Nosocomial pseudo-outbreak of Mycobacterium gordonae associated with a hospital's water supply contamination: a case series of 135 patients

AECOPD

acute exacerbation of chronic obstructive pulmonary disease

AFB

acid-fast bacilli

ATS

American Thoracic Society

IDSA

Infectious Diseases Society of America

IGRA

interferon gamma release assay

IPF

idiopathic pulmonary fibrosis

MTB

Mycobacterium tuberculosis

NTM

nontuberculous mycobacteria

TB

tuberculosis

Nontuberculous mycobacteria (NTM) encompass all Mycobacterium species other than Mycobacterium tuberculosis (MTB) complex and Mycobacterium leprae. Recently, these organisms have been coming into the limelight due to an increased isolation frequency of NTM as well as an increasing incidence of pulmonary NTM disease (Griffith et al. 2007; Marras et al. 2007; Martín-Casabona et al. 2004; Andréjak et al. 2010; Van Ingen et al. 2010). Unlike MTB, NTM are present in soil and water and infection is thought to be acquired from the environment rather than transmitted person-to-person, with very rare exceptions (Aitken et al. 2012). Due to their ubiquitous presence in municipal water supplies, exposure to NTM is common (Von Reyn et al. 1994; De Groote et al. 2006; Falkinham et al. 2008; Feazel et al. 2009; Falkinham 2011), and NTM can colonize the respiratory tract without actually causing a disease.

Currently, rapid tests for identifying MTB are not routinely available in all laboratories included in the mycobacteria diagnostics network in Croatia, and identification of clinical isolates of mycobacteria at the species level is primarily based on the characteristics of the cultured bacteria, biochemical test results and, in some cases, identification using molecular methods. These test procedures are complex and laborious, and they are usually impeded by the slow growth of mycobacteria in culture (Bang et al. 2011). Clinical symptoms combined with acid-fast bacilli (AFB)-positive sputum smear or culture do not allow accurate differentiation between MTB, NTM infection or NTM environmental contamination. Since definitive identification of mycobacterial species can take several weeks, patients who have respiratory AFB-positive specimens present a public health dilemma. This is especially true for countries with a low incidence of NTM infections where clinicians rarely think about NTM as a disease causing pathogens. It is only recently that the first population-based study of NTM isolation frequency in Croatia has been published (Jankovic et al. 2013). This study showed that the number of NTM infections, in spite of an increase in the isolation frequency, was still remarkably low when compared to tuberculosis (TB). Thus, when a positive sputum smear microscopy result is reported to clinicians, they usually assume TB infection. This possible misdiagnosis can have serious consequences for both patients and their contacts.

The aim of this study was to describe the NTM isolation frequency at General Hospital Zabok, to confirm the presumed pseudo-outbreak of M. gordonae and to assess the impact of this outbreak on patients’ treatment.

A relative increase in the isolation of M. gordonae from pulmonary samples originating from General Hospital Zabok was noted by the National Reference Laboratory (NRL), at the Croatian National Institute of Public Health, Zagreb in 2009. An epidemiological survey was conducted to assess possible contamination of patients’ samples with NTM from water supplies within the hospital. Moreover, all incident cases of respiratory infection due to NTM reported from 2007 to 2012 at General Hospital Zabok were included for retrospective review. This study protocol was approved by the Ethics Committee for the Protection of Human Subjects, by the General Hospital Zabok, Croatia (Protocol 4716, December 20th 2012).

A total of 3903 pulmonary samples suspected of containing MTB or NTM were sent from General Hospital Zabok to the NRL, in the period 2007–2012. All samples were processed according to standard procedures as previously described. After discovering AFB in the sample, preliminary results were sent back to General Hospital Zabok, while the final result followed after identification of the species. The samples of cold tap water (n = 4) were collected in 2009 from multiple sites within the hospital and sent to the NRL for isolation and identification of mycobacteria. All NTM were identified by biochemical and molecular methods (GenoType^© CM/AS; Hain Lifescience, Nehren, Germany).

The records of all patients with NTM were reviewed. The following data were recorded from the patients’ charts: age, specimen collection source, clinical presentation, underlying diseases, radiological data and treatment.

From January 2007 to December 2012, a total of 150 incident cases with pulmonary NTM isolates were identified at the General Hospital Zabok (Figure 1).

Distribution of patients with isolated NTM species throughout selected years is shown in Table 1. Out of 150 patients included in this study, 91 (60.6%) were men with median age of 70.2 years and 59 (39.3%) were women with median age of 74.1 years. Most of the isolated NTM species were those of little clinical significance, with M. gordonae being the most frequently isolated species (n = 135; 90%). Furthermore, M. gordonae was isolated from all cold tap water samples collected in different parts of the hospital. The incidence of NTM isolation peaked in 2009 when 55 patients were identified, out of which 96.4% had M. gordonae isolates. Out of 135 patients with M. gordonae, 10 (7.4%) patients had two or more M. gordonae isolates, while 92.6% had a single isolate.

