By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Objectives This study assesses the outcome of current treatment guidelines and the effect of highly active antiretroviral therapy (HAART) on survival of HIV/TB-coinfected patients in a resource-limited setting.
Methods Observational cohort study at the pediatric HIV Clinic, RML Hospital, Delhi. All HIV-infected patients who visited the clinic for the diagnosis of TB between 2002 and 2006 were observed until 31 March 2010. TB was diagnosed either at the time of enrolment or during follow-up visits. Clinical and epidemiological data were registered. We compared children who were given HAART with TB treatment at time of diagnosis [simultaneous therapy (ST)] and children who received delayed HAART. Survival was assessed by Kaplan–Meier method and Cox regression model.
Results Among the 298 children, 126 (42.2%) had TB, including 96 who received ST (76% of 126) and 30 who did not. There were no differences between the two groups except for a lower CD4 count in patients undergoing ST. ST was associated with improved survival [hazard ratio (HR), 0.35; 95% CI, 0.20–0.74; P =0.002] and so were year of TB diagnosis and other AIDS-defining conditions. Multivariate analysis revealed that ST was a powerful predictor of survival (HR, 0.30; 95% CI, 0.14–0.68; P =0.003). After adjusting for other prognostic variables such as age, gender, CD4 count at time of TB diagnosis, by Cox multivariate analysis, ST remained robustly associated with improved survival (HR, 0.32; 95% CI, 0.17–0.71; P =0.001).
Conclusions Starting HAART during tuberculosis therapy significantly improves survival and provides further impetus for the integration of TB and HIV services.
Objectifs: Cette étude évalue les résultats des directives actuelles de traitement et les effets de la thérapie HAART sur la survie des patients coinfectés par le VIH et la TB dans un cadre à ressources limitées.
Méthodes: Etude de cohorte observationnelle dans la clinique pédiatrique du VIH, à l’Hôpital RML, à Delhi. Tous les patients atteints du VIH qui ont visité la clinique avec un diagnostic de TB entre 2002 et 2006 ont été observés jusqu’au 31 mars 2010. La TB a été diagnostiquée soit au moment de l’inscription ou au cours des visites de suivi. Les données cliniques et épidémiologiques ont été enregistrées. Nous avons comparé les enfants qui ont reçu la thérapie HAART en même temps que le traitement de la TB lorsque le diagnostic a été effectué [thérapie simultanée (TS)] et ceux qui ont reçu le HAART plus tard. La survie a étéévaluée par la méthode de Kaplan-Meier et le modèle de régression de Cox.
Résultats: Parmi 298 enfants, 126 (42,2%) avaient la TB, dont 96 qui ont reçu la TS (76% de 126) et 30 qui ne l’ont pas reçu. Il n’y avait aucune différence entre les deux groupes, sauf pour une faible numération des CD4 chez les patients TS. La TS était associée à une survie améliorée (HR = 0,35; IC95%: 0,20 à 0,74, P = 0,002) de même que l’année de diagnostic de la TB et autres conditions définissant le SIDA. L’analyse multivariée a révélé que la TS était un puissant facteur prédictif de la survie (HR = 0,30; IC95%: 0,14 à 0,68, P = 0,003). Après ajustement pour les variables pronostiques tels que l’âge, le sexe, le taux de CD4 au moment du diagnostic de la TB, par l’analyse multivariée de Cox, la TS est restée robustement associée à une survie améliorée (HR = 0,32; IC95%: 0,17 à 0,71; P = 0,001).
Conclusions: Commencer la thérapie HAART avec le traitement de la TB améliore significativement la survie et donne un nouvel élan pour l’intégration des services TB et VIH.
Objetivos: Este estudio evalúa el resultado de las actuales guías de tratamiento y el efecto del TAR en la supervivencia de pacientes coinfectados con VIH-TB en un emplazamiento con recursos limitados.
Métodos: Estudio observacional de cohortes realizado en la clínica pediátrica para VIH del Hospital RML, Delhi. Todos los pacientes con VIH que visitaron la clínica con un diagnóstico de TB entre 2002 and 2006 fueron observados hasta el 31 Marzo, 2010. La TB se diagnosticó en el momento de inclusión en el estudio o durante el seguimiento. Se registraron datos clínicos y epidemiológicos. Hemos comparado a niños que recibieron TAR junto con el tratamiento para TB en el momento del diagnóstico [terapia simultánea (TS)] con niños que recibieron el TAR con retraso. La supervivencia se evaluó mediante el método de Kaplan-Meier y el modelo de regresión de Cox.
