Aliyu Gambo Gumel Permanent E-Mail Address

Pattern of Mycobacterial Infections and their Associations with HIV among Laboratory Confirmed Cases of Pulmonary Tuberculosis in Nigeria

Item Type dissertation

Authors Aliyu, Gambo Gumel

Publication Date 2012

Abstract Background: Nigeria has the fourth highest tuberculosis (TB) burden worldwide. In this study, mycobacterial agents from clinically symptomatic TB patients, regardless of HIV co-infection were isolated and characterized, and resistance to isoniazid an...

Keywords non-tuberculous; HIV Infections; Mycobacterium; Nigeria; Tuberculosis, Pulmonary

Download date 02/10/2021 08:51:47

Link to Item http://hdl.handle.net/10713/2224 CURRICULUM VITAE

Name: Aliyu Gambo Gumel

Permanent e-mail Address: [email protected]

Degree and Date to be Conferred: Ph.D., 2012

Collegiate Institutions Attended:

University of Maryland, Baltimore, MD Ph.D., Epidemiology, 2012 M.S., Clinical Research, 2008

Ahmadu Bello University, Zaria, Nigeria MB.BS, Medicine 1995

Professional Publications:

1. Aliyu G., Mohammad M., Saidu A., Mondal P., Charurat M., Abimiku A., Nasidi A., Blattner W. HIV Infection Awareness and Willingness to Participate in Future HIV Vaccine Trials across Different Risk Groups in Abuja, Nigeria. AIDS Care; July, 2010 (1-8)

2. Aliyu G ., Informed Consent of Very Sick Subjects in Nigeria. Account Res. 2011 Jul-Aug; 18:4, 289-296

3. Charurat M, Nasidi A, Delaney K, Saidu A, Croxton T, Mondal P, Aliyu GG , Constantine N, Abimiku A, Carr JK, Vertefeuille J, Blattner W. Characterization of acute HIV-1 infection in high-risk Nigerian populations, J Infect Dis. 2012 Apr; 205(8):1239-47. Epub 2012 Feb 21

4. Sani-Gwarzo Nasiru, Gambo G. Aliyu , Alex Gasasira , Muktar H. Aliyu, , Mahmud Zubair, Sunusi U. Mandawari, Hassana Waziri, Abdulsalami Nasidi, Samer S. El- Kamary. Breaking Community Barriers to Polio Vaccination in Northern Nigeria: The Impact of a Grass Roots Mobilization Campaign (Majigi). Accepted for publication in the journal of pathogens and global health on April 30, 2012

5. Charurat, M, Delaney, K, Ahmed, S, Villalba-Diebold, P, Aliyu, G., Constantine, NT, Onoja, A, Vertefeuille, J, Blattner, W, Nasidi, A. HIV testing and access to care needs of most at risk populations in Nigeria. Accepted for publication in the AIDS Care journal on April 16, 2012

Abstracts :

1. Aliyu G , Sill A, Olanrewaju A, Delany K, Croxton T, Charurat M, , Vertefeuille J, Sheppard J, Nasidi A, Abimiku AG, Constantine N, Blattner W Estimation of HIV-1 incidence in Nigeria: a prerequisite for HIV vaccine trials. Poster abstract presented at the 3 rd AAVP forum Yaoundé Cameroun, October, 2005.

2. Aliyu G , Mukhtar M, Ahmed S, Prosanta M, Charurat M, Abimiku A, Nasidi A, Blattner W. HIV disease awareness and willingness to participate in HIV vaccine trials differ across subpopulations at risk of HIV in Abuja Nigeria. Poster abstract presented at the 4 th AAVP forum in Abuja-Nigeria, November, 2007.

3. Delaney K, Charurat M, Constantine N, Owen M, Keating S, Candal D, Ethridge, Saidu A, Saleh A, Croxton T, Villalba-Diebold, Aliyu G, Vertefeuille J, Sill A, Busch M, Nasidi A, Blattner W. Evaluation of five HIV incidence assays using Nigerian seroconversion specimens. Implications for suiveillance programs in Nigeria. Presented at 2009 international AIDS society July 17-19, 2009. Capetown, South Africa.

4. Delaney KP, Charurat M, Villalba-Diebold P, Ahmed S, Aliyu G, Abimiku A, Vertefeuille J, Nasidi A, Blattner WA. HIV Counseling and Testing (HCT) of Nigerian Sex workers: increased knowledge, increased condom use, no change in risk of infection. Presented at the 2010 International AIDS Meeting in Vienna, Austria.

Membership in professional societies

2010 to present American College of Epidemiology 1995 to present Medical and Dental Council of Nigeria

Grant award W.H.O grant on HIV vaccine preparedness studies in Africa Funding: $15,000 11/20/2005 to 08/20/2006 Role: Principal Investigator

ABSTRACT

Title of Dissertation: Pattern of Mycobacterial Infections and their Associations with HIV among Laboratory Confirmed Cases of Pulmonary Tuberculosis in Nigeria

Gambo Gumel Aliyu, Doctor of Philosophy, 2012

Dissertation Directed by: Samer S. El-Kamary, M.D., M.P.H Assistant Professor Department of Epidemiology and Public Health

Background: Nigeria has the fourth highest tuberculosis (TB) burden worldwide. In this study, mycobacterial agents from clinically symptomatic TB patients, regardless of HIV co-infection were isolated and characterized, and resistance to isoniazid and rifampicin determined.

Methods : Suspected TB cases were recruited from two TB clinics into a cross-sectional study. All patients were screened for HIV, and their sputum samples were screened for

Mycobacteria using an algorithm that included smear microscopy, liquid broth and solid

media culture, TB-antigen detection assay, and molecular probe assays, to determine the

type of Mycobacterium and pattern of resistance to isoniazid and rifampicin.

Results : Of 1,603 patients screened, 466 (29%) had liquid broth culture-positive

pulmonary Mycobacterial infection. Of these, 444 (95%) had mycobacterial infections

and 22 (5%) were false-positive non-mycobacterial strains. Of the 444 cases, 375 (80%)

were infected with Mycobacterial tuberculosis (MTB) complex (354 M. tuberculosis, 1

M. bovis and 20 M. africanum) and 69 (15%) were Non-tuberculous mycobacteria

(NTM ). HIV co-infection was detected in 101 (27%) of the MTB complex and 26 (38%) of the NTM cases respectively .

Twenty-three (6.1%) of the MTB complex cases had organisms resistant to

isoniazid (3.5%), rifampicin (1.3%) or both, i.e. multi-drug resistant TB (MDR-TB,

1.3%), by molecular analysis, and of those, 8 (35%) had a prior history of TB treatment.

Currently available molecular assays are incapable of detecting drug-resistant NTM

strains. After controlling for prior treatment, cases with any resistance (i.e. at least to one

drug) were more likely to be co-infected with HIV compared to cases without any

resistance (OR=3.6, 95%CI=1.5-8.8; p=0.0039 ). Compared to M. tuberculosis , the NTM

cases were more likely to be HIV co-infected more likely to present with clinical

symptoms during the intense Harmattan dust storms from December to February

(OR=2.5, 95%CI=1.4-4.5; p=0.0034, and OR=2.2, 95%CI=1.2-3.8; p<0.0067 and respectively), and less likely to be detected by the routine sputum smear test (OR=0.05,

95%CI=0.02-0.13; p<.0001 ).

Conclusions: The high frequency of smear-negative NTM cases with HIV co-infection identified during the period of Harmattan dust storm presents a novel public health challenge. Introduction of molecular detection assays to identify smear-negative NTM

and MDR-TB is a high priority.

Pattern of Mycobacterial Infections and their Associations with HIV among Laboratory

Confirmed Cases of Pulmonary Tuberculosis in Nigeria

Gambo Gumel Aliyu

Dissertation submitted to the Faculty of the Graduate School of the

University of Maryland, Baltimore in partial fulfillment of the requirements for the

degree of Doctor of Philosophy

2012

ACKNOWLEDGEMENTS

I would like to extend sincere gratitude to my dissertation committee chair, mentor and research advisor, Dr. Samer El-Kamary for his mentorship, advise, dedication and due diligence in steering the dissertation work to a successful conclusion. I would like to equally extend my heartfelt gratitude to my boss, sponsor, mentor and committee member, Dr. William Blattner for the opportunities, faith and the confidence he has reposed in me without which this work would not have been possible. To the rest of my committee members: Dr. Alash’le Abimiku, Dr. Laura Hungerford, Dr. Kathleen Tracy and Dr. Clayton Brown. I remain sincerely grateful for your commitment, input and resourcefulness in the conduct of this dissertation.

I would like to thank the Institute of Human Virology Fogarty fellowship program for sponsoring my PhD program; mentoring and support from past and present advisors of the IHV Fogarty grant including Dr. Abdulsalami Nasidi, Dr. Manhattan Charurat and

Joyce Johnson. I also wish to thank the faculty (especially Dr. Patricia Langenberg) and staff of the Department of Epidemiology and Public Health for the support and knowledge I have acquired. The tireless assistance of the academic coordinator, Danielle

Fitzpatrick is greatly appreciated.

Finally, I thank my family and friends for their tremendous support and encouragement at the critical hours of need.

iii

TABLE OF CONTENTS I. INTRODUCTION ...... 1 A. Tuberculosis Epidemiology ...... 1 1. Mycobacterium Tuberculosis (MTB) Complex ...... 2 2. Non-tuberculous Mycobacterium ...... 4 B. HIV-Mycobacterium TB Co-infection ...... 5 C. HIV-Mycobacterium bovis Co-infection ...... 6 D. Tuberculosis Treatment ...... 6 E. Multi-Drug Resistant TB (MDR-TB) ...... 7 F. Multi-Drug Resistant TB in HIV ...... 7 G. Burden of Environment Related Mycobacterial Infections in Nigeria ...... 8 H. Research Questions, Hypotheses and Specific Aims ...... 10 I. Summary of Research Significance ...... 11 II. RESEARCH DESIGN AND METHODS ...... 13 A. Study Sites ...... 13 B. Target Population and Study Design ...... 14 C. Data Collection ...... 15 D. Laboratory testing ...... 20 E. Detailed Description of Laboratory Methods ...... 21 1. HIV status determination ...... 21 2. Sputum sample collection ...... 21 3. Sputum smear microscopy test ...... 22 4. Liquid broth mycobacterial culture ...... 23 5. Interpretation of culture results and further characterization ...... 23 6. Detection of resistance to isoniazid and rifampicin ...... 26 7. Molecular line probe assays technique ...... 28 F. Data Management and Analyses ...... 30 1. Specific Aim I ...... 30 2. Specific Aim II ...... 31 3. Specific Aim III ...... 35 III. RESULTS: DISSERTATION MANUSCRIPT I ..... Error! Bookmark not defined. 39

A. Abstract ...... 39 B. Introduction ...... 41 C. Methods ...... 43 D. Results ...... 50 E. Discussion ...... 54 IV. RESULTS: DISSERTATION MANUSCRIPT II .... Error! Bookmark not defined. 65 A. Abstract ...... 665 B. Introduction ...... 67 C. Methods ...... 69 D. Results ...... 744 E. Discussion ...... 777 V. DISCUSSION OF DISSERTATION FINDINGS ...... 833 A. Prevalence of pulmonary mycobacterium ...... 833 B. Non-tuberculosis Mycobacterial Infection ...... 855 C. Association with HIV Infection ...... 866 D. Resistance to Isoniazid and Rifampicin ...... 866 E. Limitations ...... 877 VI. APPENDIX A...... 90 VII. REFRENCES...... 97

LIST OF TABLES

Table 1. Variables in the study data and their definitions and categories ...... 16

Table 2. Grading and interpretation of smear microscopy test ...... 22

Table 3. Sample size estimation for different effect sizes and ratios of HIV negative to HIV positive for a fixed power (.80) and a two-sided error rate ( α =0.05) ...... 35

Table 4. Measured demographics and laboratory characteristics stratified by site ...... 58

Table 5. Unadjusted analysis comparing the groups with detectable M. tuberculosis, M. Africanum, NTM and NMB to the group without any detectable isolates (reference group) with respect to HIV infection and measured covariates ...... 59

Table 6. Multiple logistic regressions analyses comparing the groups with detectable M. tuberculosis, M. Africanum, NTM and UNS to the group without any detectable isolates (reference group) with respect to HIV infection and measured covariates ...... 61

Table 7. Unadjusted comparison between NTM and MTB complex infections with respect to HIV Infection and measured covariates ...... 62

Table 8. Adjusted analysis comparing NTM and M. tuberculosis among cases of pulmonary mycobacterial infection ...... 63

Table 9. Demographic and covariates of the pulmonary TB cases by site ...... 79

Table 10. Unadjusted analysis comparing cases with resistance to isoniazid alone, rifampicin alone, both (MDR-TB) and those with any resistance to those without any drug resistance ...... 80

Table 11. Adjusted associations between HIV co-infection and prior TB treatment with resistance type versus no resistance ...... 81

LIST OF FIGURES

Figure1. Flow diagram illustrating steps used to detect and characterize mycobacterial isolates...... 25 Figure 2. Mycobacterial tuberculosis complex strains banding pattern in Genotype MTBC assay...... 26 Figure 3. Genotype MTBDR plus strip reaction zones for detection of mutations to rifampicin and isoniazid ...... 27 Figure 4. Pattern of occurrence of tuberculosis and non-tuberculous mycobacterial infections over a period of 12 calendar months in Nigeria: solid lines indicate monthly proportions of NTM and MTB among all subjects screened while dotted lines are the 95% confidence interval of the proportion ...... 64

vii

I. INTRODUCTION

Nigeria is located in West Africa at the Gulf of Guinea. It has a total land mass of

923,768 km 2 and a population of 140 million (Nigerian National Population Census

2006), and has the fourth highest tuberculosis (TB) burden worldwide. In 2007, there

were 460,000 new tuberculosis cases making it the highest in Africa, and second only to

the republic of South Africa in total number of HIV co-infected TB cases . In addition to

TB, Africa is the epicentre of the HIV pandemic with 80% of the world’s HIV positive

TB cases with Nigeria accounting for 10% of the global total [1].

A. Tuberculosis Epidemiology

There are two broad groups of mycobacterial diseases: tuberculous and non-

tuberculous mycobacterial infections. Tuberculous mycobacterial infections which cause

tuberculosis disease (TB) in humans and animals are the most important preventable

communicable diseases with a significant morbidity and mortality. While TB is

completely curable, it has persisted for over two centuries since its discovery in 1882.

Worldwide, 1 in 3 persons is infected with TB and the disease kills more than three

million people annually.[2, 3] About 80% of global TB infections are concentrated in 22

nations. The highest ranking five are: India, China, Indonesia, Nigeria and South

Africa.[1] The disease is largely caused by Mycobacterium tuberculosis and is

propagated by airborne transmission when an infected individual coughs, sneezes, speaks

or spits. It was reported that in a single sneeze, up to 40,000 droplets may be released and

inhaling as low as ten bacteria may eventually cause an infection.[4-6] The bacterium is

an aerobic organism that primarily infects the respiratory system. It is a slow-growing

organism (15-20 hours for each division), and its unusual waxy coat makes it resistant to

digestion by macrophages as well as decolourization with acid or alcohol.

The majority of infections in humans are latent as only a small percentage

progress to active disease, which has devastating consequences if left untreated. Disease

development is dependent on the interplay between the host cellular immunity and

genetics, and the dose and virulence of the transmitted bacteria.[7-9] A person with active

TB typically presents with symptoms including prolonged and productive cough which

often consists of blood, fever, night sweats, weight loss and fatigue. Individuals at high

risk for TB infection include those who are in contact with active TB patients, those with

diabetes mellitus, silicosis, chronic renal failure and malnutrition. The elderly, the poor,

migrants, refugees and some travellers are also at risk. Other factors observed to facilitate

transmission are cigarette smoking, alcohol intake,[10, 11] and overcrowding.[12]

1. Mycobacterium Tuberculosis (MTB) Complex

The MTB complex consists of mycobacterial species that cause tuberculosis

disease in humans and animals, including: Mycobacterium (M.) tuberculosis, M. bovis ,

M. africanum , M. microti and M. canetti ,[13] and a recent addition to this complex is M. mungi.[14] Members of this group are morphologically characterized by their slow growth on both solid and broth media as well as the presence of MPB70 antigen, IS6110 element and 16S rDNA on polymerase chain reaction (PCR) testing. Within the group, species differentiation is made by identification of specific sequence single-nucleotide polymorphism (SNPs).[14]

M. tuberculosis is the commonest member of the MTB complex and is responsible for more than 75% of cases of pulmonary TB globally. The host is mainly human and treatment is with conventional anti-TB drugs (Isoniazid, Rifampin,

Ethambutol and Pyrazinamide) for a period of six to 18 months.

M. africanum consists of two distinct clades: M. africanum type I seen

predominantly in West Africa and M africanum type II found in East Africa. In some

West African countries, this species accounts for more than 20% of cases of pulmonary

TB.[15] It causes disease in humans and animals (cattle in particular), and produces biochemical test results with an intermediate pattern between those of M tuberculosis

(sensitivity to pyrazinamide) and M. bovis. M. africanum is nitrate negative, a weak producer of niacin and sometimes grows in media supplemented with pyruvate,[16] and is transmitted from person to person by inhalation, though progression to disease after infection is slow compared to M. tuberculosis .[17]

M. bovis is the causative agent of tuberculosis in cattle (more commonly known as bovine TB). It is related to M. tuberculosis, and can infect humans. Effective prevention and control measures have rendered the disease rare in developed nations but the disease is still prevalent in developing countries. The uncontrolled movements of animals and overcrowding; farming practices that allow contact between animals from different herds; insufficient veterinary services; lack of proper disease surveillance; and intensive dairy farming practices were some of the factors identified in the persistence of disease in animals of developing countries.[18] In humans, especially in Sub-Saharan

Africa, cattle farming, mixed herding systems, age, sex (herding done mainly by children and men while the milking is mostly done by women) and consumption of raw milk and

meat were identified as risk factors for bovine TB infection.[19] Clinical signs,

radiological and pathological features of M. bovis are not different from those of M. tuberculosis making assessment of the extent of the burden from bovine TB infection in

Africa difficult. [20] Considering the routes of transmission, M. bovis is more likely to cause non-respiratory disease, particularly gastrointestinal TB.

2. Non-tuberculous Mycobacterium

Closely related to the pulmonary infection caused by tuberculous mycobacterium in both clinical presentation and laboratory outcome is pulmonary infection due to non- tuberculosis mycobacterium (NTM) . The NTMs are widely distributed in the environment and are increasingly being identified as common causes of pulmonary diseases in industrialized nations. [21, 22] The prevalence of pulmonary mycobacterial infection from NTM is on the rise in Japan, the United States and some countries in Europe and

Asia. [22-25] There are over 125 species identified to date and the list keeps expanding

with increasing breakthroughs in species isolation and identification. [22]

Mycobacterium-avium complex (MAC) is the dominant NTM infection in both the

United States and Japan while M. kansasii and malmoense are reportedly more common

in the United Kingdom and Scotland. [21, 22] The average age of patients with NTM infection in the U.S and Japan is over 50 years. Pulmonary NTM disease is found to be associated with HIV although studies have reported a rising incidence of disease among

HIV negative populations. [21, 26, 27] HIV-infected NTM subjects have associated fever,

anaemia, low CD4 counts and long standing symptoms. Radiological findings are similar

to those of pulmonary tuberculosis with predominance of interstitial infiltrates and cavity

lesions. [26, 27]

Information on the burden of NTM disease in Africa is lacking. Limited tools are available for mycobacterial species identification and such services are not routinely provided to patients. Standard of care practice only requires a microscopic examination of expectorated sputum after staining with a Ziehl-Neelsen stain to detect tuberculous acid fast bacillus (AFB). However, even in very good centers, expectorated smear AFB detection identifies only 40-50% of cases compared to an 80% yield by culture [1, 28].

Earlier studies conducted by the World Health Organization (WHO) as far back as the late 1950s and early 1960s have shown the presence of NTM in Africa. Using traditional tools to identify mycobacterial groups based on certain characteristics like speed of growth and morphology, NTMs were found in both tuberculosis patients and the general public in some African countries including Nigeria. [29, 30] According to Zykov and colleagues, the prevalence of NTM among tuberculosis patients in the countries surveyed was 1.1%.[29]

B. HIV-Mycobacterium TB Co-infection

HIV infection drives the TB epidemic in developing nations where the burden of

HIV disease is high.[31, 32] Worldwide, it is estimated that 9% of all new TB cases in adults (aged 15-49 years) are attributable to HIV infection with the highest proportion of coinfected cases being in Sub-Saharan African (31%), where HIV prevalence is the highest in the world.[33] There is a 1-in-10 annual risk of developing TB among HIV- infected persons which is approximately equivalent to the life-time risk of developing TB among HIV-uninfected persons. The life-time risk increases among HIV-infected persons with the likelihood of nearly 1-in-2 HIV-infected persons developing the disease in their life time.[34] HIV-infected TB cases are twice as likely to die compared to those who are

not HIV-infected [35]. It is estimated that the prevalence of HIV co-infection among TB cases in Nigeria is 30%; however, HIV screening is available to only 32% of TB cases and mostly at major health facilities[1].

C. HIV-Mycobacterium bovis Co-infection

Isolated clinical cases of HIV co-infection with M bovis have been reported, including an outbreak involving drug-resistant strains of M bovis [36, 37]. In a study in

Mexico, M. bovis strains were isolated from clinical specimens of HIV-infected persons.

[38] However, as of the time of designing this study we have not found other studies that established a similar link between either M. bovis or M. africanum and HIV as in the case of M. tuberculosis .

