Article Text

Original research
Testing and treatment for latent tuberculosis infection in people living with HIV and substance dependence: a prospective cohort study
  1. Tessa Runels1,
  2. Elizabeth J Ragan1,
  3. Alicia S Ventura2,3,
  4. Michael R Winter4,
  5. Laura F White5,
  6. C Robert Horsburgh1,5,6,
  7. Jeffrey H Samet2,3,
  8. Richard Saitz2,3,
  9. Karen R Jacobson1
  1. 1Section of Infectious Diseases, Boston Medical Center, Boston, Massachusetts, USA
  2. 2Section of General Internal Medicine, Boston Medical Center, Boston, Massachusetts, USA
  3. 3Clinical Addiction Research and Education Unit, Boston University School of Medicine, Boston, Massachusetts, USA
  4. 4Data Coordinating Center, Boston University School of Public Health, Boston, Massachusetts, USA
  5. 5Department of Biostatistics, Boston University School of Public Health, Boston, Massachusetts, USA
  6. 6Department of Global Health, Boston University School of Public Health, Boston, Massachusetts, USA
  1. Correspondence to Dr Karen R Jacobson; kjacobso{at}


Objective To quantify the proportion of people living with HIV (PLWH) with other tuberculosis (TB) risk factors that completed the latent tuberculosis infection (LTBI) care cascade and describe factors associated with attrition. The care cascade was defined as follows: (1) receipt of an LTBI test and result, (2) initiation of LTBI treatment and (3) completion of LTBI treatment.

Design Prospective cohort study.

Setting Reactivation of LTBI remains a large source of active TB disease in the USA. PLWH and those who use substances are at greater risk and are harder to engage and retain in care.

Participants Participants enrolled in a Boston cohort of PLWH from 2012 to 2014.

Primary and secondary outcome measures Our primary outcome was the number and proportion of participants who completed each stage of the cascade and the factors associated with completing each stage. Our secondary outcomes were differences between participants tested with an interferon gamma release assay (IGRA) versus tuberculin skin test and differences between participants who tested positive versus negative for LTBI.

Results Only 189 of 219 (86.3%) participants completed testing. Five of the 11 with LTBI initiated and three completed treatment. Participants tested with an IGRA were more likely to complete testing (OR 3.87, 95% CI 1.05 to 14.30) while among participants successfully tested, being foreign-born was associated with a positive test result (OR 3.95; 95% CI 1.13 to 13.77).

Conclusions Although the majority completed LTBI testing, our findings warrant further investigation in a larger cohort to better understand factors that lead to suboptimal treatment initiation and completion in a low-burden country.

  • HIV & AIDS
  • tuberculosis
  • substance misuse

Data availability statement

Data are available upon reasonable request. Data collected for the study are available to interested investigators in the URBAN ARCH Repository:

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Strengths and limitations of the study

  • While other studies have investigated predictors of latent tuberculosis infection (LTBI) care cascade completion in the USA, our analysis uses predictors derived from validated biobehavioural instruments (eg, Mini-International Neuropsychiatric Interview V.6.0, Addiction Severity Index and 30-day Timeline Followback), especially capturing substance use and dependence, reducing misclassification bias of these risk factors.

  • A unique element of our study is that we looked at individuals engaged in HIV care, that is, a group that, per guidelines, should be receiving LTBI testing in the clinical setting and therefore provides an opportunity to understand how well testing is done under these conditions.

  • Our study is limited by its small sample size, preventing us from further assessing risk factors for non-completion along the entire cascade of LTBI care.

  • Our study is limited in its generalisability: our participants were all engaged in HIV care at the time of enrolment, potentially increasing the likelihood that they received LTBI testing compared with the broader population of people living with HIV.

  • Our study was not randomised, which may lead to unmeasured confounding and impact validity of findings.


Reactivation of tuberculosis (TB) infection remains a source of active TB disease in the USA. An estimated 13 million people in the USA have latent TB infection (LTBI).1 To reduce TB disease incidence and mortality, TB testing and treatment is prioritised for individuals at increased risk for progression to active disease, including persons living with HIV (PLWH) and people who inject drugs (PWID).1 2

In PLWH, TB preventive therapy has been shown to reduce the risk of progression to active disease by up to 62%.3 HIV is one of the greatest risk factors for active TB development; PLWH are 20–30 times more likely to develop TB than individuals without HIV.4 Injection drug use (IDU) has also been associated with higher LTBI risk, with an estimated 10%–59% of PWID to be infected with TB.5 Alcohol use is also a risk factor for TB disease, with individuals who consume ≥40 g of alcohol daily having a nearly threefold increased risk.6 For these reasons, PLWH and individuals who use drugs and alcohol have been prioritised for LTBI testing and TB preventive therapy.

