Background
Intermittent therapy has been widely used in the continuation phase of first-line tuberculosis (TB) treatment, reducing medication and healthcare worker costs for TB programs, and facilitating global scale-up of directly observed therapy. [
1‐
4] However, given the emergence of TB drug resistance and concerns regarding TB relapse, the World Health Organization (WHO) has recommended daily therapy as the preferred intermittency for TB treatment since 2008 [
5,
6]. In addition, the American Thoracic Society (ATS), Centers for Disease Control and Prevention (CDC), and Infectious Diseases Society of America (IDSA) recommended daily TB treatment for HIV-positive persons in 2016 [
7]. However, guidelines vary across regions and countries, and treatment may be given 7, 5, 3, or 2 days/week in the continuation phase. For example, in Mexico and Argentina, TB treatment guidelines for HIV-positive persons recommend daily therapy during the initial two-month intensive phase and thrice-weekly therapy in the continuation phase [
8‐
10]. In Brazil and Haiti, national recommendations include daily therapy in the continuation phase [
11‐
14]. In Peru, previous national guidelines recommended intermittent treatment in the continuation phase, but more recently, daily therapy has been recommended [
14].
Few studies have described the effect of TB treatment intermittency in the continuation phase on mortality in HIV-positive persons, particularly in persons receiving concomitant antiretroviral therapy (ART). In TB patients without HIV, systematic reviews have evaluated clinical outcomes according to different dosing frequencies. [
15‐
17] Johnston and collaborators published a meta-analysis demonstrating that thrice-weekly dosing throughout TB treatment was associated with increased mortality compared to daily therapy. [
16] Less frequent dosing schedules were significantly associated with an increased risk of TB relapse. In the meta-regression, they were also significantly associated with higher relapse rates, failure, and acquired drug resistance [
16].
Several randomized, controlled trials (RCTs) have been published addressing first-line standard TB treatment, with few focusing on the intermittency of continuation phase treatment in HIV-negative patients [
18‐
21]. More recently, a randomized clinical trial performed at several sites in South India concluded that in HIV-positive patients, daily anti-TB therapy throughout treatment for pulmonary TB was superior to thrice-weekly treatment throughout therapy, and thrice-weekly treatment in the continuation phase (with daily therapy in the intensive phase) [
22]. However, the sample size was small, and hepatoxicity risk was higher with daily therapy. Additionally, a prospective observational study concluded that thrice-weekly anti-TB therapy was effective in HIV- negative but not HIV-positive patients [
23]. Patients who received directly observed intermittent anti-TB therapy during the intensive phase had a 40% higher risk of mortality than patients who received an unsupervised daily regimen [
24]. Our group published an observational study assessing factors associated with mortality in TB/HIV co-infected patients. Duration of more than 6 months of TB treatment was associated with better survival than 6 months of treatment [
25]. However, the description of frequency of intermittency in the continuation phase of TB treatment in HIV-positive persons in Latin America and its impact in survival has not been evaluated; data from large cohorts could provide important information for all settings, to confirm the data from the one clinical trial above. Given current recommendations that TB and HIV should be treated concomitantly, we focused on persons who received TB treatment and ART.
Discussion
In the present study, we assessed for differences in mortality in HIV-positive patients with TB depending on the intermittency of TB treatment in the continuation phase. There was a decreased mortality risk in patients who received treatment 5–7 days/week compared to 2–3 days/week, as evidenced by the crude risk of death in the Kaplan–Meier curves, and the multivariable Cox model that adjusted for age, sex, CD4 count, ART use at TB diagnosis, site of TB disease, and year of TB diagnosis. However, after stratifying by study site, there was no longer a protective effect. This suggests that there may have been other differences among sites that accounted for the difference in survival; this could include factors such as differences in ART, host factors, or access to specialized care. As expected, increased age and decreased CD4 count at diagnosis were statistically significant risk factors for mortality across all analyses. The risk of death increased even after TB treatment completion, which we have noted previously [
26].
