Liver International ISSN 1478-3223
CIRRHOSIS AND LIVER FAILURE
The chronic use of beta-blockers and proton pump inhibitors may affect the rate of bacterial infections in cirrhosis Manuela Merli1, Cristina Lucidi1, Vincenza Di Gregorio1, Valerio Giannelli1, Michela Giusto1, Giancarlo Ceccarelli2, Oliviero Riggio1 and Mario Venditti2 1 Gastroenterology, Department of Clinical Medicine, ‘Sapienza’ University of Rome, Rome, Italy 2 Department of Infectious Disease, ‘Sapienza’ University of Rome, Rome, Italy
Keywords drugs – non-absorbable antibiotics – nonabsorbable disaccharides – sepsis Abbreviations BB, beta-blocker; MDI, microbiologically documented infections; NGTF, non-typical gut flora; PPI, proton pump inhibitor; SBP, spontaneous bacterial peritonitis; TGF, typical gut flora. Correspondence Professor Manuela Merli, MD II Gastroenterologia, Dipartimento di Medicina Clinica ‘Sapienza’ Universit a di Roma, Viale dell’Universit a 37, 00185 Roma Tel: +390649972001 Fax: +39064453319 / +39064440806 e-mail: [email protected] Received 2 February 2014 Accepted 11 May 2014 DOI:10.1111/liv.12593
Abstract Background & Aims: Bacterial infections are among the most common and life-threatening complications in cirrhosis. Qualitative and quantitative modifications of the gut microbiota, dysfunction of the intestinal barrier and multiple immune defects are factors that contribute to a pathological ‘bacterial translocation’ (BT), leading to a higher susceptibility to infections in cirrhotic patients. Long-term therapies, commonly adopted in cirrhotic patients, may influence BT and modify the risk of infection in these patients. To investigate the influence of chronic therapies on the prevalence and microbiological characteristics of infections in cirrhosis. Methods: Consecutive cirrhotic patients hospitalised from 2008 to 2013 were enrolled. All previous treatments were carefully recorded. Infections were actively sought out, patients were actively monitored for infection, and possible risk factors were evaluated. Results: Four hundred cirrhotic patients were included. The most frequent therapies were proton pump inhibitors (PPIs) (67%), non-absorbable-disaccharides (44%), beta-blockers (BBs) (39%) and non-absorbableantibiotics (10%). Child-Pugh C (P < 0.001; OR 5; 95%CI: 2.6–9.9) and PPI therapy (P = 0.008; OR 2; 95% CI: 1.2–3.2) were found to be independent predictors of infection, and the use of BBs was a protective factor (P = 0.001; OR 0.46; 95%CI: 0.3–0.7). Cirrhotic patients with bacterial infection showed lower morbidity and mortality when taking BBs. Conclusions: Proton pump inhibitors increase the risk of infection in cirrhosis and should not be prescribed in these patients without specific indications. In contrast, the use of BBs is associated with a lower rate of infection and attenuates the consequences of infections in cirrhotic patients.
Bacterial infection represents a severe complication with a significant impact on morbidity and mortality in cirrhotic patients (1–4). A relevant pathogenic factor contributing to the high susceptibility to infections in cirrhosis is pathological gut bacterial translocation (BT). This process, defined as the migration of viable microorganisms and microbial products from the gut to the portal venous system, mesenteric lymph nodes and other extra-intestinal sites, may occasionally also occur in healthy individuals but is significantly more common under pathological conditions (5). In patients with advanced liver disease, BT is increased because of multiple factors, e.g. quantitative and/or qualitative imbalance of intestinal microbiota, local immune deficiency and alterations in the intestinal barrier (6). Small intestinal bacterial overgrowth (SIBO), defined as a bacterial load greater than 105 CFU/ml in small intestine aspirate, has been frequently reported in Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
cirrhotic patients (7). Small intestinal bacterial overgrowth may be favoured by alterations in the mechanisms able to restrict bacterial colonisation in the gut; in fact, a failure of the gastric acid barrier may lead to an overgrowth of ‘upper respiratory tract’ flora (characterised by a high load of Gram-positive bacteria), and a decrease in intestinal motility may facilitate an overgrowth of intestinal flora (characterised by a high load of Enterobacteriaceae and Enterococcus) (8). It is well known that some medications may interact with the above reported mechanisms. Different treatments modulate the intestinal microbial composition by decreasing the oro-caecal transit time (intestinal prokinetics and beta-blockers) (9, 10), by modifying the prevalence of different bacterial species (non-absorbable or systemic antibiotics, probiotics and prebiotics) (11–15) or through both these mechanisms (non-absorbable disaccharides such as lactulose) (16).