D. Tuberculosis Treatment

Treatment of uncomplicated tuberculosis requires at least six to nine months with a cocktail of antibiotics. The dosage and length of treatment depends on individual’s age, body weight, resistance to therapy and bacterial location in the body. The common combination drugs used for a standard treatment include: Isoniazid, Rifampin,

Ethambutol and Pyrazinamide. Treatment options and durations may vary in the presence of HIV co-infection (due to potential drug interactions); multi-drug resistant TB (MDR-

TB); and some infections involving M. bovis or NTM . Directly Observed Treatment short

(DOTS) course is the treatment strategy promoted by the WHO for the global control of

TB mainly to ensure adherence and minimize development of resistant strains. It requires drug intake to be directly observed by a healthcare worker or a community health worker for at least the first two months of therapy.[39]

E. Multi-Drug Resistant TB (MDR-TB)

This is defined as resistance to at least the two most powerful first-line anti-TB drugs; isoniazid (INH) and rifampicin (RMP) with or without resistance to any other antituberculosis drugs. Extensively drug resistant tuberculosis (XDR-TB) is defined as TB that has developed resistance to rifampicin and isoniazid (first line anti-TB drugs), as well as to any member of the quinolone family and at least one of the second-line anti-TB injectable drugs (kanamycin, capreomycin, or amikacin). This definition of XDR-TB was agreed upon by the WHO Global Task Force on XDR-TB in October 2006.

F. Multi-Drug Resistant TB in HIV

The association of TB with HIV infection and the increasing prevalence of

MDR-TB are threatening the global control of tuberculosis. The risk of death is even higher in the dual presence of HIV and MDR-TB. More than 50% of HIV co-infected

MDR-TB positive patients in a study in Peru died within 8 weeks of diagnosis. [40] In another study in South Africa, 98% of HIV co-infected patients with extensive drug resistant TB (XDR-TB) had a median survival of 16 days from the day of diagnosis. [41]

There is currently little primary data on the prevalence of MDR-TB in Nigeria, and many practitioners suspect the prevalence of primary MDR-TB to be much higher than the

WHO estimated 1.8%.[1] A recently concluded pilot survey on a small non- representative sample of TB cases with previous TB treatment found an MDR-TB prevalence of 13%.[42]

G. Burden of Environment Related Mycobacterial Infections in Nigeria

Pulmonary mycobacterial disease caused by NTM , also known as environmental mycobacteria, is the main form of environmentally related disease. However, pulmonary disease due to M. bovis is also considered environmental given that it can be acquired directly from cattle with respiratory disease and via aerosolized milk during milking and preparation of milk products. In Nigeria, M. bovis infection has been reported in recent surveys generating concern about its potential risk for HIV-infected persons. The prevalence of pulmonary tuberculosis (regardless of HIV co-infection) from M. bovis varies from 10% in the southern part of the country to about 15% in the more nomadic north [43, 44]. A recently concluded unpublished dissertation work titled: The

Epidemiology of Bovine and Human Tuberculosis in the Federal Capital Territory and

Kaduna State of Nigeria, PhD Thesis by Aisha Abubakar of University of Plymouth,

United Kingdom, 2007 compared the species of M. bovis isolated from suspected human and cattle TB cases and found the isolates to be genetically identical.

There is limited knowledge regarding risk of human infection with bovine TB in

Africa due to insufficient data on the disease in cattle and the risks associated with milk consumption. Like most African communities, milk in Nigeria is consumed either raw or fermented without pasteurization or purification. Furthermore, the practice of pooling milk together from different animals increases the risk of exposure to the milk of an infected cow. [45-47]. Of additional interest, is the recent discovery of a new colonal complex of M. bovis named “Africa 1” which is found in high frequency in four West

African countries including Nigeria [48].

The reported increase in NTM prevalence among TB patients in the industrialized nations, along with the epidemic in Sub-Saharan Africa and the lack of information on the burden of pulmonary disease caused by NTM in Africa is a major concern. This gap in the literature highlights the acute need for information on the burden, distribution, species identification and correlates of NTM infection in Africa.

The purpose of this study is to: (1) assess the proportion of pulmonary M. bovis ,

M. africanum and NTM among laboratory confirmed cases of pulmonary mycobacterial infection with and without HIV co-infection; (2) evaluate associations between HIV and

M. bovis , M. africanum and NTM ; (3) assess levels of resistance to isoniazid and rifampicin and (4) compare the demographic and clinical characteristics of pulmonary infection with NTM and MTB complex. While standard of care requires only the microscopic examination of an expectorated sputum sample, we used a complex algorithm that included smear microscopy, liquid broth and solid media culture, TB- antigen assay detection, and molecular probe assays, to determine the type of

Mycobacterium and pattern of resistance to isoniazid and rifampicin.

H. Research Questions, Hypotheses and Specific Aims

Questions

1. Among suspected cases of pulmonary mycobacterial infection, is there an

association between infection with Mycobacterium tuberculosis (MTB) complex

organisms ( M. tuberculosis, M. bovis, M. africanum), and Non tuberculous

mycobacterium (NTM) infection with HIV infection?

2. Is there a difference in drug resistance to TB therapy among TB positive/HIV-

positive cases compared to TB positive/HIV negative cases?

Hypothesis 1: HIV co-infected mycobacterial cases are more likely to have co-infection with MTB and NTM than are HIV uninfected mycobacterial cases.

Specific Aim I : To determine the proportion of confirmed mycobacterial cases caused by

MTB complex and NTM using an established algorithm that combines culture and

molecular line probe assay to determine the type of pulmonary mycobacterial infection.

Specific Aim II : To compare the proportion of MTB complex and NTM among HIV positive and HIV negative cases, and contrast the characteristics of pulmonary NTM infection with those of pulmonary TB infection caused by MTB complex.

Hypothesis 2: TB cases infected with HIV are more likely to have resistance to TB therapy compared to TB cases that are not infected with HIV.

Specific Aim III: To measure the proportion of cases with resistance to the first line anti-

TB drugs (isoniazid and rifampin), also known as multi-drug resistant TB (MDR-TB)

comparing HIV-positive to HIV-negative TB cases.

I. Summary of Research Significance

Identification of the specific species involved in mycobacterial infections in

Africa has largely employed the use of less sophisticated techniques with limited

sensitivity and potential for misclassifications. The first specific aim assesses the proportion of MTB complex organisms ( M. tuberculosis, M. bovis, M. africanum) and

NTM among suspected TB cases using the highly sensitive molecular diagnostic tools.

The outcomes of this aim may influence TB clinical care services as well as TB control strategies in Nigeria. Identifying factors associated with these infections will allow health personnel the opportunity to counsel populations at risk of infection. The government may then target specific population groups such as livestock farmers, livestock merchants, slaughter house workers and meat vendors for health education on the importance of livestock handling, milk pasteurization as well as surveillance and reporting of bovine TB cases. These were measures that were effective in reducing the prevalence of bovine TB in other communities.[49-51] New policy measures may be undertaken to reduce environmental contamination with NTM including health education and provision of screening tools at TB treatment facilities.

The impact of HIV infection on M. tuberculosis has been extensively studied and reported,[52-54] however, little is known about the impact of HIV on human TB in

Nigeria caused by other species of mycobacteria. Few studies, including a report of a nosocomial outbreak of pulmonary TB of a case co-infected with HIV and a drug resistant strain of M. bovis, have demonstrated the likelihood that an HIV-infected patient may host M. bovis infection.[37, 55] The second specific aim evaluates possible

associations between MTB complex and NTM and HIV infection . Counseling HIV- infected individuals against potential sources of exposures to these bacteria could improve survival and minimize morbidity. Livestock keeping and consumption of locally produced milk or milk products are popular with the populace. Data analyses for this specific aim may provide insight into epidemiologic factors that uniquely define environmentally acquired mycobacterial infections in Nigeria. This information may enhance management and control of mycobacterial infections.

The third specific aim assesses the association between HIV and resistance to therapy with a focus on MDR-TB, which is characterized by resistance to the two, top first line drugs for TB treatment: rifampicin and isoniazid. Patients with MDR-TB have to be treated with the much more expensive second-line drugs, and if not treated quickly, can result in a high fatality rate and transmission to others. With the high rate of pulmonary TB and HIV co-infection; circulation and transmission of resistant strains of mycobacterium in this densely populated nation of Africa will adversely affect the global

TB control effort. The presence of MDR-TB in HIV co-infected individuals will make a bad situation worse. For a resource limited nation, the consequences could be catastrophic. Furthermore, M. bovis species bovis is inherently resistant to pyrazinamide, and cannot be treated with it. The presence of resistance to any or both the first line drugs by this strain of M. bovis will automatically compromise the most available and affordable first line treatment.

II. RESEARCH DESIGN AND METHODS

A. Study Sites

The two Nigerian sites for this study are located within the north-central state of

Kaduna. The National TB and Leprosy Training Center (NTBLTC) is located in Zaria while the Barau Dikko Hospital (BDH) is in the metropolitan Kaduna City, Nigeria.

Kaduna state has an estimated population of a little over 6 million people according to the

2006 census report. There were 58 urban and rural centers across where TB cases were detected. In 2007, about 5,000 new TB cases were detected across the state (personal communication with the state’s TB coordinator: Dr. Gajere). Five centers, namely:

Dantsoho, Barau Dikko, Forty-four Army Reference Hospital, Gwamna Awan and the

NTBLTC account for about 40-50 percent of new TB cases detected in the state annually.

The study center (NTBLTC) accounted for about 25% of the newly detected cases in the state in 2007 (record inspection at the NTBLTC). The center provides TB, HIV and leprosy care to individuals from within the state of Kaduna and from all other states in the northern Nigerian region. The centre is supported by the Institute of Human Virology,

University of Maryland, Baltimore’s (IHV-UMB) U.S President’s Emergency Plan for

AIDS Relief (PEPFAR) program whose primary focus was to prevent and treat

HIV/AIDS, and combat opportunistic infections including tuberculosis. The IHV-UMB established and equipped state-of-the-art TB bio-safety-level 3 (BSL-3) diagnostic laboratories at the NTBLTC. This centre is situated about 40 miles north of Kaduna City.

The Barau Dikko Hospital (BDH) is located within the city of Kaduna and is the second largest TB detection center in the state of Kaduna. It accounts for about 8% of newly detected cases in 2007. Patients from the metropolitan and outskirts of Kaduna

receive care at this facility which in addition to a TB clinic has a PEPFAR clinic for HIV treatment.

Approvals for the conduct of this study at these sites were granted by the

University of Maryland Institutional Review Board and the Nigeria National Health

Research Ethics committee with written expressions of support by the directors of the study sites.

B. Target Population and Study Design

This was a cross-sectional study where individuals visiting the directly observed therapy short course (DOTS) clinics located within the selected sites were recruited into the study through their care providing physicians. Suspected cases of TB who met the eligibility criteria were enrolled from August 2010 through August 2011 at the NTBLTC in Zaria, and from December 2010 through August 2011 at the BDH in Kaduna City.

Participants were invited to sign an informed consent for their participation in the study that included the performance of additional tests on the sputum samples they provided to the facility for routine clinical care as well as access to their medical records. HIV screening was done routinely as a policy for all suspected cases of TB. Participants were included if they had already accepted routine HIV counseling and testing; gave an informed consent; spoke either the native Hausa or English language; and were aged 18 years or older or were emancipated adults.

C. Data Collection

Data for this research were collected from the two Ministry of Health TB clinics

at the BDH in Kaduna and at the NTBLTC in Zaria. Participants who gave informed

consent were then surveyed using a pilot-tested survey instrument (Appendix A) that

contained 43-items numbered from 005 to 007 and from 101 to 140: 8 items were

baseline demographics including weight and height; 4 items were related to the risk

factors for HIV infection, the remaining 31 items probed participants on potential risk

factors for mycobacterial infection including M. bovis . All surveys were administered by trained interviewers.

Each participant was measured to assess height and weight, and these measurements were recorded on the survey instrument. The scales used for measuring height and weight were of the same make and type for the two study sites. Most of the questions asked were close-ended with response categories coded: 0 for a ‘no’ response,

1 for ‘yes’, 97 for ‘don’t know or not sure,’ and 99 for a ‘refused response’. Some questions had options for the different item with response categories code ranging from 2,

3, 4, 5 or 6 depending on the number of options provided. Table 1 provides the list of

collected variables, their definition

Table 1: Variables in the study data and their definitions and categories Variable Definition Category Demographics Sex Participant’s sex Male (M) or Female (F) Age Age in years as of the date of enrollment Level of education Level of formal education None completed by participant Primary (elementary) Junior secondary (Middle school) Senior secondary (High school) Diploma (College) University Ethnicity Participant’s Ethnic group or Hausa Tribe Fulani Yoruba Igbo Other Marital Status Participant’s marital status Married Not married Widowed Divorced Single Occupation Participant’s occupation Full time farming Part time farming Animal health worker Human health worker Abattoir (slaughterhouse) worker Meat vendor Other Unemployed Height Height in Meters Weight Weight in Kilograms Mycobacterial infection risk factors Number per house Number of people living in One to five hold participant’s house hold Six to ten Greater than ten Number per room Number of persons sleeping in a One to two room Three to five

Greater than five Windows per Number of windows per room No window room One window Two or more windows Keep livestock Engagement in cattle rearing or No, Yes, Don’t know or other livestock raising Refused

Time spent with Average number of hours of the No time spent livestock day spent with livestock The whole day One hour or less More than one hour Type of livestock The type of livestock participant None spends time with, if any Cattle Sheep Goat Mixed Herd size The number of livestock None participant spends time tending to Less than ten in a day, for participants that keep Between ten to twenty livestock Between twenty-one to thirty More than thirty Sick cattle Presence of sick cattle in the No, Yes, Don’t know or participant’s herd or Refused neighborhood in the last three months Cattle with cough Presence of any cattle with cough No, Yes, Don’t know or in participant’s herd or Refused neighborhood in the last three months Cattle with TB Presence of any cattle diagnosed No, Yes, Don’t know or with TB in the participant’s herd Refused or neighborhood in the last three months Cattle with Presence of any cattle in the No, Yes, Don’t know or swollen neck participant’s herd observed to Refused have swollen neck in the last three months New cattle in the The presence of new cattle in the No, Yes, Don’t know or herd herd introduced in the last three Refused months Milk lactating Milking of lactating livestock for No, Yes, Don’t know or livestock any purposes Refused Milk consumption Consumption of locally expressed No, Yes, Don’t know or milk or meal consisting of this Refused type of milk Consumption rate The frequency of milk Not at all

consumption by the participant Every day Some days Don’t know Refused Form of milk Locally produced milk is None consumed consumed in different forms: Unboiled fresh or locally treated (boiled) Boiled Both Don’t know Refused Raw meat intake For people that eat uncooked No, Yes, Don’t know or meat or drink blood Refused Current TB Participants already started on TB No, Yes, Don’t know or treatment drugs at the time of enrollment Refused Previous TB Previous TB treatment either No, Yes, Don’t know or treatment completed or disrupted Refused Family or Current case or cases of TB in No, Yes, Don’t know or contact’s history the participant’s family or Refused of TB contacts Family or Previous case or cases of TB in No, Yes, Don’t know or contact’s the participant’s family or contact Refused previous history of TB Neck swelling Presence of neck swelling on any No, Yes, Don’t know or member of the family or close Refused contacts Member tending Member of the family or contact No, Yes, Don’t know or to cattle with TB symptoms that tends to Refused cattle BCG vaccination Participant’s receipt of BCG No, Yes, Don’t know or vaccination against TB Refused History of Diagnosis of diabetes mellitus in No, Yes, Don’t know or Diabetes mellitus the participant prior to enrollment Refused Cigarette Participant’s history of previous No, Yes, Don’t know or Smoking or current smoking Refused 100 cigarette Estimated number of cigarettes No, Yes, Don’t know or sticks smoked so far equals or exceeds Refused 100 Sticks per day Estimated number of cigarettes 0, 1-2, 3-5, > 5, Don’t know or currently smoked on daily bases Refused Alcohol intake History of ever drinking alcohol No, Yes, Don’t know or Refused Recent alcohol History of drinking alcohol in the No, Yes, Don’t know or intake last 30 days Refused HIV infection risk factors

Sexually Participant previous treatment for No, Yes, Don’t know or transmitted Gonorrhea, Syphilis or HPV Refused infection Unprotected sex Extra marital sex with multiple No, Yes, Don’t know or partners without condom in the Refused last one year Sex with HIV- Occurrence in the past of No, Yes, Don’t know or infected person unprotected sex with HIV- Refused infected person Blood transfusion Participant previous transfusion No, Yes, Don’t know or with blood, if any Refused Isolate Variables Sputum smear Outcome of participant’s sputum Positive or Negative Microscopy sample evaluation with Ziehl Neelsen stain MGIT culture Growth of Mycobacterium from Positive, Negative or sputum sample after 42 days Contaminant incubation in Mycobacterium Growth Incubator Tube System SD bioline Confirmation of M. tuberculosis MTC or NTM suspect complex isolate from a positive MGIT growth MTBC assay Hain molecular assay results for M.TB , M.bovis or M. africanum MTC isolates characterization MTBDRplus Hain molecular assay results for No resistance assay MTC resistance to isoniazid and Resistance to rifampicin rifampicin Resistance to isoniazid Resistance to both (MDR) AS/CM assay Hain molecular assay results for Not mycobacterium, NTM isolates characterization NTM characterized NTM uncharacterized HIV status Outcome of HIV rapid assay Positive or Negative screening at enrollment from participant’s medical records Chest-x-ray Chest-x-ray findings for Normal participant’s enrolled at NTBLTC Abnormal consistent with TB Zaria Abnormal not consistent with TB

D. Laboratory testing

The laboratory data consisted of HIV blood test results, and the TB test results on participants’ sputum samples. HIV testing was completed locally at each of the two clinic sites (BDH and NTBLTC) as per standard of care. The microscopic examination of a sputum smear sample for acid-fast tubercle bacilli using a Ziehl-Neelsen stain was conducted at each site as per standard of care; sputum smear examination were repeated again to confirm the presence of AFB after culture growth.. Sputum samples collected from both sites were processed only at the three NTBLTC laboratories: (1) Ministry of

Health (MoH) standard TB laboratory, (2) portable BSL-2 and (3) portable BSL-3 IHV-

UMB laboratories. Samples from BDH Kaduna were transported to the NTBLTC 24 to

48 hours after collection with maintenance of the cold chain and were processed immediately upon receipt. Samples were processed as follows at NTBLTC:

1. Routine smear microscopy tests were done for all patients presenting to both clinics at

the standard MoH laboratory and results were immediately made available to the

clinics for the purpose of care.

2. Sputum samples from patients, who consented to participate in our study, were

moved to the BSL-2 laboratory for initial culture in liquid broth media (MGIT) and

subsequent subculture in a solid Lowenstein Jensen (LJ) media.

3. Additional molecular assays were completed at the BSL-3 laboratory.

Completed data for a positive result took a minimum of 3 weeks to be generated.

E. Detailed Description of Laboratory Methods

1. HIV status determination

Evidence of HIV infection was detected by the standard HIV screening algorithm

recommended by the Center for Disease Control and Prevention (CDC) office in Nigeria,

and the Nigerian Ministry of Health (MoH) for all PEPFAR program sites. Blood was

serially tested with two HIV rapid assays (Trinity Biotech Unigold and Abbott

Determine). Detection of HIV antibody in both was required to consider a suspected

sample HIV positive. A tie breaker assay-Chembio Diagnostic STAT PAK was used in

situations where the first two assays had discordant outcomes. A positive test with the

third assay confirmed positive HIV status; otherwise, the case was regarded as HIV

negative for the purpose of this survey. The Abbott Determine assay has a reported

sensitivity of 99.4% and a specificity of 99.6%; similarly, the Trinity Biotech Unigold

assay has 98.5% and 99.5% sensitivity and specificity respectively, for HIV antibody

while the tie breaker STAT PAK has sensitivity of 99.8% and a specificity of 99.9%.[56-

58]

2. Sputum sample collection

Sputum smear microscopy of expectorated sputum was used at the laboratory to detect cases of TB according to the WHO guideline. This involved examination of three collected sputum samples within 24 hours: (1) ‘on the spot’ sputum sample under supervision during the first visit; (2) an ‘at home early morning’ sample the next morning without supervision; then (3) another ‘on the spot’ sample under supervision when the ‘at home’ sample was collected.

3. Sputum smear microscopy test

Sputum samples were collected in sterile containers and then registered and assigned accession numbers. Smears of size 1 x 2cm were made on a new grease-free slide and allowed to air dry. The air dried smears were then fixed by gently passing them on a flame 2-3 times. The smears were then stained with Ziehl-Neelsen (ZN) technique as described elsewhere[59] . Briefly, 1% strong carbol fuchsin was applied to the slides and heated with a Bunsen flame intermittently 3 times and allowed to stain for 15 min. The stain was washed off with gentle tap water, 3% acid alcohol was then applied on the smears for 3 minutes. Methylene blue (0.3%) was then added and left to stand for 1 minute before it was washed off with tap water. Positive and negative controls were included in the staining process. Examination of the slides was done under a high magnification (x100) microscope to determine the presence of acid fast bacilli (AFB) and count them. For the purpose of this study, smears were graded and interpreted as shown in Table 2 below.

Table 2. Grading and interpretation of smear microscopy test

Number of AFB seen Recorded as Reported as

No AFB in at least 100 microscopic fields 0 Negative 1-9 in 100 fields scanty Positive 10-99 in 100 fields +1 Positive 1-10 AFB per field in at least 50 fields +2 Positive > 10 AFB per field in at least 20 fields +3 Positive

4. Liquid broth mycobacterial culture

The early morning home collected sputum samples were transferred to the BSL-2

laboratory in 50 ml falcon tubes to a class II biosafety cabinet (BSC). These samples

were excluded from performance of the routine smear test before their incubation to

minimize contamination which was high at the commencement of the study. They were

then cultured in the liquid broth culture medium as described elsewhere [60] . Briefly, the

sample was treated with BD Mycoprep TM (Beckton Dickinson Diagnostic Systems,

Sparks, Maryland, USA) which consists of 1% N-acetyl-L-cysteine (NALC), 4% sodium hydroxide and 2.9% sodium citrate. An equal amount of Mycoprep was added to the sputum, homogenized, and allowed to act for 15 minutes. Phosphate buffered saline was added to stop the decontamination reaction. This was then centrifuged under refrigerated conditions at 3000 rpm for 15 minutes. The supernatant was discarded and the pellet was reconstituted with 2.5 ml phosphate buffered saline. Supplemented 7 ml Mycobacterium

Growth Indicator Tube (MGIT) tubes were inoculated with 0.8 mls of the pellet and incubated in the automated BACTEC MGIT 960™ machine (Becton Dickinson

Diagnostic Instrument Systems), which monitors growth.

