Treatment Outcomes and Predictive Factors for Multidrug-Resistant TB and HIV Coinfection in Rio De Janeiro State, Brazil

Background: This study aims to identify and compare factors associated with the unfavorable treatment of MDR-TB patients with HIV coinfection in the of (RJ), Brazil. Methods: A retrospective cohort study with data from TB-MDR patients notied between 2000 and 2016 in RJ. Clinical and epidemiological characteristics and treatment outcomes of patients with HIV positive, negative, and unknown were compared. Bivariate and multivariate Cox hazard regression analysis was conducted to identify factors associated with unfavorable treatment results in patients MDR-TB / HIV positive. Results: Of 2,269 cases, 1,999 (88.1%) had a negative result for HIV, 156 (6.9%) positive and 114 (5.0%) unknown. The proportion of unfavourable outcomes was 43.7% among HIV negative, 52.6% among HIV positive and 43.9% among HIV unknown. Regarding MDR-TB / HIV positive cases, the unfavorable outcomes more prevalent were loss to follow up (24.4%) and death (23.1%). Previous MDR-TB treatment was associated with unsuccess and death. Illicit drugs use was a risk factor for unsuccess and loss to follow up. The six-month culture conversion was a protective factor for unsuccess and death. Administration of antiretroviral therapy was a protective factor for unsuccess, death and failure. The risk of XDR-TB cases for failure was six times higher than that of MDR-TB patients. All XDR-TB / HIV positive cases had unfavorable outcomes. Conclusions: Early identication and immediate initiation of appropriate treatment are key to reducing unfavorable outcomes among MDR-TB / HIV positive. The high proportions of loss to follow-up and death may reect diculties in monitoring treatment, due to poor integration between the TB and HIV programs. In addition, none of XDR-TB / HIV positive patient had therapeutic success, highlighting the urgency to incorporate new drugs for the treatment of multidrug-resistant TB by Brazil's Ministry of Health.

recommendation is for patients with CD4 < 50 cells/mm. In other patients, ART should begin on the 8th week. Concomitant initiation of treatment for both conditions is not recommended (10).
Outcomes: de nition of key terms Treatment outcomes were classi ed according to WHO recommendations (11). The former are mutually exclusive.
"Cure" occurs when the patient has at least three negative cultures after the 12th month of treatment. "Complete treatment" is the case when the patient has completed the stipulated time for treatment, with favorable clinical and radiological evolution, but without the accompanying cultures. "Loss of follow-up" takes place when the treatment was interrupted for 2 consecutive months or more. "Failure" is the outcome if the patient had two or more positive cultures in three recommended ones after the 12th month of treatment or three consecutive positive cultures after the 12th month of treatment, at least 30 days apart. Failure can also be the case according to medical evaluation and decision to change treatment early due to clinical and radiological worsening. "Death" is the outcome whenever the patient died for any reason during treatment. The "unfavorable" outcome is the sum of patients who had results classi ed as death, failure or loss of follow-up. "Success" of treatment is the sum of patients who had the result classi ed as complete cure and treatment.

Independent variables
The following variables were included in the regression analyses: sex, being under 37 years old, having less than 8 years of schooling, race/color, diabetes, comorbidities (viral hepatitis, renal failure, neoplasia, silicosis, transplant, mental disorder, use long-term use of corticosteroids, users of TNF alpha inhibitors, seizures and other unde ned drugs, use of illicit drugs, alcohol dependence, smoking, unemployment, drug resistance category (MDR-TB or XDR-TB), treatment regimen (standardized or individualized), type of drug resistance (primary, that is, patients with no history of previous treatment; or acquired, that is, patients already treated for TB for one month or more) (12), disease extension (presence of chest cavity and/or bilateral disease), previous MDR-TB treatment, use of ART, and six-month culture conversion (de ned as those patients with at least two negative cultures by the sixth month of treatment).
The standardized treatment regimen for MDR-TB is recommended and applied in Brazil and must include at least one rst-line oral drug, if susceptible, a uoroquinolone, an injectable drug along with terizidone. Individualized regimens may include other drugs, such as clofazimine, linezolid, imipenem and isoniazid in high doses. They are restricted to patients with additional resistance to rst-line drugs, pre-XDR, XDR-TB, and to patients who have had adverse events with standardized regimen (7).
Patients who had more than one treatment for MDR-TB registered with SITETB were considered to have had previous treatment for MDR-TB.
Only variables that had a maximum of 10% of missing values were selected. For this reason, the results of counting CD4, CD8 and HIV viral load were only described.

Statistical methods
Given the nature of the dependent variable, number (frequency) and median (interquartile range [IQR] 25% -75%) were used to describe the characteristics of the patients and the treatment outcome according to coinfection status (TB-MDR HIV negative, positive and unknown). Treatment outcomes were also described according to the drug resistance pattern (MDR or XDR-TB), and coinfection. The chi-square test (χ2) or Fisher's exact test were used to compare the proportions between groups.
Cox proportional hazards models were implemented to estimate hazard rates (HR) between each of the treatment outcomes and the covariates in the MDR-TB HIV positive patient group. Variables with signi cance levels ≤0. 20  Failure was twice as high among HIV negative patients (10.5%) as among HIV positive patients (5.1%).

