Date post: | 18-Dec-2015 |
Category: |
Documents |
Upload: | kelley-osborne |
View: | 212 times |
Download: | 0 times |
CCEB
Impact of HIV on Early MDR-TB Treatment Outcomes in
Botswana
Jeffrey Hafkin MD
Botswana-UPenn PartnershipCenter for Clinical Epidemiology and Biostatistics
Infectious Disease Division, Department of MedicineUniversity of Pennsylvania School of Medicine
CCEB
Background: MDR-TB and HIV-infection
• In the pre-HAART era, MDR-TB with HIV infection had worse treatment response and higher mortality compared to HIV-uninfected MDR-TB
• Few studies of MDR-TB outcomes in HIV+ with access to HAART, particularly in sub-Saharan Africa
Seung et al. PloS One 2009Brust et al. PloS One 2011
CCEB
Background: Epidemiology of HIV and TB in Botswana
• Prevalence of HIV infection = 25% (UNAIDS 2009)
• Incidence of TB disease = 503 cases per 100,000 population (WHO 2010)
• Prevalence of MDR-TB during 1995-2008:– 0.2% to 3.4% (treatment naïve)– 6.1% to 13.1% (treatment experienced)
• Start of HAART roll-out in Botswana in 2003
Nelson et al. Lancet 2005WHO. M/XDR-TB: 2010 global report on surveillance and response
CCEB
Study Design
• Cohort study– MDR-TB patients receiving individualized,
integrated, ambulatory care at a two public clinics in Botswana
• Exposure– HIV infection (two parallel HIV ELISA assays)
• Outcome– Sputum culture conversion (two consecutive
negative sputum cultures at least one month apart)
CCEB
Endpoints
• Primary:– Time to initial sputum culture clearance
after starting MDR-TB treatment – Proportion of patients converting sputum
cultures• Secondary:
– Proportion of patients with ototoxicity, peripheral neuropathy, and renal toxicity
CCEB
Data Collection
• Subject population: – All confirmed MDR-TB patients who were
started on anti-MDR-TB treatment prior to September 2008
– Excluded from analysis if no quantifiable culture follow up time after start of treatment
CCEB
Statistical Analysis
• Unadjusted Analysis:– Time to culture conversion by HIV status
• Log rank test and Kaplan-Meier curves
– Proportion of patients with sputum culture conversion by HIV status• Chi-Square test
• Adjusted Analysis:– Cox proportional hazard models controlling for
age, sex, number of active anti-TB agents and TB treatment history
CCEB
Baseline Characteristics
• 74 patients with culture-confirmed MDR-TB were identified– 4 were excluded
• no quantifiable culture f/u time after start of treatment
• 70 had complete data for analysis– 40 (57%) HIV-infected– 30 (43%) HIV-uninfected
CCEB
HIV-Related Characteristics• Baseline CD4+ count = 158 (IQR 88-347)
3m f/u CD4+ count = 262 (IQR 129-382)
• 28 (69%) on HAART prior to start of MDR-TB treatment and 36 (90%) during treatment
• Most regimens consisted of 2NRTI + NNRTI– 19 (53%) zidovudine/lamivudine– 8 (22%) stavudine/lamivudine– 8 (22%) tenofovir/lamivudine– 4 (11%) abacavir/lamivudine
CCEB
Baseline Characteristics by HIV Status
Characteristic HIV+N=40
HIV-N=30
P-value
Median Age (years, IQR) 39 (30-45) 33 (24-59) 0.4
Sex, n (%) Male Female
29 (73)11 (28)
22 (73)8 (27)
>0.5
TB Treatment History, n (%) Any 1st line therapy Any 2nd line (plus 1st line) Unknown
29 (73)7 (18)4 (10)
23 (77)3 (10)4 (13)
>0.5
Resistant Drugs at Baseline(median, IQR) 5 (4-5) 5 (5-5) >0.5
Number Active Drugs(median, IQR) 4 (4-4) 4 (4-4) >0.5
CCEB
1st Line Anti-TB Drug Resistance by HIV Status
Drug n (%)
HIV+N=40
HIV-N=30
P-value
Ethambutol Yes No Unknown
