Use of High-Resolution Geospatial and Genomic Data to Characterize Recent Tuberculosis Transmission, Botswana.

Combining genomic and geospatial data can be useful for understanding Mycobacterium tuberculosis transmission in high-burden tuberculosis (TB) settings. We performed whole-genome sequencing on M. tuberculosis DNA extracted from sputum cultures from a population-based TB study conducted in Gaborone, Botswana, during 2012-2016. We determined spatial distribution of cases on the basis of shared genotypes among isolates. We considered clusters of isolates with ≤5 single-nucleotide polymorphisms identified by whole-genome sequencing to indicate recent transmission and clusters of ≥10 persons to be outbreaks. We obtained both molecular and geospatial data for 946/1,449 (65%) participants with culture-confirmed TB; 62 persons belonged to 5 outbreaks of 10-19 persons each. We detected geospatial clustering in just 2 of those 5 outbreaks, suggesting heterogeneous spatial patterns. Our findings indicate that targeted interventions applied in smaller geographic areas of high-burden TB identified using integrated genomic and geospatial data might help interrupt TB transmission during outbreaks.

Using genetic data to identify transmission risk factors: Statistical assessment and application to tuberculosis transmission.

Identifying host factors that influence infectious disease transmission is an important step toward developing interventions to reduce disease incidence. Recent advances in methods for reconstructing infectious disease transmission events using pathogen genomic and epidemiological data open the door for investigation of host factors that affect onward transmission. While most transmission reconstruction methods are designed to work with densely sampled outbreaks, these methods are making their way into surveillance studies, where the fraction of sampled cases with sequenced pathogens could be relatively low. Surveillance studies that use transmission event reconstruction then use the reconstructed events as response variables (i.e., infection source status of each sampled case) and use host characteristics as predictors (e.g., presence of HIV infection) in regression models. We use simulations to study estimation of the effect of a host factor on probability of being an infection source via this multi-step inferential procedure. Using TransPhylo-a widely-used method for Bayesian estimation of infectious disease transmission events-and logistic regression, we find that low sensitivity of identifying infection sources leads to dilution of the signal, biasing logistic regression coefficients toward zero. We show that increasing the proportion of sampled cases improves sensitivity and some, but not all properties of the logistic regression inference. Application of these approaches to real world data from a population-based TB study in Botswana fails to detect an association between HIV infection and probability of being a TB infection source. We conclude that application of a pipeline, where one first uses TransPhylo and sparsely sampled surveillance data to infer transmission events and then estimates effects of host characteristics on probabilities of these events, should be accompanied by a realistic simulation study to better understand biases stemming from imprecise transmission event inference.

The association of household food insecurity and HIV infection with common mental disorders among newly diagnosed tuberculosis patients in Botswana.

OBJECTIVE: To determine the association between food insecurity and HIV infection with depression and anxiety among new tuberculosis (TB) patients. DESIGN: Our cross-sectional study assessed depression, anxiety and food insecurity with Patient Health Questionnaire (PHQ-9), Zung Anxiety Self-Assessment Scale (ZUNG) and Household Food Insecurity Access Scale, respectively. Poisson regression models with robust variance were used to examine correlates of depression (PHQ-9 ≥ 10) and anxiety (ZUNG ≥ 36). SETTING: Gaborone, Botswana. PARTICIPANTS: Patients who were newly diagnosed with TB. RESULTS: Between January and December 2019, we enrolled 180 TB patients from primary health clinics in Botswana. Overall, 99 (55·0 %) were HIV positive, 47 (26·1 %), 85 (47·2 %) and 69 (38·5 %) indicated depression, anxiety and moderate to severe food insecurity, respectively. After adjusting for potential confounders, food insecurity was associated with a higher prevalence of depression (adjusted prevalence ratio (aPR) = 2·30; 95 % CI 1·40, 3·78) and anxiety (aPR = 1·41; 95 % CI 1·05, 1·91). Prevalence of depression and anxiety was similar between HIV-infected and HIV-uninfected participants. Estimates remained comparable when restricted to HIV-infected participants. CONCLUSIONS: Mental disorders may be affected by food insecurity among new TB patients, regardless of HIV status.

