Treatment outcome of tuberculosis patients at National Hospital Abuja Nigeria: a five year retrospective study

Abstract

Background: The HIV/AIDS epidemic in Nigeria has militated against efforts to control TB. This study determined the differences in TB treatment outcome between patients with TB/HIV co-infection and those without HIV co-infection in National Hospital Abuja, Nigeria.

Methods: This was a retrospective cohort study on pulmonary TB patients that were treated in National Hospital Abuja, Nigeria from January 2007 to December 2012. Data was collected from the Directly Observed Treatment Shortcourse (DOTS) clinic and hospital records and analysed using stata version 12. The HIV sero-prevalence was determined and TB treatment outcomes in patients co-infected with TB and HIV was compared to that of HIV negative patients. The t-test and chi square test was used to verify differences in means and proportions and multiple logistic regression to adjust for potential confounders. Sensitivity analysis was done to address the problems of loss to follow up and missing data.

Results: A total of 389 cases were assessed. The HIV sero-prevalence rate in the study was 42.7%. Fifty-three percent of the study population were within the age category 30–49 years. There was strong evidence of an association between TB/HIV co-infection rate among the various age categories (p = 0.01). HIV positive cases had a lower treatment success rate, 48.8% vs. 78.5% (p < 0.001), a higher rate of treatment failure, 10.8% vs. 4% (p = 0.01), and a higher rate of default, 38.6% vs. 17% (p < 0.001).

Conclusion: The study provides evidence that TB/HIV co-infection impacts negatively on TB treatment outcome.

Keywords:

• DOTS
• TB treatment outcome
• TB/HIV co-infection
• tuberculosis

Introduction

Tuberculosis (TB) is a curable infectious disease caused by the mycobacterium tuberculosis.¹ It is an airborne infectious disease that primarily affects the lungs (pulmonary) but can affect other parts of the body like the kidneys, lymph nodes, spinal cord and the abdomen (extra pulmonary).¹

HIV/AIDS, on the other hand, is a non-curable disease caused by the human immunodeficiency virus (HIV).² It is transmitted via body fluids, for example, blood, seminal and vaginal fluids.² It affects the human immune system destroying them and impairing their function and progressively diminishing the body’s ability to fight infections and certain cancers.²

The intricate linkage of tuberculosis (TB) and human immunodeficiency virus (HIV) infection for nearly the past 3 decades poses a major threat to the international community’s effort to achieve the health-related United Nations Millennium Development Goals for TB and HIV infection.¹ These two diseases are a deadly combination and are far more dangerous occurring together than either disease alone. According to the World Health Organization (WHO) TB/HIV co-infection can be referred to as two monsters working against humanity.

With the increasing number of individuals co-infected with HIV and TB, early recognition, diagnosis, prevention, prophylaxis and treatment of TB has become more challenging. TB/HIV co-infection may affect the efficacy, safety and tolerability of treatment, including differences in clinical presentation and differences in bacteriological responses to treatment.²

Co-infection is said to occur when both diseases occur in the same individual at a point in time irrespective of which disease preceded the other. Concomitant HIV infection and TB is a lamentable medical phenomenon with dreadful social and economic impact across the globe, aptly described as the ‘accursed duet’.³ HIV and mycobacterium TB have a synergistic interaction; each accentuates progression of the other.⁴

TB/HIV co-infection poses an enormous challenge to TB control especially in resource poor settings like Nigeria where the majority of urban population live in slums. Key features of slum life, such as crowded housing, working conditions with poor ventilation, poor nutrition, and lack of access to quality health care, continue to drive TB transmission.⁵ Nigeria’s HIV epidemic is largely fuelled by heterosexual and mother-to-child transmission.⁶

Globally, the TB epidemic is fuelled by the HIV epidemic. Even if all new HIV infections were prevented, TB incidence would increase due to the high risk of TB progression among prevalent HIV cases.⁷ The highest rates of HIV co-infection in TB patients are in the African Region in contrast to the rates being recorded in the more industrialised parts of the world. In New York City (NYC), United States of America, HIV co-infection among TB cases has decreased from 34% to 13% since 1992.⁷ In Nigeria, the TB burden is compounded by a high prevalence of HIV in the country which stands at about 4.1% in general population.⁸

