Clinical and Bacteriological features of spontaneous bacterial peritonitis in Malagasy patients with Cirrhosis

Background: Spontaneous bacterial peritonitis (SBP) represent frequent and serious complications in cirrhosis patients with ascites. Our aim was to describe the clinical and bacteriological characteristics of SBP in Madagascar. Methods: This is a 21-month prospective study between January 2018 and October 2019, including hospitalized patients with cirrhosis, with clinical and biological symptoms of SBP. Results: Thirty-three patients were included. The mean age was 48.09 ± 13.55 years (extremes: 19 – 78 years), the sex ratio was 3.12. Abdominal pain (55%), fever (36%), diarrhea (6%), hepatic encephalopathy (18%) are the most common symptoms. Gastrointestinal bleeding (18.18%) was the main risk factor to SBP. SBP was community-acquired in 87.88% of cases. A culture of ascites uid was positive for 9 patients (27.27%). The infectious agents found were Escherichia coli (12.10%), Klebsiella pneumoniae (3%), Pseudomonas (3%), Streptococcus mitis (9.1%). Escherichia coli were wild with one case resistant to Ceftriaxone. The Klebsiella were multidrug resistant. The other two pathogens did not show resistance. After antibiotic therapy adapted to the antibiogram, healing was observed in 26 patients (78.78%). Seven patients (21.22%) died from various complications. All deceased patients had bacteria identied in ascites uid. Conclusion: SBP dened according to clinical and biological criteria is apparently sterile in the majority of cases. Gram-negative bacteria were the major pathogens involved in SBP in cirrhotic patients. Escherichia coli and streptococcus were the most common pathogen isolated. Bacteriological study is essential to adapt antimicrobial to multidrug-resistant bacteria.


Background
Spontaneous bacteria peritonitis (SBP) is a common and serious infectious complication of advanced liver disease with portal hypertension. In-hospital prevalence varies from 10 to 30% [1,2]. SBP is a cause of high morbidity and mortality in patients with cirrhosis. Mortality during SBP has decreased markedly in recent years, but remains high with an in-hospital mortality rate ranging from 10 to 50% [2][3][4][5][6]. Early treatment of SBP could decrease the death rate of cirrhotics by 30% [2,5,6]. Knowledge of bacterial ecology is essential for a well-adapted antibiotic therapy. No previous studies have been done on the pathogens commonly responsible for SBP in Madagascar. The aim of this study was to describe the clinical and bacteriological characteristics of SBP in Madagascar.

Materials And Methods
This is a prospective and descriptive study over 21 months (January 2018 to October 2019), carried out at University Hospital of Joseph Raseta Befelatanana, Antananarivo. All patients with hospitalized decompensated cirrhosis with clinical symptoms in favor of SBP were recruited. the study included, all patients with clinical symptoms of SBP and a polymorphonuclear neutrophils leukocytes (PMN) count in ascitic uid greater than 250 cells/mm3 or bacterascites associated with clinical symptoms of SBP.
Bacterascites is de ned by a PMN count of less than 250 cells/mm with a positive culture. All patients with secondary bacterial peritonitis were excluded. Demographic data (gender and age), clinical data (antecedents, causes of cirrhosis, clinical symptoms, factors risk SBP, complications, communityacquired or nosocomial SBP), biological, cytological and bacteriological data of ascites uid (serum creatinine, Child-Pugh classi cation, ascites uid protein level, bacterascites, ascites uid culture and antibiogram) and the therapeutic data were collected. Community-acquired SBP was de ned as an infection diagnosed within the rst 48h of admission to hospital without any prior contact with health care within 90 days, where as a diagnosis made more than 48h after hospitalization was de ned as nosocomial SBP.
Diagnostic paracentesis was been done immediately of any suspicion of SBP, under aseptic conditions.
The ascites uid was received on 2 dry tubes for a biochemical, cytological and bacteriological examination. The bacteriological examination was carried out on solid media (chocolate agar and blood incubated under CO2 at 37 ° C) and liquids media (brain heart broth). Concomitant culture was done on a blood culture bottle. The samples are sent to the laboratory in less than 2 h. In the event of a positive culture, the identi cation of the pathogen and the antibiogram was carried out. In case of negative culture within 48-72h, the vials were preserved until the 5th day.
The data were analyzed on the software Epi-Info version 7.2.

