 pneumococcal blood isolates during 2 different periods. There was increased prevalence of penicillin, cephalosporin, and macrolide resistance and decreased tetracycline and chloramphenicol resistance.
We suggest a comprehensive classification of pneumococcal peritonitis (Table 4). Most patients with pneumococcal peritonitis can be included in 1 of 3 categories: (1) primary peritonitis associated with liver cirrhosis, nephrotic syndrome, or chronic renal failure and continuous ambulatory peritoneal dialysis; (2) secondary (or tertiary) peritonitis associated with gastrointestinal disease (or after surgery); and (3) peritonitis in young women with or without an apparent genitourinary focus.
To date, most cases of pneumococcal peritonitis have occurred in patients with liver cirrhosis.7,9 According to different studies,10,11,39 of all cases of primary peritonitis in patients with cirrhosis, 1% to 11% are caused by S pneumoniae. However, pneumococcus rarely causes secondary (or tertiary) peritonitis, and there have been only anecdotal reports of well-documented secondary (or tertiary) pneumococcal peritonitis associated with gastrointestinal disease8,9,29,30 or in young women.5,22- 28,33
Herein, we found that in patients with cirrhosis, the mechanism of spread and prognosis are the 2 most important differences between primary peritonitis due to S pneumoniae and primary peritonitis due to E coli.
In patients with cirrhosis and primary peritonitis due to E coli, the most important mechanism is thought to be translocation from the gastrointestinal tract, and, in some cases, it can be due to a bacteremic spread from a distant focus (eg, urinary tract infections). On the other hand, the mechanism of primary pneumococcal peritonitis in patients with cirrhosis is controversial because S pneumoniae is not found in the gastrointestinal tract, probably because this microorganism is soluble in bile salts that avoid bacterial growth.5 However, we can speculate that hematogenous spreading might be important. Thus, in our study, 16 patients (36%) had associated pneumonia, and, in these cases, hematogenous spread from the respiratory tract could be suggested.7 In addition, other patients who develop pneumococcal peritonitis are those without an apparent respiratory tract infection but who are at higher risk of S pneumoniae oropharyngeal colonization (eg, smokers and alcoholics) and who were recently subjected to previous endoscopic procedures. Transient bacteremia can occur after endoscopy.40
Recent studies41- 46 have shown that short-term prognosis in patients with cirrhosis and primary bacterial peritonitis (most cases caused by Enterobacteriaceae) has improved during the past decade, probably as a consequence of early diagnosis and better therapeutic approaches. Thus, mortality from primary bacterial peritonitis (all cases) decreased from more than 75% in the 1970s to about 40% in the 1990s.13,14,17,21,22,47 We also found a decreased mortality rate during recent years in patients with either pneumococcal or E coli primary peritonitis.
In our study, we found a higher 30-day mortality rate in patients with pneumococcal peritonitis compared with those with E coli peritonitis, but this difference did not reach statistical significance (Table 1). Moreover, patients with pneumococcal peritonitis had higher attributable mortality (27%) than those with peritonitis due to E coli (9%). We can speculate that S pneumoniae might be more virulent than E coli and that pneumococcal peritonitis is often associated with pneumonia, which can be complicated with respiratory failure and septic shock. We did not find that patients with pneumococcal peritonitis had a more advanced stage of cirrhosis.
Of note is the high number of nosocomial-acquired pneumococcal and E coli peritonitis in our study. Despite the fact that these microorganisms are usually components of the patient's own flora (acquired in the community), in recent years patients with cirrohsis have a long life expectancy and might develop peritonitis during hospitalization for other reasons. However, we cannot exclude that some of these microorganisms were transmitted from inpatients or hospital personnel.
Table 3 shows trends in antibiotic resistance during the past 2 decades. We observed significant increases in cases with intermediate or resistant penicillin, cephalosporin, and macrolide. We did not find a correlation between the minimal inhibitory concentrations of penicillin and cephalosporin and mortality rates (Table 2). This is also found with other nonmeningeal pneumococcal infections48 and might be because levels of β-lactams achieved in serum and ascitic fluid are much higher than minimal inhibitory concentrations.
In cases of secondary peritonitis, the infection might occur as a result of loss of integrity of the gastrointestinal tract wall (eg, perforation of gastric ulcer or colonic neoplasm).14 Secondary peritonitis caused by S pneumoniae has rarely been reported.9 We identified 19 cases of secondary (or tertiary) pneumococcal peritonitis associated with upper or lower gastrointestinal tract diseases.
Although S pneumoniae is not a component of gastrointestinal tract flora, we can speculate that in some circumstances the microorganisms might become part of the transient gastric flora. Despite not being well documented, upper abdominal transient colonization by S pneumoniae might occur in patients with low gastric acidity who had oropharyngeal colonization by S pneumoniae, and particularly in patients subjected to endoscopic procedures.
In lower abdominal infections, the pneumococcal colonization is more difficult to explain, and we found in the literature only a few cases of acute appendicitis caused by S pneumoniae.9,29,30 In our study, we identified 6 patients with colonic diseases and pneumococcal peritonitis, and in all cases the infection appeared after surgery. In cases of pneumococcal peritonitis that occur after surgery (tertiary peritonitis), one can speculate that the microorganism might also be inoculated during the surgical procedure from the oropharyngeal flora of the surgical personnel.
In conclusion, the main characteristics of primary pneumococcal peritonitis in patients with cirrhosis (compared with E coli peritonitis) were the hematogenous spread from the respiratory tract and the higher attributable mortality rate. Perhaps, in some cases, S pneumoniae might directly reach the gastrointestinal tract from the oropharynx favored by endoscopic procedures or achlorhydria. In secondary (and tertiary) pneumococcal peritonitis, the origin of the pneumococcus is unclear, but transient gastrointestinal tract colonization and inoculation during surgical procedures might be the most important causes. Mortality was not related to penicillin and cephalosporin resistance.
Accepted for publication December 4, 2000.
This study was supported in part by grant FIS 97/0716 from the National Health Service, Madrid, Spain, and from the Agencia d'Avaluacio de Tecnologia Medica, Generalitat de Catalunya, Barcelona, Spain.
Corresponding author: Roman Pallares, MD, Fundacio August Pi i Sunyer, Hospital de Bellvitge, Feixa Llarga s/n, 08907 L' Hospitalet, Barcelona, Spain (e-mail: firstname.lastname@example.org).