Studies done in Oyo and Gombe states in Nigeria^9,10 showed poorer TB treatment outcomes in terms of cure and mortality rates, in patients dually infected with HIV and TB compared to those without HIV infection. Similar results were obtained in another study done in Togo.¹¹ Factors identified to be responsible for these poor outcomes include low CD4 counts and availability of health care services including the highly active antiretroviral therapy (HAART). Older age, low haemoglobin levels, higher viral load, and the presence of other opportunistic infections like oral thrush at baseline are predictors of mortality among patients dually infected with TB and HIV.¹² HIV positive patients with a history of homelessness before or during treatment with anti-tuberculosis medications and those who are sputum AFB smear positive have also been said to be less likely to have treatment success.⁷

Justification of the study

Tuberculosis is almost always curable if patients with drug susceptible organisms are given sufficient uninterrupted therapy. If untreated, 50%–80% of patients with smear positive TB die and in a poorly implemented TB programme, as many as 30% of patients with smear positive TB die.¹² Despite the fact that TB is treatable and curable, it has proved impossible to eliminate and this has been worsened by the HIV/AIDS epidemic. Although there has been recent progress of global efforts, TB is still one of the leading causes of mortality and morbidity worldwide by an infectious agent and still remains a major public health burden globally and especially in developing countries like Nigeria.

The National Hospital Abuja, Nigeria (NHA) is one of the top ranking hospitals in the country with a very functional and fully equipped DOTS clinic for the effective management of TB patients. It is one of the many DOTS centres in the country. Evidence abounds from within and outside the National TB programme that HIV impacts negatively to TB treatment outcome. To address these challenges, the NHA strengthened its TB/HIV collaborative activities by setting up a hospital based TB/HIV team to ensure cross referral of TB/HIV patients and timely provision of treatment, care and other services. This has made TB treatment readily available to the dually infected patients. Since the inception of DOTS clinic as well as the increased collaborative activities in NHA, no study has been done to assess the TB treatment outcomes in HIV positive or HIV negative patients. Periodic retrospective analysis of outcome after treatment of smear positive pulmonary TB cases has been recommended as the standard method of evaluating TB control activities through the calculation of ‘rates’ corresponding to the various TB treatment outcomes.¹³ Treatment outcome results serve as a proxy of the quality of TB treatment provided by the health care system.¹⁴ This study would help highlight any existing differences in TB treatment outcome in both groups of patients and would inform the proper future management of these patients.

Aim of the study

To determine the differences in TB treatment outcomes between patients with TB/HIV co- infection and those without HIV co-infection in National Hospital Abuja, Nigeria

Study objectives

1. To determine the prevalence of HIV among the new smear positive pulmonary TB patients.

2. To compare the age and sex distribution among the TB patients with HIV co-infection and those without HIV co-infection.

3. To compare the mortality rate among the HIV positive TB patients and those who are HIV negative, controlling for known confounders.

4. To establish the difference in treatment success rate (cure and treatment completion) between the TB patients co-infected with HIV and those without HIV co-infection when treated with short course chemotherapy, controlling for known confounders

Methodology

Study design

A retrospective hospital based cohort study.

Study population

The study looked at all new sputum category 1 smear positive acid fast bacilli (AFB) patients who commenced TB treatment in the NHA DOTS clinic over the past five years (January 2007–December, 2012). The National Hospital Abuja (NHA) is a tertiary health care facility situated in the Federal Capital Territory, Abuja, Nigeria.

Sampling method

Five year data (January 2007–May 2012) of all patients newly diagnosed with pulmonary TB in the DOTS clinic of National Hospital Abuja, Nigeria were assessed.

Sources of data and method of data collection

The TB treatment register and hospital patient folders were the primary sources of data in this study. A specially designed data collection tool was used to collect the data (Appendix 1).

Statistical analysis

Data was analysed using Stata version 12 variables were imported into Stata, cleaned and processed. Continuous explanatory variable (age) was categorised. Student’s t-tests were used to obtain differences between means (mean age within each sex). To assess the distribution of HIV status amongst age groups, sex and other categorical variables including the outcome variables, the chi-squared tests was used. The level of significance was set at 5%.