5. Interpretation of culture results and further characterization

Figure 1 summarizes the steps involved in the detection and species characterization of

TB cases in this study. Samples that failed to show any growth in the MGIT tube after 42 days of incubation in the machine were removed and classified as negative. Samples that indicated positive growth were removed from the machine and inoculated on blood agar to check for non-tubercle bacterial contamination. If the sample showed non- mycobacterial growth on blood agar, then smear microscopy with ZN stain was

performed on this sample, and if it was negative for AFB, then the growth was

considered a contaminant, and the sample was excluded from the study.

Among those with positive MGIT, with and without growth on blood agar, and positive

for AFB with a ZN microscopic examination, a TB antigen assay (SD Bioline Ag MPT64

Rapid™ assay, Standard Diagnostics, Kyonggi-do, Korea) was used for the rapid

differentiation between MTB complex and NTM . The MTB complex positive samples were then speciated using the GenoType MTBC test (Hain Lifescience, Nehren,

Germany) to identify the different MTB complex organisms (M. tuberculosis, M. bovis and M. africanum). Samples without non-mycobacterial growth but negative by ZN staining were incubated for another ten days and checked by ZN staining every five days, and if still negative by then, they were classified as negative.

Samples positive for AFB on the ZN stain and negative for MTB complex on the SD-

Bioline test were considered as NTM and then were further sub-cultured on the

Lowenstein Jensen (LJ) solid culture medium for a maximum of 42 days. A positive growth on the LJ solid culture was sent for speciation with the GenoType Mycobacterium

CM/AS assay (Hain Lifescience, Nehren, Germany). If these samples were negative on the CM/AS assay the culture growth was considered a non-mycobacterial growth (NMB).

Sputum samples that were negative on LJ medium were also considered as contaminants.

Figure1. Flow diagram illustrating steps used to detect and characterize mycobacterial isolates

HIV antibody and sputum smear tests NTBLTC Zaria Site A Site B BDH Kaduna

Sputum samples incubated in Bactec MGIT system 6

POS NEG after 42 days STOP Blood Agar

POS NEG ZN Stain ZN Stain

Inc. for POS NEG POS NEG < 42 ZN Stain days STOP

Contaminant POS NEG

STOP

MGIT MYCOBACTERIA POSITIVE STOP SD bioline TB antigen assay NEG Contaminant POS NEG LJ Medium <42 days

Hain MTBC POS STOP Hain CM/AS assay STOP assays

Both the Hain assays (MTBC and CM/AS ) used DNA strips that contain specific probes targeted to specific regions on the genes shared by various strains of the mycobacterial isolates. Pieces of DNA formed from polymerase chain reaction (PCR) react with these probes and anneal to their complementary strand in the probes leading to the observable bands on the strip as indicated in Figure 2.

Figure 2 Mycobacterial tuberculosis complex strains banding pattern in Genotype MTBC assay (source: Hain Lifescience)

6. Detection of resistance to isoniazid and rifampicin

Resistance to isoniazid and rifampicin was tested among the MTB complex

isolates using the Hain MTBDR plus assay (Hain Lifescience, Nehren, Germany) . The

probes in the DNA strips for the MTBDR plus assay target important mutations within

specific genes encoding for critical enzymes in the metabolism of rifampicin and

isoniazid drugs.

As indicated in Figure 3 , the Genotype MTBDR plus assay strip has a total of 27

reaction zones. Twenty-one reaction zones are probes for mutations while remaining six

are control probes for verification of the assay procedures. The control probes consist of a

conjugate control, and amplification control, an MTB complex-specific control, an rpoB amplification control, a katG amplification control, and an inhA amplification control.

For rifampicin resistance, the probes target the rpoB gene, while for isoniazid resistance the specific probes locate the katG and inhA genes.

Resistance to a particular antibiotic is indicated by the absence of at least one of the wild-type bands or the presence of bands indicating a mutation in the drug’s resistance-related gene. When all the wild-type probes of a gene are positive and there is no detectable mutation within the region examined, the sample is sensitive to the respective drug. For a valid result, all six expected control bands should appear correctly.

Otherwise, the result is considered invalid. The sensitivities as well as specificities of the

Hain line probe assays were previously validated and documented [61-63].

Figure 3. Genotype MTBDR plus strip reaction zones for detection of mutations to rifampicin and isoniazid (source: Hain Lifescience)

katG RMP INH T T T U T T

U U t t e n n

# v M W M W M i a a

t t t

i WT1 WT2 WT3 WT4 WT5 WT6 WT7 WT8 MUT1 MUT2A MUT2B MUT3 B s s G G A A MUT1 MUT2 WT1 WT2 MUT1 MUT2 MUT3A MUT3B s B WT i i t t o n h h s s U a a p e

e e n n T r k k i i s r sensitive r CC AC TUB rpoB rpoB rpoB rpoB rpoB rpoB rpoB rpoB rpoB rpoB rpoB rpoB rpoB katG katG katG katG inhA inhA inhA inhA inhA inhA inhA 1 ++- +- +- + +

2 +- +- ++- + + 3 ++- +- - + + +

4 +- +- +- + + + 5 +- - - - +- + + rpoB inhA 27

7. Molecular line probe assays technique

The three molecular line probe assays described above (MTBC, CM/AS and

MTBDR plus assays; Hain Lifescience, Nehren, Germany), were essentially the same for

all tests, differing only on the primer nucleotide probe mix that varies according to the

tests. An aliquot of 1000 ul from the positive mycobacterium growth indicator tube liquid

medium (MGIT) or LJ medium was added into a 2 ml tube. The mycobacteria were

pelleted by spinning for 15 minutes in an aerosol tight micro centrifuge at 10,000 rpms.

The supernatants were discarded and the sediments re-suspended in 100 ul of molecular

grade water by spiral motion. The mycobacterial suspension was then incubated for 20

minutes in a water bath at 95 0C followed by incubation in an ultrasonic bath for 15

minutes. This was followed by centrifugation at full speed (10,000g) for 5 minutes and

the supernatant containing the DNA was aliquoted and stored at -20 0C if not amplified

immediately.

An amplification mix containing 35 ul of Primer Nucleotide Mix, 5 ul of 10X

PCR buffer, 2 ul of 25 nM MgCl 2, 0.2 ul of HotStar Taq and 3 ul of water were prepared

per sample. Five microliters of the DNA was added to this amplification mix and

amplified at the following heating cycles: a) 1 cycle at 950C for 15 minutes b) 10 cycles

at 95 0C for 30 seconds then at 58 0C for 2 minutes, c) 20 cycles at 95 0C for 25 seconds,

then at 40 0C for 40 seconds and d) 1 cycle at 70 0C for 8 minutes. Amplification products were stored at +4 to -20 0C if not detected immediately. Detection of the amplicons was

done by hybridization technique using the twincubator hybridization tray. Some 20 ul of

the amplicons was added to 20 ul of the denaturation solution in a tray and incubated for

5 minutes. Then, 1 ml of pre-warmed hybridization buffer was added to the tray and

rocked for homogeneity. Labeled DNA strips were immersed into each well of a tray

placed on the shaking twincubator.

The incubation lasted for 30 minutes at 45 0C and was followed by aspiration of

the buffer and the addition of 1 ml of stringent wash solution. It was incubated again on

the twincubator for 15 minutes at 45 0C. The stringent wash solution was completely removed, and 1 ml of rinse solution was added for 1 minute. Some 1 ml of diluted conjugate was added to each strip and incubated for 30 minutes on the twincubator. The solution was removed and washed twice for 1 minute with 1 ml rinse solution and 1 ml of distilled water on the twincubator. The solutions were poured out each time, and 1 ml of diluted substrate was added to each strip and incubated while protected from light without shaking for 3-20 minutes for color development. Reactions were stopped by briefly rinsing twice with distilled water. The strips were removed with tweezers, dried, pasted and interpreted with the Hain interpretation chart as per the manufacturer’s instructions.

F. Data Management and Analyses

Microsoft Access XP database was used to capture the data that was collected into different tables that were linked by participants’ identification (PID). Data were entered at least twice with applicable error checking and data validation rules enforced.

Statistical analysis software (SAS Institute, Inc., Cary, North Carolina) version

9.2 was used for the analysis. Data were further examined for outliers and missing values.

The frequency distributions of categorical variables were examined and missing values were further identified, checked and corrected. The continuous variables age, height and weight were assessed for outliers on box plots. An additional continuous variable: body mass index (BMI) was created from variables height and weight. Groups of variables for age and BMI were then created based on their mean or median values. The two study sites were compared with respect to participants’ baseline demographic and selected laboratory characteristics.

1. Specific Aim I

To determine the proportion of confirmed mycobacterial cases caused by MTB complex and NTM using an established algorithm that combines culture and molecular line probe assay to determine the type of pulmonary mycobacterial infection.

The proportions and their 95% confidence intervals of MTB complex organisms

(M. tuberculosis, M. bovis , M. Africanum ) and NTM were calculated among all patients screened and among confirmed cases of pulmonary mycobacterial infections. Then, among all screened patients stratified by the study sites. Within the confirmed cases of

Mycobacterial infection, the proportions of tuberculous and non-tuberculous groups were estimated. In each of these groups, proportions of isolates were determined with the

group’s total as denominator.

In calculating the sample size for estimating the proportion of TB cases in the

study population, a 95% confidence interval was used with accuracy of estimation within

3% margin of error of the true proportion. The previously reported estimate of prevalence

of M. bovis among suspected TB cases with unknown HIV status was 15% [44]. Based on this prevalence, the minimum number of TB cases required using standard errors based on binomial distribution was 544. To obtain this number of TB cases, we estimated that 1,830 patients suspected of having pulmonary TB will need to be screened. The study aims were modified along the way to accommodate the emerging evidence of significant number of cases of TB due to M. africanum and NTM in this population from our molecular analysis. Although instances of M. africanum and NTM infections among pulmonary TB cases were earlier reported in Nigeria,[64, 65] these were not used to determine the sample size due to the post-hoc nature of our study findings.

2. Specific Aim II

To compare the proportion of MTB complex and NTM among HIV positive and HIV negative cases, and contrast the characteristics of cases with pulmonary NTM infection with those of cases with pulmonary TB infection caused by MTB complex.

In preliminary bivariate analyses, Pearson’s chi-square tests or Fisher’s exact test were used to examine the unadjusted associations between the outcomes of interest

(presence of infection with MTB complex organisms [ M. tuberculosis, M. bovis , M. africanum ] or NTM ) and the primary exposure (HIV infection) and other selected

covariates. Odds ratios and 95% confidence intervals were reported, and p-values 0.05 or less were considered statistically significant.

Covariates evaluated for associations with outcome variables included livestock farming, livestock type, time spent with livestock and intake of unpasteurized milk.

Others included age, sex, cigarette smoking, alcohol consumption and BMI. Additional factors included family history of TB; history of diabetes mellitus; overcrowding (more than 3 persons in a living room); and marital status. Each of these factors was found to be independently associated with pulmonary TB in previous studies [10-12, 66, 67]. Age, sex, occupation and marital status also predict HIV infection [68]. Other HIV associated risk factors evaluated were: history of sexually transmitted infections (STI); unprotected sex; prior blood transfusion; and sex with a known HIV-infected person. Participants’ level of education and ethnic group were also among the selected covariates.

Proportions of HIV infection in each of the outcomes: M. tuberculosis , M.bovis ,

M.africanum, and NTM cases were compared with the proportion of HIV infection among mycobacterium culture-negative cases within the dichotomized levels of the selected covariates for evidence of confounding or effect modifications. Confounding and effect modification were tested in stratified analyses. We took the following 2 steps to evaluate covariates as potential confounders. First, we evaluated each covariate, one at a time, by comparing the crude odds ratio (between HIV status and mycobacterial group) to the Mantel Haenszel estimate stratified by the covariate. A difference of 10% or more between the adjusted Mantel Haenszel and the unadjusted odds ratio was considered potential confounding.

Second, multiple logistic regressions were used to adjust for confounding of the associations between the different mycobacterial groups and HIV in which odds ratios and 95% confidence intervals were reported. Each potential confounder identified in the first step was added one-by-one to the simple model consisting of the outcome of interest and the main predictor (HIV). A covariate was retained in the model if it was significant

(p < .05) or if it was considered an important covariate due to biologically plausible relationships. A variable was not retained in the model if it was not significant (p > .05) and its removal did not significantly reduce the model fit (as assessed by -2 log likelihood).

Likewise, we used a two-step process for evaluating effect modification. First, potential effect modifiers were identified using the Breslow-Day test for homogeneity of the odds ratios over different strata of the covariate. A p-value of 0.05 or less was considered significant modification.

Second, for those covariates identified as potential effect modifiers, effect modification was tested by comparing two logistic regression models – one model with the interaction term versus the model without the interaction term. The log likelihood ratio test was used to assess model goodness of fit improved with the interaction term.

Statistical analysis software (SAS Institute, Inc., Cary, North Carolina) version 9.2 was used for the analysis. Two-sided P-values of 0.05 or less were considered statistically significant.

Comparison of characteristics between MTB complex cases and NTM was done using a chi-square test in which the odds ratio and 95% confidence intervals were reported and a p-value of 0.05 or less was considered significant.

Power calculations were performed to assess the number of subjects needed to have sufficient power to detect a significant difference in the proportion of cases with M. bovis infection. We used the reported 15% among cases of pulmonary TB with unknown

HIV status [43, 44, 64] as the estimate of prevalence of M. bovis and M. africanum

among HIV negative TB cases. Using a fixed power of 80% and a two-sided error rate of

0.05; and the ratio of TB cases with HIV to TB cases without HIV as well as the

proportion of M. bovis was allowed to vary, the minimum sample sizes of TB cases with and without HIV required to detect a difference were given in Table 3 . For this specific aim, a 10 percentage point difference was defined as clinically important; hence we powered the study to detect this difference. With a ratio 2:1 of two HIV unexposed to one

HIV exposed, the minimum number of confirmed TB cases required was 546 of which at least 182 would be HIV positive. To obtain the required sample size, we estimated that about 1,835 suspected cases will need to be screened.

Table 3 . Sample size estimation for different effect sizes and ratios of HIV negative to

HIV positive for a fixed power (.80) and a two-sided error rate ( α =0.05)

Total confirmed TB cases (N), number of HIV Positive TB cases from the total confirmed cases (n) and corresponding ratio

% M. bovis in Expected HIV Unexposed detectable % of 1:1 ratio 2:1 ratio 3:1 ratio TB cases M. bovis in HIV Exposed TB N n N n N n cases

10 15 1372 686 1506 502 1760 440 10 20 400 200 429 143 496 124 10 25 200 100 213 71 244 61

15 20 1812 906 2007 669 2356 589 15 25 500 250 546 182 640 160 15 30 240 120 261 87 304 76

The association of HIV with NTM and the comparison with MTB complex were assessed based on the numbers already identified since these were not in the primary aims at the onset of the study.

3. Specific Aim III

To measure the proportion of cases with resistance to the first line anti-TB drugs

(isoniazid and rifampicin) also known as multi-drug resistant TB (MDR-TB) comparing

HIV-positive to HIV-negative TB cases.

This analysis included only the members of MTB complex organisms since the testing assay’s ability to detect anti-TB drug resistance in NTM cases was limited, and

hence resistance testing was done only on MTB complex organisms. The proportion of 35

cases with resistance to isoniazid, rifampicin and to both drugs was calculated and patterns of resistance described. Crude associations comparing cases with any resistance: all forms of resistance (resistance to isoniazid alone, or to rifampicin alone, or to both combined), resistance to both drugs combined (MDR-TB), isoniazid (INH) only and rifampicin (RIF) only resistance to cases without any resistance with respect to HIV infection were evaluated. This comparison between cases with different resistance outcome and cases without any resistance was repeated with respect to measured covariates that included: prior TB treatment, age, sex cigarette smoking alcohol intake and diabetes mellitus.

Fisher’s exact test was used to examine associations between the outcome groups, the predictor and the measured covariates. Odds ratios and 95% confidence intervals were computed. Covariate evaluated and found to be associated with any of the outcome group and/or HIV was tested for confounding and effect modification in stratified analyses. Like in the specific aim II above we took the following 2 steps to evaluate covariates as potential confounders. First, we evaluated each covariate, one at a time, by comparing the crude odds ratio (between HIV status and mycobacterial group) to the Mantel Haenszel estimate stratified by the covariate. A difference of 10% or more between the adjusted Mantel Haenszel and the unadjusted odds ratio was considered potential confounding.

and the main predictor (HIV). A covariate was retained in the model if it was significant

Statistical analysis software (SAS Institute, Inc., Cary, North Carolina) version 9.2 was used for the analysis. Two-sided P-values of 0.05 or less were considered statistically significant.

Power calculations were performed to assess the number of subjects required to detect a significant difference in frequency of resistance between HIV positive and HIV negative cases. As with the specific aim II above, power and error rate were fixed at 80% and 0.05 respectively. The estimated MDR-TB prevalence of 10% was based on a reported pilot survey that found an MDR-TB prevalence of 13% among cases of TB receiving treatment in Nigeria [42]. The estimate was also based on a 2:1 ratio of two

HIV negative TB cases to one HIV positive case. As shown in Table 3 above, a 10% point difference between exposed and unexposed using a baseline prevalence of 10% among the HIV unexposed and a 2:1 ratio (unexposed: exposed); the minimum cases of

TB needed were 429; at least 143 of them with HIV co-infection.

III. High Prevalence of HIV and Environmentally Acquired Mycobacterium Infection in Nigeria

G.G. Aliyu, S.S. El-Kamary, A. Abimiku, C. Brown, K. Tracy, L. Hungerford and W. Blattner

A. Abstract

Background : Nigeria has the fourth highest tuberculosis (TB) burden worldwide, and the highest in Africa. It also has the second highest HIV-TB co-infection in Africa. There is limited information on the burden of environmentally acquired mycobacterial infection from this nation. Recent reports from other countries indicated a high level of known environmental pathogens such as non-tuberculous mycobacteria (NTM) and

Mycobacterium bovis (M. bovis ) infections among pulmonary disease cases raising concern about the risk for HIV-infected patients. We investigated culture positive cases of pulmonary mycobacterial infections for association with HIV from two tuberculosis treatment centers in Kaduna state, Nigeria.

Methods : Sputum samples of 1,603 consecutive cases suspected to have tuberculosis

(TB) were collected. HIV testing was performed on all cases as part of routine standard of care screening. All sputum samples were cultured in liquid culture Mycobacteria

Growth Indicator Tube (MGIT) system, and culture-positive samples suggestive of NTM were sub-cultured in solid Lowenstein Johnson (LJ) medium. Genotype MTBC , AS and

CM assays (Hains Life science, Nehren, Germany) were performed as appropriate on the culture positive isolates, to further characterize isolates.

Results : Of 1,603 patients screened, 444 (28%) culture-positive cases of pulmonary tuberculosis were identified. Of these, 375 (84%) were due to strains of mycobacterial

tuberculosis (MTB) complex (354 cases of M. tuberculosis , 20 M. africanum and one case of M. bovis ) and 69 (16%) were due to infection with NTM . The rate of HIV co- infection was 23% among MTB complex cases and 38% among the NTMs . In contrast to the MTB complex cases, the NTM cases were more likely to be diagnosed during the season of Harmattan (intense dust storms occurring during the months of December to

February).

Interpretation: While infections with M. bovis are rare, the high frequency of cases with clinical NTM infection and its association with seasonal dust exposure and HIV co- infection present a novel public health challenge for prevention and treatment, given the known suboptimal efficacy of standard-of-care TB regimens in these patients.

Introduction of molecular detection and screening assays to address the rapid identification of NTM and drug resistant TB is a high priority for strengthening the public health response.

B. Introduction

The constant interaction with environment and the effect of climate change increase

human exposure to environmental pathogens.[69] Modification of the environment

through agricultural practices and domestication of animals may have facilitated the

transmission to human of pathogenic organisms of the wildlife.[70] While

Mycobacterium tuberculosis (M. tuberculosis) is the commonest cause of pulmonary TB in humans, Non-tuberculous mycobacteria (NTM) and M. bovis have emerged as

common causes of environmentally-acquired pulmonary diseases.[71, 72]

Non-tuberculous mycobacterium is frequently mistaken for pulmonary

tuberculosis, particularly in developing countries, where laboratory identification is not

commonly available. They are widely distributed in the environment and are also being

identified as common causes of pulmonary diseases in industrialized nations.[71, 72]

However, this pattern of distribution for NTM disease may soon change as emerging

evidence from the Sub-Saharan Africa shows that NTMs are increasingly being isolated

among HIV positive and negative TB cases.[73-75] A prospective evaluation of a cohort

of 721 HIV positive patients in Abidjan, Cote d’Ivoire, Sub-Saharan Africa found a rate

of NTM infection 9.7 times higher among patients with baseline CD4 cell counts less than

100 cells/mm 3 compared to patients with CD4 cell counts above 100 cells/mm 3.[76]

With molecular diagnostic capacities for rapid identification of mycobacterial

species (routinely used in industrialized nations) becoming increasingly available in Sub-

Saharan Africa, the frequency of isolating pathogenic NTM is expected to increase given

the very high prevalence of coexisting HIV infection.

