An assessment of the risk factors for pulmonary tuberculosis among adult patients suffering from human immunodeficiency virus attending the Wellness Clinic at Themba Hospital

Abstract

Background: In South Africa, the data about risk factors for pulmonary tuberculosis (PTB) is limited. Therefore, the assessment of risk factors for PTB is an important step to identify which risk factors are unique in every specific population context and in this way gain a better understanding of them. The aim of this research was to assess contributory risk factors for PTB among adult patients suffering from Human Immunodeficiency Virus (HIV) attending the Wellness Clinic at Themba Hospital.

Method: A cross sectional descriptive design was employed. A total of 300 participants were interviewed using a structured questionnaire to collect data on socio-demographic information, behavioural factors and medical history of patients in the research. Data were analysed using Stata Release 11 software. Univariate and multivariable logistic regression models were used to determine factors associated with PTB.

Results: The results of this investigation showed that being of single marital status (OR = 1.96; 95% CI = 1.01–3.79), moderate/heavy drinkers (OR = 3.46; 95% CI = 1.56–7.69) and living in a dusty outdoor environment (OR = 2.05; 95% CI = 1.16–3.61) were statistically associated with pulmonary tuberculosis for the study participants in multivariable logistic regression models.

Conclusion: Single marital status, dusty outdoor environment and moderate/heavy drinkers were strongly associated with an increased risk of PTB among HIV positive patients. Pulmonary tuberculosis control and prevention strategies among HIV positive patients should be focused on reducing or limiting the impact of these risk factors.

Keywords:

• risk factors for TB
• TB and age
• TB and alcohol
• TB and gender
• TB and HIV
• TB and smoking
• TB and the environment

Introduction

Tuberculosis (TB) control and management worldwide possess a huge medical and social challenge. In South Africa, the data about risk factors for pulmonary tuberculosis (PTB) among HIV-positive patients is limited. Therefore, the assessment of risk factors for PTB is an important step to identify which risk factors are unique in every specific population context and in this way gain a better understanding of them.

According to Lienhardt, factors reported to influence the risk of mycobacterial infection include age, sex, crowding, socio-economic conditions, urbanisation, racial/ethnic group, and human immunodeficiency virus (HIV) infection. The author suggest that any condition modifying the balance established in the body between the tubercle bacilli and the host’s immune defences can have an impact on the risk of developing the disease.¹

As reported by many researchers, the critical TB epidemic in Africa is attributable to multiple factors, where HIV infection has been recognised as the most important. Even though according to Dye, HIV infection has clearly had a profound effect on TB epidemiology. However, other potentially important risk factors have not been extensively investigated.²

This research aimed to assess contributory risk factors for pulmonary tuberculosis among adult patients suffering from HIV/AIDS attending the Wellness Clinic at Themba Hospital.

Method

Study design

A cross-sectional descriptive design was employed in this study.

Study site

This study was conducted at the Wellness Clinic of Themba Hospital, situated in Kabokweni Township, Ehlanzeni District, Mpumalanga Province in South Africa. Themba Hospital serves a diverse population of approximately 350 000 people. The Wellness Clinic provides a range of health services to a significant number of HIV/AIDS patients on weekdays (Monday to Friday).

Study population

The study population consisted of adult patients (age ≥ 18 years) who were registered for clinical care in the Wellness Clinic at Themba Hospital between January and March 2009.

Sample size

The sample size was estimated based on the study objective to determine the prevalence of PTB among HIV patients. An accuracy of within 5% was required. It was then estimated to be a sample size of 303 patients. This estimate is expected to give a prevalence of 27% with an accuracy of 5% and 95% confidence interval (nQuery Advisor 7 Software). Furthermore, by convention, 10–15 subjects are needed for inclusion in the study when determining the risk factors in a multivariable analysis (logistic regression). Thus, in this study, the researchers collected data from a total of 300 participants which was thought to be adequate to carry out this project.

Participants who are HIV positive and aged 18 years or older attending the Wellness Clinic at Themba Hospital at the time of the interview who had given informed consent to participate in the research were included. Patients below 18 years, HIV negative, too ill or mentally impaired, withheld consent to participate and those diagnosed of extra-pulmonary tuberculosis were excluded from the study.

