In the present observational study, a total of 88 cases suspected with pneumonia, admitted to Tertiary Hepatobiliary Centre, Institute of Liver and Biliary Sciences Hospital, New Delhi, India, from April to September 2018 were included. Ethical clearance was taken from the Institute Ethics Committee (IEC/2018/61/MA15) before commencement of the study. Cirrhotic patients with features suggestive of pneumonia were selected. Patients with: (i) Disseminated Intravascular coagulation/extremely moribund individuals with expected survival <48 hours; (ii) Inappropriately collected samples which fall under rejection criteria; (iii) Smear positive/culture positive/Cartridge Based Nucleic Acid Amplification Tests (CBNAAT) positive samples for M. tuberculosis were excluded from the study. After obtaining an informed written consent, BAL/ mini-BAL samples were collected from all the participants for routine laboratory work-up as per clinician request.

Both conventional culture media and multiplex real time PCR were used to detect bacterial pathogens. Culture media for bacterial growth included Blood and MacConkey agar [4]. The inoculated mediums were incubated for 24-48 hours in a 5% CO₂ incubator. For fungal examination, all the samples were inoculated on SDA media and incubated for four weeks [4]. For molecular analysis, DNA was extracted using QIAamp DNA mini kit (Qiagen, Germany). Multiplex PCR was carried out using real time thermocycler (Rotor-Gene Q, Qiagen) and standard detection kit (FTD Bacterial pneumonia_CAP, Fast Track Diagnostics, Luxemburg, catalogue no. FTD-29.1-32). Organisms targeted by multiplex real time PCR included Streptococcus pneumoniae, Staphylococcus aureus, Legionella pneumophila/longbeachae, Moraxella catarrhalis, Mycoplasma pneumoniae and Chlamydophila pneumoniae. Results were analysed on the basis of aetiological categorisation of liver diseases, radiological and microbiological results.

In the present study, samples from 69 men and 19 women aged 2-71 years were analysed. The diagnosis of patients included CLD (n=41), ACLF (n=26), Acute Liver Failure (ALF) (n=11), Non Cirrhotic Portal Fibrosis (NCPF) (n=1), and other condition: Pancreatitis, Secondary HLH, Kidney transplant, Carcinoma, Biliary atresia, Visceral Leishmaniasis (n=11). The aetiologies of the liver disease have been summarised in [Table/Fig-1]. All these patients with clinical suspicion of pneumonia underwent radiological examination, the findings of which have been categorised in [Table/Fig-2]. In the microbiology laboratory, a total of 9 out of 88 cases were positive either by culture or PCR [Table/Fig-3]. Culture results showed growth of Klebsiella pneumoniae, Acinetobacter baumannii, Staphylococcus aureus, Aspergillus flavus and Aspergillus fumigatus. Three samples were positive by PCR, among them one sample was positive for Legionella species on PCR and on culture this sample showed growth of A. flavus. The second sample was positive for both Legionella species and S. aureus on PCR. The third sample was positive for S. aureus on PCR and grew MSSA on culture.

Infections in CLD patients trigger rapid deterioration and liver disease progression. In 25% to 47% of hospitalised cirrhosis patients, the precipitating factor leading to acute decompensation are bacterial infections [5]. The condition of most of these patients rapidly deteriorate, ultimately progressing to multiple organ failure and septic shock. Pneumonia is a major cause of morbidity and mortality in cirrhotic patients [5]. Chest radiography is usually the first examination to be obtained for the evaluation of a patient with a suspicion of pneumonia. An overlapping pattern caused by atypical pathogens makes it essential to have a complete diagnostic armamentarium targeting typical and atypical pathogens [6]. Atypical pathogens are fastidious organisms that are difficult to culture. In our study, in addition to conventional culture methods, the respiratory panel was used to target the atypical pathogens that helped us to broaden the targeted organism for diagnosis and understand the respiratory pathogen profile in CLD patients.

PCR based nucleic acid detection methods have been proven to be more sensitive as compared to culture and serology testing. In a study by Pinar A et al., the multiplex methods used for diagnosis could identify 41.4% of cases as opposed to 23.4% using conventional methods and the most prevalent pathogens included S. pneumoniae, H. influenzae and M. pneumoniae [7]. In another study by Aydemir O et al., on BAL samples, reported better diagnostic value of multiplex PCR with 63.5% cases identification in comparison to culture which only yielded 31.5% of result positivity [8].