Population-Based Geospatial and Molecular Epidemiologic Study of Tuberculosis Transmission Dynamics, Botswana, 2012-2016.

Tuberculosis (TB) elimination requires interrupting transmission of Mycobacterium tuberculosis. We used a multidisciplinary approach to describe TB transmission in 2 sociodemographically distinct districts in Botswana (Kopanyo Study). During August 2012-March 2016, all patients who had TB were enrolled, their sputum samples were cultured, and M. tuberculosis isolates were genotyped by using 24-locus mycobacterial interspersed repetitive units-variable number of tandem repeats. Of 5,515 TB patients, 4,331 (79%) were enrolled. Annualized TB incidence varied by geography (range 66-1,140 TB patients/100,000 persons). A total of 1,796 patient isolates had valid genotyping results and residential geocoordinates; 780 (41%) patients were involved in a localized TB transmission event. Residence in areas with a high burden of TB, age <24 years, being a current smoker, and unemployment were factors associated with localized transmission events. Patients with known HIV-positive status had lower odds of being involved in localized transmission.

Vulvar cancer in Botswana in women with and without HIV infection: patterns of treatment and survival outcomes.

OBJECTIVES: Vulvar cancer is a rare gynecological malignancy. However, the incidence of human papillomavirus (HPV)-associated vulvar disease is increasing, particularly in low- and middle-income countries. HIV infection is associated with an increased risk of HPV-associated vulvar cancer. We evaluated treatment patterns and survival outcomes in a cohort of vulvar cancer patients in Botswana. The primary objective of this study was to determine overall survival and the impact of treatment modality, stage, and HIV status on overall survival. METHODS: Women with vulvar cancer who presented to oncology care in Botswana from January 2015 through August 2019 were prospectively enrolled in this observational cohort study. Demographics, clinical characteristics, treatment, and survival data were collected. Factors associated with survival including age, HIV status, stage, and treatment were evaluated. RESULTS: Our cohort included 120 women with vulvar cancer. Median age was 42 (IQR 38-47) years. The majority of patients were living with HIV (89%, n=107) that was well-controlled on antiretroviral treatment. Among women with HIV, 54.2% (n=58) were early stage (FIGO stage I/II). In those without HIV, 46.2% (n=6) were early stage (stage I/II). Of the 95 (79%) patients who received treatment, 20.8% (n=25) received surgery, 67.5% (n=81) received radiation therapy, and 24.2% (n=29) received chemotherapy, either alone or in combination. Median follow-up time of all patients was 24.7 (IQR 14.2-39.1) months and 2- year overall survival for all patients was 74%. Multivariate analysis demonstrated improved survival for those who received surgery (HR 0.26; 95% CI 0.08 to 0.86) and poor survival was associated with advanced stage (HR 2.56; 95% CI 1.30 to 5.02). Survival was not associated with HIV status. CONCLUSIONS: The majority of women with vulvar cancer in Botswana are young and living with HIV infection. Just under half of patients present with advanced stage, which was associated with worse survival. Improved survival was seen for those who received surgery.

Chemoradiation versus radiation alone in stage IIIB cervical cancer patients with or without human immunodeficiency virus.