To investigate the differences in treatment outcome between HIV positive and HIV negative cases, the Odds ratio was determined for each outcome individually in the crude form, and later adjusted for the potential confounders-age, sex, marital status and level of education.

Multiple logistic regression models were also fitted to examine the relationship between the HIV status and treatment outcome, taking into account all the potential confounders. The complete case analysis was used to handle the missing values as they had already been determined to be missing at random.

Ethical considerations

The study protocol was submitted to the National Hospital Abuja Ethics Committee for approval before data collection. To ensure confidentiality, anonymity was maintained from data entry into the data collection tool and into the computer before data analysis.

Results

This study comprises 389 adults in total. There were 234 (60%) males and 155 (40%) females with ages ranging between 18 and 77 years. The mean age in males was 37.12 ± 10.94 years and 32.64 ± 10.64 years in females. There was a statistically significant difference in the mean ages of males and females (p < 0.001). One hundred and thirty-six cases (35%) were below 30 years, 205 (53%) within the 30 to 49 age category, 38 (10%) between 50 and 59 years and 10(2%) above 60 years old.

Ninety-nine (26%) of the cases had no education. Thirty-seven (10%), 141 (36%) and 99 (26%) had primary, secondary and tertiary level of education respectively. Thirteen (3%) of the cases had this variable missing and it was determined to be missing at random. Two hundred and twenty (57%) of the study population were married. One hundred and thirty-four (35%), 19 (4%) and 13 (3%) were single, separated and divorced respectively. Three (1%) of the cases had this variable missing and was also determined to be missing at random. Of the 389 adult cases studied, 166 (42.7%) were confirmed HIV positive.

The highest percentage of cases with TB/HIV co infection was in the age group 30 to 49 years accounting for 49.8% of the cases. There was quite strong evidence of a difference in the co infection rate among different age groups (p = 0.01). Of the 234 males, 134 (57.3%) were HIV negative and 100 (42.7%) were positive. Sixty-six (42.6%) of the 155 females were HIV positive and 89 (57.4%) were negative. There was no evidence of a difference in sex of the cases in relation to their HIV status (p = 0.98).

Table 1 shows the age, sex, marital status and level of education distribution in relation to HIV status in this study. There is no evidence of an association between HIV status and sex (p = 0.9), very little evidence of an association between HIV status and marital status (p = 0.1), and some evidence of an association between HIV status and level of education (p = 0.09).

Table 1: Age, sex, marital status and level of education of study population in relation to HIV status

Characteristic	HIV positive [n (%)]	HIV negative [n (%)]	Total no. of TB cases	p-value
Age Group (years)
<30	50 (36.8)	86 (63.2)	136	0.01^*
30–49	102 (49.8)	103 (50.2)	205
50–59	13 (34.2)	25 (65.8)	38
≥ 60	1 (10.0)	9 (90.0)	10
Sex
Male	100 (42.7)	134 (57.3)	234	0.98
Female	66 (42.6)	89 (57.4)	155
Total	166	223	389
Marital Status
Married	85 (38.6)	135 (61.4)	220	0.10
Single	66 (49.3)	68 (50.7)	134
Separated	11 (57.9)	8 (42.1)	19
Divorced	4 (30.8)	9 (69.2)	13
Total	166	220	386
Level of Education
Nil	40 (40.4)	59 (59.6)	99	0.09
Primary	23 (62.2)	14 (37.8)	37
Secondary	58 (41.1)	83 (58.9)	141
Tertiary	39 (39.4)	60 (60.6)	99
Total	160	216	376

^* Statistically significant (chi-square test).

A cure rate of 48.8% was obtained in this study. Sixty-six (17.0%) cases of the study population completed treatment. Overall, a treatment success rate of 256 (65.8%) was obtained. Twenty-seven (6.9%), 102 (26.2%) and four (1.0%) cases were the recorded for treatment failure, default and dead respectively. Table 2 shows TB treatment outcomes by HIV status. There was very strong evidence of an association between HIV status and cure rate and default with p-values of less than 0.001 in treatment outcomes. With a p-value of 0.01, the evidence of an association between HIV status and treatment failure was strong. The odds of achieving treatment cure, with TB/HIV co-infection, is 0.3 compared to if the patient is not infected with HIV. If a patient is HIV positive, he/she is three times more likely to have treatment failure and three times more likely to default, compared with if he/she is HIV negative.

