The intestinal carriage of nontuberculous mycobacteria (NTM) is associated with disease in both immunocompetent as well as immunocompromised individuals [1]. Nontuberculous mycobacteria (NTM) are a group of Mycobacterium species different from those belonging to the Mycobacterium tuberculosis complex (MTB) and M. leprae. There are currently over 150 NTM species recognised [2]. The organisms, although frequently dismissed as mere contaminants, have been known to cause various types of illnesses affecting the skin, lungs, intestines, the joints as well as the blood stream. There has been an increased incidence of NTM since the advent of HIV, partly due to HIV but also because of increased awareness [3–5].

Many NTM are free-living organisms which have been isolated in a wide range of environments, including water, soil, dust and aerosols. In general, disease due to NTM is more common in individuals with suppressed local or systemic immunity than those with normal immunity. However, some studies have found an increased incidence of NTM disease in the elderly [6]. Some of the conditions predisposing to systemic, respiratory and abdominal NTM disease include HIV, emphysema, pneumoconiosis, cystic fibrosis, previous gastrectomy, continuous ambulatory dialysis and chronic alcoholism [7–9]. In the gastrointestinal tract, both M. tuberculosis and NTM can cause abdominal pain, ascites, chronic diarrhoea and other clinical manifestations [10].

Gastrointestinal (GI) disease is characterised by invasion of Peyer’s patches and mesenteric lymph nodes of the small intestine, leading to foamy histiocytes and mycobacteria-laden macrophages within the lamina propria of the intestinal mucosa. Symptoms include diffuse abdominal pain, weight loss, fever and/or diarrhoea. Findings on endoscopy include multiple raised nodules or normal-appearing mucosa. Other findings may include ulceration, erythema, oedema, friability, reduced mucosal vascularity, stricture, and aphthous erosions [11, 12]. The prevalence of NTM in intestinal specimens such as stool, intestinal lavage and colon biopsies is unknown.

Little is known about the role of NTM in the aetiology of abdominal disease in Zambian patients. Further it has been suggested that the burden of nontuberculous mycobacteria in Zambia, some of which may present with abdominal symptoms, may be underestimated [13]. This article presents the results of a descriptive study of carriage rates of NTM as well as factors associated with their carriage in patients booked for endoscopy at the University Teaching Hospital (UTH) in Lusaka, Zambia.

In this study, we found evidence of NTM among patients booked for colonoscopy and flexible sigmoidoscopy at UTH. This observed prevalence is higher than a previous report in the stools of patients with chronic diarrhoea [18] in the same hospital. The reasons for this are unclear and were beyond the scope of this study, but the development of MGIT detection systems in the interval between these studies may partly explain the higher apparent prevalence. However, much higher prevalences have been detected from other Zambian patient populations using sputum samples, where the prevalence of NTM was as high as 56% [13, 19]. Given the known biological characteristics of these micro-organisms, their presence in this population also suggests an environmental contamination.

This being a hospital based study, there are inherent external validity reservations for such findings, as this study was limited to a highly selected population scheduled for endoscopy. There may be arguments supporting the thinking that external validity can only be guaranteed if such a study should have been done in the general population in Lusaka. We argue that the great majority of our participants were generally healthy, apart from having abdominal symptoms for which they sought endoscopy services. It would not be possible to justify colonoscopy in unselected patients in the general population, merely for the purpose of obtaining biopsies. Notwithstanding these limitations, we think that these findings may even be an under-estimate of the burden of NTM. This is due to the use of the culture methods without special incubation temperatures or additional nutritional supplements to culture more fastidious NTM [20] from both the clinical and water samples. Further, the use of N-acetyl l-cysteine—6% NaOH (NALC–NAOH) for decontamination may have been overly harsh for some samples with low levels of organisms [21]. Our results were not affected by non-participation due to refusal to give consent as this remained below 5%. We consider it likely that this population has a measurable carriage rate of NTM in the intestine.

Finding that the carriage of NTM was not associated with the gender of the respondent, age, employment status, or presenting symptoms was not surprising given the ubiquitous nature of these organisms. These results illustrate the abundant nature of NTM, which are found in soil, dust and water systems [22]. This might explain the general exposure that exists irrespective of the age, sex and other patient characteristics. However, previous studies elsewhere have shown a relationship between isolation of NTM and both gender and age of the patient [6, 23] and type of occupation [24]. Tests for NTM done on the hospital water systems during the period of the study were found to be negative. Given that the water samples and the clinical samples were processed in the same way, our findings suggest that the NTM were not due to contamination.

It has been suggested that NTM, such as the Mycobacterium avium complex, may be responsible for some non-specific abdominal symptoms such as anorexia, weight loss, dysphagia, odynophagia and abdominal pain [11]. While abdominal pain was the most common presenting symptom in our patients (72.0%), there was no relationship between the carriage of NTM and presenting symptoms such as abdominal pain, weight loss, vomiting, nausea, dizziness and anorexia. We did not isolate Mycobacterium avium species in our patients.

We found a higher diagnostic yield of NTM from colonic biopsies compared to stool and lavage samples, although this was not statistically significant. Mycobacterium avium is known to be the most frequently isolated NTM from the gastrointestinal tract among immunocompromised patients [25]. However, this was not the case in this study because Mycobacterium gordonae was the most frequently isolated species in our patients, followed by M. Kansasii and M. genavense. All the isolated species were potentially pathogenic [26]. Reports of increasing incidence of nontuberculous mycobacteria point to the need for further attention to these organisms [3, 27].

It is well known that immune-compromised patients are more likely to suffer from disease caused by NTM than the general population. Carriage rates of NTM were not associated with levels of haemoglobin and HIV status in our patients. The HIV prevalence was 19.3%, and the patients were generally healthy with very few who were chronically ill or had other debilitating illnesses. Compared to the general Zambian adult population, whose HIV prevalence according to the Zambia Demographic and Health Survey (ZDHS) [28] was 13.3% in 2014, the HIV prevalence in our population was higher.

Some of our patients had M. tuberculosis isolates. However, only one patient (who was not suspected to have TB at the time) showed endoscopic findings that were consistent with abdominal tuberculosis in the ileo-caecal area. Culture of the caecal biopsy was positive for M. tuberculosis and he was subsequently treated for abdominal TB. Other patients were referred to the appropriate TB services for further evaluation.

Our results have demonstrated presence of NTM in this population, suggesting an environmental contamination of the gut by potentially pathogenic nontuberculous mycobacteria that was not associated with any symptoms or demographic status. The extent to which these organisms are responsible for some of the morbidity in our patients remains to be unravelled. Given that this may be an indicator or a bigger burden of NTM in this population, there is thus need to further explore this burden and the contribution it could have on the disease in the intestine as well as critically examining potential factors that might be important predictors. We found that descending colon biopsies gave the highest yield for detection of NTM.