1
Merli et al.
Chronic therapies and infections in cirrhosis
Beta-blockers (BBs), which are b-adrenergic receptor antagonists, are known to shorten the intestinal transit time and to normalise intestinal permeability by reducing portal pressure (17–19). These drugs are largely prescribed in cirrhotic patients for first or secondary prophylaxis of variceal bleeding, although it has been suggested that caution should be exercised in their use in advance liver disease because of a possible impairment of the haemodynamic response during acute complications (20, 21). Proton pump inhibitors (PPIs) are also frequently included in the therapeutic armamentarium of cirrhotic patients. In the general population, PPIs may increase the risk of infection (22–24) through a reduction in gastric acidity, a delay in gastric emptying and a direct effect on the immunological system (inhibition of neutrophil, cytotoxic T lymphocyte and natural killer cell activity) (25–28). It has been suggested that PPIs increase the risk of SBP in cirrhotic patients (29–35), but this was not clearly confirmed in all studies (36, 37). The present cross-sectional study aimed to investigate the possible association between chronic therapies likely to influence the above-cited mechanisms and bacterial infections in cirrhosis. For this purpose, we considered a large cohort of hospitalised cirrhotic patients prospectively enrolled in an infection surveillance programme. Patients and methods Patients
All cirrhotic patients consecutively admitted at our Department University hospital from October 2008 to January 2013 were considered for enrolment. The exclusion criteria were concomitant HIV infection, high-dose corticosteroid treatment, immunosuppressive therapy, hepatocellular carcinoma out of the Milan criteria and systemic antibiotic therapy in the last 4 weeks. The chronic consumption of non-absorbable antibiotics and quinolone prophylaxis for SBP were not a cause of exclusion. At admission, demographical, clinical and biochemical data were recorded. The severity of liver disease was assessed according to Child-Pugh and model of endstage liver disease (MELD) scores. This study was approved by the local Ethical Committee Review Board, and informed consent was obtained by all the participants to allow the collection of their medical records.