There was substantial variation across our sites in access to ART, treatment practices, and resources for treating HIV and TB co-infection. For example, sites had varying access to ART drug classes such as integrase strand transfer inhibitors, including raltegravir and more recently, dolutegravir. In addition, differences in HIV drug resistance patterns and resultant ART regimens could have also affected adherence and clinical outcomes [
27‐
29]. Furthermore, local guidelines for treating TB and provider adherence to standards of care could vary within the region [
9‐
13]. For instance, in Peru guidelines regarding treatment intermittency in the continuation phase changed over the study period—from 3 times per week to daily dosing [
14]. In addition, since patients were not randomized to the different dosing intervals in our study, there may have been unmeasured confounding variables associated with mortality—as can occur in all observational studies [
30]. Finally, there were differences in how many patients each center contributed to the study: Haiti (68%), Peru (18%), Brazil (11%), Chile (1.5%), Mexico (1.2%), and Honduras (1.0%) as well as differences in treatment frequency by site with the two largest contributors (Haiti and Brazil) administering treatment 5–7 days/week to all patients, consistent with their local guidelines [
11‐
13].
Other studies have demonstrated that TB treatment given 2–3 days/week in the continuation phase is associated with an increased risk of relapse, treatment failure, and acquired drug resistance compared to treatment 5–7 days/week, most of which were conducted in HIV-negative persons [
2,
15‐
17]. To date, only one randomized controlled clinical trial of TB treatment intermittency among HIV-positive persons has been conducted; it was performed in India and demonstrated superior efficacy and decreased emergence of drug resistance with daily dosing of TB medications throughout treatment [
22]. This is also consistent with evidence from observational studies of HIV-negative persons. [
15‐
17]. A retrospective cohort study in India among 292 HIV-TB coinfected patients on atazanavir/ritonavir compared outcomes among persons receiving daily rifabutin or thrice-weekly rifabutin during the TB treatment continuation phase. More individuals in the daily group achieved clinical cure (73.0% vs. 44.1%, P < 0.001), with no significant differences in relapse/recurrence or all-cause mortality between groups. [
31] The lack of a difference in all-cause mortality was similar to our findings in the multivariable analysis stratified by study site.
There were several limitations of our study. First, it was an observational cohort study. Although we performed multivariable analyses to adjust for potentially confounding variables, there may have been residual or unmeasured confounding that could affect study findings. There are other ways that we could have adjusted for confounding variables (e.g., inverse probability weighting), but these other approaches are also potentially sensitive to unmeasured confounding. We did not have information regarding TB treatment failure, TB relapse, nor cause of death. In addition, Latin America and the Caribbean include countries that span from low to high-income levels [
32,
33], representing a heterogeneous group with varying access to TB diagnostic tests, TB treatment guidelines (even within countries during the period of study, e.g., Peru), and availability of treatment and care resources. However, this reflects the reality of clinical settings in Latin America and throughout the world. Our results are consistent with recent studies [
16,
22,
31‐
33], but additional data are needed to inform the care of persons with HIV-related TB optimally. Prospective, randomized controlled trials of different TB treatment dosing intervals that include different regions in the globe affected by tuberculosis, would help identify the most effective and cost-effective regimens for TB treatment in persons with HIV who concomitantly receive ART.