After reviewing the medical records, we found that none of the 150 cases met the criteria of American Thoracic Society (ATS) and Infectious Diseases Society of America (IDSA) for pulmonary NTM disease. The vast majority of patients had underlying pulmonary diseases with acute exacerbation of chronic obstructive pulmonary disease (AECOPD) being the most common one (Table 2). History of prior TB treatment was noted in 17 (11.3%) patients, while concomitant MTB infection was confirmed in only 6 (4%) patients. Still, in a significant portion of patients (n = 64; 42.6%), the anti-TB treatment was initiated. The majority of cases in which the treatment was initiated occurred in the period from 2008 throughout 2010 (Figure 2).

After observing increased M. gordonae isolation in patients’ samples from the General Hospital Zabok in 2009, a pseudo-outbreak in a newly constructed hospital building was suspected; namely, the building where the Department for Respiratory Diseases is situated which was opened in February 2008. Epidemiological Service (part of the Croatian National Institute of Public Health) performed an inspection and confirmed contamination of cold tap water with M. gordonae in multiple sites throughout the plumbing system of the new building. Municipal water supplies have already been recognized as a major niche for NTM, as well as the major reservoir for most nosocomial outbreaks and pseudo-outbreaks caused by these organisms (Wallace et al. 1998; Philips & Fordham von Reyn 2001). NTM are tolerant to a much wider pH and temperature range than most other bacterial pathogens detected in municipal water supplies. Their general tolerance to chlorine makes them potentially more difficult to eliminate (Carson et al. 1978; Hall-Stoodley et al. 1999; Le Dantec et al. 2002). The presence of some of the nonthermophilic NTM, such as M. kansasii, in hospital water systems has been reported worldwide (Bailey et al. 1970; McSwiggan & Collins 1974; Picardeau et al. 1997). Conversely, M. gordonae has not been studied extensively yet. Two recent studies from Turkey (Genc et al. 2013) and Korea (Shin et al. 2008) have shown the presence of M. gordonae in hospital tap water samples without consequences for staff or patients treated in those hospitals.

The tap water is potable in Croatia, hence it was presumed that drinking and/or gargling was the reason for the rise and pseudo-outbreak of M. gordonae infections in our patients.

Accordingly, guidelines for physicians, nurses and patients were issued, advising avoidance of drinking the tap water or even rinsing the mouth prior to giving the sputum sample. The potential weakness of the report was that the clonality of the isolates was not assessed but the implementation of measures was followed by a decreasing isolation trend of M. gordonae from respiratory samples with a steep decline achieved in 2012 and maintained up to date. Even though TB was later confirmed in only 6 (4%) patients, in more than 40% of all patients, treating physicians initiated the anti-TB treatment immediately after notification of positive AFB in the specimen. This unnecessary drug administration could lead to potentially severe adverse reactions. According to standard procedure in our country, patients were mostly transferred to a hospital which specialized in treating TB for prolonged hospitalization and/or isolation. Furthermore, it can be assumed that contact tracing was carried out, including unnecessary physical exams, chest X-rays, interferon gamma release assay (IGRA) testing and preventive treatment, especially in children under 5 years of age. Moreover, a false diagnosis of TB infection could have delayed the recognition of the real underlying condition and postponed specific treatment for a number of patients.

Some of the reasons for such a great number of unnecessarily treated patients are: the burden of TB in Croatia; existing respiratory symptoms in patients (mostly due to other underlying diseases); and the lack of knowledge about the NTM, even among respiratory physicians. Croatia is a country that still has a medium TB incidence (23/100,000 in 2008). At the time, the incidence of NTM isolation was still considered quite low and there were only occasional reports regarding the significance of these pathogens (Katalinic-Jankovic et al. 2007). The percentage of treated patients was highest in the period from 2008 to 2010, whereupon both the number of M. gordonae isolates as well as the fraction of treated patients started to decline due to the raised awareness about this pathogen among physicians.

We have confirmed a pseudo-outbreak of M. gordonae infection originating from a hospital's plumbing system and evaluated its impact on patients. This is a rare study of a large pseudo-outbreak of M. gordonae and its high influence on misdiagnosis of TB. The low awareness about the NTM in general, resulted in a significant number of inappropriately treated patients. TB treatment might have caused increased risk of drug-related adverse events but also unnecessary expense for the patient, hospital and the state. Moreover, patients’ contacts were probably subjected to physical exams, laboratory analyses, and in some cases, prescribed with TB preventive medication. This study points out the importance of a faster NTM identification in patient samples as well as the importance of education and good communication between laboratory personnel and physicians when evaluating the significance of the isolated NTM.

The authors whose names are listed certify that they have no affiliations with or involvement in any organization or entity with any financial or non-financial interest in the subject matter or materials discussed in this manuscript.