Due to limited number of studies on infections in liver diseases, the spectrum of respiratory pathogens in pneumonia patients is not well documented. Some of the most prevalent bacterial pathogens in liver disease patients with pneumonia include Enterococci, S. pneumoniae, H. influenzae, M. pneumoniae, Legionella, Enterobacteriaceae, P. aeruginosa and S. aureus [5]. In our study, the aetiological agents were Legionella (detected by PCR), S. aureus (detected by both PCR and culture), Enterobacteriaceae and Acinetobacter baumannii (detected by culture) and A. flavus and A. fumigatus (detected by culture). In addition, co-infections of Legionella infections with MSSA and A. flavus were also observed. However, unlike previous studies, the common respiratory pathogens such as S. pneumoniae, H. influenzae and M. pneumoniae were detected in our study population.

In our study, the two positive cases of Legionella species helped in identification of patients with atypical pathogens and early initiation of target therapy. The data on Legionellosis in liver disease from India is limited. Till date, no cases of Legionella in a cirrhotic patient have been reported. The bacterium Legionella has a unique susceptibility profile to fluoroquinolones and macrolides only. Non-specific clinical and radiological presentation makes it important to identify patients at risk of Legionella infection. Risk factors associated with Legionella infection includes people with age 50 years or older, smokers, COPD patients or immunosuppressed individuals, with underlying co-morbid conditions such as diabetes, renal failure or liver failure [9]. Thus, our study emphasises the need to test cirrhotic patients for Legionella infection. With more than 15 serotypes, L. pneumophila require special culture media like Buffered charcoal yeast extract agar, that requires 4-10 days to grow and this method is incapable of detecting ‘Viable But Not Culturable’ (VBNC) cells [10]. Legionella antigen testing, a rapid diagnostic test for L. pneumophila are used on urine samples but they can detect only L. pneumophila (serotype 01) [11]. Thus, PCR has an added advantage of detecting L. pneumophila non1 serotype, as in immunocompromised population non L. pneumophila is more prevalent [12]. Qin T et al., reported four cases of cirrhosis with Legionnaire’s disease from China where all the four cases were diagnosed using PCR and were successfully treated [10].

Fungal pathogens also pose a similar threat to the CLD patients [13]. Prolonged hospital stay, cirrhosis and antimicrobial therapy often leads to emergence of fungal infections in liver disease patients. In our study, 4 out of 88 cases had positive culture for Aspergillus species including both A. fumigatus and A. flavus. Liver diseases serve as a risk factor for invasive aspergillosis [14]. Differentiation between colonisation and Invasive Pulmonary Aspergillosis (IPA) mandates a complete clinical and radiological evaluation. IPA is a fatal cirrhosis associated complication with mortality rate exceeding 50% because of the immunosuppression in End Stage Liver Disease [ESLD] [14]. In a study by Prattes J et al., prevalence rate of IPA was found to be 1.3% [14]. In another study, fungal colonisation was found to be 25% in ESLD patients of which Aspergillosis constituted upto 3% [15]. Increased prevalence of IPA or aspergillus colonisation has been observed in the past decade, which mandates the need to screen fungal infections in liver disease patients.

In the present study, the sample size was relatively small. Further studies on larger study cohort are required to validate these findings. In addition, while multiplex PCR was used for all the samples, culture and serology for Legionella, Mycoplasma pneumoniae and Chlamydophila pneumoniae could not be done.

Liver disease patients need to be treated as a separate group of patients due to their predisposition to infections. An immunosuppressive state mandates the judicious use of antimicrobial prophylaxis to avoid development of antimicrobial resistance. Use of molecular techniques like PCR helps in identifying the atypical pathogens which is often neglected. Our study indicates that Legionella infection is a risk factor in liver disease patients. With a different susceptibility profile including macrolides and fluoroquinolones, early diagnosis help in limiting the use of indiscriminate antimicrobials, and timely initiation of targeted antimicrobial treatments.