Clinically relevant pulmonary NTM infections are prevalent among HIV-infected subjects in particular.[77-80] Although there is conflicting evidence on the effect of HIV infection on alveolar macrophages,[81-83] there is emerging evidence of an interplay between environment and genetic factors in the facilitation of susceptibility to mycobacterial infection in HIV patients.[84] Occupational exposure to dust in HIV positive subjects accelerates the risk of pulmonary infection by tuberculosis and

NTM .[85, 86]

Environmental acquisition of M. bovis is higher among livestock farming communities where mycobacteria shed by infected animals is inhaled, or unpasteurized

milk of the infectious animal is consumed.[87-89] In most African countries including

Nigeria, milk is consumed either raw or fermented without pasteurization. Furthermore,

the practice of pooling milk together from different animals increases the risk of exposure

to milk of infected cattle.[90, 91] M.bovis is estimated to be responsible for 2.1% of all

cases of pulmonary TB with prevalence as high as 15% reported from northern

Nigeria.[64, 92, 93] Given the 30% prevalence of HIV co-infection with M. tuberculo sis

in Nigeria, there is a growing concern that NTM, M. bovis and other mycobacterial

infections could be underdiagnosed causes of pulmonary TB in HIV co-infected

persons.[1] This study reports the prevalence of environmentally acquired mycobacterial

infections among TB suspected cases with or without associated HIV infection and the

observed likely influence of the annual Harmattan dust season on the pattern of

occurrence of pulmonary mycobacterial infections in Nigeria.

C. Methods

Participants were enrolled into this cross-sectional study from two TB clinics in the state of Kaduna, Nigeria: the National TB and Leprosy Training Center (NTBLTC),

Zaria, from August 2010 through August 2011, and at the Barau Dikko Hospital (BDH), in Kaduna City, from December 2010 through August 2011. Participants aged 18 years or older who agreed to take the routine HIV test and provided informed consent to participate in the study were asked to respond to an itemized survey instrument. Only cases with previously unknown HIV status were enrolled to avoid over-representation of

HIV cases if suspected TB cases with known HIV positive status were enrolled. Details of participants’ enrollment into this study, including the ethical challenges, were described in a previous article [94]. The protocol for this study was reviewed and approved by the University of Maryland’s Institutional Review Board and the Nigerian

Health Research Ethics Committee.

Smear Microscopy

Three sputum samples were received from each participant in the following sequence: (1) a supervised ‘on the spot’ sample (first day); (2) an unsupervised early morning home-collected sputum sample; and (3) a supervised ‘on the spot’ sample upon return to the clinic the next day. Samples were registered and assigned accession numbers. Smears of size 1 x 2cm were made on a new grease-free slide and allowed to air dry. The air dried smears were then fixed by gently passing them over a flame 2-3 times.

The smears were then stained with Ziehl- Neelsen (ZN) technique as follows: 1% strong carbol fuchsin was applied to the slides and heated with a Bunsen flame intermittently 3 times and allowed to stain for 15 min. The stain was washed off with gentle tap water and

3% acid alcohol was applied on the smears for 3 minutes. Methylene blue (0.3%) was

then added for up to 1 minute and washed off with tap water. Positive and negative

controls were included in the staining process. Examination of the slides was done under

x100 microscope magnification. For the purpose of this study, smears were graded as

positive when five or more bacilli were detected in at least 100 microscopic fields.

Mycobacterial Culture

Of the three sputum samples collected, only the early morning home-collected

smear sample was cultured, given that this sample is likely to yield the highest

concentration of tubercle bacilli, and is the least likely to be contaminated with other

bacteria as it was not manipulated to provide the smears used in the routine pre-

enrollment ZN staining and microscopic examination. The early morning sputum

samples were transferred to 50 ml falcon tubes in a class II biosafety cabinet. They were

cultured in liquid Mycobacterium Growth Indicator Tubes (MGIT) as previously

described.[60] Briefly, the samples are first treated with BD Mycoprep TM (Beckton

Dickinson Diagnostic Systems, Sparks, Maryland, USA) which consists of 1% N-acetyl-

L-cysteine (NALC), 4% sodium hydroxide and 2.9% sodium citrate. An equal amount of

Mycoprep was added to the sputum, homogenized and allowed to act for 15 minutes.

Phosphate buffered saline was added to stop the decontamination reaction. This was then centrifuged under refrigerated conditions at 3000 rpm for 15 minutes. The supernatant was discarded and the pellet was reconstituted with 2.5 ml phosphate-buffered saline.

Supplemented 7 ml MGIT tubes were inoculated with 0.8 mls of the pellet and incubated in the automated BACTEC MGIT 960™ machine (Becton Dickinson Diagnostic

Instrument Systems), which monitors growth.

. Samples that failed to show any growth after 42 days of incubation in the

machine were removed and classified as negatives. Samples with positive growth were

removed from the machine and inoculated on blood agar to check for non-mycobacterial

contamination. Then, a Ziehl-Neelsen (ZN) stain was performed to check for the presence

of tuberculous acid fast bacilli (AFB). Samples with non-mycobacterial growth in the

absence of AFB by ZN stain were considered contaminated and excluded from the study.

Samples with or without bacterial growth and a positive AFB by ZN stain were

considered mycobacterium culture positive and further tested with a TB antigen rapid

assay (SD-Bioline Ag MPT64 Rapid™ assay; Standard Diagnostics, Kyonggi-do, Korea)

for the identification of MTB complex organisms (M. tuberculosis, M. bovis and M. africanum). Samples without non-mycobacterial growth but negative by ZN staining were further incubated for a maximum of 42 days - and checked by ZN staining every five days, if still negative after the maximum incubation period, they were classified as negative.

Samples were considered positive for MTB complex if, in addition to the positive growth on the MGIT, they showed the presence of AFB on ZN stain and tested positive on the SD-Bioline. The culture confirmed MTB complex positives were then characterized with Genotype MTBC test (Hain Lifescience, Nehren, Germany) to further identify the different MTB complex organisms ( M. tuberculosis, M. bovis and M. africanum ). Samples positive for AFB on the ZN stain and negative for MTB complex on

the SD-Bioline were identified as potential NTMs . Those samples were further sub-

cultured on the Lowenstein Jensen solid culture medium for a maximum of 42 days. A

positive growth on the LJ solid culture medium was characterized with the GenoType

Mycobacterium CM/AS assay (Hain Lifescience, Nehren, Germany). If any of these samples tested negative on the CM/AS assay, the culture growth was considered acid fast positive unspecified (UNS). Sputum samples that were negative on LJ medium were considered as contaminants. The summary of the testing algorithm used for the detection and characterization of mycobacterial isolates in the study is provided in Figure 1 .

Genotype MTBC; CM and AS Assays

An aliquot of 1000 ul from the positive mycobacterium positive growth liquid medium (MGIT) or LJ medium was added into a 2 ml tube. The mycobacteria were pelleted by spinning for 15 minutes in an aerosol tight micro centrifuge at 10,000 RPM.

The supernatants were discarded and the sediments re-suspended in 100 ul of molecular grade water by spiral motion. The mycobacterial suspension was incubated for 20 minutes in a water bath at 95 0C followed by incubation in an ultrasonic bath for 15

minutes. This was followed by centrifugation at full speed (10,000 RPM) for 5 minutes

and the supernatant containing the DNA was aliquot and stored at -20 0C if not amplified

immediately.

Amplification mix containing 35 ul of Primers Nucleotide Mix, 5ul of 10X PCR

buffer, 2 ul of 25 nm MgCl 2, 0.2 ul of HotStar Taq and 3 ul water were prepared per sample. Five microliters of the DNA were added to this amplification mix and amplified as follows: a) 1 cycle at 95 0C for 15 minutes b) 10 cycles at 95 0C for 30 seconds then at

58 0C for 2 minutes, c) 20 cycles at 95 0C for 25 seconds, then at 40 0C for 40 seconds and d) 1 cycle at 70 0C for 8 minutes. Amplification products were stored at +4 to -20 0C if not detected immediately. Detection of the amplicons was done by hybridization technique using the twincubator hybridization tray. Some 20 ul of the amplicons was added to 20 ul

of the denaturation solution in a tray and incubated for 5 minutes. Then 1 ml of pre-

warmed hybridization buffer was added to the tray and rocked for homogeneity. Labeled

DNA strips were immersed into each well of a tray placed on the shaking twincubator.

The incubation lasted for 30 minutes at 45 0C and was followed by aspiration of

the buffer and the addition of 1 ml of stringent wash solution. It was incubated again on

the twincubator for 15 minutes at 45 0C. The stringent wash solution was completely removed and 1 ml of rinse solution was added for 1 minute. Some 1 ml of diluted conjugate was added to each strip and incubated for 30 minutes on the twincubator. The solution was removed and washed twice for 1 minute with 1 ml rinse solution and 1 ml of distilled water on the twincubator. The solutions were poured out each time and 1 ml of diluted substrate was added to each strip which was then incubated while protected from light without shaking for 3-20 minutes for color development. Reactions were stopped by briefly rinsing twice with distilled water. The strips were removed with tweezers, dried, pasted and interpreted with the Hain interpretation chart as per the manufacturer’s instructions.

Data Analysis

Missing values and outliers were checked for errors and were corrected as appropriate. Frequencies and proportions of demographic and selected covariates were evaluated. Associations between categorical variables were assessed using Pearson’s

Chi-square or Fisher’s exact test where appropriate. Unadjusted odds ratio for the association between mycobacterial infections and HIV were calculated. Covariates including age, gender, ethnic group, body mass index and educational level among others were evaluated and those found to be associated with any of the mycobacterial infection

groups and HIV were tested for confounding and effect modification in stratified analyses. Potential confounders were evaluated as follows: each covariate was tested, one at a time, by comparing the crude odds ratio (between HIV status and mycobacterial group) to the Mantel Haenszel estimate stratified by the covariate. A difference of 10% or more between the adjusted Mantel Haenszel and the unadjusted odds ratio was considered potential confounding.

Multiple logistic regressions were then used to adjust for confounding of the associations between the different mycobacterial groups and HIV in which odds ratios and 95% confidence intervals were reported. Each potential confounder identified in the first step was added one-by-one to the simple model consisting of the outcome of interest and the main predictor (HIV). A covariate was retained in the model if it was significant

Likewise, effect modification was evaluated by identifying the potential effect modifiers using the Breslow-Day test for homogeneity of the odds ratios over different strata of the covariate. A p-value of 0.05 or less was considered significant modification.

For those covariates identified as potential effect modifiers, effect modification was tested by comparing two logistic regression models – one model with the interaction term versus the model without the interaction term. The log likelihood ratio test was used to assess model goodness of fit improved with the interaction term. Statistical analysis

software (SAS Institute, Inc., Cary, North Carolina) version 9.2 was used for the analysis.

Two-sided P-values of 0.05 or less were considered statistically significant.

D. Results

Of 1,657 patients presenting to both clinics, a total of 1,603 (96.7%) subjects

provided informed consent to participate in the study: 1,391 (87%) enrolled at the

NTBLTC, Zaria, and 212 (13%) were from BDH, in Kaduna City. The mean age in years

and standard deviation (SD) of the study sample was 37.0 (13.8) with 43.8% females, and

their mean body mass index (SD) was 19.2 (4.6). About 52% of the study population

were underweight (BMI<18.5) according to the international WHO criteria for BMI. The

majority, 963 (60%) were of the Hausa ethnic group, while the rest were from other

ethnic groups including the Fulani (18.1%), Yoruba (2.9%), Igbo (17.8%) and others

(1.1%). About 910 (57%) patients were 35 years or younger. Some 437 (27%) engaged in

livestock farming and up to 1,418 (88%) consumed locally produced milk; 309 (22%) on

a daily basis. A summary of selected demographic and laboratory characteristics of the

study population is provided in Table 4.

There were 444 (28%) cases of mycobacterial infections with MGIT positive

samples, of which 375 (84%) were confirmed as MTB complex isolates by the presence

of MPT 64 antigen and positive on SD-bioline test, and on further speciation using

GenoType MTBC test. Among the 375 MTB complex cases, 354 cases (94.4%) were M.

tuberculosis; 20 cases (5.3%) were M. africanum ; and one case (0.3%) was M. bovis

(Table 4).

There were 91 (20%) patients negative on SD-bioline but positive on Lowenstein

Jensen solid culture medium of which 69 (16%) were identified as NTM isolates by the

GenoType Mycobacterium CM/AS assay. The remaining 22 were acid fast positive, culture positive but not identified, i.e. unspecified (UNS) by the same assay. Of the 69

identified NTM cases, 23 (33.3%) could not be classified by the GenoType

Mycobacterium CM/AS assay and considered uncharacterized, while the remainder were

divided between M. intracellulare 21 (30.4%), M. abscessus 8 (11.6%), M. fortuitum 4

(5.8%), M. sucrofulaceum 3 (4.3%), M. gordonae 4 (5.8%), M. malmoense 2 (2.9%), M.

kansasii 1 (1.4%), M. interjectum 1 (1.4%), M. peregrinum 1 (1.4%) and M. xenopi 1

(1.4%) species (Table 4).

There were 234 (14.6%) contaminated samples and 903 samples of clinically

symptomatic patients without any detectable isolates by laboratory testing.

Among all 444 culture positive cases, 127 (29%) were HIV co-infected, with 101

(27%) among 375 MTB complex cases, and 26 (38%) of 69 NTM cases. Another 9 (41%)

of 22 of UNS cases, 57 (24%) of 234 of contaminated cases, and 185 (20%) of 903

culture-negative cases were also HIV co-infected, bringing the total to 378 cases.

Pulmonary Infections and HIV

The unadjusted outcomes of group comparisons between cases of M. tuberculosis, M.

Africanum, NTM and UNS to cases without any detectable isolates in the culture as regards HIV infection and several covariates is provided in Table 5 . M. bovis was not

included in this analysis or in Table 5 given that there was only one case and could not be

compared to other groups. HIV infection was significantly associated with M. tuberculosis (OR=1.4; 95%CI=1.1-1.9) , NTM (OR=2.3; 95%CI=1.4-3.9) , and UNS

(OR=2.9; 95%CI=1.2-6.1) , but not with M. africanum (OR=1.0; 95%CI=0.3-2.9). One of

the more striking findings is that the standard of care pre-enrollment smear test could

only detect 60.5% of M. tuberculosis cases, 70% of M. africanum cases, and falsely

classified 7.3% of NTM cases, 18.2% of UNS cases, and 4.7% of isolate-negative cases

as positive for pulmonary tuberculosis (Table 5).

After controlling for age, cigarette smoking and site the association between HIV

and M. tuberculosis was marginally significant (OR=1.3; 95%CI=1.0-1.8; p-

value=0.0544). The association between HIV and NTM however, remain strong after

controlling for age and Harmattan dust storm season (OR=2.2; 95%CI=1.3-3.8; p-

value=0.0027). The association between HIV and the unspecified isolates (UNS) was not

statistically significant after controlling for the Harmattan dust storm season (OR=2.1;

95%CI=0.9-5.2; p-value=0.0936) (Table 6 ).

An unadjusted analysis within patients with pulmonary mycobacterial infection (

NTM versus MTB complex cases ), cases with NTM differed significantly from MTB

complex cases as regards positive pre-enrollment smear test, age, and seasonal changes,

and only marginally with respect to HIV infection (OR= 1.6, 95%CI=, 0.9-2.8,

p=0.0694 ) and BMI (OR= 0.6, 95%CI=, 0.4-1.0, p=0.0705 ). NTM cases were more

likely to be older with lower BMI, test negative on standard of care smear and present

with symptoms in the Harmattan season (Table 7 ). In an adjusted analysis involving

covariates of age, pre-enrollment smear test, BMI, history of diabetes mellitus and

season, the odds of HIV infection was 2.5 times higher among NTM cases compared to

M. tuberculosis cases (95%CI, 1.4-4.5). The adjusted differences between NTM and M.

tuberculosis with respect to the other covariates are shown in Table 8.

Characteristics of NTM Infections

In addition to the high burden of HIV infection among NTM infected cases with

26 (38%) of them HIV positive compared to 101 (27%) cases co-infected with HIV among the MTB complex infected group; the majority of NTM cases were males (n=39,

57%), but HIV co-infection was higher among females (n=18, 69%) versus males (n=8,

31%). Of the 69 NTM cases identified over a period of twelve calendar months, 29 (42%) occurred between the months of January and February (peak of Harmattan dust season in

Nigeria). A plot of the proportion of monthly detected cases over a year period shows the proportion of NTM cases approaching that of MTB cases only at the peak of the

Harmattan season with a unique pattern similar to a single source outbreak with a peak at

January-February (Figure 4). The MTB cases seem to vary throughout the year without any specific pattern across the 12 calendar months.

E. Discussion

This study shows that pulmonary tuberculosis in our study sites in Nigeria are predominantly caused by M. tuberculosis with a few cases caused by M. africanum , and only one case by M. bovis , while an unexpectedly high number were caused by NTM infection that seemed to coincide with the brief period of intense Harmattan dust season.

Harmattan is a West African trade wind that occurs during the winter and is characterized by heavy amounts of dust in the air, low humidity and reduced visibility.[95] This wind regularly drives the outbreak of highly contagious bacterial meningitis (meningococcal meningitis) in the West African sub-region, the largest in recent times being the 1996 epidemic.[96, 97]

The intense air pollution and the irritant effect of heavy doses of coarse and fine particulates on the mucosa as well as alveolar macrophages could aid invasion and infection by pathogens suspended in the dust. Environmentally acquired pulmonary mycobacterial infections frequently occur in individuals with occupational exposures to dust and in HIV-infected subjects.[86, 98, 99] Of all dust types, silica dust, consisting of fine crystalline quartz carries a higher risk of predisposition to infection by opportunistic

Mycobacteria.[100, 101] Quartz (silicon oxide) binds to alveolar macrophages (a key component of pulmonary host defense system) to trigger a chain of reactions that destroy the macrophages making the host vulnerable to infection by less virulent pathogens.[102,

103] Silica is abundantly available in the earth’s crust including sand, soil and rocks, from which quartz are generated during mining, drilling and construction related activities.[101]

The coincidence of a high number of NTM cases with the dust season strongly suggests a possible role of the Harmattan dust in the spread of these known environmental pathogens. While it is possible that these cases were previously infected with NTMs and became symptomatic during the dry dusty weather, it is also possible that they acquired the infection during this three-month long season (December to

February). Since our study is a cross-sectional study, it is not possible to determine causality.

A follow up of these patients could reveal a long lasting disease and possibly more adverse outcome depending on the organisms’ response to regular anti-TB drugs.

The presence of TB-like symptoms at the peak of the Harmattan season in patients with negative TB tests on smear microscopy should prompt the suspicion of a possible NTM or even fungal infections especially among HIV positive subjects. Airborne fungi were previously isolated from Harmattan dust in Zaria[104] and cases of pulmonary blastomycosis with TB-like features were reported in the past.[105]

The only one case of M. bovis infection was in an HIV co-infected patient. This patient neither kept nor tended to livestock but occasionally consumed locally produced milk. It will be hard to deduce any association between this single case and HIV. The very low prevalence (0.3%) of M. bovis pulmonary TB identified in our study is in agreement with the findings of similar recently reported studies for the identification of mycobacterial strains in the neighboring west African countries of Ghana, Mali,

Cameroun and Burkina Faso [15, 106-108] which either failed to detect any M. bovis or identified it in very low proportion (3% from Ghana, 0.8% from Mali, 0.2% from

Cameroun and none from Burkina Faso) but identified high proportions of M. Africanum

(9 to 28% among the mycobacterial strains).

Our findings however, contradicted previous studies from Nigeria on similar population

groups (facility based cases of TB) which reported a prevalence of M. bovis in the range

of 10-15% [44, 64] The difference may be explained by the fact that some of these

studies relied solely on the organism’s morphology and biochemical reactions to

distinguish M. bovis from other isolates of MTB complex. However, M. bovis and M.

africanum share the same dysgonic growth pattern on culture and the two may be hard to distinguish in biochemical reactions.[16, 109] It is possible that M. africanum strains

were mischaracterized and reported as M. bovis given the prevalence of M. africanum

among cases of pulmonary tuberculosis. Another recently concluded study within Nigeria

in a center that serves patients from both northern and southern regions involving 100

samples of smear positive, culture positive cases of pulmonary tuberculosis in Nigeria

failed to identify any M. bovis among the isolates characterized by a combination of conventional methods and molecular line probe assays.[110] Mycobacterium bovis may still exist in the livestock in Nigeria and establish disease when transmitted through routes other than the respiratory system. Sampling suspected cases of abdominal TB might have yielded a different outcome.

The burden of M. africanum in this population is lower compared to a previous

report on this species from southern Nigeria that found 13% of pulmonary TB due to M.

africanum . [64] This disparity may reflect a regional difference in the prevalence of M.

africanum within Nigeria since to our knowledge; our study is the first to report the

burden of M. africanum from the northern Nigeria region.

Our study is a cross-sectional study and hence limited in its ability to evaluate causal and temporal associations between HIV, dust exposures and the outcomes of interest. The findings may only represent facility-based TB patients whereas others who do not visit these healthcare facilities may differ. However, bias in this study was minimized by enrolling a large majority of suspected cases (96.7%) visiting the facilities for the first time. The comparison group consisted of patients from within the study population who did not have mycobacterial infection. However, this group was still sick as indicated by the high frequency (more than 50%) of underweight individuals among them as classified by the WHO BMI cut-off of <18.5. Comparison of covariates like body mass index between those infected with mycobacteria and the uninfected comparison group may underestimate the difference, specially if compared to the normal Nigerian population where the proportion of underweight individuals is less than 8%.[111]

The use of liquid broth medium for the growth of MTB complex isolates is more sensitive, and unlike the solid culture medium allows unhindered growth of a wide range of mycobacteria including M. bovis , has added to the strength of this study.[112-114]

While we opted to culture only one out of the three sputum samples, namely the early morning sample, this is unlikely to have reduced our ability to detect positive cases, given that the early morning sample is known to have the highest concentration of mycobacteria in patients, and it also reduces the likelihood of contamination that can occur as a result of smears taken from the other two samples during the clinic’s standard of care. Furthermore, a single sample liquid broth culture is reported to have a yield comparable to that from combining the three sputum sample and using solid

cultures.[115] The performance of molecular line probe assays (Hain assays) on cultured

specimens increases the sensitivity of isolating the mycobacteria. These assays have

established superiority over the conventional methods for identification and

characterization of Mycobacterial isolates.[116-118]

Finally, while MTB complex strains remain the predominant cause of

pulmonary tuberculosis in this population, M. bovis is rare and M. africanum as previously reported in West Africa is frequent. The high prevalence (16%) with clinical pulmonary TB due to NTM linked to Harmattan dust exposure, and to HIV co-infection

(38%) presents a novel public health challenge for prevention and treatment of these

patients and understanding the efficacy of the standard TB regimens since their responses

to these regimens are known to vary from that by M. tuberculosis . The comparable

ineffectiveness of the standard of care test (pre-enrollment sputum smear microscopy)

where only 60% of M. tuberculosis and 70% of M. africanum were detected and many

non-tuberculous cases were misdiagnosed as false negatives, underscores the need for a

more sensitive and specific TB screening algorithm for effective disease control.