OBJECTIVE: Cervical cancer remains the most common cancer among women in sub-Saharan Africa and is also a leading cause of cancer related deaths among these women. The benefit of chemoradiation in comparison with radiation alone for patients with stage IIIB disease has not been evaluated prospectively in women living with human immunodeficiency virus (HIV). We assessed the survival of chemoradiation versus radiation alone among stage IIIB cervical cancer patients based on HIV status. METHODS: Between February 2013 and June 2018, patients with International Federation of Gynecology and Obstetrics (FIGO) 2009 stage IIIB cervical cancer with or without HIV and treated with chemoradiation or radiation alone, were prospectively enrolled in an observational cohort study. Overall survival was evaluated using the Kaplan-Meier method. Cox proportional hazards modeling was used to analyze associations with survival. RESULTS: Among 187 patients, 63% (n=118) of women had co-infection with HIV, and 48% (n=69) received chemoradiation. Regardless of HIV status, patients who received chemoradiation had improved 2 year overall survival compared with those receiving radiation alone (59% vs 41%, p<0.01), even among women living with HIV (60% vs 38%, p=0.02). On multivariable Cox regression analysis, including all patients regardless of HIV status, 2 year overall survival was associated with receipt of chemoradiation (hazard ratio (HR) 0.63, p=0.04) and total radiation dose ≥80 Gy (HR 0.57, p=0.02). Among patients who received an adequate radiation dose of ≥80 Gy, adjusted overall survival rates were similar between chemoradiation versus radiation alone groups (HR 1.07; p=0.90). However, patients who received an inadequate radiation dose of <80 Gy, adjusted survival was significantly higher in chemoradiation versus radiation alone group (HR 0.45, p=0.01). CONCLUSIONS: Addition of chemotherapy to standard radiation improved overall survival, regardless of HIV status, and is even more essential in women who cannot receive full doses of radiation.

Possible Transmission Mechanisms of Mixed Mycobacterium tuberculosis Infection in High HIV Prevalence Country, Botswana.

Tuberculosis caused by concurrent infection with multiple Mycobacterium tuberculosis strains (i.e., mixed infection) challenges clinical and epidemiologic paradigms. We explored possible transmission mechanisms of mixed infection in a population-based, molecular epidemiology study in Botswana during 2012-2016. We defined mixed infection as multiple repeats of alleles at >2 loci within a discrete mycobacterial interspersed repetitive unit-variable-number tandem-repeat (MIRU-VNTR) result. We compared mixed infection MIRU-VNTR results with all study MIRU-VNTR results by considering all permutations at each multiple allele locus; matched MIRU-VNTR results were considered evidence of recently acquired strains and nonmatched to any other results were considered evidence of remotely acquired strains. Among 2,051 patients, 34 (1.7%) had mixed infection, of which 23 (68%) had recently and remotely acquired strains. This finding might support the mixed infection mechanism of recent transmission and simultaneous remote reactivation. Further exploration is needed to determine proportions of transmission mechanisms in settings where mixed infections are prevalent.

A Neighbor-Based Approach to Identify Tuberculosis Exposure, the Kopanyo Study.

Contact investigation is one public health measure used to prevent tuberculosis by identifying and treating persons exposed to Mycobacterium tuberculosis. Contact investigations are a major tenet of global tuberculosis elimination efforts, but for many reasons remain ineffective. We describe a novel neighbor-based approach to reframe contact investigations.

Human papillomavirus genotypes in women with invasive cervical cancer with and without human immunodeficiency virus infection in Botswana.

Cervical cancer remains a significant cause of morbidity and mortality in women worldwide and is the leading cause of cancer-related death in Botswana. It is well established that women with HIV have a higher risk of persistent HPV infection leading to cervical cancer. We assessed HPV prevalence and genotype distribution in 126 tissue specimens from confirmed invasive cervical cancer cases using Abbott real-time PCR assay. Overall, 88 (69.8%) women were HIV-infected. Fifty-seven (64.8%) of the HIV-infected women had a baseline CD4+ count ≥350 cells/µl, and 82 (93.2%) were on antiretroviral therapy at the time of cervical cancer diagnosis. The median age of HIV-infected patients was significantly younger than that of HIV-uninfected patients (p < 0.001). HPV DNA was detected in all of 126 (100%) of tissues analyzed in our study. The HPV genotypes identified included the HPV-16 (75.4%), HPV-18 (28.6%) and other high-risk (hr) HPV genotypes (16.7%). HIV infection was positively associated with the presence of the HPV-16 genotype (p = 0.036), but not with HPV-18 or with other high-risk (hr)-HPV genotypes. Thirty-three percent of the patients had multiple hr-HPV genotypes, with higher rates in HIV-infected women. These results highlight the importance and potential impact of large-scale HPV vaccination programs covering HPV-16 and HPV-18 genotypes in countries like Botswana with high burden of HIV infection.