PLWH are a particularly high priority group for TB testing because for those with LTBI, TB preventive therapy decreases their risk of progressing to active TB, development of immune reactivation inflammatory syndrome and overall HIV-associated morbidity and mortality.3 7 8 The reduced risk of TB disease and mortality attributable to TB preventive therapy among PLWH is independent of whether an individual is on antiretroviral therapy (ART).9

The two main diagnostic tests for LTBI are the Mantoux tuberculin skin test (TST) and the interferon-gamma release assay (IGRA), the former an intradermal hypersensitivity test and the latter a whole blood immune assay.2 The care cascade for TB testing includes: (1) being successfully tested for TB (ie, receiving a result), (2) initiating TB preventive therapy and (3) completing treatment. Currently, 3 months of weekly isoniazid (INH) plus rifapentine, 4 months of daily rifampin or 3 months of daily INH plus rifampin are standard regimens in the USA. Six or 9 months of daily INH can also be offered, the latter which was standard at the time of this study.1 A meta-analysis on the LTBI cascade found that among individuals intended for screening in high-income countries, 76% received a test result, 65% accepted and started treatment and 23% completed treatment.10 Another meta-analysis of the LTBI cascade among PLWH found treatment completion to be 37.9% in high-income countries.11 Populations at higher risk for LTBI and HIV, including people who are foreign-born, homeless, have a history of incarceration and who use substances,2 are often harder to engage and retain in care. US-based studies report that between 54% and 79% of eligible PLWH receive TB testing which, although greater than the general population, falls short of the 100% goal.12 13 One US-based study found that individuals who reported consuming ≥3 alcoholic drinks daily were half as likely to initiate treatment than those who did not use alcohol.14 Treatment completion rates are also lower than desired among PLWH, people with substance dependence, who are foreign-born and who are homeless.15–24 Increased understanding of the drivers of failure to complete each stage of the LTBI cascade could improve interventions for these priority populations.

In this prospective study of a well-characterised US-based cohort of PLWH with substance dependence (Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV)) history, we aimed to quantify the percentage that completed each stage of the LTBI care cascade and to describe factors associated with failure to complete each stage.


Study setting and population

We conducted a secondary analysis of the prospective Boston Alcohol Research Collaboration on HIV/AIDS (ARCH) cohort, a study of individuals with multiple risk factors for both LTBI and poor completion of the LTBI cascade of care.25 All cohort participants were recruited from the Boston Medical Center infectious disease and HIV clinic or the Boston Healthcare for the Homeless Programme’s HIV primary care clinic between September 2012 and November 2014. All participants were actively receiving HIV care at the time of cohort enrolment, enabling us to identify dropout from the LTBI cascade among those already engaged in care. Eligibility criteria for the cohort have been described previously.25 Additionally for our analysis, participants were excluded if they had past active TB disease or documentation of previous TB preventive therapy, as these individuals would no longer be eligible for repeat testing or treatment. TB testing and treatment were conducted by participants’ routine care providers and were not done as part of ARCH cohort participation. Those who test positive were offered 9 months INH with vitamin B6 and, if they agreed to start treatment, were seen monthly by clinical nurses to confirm medication tolerance. One participant received experimental LTBI treatment as part of a trial. TB testing was either by TST or QuantiFERON TB-Gold In-Tube.

Data collection

Demographic characteristics of study participants were obtained through research assistant administered interviews and medical record review. Alcohol and other drug use were assessed using validated questionnaires by Boston ARCH study personnel. Information on TB testing and treatment and history of active TB was abstracted from the electronic medical record.


