Molecular characterization of nontuberculous mycobacteria in hospital waters: a two-year surveillance study in Tehran, Iran

Infections caused by nontuberculosis mycobacteria (NTM) have been recently reported as an important public health problem in many parts of the world (Prevots & Marras 2015) NTM infections are rarely spread from person to person and are mainly acquired from environmental sources including water, drinking water supplies, soil, etc. (Prevots & Marras 2015). They can cause serious infections in immunocompromised patients, including pulmonary disease which may resemble tuberculosis (TB) as well as cervical lymphadenopathy and skin and soft tissue infections (Griffith et al. 2007). NTM are often resistant to disinfectants and antibiotics, including anti-TB drugs, and treatment for one species may not be effective for others (Soni et al. 2016; Aguilar-Ayala et al. 2017). As NTM are usually resistant to disinfectants, it is not uncommon to isolate NTM from water distribution systems (Crago et al. 2014). NTM colonization and persistence within a hospital water distribution system may lead to dissemination and infection in patients (Vaerewijck et al. 2005; Castillo-Rodal et al. 2012; Mullis et al. 2013; Crago et al. 2014; Falkinham 2015, 2016). In countries such as Iran, where TB is endemic, we have suggested that there is an increasing prevalence of NTM disease, which is of concern (Nasiri et al. 2015, 2018a, 2018b). While monitoring hospital water distribution systems for NTM is not a routine practice in Iran and because of the significant impact that these infections may have on patients, the investigation of NTM seems to be important. Thus, the present study aimed to evaluate the frequency of NTM in hospital waters in Tehran, Iran.

A total of 218 samples from January 2016 to December 2017 was collected from tap water and medical devices (such as oxygen manometers, dialysis devices, nebulizers, and dental units) of different wards of the six hospitals in Tehran, the capital of Iran. Approximately 50 mL of each sample was collected in a sterile glass bottle, transferred to the laboratory in an icebox and analyzed within 24 hr. Stranded procedure recommended by the Centers for Disease Control and Prevention (CDC) was used to produce water for hemodialysis machines (Sehulster et al. 2003).

Water samples were centrifuged for 30 min at 3,000 rpm. The sediment (3 mL) was transferred to two sterile containers, and was decontaminated by our method (1.5 mL of NaOH 1% and SDS 3%) for 30 min at room temperature. Phenolphthalein and phosphoric acid 40% was used for the neutralization step. The samples were then centrifuged for 30 min at 3,000 rpm. The supernatant was discarded and the sediments of each treated sample were used to prepare a Ziehl–Neelsen smear and inoculated into two separate batches of Lowenstein Jensen (LJ) medium containing cyclohexamide 0.5 g and incubated at temperatures 37 °C and 25 °C for two months (Khosravi et al. 2016).

All mycobacterial isolates were grown on LJ medium and examined for growth rate, macroscopic and microscopic morphological features, growth at different temperatures, and also a set of biochemical tests including Tween 80 hydrolysis, nitrate reduction, niacin production, arylsulfatase, urease production, tellurite reduction, salt tolerance, and catalase production according to standard procedures (Nasiri et al. 2018a, 2018b).

IS6110-based polymerase chain reaction (PCR) assay was used for differentiation of NTM from Mycobacterium tuberculosis complex. Genomic DNA, for IS6110-based PCR assay, was extracted using QIAamp DNA Mini Kit (QIAGEN, USA) according to kit instructions. A 123 bp fragment of insertion element IS6110 of the M. tuberculosis complex was used as a target and amplified using previously described PCR primers (Eisenach et al. 1990). Genomic DNA of M. tuberculosis H37Rv (ATCC27294) and M. fortuitum (ATCC 49404) were used as positive and negative controls, respectively. The assay is negative for NTM species due to the absence of IS6110 (Eisenach et al. 1990).