Tumor microbial diversity and compositional differences among women in Botswana with high-grade cervical dysplasia and cervical cancer.

INTRODUCTION: We characterized the cervical 16S rDNA microbiome of patients in Botswana with high-grade cervical dysplasia and locally advanced cervical cancer. METHODS: This prospective study included 31 patients: 21 with dysplasia and 10 with cancer. The Shannon diversity index was used to evaluate alpha (intra-sample) diversity, while the UniFrac (weighted and unweighted) and Bray-Curtis distances were employed to evaluate beta (inter-sample) diversity. The relative abundance of microbial taxa was compared among samples using linear discriminant analysis effect size. RESULTS: Alpha diversity was significantly higher in patients with cervical cancer than in patients with cervical dysplasia (P<0.05). Beta diversity also differed significantly (weighted UniFrac Bray-Curtis, P<0.01). Neither alpha diversity (P=0.8) nor beta diversity (P=0.19) varied by HIV status. The results of linear discriminant analysis effect size demonstrated that multiple taxa differed significantly between patients with cervical dysplasia vs cancer. Lachnospira bacteria (in the Clostridia class) were particularly enriched among cervical dysplasia patients, while Proteobacteria (members of the Firmicutes phyla and the Comamonadaceae family) were enriched in patients with cervical cancer. DISCUSSION: The results of our study suggest that differences exist in the diversity and composition of the cervical microbiota between patients with cervical dysplasia and patients with cervical cancer in Botswana. Additional studies are warranted to validate these findings and elucidate their clinical significance among women living in sub-Saharan Africa, as well as other regions of the world.

Gene expression profiling informs HPV cervical histopathology but not recurrence/relapse after LEEP in ART-suppressed HIV+HPV+ women.

Identification of factors associated with human papillomavirus (HPV) cervical histopathology or recurrence/relapse following loop electrosurgical excision procedure (LEEP) would allow for better management of the disease. We investigated whether gene signatures could (i) associate with HPV cervical histopathology and (ii) identify women with post-LEEP disease recurrence/relapse. Gene array analysis was performed on paraffin-embedded cervical tissue-isolated RNA from two cross-sectional cohorts of antiretroviral therapy (ART)-suppressed HIV+HPV+ coinfected women: (i) 55 women in South Africa recruited into three groups: high risk (HR) (-) (n = 16) and HR (+) (n = 15) HPV without cervical histopathology and HR (+) HPV with cervical intraepithelial neoplasia (CIN) grade 1/2/3 (n = 24), (ii) 28 women in Botswana with CIN2/3 treated with LEEP 12-month prior to recruitment and presenting with (n = 13) and without (n = 15) lesion recurrence/relapse (tissue was analyzed at first LEEP). Three distinct gene expression signatures identified were able to segregate: (i) HR+ HPV and CIN1/2/3, (ii) HR HPV-free and cervical histopathology-free and (iii) HR+ HPV and cervical histopathology-free. Immune activation and neoplasia-associated genes (n = 272 genes; e.g. IL-1A, IL-8, TCAM1, POU4F1, MCM2, SMC1B, CXCL6, MMP12) were a feature of cancer precursor dysplasia within HR HPV infection. No difference in LEEP tissue gene expression was detected between women with or without recurrence/relapse. In conclusion, distinctive gene signatures were associated with presence of cervical histopathology in tissues from ART-suppressed HIV+/HPV+ coinfected women. Lack of detection of LEEP tissue gene signature able to segregate subsequent post-LEEP disease recurrence/relapse indicates additional factors independent of local gene expression as determinants of recurrence/relapse.