24 (60)16 (40)
0 (0)
23 (77)7 (23)0 (0)
0.14
Streptomycin Yes No Unknown
25 (63)13 (33)
2 (5)
19 (63)9 (30)2 (7)
>0.5
CCEB
2nd Line Anti-TB Drug Resistance by HIV Status
Drug n (%)
HIV+N=40
HIV-N=30
P-value
Kanamycin Yes No Unknown
1 (3)18 (45)21 (53)
3 (10)12 (40)15 (50)
0.4
Ofloxacin Yes No Unknown
2 (5)16 (40)22 (55)
3 (10)10 (33)17 (57)
>0.5
Ethionamide Yes No Unknown
7 (18)12 (30)21 (53)
6 (20)9 (30)
15 (50)
>0.5
CCEB
Anti-TB Drugs Used by HIV Status
Drug, n (%) HIV+ (n=40) HIV- (n=30) P value
Ethambutol 12 (30%) 8 (27%) >0.5
Pyrazinamide 38 (95%) 29 (97%) >0.5
Ciprofloxacin 37 (93%) 27 (90%) >0.5
Ofloxacin 0 (0%) 1 (3%) 0.4
Moxifloxacin 2 (5%) 3 (10%) 0.4
Streptomycin 2 (5 %) 1 (3%) >0.5
Amikacin 35 (88%) 20 (83%) >0.5
Capreomycin 2 (5%) 3 (10%) 0.4
Cycloserine 27 (67%) 20 (67%) >0.5
Terizidone 0 (0%) 1 (3%) 0.4
PAS 2 (5%) 3 (10%) 0.4
Amox/Clav 2 (5%) 6 (20%) 0.05
Clarithromycin 3 (8%) 1 (3%) 0.5
CCEB
Follow-up
• Median on treatment follow up time:– 82 days (IQR 52-133)
• Median duration aminoglycoside: – 8 mos (no difference by HIV status, p=0.48)
• Sputum culture conversion:– Overall: 59 of 70 (84%)– HIV-infected: 34 of 40 (85%)– HIV-uninfected: 25 of 30 (83%)
(p>0.5)
CCEB
Follow-up
• 4 patients died following enrollment:– 2 following sputum culture conversion (both
HIV-infected)– 2 prior to culture conversion (both HIV-
uninfected)
• 2 patients identified with XDR TB during follow-up (one HIV+, one HIV-)
• No patients defaulted from care during the study period
CCEB
Time to Sputum Culture Conversion
• Median time to sputum culture conversion: – 78 days (IQR 42-186) for HIV-infected– 95 days (IQR 70-133) for HIV-uninfected
• (log rank p >0.5)
• Unadjusted HR = 0.9 (95% CI: 0.5 to 1.5)
• Adjusted HR = 0.8 (95% CI: 0.4 to 1.4)– adjusting for age, gender, TB treatment history,
and number of active agents
CCEB
HIV neg
HIV posLogrank Test, p>0.5
0.00
0.25
0.50
0.75
1.00
Pro
port
ion
with
Cul
ture
Cle
aran
ce
40 24 14 10 8 4 3HIV = Pos30 23 11 3 2 1 1HIV = Neg
Number at risk
0 60 120 180 240 300 360
Analysis Time (Days)
Kaplan-Meier Clearance Estimates
CCEB
Toxicity
Type of toxicity
HIV positiven (%)
HIV negativen (%)
P-value
Neuropathy 16 (40) 3 (10) p<0.01
Nephropathy 10 (25) 2 (7) p=0.04
Ototoxicity 21 (53) 21 (70) p=0.14
Neuro-psychiatric
3 (8) 3 (10) p>0.5
CCEB
Limitations
• Sample size• Generalizability:
– Botswana vs. Sub-Saharan Africa• Selection bias:
– Specialist care in a referral center– Missing culture/DST data
• Misclassification bias:– Characterization of toxicity
• Unmeasured confounding
CCEB
Conclusions
• HIV infection did not impact time to culture conversion in cohort of adults with MDR-TB from Botswana
• Suggests that in resource-limited settings with broad access to ART and individualized MDR-TB care, short term microbiologic outcomes may be comparable in HIV-infected and uninfected patients
CCEB
Conclusions
• Furthermore, high rates of drug toxicity overall and more likely to be associated with co-infection
CCEB
Acknowledgements
• Univ of Pennsylvania– Chawanga Modongo– Lephata Molopisi– Craig Newcomb– Elizabeth Lowenthal– Andrew Steinhoff– Rob Roy MacGregor– Harvey Friedman– Gregory Bisson
• Botswana NTP– Tore Steen, Howard
Moffet
• Botswana TB Reference Lab– Koobiditse Radisowa– Valentina Anisimova
• CDC-Atlanta– Victoria Gammino
• CDC-Botswana– Robert Makombe
• Funding Support:– NIH T32AI055435– Penn CFAR NIH
P30AI045008