According to their epidemiological characteristics, infections were classified as follows: (1) Hospital acquired if the diagnosis of infection was made after more than 48 h of hospital stay. (2) Healthcare associated if the diagnosis was made within 48 h of hospitalisation in patients with any of the following criteria: (i) had attended a hospital or a haemodialysis clinic, (ii) was hospitalised for at least 2 days or had undergone surgery during the 180 days before infection or (iii) had resided in a nursing home or a long-term care facility. (3) Community acquired if the diagnosis of infection was made within 48 h of hospitalisation and the patient did not fulfil the criteria for healthcare-associated infection (i.e. had no recent contact with the healthcare system and had not been hospitalised in the last 6 months) (38). Isolated pathogens were classified as ‘typical gut flora’ (TGF) (Enterobacteriaceae, Enterococcus and Streptococcus bovis) and ‘not typical gut flora’ (NTGF) (oral streptococci, Streptococcus pyogenes and Staphylococcus). As it has been shown that chronic quinolone prophylaxis increases Staphylococcus spp. in the gut, these bacteria were also considered TGF in these patients (39). Sepsis was diagnosed according to the international guidelines as systemic inflammatory syndrome associated with infection and was classified in three stages of severity, namely, ‘sepsis’, ‘severe sepsis’ (when acute organ failure is attributed to sepsis) and ‘septic shock’ (when hypotension unresponsive to intravascular volume loading and peripheral perfusion abnormalities require inotropic support) (40, 41). Chronic therapies
Therapeutic regimens at baseline were carefully noted with regard to dose and duration of therapy. Patients were arbitrarily considered ‘chronic drug users’ when the treatment started at least 4 weeks prior to the admission. This timing was chosen according to pharmacodynamics and in keeping with the available literature. Regarding PPIs, we defined as ‘standard dosages’ the daily administration of 20 mg of omeprazole, 30 mg of lansoprazole and 40 mg of pantoprazole or esomeprazole. Results were reported for drugs more frequently utilised, as follows: PPIs, BBs, prokinetics, non-absorbable disaccharides, quinolones, non-absorbable antibiotics and laxatives. Statistical analysis
Assessment of infections and sepsis
Episodes of infections were always actively sought out either at admission or during the hospital stay. The diagnosis of bacterial infection was based on previously reported standard criteria (2). Infections with a positive culture were defined as ‘microbiologically documented infections’.
2
All the values are reported as the means ± SD, and P values < 0.05 were considered significant. The data were analysed as continuous or categorical using the Student’s t test for parametric data and the Mann– Whitney U test or Wilcoxon test for non-parametric data. The Chi-square test was used for the comparison of dichotomous data. A univariate analysis (Log Rank) Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Merli et al.
Chronic therapies and infections in cirrhosis
was used to identify drugs associated with the development of infections, and the variables selected by this analysis were included in a multivariate analysis according to a Cox regression model. The software used for the analysis was NCSS (Number Cruncher Statistical System) 2007. Results Patients, infections and outpatient therapies
Thirty-one patients were excluded, five because they were taking immunosuppressive medications, 12 for a diagnosis of hepatocellular carcinoma out of the Milan criteria and 14 because they reported the consumption of antibiotics in the last month for the treatment of an infectious episode. Four hundred patients were enrolled, 119 women and 281 men, with a mean age of 61.5 ± 13 years. The aetiology of cirrhosis was more frequently hepatitis C (47.5% of cases), alcohol abuse (22.5%) and hepatitis B (9%). The Child-Pugh class was A in 24%, B in 49% and C in 27% of the patients. The mean MELD score at admission was 13.6 ± 5.5. An episode of bacterial infection was documented in 140 patients (35%); of them, 27 patients had two infectious episodes. Seventy-four cases were complicated by sepsis. The main sites of infections were urinary tract (43%), respiratory tract (19%) and ascites (13%). The demographic, clinical and biochemical characteristics of the patients included in this study according to the diagnosis of infection are shown in Table 1. Age, sex, prevalence of comorbidities (diabetes and organic renal failure), origin of liver disease and prevalence of gastrointestinal bleeding during the hospitalisation were
similar in the two groups. Patients with infections had more severe liver disease, as documented by the ChildPugh and MELD scores. With regard to chronic therapeutic regimes, the class of drugs more frequently utilised was PPIs (269 patients; 67%). Within this class, the majority of the patients (86%) were taking medications at a ‘standard dosage’. An appropriate indication was lacking in a considerable number (41%) of patients taking PPI therapy. Other drugs largely utilised were non-absorbable disaccharides (178 patients), BBs for primary or secondary prophylaxis for variceal bleeding (156 patients) (98% propranolol) and non-absorbable antibiotics (rifaximin) (42 patients). Thirteen patients were taking osmotic laxatives, 12 norfloxacin as secondary prophylaxis of SBP, six prokinetics such as metoclopramide and five probiotics. Three patients had initiated lactulose therapy only 10 days before the hospital admission and were therefore excluded from the chronic lactulose user group, and two patients were taking PPI therapy for 7 days and were not considered to be chronic PPI users. As shown in Table 1, patients with infections were more frequently PPI users (76%) and less frequently BB users (26%). To exclude the role of confounding factors, demographic, clinical characteristics and comorbidities of the cirrhotic patients were evaluated according to the consumption of these two drugs (Tables 2 and 3). The clinical characteristics, severity of liver disease and comorbidities were not significantly different. As expected, BB users had a higher prevalence of oesophageal varices. It is worth noting that PPI users had a higher prevalence of infections despite a trend toward a better liver function (Child-Pugh score was lower in PPI users vs. nonusers; P = 0.062).