Table 2: Treatment outcome in relation to HIV status

Outcome	HIV positive (n = 166)	HIV negative (n = 223)	OR (95% CI)	P value
Cured	53 (31.9%)	137 (61.4%)	0.29 (0.19, 0.46)	<0.001
Treatment Completed	28 (16.9%)	38 (17.0%)	0.99 (0.58, 1.69)	0.960
Treatment Success	81 (48.8%)	175 (78.5%)	0.26 (0.16, 0.42)	<0.001
Treatment Failure	18 (10.8%)	9 (4.0%)	2.89 (1.25, 6.67)	0.010
Died	3 (1.8%)	1 (0.5%)	4.09 (0.42, 39.95)	0.190
Default	64 (38.6%)	38 (17.0%)	3.05 (1.88, 4.95)	<0.001

The significant relationship between default and HIV status remained after adjusting for age and marital status which appeared to be confounders. Although there was no evidence of an association between HIV status and treatment completed (p = 0.96), there is very strong evidence of an association between HIV status and the overall treatment success (the sum of treatment completed and cure) with a p-value of less than 0.001. The treatment success rate for the HIV positive patients was 48.8% (80) and for the HIV negative cases it was 78.5% (175). As shown in Table 3, multiple regression highlighted the significant relationship between treatment success and HIV status (p < 0.001) after adjusting for sex and marital status which appeared to be confounders.

Table 3: Regression coefficients and p values of the relationship between treatment outcomes and explanatory variables

	Treatment outcome [regression coefficient (p-values)]
Variables	Treatment cure	Treatment completed	Treatment success	Treatment failure	Dead	Default
Age	−0.002 (0.90)	−0.001 (0.87)	−0.001(0.80)	−0.001(0.26)	0.005(0.24)	−0.001(0.51)
Sex	0.115 (0.02)^*	−0.029 (0.46)	0.085(0.08)	−0.054(0.08)	−0.004(0.68)	−0.043(0.43)
Level of education	0.041 (0.07)	−0.018 (0.31)	0.015(0.28)	−0.014(0.37)	−0.010(0.20)	0.011(0.70)
Marital status	−0.120 (<0.001)^*	0.050 (0.10)	−0.065(0.06)	−0.023(0.39)	−0.010(0.20)	0.112(0.01)^*
HIV status	−0.295 (<0.001)^*	−0.002 (0.96)	−0.297(<0.001)^*	0.068(0.01)^*	0.014(0.19)	0.215(<0.001)^*

^* Statistically significant.

A higher mortality rate of 1.8% (3) was recorded amongst the HIV positive cases compared with HIV negative cases where the mortality rate obtained was 0.5% (1). With a total mortality rate of just 1% in the entire study, it was difficult to compare the difference in mortality rate in HIV positive and HIV negative cases after controlling for potential confounders. There appeared to be very little evidence of an association between HIV status and the outcome dead [p = 0.19, OR (95% CI) = 4.09 (0.42, 39.95)].

Definition of terms

Cured: a TB patient who was smear positive at diagnosis, who completed the course of 8 months treatment and was smear negative at the end of the 7th month of treatment.

Treatment completed: any patient who was smear positive at diagnosis and who completed treatment but in whom smear examination results were not available at the end of treatment. This includes all smear-negative and extra-pulmonary patients who completed treatment.

Treatment success: This is the sum total of all the patients that were cured and all those that completed treatment.

Treatment failure: A smear positive TB patient who while on the first line of treatment remained or becomes smear positive again 5 months or later after commencement of treatment.

Died: Any patient who dies for any reason during the course of his or her treatment.

Default: A TB patient who completed at least one month of treatment and returns after interrupting treatment for two months or more.