Resultados: De los 298 niños, 126 (42.2%) tenían TB, de los cuales 96 recibieron TS (76% de 126) y 30 no lo recibieron. No había diferencias entre los dos grupos excepto con respecto a un menor conteo de CD4 en pacientes con TS. La TS estaba asociada con una supervivencia mejorada (HR 0.35; IC 95% 0.20-0.74, P = 0.002) al igual que el año de diagnóstico de la TB y de otras condiciones indicativas del SIDA. Un análisis multivariado reveló que la TS era un poderoso vaticinador de supervivencia (HR 0.30; IC 95% 0.14-0.68, P = 0.003). En un análisis multivariado de Cox, después de ajustar para otras variables de pronóstico tales como la edad, el género, el conteo de CD4 en el momento del diagnóstico de TB, la TS continuaba estando fuertemente asociada con una supervivencia mejorada (HR 0.32; IC 95% 0.17-0.71, P = 0.001).
Conclusiones: Comenzar el TAR durante la terapia anti-tuberculosa mejora significativamente la supervivencia y da un mayor ímpetu para la integración de los servicios de TB y VIH.
Tuberculosis (TB) is a major cause of death among children with human immune deficiency virus (HIV) worldwide, especially in low-income countries. Mortality rates are high among HIV/TB-coinfected patients with advanced disease, but the optimal timing for ART initiation in relation to TB therapy has been controversial. The interaction between HIV and TB infections is bidirectional. HIV infection increases the risk of both primary and reactivation TB (Daley et al. 1992), and this risk increases markedly with advancing HIV infection (Wolday et al. 2003). The case fatality rates of HIV-associated TB are high; the estimated aggregate case fatality rate of HIV-infected TB is about 40% and may exceed 50% in many developing countries (Mukadi et al. 2000; Corbett et al. 2003). TB case fatality rates appear to be closely related to the prevalence of HIV infection, and HIV-related conditions may be the main cause for the increased death rate associated with HIV/TB coinfection (Churchyard et al. 2000; Mukadi et al. 2000; UNAIDS/WHO 2003). While deaths in the first month of TB treatment may be due to TB, late deaths in coinfected persons are attributable to HIV infection progression (Churchyard et al. 2000; Mukadi et al. 2000).
The immune pathogenesis of HIV-associated TB (Cowley & Elkins 2003; Toossi 2003) suggests that by inhibiting HIV replication and allowing for CD4+ T-cell-related immune reconstitution, highly active antiretroviral therapy (HAART) will reduce both the incidence of TB and mortality. The use of HAART in TB-endemic areas has been associated with more than 80% reduction in the incidence of HIV-associated TB; the protective effect of HAART was seen at all stages of HIV disease, but was greatest in symptomatic patients and those with advanced disease (Badri et al. 2002; Santoro-Lopes et al. 2002). Several observational studies on adults have found that the use of concurrent HAART in coinfected patients during TB treatment is associated with reduced mortality (García de Olalla et al. 2002; Santoro-Lopes et al. 2002). These studies suggest that HAART has promise in reducing the high morbidity and mortality associated with HIV/TB-coinfection.
In resource-limited settings, the diagnosis of and treatment for TB in HIV-infected children is very complicated and challenging. In India, where the prevalence of TB is very high, the difficulties become more pronounced as in children the signs and symptoms of TB are non-specific and are diagnosed very late in the course of the disease. Moreover, there is no clear guidance on INH prophylaxis in HIV-infected children. We planned this observational study to analyse the impact of simultaneous HAART vs. delayed HAART in HIV-positive children diagnosed with TB on their survival.
An observational cohort study was undertaken at Dr Ram Manohar Lohia Hospital, New Delhi, India, at the HIV paediatric clinic. The study was approved by the Ethics Committee of Dr Ram Manohar Lohia Hospital, New Delhi. All HIV-positive patients who visited the clinic for the diagnosis of TB during 2002–2006 were selected. TB was diagnosed on National IAP guidelines and treated according to Revised National Tuberculosis Control Programme (RNTCP) guidelines (TB India, RNTCP Status Report 2007). The study group suffering from chronic cough (>3 weeks) or with a body weight <80% of the normal weight for age was evaluated for possible TB with the following tests: clinical presentation, Ziehl–Neelsen staining and microscopy, chest X-ray (PA view), Mantoux test (1 TU PPD RT23 with Tween 80, positive if induration >5 mm after 48–72 h). The patients were followed up to 31 March 2010 and underwent clinical and immunological assessments at 6-month intervals.