Table 1. Demographical, clinical and biochemical characteristics of patients with and without infections included in this study Variables Age (years) Male/Females Diabetes mellitus, n (%) Organic renal failure, n (%) Origin of liver disease Alcohol, n (%) HCV-related, n (%) HBV-related, n (%) Others, n (%) Ascites, n (%) Child-Pugh Score MELD Proton pump inhibitors users, n (%) Beta-blockers users, n (%) Prokinetics users, n (%) Non-absorbable disaccharides users, n (%) Quinolone users, n (%) Non-absorbable antibiotics users, n (%) Laxatives users, n (%)
Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Patients without infections (n = 260)
Patients with infections (n = 140)
P
61.6 ± 12 185/75 74 (29) 20 (8)
61 ± 13 96/44 50 (36) 18 (14)
Ns Ns Ns Ns
56 (22) 131 (50) 23 (8) 50 (20) 114 (44) 7.6 ± 2 12.4 ± 4.4 163 (63) 119 (46) 5 (3) 107 (41) 6 (4) 24 (15) 10 (6)
34 (24) 59 (42) 13 (9) 106 (25) 92 (60) 8.8 ± 1.9 15.7 ± 6.6 106 (76) 37 (26) 1 (1) 71 (50) 6 (7) 18 (20) 3 (4)
Ns
CIRRHOSIS AND LIVER FAILURE
The chronic use of beta-blockers and proton pump inhibitors may affect the rate of bacterial infections in cirrhosis Manuela Merli1, Cristina Lucidi1, Vincenza Di Gregorio1, Valerio Giannelli1, Michela Giusto1, Giancarlo Ceccarelli2, Oliviero Riggio1 and Mario Venditti2 1 Gastroenterology, Department of Clinical Medicine, ‘Sapienza’ University of Rome, Rome, Italy 2 Department of Infectious Disease, ‘Sapienza’ University of Rome, Rome, Italy
Keywords drugs – non-absorbable antibiotics – nonabsorbable disaccharides – sepsis Abbreviations BB, beta-blocker; MDI, microbiologically documented infections; NGTF, non-typical gut flora; PPI, proton pump inhibitor; SBP, spontaneous bacterial peritonitis; TGF, typical gut flora. Correspondence Professor Manuela Merli, MD II Gastroenterologia, Dipartimento di Medicina Clinica ‘Sapienza’ Universit a di Roma, Viale dell’Universit a 37, 00185 Roma Tel: +390649972001 Fax: +39064453319 / +39064440806 e-mail: [email protected] Received 2 February 2014 Accepted 11 May 2014 DOI:10.1111/liv.12593
Abstract Background & Aims: Bacterial infections are among the most common and life-threatening complications in cirrhosis. Qualitative and quantitative modifications of the gut microbiota, dysfunction of the intestinal barrier and multiple immune defects are factors that contribute to a pathological ‘bacterial translocation’ (BT), leading to a higher susceptibility to infections in cirrhotic patients. Long-term therapies, commonly adopted in cirrhotic patients, may influence BT and modify the risk of infection in these patients. To investigate the influence of chronic therapies on the prevalence and microbiological characteristics of infections in cirrhosis. Methods: Consecutive cirrhotic patients hospitalised from 2008 to 2013 were enrolled. All previous treatments were carefully recorded. Infections were actively sought out, patients were actively monitored for infection, and possible risk factors were evaluated. Results: Four hundred cirrhotic patients were included. The most frequent therapies were proton pump inhibitors (PPIs) (67%), non-absorbable-disaccharides (44%), beta-blockers (BBs) (39%) and non-absorbableantibiotics (10%). Child-Pugh C (P < 0.001; OR 5; 95%CI: 2.6–9.9) and PPI therapy (P = 0.008; OR 2; 95% CI: 1.2–3.2) were found to be independent predictors of infection, and the use of BBs was a protective factor (P = 0.001; OR 0.46; 95%CI: 0.3–0.7). Cirrhotic patients with bacterial infection showed lower morbidity and mortality when taking BBs. Conclusions: Proton pump inhibitors increase the risk of infection in cirrhosis and should not be prescribed in these patients without specific indications. In contrast, the use of BBs is associated with a lower rate of infection and attenuates the consequences of infections in cirrhotic patients.