Discussion

This study has established that being co-infected with HIV and TB potentially increases the risk of a poorer TB treatment outcome compared with if the patient with TB is HIV negative. This study revealed that there is an increased chance of treatment failure, default and death if the patient is dually infected with TB and HIV, but more likely to achieve treatment success (the sum of treatment cure and treatment completed) if HIV negative. The findings of this study are in line with the current notion that HIV positivity has a negative impact on TB treatment outcome.

The prevalence of HIV among TB patients in Nigeria increased from 2.2% in 1991 to 19.1% in 2001 and 25% in 2010.⁶ This is different from what has been obtained in the developed parts of the world where the prevalence of HIV in TB patients is on the decline. In the United States of America, HIV co-infection among TB cases has decreased from 34% to 13% since 1992.⁷ This shows the TB situation in low resource countries like Nigeria may be explained by the HIV pandemic. Today 50% or more of new TB cases are also HIV co-infected in southern and eastern Africa, which is the centre of HIV/TB epidemic.¹⁵

The seroprevalence of HIV among TB patients recorded in this study was 42.7%. This is higher than the findings in other Nigerian studies.^9,10 The higher rate in this study can be explained by fact that NHA is a flagship referral centre where ART and DOTS services are offered and as such, most co-infected patients, residing in the federal capital territory and those in neighbouring states, would seek medical care in this hospital. However, a study in Accra, Ghana,¹⁶ reported similar findings, and another in Zimbabwe reported a much higher rate of 63.3%.¹⁷

HIV/TB co-infection is more common in the sexually active age group.¹⁸ Results of this study show the highest percentage of cases with TB/HIV co-infection was in the age group 30–49 years. This age group is a sexually active age group and can explain the high percentage of TB/HIV co-infection found in this age group. This finding is similar to findings in some studies,¹⁰ but other studies⁹ have recorded a high percentage of TB/HIV co- infection in a lower age group (20–29 years). This disparity can be explained by the fact that, most of the participants in this study fall into the age group 30–49 and hence can explain the high percentage of dual infection recorded in this age group.

In sub-Saharan Africa, gender inequality is an important driver of the HIV epidemic accounting for 60% of the infection.⁸ Various reviews suggest that women in many parts of the developing world are less likely to control how, when and where sex takes place thereby increasing the likelihood of HIV infection.¹⁰ The prevalence of HIV has been reported to be higher among females but findings in this study show an almost equal prevalence in both males and females (42.7% and 42.6% respectively).This may be a reflection of the increasing risky behavioural patterns in African men. This finding is also similar to that of other studies.^9,19 However, a study conducted in South Africa²⁰ found a higher rate of HIV infection in males.

The treatment success rate (a sum of the rates for treatment cure and completed) recorded in this study was 78.5% and 48.8% in the HIV negative and HIV positive cases respectively and there was strong evidence of its association with HIV status with P values less than 0.001. These percentages obtained are below the 85% target for tuberculosis control set by the World Health Assembly. Similar findings have been obtained in other low and middle income countries.^9,11 Studies done in high income countries,^7,21 however, have reported high success rates even in HIV positive individuals. This emphasizes that TB has remained a disease of poverty and disproportionately affects the poor countries and marginalised population and this is compounded by the HIV/AIDS pandemic in these parts of the world.

The cure rate obtained in this study was 31.9% in the HIV positive cases and 61.4% in the HIV negative. This difference was statistically significant as shown in Table 4. Although some studies have reported no difference in the cure rates in both groups,² several studies have also reported lower cure rates in HIV positive individuals compared to negative ones.^10,22 The marked difference in TB treatment success rate, in patients not HIV positive and those dually infected, reflects the enormous contribution that HIV co-infection has towards obtaining a low treatment success of anti-tuberculosis treatment.

