Nosocomial Bacterial and Fungal Chest Infections in Cirrhotic Patients ‎

INTRODUCTION

In hospitalized cirrhotic patients, nosocomial chest infections (NCI) ‎are a frequent complication. Any chest infection contracted by a patient in a ‎hospital at least 48–72 hours after admission is considered to be NCI. In ‎cirrhotic patients, infectious complications are the most common cause of ‎death. Antibiotics should be started as soon as possible based on the culture ‎results [1].‎

The high risk of infection in patients with liver cirrhosis (LC) is ‎caused by defects in natural defense mechanisms, such as acquired and ‎progressive defects of the innate immune and reticuloendothelial systems, ‎which are exacerbated by the presence of alcohol (alcohol consumption; ‎alterations in the enteric flora; intrinsic cellular defects and the increasing ‎use of invasive procedures) [2].‎

The seriousness of the disease, pre-existing conditions, and the host's ‎reaction to infection are leading factors in poor outcome of these patients. ‎Infections are a common cause of morbidity and mortality in patients with ‎chronic liver disease [3].‎

NCI is considered a fatal complication in LC patients, but there are ‎few studies in this area. As a result, we created this study to examine the ‎pattern of nosocomial infection in cirrhotic patients, including the ‎incidence, organisms, and potential predictors of NCIs‎.

Study setting and design

A cross sectional hospital study was prospectively conducted at Intensive Care Unit (ICU) of the Tropical Medicine and Gastroenterology Department at , Al Rajhi Liver Hospital, Assiut University Hospitals, Assiut,  Egypt over one year duration between January and December 2019.

Inclusion& exclusion criteria

All patients with LC who were admitted to ICU during the study period were enrolled.  Patients with known chronic chest diseases as bronchial asthma or chronic obstructive lung disease, human immunodeficiency virus, history of organ transplantation with immunosuppression, and liver cirrhosis with Child A classification were excluded.

Patients

A total of 845 cirrhotic patients were admitted during the study period. 345  patients  were  excluded  (250  patients  had  compensated  LC (Child A))  and  95 patients had chronic chest disease). So a total 500 patients had met the inclusion criteria. Out of enrolled patients 100/500 (20%) patients developed nosocomial chest infection while 400/100 (80%) patients didn’t develop (figure 1).

Methods

Careful history taking and clinical evaluation were done to all enrolled patients. The clinical evaluation was repeated after two days to detect any manifestations of NCIs as development of fever, cough, dyspnea, and/or expectoration. Baseline laboratory assessment included liver and kidney functions, coagulation profile, complete blood count, and random blood sugar were achieved.

Plain chest radiograph was done at baseline and later on in case of suspicious of NCIs. Chest sonar was required if there was suspicious of underlying chest disease (if present, patient was excluded). All patients were assessed by abdominal ultrasound at baseline.

A sputum sample was obtained from cases with suspected NCIs for culture and antimicrobial sensitivity. In most cases, the sputum sample was collected early in the morning before patient ate or drunk anything after rinsing  the mouth with water. In case of very sick or unconscious patients, the sample was obtained by suction through nasotracheal catheter.

The following media were used for culture: chocolate agar (non-selective growth media), blood agar (isolation of streptococci), Macconky agar (isolation of gram negative bacilli), Mannitol salt agar (isolation of staph aureus), and Sabroud dextrose agar was used for isolation of fungi. Also, blood culture was performed in case of suspicious of NCIs on the same media of sputum culture. Any detected organism in the culture was defined under a microscope or by chemical tests. Sensitivity testing was performed to know the effective antimicrobial agents.

Statistical analysis

Data was analyzed those using SPSS (Statistical Package for the Social Science, version 20, IBM, and Armonk, New York). Continuous  data  was  expressed  in  form  of  mean  ±  SD while nominal data was expressed in form of frequency  (percentage). Chi²-test was used to compare the nominal data of patients with NCI and those without NCI while continuous data of both groups were compared with Student t test.  Precitors of NCI among hospitalized decompensated cirrhotic patients were determined by multivariate regression analysis. Level of  confidence  was  kept  at  95%  and  hence,  P  value  was  considered significant if <0.05.

RESULTS

Demographic and clinical data of enrolled patients (table 1):

It was noticed that patients with NCI had significantly higher mean age in comparison to those without NCI (63.68 ± 9.30 vs. 54.56 ± 10.34; P< 0.001). Also, majority (95%) of patients who developed NCI was 40 years old or more while majority of patients without NCI were < 40 years old.

Mean duration of liver cirrhosis was significantly higher among patients with NCI (7.89 ± 2.34 vs. 4.55 ± 1.20 years; P= 0.04). Frequency of hematemesis/ melena alone was significantly higher among those without NCI (200 (50%) vs. 25 (25%); P= 0.04) but patients with NCI had significantly higher frequency of hepatic encephalopathy (45 (45%) vs. 80 (20%); P= 0.03), spontaneous bacterial peritonitis (50 (50%) vs. 120 (30%); P= 0.01). Both groups showed no significant diffrences as regard baseline clinical evaluation (P> 0.05).

Baseline laboratory data and abdominal ultrasound of enrolled patients (table 2):

Both groups of patients had insignificant differences as regard baseline laboratory data and abdominal ultrasound with exception of significantly lower serum albumin (23.06 ± 6.18 vs. 29.11 ± 2.22 mg/dl; P= 0.03) in patients who developed NCI.

Radiological findings and cultures among patients with NCI (table 3):

Out of studied patients with NCI; 40 (40), 35 (35%) and 65 (65%) patients had pleural effusion, bronchopneumonia, and pneumonia, respectively. Ninety five patients had no growth in blood culture. Enterococcus, Actinobacter species, E.coli and Staph.aureus were observed in 2 (2%), 1 (1%), 1 (1%) and 1 (1%) patient, respectively.

The most detected organisms in sputum culture were Klebsiella species (32%), Candida albicans (19%), Strept.pneumoniae (12%) and Staph.aureus, respectively. It was noticed that Pseudomonas species was observed in 9(9%) while 8 (8%) patients had MRSA

Hospital stay and outcome of enrolled patients (table 4):

Patients with NCI had significantly longer duration of hospital stay (11.34 ± 2.34 vs. 8.34 ± 1.13 days; P= 0.01). It was noticed that majority (65%) of patients with NCI were dead while majority (60%) of patients with no-NCI were alive.

Predictors of NCI among enrolled patients (table 5):

Based on the current study, predictors for NCI in patients with advanced LC were; old age (OR=2.33, 95%CI= 1.09-4.65, p< 0.001), spontaneous bacterial peritonitis (OR=2.11, 95%CI= 2.01-5.67, p< 0.001), and low serum albumin (OR=4.01, 95%CI= 3.1-10.11, p<0.001).