Bacterial infection represents a severe complication with a significant impact on morbidity and mortality in cirrhotic patients (1–4). A relevant pathogenic factor contributing to the high susceptibility to infections in cirrhosis is pathological gut bacterial translocation (BT). This process, defined as the migration of viable microorganisms and microbial products from the gut to the portal venous system, mesenteric lymph nodes and other extra-intestinal sites, may occasionally also occur in healthy individuals but is significantly more common under pathological conditions (5). In patients with advanced liver disease, BT is increased because of multiple factors, e.g. quantitative and/or qualitative imbalance of intestinal microbiota, local immune deficiency and alterations in the intestinal barrier (6). Small intestinal bacterial overgrowth (SIBO), defined as a bacterial load greater than 105 CFU/ml in small intestine aspirate, has been frequently reported in Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
cirrhotic patients (7). Small intestinal bacterial overgrowth may be favoured by alterations in the mechanisms able to restrict bacterial colonisation in the gut; in fact, a failure of the gastric acid barrier may lead to an overgrowth of ‘upper respiratory tract’ flora (characterised by a high load of Gram-positive bacteria), and a decrease in intestinal motility may facilitate an overgrowth of intestinal flora (characterised by a high load of Enterobacteriaceae and Enterococcus) (8). It is well known that some medications may interact with the above reported mechanisms. Different treatments modulate the intestinal microbial composition by decreasing the oro-caecal transit time (intestinal prokinetics and beta-blockers) (9, 10), by modifying the prevalence of different bacterial species (non-absorbable or systemic antibiotics, probiotics and prebiotics) (11–15) or through both these mechanisms (non-absorbable disaccharides such as lactulose) (16).
1
Merli et al.
Chronic therapies and infections in cirrhosis
Beta-blockers (BBs), which are b-adrenergic receptor antagonists, are known to shorten the intestinal transit time and to normalise intestinal permeability by reducing portal pressure (17–19). These drugs are largely prescribed in cirrhotic patients for first or secondary prophylaxis of variceal bleeding, although it has been suggested that caution should be exercised in their use in advance liver disease because of a possible impairment of the haemodynamic response during acute complications (20, 21). Proton pump inhibitors (PPIs) are also frequently included in the therapeutic armamentarium of cirrhotic patients. In the general population, PPIs may increase the risk of infection (22–24) through a reduction in gastric acidity, a delay in gastric emptying and a direct effect on the immunological system (inhibition of neutrophil, cytotoxic T lymphocyte and natural killer cell activity) (25–28). It has been suggested that PPIs increase the risk of SBP in cirrhotic patients (29–35), but this was not clearly confirmed in all studies (36, 37). The present cross-sectional study aimed to investigate the possible association between chronic therapies likely to influence the above-cited mechanisms and bacterial infections in cirrhosis. For this purpose, we considered a large cohort of hospitalised cirrhotic patients prospectively enrolled in an infection surveillance programme. Patients and methods Patients
All cirrhotic patients consecutively admitted at our Department University hospital from October 2008 to January 2013 were considered for enrolment. The exclusion criteria were concomitant HIV infection, high-dose corticosteroid treatment, immunosuppressive therapy, hepatocellular carcinoma out of the Milan criteria and systemic antibiotic therapy in the last 4 weeks. The chronic consumption of non-absorbable antibiotics and quinolone prophylaxis for SBP were not a cause of exclusion. At admission, demographical, clinical and biochemical data were recorded. The severity of liver disease was assessed according to Child-Pugh and model of endstage liver disease (MELD) scores. This study was approved by the local Ethical Committee Review Board, and informed consent was obtained by all the participants to allow the collection of their medical records.