All diagnoses associated with HIV–TB were recorded on admission and included date of diagnosis of TB and HIV; type of TB; age and sex; number of antiretroviral drugs (if any) used before the diagnosis of TB and at the end of follow-up; last count of CD4+ T lymphocyte before the diagnosis of TB and at the end of follow-up, date of first visit; date of last visit; and survival status. Patients who missed their follow-up appointments for more than 10 months were regarded as lost to follow-up and not included in the analysis. Also, patients were excluded if culture results yielded non-tuberculous mycobacterium or if they initiated HAART more than 2 months before the TB diagnosis. Z-scores were calculated for CD4+ T-lymphocyte counts to correct for age-related differences.
Guidelines of the Indian National Tuberculosis Control Programme recommend the directly observed treatment and short-course for the treatment of paediatric TB, which typically lasts for 6–8 months (RNTCP Status Report 2007). Based on the nature/severity of the disease and the patients’ exposure to previous antitubercular treatments, RNTCP classifies tuberculosis patients into three treatment categories (RNTCP Status Report 2007). The antituberculous therapy for Category 1 comprised of isoniazid, rifampicin, pyrazinamide and ethambutol for 2 months followed by isoniazid and rifampicin for 4 months. For Category 2, the therapy comprised of streptomycin, isoniazid, rifampicin, pyrazinamide and ethambutol for 3 months followed by isoniazid, rifampicin and ethambutol for 5 months. For Category 3, treatment consisted of isoniazid, rifampicin and pyrazinamide for the 2 months followed by isoniazid and rifampicin for 4 months.
Highly active antiretroviral therapy was defined as receipt of one non-nucleoside reverse transcriptase inhibitor and two nucleoside reverse transcriptase inhibitors. Efavirenz replaced nevarapine when rifampicin was given as TB treatment, as per the national guidelines (RNTCP Status Report 2007). In our study group, patients with confirmed HIV/TB coinfection were assigned to one of the two study groups. In the first group, HAART was initiated within 2 months after the start of TB therapy [simultaneous therapy (ST) group]. In the second group, HAART was initiated after 2 months, namely after the completion of the intensive phase of TB therapy [delayed therapy (DT) group]. Patients older than 5 years with CD4+ T-lymphocyte counts <200 cells/mm3, patients younger than 5 years with CD4+ T-lymphocyte counts <500 cells/mm3 (WHO 2003), and patients with advanced AIDS were given ST as early as possible. Individuals older than 5 years with CD4+ T-lymphocyte counts >200 cells/mm3 and under 5 years with CD4+ T-lymphocyte counts >500 cells/mm3 were assigned to each of the therapy groups based on symptoms of further AIDS-defining conditions, CD4+ T-lymphocyte counts and rate of decline (if available), assessments of potential drug toxicities and drug–drug interactions.
The primary end point in this study was all-cause mortality. For individuals reported deceased by family members, the precise date of death along with the date of entry was used to calculate the event time. Those who survived were defined as censored, and their data were censored on 31 March 2010. Comparison of variables between the ST and DT groups assessed at baseline and follow-up were made by chi-squared test for categorical variables and Student’s t-tests for continuous variables. Fisher’s exact test was used in which any of the cell frequency was <5. The rate of change in the count of CD4+ T lymphocyte per year between the baseline and follow-up points was calculated by dividing the differences in the count of CD4+ T lymphocyte between the baseline and follow-up points by the time interval expressed in years (i.e. 4 years in this study; Chattopadhya et al. 2002). The correlation of duration of survival with CD4+ T-lymphocyte z-scores was made by Spearman’s rank correlation coefficient analysis. Survival was assessed by the Kaplan–Meier method (univariate) and Cox proportional hazards regression model. Survival times for the ST and DT groups were compared using log-rank chi-squared test.
While assembling the study cohort, 298 HIV-1-infected children were screened, of whom 126 (43.2%) experienced at least one episode of TB. At the time of TB treatment, ST was received by 96 (76% of 126) patients. We excluded eight patients from the study because they had received HAART for more than 2 months when diagnosed with TB. Their mean age was 7.1 years (SD 2.67), and 75% were men. Seventeen children (13.5%) were reported dead by caregivers; the cause of death was possible other opportunistic infections (OIs) that manifested during follow-up.