Sampling method

The researchers used a systematic sampling method. The average number of patients receiving care monthly in the Wellness Clinic was 850. This number was then divided by the sample size of 300. The systematic sampling included every third patient in the queue. The first participant was selected using a random number generator and thereafter every third patient who met the research criteria was invited to participate and when a patient declined participation, the researchers selected the next patient in the queue.

Measuring tool or instrument

The investigators used a structured questionnaire to interview participants and the data were collected by means of the closed ended questions contained in the questionnaire. The questionnaire assessed several of the known important risks factors associated with pulmonary tuberculosis such as demographic variables, socio-economic status, behavioural habits, medical history and environmental factors.

Data collection

A total number of 300 participants were interviewed on a one-to-one basis and questionnaires were administered to collect data from participants. Then data collected were transferred into the pre-built data sheet in the Epi Info programme version 3.5.3 for further statistical analysis.

Every participant interviewed had data on PTB positive (+) or negative (−) status obtained from the medical file. Data on PTB signs and symptoms such as chronic coughs, night sweats, blood stained sputum, and loss of weight were also obtained from the patients during the interviews.

Patients with signs and symptoms of PTB without the diagnosis of pulmonary tuberculosis documented in their medical files, TB screening tests were performed in line with the South African National Tuberculosis diagnosis and management guidelines for TB diagnosis in HIV positive patients.

Pilot study

A pilot study was conducted between January 6 and 8, 2009, in the Wellness Clinic, at Themba Hospital to test the measuring tool and the sampling methods. Ten patients who were not part of the original research were interviewed using the same questionnaire for the study. This assessment helped the researchers to identify the strengths and weaknesses of the questionnaire. It assisted in modifying some of the questions and estimating the average interview duration.

Ethics

The study was approved by the Human Research Ethics Committee (Medical) of the University of the Witwatersrand and approvals to conduct the research were obtained from the Chief Executive Officer of Themba Hospital and the Clinical Manager of the Wellness Clinic of Themba Hospital. The researchers were allocated a private room at the Wellness Clinic to guarantee privacy. All patients who participated in the study were provided with written information and consent forms which were explained to them. Only those patients who consented and signed the consent forms where allowed to proceed with the interview process. The questionnaires were anonymous and could not be linked to the patients.

Data analysis

Data were entered using the Epi Info programme version 3.5.3 which generated an Excel® sheet that was transferred to Stata Release 11 software for analysis. Frequencies, rates and proportions were calculated for the categorical variables (sex, age group, marital status, etc.). A Pearson Chi-square (χ²) test was applied to test for an association between risk factors for PTB among HIV positive patients.

Univariate and multivariable logistic regression models were used to investigate factors associated with PTB. Estimated odds ratios were reported with 95% confidence intervals and in all the analyses a p-value of < 0.05 was taken to be significant.

Results

Socio-demographic information on the participants

Table 1 presents the demographic characteristics of the study sample. The study comprised n = 300 participants. The mean age of patients was 38.2 years with standard deviation (SD = 10.9 years) and ranged from 18–86 years of age. Forty per cent of the participants were aged 31–40 years, with 63.6% of the subjects aged under 41 years. More than half (67.0%) of the study participants were females. In terms of educational level, more than two thirds of the participants (69.3%) had low level educational backgrounds of mainly primary and secondary education. With regard to marital status, more than half (65.7%) of the participants reported being single. The majority of participants (59.7%) were unemployed. A large proportion of the patients (63.3%) reported living in a non-overcrowded environment. Finally, the majority of the study participants (64.7%) reported living in dusty outdoor conditions.

Table 1: Socio-demographic information of the participants

Variables	Frequency (%)
Age group in years
18–30	70 (23.3%)
31–40	121 (40.3%)
41–50	64 (21.3%)
> 50	45 (15.1%)
Sex
Female	201 (67.0%)
Male	99 (33.0%)
Level of education
High School/University degree	68 (22.7%)
Primary/Secondary	208 (69.3%)
Illiterate/No formal education	24 (8.0%)
Marital status
Married/cohabitation	72 (24.0%)
Widowed/divorced	31 (10.3%)
Single	197 (65.7%)
Employment status
Presently employed	93 (31.0%)
Pensioner	28 (9.3%)
Unemployed	179 (59.7%)
Household class
Normal	190 (63.3%)
Overcrowded	110 (36.7%)
Outdoor environment
Clear	106 (35.3%)
Dusty	194 (64.7%)

Participants’ behavioural habits

Table 2 shows the smoking habits and alcohol consumption among the participants. The result indicates that the majority of the patients (75%) were non-smokers while most of the participants (89.7%) were in the category of non-drinkers/light drinkers.