The Mini-International Neuropsychiatric Interview V.6.0 was used to assess current alcohol use disorder (past 12 months) as well as current alcohol and drug dependence using DSM-IV criteria. History of IDU, current use of ART, time on ART, country of birth and smoking status were ascertained via self-report. Past 30-day IDU was measured using questions from the Addiction Severity Index (ASI) while current prescription of methadone or buprenorphine (treatment for opioid addiction) was measured using the ASI in conjunction with the medical record. Current alcohol use was measured using a 30-day Timeline Followback assessment with drinking behaviour categorised as abstinent, not heavy (not exceeding the National Institute on Alcohol Abuse and Alcoholism (NIAAA) daily or weekly limits by gender), heavy (exceeding NIAAA daily or weekly limits by gender, but <5 heavy days in past month) or frequent heavy (≥5 heavy days in past month). Lifetime alcohol consumption was ascertained using the lifetime drinking history instrument.

Homelessness was defined as self-reported one or more nights on the street or in a shelter in the past 6 months. History of incarceration was abstracted from the medical record. Diabetes without chronic complications was measured according to the Charlson Comorbidity Index. Being foreign-born was defined as being born outside of the USA. Each participant’s medical record was reviewed at the time of enrolment and, if available, within the past 3 months for HIV viral load. An undetectable viral load was defined as ≤50 copies/mL. If the medical record lacked documentation, HIV viral load was tested at the time of enrolment. Adherence to ART was defined as self-reported adherence of ≥95% of doses in the last month using a Visual Analogue Scale. Viral hepatitis was defined as having either a positive hepatitis C virus (HCV) antibody result, an HCV viral load of greater than zero, or a positive hepatitis B virus antigen test. Variables collected as part of this study, as defined above, that were comparable to US Centers for Disease Control and Prevention (CDC)-defined risk factors for LTBI were summed to create a final cumulative count, and included: being foreign born; ever IDU; alcohol or drug dependence in the past 12 months; HIV; homelessness; ever incarceration; diabetes; and liver disease. All study participants had a minimum of two risk factors (ie, HIV infection and substance dependence).

The LTBI care cascade included the following: (1) completed TB testing in the 10 years prior to enrolment in the Boston ARCH cohort or up to 12 months thereafter; (2) initiation of TB preventive therapy; and (3) treatment completion (figure 1). TST was interpreted as negative if induration was <5 mm and positive if ≥5 mm, per guidelines.26 IGRA was recorded as positive, negative or unsuccessful based on standard cut-off of Nil ≥0.35 IU/mL.27 TB testing completion was defined as receiving a valid TST or IGRA result. Initiation of TB preventive therapy was defined as having a physician visit where TB preventive therapy was prescribed. Information on adherence was not available. Treatment completion was defined as documented completion of a TB preventive therapy regimen in the medical record.

Figure 1

Flowchart of the latent tuberculosis infection (LTBI) cascade of care: (1) completion of tuberculosis (TB) testing; (2) initiation of TB preventive therapy; and (3) completion of TB preventive therapy.

Data analysis

We quantified participants who completed each stage of the cascade, and, when sample size allowed, compared factors associated with completing each stage. We assessed differences in demographics, including number of CDC-defined risk factors, cascade completion and test positivity between participants tested with an IGRA versus TST. Lastly, we looking at differences between participants who tested positive versus negative for LTBI. Measures of central tendency were used to describe the distribution of continuous variables. We employed χ2 and Fisher’s exact tests for categorical variables and t-tests for continuous variables to assess the association between cohort characteristics and completion of each stage of the cascade. Wilcoxon rank-sum tests were used for non-normal data. SAS University Edition was used for all statistical analyses with two-tailed significance defined at p<0.05. We report ORs with 95% CIs.

Patient and public involvement

Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.


Cohort characteristics

Of the 250 Boston ARCH cohort participants, 1 had a previous history of active TB and 30 had already been treated with TB preventive therapy, leaving 219 eligible for TB testing (table 1). More than half (51.6%) reported concurrent alcohol and drug dependence in the past 12 months, with 11.4% reporting IDU and 53% reporting heavy or frequent heavy alcohol drinking in the 30 days prior to assessment. Including HIV-infection and substance use, half (50.7%) had four or more CDC-defined risk factors for LTBI.