Table 4: Results of treatment outcome of study subjects

Outcome	HIV positive(n=166)	HIV negative(n=223)	OR (95% CI)	p-value
Treatment Cure	53(31.9%)	137(61.4%)	0.29(0.19,0.46)	<0.001
Treatment completed	28(16.9%)	38(17.0%)	0.99(0.58,1.69)	0.96
Treatment success	81(48.8%)	175(78.5%)	0.26(0.16,0.42)	<0.001
Treatment failure	18(10.8%)	9(4.0%)	2.89(1.25,6.67)	0.01
Died	3(1.8%)	1(0.5%)	4.09(0.42,39.95)	0.19
Defaulted	64(38.6%)	38(17.0%)	3.05(1.88,4.95)	<0.001

A study carried out in another part of the country¹⁰ suggested that a better treatment success rate would be obtained if the drug regimen for TB treatment is changed from the regimen that consists of a 2-month intensive phase of Rifampicin, Isoniazide, Pyrizinamide and Ethambutol followed by a 6-month continuation phase of Ethambutol and Isoniazide (2RHZE/6EH) to the regimen that consists of a 2-month intensive phase of Rifampicin, Isoniazide, Pyrizinamide and Ethambutol then a 4-month continuation phase of Rifampicin and Isoniazide (2RHZE/4RH) used in most developed countries. The NHA DOTS clinic commenced the suggested 6 month course regimen in December 2011. All the patients recruited in this study underwent the 8-month course, so the low treatment success rate obtained in this study may be due to the drug regimen used then.

Alarmingly, African countries’ HIV positive TB mortality rates are said to be 29.9 times higher than non-African countries.²³ In this study, mortality rate was higher in the HIV positive individuals compared to their negative counterparts. This finding is similar to that from several African^11–13 and non-African studies.^21,24,25 Though mortality rate from HIV associated TB in developing countries is high, it is not clear whether it is due to failure of anti-TB treatment or complications of HIV.²⁶ Factors that have been attributed to the higher mortality rate amongst the TB/HIV patients include immunosuppression,²⁴ unavailability/inaccessibility of ARTs,⁷ late diagnosis of HIV,¹⁹ smear negative pulmonary TB or extra pulmonary TB¹³ and not having a treatment supporter.¹⁶ The presence of other morbidities like neoplastic diseases,²² respiratory, and cardiovascular diseases²⁴ as well as diabetes mellitus²⁴ in HIV positive TB patients, have also been implicated in the increased mortality in these patients.

In this study, an overall mortality rate of just 1% was obtained and thus comparing the rates in both HIV positive and negative groups after adjusting for confounders was difficult. This low mortality rate obtained in this study is very likely due to the fact that most deaths in this part of the world occur outside the hospitals and there is hardly any system in place to record these deaths that occur anywhere but in hospitals. This highlights the paucity of these data used in this study. Reducing mortality due to TB in endemic countries in programmes that have fully implemented the DOTS strategy may require additional interventions that are tailored to the local situation.¹⁶ This would help in the proper management of TB patients who are HIV positive as well as those who are HIV negative.

Comparing the HIV positive and negative cases in this study, a higher default rate was obtained in the HIV positive group, 17% in the HIV negative cases and 38.6% in the HIV positive ones. This difference was statistically significant and again, it reflects the poorer TB treatment outcomes in HIV positive individuals. This finding is similar to that of several other studies.^9,10,22 The default rate established, in the HIV positive TB cases in this study, is higher than the rates reported in patients dually infected with TB and HIV in other countries with high TB burden like Nigeria. In Brazil, one of the 22 countries with a high TB burden, the percentage of TB treatment default reported in HIV positive subjects is 14%.²⁷ This disparity in the rates of default may reflect a weakness in the TB/HIV collaborative services in the National Hospital Abuja, Nigeria. This calls for an urgent review of the existing health services available to patients co-infected with TB and HIV.

A study carried out in The Gambia established that patients with high bacterial load in initial sputum smears were more likely to default from treatment,¹³ and in another study carried out in Brazil risk factors identified for TB treatment default were male sex, smoking and a CD4 count of less than 200 cells/mm³ while a complete or incomplete secondary or university education was associated with a lower risk of defaulting from TB treatment.²⁷ Enhanced patient counselling and education on TB has been suggested to help reduce the rates of default in TB patients.²⁸ Patients who default tend to be at higher risk of treatment failure. Default from TB treatment is associated with a longer period of TB transmission and mortality.²⁷