Introduction of molecular detection and screening assays to address the rapid

identification of NTM and drug resistant TB should be a high priority for strengthening

the public health response.

Table 4. Measured demographics and laboratory characteristics stratified by site Characteristics NTBLTC at Zaria BDH at Kaduna n = 1391 (%) n= 212 (%) Age: Mean (SD) 37.0 (13.9) 36.6 (13.5) BMI: Mean (SD) 19.0 (4.6) 20.8 (4.4) Female gender 591 (42.5) 111 (52.6) Education: 8th Grade or less 900 (64.7) 80 (37.9) Married 991 (71.2) 122 (57.8) Hausa or Fulani Ethnic group 1148 (82.5) 104 (49.3) Prevalence of HIV 312 (22.4) 66 (31.3) Prevalence of MTB complex infection 315 (22.6) 60 (28.3)

Mycobacterium tuberculosis 298 (21.4) 56 (26.4)

Mycobacterium africanum 16 (1.2) 4 (1.9)

Mycobacterium bovis 1 (0.0) 0 (0.0) Prevalence of Non-tuberculosis 59 (4.2) 10 (4.7) Mycobacterium (NTM) infection

Mycobacterium intracellulare 16 (1.2) 5 (2.4) Mycobacterium abscessus 8 (0.6) 0 (0.0) Mycobacterium fortuitum 4 (0.3) 0 (0.0) Mycobacterium sucrofulaceum 3 (0.2) 0 (0.0) Mycobacterium gordonae 4 (0.3) 0 (0.0) Mycobacterium malmoense 2 (0.1) 0 (0.0) Mycobacterium kansasii 0 (0.0) 1 (0.5) Mycobacterium interjectum 1 (0.0) 0 (0.0) Mycobacterium peregrinum 1 (0.0) 0 (0.0) Mycobacterium xenopi 1 (0.0) 0 (0.0)

N (%) N (%) 41 (4.5) 41 (4.5) 40 (4.4) N=903 N=903 185 (20.5) 185 (20.5) 428 (47.4) 488 (54.0) 446 (49.4) 222 (24.6) 256 (28.4) 101 (11.2) 152 (16.8) 803 (88.9) 207 (22.9) 723 (80.1.) (80.1.) 723 No Isolates [Reference] [Reference]

P- value 0.0131 0.7084 0.8796 0.0047 0.9125 0.1638 0.3381 0.6427 0.7931 0.9935 0.0013 0.0128 <.0001 OR, 95%CI OR, 95%CI 1.2-6.7 2.9, 0.4-2.0 0.8, 0.4-2.2 0.9, 1.4-12.9 4.3, 0.3-2.5 0.9, 0.1-1.5 0.4, 0.2-1.6 0.6, 0.4-4.6 1.3, 0.4-3.5 1.2, 0.1-7.7 1.0, 1.7-16.7 5.3, 0.1-0.8 0.3, 2.9-17.9 7.2,

UNS N=22 N (%) 9 (40.9) 9 (40.9) 11 (50.0) 17 (77.2) 5 (22.7) 3 (13.6) 15 (68.2) 3 (13.6) 4 (18.2) 1 (4.5) 4 (18.2) 15 (68.2) 15 (68.2) P- value value 0.0009 0.5797 0.0541 0.4438 0.1198 0.1398 0.1398 0.1128 0.3030 0.5128 0.2398 0.3861 <.0001 <.0001

3.9 1.4 1.0 1.3 2.6 1.1 3.3 2.5 2.6 4.7 1.5 5.1 OR, 95%CI 95%CI NTM 2.3, 1.4-2.3, 0.5-0.9, 0.4-0.6, 0.5-0.8, 0.9-1.5, 0.4-0.7, 0.9-1.6, 0.8-1.4, 0.1-0.6, 0.7-1.8, 0.4-0.7, 1.9-3.1, N=69 1.0, 0.6- 1.8 N (%) 2 (3.0) 2 (7.3) 5 15 (2.7) 26 (37.7) 26 (37.7) 30 (43.5) 29 (42.0) 31 (44.9) 23 (33.3) 50 (72.5) 12 (17.4) 59 (85.5) 33 (47.8) 20(29.0) 20(29.0)

P- value value 0.2884 0.9327 0.0288 0.0405 0.4946 0.0291 0.5799 0.7127 0.2771 0.2711 0.7581 0.9542 0.9542 <.0001

M. tuberculosis, M. Africanum, M. Africanum, tuberculosis, M. 2.5 0.9 6.0 3.5 0.9 4.9 3.7 1.5 2.5 OR, 10.0 10.0 53.0, 19.3- 145.7 145.7

95%CI 95%CI 2.5, 1.0 1.0, 0.4- 0.3, 0.1- 1.4, 0.5- 0.4, 0.2- 1.4, 0.4- 1.2, 0.4- 2.2, 0.5- 0.5, 0.2- 0.8, 0.3- 1.0, 0.3-1.0, 2.9 0.2-0.6, 1.5 ectable ectable N=20 s (reference group) with respect to HIV to HIV respect with group) (reference s M .africanum io; CI = confidence interval interval =confidence CI io; N (%) N 4 (20.0) (20.0) 4 (35.0) 7 (25.0) 5 (45.0) 9 (35.0) 7 (15.0) 3 (20.0) 4 (10.0) 2 (20.0) 4 11 (55.0) 11 (55.0) 12 (60.0) 14 (70.0) 16 (80.0) = Non-tuberculosis Mycobacterium; No No Isolates: Mycobacterium; = Non-tuberculosis NTM P- value value 0.0120 0.3042 0.0389 0.1881 0.0032 0.7413 0.0292 <.0001 <.0001 <.0001 <.0001 <.0001 <.0001 0.0220

1.9 2.7 0.7 1.5 1.0 1.1 2.4 2.6 1.7 0.9 1.8 OR, 51.4 51.4 34.7, 23.4- 95%CI 1.4, 1.1-1.4, 1.6-2.1, 0.4-0.5, 0.9-1.2, 0.6-0.7, 0.6-0.8, 1.2-1.7, 1.6-2.0, 0.5-0.9, 0.5-0.7, 1.0-1.4, 0.6, 0.4- 0.7 N=354 M. tuberculosis M. N (%) N (%) 15 (4.2) 15 (4.2) 96 (27.1) 96 (27.1) 98 (27.7) 79 (22.3) 61 (17.2) 119 (33.6) 119 (33.6) 251 (70.9) 121 (34.2) 271 (76.6) 101 (28.5) 214 (60.5) 298 (84.2) 102 (28.8) . Unadjusted analysis comparing the groups with det with thegroups comparing analysis Unadjusted . 5

and UNS to the group without any detectable isolate detectable any without the group to andUNS No detectable mycobacterial isolates; OR = Odds rat =Odds OR isolates; mycobacterial No detectable Table Characteristic HIV infection SexFemale <=35 Age years BMI<=19.2 Farming Keeping livestock Majority Ethnic group Alcohol intake Cigarette Smoking of History Diabetes mellitus. Positive pre- enrollment sputum smear Site(Zaria) A Harmattan Season isolates; (unspecified) Unidentified UNS= infection and measured covariates covariates andmeasured infection NTM

Table 6: Multiple logistic regressions analyses comparing the groups with detectable M. tuberculosis, M. Africanum, NTM and UNS to the group without any detectable isolates (reference group) with respect to HIV infection and measured covariates Characteristic M. tuberculosis M. africanum NTM UNS

OR, P- OR, P- OR, P- OR, P-value 95%CI value 95%CI value 95%CI value 95%CI HIV Infection No 1 Yes 1.3, 1.0- 0.0544 0.9, 0.3- 0.8790 2.2, 1.3- 0.0027 2.1, 0.9- 0.0936 1.8 2.8 3.8 5.2 Age Age > 35 years 1 Age < 35 2.1, 1.6- <.0001 1.1, 0.4- 0.8715 0.5, 0.3- 0.0195 * * years 2.8 2.6 0.9

Cigarette Smoking No 1 Yes 2.1, 1.5- <.0001 1.3, 0.4- 0.6895 * * * * 2.8 3.8 Site Kaduna 1 Zaria 0.7, 0.5- 0.0374 0.4, 0.1- 0.1708 * * * * 1.0 1.4

Harmattan season Other period 1 Dec-Jan * * 0.7, 0.2- 0.5405 3.1, 1.8- <.0001 6.2, 2.4- <.0001 2.2 5.2 16.0 1=Reference category: * = Variables not in the best fitted model; OR= Odds ration; CI = confidence interval UNS= Unidentified (unspecified) isolates; NTM = Non-tuberculosis Mycobacterium

Table 7. Unadjusted comparison between NTM and MTB complex infections with respect to HIV Infection and measured covariates Characteristics NTM MTB Complex Odds Ratio P-value N=69 N=375 95%CI N (%) N (%) HI V infection 26 (37.8) 101 (26.9) 1.6 (0.9-2.8) 0.0694 Female sex 30 (43.5) 126 (33.6) 1.5 (0.9-2.6) 0.1144 Age < 35 years 29 (42.0) 262 (69.9) 0.3 (0.2-0.5) <.0001 BMI < 19.2 31 (44.9) 126 (33.6) 1.6 (0.9-2.7) 0.0705 Farming 23 (33.3) 107 (28.5) 1.3 (0.7-2.2) 0.4207 Keep animals 20 (29.0) 86 (22.9) 1.4 (0.8-2.4) 0.2787 Majority Ethnic 50 (72.5) 284 (75.7) 0.8 (0.5-1.5) 0.5632 Alcohol intake 12 (17.4) 65 (17.4) 1.0 (0.5-2.0) 0.9981 Cigarette smoking 15 (21.7) 105 (28.0) 0.7 (0.4-1.3) 0.2818 History of diabetes mellitus 2 (3.0) 17 (4.6) 0.6 (0.1-2.8) 0.2448 Sex with HIV+ 1 (1.5) 19 (5.2) 0.3 (0.0-2.1) 0.1874 Pre-enrollment smear test + 5 (7.3) 229 (61.1) 0.0 (0.0-0.1) <.0001 Site A (Zaria) 59 (85.5) 315 (84.0) 1.1 (0.5-2.3) 0.7522 Harmattan dust season 33 (47.8) 107 (28.5) 2.3 (1.4-3.9) 0.0015

NTM : Non tuberculous mycobacterium MTB complex: Mycobacterial tuberculosis complex, i.e. M. tuberculosis, M. Africanum and M. bovis

Table 8. Adjusted analysis comparing NTM and M. tuberculosis among cases of pulmonary

mycobacterial infection

Characteristics Odds ratio 95% CI p-value

HIV infection 2.2 (1.1-4.3) 0.0241

Age <= 35 years 0.3 (0.1-0.5) <.0001

Pre-enrollment smear test 0.5 (0.0-0.1) <.0001

History of Diabetes Mellitus 0.1 (0.0-1.1) 0.0648

Harmattan dust storm season 2.2 (1.2-3.8) 0.0280

NTM = Non-tuberculosis Mycobacterium

Figure 4. Pattern of occurrence of tuberculosis and non-tuberculous mycobacterial infections over a period of 12 calendar months in Nigeria: solid lines indicate monthly proportions of NTM and MTB among all subjects screened while dotted lines are the 95% confidence interval of the proportion

IV. HIV Co-infection and Resistance to Isoniazid and Rifampicin in Cases of Pulmonary Tuberculosis

G.G. Aliyu, S.S. El-Kamary, N. Ezati, I. Mosunmola, J. Obasanya, A. Abimiku, L. Hungerford, C. Brown, K. Tracy and W. Blattner

A. Abstract

Background : HIV and tuberculosis adversely affect the course of each other and often alter the drug treatment choice when the two are treated simultaneously. Isoniazid and rifampicin are powerful drugs in the treatment of tuberculosis. We investigated the association between HIV co-infection and pattern of resistance to rifampicin and isoniazid among cases of pulmonary tuberculosis.

Methods : Genotype MTBDR plus assay (Hain Lifescience GmbH, Nehren, Germany) was used to detect resistance to rifampicin and isoniazid from cultured specimen of 375 cases of pulmonary tuberculosis. Records of patient’s HIV status determined by a serial rapid assay algorithm consisting of Trinity Biotech Unigold and Abbott Determine were obtained. Fisher’s exact tests in addition to multivariable logistic regression were used for the analysis.

Results : One-hundred-one (27%) of the pulmonary tuberculosis cases had co-infection with

HIV. In 23 cases (6.2%), the mycobacterium was resistant to at least one drug. Among the resistant cases; 13 (56%) were single resistance to isoniazid, 5 (22%) were rifampicin alone resistance while the remaining 5 (22%) had resistance to both (MDR-TB). More than 50% of all resistant cases were treatment naïve. After controlling for prior TB treatment, cases resistant to at least one drug were more likely to have co-infection with HIV compared to cases without any resistance (OR=3.6, 1.5-8.8; p=0.0039). Similarly, cases with single resistance to isoniazid were

3 times more likely to have HIV co-infection than cases without (OR= 3.3, 1.0-10.7; p=0.0472).

Conclusion : TB cases with HIV co-infection show increased tendency to have resistance to least one of the two first-line anti-TB drugs. Isoniazid resistant tuberculosis is more common in HIV- infected cases. Increased adherence to directly observed therapy short course for tuberculosis in this population is recommended, and drug resistance should be suspected if treatment fails.

Introduction of molecular detection and screening assays to detect drug resistant TB should be a high priority for strengthening the public health response.

B. Introduction

Isoniazid and rifampicin are critical first line drugs in the treatment of tuberculosis

(TB).[119] Isoniazid is recommended by the WHO for use solely or in combination with anti- retroviral therapy (ART) with intensified case finding for the prevention of TB in HIV-infected subjects.[120-122] Combined resistance to the first-line drugs isoniazid and rifampicin (MDR-

TB), in most cases requires treatment with second-line drugs and carries a higher risk of death.[123, 124] HIV-infected TB cases are twice as likely to die compared to those who are not

HIV-infected,[35] and in the presence of MDR-TB, the risk of death is even higher.[125, 126] In a study of TB drug resistance and mortality in Peru involving 287 TB patients, 17 of the 31 HIV-

MDR-TB patients died before the conventional drug susceptibility test identified their MDR-TB status [40].

Isoniazid only resistance is reported by the World Health Organization (WHO) in significant proportions in 93 settings in 82 countries surveyed from 2002 to 2006; however, in

80% of the settings, rifampicin alone resistance was reported to be less than 1%.[127] The global population weighted means of resistance to isoniazid and rifampicin among new cases of

TB in the WHO report were 10.3 and 3.7 respectively while the mean MDR-TB among new cases was 2.9. The tendency to resist antibiotics increased with treatment failure and treating isoniazid resistance with the standard regimen were observed to produce poor outcomes in previous reports [128].

The association between HIV and MDR-TB or resistance to any TB drug has not been sufficiently examined in most of the settings surveyed by the WHO because of the low number of HIV positive cases with resistance to treatment in the majority of the settings. However, for the few settings with enough cases, HIV infection was significantly associated with MDR-TB

and any anti-TB drug resistance [127]. A comprehensive review of the literature on the association between HIV and MDR-TB indicated that some studies from North America and

Europe found associations between HIV and MDR-TB.[129] In Africa, few studies looked at the possible link between MDR-TB and HIV but failed to definitively demonstrate an association. In some, there were fewer HIV-infected cases with resistant organism while in others the associations were not statistically significant.[42, 130-132] Studies from Nigeria reported MDR-

TB among cases of TB receiving treatment with first line tuberculosis drugs,[42, 133, 134] but none of them identified HIV as a risk factor despite the relatively high prevalence of MDR-TB identified among previously treated patients in all the studies. The dual epidemics of HIV and

TB in Sub-Saharan African and the high mortality among HIV patients attributable to TB in this region underscores the importance of monitoring the pattern of resistance to TB drugs and its association with HIV. [135-137] In this study we examined the association between HIV infection and resistance to the most powerful first line drugs: isoniazid and rifampicin among facility based cases of pulmonary tuberculosis.

C. Methods

Participants were enrolled into this cross-sectional study from two TB clinics in the state

of Kaduna, Nigeria: the National TB and Leprosy Training Center (NTBLTC), Zaria, from

August 2010 through August 2011, and at the Barau Dikko Hospital (BDH), in Kaduna City,

from December 2010 through August 2011. The early morning sputum samples were cultured in

liquid Mycobacterium Growth Indicator Tubes (MGIT) in the automated BACTEC MGIT 960™

machine (Becton Dickinson Diagnostic Instrument Systems), which monitors growth. Samples

that failed to show any growth after 42 days of incubation in the machine were removed and

classified as negatives.

Samples with positive growth were removed from the machine and inoculated on blood

agar to check for non-mycobacterial contamination. Then, a Ziehl-Neelsen (ZN) stain was

performed to check for the presence of acid fast bacilli (AFB). Samples with non-mycobacterial

growth that were negative of AFB by ZN stain were considered contaminated and excluded from

the study. Samples without non-mycobacterial growth but negative of AFB by ZN stain were

further incubated for a maximum of 42 days and checked by ZN stain every five days, and if still

negative after the maximum incubation period, they were classified as negative. Samples with or

without the non-mycobacterial growth that were positive of AFB by ZN stain were considered

mycobacterium culture positive and further tested with a TB antigen rapid assay (SD-Bioline Ag

MPT64 Rapid™ assay; Standard Diagnostics, Kyonggi-do, Korea) for the identification of MTB

complex organisms ( notably: M. tuberculosis, M. bovis and M. africanum).

Samples were considered positive for MTB complex if, in addition to the positive growth on the MGIT, they showed the presence of AFB on ZN stain and tested positive on the SD-

Bioline. The culture confirmed MTB complex positives were then characterized with Genotype

MTBC test (Hain Lifescience, Nehren, Germany) to further identify the different MTB complex organisms.

Detection of Resistance to Isoniazid and Rifampicin

Culture confirmed MTB complex positive sputum specimens were assayed for evidence of resistance to isoniazid and rifampicin using a Genotype MTBDR plus (Hain Lifescience,

Nehren, Germany). This line probe assay was performed according to the manufacturer’s instructions and is briefly described as follows: an aliquot of 1000 ul of bacteria grown on liquid medium (MGIT 960) was added into a 2 ml tube. The mycobacteria were pelleted by spinning for 15 minutes in an aerosol tight micro centrifuge at 10,000 RPM. The supernatants were discarded and the sediments re-suspended in 100 ul of molecular grade water by spiral motion.

The mycobacterial suspension was incubated for 20 minutes in a water bath for 95 0C followed by incubation in an ultrasonic bath for 15 minutes. This was followed by centrifugation at full speed (10,000 RPM) for 5 minutes and the supernatant containing the DNA was aliquot and stored at -20 0C if not amplified immediately.

Amplification mix containing 35ul of Primers Nucleotide Mix, 5ul of 10X PCR buffer, 2 ul of 25 nM MgCl 2, 0.2 ul of HotStar Taq and 3 ul water were prepared per sample. Five microliters of the DNA were added to this amplification mix and amplified as follows; a) 1 cycle at 950C for 15 minutes b) 10 cycles at 95 0C for 30 seconds then at 58 0C for 2 minutes, c) 20 cycles at 95 0C for 25 seconds, then at 40 0C for 40 seconds and d) 1 cycle at 70 0C for 8 minutes

Amplification products were stored at +4 to -20 0C if not detected immediately. Detection of the amplicons was done by hybridization technique using the twincubator hybridization tray. Some

20 ul of the amplicons was added to 20 ul of the denaturation solution in a tray and incubated for

5 minutes. Then 1 ml of pre-warmed hybridization buffer was added to the tray and rocked for

homogeneity. Labeled DNA strips were immersed into each well of a tray placed on the shaking twincubator.

The incubation lasted for 30 minutes at 45 0C and was followed by aspiration of the buffer and the addition of 1 ml of stringent wash solution. It was incubated again on the twincubator for

15 minutes at 45 0C. The stringent wash solution was completely removed and 1ml of rinse solution was added for 1 minute. Some 1 ml of diluted conjugate was added to each strip and incubated for 30 minutes on the twincubator. The solution was removed and washed twice for 1 minute with 1 ml rinse solution, 1 ml of distilled water on the twincubator. The solutions were poured out each time and 1ml of diluted substrate was added to each strip and incubated while protected from light without shaking for 3-20 minutes for color development. Reactions were stopped by briefly rinsing twice with distilled water.

The strips were removed with tweezers, dried, pasted and interpreted based on the presence or absence of mutations that characterize resistance to either isoniazid or rifampicin or both, as per the manufacturer’s instructions. The assay strip had a total of 27 reaction zones.

Twenty-one zones probed mutations and the remaining 6 were control probes for verification of the assay procedures. The control probes consisted of a conjugate control, and amplification control, an MTB complex-specific control, an rpoB amplification control, a katG amplification control, and an inhA amplification control. Rifampicin resistance was marked by the rpoB gene, while isoniazid resistance was marked by the katG and inhA genes.