An approximately 441 bp fragment of hsp65 gene was amplified by PCR using two specific primers Tb11 (50-ACCAACGATGGTGTGTCCAT-30) and Tb12 (50-CTTGTCGAACCGCATACCCT-30). PCR products were digested with 5 U of restriction enzyme HaeIII and BstII for 24 hours at 37 °C (Telenti et al. 1993). The pattern of digested products was analyzed using 8% polyacrylamide gel. M. fortuitum (ATCC 49404) and double distilled water were used as positive and negative control in all PCR experiments, respectively. Species identification was performed using algorithms proposed by Roth et al. (2000) and Telenti et al. (1993).

A 441 bp fragment of hsp65 gene was amplified by PCR with two specific primers Tb11 (5′-ACCAACGATGGTGTGTCCAT-3′) and Tb12 (5′-CTTGTC GAACCGCA-TACCCT-3′) (Telenti et al. 1993). The cycling condition was 95 °C for 5 min, followed by 35 cycles of 94 °C for 30 s, 62 °C for 30 s, and 72 °C for 1 min and finalized with 72 °C for 5 min. Distilled water was used as a negative control. PCR products of hsp65 genes were analyzed on 1% agarose gel electrophoresis. For confirmation of results, a number of samples were randomly examined for rpoB gene by using two specific primers MycoF (50-GGCAAGGTCACCCCGAAGGG-30) and MycoR (50-AGCGGCTGCTGGGTGATCATC-30) (Adékambi et al. 2003).

The obtained sequences for each isolate were aligned separately and compared with all existing relevant sequences of mycobacteria retrieved from GenBank database at the NCBI website via the nucleotide BLAST search.

Out of 218 samples from different water sources of hospitals in Tehran, 85 (39.0%) samples at 37 °C and 87 (40.0%) samples at 25 °C were identified as NTM using conventional and molecular methods (Tables 1 and 2). Using hsp65-PRA and sequencing methods, M. lentiflavum (n: 72, 84.7%) was the most frequently encountered, followed by M. gordonae (n: 11, 13.0%) and M. paragordonae (n: 2, 2.3%) at 37 °C (Table 3). No significant difference was seen in frequency and species in mycobacteria isolated at temperatures of 37 °C and 25 °C. Randomly selected NTM isolates were also confidently identified by rpoB gene.

The frequency of NTM species in different hospital water sources are presented in Tables 4 and 5. Humidifying cup of oxygen manometer had the most contaminated water among the investigated water distribution systems in the hospitals.

There are several reports of nosocomial infections caused by waterborne NTM, indicating that this route of transmission may cause significant impact on patients (Crago et al. 2014). This study found that an unexpected number of the examined isolates throughout the hospital water distribution systems over the two-year surveillance were NTM, in particular M. lentiflavum and M. gordonae. Although M. lentiflavum and M. gordonae are generally considered non-pathogenic, there are cases describing infections caused by these mycobacteria. Similar findings were reported by Khosravi et al. (2016) in Khuzestan province of Iran, in which of 77 culture-positive mycobacteria isolated from 258 hospital water samples, M. fortuitum and M. gordonae were among the most prevalent isolates. Likewise, in other studies conducted in Iran, M. gordonae and M. lentiflavum were among the most prevalent NTM isolated from water samples (Bahram et al. 2012; Moghim et al. 2012; Azadi et al. 2016). Our results were also relatively similar to other investigations on hospital waters in the world (Angenent et al. 2005; Hussein et al. 2009; Genc et al. 2013). In a study in Turkey, a total of 160 water samples from hot and cold water in two hospitals was examined (Genc et al. 2013). Out of 33 isolated strains, 20 (60/6%) were M. lentiflavum and the remainder were M. gordonae (30/3%) and M. peregrinum (9/1%) (Genc et al. 2013). Furthermore, in our recent publication we showed that M. gordonae was among the dominant pathogens recovered from clinical samples, highlighting the role of hospital tap water as a vector for transmission of the bacteria (Nasiri et al. 2018a, 2018b).