The ST and DT groups did not differ significantly in baseline characteristics regarding age, sex, presence of other AIDS-defining illnesses and proportion of extrapulmonary TB, except for a lower CD4+ T-lymphocyte count in the simultaneous therapy group. Of the 126 children with TB, 101 (80%) children had pulmonary TB, 10% miliary TB, 4.2% disseminated TB, 3.4% central nervous system TB and 2.4% abdominal TB. Other AIDS-defining illnesses were present in 17% of the cases. CD4+ T-lymphocyte count at the time of diagnosis of TB was (276 SD 242.13), median 200, IQR (127–310).
The CD4+ T-lymphocyte z-scores in the ST and DT groups showed no difference in terms of mean values at baseline (0.87 ± 1.11 and 0.91 ± 0.99 respectively, P = NS). At the follow-up assessment, after a 4-year interval, the DT group showed minimal significant change in CD4+ T-lymphocyte z-score over baseline and the ST group showed a significant increase in values over baseline (1.86 ± 1.02 and 1.36 ± 1.01 respectively, P =0.001; Table 1, Figure 1). The rates of increase in the CD4+ T-lymphocyte z-score per year, as evident from the range and median values of the β-slopes, were marked in the ST group compared with the DT group. Moreover, in our study group, none of the patients in the ST group showed any sign of relapse.
Table 1. Characteristics of patients with tuberculosis undergoing simultaneous therapy or delayed therapy
TB-HAART treatment, n = 126
Simultaneous therapy (n = 96)
Delayed therapy (n = 30)
CD4+ T-lymphocyte count: cells per cubic millilitre.
HAART, highly active antiretroviral therapy.
7.5 ± 2.76
6.7 ± 2.17
Sex (% Male)
CD4+ T-lymphocyte count at diagnosis of TB, median (IQR)
Extra pulmonary TB %
CD4+ T-lymphocyte count at the end of follow-up, median (IQR)
Other AIDS-defining conditions %
The correlation of CD4+ T-lymphocyte z-score, assessed at various points, with the overall survival period showed that despite higher values than the DT group at follow-up, there was no correlation of the CD4+ T-lymphocyte z-score with overall survival in the ST group (r = 0.045; P =0.775, Figure 2a). However, there was a significant positive correlation between the rate of increase, i.e. β-slope of the CD4+ T-lymphocyte z-score and overall survival period, in the ST group (r = 0.629; P <0.0001, Figure 2b).
According to univariate Cox analysis, year of diagnosis of TB, presence of other AIDS-defining conditions and ST were the only significant variables associated with poor survival (Table 2). Patients undergoing ST were 65% more likely to survive than those undergoing DT [hazard ratio (HR), 0.35; 95% confidence interval (CI), 0.20–0.74; P =0.002]. In a multivariate analysis including all nearly significant variables (P <0.05), ST remained as a powerful predictor of survival (HR, 0.30; 95% CI, 0.14–0.68; P =0.003).
Table 2. Cox analysis (univariate and multivariate) of survival
Univariate Cox analysis
Multivariate Cox model
Multivariate Cox model adjusted*
*Adjusted for age, sex, CD4+ T-lymphocyte z-score at the time of the diagnosis of TB.
HR, hazard ratio.
Year of TB diagnosis
Other AIDS-defining illnesses
A Cox proportional hazards regression model adjusted for age, gender and CD4+ T-lymphocyte z-score at TB diagnosis determined that ST and later year of TB diagnosis independently lowered the risk of death, while other AIDS-defining illnesses tended to raise the risk of death.
At 3 years of follow-up, the hazard ratio for survival with ST was 0.24 (95% CI, 0.062–0.926; P =0.01), and at 5 years, the HR was 0.33 (95% CI, 0.11–0.78). A similar survival advantage with ST was maintained through the end of follow-up. Kaplan–Meier analysis through 7 years of follow-up confirmed shorter survival in children who started HAART after completing anti-TB therapy (P =0.033, Figure 3).
This study showed that initiation of HAART simultaneously along with TB therapy in patients with HIV/TB coinfection improved the overall survival of the patients. A lack of correlation between the baseline, as well as follow-up CD4+ T-lymphocyte z-scores, and subsequent duration of survival in the ST group reveals a limitation of the predictive value of such single-point assessment and tends to support the view of other workers who have documented similar limitations (Peterson et al. 1989; Chattopadhya et al. 2002). A significant positive correlation was recorded between the rate of increase (β-slope) and subsequent duration of survival (r = 0.629). This shows that more valuable information may be obtained by recording changes in CD4+ T-lymphocyte z-scores rather than a one-point assessment. In this study, the rate of increase in CD4+ T-lymphocyte z-scores among the ST group was faster compared than that in the DT group, thereby providing compelling evidence of the benefit of initiating antiretroviral therapy during TB therapy in patients with HIV coinfection. The findings also support recommendations by the WHO and others for the integration of TB and HIV care.