Table 1

Characteristics of study participants eligible for TB testing (n=219)

Completion of the cascade of care

Of the 219 participants eligible for this analysis, 203 (92.7%) were tested for LTBI and 189 (86.3%) completed testing (figure 2). Participants tested with IGRA (n=100, 49.3%) were four times more likely to complete testing than those tested with TST (OR=3.87, 95% CI 1.05 to 14.30). Of those who completed testing, more of those tested with TST compared with those tested with IGRA required more than one test (8.8% vs 1.0%, p=0.02). Cumulative number of CDC-defined TB risk factors was not associated with whether a participant was tested for LTBI (table 2), nor was it associated with which test was used (online supplemental additional file 1). Substance dependence was also not associated with lack of TB test completion (OR=1.11, 0.37–3.31). Individuals with alcohol dependence were more likely to be tested with TST (OR=2.15, 1.16–4.00) (online supplemental additional file 1).

Figure 2

Flowchart of study participants who were eligible for TB testing, received a test (IGRA or TST), were successfully tested (ie, received a valid result), tested positive for TB infection, initiated treatment and completed treatment (N=250). ARCH, Alcohol Research Collaboration on HIV; IGRA, interferon-gamma release assay; LTBI, latent TB infection; TB, tuberculosis; TST, tuberculin skin test.

Table 2

Comparison of participants who completed testing* versus those who did not (n=219)

Eleven (5.8%) individuals tested positive for LTBI. Among participants successfully tested (n=189), being foreign-born (OR=3.95; 1.13–13.77) was associated with having a positive test result (table 3). Of the 11 participants who tested positive for LTBI, 5 received a TST and 6 received an IGRA. Of those who received a TST, three had information on their induration ranging from 10 mm to 30 mm while one participant originally tested with IGRA was retested with TST and had a recorded TST induration of 25 mm. Five individuals initiated TB preventive therapy even though two patient records detailed a history of liver disease or the need to abstain from drinking. Four participants initiated the standard 9-month INH regimen and one initiated a 1 month short course of a rifamycin and INH as part of a clinical trial. Three individuals completed treatment. Of those that discontinued treatment, one participant cited abdomen cramping and seizures while the other did not have details other than the provider noting that stopping treatment was safest in light of chronic liver disease.

Table 3

Participants with a positive result on TST/IGRA compared with participants with a negative test result (n=189)


For this cohort of PLWH and substance dependence engaged in clinical care, we found a high proportion were successfully tested for LTBI, higher than previously reported by two studies of HIV-infected populations with lower levels of substance use, which may reflect that our cohort was more engaged in HIV care (ie, all on ART, median CD4 count >500).28 29 Substance dependence was not predictive of whether an individual successfully completed testing. However, our findings show that even with successful engagement in the beginning of the cascade, retention through the treatment phase of the LTBI cascade was poor. All participants should have been tested for LTBI, and all those who tested positive should have initiated and completed TB preventive therapy.1 2 Only a quarter of participants eligible completed the full cascade. No known risk factors nor the cumulative number of risk factors were associated with attrition from TB testing. Trends in LTBI and TB disease in the USA continue to support the prioritisation of the testing and treatment of populations such as our cohort. A review of TB outbreaks in the USA from 2002 to 2011 reported that substance use disorder, homelessness and incarceration were significant drivers in outbreak development.30

Among the study participants, 5.8% of those successfully tested had a positive result. A 2010–2012 analysis of a representative nationwide US database of PLWH showed 6.9% had evidence of LTBI.31 Consistent with both national and state trends and a well-recognised risk factor for LTBI,2 32 foreign-born participants were more likely to test positive. Substance dependence did not increase individual risk for LTBI in our study population.

LTBI diagnostics may have an impact on cascade retention. Participants tested with TST were less likely to be successfully tested, reflecting the challenges associated with this test. TST is susceptible to operator performance variability and requires a second visit to obtain a result.2 33 IGRAs are more specific and require only one visit, but are costly and discordance with TST and higher frequency of positive results raises concerns about interpretation.33 Current guidelines suggest the two tests can be used interchangeably, but that IGRA is preferentially used for individuals unlikely to return for TST reading.2 A previous study found that patients at an urban HIV clinic were 1.5 times more likely to receive care adherent to TB treatment guidelines after the clinic transitioned to using an IGRA.29 Another study found being tested with an IGRA to be significantly associated with treatment completion.34 Findings such as ours highlight the potential benefit of IGRAs for patients at-risk of not returning for TST reading. Interpretation of LTBI test results can be challenging and may ultimately impact likelihood of providers recommending and patients subsequently initiating treatment.