Due to the risk of NTM-contaminated hospital water supplies for nosocomial infections, preventive measures should be considered. For water supplies, it is suggested to prevent water stagnating in water supplies and to regularly disinfect regions at risk (i.e., reservoirs and drinking fountains) (Dailloux et al. 1999). Furthermore, in the tap water used to rinse haemodialyzers, the use of water inlet filters is recommended (Dailloux et al. 1999). For healthy people, a number of common measures should be indicated, namely, disinfection of wounds in the case of an accidental injury and protecting them from water.

M. gordonae was the second most common NTM in our study. This mycobacterium is usually found in water, soil, and raw milk and is known as saprophytic and a rarely pathogenic bacterium in humans (Wolinsky 1979; Jarikre 2011). M. gordonae (the tap water bacillus) has been reported responsible for infections including skin and soft tissues infections (Gengoux et al. 1987; Modilevsky et al. 1989), lung disease (de Gracia et al. 1989), and liver and peritoneal infections (Kurnik et al. 1983). Based on recent studies, M. gordonae is one of the most common mycobacterium isolated in hospital waters (Angenent et al. 2005; Genc et al. 2013).

M. lentiflavum, another prevalent NTM in the current study, was first described in 1996 (Springer et al. 1996). This mycobacterium has been shown to grow in a wide range of temperatures (22 to 37 °C), and can cause human disease in both immunocompetent and immunocompromised individuals (Springer et al. 1996; Cabria et al. 2002; Tortoli et al. 2006; Marshall et al. 2011). In a study conducted in Australia in 2001 and 2008, the relationship between genotype and geographical similarities between M. lentiflavum isolates from patients and drinking water was evaluated (Marshall et al. 2011). Based on their results, genotypes of the environmental clone of M. lentiflavum were close to the strains that were isolated from patients. This finding suggests that drinking water can be the source of infection for M. lentiflavum. Similarly, in the study of Torvinen et al. (2004) over 90% of the mycobacteria isolated from water belonged to M. lentiflavum and M. gordonae. Likewise, in South Korea, Lee et al. (2008) showed that 65% of the samples obtained from drinking water were M. lentiflavum. Although for setting up the decontamination method we evaluated several protocols, we should mention that due to the method used for decontamination, some species might not have been isolated.

The rising number of NTM isolated from hospital waters in Iran may have several negative effects on public health status. Importantly, most TB laboratories in Iran are not equipped to perform NTM species identification; thus, culture reported positive for TB may likely be the result of laboratory contamination with environmental NTM. Consequently, NTM may be misdiagnosed as TB which results in unnecessary anti-TB treatment (Nasiri et al. 2015).

Another important finding in our study is the contamination of medical devices such as oxygen manometer (humidifying cup), nebulizers and hemodialysis fluid with mycobacteria which is a significant health risk for hospitalized patients. Recently, Heidarieh et al. (2016) in Iran indicated that a diverse bacterial community, containing predominantly mycobacteria, was detected in a hospital's hemodialysis distribution system. Due to poor immune system in these patients, microbiological monitoring of water used for hemodialysis and other medical devices is very important for the prevention of NTM-associated illness (Montanari et al. 2009; Heidarieh et al. 2016). Recently, transmission of M. chimaera from heater–cooler units during cardiac surgery has been reported (Sax et al. 2015; Sommerstein et al. 2016). Thus, it is recommended that standard protocols for the evaluation of microbial contamination should be used for regular monitoring and identification of NTM in medical devices.

Isolation of NTM from hospital water sources is a serious public health problem in Iran and merits further attention by health authorities. The establishment of microbiological monitoring systems for hospital waters, and expanding the number of facilitated laboratories are strongly recommended.

This study has been supported by Deputy of Research and Technology, Iran University of Medical Sciences, Grant No. 25606–30-01-94. All contributing authors declare no conflicts of interest.