|Home | About | Journals | Submit | Contact Us | Français|
The isolation of nontuberculous mycobacteria (NTM) raises the question of their clinical significance, especially in an African setting. We found a high percentage of NTM isolated from various specimens, including ones that are normally sterile, among 213 patients in Zambia. Because tuberculosis can affect all parts of the body, we decided to include patients who had signs and symptoms in any part of the body for more than 2 weeks. Most patients had tractus respiratorius (80%) and tractus digestivus (10%) symptoms. During three consecutive days, sputum was collected and two separate sputum specimens were cultured for mycobacteria. Depending on the clinical picture, pleural effusion, ascites, abscess material, or enlarged lymph nodes were also cultured for mycobacteria. A specimen from one sterile body site was collected from 25 patients (60% human immunodeficiency virus [HIV] positive). NTM were isolated from 8 of these 25 specimens. Mycobacterium lentiflavum was isolated from four patients, and Mycobacterium goodii was isolated from one patient. In order to exclude the possibility of laboratory cross-contamination, a novel amplified fragment length polymorphism DNA typing method for M. lentiflavum was developed. Genetic variation was detected, rendering the likelihood of laboratory contamination unlikely. Clinically relevant infection due to NTM occurs in both HIV-positive and HIV-negative patients in Zambia, and their clinical impact seems to be underestimated. This is the first report of M. lentiflavum and M. goodii infections in Africa.
Mycobacterium tuberculosis and human immunodeficiency virus (HIV) coinfections are increasingly noted in third-world countries (8). In developed countries, it has been reported that HIV-infected patients are susceptible to infections by nontuberculous mycobacteria (NTM) (14). The isolation of NTM generally raises questions of their clinical significance, especially in an African setting, and assessment of this significance should be guided by the diagnostic criteria of the American Thoracic Society (1). According to these criteria, the diagnosis of NTM pulmonary disease must be based on solid clinical, radiographic, and bacteriologic factors. A culture of a tissue biopsy is one of the bacteriologic criteria and is considered positive if there is any growth from a usually sterile extrapulmonary site. Many infections occur without the identification of causal microorganisms, especially in African settings. An improved understanding of the etiology of these infections and which microorganisms play a role is of the utmost importance. Because only microscopic examination is usually available to confirm the diagnosis of tuberculosis in Africa, the involvement of acid-fast NTM in tuberculosis-like syndromes might result in the misdiagnosis of tuberculosis.
Mycobacterium avium complex bacteria, as well as other NTM, are widely distributed in the environment, and colonization of humans appears to be common (16). However, genuine infection by NTM has been reported almost exclusively from developed countries and not from developing countries, although the prevalence of HIV infection is generally much higher in the latter settings (7). Disseminated disease due to M. avium complex bacteria does occur in HIV-positive Africans but has been reported more frequently in the United States and Europe (2). This may be due to a lack of attention to this phenomenon in Africa, where possibilities for identification of mycobacteria are generally limited.
During a tuberculosis surveillance study prior to the present study, we regularly isolated NTM from various clinical specimens from 84 Zambian patients with pulmonary syndromes. A positive Mycobacterium culture was found in 51% of the included patients who were clinically suspect for tuberculosis, and 73% of these mycobacteria were NTM. The clinical significance of some of these NTM was questionable, and the aim of the present study was to resolve this enigma. To examine the clinical relevance of NTM in Zambian patients, eight patients from whom NTM were cultured from sterile body sites are described in detail.
A 67-year-old woman who was HIV negative was admitted with a painful abdominal swelling on the left side. Pain started a month earlier, after a herpes zoster infection. On physical examination, an enlarged axillary lymph node and enlarged liver and spleen, but no chest abnormalities, were found. Mycobacterium lentiflavum was cultured from the lymph node biopsy and from two sputum samples.