The length of the standard regimen for treating LTBI in PLWH is another barrier to retention. New shorter rifapentine-containing regimens have shown similar efficacy to and higher completion rates than 9 months of INH,19 35 36 providing a treatment alternative for groups difficult to retain in care. However, rifamycin-containing regimens require coordination with HIV medications to avoid drug interactions and recent studies may suggest needs for alternate dosing for PLWH.37 Directly observed therapy, patient incentives and social interventions during treatment, such as education, adherence coaching and peer counselling, have been shown to improve adherence and could also improve completion rates.38 39 The overwhelming summary is that successful interventions depend on the target population, and further research is required to inform strategies for high-risk populations.

Our study has limitations. First, the sample size was too small to assess risk factors along the entire cascade of LTBI care. This limited our ability to draw conclusions about who begins and completes treatment for LTBI. Second, our study is limited in its generalisability: our participants were all engaged in care at the time of enrolment, potentially increasing the likelihood that they received LTBI testing compared with the broader population of PLWH. However, even with our smaller sample size and focus on those engaged in care, we did not find 100% LTBI testing and our low treatment initiation and completion numbers demonstrate a persistent gap.


Our study demonstrates that high testing coverage for LTBI is possible among PLWH with numerous other TB risk factors, including high rates of substance use, incarceration and homelessness. However, we also demonstrate persistent challenges in retaining patients through treatment completion. Our findings reaffirm that high-risk populations, such as PLWH and PWID, can be successfully tested, but reasons for attrition from initiation and completion of preventive therapy need further investigation.

Data availability statement

Data are available upon reasonable request. Data collected for the study are available to interested investigators in the URBAN ARCH Repository:

Ethics statements

Patient consent for publication

Ethics approval

The study protocol was approved by the Boston University Medical Campus’ Institutional Review Board (application #H-31295) and written informed consent was obtained from all individual participants included in the study. The NIAAA provided a Certificate of Confidentiality to further protect participants. All data collection methods were done in conformance with Good Clinical Practice (GCP) standards, including automated validation and quality control checks in the electronic database. Participants gave informed consent to participate in the study before taking part.


The authors would like to thank the Boston ARCH Cohort study participants for their time and contributions. We want to acknowledge Margo Godersky, Kate Haworth, Keshia Toussaint and Laura Vercammen for their work in recruitment, data collection and participant retention, Seville Meli for her oversight of study operations and Carolina Geadas for her assistance in formatting.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • TR and EJR are joint first authors.

  • Twitter @AliciaSVentura

  • Deceased Dr. RS deceased

  • Contributors TR contributed to the study design and methods, aspects of data collection, data analysis, interpretation of results and drafting and editing the manuscript. EJR contributed to the study design and methods, data analysis, interpretation of results and drafting and editing the manuscript. ASV contributed to the study design and methods, aspects of data collection, data analysis, interpretation of results and drafting and editing the manuscript. MRW contributed to the study design and methods, aspects of data collection, data analysis and drafting and editing the manuscript. LW contributed to the study design and methods, data analysis, interpretation of results and drafting and editing the manuscript. CRH contributed to the study design and methods, interpretation of results and drafting and editing the manuscript. JHS contributed to the study design and methods, interpretation of results and drafting and editing the manuscript. RS contributed to the study design and methods, interpretation of results and drafting and editing the manuscript. KRJ contributed to the study design and methods, data analysis, interpretation of results and drafting and editing the manuscript, and is the guarantor of the study. All authors read and approved the final manuscript.

  • Funding RS, JHS, ASV and MRW were funded by the National Institute on Alcohol Abuse and Alcoholism (award numbers U01AA020784, U24AA020779, U24AA020778) and by the National Center for Advancing Translational Sciences (UL1TR001430). KRJ, CRH, JHS and LW were supported by the National Institute of Allergy and Infectious Diseases (NIAID) through the Providence/Boston Center for AIDS Research (P30AI042853) and KRJ, CRH and LW were supported by the NIAID through the Boston University/Rutgers Tuberculosis Research Unit (U19AI111276). KRJ, CRH and EJR were supported by the NIAID (R01AI119037).

  • Competing interests RS reports that Alkermes provides medication for a comparative effectiveness study funded by NIAAA/NIH for which he is the principal investigator. Authors report no other conflicts.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.