HIV infection is not associated with the initiation of curative treatment in women with cervical cancer in Botswana.

BACKGROUND: Cervical cancer is the leading cause of cancer death in Sub-Saharan Africa. The risk of developing cancer is increased for women living with human immunodeficiency virus (HIV) infection. It is unknown which factors predict the initiation of curative chemoradiotherapy (CRT) in resource-limited settings and whether HIV is associated with initiating curative CRT in settings with a high HIV burden. METHODS: All women living with and without HIV infection who were initiating curative and noncurative CRT for locally advanced cervical cancer in Botswana were prospectively enrolled in an observational study. The factors associated with receiving CRT were evaluated in all patients and the subgroup of women living with HIV. RESULTS: Of 519 enrolled women, 284 (55%) initiated CRT with curative intent. The curative cohort included 200 women (70.4%) who were living with HIV and had a median CD4 count of 484.0 cells/µL (interquartile range, 342.0-611.0 cells/µL). In the noncurative cohort, 157 of 235 women (66.8%) were living with HIV and had a median CD4 count of 476.5 cells/µL (interquartile range, 308.0-649.5 cells/µL). HIV status was not associated with initiating curative CRT (odds ratio [OR], 0.95; 95% confidence interval [CI], 0.58-1.56). The factors associated with receiving curative CRT treatment on multivariable analysis in all patients included baseline hemoglobin levels ≥10 g/dL (OR, 1.80; 95% CI, 1.18-2.74) and stage I or II versus stage III or IV disease (OR, 3.16; 95% CI, 2.10-4.75). Women aged >61 years were less likely to receive curative treatment (OR, 0.43; 95% CI, 0.24-0.75). Among women who were living with HIV, higher CD4 cell counts were associated with higher rates of CRT initiation. CONCLUSIONS: The initiation of CRT with curative intent does not depend on HIV status. Significant predictors of CRT initiation include baseline hemoglobin level, disease stage, and age.

Utility of Targeted, Amplicon-Based Deep Sequencing To Detect Resistance to First-Line Tuberculosis Drugs in Botswana.

Multidrug-resistant tuberculosis (TB) is an alarming threat, and targeted deep sequencing (DS) may be an effective method for rapid identification of drug-resistant profiles, including detection of heteroresistance. We evaluated the sensitivity and specificity of targeted DS versus phenotypic drug susceptibility testing (pDST) among patients starting first-line anti-TB therapy in Botswana. Overall, we found high concordance between DS and pDST. Lower sensitivity of DS, which targets established high-confidence resistance variants, was observed for detecting isoniazid resistance among HIV-infected patients.

Optimizing ethambutol dosing among HIV/tuberculosis co-infected patients: a population pharmacokinetic modelling and simulation study.

BACKGROUND: Reduced ethambutol serum concentrations are commonly observed among TB patients co-infected with HIV and may lead to treatment failure. OBJECTIVES: To perform a population pharmacokinetic study of ethambutol in HIV/TB patients, and to evaluate an intensified ethambutol weight-based dosing strategy to support pharmacokinetic target attainment. METHODS: We conducted a prospective study of ethambutol pharmacokinetics among HIV/TB patients administered first-line TB treatment in Botswana, with study visits before and after initiation of ART. Clinical and disease status markers, including HIV-associated systemic immune activation and gut dysfunction biomarkers, were evaluated as covariates of ethambutol pharmacokinetic parameters in non-linear mixed effects analysis. Monte Carlo simulations were performed to compare pharmacokinetic target attainment under standard and intensified weight-based ethambutol dosing strategies. RESULTS: We studied 40 HIV/TB patients prior to initiation of ART, of whom 24 returned for a second visit a median of 33 days following ART initiation. Ethambutol serum concentrations were best explained by a two-compartment model with first-order elimination, with a significant improvement in oral bioavailability following ART initiation. In Monte Carlo simulations, a supplementary ethambutol dose of 400 mg daily led to >2-fold improvements in pharmacokinetic target attainment probabilities in lung tissue, both before and after ART initiation. CONCLUSIONS: Low serum ethambutol concentrations were commonly observed among HIV/TB patients in Botswana, and the oral bioavailability of ethambutol increased following ART initiation. Supplementary ethambutol dosing among HIV/TB patients may provide a strategy to optimize anti-TB treatment regimens in this high-risk population.