Results
During the study period (January 2018 to October 2019), we collected 129 patients with cirrhosis of which 46 patients had presented one or more clinical symptoms of SBP. We excluded 13 patients, of which 4 of the excluded patients had a secondary bacterial peritonitis and the 9 others presented symptoms in favor of SBP but the PMN leukocytes count < 250 cells/mm3 without bacterascites. The owchart for patient enrollment was shown in Fig.1. Thirty-three patients were selected, giving an inhospital prevalence of 25.58%.

Bacteriological characteristics and antimicrobial susceptibility
Culture of ascites uid was sterile in 72.73% (24 patients). Nine patients (27.27%) had a positive culture. The proportion of Gram-negative bacteria and Gram-positive bacteria were 66.67% (6 patients) and 33.33% (3 patients) respectively. Escherichia coli was the major pathogen agent (4 patients, 44.44%), followed by Streptococcus mitis (3 patients, 33.33%), Klebsiella pneumoniae (1 patient, 11.11%) and Pseudomonas (1 patient, 11.11%). Escherichia coli and Streptococcus were sensitive to rst-line antimicrobial for SBP treatment (Quinolones, Cephalosporins and Penicillins). A case of Escherichia coli resistant to cephalosporins (25%) was observed. Pseudomonas was sensitive to quinolones and penicillins but resistant to cephalosporins. Klebsiella was multidrug resistant. Bacteriological characteristics and antimicrobial susceptibility are shown respectively in Table 2 Table 4.