A 37-year-old female HIV-negative patient who was suffering for 2 months with a painfully swollen abdomen was admitted. She was also complaining about night sweats and a poor appetite. The chest X ray revealed an enlarged heart and pleural effusion at the left and right sites. The ultrasound showed an enlarged liver, ascites, a splenic abscess, and pericardial fluid. The splenic abscess and heart failure were treated with metronidazole and diuretics. After 2 weeks of treatment, the splenic abscess was reduced in size, but the liver seemed to have enlarged further. Treatment with rifampin, isoniazid, and pyrazinamide was started. Mycobacteria were cultured from the ascites and subsequently identified by 16S rRNA gene sequencing as an unknown mycobacterium phylogenetically related to Mycobacterium gilvum (Fig. (Fig.1).1). One month later, the patient was readmitted because of heart failure.
The third case concerns a 38-year-old HIV-negative woman. She complained about chills and abdominal distension that started 7 months earlier. She had already been treated with diuretics without improvement. The white blood cell count was 6.200/mm3 with 18% eosinophils. No abnormalities were seen on the chest X ray. The patient was treated with antibiotics, anthelminthics, diuretics, and draining of the ascites. M. lentiflavum was cultured from the ascites.
A 21-year-old female HIV-positive patient presented herself with a distended abdomen and edema of the ankles that started 3 weeks earlier. Furthermore, she was having a productive cough and pain passing urine for 4 months. On physical examination, an icteric patient was seen with enlarged cervical, supraclavicular, and submandibular lymph nodes and abdominal ascites. The chest X ray did not show any abnormality. M. tuberculosis was isolated from sputum and urine, and Mycobacterium fortuitum was isolated from the ascites. The patient died 4 days later.
The patient was a 36-year-old HIV-positive woman who suffered from recurrent abscesses for 2 years. The patient presented with fever and multiple cutaneous abscesses on both legs and in the neck area. M. lentiflavum was isolated from the pus and from one of two sputum samples.
A 35-year-old HIV-negative woman was admitted previously because of cough and left pleural effusion. Symptomatic improvement was achieved with amoxillin. She subsequently presented with a nonproductive cough and left chest pain. The radiographic investigation of the chest revealed a pleural effusion and alveolar consolidation on the left side. The culture of the pleural effusion proved the presence of M. lentiflavum. The patient improved upon treatment with rifampin, isoniazid, pyrazinamide, and ethambutol.
An HIV-negative 66-year-old man presented with a productive cough, chest pain, and diarrhea for 3 days. The patient's medical history showed that he had pneumonia 1 year ago which improved with amoxicillin treatment. Pleural effusion, a left alveolar infiltration, and peribronchial pathology were seen on the chest X ray. Hookworms were found in the stool. The culture of the pleural effusion indicated the presence of M. goodii.
The patient was a 26-year-old HIV-positive woman who complained of a productive cough with hemoptysis and chest pain of 1-month duration. She had been treated for tuberculosis twice in the past. Her chest X ray showed pleural effusion and possible interstitial pathology. Mycobacteria were cultured from the pleural effusion and could not be identified to the species level (Fig. (Fig.1).1). The highest phylogenetic similarity, based on rRNA gene sequences, was demonstrated with M. parafinicum and M. scrofulaceum.
The study population consisted of 213 Zambian patients, 99 female and 114 male adults over 15 years of age. All patients who were admitted with chronic complaints to St. Francis Hospital in Katete (Eastern province), Yeta District Hospital in Sesheke (Southern province), or Our Lady's Hospital in Chilonga (Northern province) during the period of March to August 2001 were included in the study. Because tuberculosis can affect all parts of the body, we decided to include patients who had signs and symptoms in any part of the body for more than 2 weeks. Most of the included patients had tractus respiratorius (80%) and tractus digestivus (10%) symptoms. The other 10% of the patients had skin infections/abscesses, lymphadenopathy, or tractus urogenitalis or central nervous system complaints.