Mixed Mycobacterium tuberculosis-Strain Infections Are Associated With Poor Treatment Outcomes Among Patients With Newly Diagnosed Tuberculosis, Independent of Pretreatment Heteroresistance.

Background: Heteroresistant Mycobacterium tuberculosis infections (defined as concomitant infection with drug-resistant and drug-susceptible strains) may explain the higher risk of poor tuberculosis treatment outcomes observed among patients with mixed-strain M. tuberculosis infections. We investigated the clinical effect of mixed-strain infections while controlling for pretreatment heteroresistance in a population-based sample of patients with tuberculosis starting first-line tuberculosis therapy in Botswana. Methods: We performed 24-locus mycobacterial interspersed repetitive unit-variable number tandem-repeat analysis and targeted deep sequencing on baseline primary cultured isolates to detect mixed infections and heteroresistance, respectively. Drug-sensitive, micro-heteroresistant, macro-heteroresistant, and fixed-resistant infections were defined as infections in which the frequency of resistance was <0.1%, 0.1%-4%, 5%-94%, and ≥95%, respectively, in resistance-conferring domains of the inhA promoter, the katG gene, and the rpoB gene. Results: Of the 260 patients with tuberculosis included in the study, 25 (9.6%) had mixed infections and 30 (11.5%) had poor treatment outcomes. Micro-heteroresistance, macro-heteroresistance, and fixed resistance were found among 11 (4.2%), 2 (0.8%), and 11 (4.2%), respectively, for isoniazid and 21 (8.1%), 0 (0%), and 10 (3.8%), respectively, for rifampicin. In multivariable analysis, mixed infections but not heteroresistant infections independently predicted poor treatment outcomes. Conclusions: Among patients starting first-line tuberculosis therapy in Botswana, mixed infections were associated with poor tuberculosis treatment outcomes, independent of heteroresistance.

Molecular, Spatial, and Field Epidemiology Suggesting TB Transmission in Community, Not Hospital, Gaborone, Botswana.

During 2012-2015, 10 of 24 patients infected with matching genotypes of Mycobacterium tuberculosis received care at the same hospital in Gaborone, Botswana. Nosocomial transmission was initially suspected, but we discovered plausible sites of community transmission for 20 (95%) of 21 interviewed patients. Active case-finding at these sites could halt ongoing transmission.

Markers of gut dysfunction do not explain low rifampicin bioavailability in HIV-associated TB.

Background: Rifampicin is the key drug responsible for sterilizing activities in the first-line TB treatment regimen. Damage to the gut during acute and chronic HIV infection may inhibit drug absorptive capacity. We sought to test the hypothesis that markers of intestinal damage, bacterial translocation and systemic immune activation would relate to rifampicin bioavailability among HIV/TB patients. Patients and methods: We conducted a prospective cohort study of rifampicin pharmacokinetics in HIV/TB patients in Gaborone, Botswana. We performed two intensively sampled pharmacokinetic visits, before and after ART initiation. Non-linear mixed-effects modelling was performed to determine whether variability in markers of gut damage, microbial translocation or systemic immune activation contributed to variability in rifampicin bioavailability before and after the initiation of ART. Results: We enrolled 40 HIV/TB patients in the first pharmacokinetic visit and 24 patients returned for the second pharmacokinetic visit after initiating ART. Low rifampicin exposure, as defined by the maximum serum concentration, was observed in 40% of patients prior to initiating ART and 46% of patients after initiating ART. In the non-linear mixed-effects model, we did not observe significant covariate effects of markers of gut damage, microbial translocation or immune activation on rifampicin bioavailability before and after ART initiation. Discussion: Markers of intestinal damage, microbial translocation and systemic immune activation did not explain variability in rifampicin bioavailability. The a priori identification of HIV/TB patients at risk for low rifampicin concentrations remains a challenge, supporting a role for therapeutic drug monitoring during HIV/TB therapy.