Apart from the timing of initiation of HAART along with anti-TB treatment in HIV/TB-coinfected patients, survival is affected by several predictors such as type of AIDS-defining illness (Hoover et al. 1995), increased CD4+ T-lymphocyte count and time of diagnosis of TB. In HIV-infected children, the measure for disease progression is mainly given by CD4+ T-lymphocyte count in the peripheral blood and other AIDS-defining illnesses. The patients in our study group were all coinfected with TB; therefore, the effect of an AIDS-defining illness on the survival prognosis was attenuated. In our cohort, CD4+ T-lymphocyte count at TB diagnosis was not associated with survival, in agreement with recent reports (Couzigou et al. 2007). Year of diagnosis of TB was associated with better survival. This is probably so because patients who were earlier diagnosed with TB were not subjected to HAART during the intensive phase of TB therapy, and with the advent of the RNTCP guidelines, HAART was administered along with anti-TB therapy.
Although our results show that other AIDS-defining illness and year of diagnosis of TB were good measures of disease progression, focusing the benefits of a full treatment requires simultaneous analysis of the disease progression and survival. This will require us to study the relationship between the multivariate longitudinal process and time to event. However, the modelling considerations may become more complex, involving the joint distribution of several longitudinal processes (Henderson et al. 2000; Tsiatis & Davidian 2004).
Mortality among patients with HIV and TB coinfection is high, despite the use of effective TB therapy (Mukadi et al. 2000). Most of the patients in our research group died of non-TB conditions. Previous studies have demonstrated that death within the first few months of TB treatment may be related to TB, whereas late deaths are attributable to HIV infection progression (Churchyard et al. 2000; Mukadi et al. 2000). Differences in baseline characteristics of patients in the two groups of simultaneous and delayed therapy may result in differences in mortality in the patients. However, we did not find any differences except for a slightly lower initial CD4+ T-lymphocyte count in the ST group. This is in accordance with a previous study that has reported significantly lower survival for HIV/TB-coinfected children with a CD4 percentage of <10% (Nahid et al. 2007). In our research, ST continued to be robustly associated with survival even after including predictor variables such as CD4+ T-lymphocyte count and other AIDS-defined conditions in the multivariate analysis.
Previous studies have shown a better survival effect in patients treated with HAART during TB treatment (Dheda et al. 2004; Manosuthi et al. 2006). In our study, we included patients who started HAART within 2 months of anti-TB treatment. Reports of clinical trials evaluating the effect of simultaneous therapy on survival of children are scarce. However, a few clinical trials on adults have shown that the initiation of antiretroviral therapy during tuberculosis therapy in patients with confirmed TB and HIV coinfection reduces mortality (Karim et al. 2010; Havlir et al. 2011). Nahid et al. (2007) noted beneficial effects of HAART during TB treatment as it was associated with more rapid conversion of smears and cultures and improved survival. None of the patients in our study who received simultaneous HAART relapsed.
Multiple complications in the comanagement of HIV and TB have been documented, including overlapping medication toxicity profiles; pharmacokinetic drug interactions; premature death from non-tuberculosis causes; development of resistant TB; and immune reconstitution reactions (Blanc et al. 2007; McIlleron et al. 2007). In our study group, we did not see any development of toxicities, immune reconstitution inflammatory syndrome (IRIS) or drug interactions in the patients undergoing simultaneous therapy. IRIS is very rare in children in most of the Indian centres, for reasons which are unclear. So in areas of high TB prevalence, simultaneous HAART and TB treatment should be initiated with the diagnosis of tuberculosis in HIV-infected patients. Delaying HAART allows the virus to proliferate and thereby increases mortality and morbidity.
There were limitations in the present study. In children under 5 years, the level of immune suppression is determined by the CD4%. However, we used absolute numbers as suggested by WHO guidelines for resource-limited countries, because we did not have facilities to record CD4%.
In conclusion, this study has demonstrated substantially increased survival in patients coinfected with HIV and TB who received simultaneous HAART and TB therapy. HAART should be initiated among advanced HIV-infected patients with TB simultaneously with TB therapy. Further prospective interventional studies on appropriate timing for initiating ART while minimizing drug toxicity, morbidity and mortality are needed in HIV- and TB-coinfected patients.