Resistance to Rifampicin was identified by the absence of at least one of the wild-type bands or the presence of bands indicating a mutation in rpoB gene. Similarly, the absence of at least one of the wild-type bands or the presence of bands indicating a mutation in either katG or inhA genes or both identified resistance to isoniazid. Joint occurrence of characteristic features

for resistance to both drugs indicated the presence of MDR-TB. A sample in which all the wild- type probes of a gene were present and there was no detectable mutation within the region examined, the sample was considered sensitive to the respective drug. A valid result had bands in all six control zones, otherwise, the result was considered invalid.

Data Management and Analyses

The purpose of this analysis was to determine the frequencies and pattern of resistance to isoniazid and rifampicin and compare the proportion of cases with different forms of resistance among HIV positive cases to HIV negative cases. Baseline demographics and selected covariates were examined for distributions and associations of resistance to isoniazid alone, rifampicin alone, or both (MDR-TB) or any resistance as well as associations with HIV. These associations were examined using Fisher’s exact test in which cases with isoniazid only resistance; rifampicin only resistance; MDR-TB resistance and any resistance were compared with cases without any resistance. Potential confounders and effect modifiers were checked in stratified analyses. We evaluated each covariate comparing the crude odds ratio (between HIV status and resistance group) to the Mantel Haenszel estimate stratified by the covariate. A difference of 10% or more between the adjusted Mantel Haenszel and the unadjusted odds ratio was considered potential confounding. Multiple logistic regressions were used to adjust for confounding of the associations between the different resistance groups and HIV in which odds ratios and 95% confidence intervals were reported. The potential confounders identified in the first step were then added one-by-one to the simple model consisting of the outcome of interest and the main predictor (HIV). A covariate was retained in the model if it was significant (p < .05) or if it was considered an important covariate due to biologically plausible relationships. A variable was not retained in the model if it was not significant (p > .05) and its removal did not significantly

reduce the model fit (as assessed by -2 log likelihood). Effect modification on the other hand was tested by identifying potential effect modifiers using the Breslow-Day test for homogeneity of the odds ratios over different strata of the covariate. A p-value of 0.05 or less was considered significant modification.

Where effect modification was found, it was tested by comparing two logistic regression models – one model with the interaction term versus the model without the interaction term. The log likelihood ratio test was used to assess model goodness of fit improved with the interaction term. Statistical analysis software (SAS Institute, Inc., Cary, North Carolina) version 9.2 was used for the analysis. Two-sided P-values of 0.05 or less were considered statistically significant.

D. Results

Of 1,657 patients presenting to both clinics, a total of 1,603 (96.7%) subjects provided informed consent to participate in the study: 1,391 (87%) enrolled at the NTBLTC, Zaria, and 212 (13%) were from BDH, in Kaduna City. The mean age in years and standard deviation (SD) of the total study sample was 37.0 (13.8) and 43.8% were female. Detailed description of their demographics and other covariates were previously described [ Manuscript # 1 ]

Of the 1,603 patients, 375 (23%) were confirmed as MTB complex isolates by the presence of

MPT 64 antigen and positive on SD-bioline test. About 57% of the 375 subjects had an eighth grade or lower level of education. The majority of cases, (76%) were of the Hausa-Fulani ethnic group. The mean age in years of cases was 33.3 (SD=12.4) and the mean BMI for cases was 18.1

(SD=3.6). Nearly one third (27%) of all the cases were co-infected with HIV. Table 9 summarizes the demographic information on these cases by study site.

Pattern of Resistance to Rifampicin and Isoniazid

Among the 375 cases of MTB complex analyzed, 354 (94.4%) were due to M. tuberculosis , 20 (5.3%) had infection with by M. africanum and 1 (0.3%) was due to M bovis .

Overall, 23 (6.1%) cases had resistance to at least one of the two drugs (any resistance); 13

(3.5%) cases had resistance only to isoniazid; 5 (1.3%) had resistance only to rifampicin; while the remaining 5 (1.3%) cases had resistance to both drugs (MDR-TB). There was no difference in the risk of infection when cases with any resistance were compared to cases without any resistance with respect to infection with MTB complex isolates.

Cases with any resistance were more likely to report prior TB treatment compared to cases without any resistance (OR=3.8; 95%CI=1.5-9.5) (Table 10). The odds of a positive

history of diabetes mellitus among cases with any resistance were 3.6 times higher than cases

without any resistance. When isoniazid only resistant cases were compared to cases without any

resistance, cases with resistance to isoniazid were more likely to have had prior TB treatment

compared to cases without any resistance (OR=4.5; 95%CI=1.4-14.3, p=0.0184 ). Repeating this

analysis with rifampicin only resistant cases versus no any resistant cases with respect to the

measured covariates, there were no associations identified. However, in the analysis comparing

MDR-TB cases to cases without any resistance, cases with MDR-TB were more likely to report

alcohol intake and history of diabetes mellitus (OR=7.1; 95%CI=1.2-43.6, p=0.0424 and

OR=15.9; 95%CI=2.5-102.9, p=0.0178 respectively) but less likely to belong to the majority

(Hausa-Fulani) ethnic group (OR=0.1 95%CI= 0.0-0.7, p=0.0155 ) Table 10 .

Association with HIV infection

The unadjusted odds ratios for outcomes comparing cases with resistance to isoniazid alone, rifampicin alone, resistance to both (MDR-TB), any resistance (i.e. resistance to one or the other or both) versus those cases without any drug resistance with respect to HIV infection and various covariates are displayed in Table 10 . Compared to 25.3% of cases without resistance to any of the two drugs, 52.2% of TB cases with any resistance (to at least one of the two drugs) were more likely to be co-infected with HIV (OR=3.2 95%CI=1.4-7.6). About 53.9% of cases with isoniazid resistance were also more likely to have co-infection with HIV compared to

25.3% of cases without resistance to any drug (OR=3.4; 95%CI=1.1-10.5). While cases with mono-resistance to rifampicin and MDR-TB did not significantly differ from cases without resistance to any drug with respect to HIV infection, this may be a function of the small sample size (3 and 2 cases respectively) rather than a true absence of association.

After controlling for potential confounding covariates, the odds of HIV co-infection were higher among cases with any resistance (OR=3.6; 95%CI=1.5-8.8, p=0.0039) and cases with isoniazid mono-resistance (OR=3.9; 95%CI=1.3-12.4, p=0.0193) compared to cases without any resistance (Table 11). The odds of drug resistance were higher in patients with prior TB treatment among those with any resistance (OR=4.4; 95%CI=1.7-11.5, p=0.0023) and with isoniazid mono-resistance (OR=5.2; 95%CI=1.6-17.2, p=0.0074), compared to cases without any resistance. There was no significant association detected between rifampicin mono-resistance and MDR-TB compared to those without any resistance including HIV co-infection and prior TB treatment.

E. Discussion

Our study revealed an increased tendency of HIV co-infected cases towards resistance to isoniazid alone or in combination with rifampicin. Although the evidence for this tendency is stronger with isoniazid than rifampicin, the difference in the strength of evidence may be due to the limited number of rifampicin alone resistant cases compared to cases with isoniazid alone resistance. Despite the high burden of TB in Nigeria [1, 138, 139] and the reported high level of resistance to other first line TB drugs [42, 133, 140], our study found the prevalence of resistance to isoniazid, rifampicin or both to be low. We did not assess resistance to other TB drugs since our original aim was to determine resistance to isoniazid and rifampicin to identify cases with

MDR-TB, and hence selected the appropriate molecular line probe assays that were designed to detect resistance to only these two drugs. The overall prevalence of resistance to other drugs may similarly be higher in this population as previously reported in some parts of Nigeria [42,

129, 136]. If this is true, then with time the failures of ethambutol and pyrazinamide (the two other less powerful first line drugs used in combination with rifampicin and isoniazid), will put more pressure on isoniazid and rifampicin and the prevalence of MDR-TB resistance may increase greatly.

Given our findings that HIV co-infected had detectable resistance to at least isoniazid, this suggests a potential increase in the rate of isoniazid resistance acquisition in this high risk group, particularly that every 3 in 10 TB cases in our study and in Nigeria are co-infected with

HIV.[1] As expected, tuberculosis treatment in the presence of isoniazid alone resistance is reportedly less effective than isoniazid-susceptible tuberculosis. [128, 141] In HIV co-infected cases every effort to prevent isoniazid failure should be attempted since this drug is currently used alone as the standard prophylaxis against TB in HIV-infected.

The association of MDR-TB with ethnicity and diabetes mellitus is interesting and deserves closer scrutiny. Despite the fewer number of cases and the fact that our study measured only a self-reported history of diabetes mellitus, cases with positive history of diabetes mellitus had an increased likelihood of resistance to isoniazid, rifampicin or both (MDR-TB). While the association with ethnicity may very likely be due to chance or some behavioral differences between the ethnic groups, the association with diabetes mellitus and alcohol intake were previously reported in patients with abdominal tuberculosis.[142] It is difficult to explain these associations, though both increased alcohol intake and diabetes are associated with reduced immunity which could increase acquisition of resistant MTB complex organisms from other infected contacts.

Our findings of the high frequency of any drug resistance cases among those with HIV co-infection are consistent with the outcomes of a previous retrospective survey that reported a higher proportion of any drug resistance among HIV positive cases, and found a greater proportion of any drug resistant TB among previously treated cases [143]. Our findings however, differed from that of Asch and colleagues [141] who failed to establish any association between isoniazid resistant TB and HIV infection, but unlike Asch and colleagues who drew inference from TB registry data with missing drug susceptibility test results, we drew inference from original study data designed to answer the research question of interest.

Cases of TB in this study were selected without prior knowledge of their TB or HIV status. Suspected cases visiting the facility for the first time were enrolled to avoid over representation of TB cases or our primary exposure of interest (HIV) that could occur if cases with pre-identified TB or HIV status were selected. The representativeness of our findings of may be limited to only those seeking care at TB treatment facilities, and there is a potential bias

that referrals to our clinical study sites were sent for further care after failed prior treatment.

However, this is less likely since 65% of the identified cases did not receive any treatment prior to enrollment and the proportion of cases with any resistance in our study is low.

In summary, this study found associations between single resistance to isoniazid and HIV co-infection and identified some correlates of resistance to the two most powerful first line drugs that are critical in the global control of tuberculosis. It is vital to prevent the spread of drug resistance by strengthening the DOTS program to promote adherence, treatment completion and identifying and appropriately treating cases with clinically resistant tuberculosis and their contacts. Failure to adopt aggressive strategies at this stage could result in greater morbidity and mortality, and a greater strain on the healthcare system given the higher cost of second-line anti-

TB drugs that are necessary to treat resistant cases.

Table 9. Demographic and covariates of the pulmonary TB cases by site (N=375) Characteristics NTBLTC at Zaria BDH atKaduna

n=315 n= 60

N (%) N (%)

Age: Mean (SD) 33.1 (11.7) 34.3 (15.4)

BMI: Mean (SD) 17.8(3.5) 19.7 (3.3)

Female Sex 102 (32.4) 24 (40.0)

Education <=8 th grade 192 (61.0) 21 (35.0)

Majority Ethnic group 258 (81.9) 26 (43.3)

Alcohol intake 52 (16.5) 13 (21.7)

Prior TB treatment 45 (14.3) 6 (10.0)

Diabetes Mellitus 12 (3.8) 5 (8.3)

HIV co-infection 85 (27.0) 16 (26.7)

n=352 14 (4.0) 14(4.0) 97 (27.6) 61 (17.3) 61 (17.3) 89 (25.3) 89 (25.3) 43 (12.2) 43 (12.2) category] 266 (75.6) 266(75.6) 118(33.5) 295 (83.8) 295(83.8) [Reference [Reference NO resistance NO

1.0000 1.0000 0.0799 1.0000 0.4754 1.0000 1.0000 1.0000 1.0000 0.0076 0.0076 0.0076 0.0076 P-value

3.2 3.4 3.0 4.5 7.6 9.5 n=23 13.5

95%CI 1.2, 0.4- 1.2, 0.9- 3.6, 1.4, 0.6- 1.4, 1.0, 0.3- 1.0, 1.3, 0.4- 1.3, 3.2, 1.4- 3.2, 3.8, 1.5- 3.8, 1.1, 1.1, 0.4- 2.6 ANY resistance ANY

3 8 8 4 8 18 20 12 (78.3) (34.8) (13.0) (34.8) (17.4) (17.4) (86.7) (86.7) (52.2) (52.2) (34.8) (34.8) N (%) (%) N Odds ratio,

P- value 1.000 0.0155 0.0155 0.0178 0.6685 0.1363 0.0424 0.0424 0.6050 0.6050 0.1219 0.1219

* 0.7 4.5 n=5 n=5 24.0 43.6 43.6 12.0 12.0 29.4 29.4 102.9 95%CI 0.4, 0.4, 0.1- 3.9, 0.6- 7.1, 7.1, 1.2- 2.0, 2.0, 0.3- 4.8, 4.8, 0.8- 0.1, 0.1, 0.0- 15.9,2.5-

MDR-TB resistance MDR-TB

2 1 3 1 3 2 2 5 (100) (100) (20.0) (20.0) (40.0) (60.0) (60.0) (60.0) (40.0) (40.0) (40.0) (40.0) N (%) (%) N ratio, Odds

P- value 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 0.1363 0.1103 0.1103 0.4850 0.4850 0.1926 0.1926 e to isoniazid alone, rifampicin

* isoniazid and rifampicin) rifampicin) and isoniazid 8.0 1.8 1.8 n=5 n=5 11.7 24.0 10.8 10.8 26.9 26.9 16.4 16.4 95%CI 1.3, 1.3, 0.1- 1.3, 1.3, 0.2- 3.9, 0.6- 1.2, 1.2, 0.1- 4.4, 4.4, 0.7- 0.3, 0.0- 1.8, 1.8, 0.2- both

to those without any drug resistance resistance drug any without to those

4 1 2 3 3 3 1 ce to to ce Rifampicin resistance only resistance Rifampicin , or both; NO resistance: Not resistant to either either to resistant Not resistance: both; NO or , uency cells cells uency (80.0) (40.0) (60.0) (20.0) (20.0) (60.0) (60.0) (60.0) (20.0) (20.0) N (%) (%) N ratio, Odds 0 (0.0) (0.0) 0

0.0443 0.4621 0.7680 0.5269 0.1371 0.0468 0.7013 0.0184 P-value P-value

3.9 2.2

- -

* * n =13 0.4 0.1 , , , 95%CI 1.2 0.5 1.8, 0.2-14.5 1.8, 0.2-14.5 3.4, 1.1-10.5 3.4, 1.1-10.5 2.3, 0.3-18.2 4.5, 1.4-14.3 4.5, 1.4-14.3 Isoniazid resistance only only Isoniazidresistance

2 7 5 13 12 (100) (53.9) (53.9) (92.3) (38.5) (38.5) (15.4) (15.4) 1 (7.7) 0 (0.0)

Unadjusted analysis comparing cases with resistanc

(%) N ratio, Odds

Group Alcohol Prior TB Smoking Cigarette Cigarette Treatment Treatment Female sex Female 2 Site A(Zaria) Consumption Characteristics Characteristics HIV Infection Odds ratio estimation not possible due to zero freq zero to due possible not estimation ratio Odds Majority Ethnic Majority Diabetes Diabetes Mellitus ANY resistance: Resistance to isoniazid, rifampicin to isoniazid, Resistance resistance: ANY isoniazid or rifampicin rifampicin or isoniazid * Table (resistan tuberculosis resistant Multi-drug MDR-TB: alone, both (MDR-TB) and those with any resistance resistance any those with and (MDR-TB) both alone,

P- value 0.0023 0.0023 0.0041 0.0041

3.6, 3.6, 4.4, 1.7- OR, OR, 11.5 11.5 95%CI 95%CI 1.5-8.8 1.5-8.8

P- value 0.4024 0.4024 0.0844 d d prior TB treatment with MDR-TB MDR-TB Resistance Any

2.2, 2.2, 3.7, 0.4- 0.6- OR, interval. NB: sample sizes for sizes for interval.sample NB: 13.5 13.5 22.1 95%CI 95%CI ent multivariable logistic regression logistic regression multivariable ent

P- value value 0.0984 0.0984 0.5121

alone alone

Rifampicin Rifampicin 5.0, 5.0, 1.4, 0.8- 0.2- OR, OR, 29.9 29.9 12.7 95%CI

P- value value 0.0193 0.0074

3.9, 3.9, 5.2, 1.3- 1.6- 12.4 12.4 17.2 Isoniazid alone Isoniazid 95%CI 95%CI

. . Adjusted associations between HIV co-infection an

1 1 Yes Yes Yes Table Table 11 HIV Infection No Prior OR, treatment No OR=OddsCI=Confidence ratio; category; 1=Reference fortoo effici low were MDR-TB and alone rifampicin analysis. resistance type versus no resistance noversus resistance type resistance Characteristic

V. DISCUSSION OF DISSERTATION FINDINGS

This study has made several important findings some of them previously unreported from Sub-Saharan Africa. The overall goal of conducting this thesis on TB and

HIV in this population was to contribute to the better understanding of the extent of inter- related problems of tuberculosis and HIV/AIDS in Nigeria. This densely populated nation has a high burden of HIV co-infected TB cases and remains a major target in the global control of tuberculosis and HIV. The findings of this study have identified gaps in pulmonary mycobacterial detection which may significantly influence the disease treatment and control.

A. Prevalence of pulmonary mycobacterium

Identifying the association between pulmonary mycobacterial infection and HIV was the main objective of this research work. While the original goal was to determine whether M. bovis was becoming an increased cause of pulmonary mycobacterial infection, our study found only one case. This result and the unexpectedly high prevalence of NTM identified in our study resulted in a revised focus of this dissertation to focus on pulmonary mycobacterial infection, regardless of causative organism, and to investigate its association with HIV co-infection, especially in the more nomadic northern Nigeria. The only single case of M. bovis identified in our study was found after screening over 375 culture- confirmed MTB complex isolates. However, the fact that this case is also HIV co-infected demonstrates that HIV may still be a strong risk factor for M. bovis infection but there was insufficient statistical power to prove this claim. Pulmonary tuberculosis from M. africanum is more common in this study population than M. bovis and the shared

similarities between the two often makes accurate differentiation harder with conventional methods. The very low prevalence (0.3%) of M. bovis pulmonary TB identified in our study is in agreement with the findings of similar recently reported studies for the identification of mycobacterial strains in the neighboring west African countries of Ghana,

Mali, Cameroun and Burkina Faso [15, 106-108] which either failed to detect any M. bovis or identified it in very low proportion (3% from Ghana, 0.8% from Mali, 0.2% from

Cameroun and none from Burkina Faso) but identified high proportions of M. Africanum

(9 to 28% among the mycobacterial strains).

Our findings on the prevalence of M. bovis infection differed considerably from previous surveys conducted on similar study populations that reported a high prevalence of M. bovis infection [64, 93]. It is likely that M. africanum was misclassified as M. bovis when using the conventional methods for M. bovis and perhaps may explain why these studies found a higher prevalence of M. bovis compared to our (and other) studies. It is unlikely that we misclassified M. africanum , given that after using state-of-the-art molecular probe assays,

20 cases of M. africanum were identified and only one case of M. bovis was found. The prevalence of TB among suspected cases dropped when the NTM cases were excluded.

When culture results are used as gold standard, the standard of care pre-enrollment smear microscopy test missed about one third of all TB cases. The misclassified active TB cases may have likely continued to spread the disease even as they receive ineffectual treatment for chest infections other than TB. For effective TB control a more efficient standard of care detection tool is necessary. Another study from Nigeria found an even higher proportion of missed cases than we did but our findings were similar on gender

differences: TB is more common among males, while females were more likely to be HIV positive and smear negative [144].

B. Non-tuberculosis Mycobacterial Infection

The prevalence of NTM among cases of pulmonary TB in Nigeria was not fully investigated before this study. The fact that 1 of every 5 TB cases was an NTM and 4 in 10 cases of NTMs were likely to be HIV positive showed a much stronger correlation between the disease and HIV than between M. tuberculosis and HIV; and the potential threat to the control of TB among HIV positives. It is quite likely that NTM infections are less aggressive than MTB complex organisms and tend to behave like opportunistic infections that can disproportionately affect patients with HIV infection.

Most of the NTM cases came to the TB clinics with pulmonary symptoms at the peak of the Harmattan dusty weather, at the same time when infectious organisms from the environment are carried in the air by wind. While it is possible that these cases were previously infected with NTMs and became symptomatic during the dry dusty weather, it is also possible that they acquired the infection during this three-month long season

(December to February). Since our study is a cross-sectional study, it is not possible to determine temporal association between the dust storm and NTM infection. The impact of this disease can best be seen when identified cases are followed up over time to evaluate response to treatment and duration. Some NTMs are known to produce disseminated disease and their response to conventional anti-TB drugs is poor [145]. In such cases the impact of the disease, even with therapy, is likely to be more severe than TB. Like most previous reports on NTM , we found the disease to be more common among older age groups [146-148]. However, while the average age of NTM subjects in the past studies was

between 50 to 60 years, the average age of subjects in our study was comparatively less.

Although these infections are commoner in older age groups, occurrence among children and adolescents has been reported.[149]

C. Association with HIV Infection

The prevalence of HIV infection among confirmed TB cases was only 3% higher than the prevalence among all subjects screened. The diminished margin is expected given that since facilities ran both DOTS and HIV clinics, and that cases of either disease were screened for the other as part of the standard of care, thereby increasing the chance for detection. Our findings are similar to the WHO reported figures on the prevalence of HIV among TB cases in Nigeria and also agreed with previously reported surveys from Nigeria

[1, 150]. The high prevalence of HIV among M. tuberculosis cases and that of M. tuberculosis among HIV cases has been reported for over two decades [151, 152]. This study also found similar associations between M. tuberculosis and HIV but did not find a similar association between HIV and M. africanum , which could be due to the fewer cases of M. africanum identified.