Isoniazid clearance is impaired among human immunodeficiency virus/tuberculosis patients with high levels of immune activation.

AIMS: Immune activation, which is characteristic of both tuberculosis (TB) and human immunodeficiency virus (HIV) infection, is associated with impaired drug metabolism. We tested the hypothesis that elevated levels of systemic immune activation among adults with HIV/TB initiating antiretroviral therapy (ART) would be associated with impaired clearance of isoniazid. METHODS: We conducted a prospective observational study of isoniazid pharmacokinetics (PK) and systemic immune activation prior to and 1 month after ART initiation. Nonlinear mixed effects analysis was performed to measure the covariate effect of immune activation on isoniazid clearance in a model that also included N-acetyltransferase-2 (NAT-2) genotype and interoccasional variability on clearance (thereby analyzing the PK data before and after ART initiation in a single model). RESULTS: We enrolled 40 patients in the PK visit prior to ART, and 24 patients returned for the second visit a median of 33 days after initiating antiretroviral therapy. The isoniazid concentration data were best described by a two-compartment model with first-order elimination. After accounting for NAT-2 genotype, increasing levels of CD38 and HLA-DR expression on CD8+ T cells (CD38+ DR+ CD8+ ) were associated with decreasing isoniazid clearance. CONCLUSION: HIV/TB patients with high levels of immune activation demonstrated impaired isoniazid clearance. Future efforts should determine the role of this relationship in clinical hepatotoxicity events.

Artificial Intelligence and Amikacin Exposures Predictive of Outcomes in Multidrug-Resistant Tuberculosis Patients.

Aminoglycosides such as amikacin continue to be part of the backbone of treatment of multidrug-resistant tuberculosis (MDR-TB). We measured amikacin concentrations in 28 MDR-TB patients in Botswana receiving amikacin therapy together with oral levofloxacin, ethionamide, cycloserine, and pyrazinamide and calculated areas under the concentration-time curves from 0 to 24 h (AUC0-24). The patients were followed monthly for sputum culture conversion based on liquid cultures. The median duration of amikacin therapy was 184 (range, 28 to 866) days, at a median dose of 17.30 (range 11.11 to 19.23) mg/kg. Only 11 (39%) patients had sputum culture conversion during treatment; the rest failed. We utilized classification and regression tree analyses (CART) to examine all potential predictors of failure, including clinical and demographic features, comorbidities, and amikacin peak concentrations (Cmax), AUC0-24, and trough concentrations. The primary node for failure had two competing variables, Cmax of <67 mg/liter and AUC0-24 of <568.30 mg · h/L; weight of >41 kg was a secondary node with a score of 35% relative to the primary node. The area under the receiver operating characteristic curve for the CART model was an R(2) = 0.90 on posttest. In patients weighing >41 kg, sputum conversion was 3/3 (100%) in those with an amikacin Cmax of ≥67 mg/liter versus 3/15 (20%) in those with a Cmax of <67 mg/liter (relative risk [RR] = 5.00; 95% confidence interval [CI], 1.82 to 13.76). In all patients who had both amikacin Cmax and AUC0-24 below the threshold, 7/7 (100%) failed, compared to 7/15 (47%) of those who had these parameters above threshold (RR = 2.14; 95% CI, 1.25 to 43.68). These amikacin dose-schedule patterns and exposures are virtually the same as those identified in the hollow-fiber system model.