The medical history of the patients was retrieved, and detailed physical examination was performed. During three consecutive days, sputum was collected from patients with a productive cough. The first two sputum specimens were cultured for mycobacteria, and the third one was stored at −20°C. Depending on the clinical symptoms, samples of pleural effusion, ascites, abscess material, or enlarged lymph nodes were obtained in a sterile manner. All specimens were cultured for mycobacteria. Materials needed for collection and cultivation were imported from The Netherlands, and an experienced Dutch technician performed the work in the Zambian laboratory. All possible efforts were made to prevent laboratory cross-contamination, including the collection of a specimen in a container that was not used before and that was imported from The Netherlands. The stock of decontamination fluid was sterilized twice a week, and each day a fresh, sterile tube or bottle was opened, the samples in the laboratory were processed one by one in the absence of other samples, and negative control cultures were included in each batch of samples. Microscopic slides were also prepared one at a time. All work associated with this project was conducted by the same experienced laboratory technician. All culture procedures were performed in a new class I biosafety cabinet.
A specimen from a sterile body site was collected from 25 patients. Ten of these specimens were cultured without decontamination, and 15 specimens were divided into two equal parts: one half was decontaminated with N-acetyl-l-cysteine (NALC)-NaOH and the other half was decontaminated using 6% sulfuric acid to compare decontamination procedures for the detection of mycobacteria (4). The specimens from a sterile body site were first cultured without decontamination, but to avoid any confusion in the handling of the samples in the laboratory, all collected specimens were decontaminated after the first weeks of inclusion. Serological testing for HIV was performed using a qualitative immunoassay (Abbott Determine HIV-1/2) and the Vidas HIV DUO assay (bioMérieux, Marcy l'Etoile, France). Chest X rays were taken and evaluated in The Netherlands in a blinded manner regarding the origin of the pictures and any additional clinical information. The specimens were cultured in Mycobacteria Growth Indicator tubes (Becton Dickinson Microbiology Systems, Cockeysville, Md.) according to the instructions of the manufacturer and the guidelines described previously by Master (17). Mycobacterial isolates were identified by the Accuprobe culture confirmation test for the M. tuberculosis complex (Accuprobe, bioMérieux, Marcy l'Etoile, France) or by 16S rRNA gene sequencing (15).
Molecular analysis of 12 M. lentiflavum isolates was performed by high-throughput amplified fragment length polymorphism (AFLP) (htAFLP) analysis according to a method described previously by van den Braak et al. (27) and Melles et al. (18). In short, DNA was isolated from pure cultures (28), and htAFLP was performed according to standard procedures. One enzyme combination was used to digest the mycobacterial DNA: Csp6I/NIaIII. The general primers for preamplification were 5′-GTAGACTGCGTACCTAC-3′ for Csp6I and 5′-GAGTCCTGAGCATG-3′ for NIaIII and eight different sets of primers (+G/AA, +G/AG, +G/AT, +G/CA, +C/AA, +C/AG, +C/AT, and +C/CA) suited for selective amplification of subsets of restriction fragments. Apart from the 10 Zambian M. lentiflavum isolates, we included two independent isolates from The Netherlands and M. tuberculosis control strain H37Rv (6). Nine isolates were tested in duplicate to define the test reproducibility. Banding patterns of coded samples were scored visually by an independent observer. DNA fragments from 150 to 550 bp in length were evaluated.
The study was approved by the research ethical committee of the University of Zambia and by the Central Board of Health and the Ministry of Health in Zambia. Informed written consent to participate in the study was obtained from all patients. At the time of the study, the estimated Zambian HIV seroprevalence was 20%, and the case notification rate for tuberculosis was about 500/100,000 (20).
Overall, out of 213 patients, 69% were HIV positive. M. tuberculosis was isolated from 44 patients, 77% of whom were HIV positive, and NTM were isolated from 90 patients. Without knowing the results of the study, treatment for tuberculosis was given to 73 patients, 10 of whom were treated for extrapulmonary tuberculosis.