Table 2: Participants’ behavioural habits (n = 300)

Variables	Total n (%)
Smoking
No	225 (75%)
Yes	75 (25%)
Drinking habit
Non-drinker/light drinker	269 (89.7%)
Moderate drinker/heavy drinker	31 (10.3%)

Medical history of the participants

The data shows that approximately half of the patients (52.7%) reported not to have previous medical history of PTB or treatment and more than half (65.0%) of the participants reported not to have a family history of PTB (Table 3).

Table 3: Medical history of the participants (n = 300)

Variables	Total n (%)
Previous medical history/treatment of pulmonary tuberculosis
Yes	142 (47.3%)
No	158 (52.7%)
Family history of pulmonary tuberculosis
Yes	105 (35.0%)
No	195 (65.0%)

PTB prevalence among the participants

Figure 1 shows the PTB prevalence among participants in this study. More than half (67.3%) of the patients were PTB negative and whilst 32.7% individuals were PTB+

Figure 1: Pulmonary tuberculosis prevalence among the participant (n = 300)

Comparison of risk factors among HIV patients who are PTB positive and negative

Risk factors for pulmonary tuberculosis

There was a statistically significant association between PTB status and living in a dusty outdoor environment (p-value = 0.006) (Table 4) and being a moderate drinker/heavy drinker (p-value = 0.001) as shown in Table 5.

Table 4: Socio-demographic characteristics among HIV patients who are PTB (−) and PTB (+) (n = 300)

Variable & category	PTB (−) (n = 202)	PTB (+) (n = 98)	Total (n = 300)	χ^2	p-value
	n (%)	n (%)	n (%)
Age group in years
18–30	53 (75.1%)	17 (24.9%)	70
31–40	75 (62.0%)	46 (38.0%)	121	4.85	0.183
41–50	41 (64.0%)	23 (36.0%)	64
> 50	33 (73.3%)	12 (26.7%)	45
Sex
Female	140 (69.6)	61 (30.4%)	201	1.48	0.222
Male	62 (62.6%)	37 (37.4%)	99
Level of education
High school/university degree	43 (63.2%)	25 (36.8%)	68	0.77	0.692
Primary/secondary	142 (68.3%)	66 (31.7%)	208
Illiterate/no formal education	17 (70.8%)	7 (29.2%)	24
Marital status
Married/cohabitation	56 (77.8%)	16 (22.2%)	72
Widowed/divorced	22 (71.0%)	9 (29.0%)	31	5.48	0.064
Single	124 (62.9%)	73 (37.1%)	197
Employment status
Presently employed	64 (68.8%)	29 (31.2%)	93	0.46	0.791
Pensioner	20 (71.4%)	8 (28.6%)	28
Unemployed	118 (65.9%)	61 (34.1%)	179
Household class
Normal	131 (69.0%)	59 (31.0%)	190	0.61	0.433
Overcrowded	71 (64.6%)	39 (35.4%)	110
Outdoor environment
Clear	82 (77.4%)	24 (22.6%)	106
Dusty	120 (61.9%)	74 (38.1%)	194	7.48	0.006^*

^* p < 0.05.

Table 5: Behavioural factors of HIV patients who are PTB (−) and PTB (+) (n = 300)

Variables	PTB (−) (n = 202)	PTB (+) (n = 98)	Total (n = 300)	χ^2	p-value
	n (%)	n (%)
Smoking
No	54 (68.4%)	71 (31.6%)	225
Yes	48 (64.0%)	27 (36.0%)	75	0.50	0.477
Drinking category
Non-drinker/light drinker	189 (70.3%)	80 (29.7%)	269
Moderate drinker/heavy drinker	13 (41.9%)	18 (58.1%)	31	10.1	0.001^*

^* p < 0.05.

Logistic regression analysis

Univariate logistic regression model for risk of pulmonary tuberculosis

The univariate analysis was conducted to identify significant risk factors associated with PTB in order to inform the multivariate model (Table 6). In summary, compared with individuals who self-reported as being non-drinkers/light drinkers, moderate/heavy drinkers had increased odds of pulmonary tuberculosis (OR = 3.27; 95% CI = 1.51–7.09). When compared with being married/cohabiting and widowed/divorced, single marital status increased the odds of PTB (OR = 2.06; 95% CI = 1.09–3.88). The odds of PTB were also higher for individuals living in a dusty outdoor environment (OR = 2.11; 95% CI = 1.22–3.64) (Table 6).