According to their epidemiological characteristics, infections were classified as follows: (1) Hospital acquired if the diagnosis of infection was made after more than 48 h of hospital stay. (2) Healthcare associated if the diagnosis was made within 48 h of hospitalisation in patients with any of the following criteria: (i) had attended a hospital or a haemodialysis clinic, (ii) was hospitalised for at least 2 days or had undergone surgery during the 180 days before infection or (iii) had resided in a nursing home or a long-term care facility. (3) Community acquired if the diagnosis of infection was made within 48 h of hospitalisation and the patient did not fulfil the criteria for healthcare-associated infection (i.e. had no recent contact with the healthcare system and had not been hospitalised in the last 6 months) (38). Isolated pathogens were classified as ‘typical gut flora’ (TGF) (Enterobacteriaceae, Enterococcus and Streptococcus bovis) and ‘not typical gut flora’ (NTGF) (oral streptococci, Streptococcus pyogenes and Staphylococcus). As it has been shown that chronic quinolone prophylaxis increases Staphylococcus spp. in the gut, these bacteria were also considered TGF in these patients (39). Sepsis was diagnosed according to the international guidelines as systemic inflammatory syndrome associated with infection and was classified in three stages of severity, namely, ‘sepsis’, ‘severe sepsis’ (when acute organ failure is attributed to sepsis) and ‘septic shock’ (when hypotension unresponsive to intravascular volume loading and peripheral perfusion abnormalities require inotropic support) (40, 41). Chronic therapies
Therapeutic regimens at baseline were carefully noted with regard to dose and duration of therapy. Patients were arbitrarily considered ‘chronic drug users’ when the treatment started at least 4 weeks prior to the admission. This timing was chosen according to pharmacodynamics and in keeping with the available literature. Regarding PPIs, we defined as ‘standard dosages’ the daily administration of 20 mg of omeprazole, 30 mg of lansoprazole and 40 mg of pantoprazole or esomeprazole. Results were reported for drugs more frequently utilised, as follows: PPIs, BBs, prokinetics, non-absorbable disaccharides, quinolones, non-absorbable antibiotics and laxatives. Statistical analysis
Assessment of infections and sepsis
Episodes of infections were always actively sought out either at admission or during the hospital stay. The diagnosis of bacterial infection was based on previously reported standard criteria (2). Infections with a positive culture were defined as ‘microbiologically documented infections’.
2
All the values are reported as the means ± SD, and P values < 0.05 were considered significant. The data were analysed as continuous or categorical using the Student’s t test for parametric data and the Mann– Whitney U test or Wilcoxon test for non-parametric data. The Chi-square test was used for the comparison of dichotomous data. A univariate analysis (Log Rank) Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Merli et al.