A specimen from a sterile body site was collected from 25 patients (60% HIV positive). Eight of these 25 specimens were positive for NTM (Table (Table1),1), while in another five specimens, the presence of M. tuberculosis was demonstrated. Twelve specimens were culture negative. The Ziehl-Neelsen smears were all negative except for one (case 1). In contrast, 21 of the 44 patients with a positive culture for M. tuberculosis had a positive Ziehl-Neelsen smear of their sputum. The restriction fragment length polymorphism patterns of the isolated M. tuberculosis strains were all different (data not shown), essentially excluding the possibility of laboratory cross-contamination for these strains.
Fifty-five M. lentiflavum isolates were cultured from 149 specimens obtained from 38 of the 70 patients admitted to St. Francis Hospital in Katete. This mycobacterial species has so far been rarely identified from clinical materials (see Discussion). In order to exclude the possibility that laboratory cross-contamination was the source of these positive cultures, a random set of 12 of these isolates was subjected to htAFLP analysis. The M. lentiflavum strains from cases 1 and 6 were included in this random set (strains 9, 11, and 13 in Table Table2).2). Evaluation of the DNA fragments resulted in 256 scorable DNA markers. Reproducibility of the variable markers as deduced from the duplo analyses appeared to be 100% (results not shown), and two strains from two specimens from a single patient were indistinguishable.
Although most of the markers were not variable among the M. lentiflavum isolates, a part of the AFLP restriction fragments revealed a considerable level of variation. Any variable combination of markers received a marker pattern number (I to VII) (Table (Table2).2). Consequently, a marker pattern is corroborated by various marker fragments, thereby constituting a very reliable genotype. M. tuberculosis appeared to yield completely distinct AFLP patterns (see Fig. Fig.22 for an example of the experimental output). Among the 12 M. lentiflavum strains, six different types were documented. The types obtained for the two Dutch strains (types B and F) were clearly different from those generated for the Zambian isolates (types A, C, D, and E) (Table (Table22).
Type D strains were isolated from six patients (patients 019, 001, 021, 033, 045, and 062), and the sputum and biopsy isolates from patient 033 yielded the same type. Although other strains from patients visiting the same hospital were genetically distinct, ruling out full-blown laboratory contamination, we cannot exclude such an event for the six patients infected by a type D strain. On the other hand, this region of Zambia might be an area where type D is endemic.
NTM are ubiquitously present in the environment and can therefore be associated with either colonization, serious infection, or pseudo-outbreaks with a wide variety of presentations (10, 22, 23, 25). NTM should be considered in all cases of nosocomial infection, and careful surveillance must be applied to identify possible outbreaks. Presumed nosocomial outbreaks of NTM can be investigated by molecular typing of mycobacterial isolates. In cases where identical DNA fingerprints are found, laboratory cross-contamination cannot always be excluded. However, in our study, all possible measures were taken to avoid any laboratory cross-contamination. The NTM were cultured from patients with compatible clinical syndromes and were isolated from specimens that do not normally contain microorganisms. Molecular analysis revealed that the mycobacterial isolates represented several evolutionary lineages and more subtle strain variations. Therefore, in our study, laboratory cross-contamination is considered highly unlikely.
On basis of the diagnostic criteria of the American Thoracic Society, the medical relevance of the NTM isolations reported here appears certain for at least half of the cases (cases 1, 6, 7, and 8). Their signs and symptoms were compatible with pathology found on physical examination or chest X ray, and the NTM were isolated from pleural effusions and a lymph node. In the other four cases (cases 2, 3, 4, and 5), NTM were isolated from ascites and a cutaneous abscess, and in these cases, it is not completely clear whether these mycobacteria were the most important causes of the pathologies. These last patients had other conditions that could have explained their nonspecific symptoms and the physical findings.