Table 6: Logistic regression analysis for risk factors of pulmonary tuberculosis

Variable & category	Univariate	P-value	Multivariate	p-value
	OR (95% CI)		OR (95% CI)
Age group in years
18–30	1	0.052
31–40	1.91 (0.98;3.72)	0.142
41–50	1.75 (0.82;373)	0.775
> 50	0.13 (0.48;2.68)
Marital status
Married/cohabitation	1	-	-	-
Widowed/divorced	1.43 (0.55;3.74)	0.462	1.48 (0.54;403)	0.445
Single	2.06 (1.09;3.88)	0.022^*	1.96 (1.01;3.79)	0.047^*
Drinking category
Non-drinker/light drinker	1	-	-	-
Moderate drinker/heavy drinker	3.27 (1.51;7.09)	0.001^*	3.46 (1.56;7.69)	0.002^*
Outdoor environment
Clear	1	-	-	-
Dusty	2.11 (1.22;3.64)	0.006^*	2.05 (1.16;3.61)	0.013^*

CI, confidence interval: OR, odds ratio.

^* significant at p < 0.05.

Multivariate logistic regression model for risk of pulmonary tuberculosis

In multivariate logistic regression models, an analysis of the variables were assessed. Only factors found to be significantly associated with PTB (p < 0.05) were retained in the final multivariate model. Single marital status was significantly associated with PTB [p-value = 0.047 (OR = 1.96; 95% CI = 1.01–3.79)]. Being a moderate or heavy drinker remained strongly associated with pulmonary tuberculosis [p-value = 0.002 (OR = 3.46; 95% CI = 1.56–7.69)]. Living in a dusty outdoor environment was also significantly associated with pulmonary tuberculosis [p-value = 0.013 (OR = 2.05; 95% CI = 1.16–3.61)] (Table 6).

Discussion

The most interesting findings to emerge from this study were that single marital status, a dusty outdoor environment, and being a moderate to heavy drinker were identified as risk factors significantly associated with PTB among adult HIV positive patients at this hospital.

However, contrary to expectations, age, gender, educational attainment, employment status, overcrowding, previous medical history of PTB or treatment, and a family history of PTB were shown not to be associated with pulmonary tuberculosis.

Marital status

According to Lienhardt et al. marital status has long been associated with the risk of developing TB in Africa. The authors suggest that the most plausible mechanism of this association is that family structure seems to play a role, since it determines the risk of exposure and reflects the socio-economic status of individuals.³ On the other hand, it has been indicated that marital status affects the risk of TB, with single men having a greater risk of TB than married men.^3,4

Similar findings to this investigation were reported in a study conducted in Guinea-Bissau which illustrated that being single and living alone was an independent risk factor for contracting TB.⁴ A multi-centre study conducted in three West African countries showed that single marital status was a risk factor for TB.³ Similarly, Wang et al. in a case control study conducted in Henan, China, reported that being single was a risk factor of PTB.⁵ In contrast a study conducted in South Africa found that cohabiting was statistically associated with PTB,⁶ and a study carried out in Estonia reported that any marital status other than being married was the main risk factor for contracting tuberculosis.⁷

Cain et al. suggested that being single implies a lack of family support,⁸ according to these researchers, vulnerability to TB could increase at times due to psychosocial stress. However, according to Pang et al. marital status is also closely associated with other socio-economic characteristics which are extremely difficult to distinguish from the many possible confounding variables in a simple study.⁹

Therefore, based on the results of this study, further studies exploring the direct effect of family structure at the individual level are needed to accurately determine the mechanism underlying these associations.

Outdoor environment/dusty environment

This research finding reveals that living in dusty outdoor conditions increases the risk of pulmonary tuberculosis. After an extensive review of the literature on dusty outdoor environments as a contributory risk factor for pulmonary tuberculosis, there is surprisingly scant information available on this topic in South Africa. The majority of published studies in the country were mainly related to occupational health environments and the mining industry.¹⁰ Therefore, this is a clear indication that further research is required in this particular field.