Drug resistance and treatment outcomes
Upon analysis of treatment outcomes by HIV status and drug resistance (see Table 2), MDR or XDR-TB, it was observed that MDR-TB patients, regardless of HIV serological status, have more therapeutic success. The success among XDR-TB HIV negative/unknown cases was only 20.8%, while, in the same period, there was no case of therapeutic success among XDR-TB HIV positive cases. Death was more frequent in HIV positive, especially in XDR-TB patients, while failure was practically the same among XDR-TB groups, but it was lower among MDR-TB HIV positive (2.8%) than among MDR-TB HIV negative cases (8.5%).  Among the 2,269 cases analyzed, the proportion of MDR-TB/HIV coinfection was 6.9%. In 2016, according to MoH, the national rate of HIV coinfection among new TB cases was 9.4%. In the same year, the proportion for the state of RJ was 8.7% (13). There is little data on MDR-TB/HIV coinfection in Brazil. In a study conducted with MDR-TB patients from all over the country, 9% of coinfection was observed (14), while another study conducted in a reference hospital in the city of São Paulo found 4% (15).
In general, in the present study, it was observed a therapeutic success rate of 56.3% among MDR-TB HIV negative cases and 47.4% among HIV positive cases. Recently, in a meta-analysis with 271 publications, including 3,368 MDR-TB HIV positive participants from Sub-Saharan Africa, treatment success was 44.8% (16).
Regarding unfavorable treatment outcomes, loss of follow-up was higher among HIV positive cases (24.4%) than among HIV negative cases (18.4%). According to WHO, the Americas has the highest proportion of cases with loss of follow-up, with 26% among MDR and rifampicin-resistant TB cases, while at the global level it is 14% (17).
Although treatment is free of charge for both TB and HIV/AIDS in Brazil, the treatment generates extra costs such as food and access to services that consume a signi cant percentage of the income of the poorest patients (18). In addition, the scenario is aggravated by the barriers encountered by TB patients in accessing social protection measures. A study conducted with patients at a referral center for the MDR-TB treatment showed that only 38% of the participants reported being bene ciaries of social protection (19). A meta-analysis of low-and middle-income countries showed that social protection measures are associated with successful treatment and reduced lost of follow-up, in addition to being associated with a lower risk of impoverishment (20).
In our study, being an illicit drugs user (IDU) was a risk factor for treatment failure (HR 1.83 95% CI 1.10-3.04 p = 0.019), and generated three times more risk for loss of follow-up (1.67-6.31 p = 0.001). Although there is no speci cation on the type of drug, several studies have described the association between TB, HIV, and the illicit drugs use, whether injectable or not (21)(22)(23)(24)(25). Illicit drug use is often associated with other risk factors for TB, such as smoking, alcohol abuse, and incarceration (26). Thus, important additional barriers remain in the treatment of TB in this group. Therefore, Illicit drug users have more di culty completing medical evaluations or adhering to Treatment for TB (3) and, when symptomatic, tend to wait longer to start appropriate treatment (27). A study conducted in the United States of America showed that, at the time of AIDS diagnosis, patients with a history of injecting drugs were 3.5 times more likely (95% CI 1.3-10.2) to have an opportunistic infection, including TB (28). This suggests that there is a lower demand for health care among IDUs, which, by its turn, can lead to more severe illnesses and contribute to an increase in TB transmission rates (29).
The challenge of maintaining high levels of adherence among IDUs highlights the current challenges facing TB elimination, which may require the provision and coordination of TB and Mental Health services with experience in handling illicit drug users, including targeted testing and treatment (30).
The proportion of death among MDR-TB HIV positive patients (23.1%) was also higher than among HIV negative cases (14.8%). Several studies report higher mortality and lower mean survival among MDR-TB HIV positive patients as compared to those without infection (31)(32)(33)(34) monitoring by healthcare professionals is necessary. Co-management of patients using ART and MDR/XDR drugs is complex and, therefore, it is essential that health professionals are trained to recognize potential and additive toxicity due to concomitant regimens.
In addition, measures at health services must be taken in order to promote the patient's commitment and adherence. Careful monitoring of patients is necessary to ensure that adverse effects are recognized quickly.
Finally, the present study has some limitations. As data were collected retrospectively, there was a lot of missing data on CD4, CD and HIV viral load. Besides, it was not possible to determine the period for starting ART. Also, due to the reduced number of records, it was not possible to assess the adverse events within the group of MDR/XDR HIV positive and their possible relationship with unfavorable outcome. Another fact that needs to be highlighted is that in the SITETB database, until 2015, variables related to diabetes, comorbidities, illicit drugs use, alcohol dependence, and smoking, whenever classi ed as "no," can also mean an absence of information. There is also no standardization for the classi cation of alcohol dependence, smoking, and mental health disorders. At any rate, the statistical ndings reported here are robust due to the large sample size on which they are based.

Conclusion
The unfavorable results for MDR-TB HIV positive patients in RJ in 2000-2016 were associated with the illicit drugs use and previous treatment for MDR-TB. Six-month culture conversion and use of ART were protective factors. This shows, on the one hand, that early identi cation and immediate initiation of appropriate treatment can reduce unfavorable outcomes among MDR-TB HIV positive patients. On the other, the high proportions of loss to followup (24.4%) and death (23.1%) may re ect di culties in monitoring treatment, given the lack of integration between the TB and HIV programs. In addition, the fact that none of XDR-TB HIV patient had therapeutic success re ects the limitation of treatment options, and the urgency of Brazil's health system to incorporate new drugs in the treatment of multidrug resistance.