In none of the eight patients diagnosed with NTM infection of normally sterile body sites was the diagnosis made before hospital discharge or death. No specific therapy was given for the mycobacteria isolated, except for patient 2, who received empirical treatment with rifampin, isoniazid, and pyrazinamide before the results of the culture became available. In five of the six specimens of sterile body sites, which were decontaminated with sulfuric acid and NALC-NaOH, NTM were cultured only after decontamination with sulfuric acid. However, the commonly used decontamination method is the use of NALC-NaOH. It has been shown that the decontamination method used affects the recovery of NTM (4).
Only a few cases of human disease caused by M. lentiflavum have been reported previously. The first case, concerning an 85-year-old woman with spondylodiscitis that improved upon antituberculous treatment, was reported in 1996 (24). Four children with lymphadenitis due to M. lentiflavum were described between 1997 and 2002 (5, 12, 26), as were cases of disseminated infection in an HIV-infected patient (21) and in a patient undergoing steroid therapy (13). Finally, M. lentiflavum was isolated from a patient with a chronic pulmonary disease. M. goodii was first described in 1999 and was isolated from patients with traumatic osteomyelitis following iatrogenic infections and from patients with respiratory infections (3). Recently, a patient with bursitis due to an M. goodii infection was described (9). All of these cases occurred in Europe and the United States. To our knowledge, here we describe the first cases of M. lentiflavum and M. goodii infection in African patients. In fact, no clinically relevant NTM isolations in sub-Saharan Africa were reported before the 1990s (11, 19).
Our htAFLP-mediated strain-typing method further corroborates the clinical relevance of the isolates; genetic diversity was observed among the strains. This degree of strain diversity shows that (i) the typing procedure is as adequate as can be expected when small numbers of strains are used and (ii) it is not the case that laboratory cross-contamination is the source of all of the M. lentiflavum isolates. However, multiple isolates of genotype D (Table (Table2)2) were identified. Whether this is due to laboratory contamination or a genuine NTM outbreak still needs to be resolved. It has to be emphasized that the respective type D strains were derived from patients nursed in different wards. The elevated incidence of the type D M. lentiflavum isolates (7/10 [70%] of the Zambian isolates) suggests that local dissemination of a certain type occurred or that M. lentiflavum as a species is quite clonal. The latter hypothesis is not in contradiction with the well-conserved features of the htAFLP fingerprints, even when isolates from The Netherlands and Zambia are compared (Fig. (Fig.2).2). In addition, both restriction fragment length polymorphism analysis and randomly amplified polymorphic DNA analysis performed for a subset of strains also revealed additional DNA polymorphisms among M. lentiflavum isolates (results not shown).
We conclude from our study that clinically relevant infection due to NTM seems to occur in HIV-positive as well as in HIV-negative patients in Zambia. The role of NTM in human disease in Africa may well be underestimated and should be examined in more detail and on a larger scale. Information is urgently needed with regard to the proper diagnostic procedures and the possibilities for adequate treatment of NTM-induced disease.
Financial support for this work was provided by a grant from the microbiologists of Rotterdam working in the Medical Centre Rijnmond-South and The Netherlands Society of Tropical Medicine and International Health.
We gratefully acknowledge M. van Prehn (MCRZ) for her excellent laboratory assistance and M. Bakkers for sampling a subgroup of the patients. We are grateful to the staff and patients of St. Francis Hospital in Katete, Yeta District Hospital in Sesheke, and Our Lady's Hospital in Chilonga for their teamwork and cooperation. We also acknowledge M. M. J. Petit-Vijgen for creating an excellent database. We thank A. van der Brandt, T. van der Laan, M. Dessens, and P. de Haas (RIVM) for their excellent laboratory assistance. htAFLP was performed at Pathofinder (Canisius Wilhelmina Hospital, Nijmegen, The Netherlands) under the supervision of M. Reijans and G. Simons.