A study in South Africa indicated that dust control management is an important component when attempting to decrease the impact of the PTB epidemic in South African gold mines.¹⁰ Another study reported that older in‐service gold miners in South Africa have a high prevalence of PTB, which is significantly associated with dust in the mines and silica exposure, even in the absence of silicosis.¹¹ A population-based case-control study conducted in China revealed that a poor environment and exposure to dust were risk factors for adult PTB in that region.¹²

It is also well recognised that better atmospheric conditions lead to a better quality of life, which could enhance the immunological function of patients who suffer from PTB. It follows that better environments and/or working conditions may reduce the risk of individuals developing pulmonary tuberculosis.

Drinking habits: moderate and heavy drinkers

As revealed by the analysis and the statistically significant results, adult HIV patients who are moderate/heavy drinkers were found to be more at risk for contracting pulmonary tuberculosis. This finding is consistent with other studies, which have reported alcohol use as a recognised risk factor for tuberculosis. In particular, a study conducted by Lönnroth et al. attributed the risk of active tuberculosis to be substantially elevated in people who drink more than 40 grams of alcohol per day, and/or have an alcohol use disorder. The researchers indicated that this may be due to both increased risk of infection related to specific social interaction patterns associated with alcohol use as well as the impact of alcohol on the immune system and other related health conditions.¹³

It has been suggested in the same report that these specific social interaction patterns may increase the risk of exposure to people with infectious TB disease in settings such as bars, shelters for the homeless, prisons and social institutions.¹³ A study conducted in a high incidence area of Western Cape Province in South Africa between1993 and 1996, found that most PTB transmission took place outside of households, and that 58% of the contracted PTB infections occurred outside the home whilst drinking in social groups.¹⁴

The present study’s finding is consistent with those which have reported alcohol use as a strong risk factor for TB in South Africa. An early study carried out on the Mamre population showed that alcohol use was strongly associated with TB.¹⁵ More recently, Harling et al. reported alcohol use as an independent risk factor for TB in South Africa.¹⁶

Furthermore, studies from different parts of the world have also reported that alcohol use is a risk factor for tuberculosis. A case-control study carried out in Estonia found that alcohol consumption was a risk factor for tuberculosis.⁷ A systematic review conducted by Rehm et al. reported that heavy alcohol consumption and alcohol use disorders (AUD) constitute a risk factor for the incidence of and re-infection with TB.¹⁷

Contrary to the finding of the present study, there are other studies which have indicated no association between alcohol and tuberculosis. A study in South Africa reported that alcohol use was not in fact associated with TB.⁶ In West Africa, Lienhardt et al. reported that alcohol intake was not associated with TB after adjusting for other factors.³ A study conducted in India showed no effect of alcohol consumption on TB infection; these authors suggested that the weak association between alcohol and tuberculosis could be attributed to the fact that alcohol use is associated with smoking.¹⁸

Study limitations

In view of the findings, and the method and design employed in this research, it is important to acknowledge some limitations in this kind of study. Such limitations could have influenced the data collected and the conclusions reached.

One limitation of the study could be related to the accuracy of participant feedback during interviews, as this interview data relied on self-reported information by participants and could have been influenced by different factors including inability to accurately recall experiences and possible embarrassment caused by some the questions.

However, the researchers endeavoured to mitigate these challenges as efforts were made to ensure confidentiality of the participants and detailed structured questionnaires were administered.

Conclusions

The results of this research showed that being of single marital status, living in a dusty outdoor environment and being a moderate/heavy drinker were all strongly associated with an increased risk of PTB for the study participants. These findings have obvious important implications for TB control strategies both in the region and nationally, particularly in similar environments as those of the study population and site.

Recommendations

The researchers recommend that there is a need to develop specific programmes for screening and detecting alcohol use disorders (AUD) in HIV patients to reduce PTB at the Wellness Clinic in Themba Hospital. This could also be applied to other HIV management health institutions across the country. There is a need to prevent or reduce dust emissions mainly from the unpaved roads of rural areas, which could improve the outdoor air quality and subsequently, the population’s health. Health policy-makers need to include these risk factors in the screening processes for PTB among HIV patients as part of the management guideline strategies.

Conflict of interest

The authors declare that they have no financial or personal relationships which may have inappropriately influenced them in writing this paper.

Acknowledgements

We gratefully acknowledge the advice and statistical assistance of Professor Piet J. Becker from the Biostatistics Unit of the South African Medical Research Council (MRC) in Pretoria.