In this study, the treatment failure rate was higher in the dually infected cases (10.8%) compared with those not co- infected with HIV patients (4%). HIV is associated with both malabsorption of TB drugs and higher rates of TB treatment failure and, hence, may be a risk factor for TB drug resistance.²⁹ Individuals who are dually infected with TB and HIV are more exposed to a high burden of pills and are thus more likely to stop taking their drugs and thus default from treatment and eventually have treatment failure. It is also likely that the high default and treatment failure rate obtained in this study may be related to the poor knowledge about the gravity of both infections and the attitude of these individuals to health. Patient non-adherence to treatment is interpreted as a failure of the health care system to cope with the natural tendency of humans to quit treatment as soon as they feel subjectively better, or better without treatment if adverse drug events supervene.¹⁴ The use of treatment supporters at the health facilities and at home would help ensure proper treatment adherence and hence less default, treatment failure and death amongst both HIV positive and negative tuberculosis patients.³⁰ The prevention and control of drug resistant TB would also be ensured.

Study limitations

1. In the course of data collection for the study, some variables were found to be missing. But during analysis of the data, these missing variables (<5%) were determined to be missing at random using the chi squared test. This was handled in regression model using complete case analysis.

2. The study analysis could not adjust for potential confounders (age, sex, marital status and level of education) when assessing for the mortality rate in those dually infected with TB and HIV compared to those not dually infected, due to the small number of mortality outcome in the data (overall mortality rate of just 1% in the entire study).

3. Certain potential confounders, for example the existence of co-morbidities like diabetes mellitus and neoplasms which has been noted in other studies could not be included in this study as such data were not available and this may have influenced the results of this study.

4. Since the study was a hospital-based study, generalising the results to wider populations must be done with caution.

Conclusions

This study clearly shows that HIV co-infection impacts negatively on TB treatment outcome. More dually infected patients failed treatment, died and defaulted compared with those who were not co-infected with HIV. A higher treatment success rate was obtained from the TB patients who were HIV negative compared with those who were HIV positive. The result of this research calls for strengthening of TB/HIV collaborative activities to ensure efficient tuberculosis and HIV treatment services for dually infected patients.

Recommendations

1. A repeat of this study in NHA after the new drug regimen (2RHZE/4RH) has been in place for a considerable period of time. This would help assess its impact on TB treatment outcome in both HIV positive and negative TB patients.

2. Bearing in mind that HIV positivity confers an increased likelihood of a poorer TB treatment outcome as established in this study, it is important that early diagnosis and treatment of both diseases is ensured hence a review of the diagnostic methods used in National Hospital Abuja should be done.

3. To improve the quality of data available for future studies, data collection methods as well as data storage need to be improved. This can be achieved with the use of electronic medical record keeping. This would help reduce the incidence of missing data and in turn aid in obtaining more valid and reliable results in future research studies.

4. To help curtail defaulters especially amongst HIV positive TB patients, it is paramount that proper TB health education, with emphasis on the duration, side effects and the risk associated with disrupting TB treatment, be given at the first visit of every patient to the DOTS clinic and reinforced at subsequent visits. The use of treatment supporters and early referral of patients to DOTS centres closer to their places of residence may also help in achieving a lower rate of default.

5. Strengthening of the TB/HIV collaborative services at all levels and ensure that the strategic frame work and guidelines for TB/HIV collaborative activities are in tune with current trends in implementation of TB/HIV services.

Conflict of interest

None of the authors have any commercial or other association with the work reported here that might pose a conflict of interest.

Sources of support

All the costs incurred in the course of the study were borne by the authors.

Declaration

All the information contained in this manuscript has not been presented elsewhere.

Acknowledgements

We would like to thank the National Hospital Abuja, Nigeria and specifically the DOTS clinic for providing access to the data used in this study.

Appendix 1: Data collection sheet

Record form for sputum smear positives only (tick as appropriate)

Age…………..

Sex…………..

Marital Status: Married [ ], Single [ ], Divorced [ ], Separated [ ], Widowed/Widower [ ].

Level of education: Nil [ ], Primary [ ], Secondary [ ], Tertiary [ ].

HIV Status: Positive [ ], Negative [ ].

TB treatment outcome (tick one from list below)

• Completed treatment [ ].

• Cured [ ].

• Failed treatment [ ].

• Defaulted [ ].

• Died [ ].