Latent and Subclinical Tuberculosis in HIV Infected Patients

A Cross-sectional Study

Meaghan M Kall; Katherine M Coyne; Nigel J Garrett; Aileen E Boyd; Anthony T Ashcroft; Iain Reeves; Jane Anderson; Graham H Bothamley


BMC Infect Dis. 2012;12(7) 

In This Article


Study Population, Design, and Entry Criteria

All patients (~600) attending the HIV outpatient clinic at Homerton University Hospital from April 2006 until October 2009 were eligible for screening, except those who were currently being investigated or treated for active TB and patients who ever received treatment for suspected or confirmed TB (n = 93). Recruitment was carried out in two phases. In phase one (April 2006 to April 2008) the immunospot assay was offered one day per week as a pilot study funded by the Department of Health, the results of which have been reported in an abstract.[12] In phase two, the Primary Care Trust agreed to expand screening to the remaining cohort as part of standard clinical care. Informed consent was obtained from patients who took part in the initial pilot study. Patients were asked whether they had cough, fever, night sweats or weight loss, as recommended by the WHO screening program.[13] The number of HIV/TB co-infected individuals treated at this hospital provided the denominator to measure the effect of screening.

Immunospot Assay

Venous blood samples were drawn and transported to the hospital microbiology laboratory on the day of collection. The T-SPOT®.TB assay (Oxford Immunotec, Oxford, UK) is a commercially available enzyme-linked immunospot assay which uses a fixed number of peripheral blood mononuclear cells. Blood samples were processed using standard operating procedure OXIM.SOP06–001. Positive and negative controls were included to validate the result. A result was considered positive if the number of spot-forming cells obtained from test antigens was more than twice the number of the negative control and had six or more spots than the negative control. Wherever possible, patients with indeterminate results had a second sample tested.

Clinical Management of Patients With Positive Result

Patients with positive results were referred to a TB specialist physician. All patients had a further symptom assessment, chest radiograph, and sputum (three samples if productive or at least one induced sputum with hypertonic 3% saline) sent for smear microscopy and mycobacterial culture. Further investigations were carried out in line with clinical findings. Referral to the TB specialist was delayed for three months for individuals commencing highly active antiretroviral therapy (HAART) to observe symptoms of TB that might emerge as a result of immune reconstitution.

The decision to initiate preventive treatment was made by the TB physician after discussion with the patient regarding the risks and benefits. Patients on HAART were offered six months of isoniazid as recommended in national guidelines.[14] Those not on HAART had the option of three months rifampicin and isoniazid.[15] Without HAART, rifampicin posed no risk of drug interactions and patients could be involved in the choice of treatment, a process which enhances adherence.[16] Where active tuberculosis was suspected, two months of rifampicin, isoniazid and pyrazinamide were given, with or without ethambutol depending on clinical suspicion and likelihood of drug-resistance.[17,18] If active infection was proven or likely from the response to treatment (i.e. weight gain, resolution of radiographic changes), the standard six month anti-tuberculosis regimen was continued or modified according to drug sensitivities. If active TB was ultimately excluded, the regimen of two months rifampicin and pyrazinamide was considered sufficient preventive treatment.[15] In severely immunosuppressed patients, TB prophylaxis was continued until CD4 > 200 cells/μl.[19] Those declining to attend the TB clinic were offered follow-up and preventive treatment in the HIV clinic.

Data Handling and Statistical Analysis

Patient details including age, sex, ethnicity, country of origin, CD4 count at test and receipt of antiretroviral therapy were collected. The Shapiro-Wilk test was used to test for normality of continuous variables. Differences in proportions of categorical variables were tested using χ2 test, and differences in median values of continuous variables were tested using the Kruskal-Wallis test. Univariate analysis was performed using logistic regression. Odds ratio and 95% confidence interval (OR, 95% CI) were used to measure the association between different variables and immunospot result. All statistical calculations were performed using STATA (Stata Statistical Software: Release 11. College Station, TX: StataCorp LP. 2009).


Ethical approval for the pilot study was granted by Multicentre Research Ethics Committee and the East London and City Health Authority (P/03/285: Blood tests for tuberculosis).