D. Resistance to Isoniazid and Rifampicin

The prevalence of MDR-TB as defined by resistance to both isoniazid and rifampicin is low, and is much lower than the reported prevalence from the entire first line drugs [42, 133]. This suggests that more treatment failures occur from other first line drugs than from the two most powerful drugs. If this is true, the implication is that with increasing resistance to the ethambutol and pyrazinamide (the two other first line drugs), there will be pressure on isoniazid and rifampicin with an eventual increase in the rate of

resistance development to these drugs. Hence this is the best time to prevent such a scenario in this population by strengthening programs like the directly observed treatment short (DOTS) course which promotes adherence and ensures completion of treatment. The association of MDR-TB with ethnicity and diabetes mellitus deserves closer scrutiny.

Despite the fewer number of cases and the fact that our study measured only a self- reported history of diabetes mellitus, cases with positive history of diabetes mellitus had an increased likelihood of resistance to isoniazid, rifampicin or both (MDR-TB). While the association with ethnicity may very likely be due to chance or some behavioral differences between the ethnic groups, the association with diabetes mellitus and alcohol intake were previously reported in patients with abdominal tuberculosis.[153]. It is difficult to explain these associations, though both increased alcohol intake and diabetes are associated with reduced immunity which could increase acquisition of resistant MTB complex organisms from other infected contacts.

E. Limitations

There may be bias in subject selection, laboratory testing and analyses. Bias in the selection of subjects was minimized by enrolling almost all (96.7%) suspected TB cases into the study. Those excluded from the study for failure to meet eligibility criteria or unwillingness to participate were less than 4% of the total subjects approached. Lack of probability sampling may make the findings of this study non-representative of the population of suspected TB persons in Nigeria. However, these two clinical sites account for 33% of all new TB cases in the state of Kaduna (NTBLTC: 25% and BDH: 8%), and provide treatment to patients from both urban and rural areas in this nationally representative state of approximately 6 million people, making our results generalizable to

a large proportion of Nigeria. The correctness of the self-reported data on a number of survey items like history of diabetes mellitus, prior or current treatment for TB depends on the willingness of the subjects to provide accurate information and skills and expertise of the interviewer. Misclassification and bias in the estimations of the levels of these exposures are therefore likely. However, the survey instrument for data collection was applied to all subjects by trained study staff without prior knowledge of the subjects’ HIV or TB status.

Bias and variability were further reduced by the large sample size of 1,600 from which over 450 confirmed cases of mycobacterial infections were detected. The excellent resources used in testing the subjects including the highly sensitive liquid cultures on all subjects regardless of the results of the pre-enrollment sputum smear examination, and the molecular line probe assays to detect mycobacterial infections including resistance to isoniazid and rifampicin further added to the strength of this study.

Finally, this research filled gaps in the literature by the identification of species of

NTM among cases of pulmonary TB in Nigeria together with estimated prevalence. The high frequency of NTM cases diagnosed with links to the seasonal dust exposure during

Harmattan, and the increased likelihood of HIV co-infection presents a novel public health challenge. The information on association of resistance to isoniazid, rifampicin or both with HIV infection is a new finding in this population. This study found a very low prevalence of M. bovis and a surprisingly higher prevalence of M. africanum . The proportion of undetected cases of active pulmonary tuberculosis (if only pre-enrollment sputum smears were used as per standard of care) has implications on the global control of tuberculosis, and while more expensive, introduction of molecular detection assays to

identify smear-negative NTM and MDR-TB should be a high priority given the high proportion of missed cases of active TB when using standard of care alone.

VI. APPENDIX A

BOVINE TUBERCULOSIS INFECTION AND ITS ASSOCIATION WITH HIV AMONG LABORATORY CONFIRMED CASES OF PULMONARY TUBERCULOSIS IN NIGERIA

QUESTIONNAIRE IDENTIFICATION NUMBER………………..

Date…………………………. 003 Study staff initials……….. ………………… 004 Participant’s ID (card number)………………….

005 Sex: M F

006 Weight (kilograms)……………………………..

007 Height (Meters)…………………………………………

No Questions and filters Coding categories Skip

101 What is your age? (AGE) AGE IN COMPLETED YEARS [ | ]

102 (EDU) What level of formal school education NONE 0 have you completed? PRIMARY 2 JUNIOR SECONDARY……………..3 SENIOR SECONDARY……………..4 CERTIFICATE/DIPLOMA.. 5 UNIVERSITY . 6

103 What is your tribe or ethnic group? HAUSA…………………….2 (ETC) FULANI …… 3 YORUBA 4 IGBO 5 OTHER 6 (if OTHER please write: ……

……………………………. )

104 Are you currently married, single, (MST) divorced or widowed? MARRIED……………… .2 PROBES WIDOWED 3 DIVORCED 4 If married, how many wives do you have SINGLE/NEVER MARRIED OR your husband has? 5 ………………………

If unmarried, number of concurrent partnerships………………………… 1-5 ……………………….. 2 105 How many people live in your house 6-10……………………… 3 (NPH) hold? >10 ………………………..4

1-2 ………………………...2 106 The number of persons sleeping per 3-5 ……………………… 3 (NPR) room? >5 ………………………...4

107 Number of windows per room? NO WINDOW…...…… …2 (WPR) 1 WINDOW………..……3 >2 WINDOWS…...………4

108 What is your occupation? (OCP) FULL TIME FARMING 2 PART TIME FARMING 3 ANIMAL HEALTH WORKER ...... 4 HUMAN HEALTH WORKER………. ……….5 WORK AT ABATTOIR …6 MEAT VENDOR ………..7 OTHER (specify)………... 8 UNEMPLOYED.. 9

109 IF 0 (TFM) Are you currently engaged in cattle NO 0 rearing or other livestock raising? YES 1 SKIP DON’T KNOW/NOT SURE 97 REFUSED …………..…….. 99 TO

Q 119

110 How much time of the day do you spend NO TIME SPEN.…………...0 (FTP) tending to cattle or other livestock? THE WHOLE DAY ………..2 1 HOUR OR LES.…………..3 MORE THAN 1 HOU……..4

111 Which of the following livestock do you NONE ……………..……… 0 (LST) spend the time with? CATTLE..………………... ..2 SHEEP...…………………. . 3 GOAT..……………………..4 MIXED (specifytype)………………..5

112 Livestock (herd) size you spend time 0…………………………..0 (HDS) tending to? <10 …………………………2 10-20 ……………….………3 21-30………………….…….4 >30………………………….5

113 (ASC) Were there any SICK cattle in your herd NO 0 or in the neighborhood within the last 3 YES 1 months? DON’T KNOW/NOT SURE 97 REFUSED ………………… 99

114 (CEC) Were there any cattle in the past 3 NO 0 months from your herd or in the YES 1 neighborhood that had coughed or DON’T KNOW/NOT SURE 97 exhibited substantial loss of weight? REFUSED ………………… 99

115 NO 0 (ATB) Were there any cattle in the past 3 YES 1 months from your herd or in the DON’T KNOW/NOT SURE 97 neighborhood that was found to have an REFUSED ………………… 99 animal TB? 116 Do you observe any cattle from your (SWC) herd with a swelling around the head, NO 0 neck or other parts of the body? YES 1 DON’T KNOW/NOT SURE 97 REFUSED ……………….... 99

117 NO 0

(NWC Any new cattle introduced into the herd YES 1 ) within the last 3 months? DON’T KNOW/NOT SURE 97 REFUSED ……………….. ..99

118 Do you milk breast feeding Livestock? NO .0 (MLA) YES .1 Cows? Goats? Other? DON’T KNOW/NOT SURE 97 REFUSED ……………….... 99

If NO, 119 Do you consume the locally expressed NO .0 (CSM) milk from livestock or meals consisting YES .1 SKIP of it? DON’T KNOW/NOT SURE 97 TO REFUSED ………………… 99 MEAL: ASK SPECIFICALLY ABOUT Q122 “FURA DA NONO” AND “MAN SHANU”

120 How often do you take the milk or meal NOT AT ALL………………..0 (RMC) consisting of it? EVERY DAY ……………….2 SOME DAYS…………….….3 DON’T KNOW/NOT SURE……………………...... 97 REFUSE.……………………99

121 NONE …………………..…...0 (FMC ) UNBOILED ………………...2 What type of the milk do you consume? BOILED …………………….3 BOTH BOILED AND UNBOILED …...... 4 DON’T KNOW/NOT SURE..………………....…....97 REFUSED……………...... …99

122 Do you consume raw meat or blood? NO 0 (RWM YES 1 DON’T KNOW/NOT SURE 97 REFUSED ……….……….. 99

123 (CTT) Are you currently receiving treatment NO 0 for TB? YES 1

DON’T KNOW/NOT SURE 97 REFUSED ………….…….. 99

124 NO 0 (TBT) YES 1 Have you ever received treatment for TB DON’T KNOW/NOT SURE 97 before now? REFUSED …………….….. 99

125 Does any member of your family or a NO 0 (FHT) close contact currently has a cough or YES 1 receives treatment for TB? DON’T KNOW/NOT SURE 97 REFUSED….…………….. . 99

126 (PHT) Did any member of your family or a NO 0 close contact had TB in the past? YES 1 DON’T KNOW/NOT SURE 97 REFUSED…….………….. . 99

127 Does any member of your family has a NO 0 (CVS) swelling around the neck, receive YES 1 treatment for neck swelling or had a DON’T KNOW/NOT SURE 97 swollen neck in the past? REFUSED………….…….. . 99

128 If any of the questions 125, 126 or 127 is NO 0 (PTC) true, does this person tend to cattle or YES 1 other livestock? DON’T KNOW/NOT SURE 97 REFUSED…….………….. . 99

130 A BCG immunization is usually given (VCN) only once in a person’s lifetime and is NO 0 given in the right arm. Have you ever YES 1 had a BCG immunization or a sign in DON’T KNOW/NOT SURE 97 your right arm suggestive of BCG REFUSED ……………….…99 immunization?

PROVIDE PHOTOS FOR CLARITY

131 Have you ever been told by a doctor, (HDM nurse or other health professional that NO 0 ) you have diabetes mellitus? YES 1 DON’T KNOW/NOT SURE PROBE WITH DIFFERENT NAMES 97 OF DM TO ENSURE REFUSED …………………99

UNDERSTANDING If NO 132 Have you ever smoked cigarettes in your NO 0 (SMK) life? YES 1 SKIP DON’T KNOW/NOT SURE 97 TO REFUSED ….………………99

Q135

133 Have you smoked at least 100 cigarettes NO 0 (NCG ) in your entire life? YES 1 NOTE: 5 PACKS= 100 CIGARETTES DON’T KNOW/NOT SURE 97 REFUSED ………………….99

134 If you smoked up to 100 cigarettes in NOT ANY MORE…………...0 (SFQ) your life, how many cigarettes do you 1-2………….. .……………….2 now smoke in a day? 3-5….…………………………3 MORE THAN 5….……..…….4 DON’T KNOW/NOT SURE…97 REFUSED…………………….99

If NO 135 Have you ever drunk alcohol? NO 0 (ALC) YES 1 SKIP DON’T KNOW/NOT SURE 97 TO REFUSED…………… …….99 Q137

136 In the past 30 days, have you had at NO 0 (SAC) least one drink of any alcoholic YES 1 beverage? DON’T KNOW/NOT SURE 97 REFUSED ………………….99

137 Have you been treated for any of the NO 0 (STI) following in the past one year? READ YES 1 OUT DON’T KNOW/NOT SURE 97 Gonorrhea, Syphilis, HPV (for female REFUSED ………………….99 less than 50 years) 138 (UPS ) Do you have unprotected sex with multiple partners within the last one NO 0 year? YES 1 DON’T KNOW/NOT SURE 97 Unprotected sex : having sex without REFUSED…………………..99

using a new latex or polyurethane condom every time

Multiple partners: extra marital partners

139 (SHP ) Do you have unprotected sex with NO 0 someone who is HIV-positive? YES 1 DON’T KNOW/NOT SURE 97 REFUSED ………………….99

140 (BTF ) Have you ever received a blood NO 0 transfusion or blood products? YES 1 DON’T KNOW/NOT SURE 97 REFUSED ………………….99

VII. REFRENCES .

1. World Health Organization (WHO) Global TB Report , 2009. 2. World Health Organization (WHO) Global TB Report 2010. 3. Raviglione, M.C., D.E. Snider, Jr., and A. Kochi, Global epidemiology of tuberculosis. Morbidity and mortality of a worldwide epidemic. JAMA : the journal of the American Medical Association, 1995. 273 (3): p. 220-6. 4. Behr, M.A., et al., Transmission of Mycobacterium tuberculosis from patients smear-negative for acid-fast bacilli. Lancet, 1999. 353 (9151): p. 444-9. 5. Cole, E.C. and C.E. Cook, Characterization of infectious aerosols in health care facilities: an aid to effective engineering controls and preventive strategies. Am J Infect Control, 1998. 26 (4): p. 453-64. 6. Griffith, D.E. and C.M. Kerr, Tuberculosis: disease of the past, disease of the present. J Perianesth Nurs, 1996. 11 (4): p. 240-5. 7. Cooke, G.S., et al., Mapping of a novel susceptibility locus suggests a role for MC3R and CTSZ in human tuberculosis. American journal of respiratory and critical care medicine, 2008. 178 (2): p. 203-7. 8. Dormans, J., et al., Correlation of virulence, lung pathology, bacterial load and delayed type hypersensitivity responses after infection with different Mycobacterium tuberculosis genotypes in a BALB/c mouse model. Clinical and experimental immunology, 2004. 137 (3): p. 460-8. 9. Bellamy, R., et al., Genetic susceptibility to tuberculosis in Africans: a genome- wide scan. Proceedings of the National Academy of Sciences of the United States of America, 2000. 97 (14): p. 8005-9. 10. Gajalakshmi, V. and R. Peto, Smoking, drinking and incident tuberculosis in rural India: population-based case-control study. Int J Epidemiol, 2009. 38 (4): p. 1018- 25. 11. Kolappan, C., et al., Selected biological and behavioural risk factors associated with pulmonary tuberculosis. Int J Tuberc Lung Dis, 2007. 11 (9): p. 999-1003. 12. Dong, B., N. Ge, and G. Liu, [Social economical status, behaviors and environment as the risk factors of tuberculosis in Chengdu China]. Zhonghua Liu Xing Bing Xue Za Zhi, 2001. 22 (2): p. 102-4. 13. van Soolingen, D., et al., A novel pathogenic taxon of the Mycobacterium tuberculosis complex, Canetti: characterization of an exceptional isolate from Africa. Int J Syst Bacteriol, 1997. 47 (4): p. 1236-45. 14. Alexander, K.A., et al., Novel Mycobacterium tuberculosis complex pathogen, M. mungi. Emerging infectious diseases, 2010. 16 (8): p. 1296-9. 15. Addo, K., et al., Mycobacterial species causing pulmonary tuberculosis at the korle bu teaching hospital, accra, ghana. Ghana medical journal, 2007. 41 (2): p. 52-7. 16. de Jong, B.C., M. Antonio, and S. Gagneux, Mycobacterium africanum--review of an important cause of human tuberculosis in West Africa. PLoS neglected tropical diseases, 2010. 4(9): p. e744.

17. de Jong, B.C., et al., Progression to active tuberculosis, but not transmission, varies by Mycobacterium tuberculosis lineage in The Gambia. The Journal of infectious diseases, 2008. 198 (7): p. 1037-43. 18. Humblet, M.F., M.L. Boschiroli, and C. Saegerman, Classification of worldwide bovine tuberculosis risk factors in cattle: a stratified approach. Vet Res, 2009. 40 (5): p. 50. 19. Tschopp, R., et al., Risk factors of bovine tuberculosis in cattle in rural livestock production systems of Ethiopia. Prev Vet Med, 2009. 89 (3-4): p. 205-11. 20. Grange, J.M., Mycobacterium bovis infection in human beings. Tuberculosis (Edinb), 2001. 81 (1-2): p. 71-7. 21. Amorim, A., et al., Non-tuberculous mycobacteria in HIV-negative patients with pulmonary disease in Lisbon, Portugal. Scand J Infect Dis. 42 (8): p. 626-8. 22. Griffith, D.E., et al., An official ATS/IDSA statement: diagnosis, treatment, and prevention of nontuberculous mycobacterial diseases. Am J Respir Crit Care Med, 2007. 175 (4): p. 367-416. 23. Katoch, V.M., Infections due to non-tuberculous mycobacteria (NTM). Indian J Med Res, 2004. 120 (4): p. 290-304. 24. Hosker, H.S., et al., The prevalence and clinical significance of pulmonary infection due to non-tuberculous mycobacteria in Hong Kong. Respir Med, 1995. 89 (1): p. 3-8. 25. Sakatani, M., [Nontuberculous mycobacteriosis; the present status of epidemiology and clinical studies]. Kekkaku, 1999. 74 (4): p. 377-84. 26. Payen, M.C., S. De Wit, and N. Clumeck, [Manifestations, diagnosis and treatment of non-tuberculous mycobacterial infections in patients with HIV infection]. Rev Mal Respir, 1997. 14 Suppl 5 : p. S142-51. 27. Alvarez-Uria, G., et al., Non-tuberculous mycobacteria in the sputum of HIV- infected patients: infection or colonization? Int J STD AIDS, 2009. 20 (3): p. 193-5. 28. Monkongdee, P., et al., Yield of acid-fast smear and mycobacterial culture for tuberculosis diagnosis in people with human immunodeficiency virus. Am J Respir Crit Care Med, 2009. 180 (9): p. 903-8. 29. Zykov, M.P., H. Roulet, and N. Gaya, Non-tuberculosis mycobacteria in Africa. 1. Isolation and identification. Bull World Health Organ, 1967. 37 (6): p. 927-38. 30. Zykov, M.P. and H. Roulet, Non-tuberculosis mycobacteria in Africa. 3. Formamidase activity--its evaluation and practical application. Bull World Health Organ, 1967. 37 (6): p. 947-51. 31. Idemyor, V., HIV and tuberculosis coinfection: inextricably linked liaison. Journal of the National Medical Association, 2007. 99 (12): p. 1414-9. 32. Diedrich, C.R. and J.L. Flynn, HIV-1/mycobacterium tuberculosis coinfection immunology: how does HIV-1 exacerbate tuberculosis? Infection and immunity, 2011. 79 (4): p. 1407-17. 33. Corbett, E.L., et al., The growing burden of tuberculosis: global trends and interactions with the HIV epidemic. Arch Intern Med, 2003. 163 (9): p. 1009-21. 34. Vynnycky, E. and P.E. Fine, Lifetime risks, incubation period, and serial interval of tuberculosis. American journal of epidemiology, 2000. 152 (3): p. 247-63. 35. Whalen, C., et al., Accelerated course of human immunodeficiency virus infection after tuberculosis. Am J Respir Crit Care Med, 1995. 151 (1): p. 129-35.

36. Hlavsa, M.C., et al., Human tuberculosis due to Mycobacterium bovis in the United States, 1995-2005. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America, 2008. 47 (2): p. 168-75. 37. Bouvet, E., et al., A nosocomial outbreak of multidrug-resistant Mycobacterium bovis among HIV-infected patients. A case-control study. AIDS, 1993. 7(11): p. 1453-60. 38. Lopez-Alvarez, R., et al., First insights into the genetic diversity of Mycobacterium tuberculosis isolates from HIV-infected Mexican patients and mutations causing multidrug resistance. BMC microbiology, 2010. 10 : p. 82. 39. Probandari, A., et al., Missed opportunity for standardized diagnosis and treatment among adult tuberculosis patients in hospitals involved in Public-Private Mix for Directly Observed Treatment Short-Course strategy in Indonesia: a cross-sectional study. BMC health services research, 2010. 10 : p. 113. 40. Kawai, V., et al., Tuberculosis mortality, drug resistance, and infectiousness in patients with and without HIV infection in Peru. Am J Trop Med Hyg, 2006. 75 (6): p. 1027-33. 41. Gandhi, N.R., et al., Extensively drug-resistant tuberculosis as a cause of death in patients co-infected with tuberculosis and HIV in a rural area of South Africa. Lancet, 2006. 368 (9547): p. 1575-80. 42. Lawson, L., et al., Pilot study on multidrug resistant tuberculosis in Nigeria. Annals of African medicine, 2010. 9(3): p. 184-7. 43. Cadmus, S., et al., Molecular analysis of human and bovine tubercle bacilli from a local setting in Nigeria. J Clin Microbiol, 2006. 44 (1): p. 29-34. 44. Mawak, J., et al., Human pulmonary infections with bovine and environment (atypical) mycobacteria in jos, Nigeria. Ghana Med J, 2006. 40 (4): p. 132-6. 45. Inangolet, F.O., et al., A cross-sectional study of bovine tuberculosis in the transhumant and agro-pastoral cattle herds in the border areas of Katakwi and Moroto districts, Uganda. Trop Anim Health Prod, 2008. 40 (7): p. 501-8. 46. Kang'ethe, E.K., C.E. Ekuttan, and V.N. Kimani, Investigation of the prevalence of bovine tuberculosis and risk factors for human infection with bovine tuberculosis among dairy and non-dairy farming neighbour households in Dagoretti Division, Nairobi, Kenya. East Afr Med J, 2007. 84 (11 Suppl): p. S92-5. 47. Karimuribo, E.D., et al., Studies on mastitis, milk quality and health risks associated with consumption of milk from pastoral herds in Dodoma and Morogoro regions, Tanzania. J Vet Sci, 2005. 6(3): p. 213-21. 48. Muller, B., et al., African 1, an epidemiologically important clonal complex of Mycobacterium bovis dominant in Mali, Nigeria, Cameroon, and Chad. J Bacteriol, 2009. 191 (6): p. 1951-60. 49. Armstrong, J.R., Eradication of tuberculosis from an extensively managed beef herd. Aust Vet J, 1981. 57 (5): p. 220-3. 50. Everett, R.E., Eradication of bovine TB: learning from other countries. Vet Rec, 2006. 158 (18): p. 640. 51. Kouba, V., [History of the elimination of bovine tuberculosis in the Czech Republic]. Cas Lek Cesk, 1999. 138 (15): p. 456-9. 52. Dhungana, G.P., et al., Tuberculosis co-infection in HIV infected persons of Kathmandu. Nepal Med Coll J, 2008. 10 (2): p. 96-9.