Chronic therapies and infections in cirrhosis
was used to identify drugs associated with the development of infections, and the variables selected by this analysis were included in a multivariate analysis according to a Cox regression model. The software used for the analysis was NCSS (Number Cruncher Statistical System) 2007. Results Patients, infections and outpatient therapies
Thirty-one patients were excluded, five because they were taking immunosuppressive medications, 12 for a diagnosis of hepatocellular carcinoma out of the Milan criteria and 14 because they reported the consumption of antibiotics in the last month for the treatment of an infectious episode. Four hundred patients were enrolled, 119 women and 281 men, with a mean age of 61.5 ± 13 years. The aetiology of cirrhosis was more frequently hepatitis C (47.5% of cases), alcohol abuse (22.5%) and hepatitis B (9%). The Child-Pugh class was A in 24%, B in 49% and C in 27% of the patients. The mean MELD score at admission was 13.6 ± 5.5. An episode of bacterial infection was documented in 140 patients (35%); of them, 27 patients had two infectious episodes. Seventy-four cases were complicated by sepsis. The main sites of infections were urinary tract (43%), respiratory tract (19%) and ascites (13%). The demographic, clinical and biochemical characteristics of the patients included in this study according to the diagnosis of infection are shown in Table 1. Age, sex, prevalence of comorbidities (diabetes and organic renal failure), origin of liver disease and prevalence of gastrointestinal bleeding during the hospitalisation were
similar in the two groups. Patients with infections had more severe liver disease, as documented by the ChildPugh and MELD scores. With regard to chronic therapeutic regimes, the class of drugs more frequently utilised was PPIs (269 patients; 67%). Within this class, the majority of the patients (86%) were taking medications at a ‘standard dosage’. An appropriate indication was lacking in a considerable number (41%) of patients taking PPI therapy. Other drugs largely utilised were non-absorbable disaccharides (178 patients), BBs for primary or secondary prophylaxis for variceal bleeding (156 patients) (98% propranolol) and non-absorbable antibiotics (rifaximin) (42 patients). Thirteen patients were taking osmotic laxatives, 12 norfloxacin as secondary prophylaxis of SBP, six prokinetics such as metoclopramide and five probiotics. Three patients had initiated lactulose therapy only 10 days before the hospital admission and were therefore excluded from the chronic lactulose user group, and two patients were taking PPI therapy for 7 days and were not considered to be chronic PPI users. As shown in Table 1, patients with infections were more frequently PPI users (76%) and less frequently BB users (26%). To exclude the role of confounding factors, demographic, clinical characteristics and comorbidities of the cirrhotic patients were evaluated according to the consumption of these two drugs (Tables 2 and 3). The clinical characteristics, severity of liver disease and comorbidities were not significantly different. As expected, BB users had a higher prevalence of oesophageal varices. It is worth noting that PPI users had a higher prevalence of infections despite a trend toward a better liver function (Child-Pugh score was lower in PPI users vs. nonusers; P = 0.062).
Table 1. Demographical, clinical and biochemical characteristics of patients with and without infections included in this study Variables Age (years) Male/Females Diabetes mellitus, n (%) Organic renal failure, n (%) Origin of liver disease Alcohol, n (%) HCV-related, n (%) HBV-related, n (%) Others, n (%) Ascites, n (%) Child-Pugh Score MELD Proton pump inhibitors users, n (%) Beta-blockers users, n (%) Prokinetics users, n (%) Non-absorbable disaccharides users, n (%) Quinolone users, n (%) Non-absorbable antibiotics users, n (%) Laxatives users, n (%)
Liver International (2014) © 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Patients without infections (n = 260)
Patients with infections (n = 140)
P
61.6 ± 12 185/75 74 (29) 20 (8)
61 ± 13 96/44 50 (36) 18 (14)
Ns Ns Ns Ns
56 (22) 131 (50) 23 (8) 50 (20) 114 (44) 7.6 ± 2 12.4 ± 4.4 163 (63) 119 (46) 5 (3) 107 (41) 6 (4) 24 (15) 10 (6)
34 (24) 59 (42) 13 (9) 106 (25) 92 (60) 8.8 ± 1.9 15.7 ± 6.6 106 (76) 37 (26) 1 (1) 71 (50) 6 (7) 18 (20) 3 (4)
Ns
