Optimizing MDR-TB Treatment Regimens through (Time-Varying) Analyses of Retrospective Data Carole D. Mitnick, ScD Harvard Medical School/Partners In Health 2da Jornada Científica: Investigación Peruana para el Control de la Tuberculosis 21 Marzo 2014

Summary Global burden MDR, status update Guidelines Open Questions Introduction to collaboration among SES, MINSA, HMS  Studies  Conclusions  Discussion    

% of TB cases with MDR-TB, 2012 % of previously treated cases

% of new cases

Source: WHO, 2013

MDR-TB Patients on Treatment 600000 500000

450000

400000 300000

200000 100000

94000

77000

detected MDR

reported on treatment

0 estimated cases of mdr

Global Treatment Outcomes, MDR-TB 2007-10

Global Guidance for MDRTB Treatment

Adapted from Keshavjee, N Engl J Med. 2012 Sep 6

Focus on regimen composition

Current WHO treatment guidelines (2011)  Total duration of 20 months  Intensive phase of 8 months  Use 4 drugs (likely susceptible) + PZA during the intensive phase  Use PZA, FQ, inject, thiomide, CS or PAS

Aggressive Regimens for MDRTB

 ≥5 likely effective drugs during the intensive phase (including a FQ and injectable)  ≥4 likely effective drugs during the continuation phase (including FQ)  First-line oral drugs & highend dosing used whenever possible

MDR-TB: The Disease

Examining alternative approaches to regimen composition  Mitnick CD, Franke MF, Rich ML, Alcántara Viru FA, Appleton SC, Atwood SS, et al. Aggressive regimens for multidrugresistant tuberculosis decrease all-cause mortality. PLoS ONE. 2013;8(3):e58664.  Franke MF, Appleton SC, Mitnick CD, Furin JJ, Bayona J, Chalco K, et al. Aggressive Regimens for multidrug-resistant tuberculosis reduce recurrence. Clin. Infect. Dis. 2013:56(6):770-6.  Seung KJ, Becerra MC, Atwood SS, Alcantara F, Bonilla CA, Mitnick CD. Clin Microbiol Infect. 2013: 1-6.

Objectives  Estimate the association between receiving an “aggressive” regimen and: 1. Mortality 2. Recurrence  Explore frequency and effect of standard from 2004 Lancet article  Extend to changes on treatment

Methods1,2  Study population: 673 MDR patients in Lima who received ITR 1997-2002  Excluded if not 1st ITR & if regimen info not available

 Aggressive if contained >=5 likely effective drugs (Groups 1-IV), including FQ & injectable  Classification varied by month: >=75% of aggressive days/month=aggressive month  Time-varying Cox proportional hazards models  Endpoints: death1, recurrence2  Adjusted for: age, sex, prior treatment, severity, comorbidities, resistance  Evaluated interaction with time 1Mitnick

et al, PLoS Med 2013; 2Franke et al, CID, 2013;

Results1: Baseline characteristics

Results: Outcomes* among 669 patients included in mortality analysis 82% received aggressive regimen for ≥1 month; mean duration 21 (IQR:15-26) months

Outcome

N (%)

Cured/completed

442 (66.1)

Treatment Failed

17 (2.5)

Died

129 (20.8)

Defaulted

67 (10.0)

Missing/Transferred out

4 (0.6)

Total

669 (100) *Classified per Laserson et al., IJTLD, 2005

Results1: Aggressive regimen & death (n=669)

End point: time from initiation of the individualized regimen to death from any cause, while on treatment 1Mitnick

et al, PLoS Med 2013

2: Recurrence

Methods2: Recurrence  Population: 673 Lima patients who received ITR for MDR  Aggressive: >=5 likely effective drugs, including FQ & inject  Classification: >=75% of aggressive days=aggressive month  Exposure of interest: aggressive regimen for >=18 months after conversion  Time-varying Cox proportional hazards models

 Chart reviews and home visits between 2004-8  Endpoints: recurrence after cure/completion 1. Positive culture, or 2. initiated TB treatment  Adjusted for: age, sex, prior treatment, severity, comorbidities, resistance

2Franke

et al, CID, 2013;

Results2: Baseline characteristics recurrence (n=422)

2Franke

et al, CID, 2013

Results2: Aggressive regimen & recurrence (N=402)

2Franke

et al, CID, 2013

3: Salvage

Methods3  Study population: 625 MDR patients in Lima who received ITR 97-07, persistent positivity  4 positive cultures in 180 days on treatment

 Exposure: received “salvage” therapy  ≥ 1 new drug, 30 days before or after persistent positivity  Priority to drugs NEVER used before

 Endpoints: culture conversion  Univariate analysis: odds ratio

3Seung

et al., CMI 2013

Results3: Baseline characteristics salvage (N=504) Characteristic

Salvage (n=213)

Non-salvage (n=291)

# (%)

# (%)

Age Male

Median (IQR) 28(16)

128 (60)

Median (IQR) 29(14)

174 (60)

# prev rx

2(1)

2(1)

BMI

19.2 (3.5)

19.6(4.9)

Bilateral, cav xray

80(50)

85 (46)

MDR

183 (92)

225(90)

MDR + FQ or inject

85 (50)

104 (50)

MDR + FQ + inject

19 (12)

39 (20)

3 Seung et al, CMI, 2013

Distribution of # of drugs in salvage regimens (n=213)

Results3: Salvage patient and regimen characteristics & conversion (N=213) Characteristic

OR (95% CI)

Characteristic

OR (95% CI)

Male

0.8 (0.5-1.5)

Clari

1.1 (0.5-2.2)

Age>28

0.9 (0.5-1.7)

PAS

1.0 (0.4-2.7)

Pre-XDR

0.8 (0.4-1.5)

SM

1.3 (0.5-3.3)

XDR

0.4 (0.1-1.4)

Z

0.6 (0.2-1.8)

≥2 new drugs

1.3 (0.7-2.4)

E

1.9 (0.7-5.2)

MFX

2.1 (1.1-3.8)

CFZ

2.4 (0.8-7.2)

CM

1.3 (0.7-2.5)

RFB

1.7 (0.5-5.5)

Amox/Clav

1.1 (0.6-2.2)

Thiomide

0.2 (0.03-1.9)

KM

1.4 (0.7-2.8)

3 Seung et al, CMI, 2013

Discussion/Next Steps  Aggressive regimen—containing >=5 likely effective drugs, including FQ & injectable—is associated with improved outcomes  Robust benefit across endpoints and in adjusted analyses  Other independent predictors of poor outcomes: + prior treatment exposure, comorbidities (diabetes for recurrence), and disease severity  Aggressive regimens should be standard of care for treatment and trials  Although salvage regimens can improve conversion, use strongest regimen up front

Limitations  Retrospective, unmeasured confounding  Exposure not randomized, still some misclassification  For recurrence, no distinction between relapse & reinfection; uncertainty about diabetes care after MDR-TB treatment completion  Survival cohort, generalizable to new patients?  Salvage small on able to detect large differences, univariate

Acknowledgements         

patients health promoters Pulmonologists/MINSA PIH/SES SES psycho-social support team specialists lab workers data abstraction-entry staff Just In Time

Bill & Melinda Gates Foundation, T.J. White, David Rockefeller Center for Latin American Studies at Harvard, Francis Family Foundation, Pittsfield Anti-Tuberculosis Association, Eli Lilly Foundation, Hatch Family Foundation, NIH/NIAID-NHLBI

www.unmaskstigma.org