Acknowledgements
We thank the staff at each of the CCASANET centers and especially our patients. The Caribbean, Central and South American network for HIV epidemiology (CCASAnet) includes research teams in the following sites: Fundación Huésped, Argentina: Pedro Cahn, Carina Cesar, Valeria Fink, Zulma Ortiz, Florencia Cahn, Agustina Roldan, Ines Aristegui, Claudia Frola. Instituto Nacional de Infectologia-Fiocruz, Brazil: Beatriz Grinsztejn, Valdilea G. Veloso, Paula M. Luz, Sandra Cardoso Wagner, Ruth Friedman, Ronaldo I. Moreira, Lara Esteves Coelho, Monica Derrico Pedrosa, Guilherme Amaral Calvet, Hugo Perazzo, Rodrigo Moreira, Maria Pia Diniz Ribeiro, Mario Sergio Pereira, Emilia Moreira Jalil. Universidade Federal de Minas Gerais, Brazil: Jorge Pinto, Flavia Ferreira, Marcelle Maia. Universidade Federal de São Paulo, Brazil: Regina Célia de Menezes Succi, Daisy Maria Machado, Aida de Fátima Barbosa Gouvêa, Fabiana do Carmo. Fundación Arriarán, Chile: Claudia Cortes, Marcelo Wolff, Maria Fernanda Rodriguez, Gabriel Castillo, Gladys Allendes. Les Centres GHESKIO, Haiti: Jean William Pape, Vanessa Rouzier, Adias Marcelin, Youry Macius, Stephano Saint Preux. Hospital Escuela Universitario, Honduras: Marco Tulio Luque, Diana Varela, Magda Chavez, Ada Mailhot. Instituto Hondureño de Seguridad Social, Honduras: Denis Padgett. Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, Mexico: Juan Sierra Madero, Brenda Crabtree Ramirez, Yanink Caro Vega. Instituto de Medicina Tropical Alexander von Humboldt, Peru: Eduardo Gotuzzo, Fernando Mejia, Gabriela Carriquiry. Vanderbilt University Medical Center, USA: Catherine C McGowan, Stephany N Duda, Bryan E Shepherd, Timothy Sterling, Anna K Person, Peter F Rebeiro, Jessica Castilho, William C Wester, Kate Clouse, Karu Jayathilake, Fernanda Maruri, Hilary Vansell, Marina Cruvinel Figueiredo, Cathy Jenkins, Ahra Kim, Sarah Lotspeich, Paridhi Ranadive. Vanderbilt University, USA: Kate Clouse.
“CCASAnet Region 2 of IeDEA” Consortium members: Argentina -Fundación Huésped: Pedro Cahn, Carina Cesar, Valeria Fink, Zulma Ortiz, Florencia Cahn, Agustina Roldan, Ines Aristegui, Claudia Frola.
Brazil—-Instituto Nacional de Infectologia-Fiocruz: Beatriz Grinsztejn, Valdilea G. Veloso (Author), Paula M. Luz, Sandra Cardoso Wagner, Ruth Friedman, Ronaldo I. Moreira, Lara Esteves Coelho, Monica Derrico Pedrosa, Guilherme Amaral Calvet, Hugo Perazzo, Rodrigo Moreira, Maria Pia Diniz Ribeiro, Mario Sergio Pereira, Emilia Moreira Jalil.
-Universidade Federal de Minas Gerais: Jorge Pinto, Flavia Ferreira, Marcelle Maia
-Universidade Federal de São Paulo: Regina Célia de Menezes Succi, Daisy Maria Machado, Aida de Fátima Barbosa Gouvêa, Fabiana do Carmo.
Chile—-Fundación Arriarán: Claudia Cortes (Author), Marcelo Wolff, Maria Fernanda Rodriguez, Gabriel Castillo, Gladys Allendes.
Haiti—-Les Centres GHESKIO: Jean William Pape (Author), Vanessa Rouzier, Adias Marcelin, Youry Macius, Stephano Saint Preux, Serena Koenig (Author)
Honduras—-Hospital Escuela Universitario: Marco Tulio Luque, Diana Varela, Magda Chavez, Ada Mailhot.
-Instituto Hondureño de Seguridad Social: Denis Padgett (Author)
Mexico—-Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán: Juan Sierra Madero (Author), Brenda Crabtree Ramirez (Author), Yanink Caro Vega
Peru—-Instituto de Medicina Tropical Alexander von Humboldt: Eduardo Gotuzzo (Author), Fernando Mejia, Gabriela Carriquiry (Author)
USA—-Vanderbilt University Medical Center: Catherine C McGowan (Author), Stephany N Duda, Bryan E Shepherd (Author), Timothy Sterling (Author), Anna K Person, Peter F Rebeiro, Jessica Castilho, William C Wester, Kate Clouse, Karu Jayathilake (Author), Fernanda Maruri, Hilary Vansell, Marina Cruvinel Figueiredo, Cathy Jenkins (Author), Ahra Kim, Sarah Lotspeich, Paridhi Ranadive. -Vanderbilt University: Kate Clouse.