53. Kimerling, M.E., et al., Prevalence of pulmonary tuberculosis among HIV-infected persons in a home care program in Phnom Penh, Cambodia. Int J Tuberc Lung Dis, 2002. 6(11): p. 988-94. 54. Ungvarski, P.J., Human immunodeficiency virus (HIV) and Mycobacterium tuberculosis (TB). J N Y State Nurses Assoc, 1990. 21(4): p. 7-10. 55. Hlavsa, M.C., et al., Human tuberculosis due to Mycobacterium bovis in the United States, 1995-2005. Clin Infect Dis, 2008. 47 (2): p. 168-75. 56. van den Berk, G.E., et al., Evaluation of the rapid immunoassay determine HIV 1/2 for detection of antibodies to human immunodeficiency virus types 1 and 2. Journal of clinical microbiology, 2003. 41 (8): p. 3868-9. 57. Piwowar-Manning, E.M., et al., Validation of rapid HIV antibody tests in 5 African countries. Journal of the International Association of Physicians in AIDS Care, 2010. 9(3): p. 170-2. 58. Lyamuya, E.F., et al., Evaluation of simple rapid HIV assays and development of national rapid HIV test algorithms in Dar es Salaam, Tanzania. BMC infectious diseases, 2009. 9: p. 19. 59. Selvakumar, N., et al., Inefficiency of 0.3% carbol fuchsin in ziehl-neelsen staining for detecting acid-fast bacilli. Journal of clinical microbiology, 2002. 40 (8): p. 3041-3. 60. Rodrigues, C., et al., Evaluation of the bactec MGIT 960 TB system for recovery and identification of Mycobacterium tuberculosis complex in a high through put tertiary care centre. Indian journal of medical microbiology, 2009. 27 (3): p. 217- 21. 61. Somoskovi, A., et al., Direct comparison of the genotype MTBC and genomic deletion assays in terms of ability to distinguish between members of the Mycobacterium tuberculosis complex in clinical isolates and in clinical specimens. J Clin Microbiol, 2008. 46 (5): p. 1854-7. 62. Richter, E., et al., Usefulness of the GenoType MTBC assay for differentiating species of the Mycobacterium tuberculosis complex in cultures obtained from clinical specimens. J Clin Microbiol, 2004. 42 (9): p. 4303-6. 63. Richter, E., et al., Evaluation of genotype MTBC assay for differentiation of clinical Mycobacterium tuberculosis complex isolates. J Clin Microbiol, 2003. 41 (6): p. 2672-5. 64. Cadmus, S., et al., Molecular analysis of human and bovine tubercle bacilli from a local setting in Nigeria. Journal of clinical microbiology, 2006. 44 (1): p. 29-34. 65. Zykov, M.P., H. Roulet, and N. Gaya, Non-tuberculosis mycobacteria in Africa. 1. Isolation and identification. Bulletin of the World Health Organization, 1967. 37 (6): p. 927-38. 66. Lienhardt, C., et al., Investigation of the risk factors for tuberculosis: a case- control study in three countries in West Africa. Int J Epidemiol, 2005. 34 (4): p. 914-23. 67. Lonnroth, K., et al., Drivers of tuberculosis epidemics: the role of risk factors and social determinants. Soc Sci Med, 2009. 68 (12): p. 2240-6. 68. Johnson, K. and A. Way, Risk factors for HIV infection in a national adult population: evidence from the 2003 Kenya Demographic and Health Survey. J Acquir Immune Defic Syndr, 2006. 42 (5): p. 627-36.

100

69. Bezirtzoglou, C., K. Dekas, and E. Charvalos, Climate changes, environment and infection: facts, scenarios and growing awareness from the public health community within Europe. Anaerobe, 2011. 17 (6): p. 337-40. 70. Boxall, A., et al., Impacts of climate change on indirect human exposure to pathogens and chemicals from agriculture. Ciencia & saude coletiva, 2010. 15 (3): p. 743-56. 71. Griffith, D.E., et al., An official ATS/IDSA statement: diagnosis, treatment, and prevention of nontuberculous mycobacterial diseases. American journal of respiratory and critical care medicine, 2007. 175 (4): p. 367-416. 72. Amorim, A., et al., Non-tuberculous mycobacteria in HIV-negative patients with pulmonary disease in Lisbon, Portugal. Scandinavian journal of infectious diseases, 2010. 42 (8): p. 626-8. 73. Buijtels, P.C., et al., Isolation of non-tuberculous mycobacteria at three rural settings in Zambia; a pilot study. Clinical microbiology and infection : the official publication of the European Society of Clinical Microbiology and Infectious Diseases, 2010. 16 (8): p. 1142-8. 74. Crump, J.A., et al., Invasive disease caused by nontuberculous mycobacteria, Tanzania. Emerging infectious diseases, 2009. 15 (1): p. 53-5. 75. Buijtels, P.C., et al., Isolation of nontuberculous mycobacteria in Zambia: eight case reports. Journal of clinical microbiology, 2005. 43 (12): p. 6020-6. 76. Bonard, D., et al., High incidence of atypical mycobacteriosis in African HIV- infected adults with low CD4 cell counts: a 6-year cohort study in Cote d'Ivoire. AIDS, 2004. 18 (14): p. 1961-4. 77. Lan, R., et al., Mycobacterium tuberculosis and non-tuberculous mycobacteria isolates from HIV-infected patients in Guangxi, China. The international journal of tuberculosis and lung disease : the official journal of the International Union against Tuberculosis and Lung Disease, 2011. 15 (12): p. 1669-75. 78. Herzmann, C., S. Esser, and C. Lange, [Infections with non-tuberculous mycobacteria in HIV-infected patients]. Der Hautarzt; Zeitschrift fur Dermatologie, Venerologie, und verwandte Gebiete, 2011. 62 (4): p. 272-9. 79. Shojaei, H., et al., Species identification of neglected nontuberculous mycobacteria in a developing country. Japanese journal of infectious diseases, 2011. 64 (4): p. 265-71. 80. Dos Santos, R.P., K. Scheid, and L.Z. Goldani, Disseminated nontuberculous mycobacterial disease in patients with acquired immune deficiency syndrome in the south of Brazil. Tropical doctor, 2010. 40 (4): p. 211-3. 81. Mwandumba, H.C., et al., Alveolar macrophages from HIV-infected patients with pulmonary tuberculosis retain the capacity to respond to stimulation by lipopolysaccharide. Microbes and infection / Institut Pasteur, 2007. 9(9): p. 1053- 60. 82. Joshi, P.C., et al., HIV-1-transgene expression in rats decreases alveolar macrophage zinc levels and phagocytosis. American journal of respiratory cell and molecular biology, 2008. 39 (2): p. 218-26. 83. Patel, N.R., et al., Impaired M. tuberculosis-mediated apoptosis in alveolar macrophages from HIV+ persons: potential role of IL-10 and BCL-3. Journal of leukocyte biology, 2009. 86 (1): p. 53-60.

101

84. Raghavan S, A.K., Selvaraj P, Immunogenetics of HIV and HIV associated tuberculosis. Tuberculosis, 2012. 92 (1): p. 18-30. 85. Park, H.H., et al., Incidence of tuberculosis and HIV and progression of silicosis and lung function impairment among former Basotho gold miners. American journal of industrial medicine, 2009. 52 (12): p. 901-8. 86. Corbett, E.L., et al., HIV infection and silicosis: the impact of two potent risk factors on the incidence of mycobacterial disease in South African miners. AIDS, 2000. 14 (17): p. 2759-68. 87. Doran, P., et al., An outbreak of tuberculosis affecting cattle and people on an Irish dairy farm, following the consumption of raw milk. Irish veterinary journal, 2009. 62 (6): p. 390-7. 88. Alemayehu, R., M. Girmay, and A. Gobena, Bovine tuberculosis is more prevalent in cattle owned by farmers with active tuberculosis in central Ethiopia. Veterinary journal, 2008. 178 (1): p. 119-25. 89. de la Rua-Domenech, R., Human Mycobacterium bovis infection in the United Kingdom: Incidence, risks, control measures and review of the zoonotic aspects of bovine tuberculosis. Tuberculosis, 2006. 86 (2): p. 77-109. 90. Inangolet, F.O., et al., A cross-sectional study of bovine tuberculosis in the transhumant and agro-pastoral cattle herds in the border areas of Katakwi and Moroto districts, Uganda. Tropical animal health and production, 2008. 40 (7): p. 501-8. 91. Karimuribo, E.D., et al., Studies on mastitis, milk quality and health risks associated with consumption of milk from pastoral herds in Dodoma and Morogoro regions, Tanzania. Journal of veterinary science, 2005. 6(3): p. 213-21. 92. Cosivi, O., et al., Zoonotic tuberculosis due to Mycobacterium bovis in developing countries. Emerging infectious diseases, 1998. 4(1): p. 59-70. 93. Mawak, J., et al., Human pulmonary infections with bovine and environment (atypical) mycobacteria in jos, Nigeria. Ghana medical journal, 2006. 40 (4): p. 132-6. 94. Aliyu, G.G., Informed consent of very sick subjects in Nigeria. Accountability in research, 2011. 18 (4): p. 289-96. 95. Weinstein, J.P., S.R. Hedges, and S. Kimbrough, Characterization and aerosol mass balance of PM2.5 and PM10 collected in Conakry, Guinea during the 2004 Harmattan period. Chemosphere, 2010. 78 (8): p. 980-8. 96. Roberts, L., Infectious disease. An ill wind, bringing meningitis. Science, 2008. 320 (5884): p. 1710-5. 97. Sultan, B., et al., Climate drives the meningitis epidemics onset in west Africa. PLoS medicine, 2005. 2(1): p. e6. 98. Mohapatra, H., S. Goswami, and D. Dey, Coalmine dust concentration and rate of tuberculosis infection around Ib Valley Coalfield, Orissa, India. Journal of environmental biology / Academy of Environmental Biology, India, 2010. 31 (6): p. 953-6. 99. Tiwari, R.R., Y.K. Sharma, and H.N. Saiyed, Tuberculosis among workers exposed to free silica dust. Indian journal of occupational and environmental medicine, 2007. 11 (2): p. 61-4.

102

100. Pasula, R., et al., Airway delivery of silica increases susceptibility to mycobacterial infection in mice: potential role of repopulating macrophages. Journal of immunology, 2009. 182 (11): p. 7102-9. 101. Rees, D. and J. Murray, Silica, silicosis and tuberculosis. The international journal of tuberculosis and lung disease : the official journal of the International Union against Tuberculosis and Lung Disease, 2007. 11 (5): p. 474-84. 102. Ghio, A.J., et al., Hypothesis: is lung disease after silicate inhalation caused by oxidant generation? Lancet, 1990. 336 (8721): p. 967-9. 103. Hamilton, R.F., Jr., S.A. Thakur, and A. Holian, Silica binding and toxicity in alveolar macrophages. Free radical biology & medicine, 2008. 44 (7): p. 1246-58. 104. Lawande, R.V. and G.C. Onyemelukwe, Airborne fungi during harmattan in Zaria, Nigeria. Annals of allergy, 1984. 52 (1): p. 47-9. 105. Anjorin, F.I., et al., A case of blastomycosis from Zaria, Nigeria. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1984. 78 (5): p. 577-80. 106. Niobe-Eyangoh, S.N., et al., Genetic biodiversity of Mycobacterium tuberculosis complex strains from patients with pulmonary tuberculosis in Cameroon. Journal of clinical microbiology, 2003. 41 (6): p. 2547-53. 107. Traore, B., et al., Molecular strain typing of Mycobacterium tuberculosis complex in Bamako, Mali. The international journal of tuberculosis and lung disease : the official journal of the International Union against Tuberculosis and Lung Disease, 2012. 108. Gomgnimbou, M.K., et al., Spoligotyping of Mycobacterium africanum, Burkina Faso. Emerging infectious diseases, 2012. 18 (1): p. 117-9. 109. Frothingham, R., et al., Phenotypic and genotypic characterization of Mycobacterium africanum isolates from West Africa. Journal of clinical microbiology, 1999. 37 (6): p. 1921-6. 110. Nwofor A., I.M., Parsons L., Peters S., Ndembi N., Gidado M., Blattner W., Abimiku A. and Adbede O., Comparison of three methods in the characterization of mycobacterium isolates from sputum smear positive HIV and non HIV infected Persons in Ilorin Nigeria , 2011: Abuja, Nigeria. 111. Mbada C.E. , A.R.A., Odejide A.S., Relationship Between Socioeconomic Status and Body Mass Index Among Adult Nigerians. AJPARS, 2009. 1(1): p. 1-6. 112. Chew, W.K., et al., Clinical evaluation of the Mycobacteria Growth Indicator Tube (MGIT) compared with radiometric (Bactec) and solid media for isolation of Mycobacterium species. Journal of medical microbiology, 1998. 47 (9): p. 821-7. 113. Tortoli, E., et al., Use of BACTEC MGIT 960 for recovery of mycobacteria from clinical specimens: multicenter study. Journal of clinical microbiology, 1999. 37 (11): p. 3578-82. 114. Pfyffer, G.E., et al., Comparison of the Mycobacteria Growth Indicator Tube (MGIT) with radiometric and solid culture for recovery of acid-fast bacilli. Journal of clinical microbiology, 1997. 35 (2): p. 364-8. 115. Monkongdee, P., et al., Yield of acid-fast smear and mycobacterial culture for tuberculosis diagnosis in people with human immunodeficiency virus. American journal of respiratory and critical care medicine, 2009. 180 (9): p. 903-8. 116. Seagar, A.L., et al., Evaluation of the GenoType Mycobacteria Direct assay for the simultaneous detection of the Mycobacterium tuberculosis complex and four

103

atypical mycobacterial species in smear-positive respiratory specimens. Journal of medical microbiology, 2008. 57 (Pt 5): p. 605-11. 117. Somoskovi, A., et al., Direct comparison of the genotype MTBC and genomic deletion assays in terms of ability to distinguish between members of the Mycobacterium tuberculosis complex in clinical isolates and in clinical specimens. Journal of clinical microbiology, 2008. 46 (5): p. 1854-7. 118. Kiraz, N., et al., Evaluation of the GenoType Mycobacteria Direct assay for direct detection of the Mycobacterium tuberculosis complex obtained from sputum samples. Journal of medical microbiology, 2010. 59 (Pt 8): p. 930-4. 119. Shi, R., N. Itagaki, and I. Sugawara, Overview of anti-tuberculosis (TB) drugs and their resistance mechanisms. Mini reviews in medicinal chemistry, 2007. 7(11): p. 1177-85. 120. Samandari, T., et al., 6-month versus 36-month isoniazid preventive treatment for tuberculosis in adults with HIV infection in Botswana: a randomised, double-blind, placebo-controlled trial. Lancet, 2011. 377 (9777): p. 1588-98. 121. Akolo, C., et al., Treatment of latent tuberculosis infection in HIV infected persons. Cochrane database of systematic reviews, 2010(1): p. CD000171. 122. WHO, Guidelines for intensified tuberculosis case-finding and isoniazid preventive therapy for people living with HIV in resource-constrained settings. , 2011. 123. Drobniewski, F., et al., A national study of clinical and laboratory factors affecting the survival of patients with multiple drug resistant tuberculosis in the UK. Thorax, 2002. 57 (9): p. 810-6. 124. Balabanova, Y., et al., Survival of civilian and prisoner drug-sensitive, multi- and extensive drug- resistant tuberculosis cohorts prospectively followed in Russia. PloS one, 2011. 6(6): p. e20531. 125. Chakroborty, A., Drug-resistant tuberculosis: an insurmountable epidemic? Inflammopharmacology, 2011. 19 (3): p. 131-7. 126. Sungkanuparph, S., et al., Impact of drug-resistant tuberculosis on the survival of HIV-infected patients. The international journal of tuberculosis and lung disease : the official journal of the International Union against Tuberculosis and Lung Disease, 2007. 11 (3): p. 325-30. 127. W.H.O, Anti-tuberculosis Drug Resistance in the World Report No. 4 , 2008. 128. Lew, W., et al., Initial drug resistance and tuberculosis treatment outcomes: systematic review and meta-analysis. Annals of internal medicine, 2008. 149 (2): p. 123-34. 129. Suchindran, S., E.S. Brouwer, and A. Van Rie, Is HIV infection a risk factor for multi-drug resistant tuberculosis? A systematic review. PloS one, 2009. 4(5): p. e5561. 130. Hesseling, A.C., et al., High prevalence of drug resistance amongst HIV-exposed and -infected children in a tuberculosis prevention trial. The international journal of tuberculosis and lung disease : the official journal of the International Union against Tuberculosis and Lung Disease, 2012. 16 (2): p. 192-5. 131. Chum, H.J., et al., An epidemiological study of tuberculosis and HIV infection in Tanzania, 1991-1993. AIDS, 1996. 10 (3): p. 299-309. 132. Mac-Arthur, A., et al., Characteristics of drug resistance and HIV among tuberculosis patients in Mozambique. The international journal of tuberculosis and

104

lung disease : the official journal of the International Union against Tuberculosis and Lung Disease, 2001. 5(10): p. 894-902. 133. Lawson, L., et al., Resistance to first-line tuberculosis drugs in three cities of Nigeria. Tropical medicine & international health : TM & IH, 2011. 134. Kehinde, A.O., et al., Multidrug resistance to Mycobacterium tuberculosis in a tertiary hospital. Journal of the National Medical Association, 2007. 99 (10): p. 1185-9. 135. Wood, R., et al., Tuberculosis control has failed in South Africa--time to reappraise strategy. South African medical journal = Suid-Afrikaanse tydskrif vir geneeskunde, 2011. 101 (2): p. 111-4. 136. Moodley, P., et al., Spread of extensively drug-resistant tuberculosis in KwaZulu- Natal province, South Africa. PloS one, 2011. 6(5): p. e17513. 137. Assam-Assam, J.P., et al., Mycobacterium tuberculosis complex drug resistance pattern and identification of species causing tuberculosis in the West and Centre regions of Cameroon. BMC infectious diseases, 2011. 11 : p. 94. 138. Bassey, E.B., et al., The trend of pulmonary tuberculosis in patients seen at DOTS clinics in the Federal Capital Territory, Abuja, Nigeria. Public health, 2005. 119 (5): p. 405-8. 139. Iliyasu, Z. and M. Babashani, Prevalence and predictors of tuberculosis coinfection among HIV-seropositive patients attending the Aminu Kano Teaching Hospital, northern Nigeria. Journal of epidemiology / Japan Epidemiological Association, 2009. 19 (2): p. 81-7. 140. Daniel, O. and E. Osman, Prevalence and risk factors associated with drug resistant TB in South West, Nigeria. Asian Pacific journal of tropical medicine, 2011. 4(2): p. 148-51. 141. Asch, S., et al., Relationship of isoniazid resistance to human immunodeficiency virus infection in patients with tuberculosis. American journal of respiratory and critical care medicine, 1996. 153 (5): p. 1708-10. 142. Lin, P.Y., et al., Lower gastrointestinal tract tuberculosis: an important but neglected disease. International journal of colorectal disease, 2009. 24 (10): p. 1175-80. 143. Tabarsi, P., et al., First-line antituberculosis drug resistance prevalence and its pattern among HIV-infected patients in the national referral tuberculosis centre, Iran. International journal of STD & AIDS, 2009. 20 (8): p. 566-70. 144. Lawson, L., et al., Clinical presentation of adults with pulmonary tuberculosis with and without HIV infection in Nigeria. Scandinavian journal of infectious diseases, 2008. 40 (1): p. 30-5. 145. Katoch, V.M., Infections due to non-tuberculous mycobacteria (NTM). The Indian journal of medical research, 2004. 120 (4): p. 290-304. 146. Suh, J.D., et al., Identification and treatment of nontuberculous Mycobacterium sinusitis. American journal of rhinology & allergy, 2011. 25 (6): p. 421-4. 147. Chen, C.Y., et al., Pulmonary infection caused by nontuberculous mycobacteria in a medical center in Taiwan, 2005-2008. Diagnostic microbiology and infectious disease, 2012. 72 (1): p. 47-51.

105

148. Latshang, T.D., C.M. Lo Cascio, and E.W. Russi, [Nontuberculous mycobacterial infections of the lung]. Therapeutische Umschau. Revue therapeutique, 2011. 68 (7): p. 402-6. 149. Pham-Huy, A., et al., Current trends in nontuberculous mycobacteria infections in Canadian children: A Pediatric Investigators Collaborative Network on Infections in Canada (PICNIC) study. Paediatrics & child health, 2010. 15 (5): p. 276-82. 150. Odaibo, G.N., et al., HIV infection among patients with pulmonary tuberculosis in Nigeria. African journal of medicine and medical sciences, 2006. 35 Suppl : p. 93- 8. 151. Tuberculosis and AIDS. Statement on AIDS and tuberculosis. Geneva, March 1989. Global Programme on AIDS and Tuberculosis Programme, World Health Organization, in collaboration with the International Union Against Tuberculosis and Lung Disease. Bulletin of the International Union against Tuberculosis and Lung Disease, 1989. 64 (1): p. 8-11. 152. Standaert, B., et al., The association of tuberculosis and HIV infection in Burundi. AIDS research and human retroviruses, 1989. 5(2): p. 247-51. 153. American Diabetes Association position statement: evidence-based nutrition principles and recommendations for the treatment and prevention of diabetes and related complications. Journal of the American Dietetic Association, 2002. 102 (1): p. 109-18.

106