Discussion
This was the rst study on the bacteriological pro le of SBP in Madagascar. This study was essential because it had made it possible to establish the bacteriological and clinical pro le of SBP in Malagasy patients with cirrhosis. This study had limitations. This is a single-center study and limited to those of inpatients. In-hospital prevalence of SBP was 25.58%, which corresponded to data from Western literatures (10-30%) [1,2,7]. Duah et al (Ghana, 2019) found a prevalence of 25.24%, similar to our study [8]. While, Oladimeji et al (Nigeria, 2013) reported a very high in-hospital prevalence at 67.7% [9]. These results are explained by the low number of the cirrhotic population during the study period, a high rate of culture positive (66.7%). Abdominal pain, fever and diarrhea were the most common symptoms during SBP. This nding has been reported by many authors (Setoyama et al; Shi et al) [10,11]. The presence of these symptoms should alarm the clinician, suggesting the introduction of a probabilistic antibiotic therapy, without waiting for the results of the samples. Despite the low prevalence of hepatic encephalopathy in our series (6.06%), many studies reported a high rate of it during SBP [10,11]. The presence of hepatic encephalopathy should systematically seek an SBP. Bacterial translocation is increased in gastrointestinal bleeding in patients with cirrhosis. Gastrointestinal bleeding was a risk factor to SBP [12][13][14][15]. Primary prevention of SBP with prophylactic antimicrobials should be initiated in gastrointestinal bleeding in patients with cirrhosis [2,[16][17][18]. In our study, gastrointestinal bleeding was present in only 6 patients (18.18%) with SBP.
In recent years, many authors have reported an upsurge in nosocomial SBP [11,[19][20][21]. Shi et al (2017) found a nosocomial SBP rate of 44.9% [11]. Ning et al (2018) [20,21]. Contrary to published data, the nosocomial SBP rate was low (12.12%) in this present study. These high rates of nosocomial SBP could be explained in part by the fact that the patients included in these studies had frequent previous hospitalizations and had long hospital stays. In this study, community-acquired SBP occupied the majority of cases (87.88%).
In patients hospitalized with decompensated cirrhosis, 6.06% had bacterascites. This was comparable to the data in the literature (11%) [2,16] [27]. Since SBP's mechanism in cirrhotics is bacterial translocation, this explains these ndings. The recommendations on treatment of SBP targeting Gram-negative bacteria as rst-line may therefore be applicable. Two cases of Escherichia coli resistant to antibiotics were identi ed, including one case of resistance to cephalosporins and one case to sulfamethoxazole/trimethoprim. Kirplani et al (Pakistan) [28] found cases of Escherichia coli resistant to Ceftriaxone (32%) and Sulfamethoxazole/trimethoprim (10.9%). Another study by Ding et al showed similar results [14]. The frequent and inadequate use of cephalosporins and sulfamethoxazole/trimethoprim in Madagascar may explain these facts [29]. The use of quinolones and amoxicillin-clavulanic acid are suitable alternatives in this case. Pseudomonas was resistant to cephalosporin and Sulfamethoxazole/Trimethoprim. It was most often a nosocomial germ and also had a natural resistance to certain antibiotics including cephalosporin, nalidixic acid, Fosfomycin, sulfonamides [30]. A case of nosocomial infection with Klebsiella has been identi ed. He was multidrug resistant and was only susceptible to Amikacin. Several authors (Nousbaum et al and Ariza et al) reported that 25% to 50% of germs during nosocomial SBP were resistant to the usual antibiotics [2,16,31]. This reiterates the need for a systematic performance of a bacteriological examination with antibiogram in the face of any suspicion of an SBP.
Indeed, clinicians must be vigilant in the face of the emergence of multidrug-resistant pathogens. Current recommendations called for the use of Cefotaxime (2g/8h) for 5 days, either amoxicillin/clavulanic acid (1g/8h) for 7 days, or O oxacin (400mg/day) for 7 days, in case of community-acquired SBP. Piperacillin/tazobactam or Carbapenem are recommended in case of nosocomial SBP. The initial antimicrobials will be adapted according to the antibiogram [16, 32,33]. In this study, quinolones were the most used, probabilistically, (57.58%) with a mean duration of 7 ± 2.98 days (Extreme: 2 to 14 days). Given the high prevalence of Gram-negative bacteria found in this study, the initial antimicrobial with quinolones was well suited.
The effectiveness of antibiotic therapy is de ned according to the International Ascites Club as a decrease of at least 25% in the level of PMN in ascites after 48 h of treatment [2,16]. This effectiveness will dictate the duration of treatment, explaining the variation in the duration of antibiotic therapy in this study.
In-hospital mortality rate was 15 [21]. The difference in the ndings can be explained, on the one hand, by the difference between the study populations; on the other hand, the multitude of management structures. However, SBP remains an absolute medical emergency, with a poor prognosis. The speed of the diagnosis of SBP, the susceptibility of pathogens to probabilistic antimicrobials as well as the precocity of the treatment can be key elements for the success of the management of SBP.

Conclusion
This was the rst study on the bacteriological pro le of SBP in Madagascar. In-hospital prevalence of SBP was 25.58%. The clinical symptoms are similar to those described in the literature. Abdominal pain and fever were the main warning signs for an SBP. SBP was predominantly community-based. SBP de ned according to clinical and biological criteria is apparently sterile in the majority of cases. Gramnegative bacteria were the major pathogens involved in SBP in cirrhotic patients. Escherichia coli and streptococcus were the most common pathogen isolated. The pathogen isolated were globally sensitive to the antimicrobial usually recommended by the current guidelines. A larger-scale study will be needed to develop a national SBP guideline. Declarations Ethical approval and consent to participate The study was approved by the hierarchical heads of University Hospital Joseph Raseta Befelatanana, Antananarivo. Written informed consent was obtained from all participants.

Consent for publication
Written consent was obtained from all participants.

Availability of data and materials
Data available on request from the corresponding author.

Competing interests
The author declare that they have no competing interests.

Funding
None.
Authors' contributions ALRR, CIR and ASR: were the main contributors to drafting the manuscript. JDEL: contributed to data collection. JAR, NHR, SM, BMR, THR: contributed in management of patients in hospital and performed the nal manuscript. RMR : contributed to study design and performed the nal manuscript. All authors have read and approved the manuscript.