Abstract
This retrospective study investigated outcomes among lost to follow-up (LTFU) adolescents and young adults (AYA, ages 10–24) with tuberculosis (TB) registered from 2008–2014 in Gaborone, using surveillance data. Of 68 LTFU AYA, 16 repeated treatment; 8 completed and 6 were again LTFU. Of 4 confirmed deaths, 3 had TB/HIV co-infection. Approaches to improve AYA retention in TB care are needed.
Keywords: adolescent, TB, HIV, retention, HIV testing
Tuberculosis (TB) is a leading cause of death among adolescents and young adults (AYA, ages 10–24) and the predominant cause of death among persons with HIV.1,2 In 2012, 1.78 million AYA developed TB, representing 17% of new TB cases globally.3 Emerging data describe poor outcomes among AYA. In Gaborone, Botswana, AYA are at increased risk for LTFU from TB treatment, compared with adults.4 LTFU is highest among adolescents with TB/HIV co-infection (22.9%).4 In South Africa, highest LTFU rates were among 15–19 and 20–24 year-olds with TB/HIV co-infection: 15% and 16%, respectively.5
Reasons for LTFU among AYA with TB are unknown, but given particular challenges for co-infected AYA, these may overlap with complex barriers to retention for AYA living with HIV.4–6 Disengagement from TB treatment may result in resistant TB or mortality. There is increasing mortality among adolescents with HIV, which is exacerbated by poor retention.6 It is unknown whether LTFU of AYA from TB treatment may include unascertained deaths.
We sought to determine follow-up events (re-registration in TB care, retreatment outcomes, or deaths) for a cohort of AYA LTFU from TB care in Gaborone, a high TB- and HIV-prevalent setting. Our hypothesis was that some AYA had died after LTFU from TB treatment.
MATERIALS AND METHODS
We performed a retrospective cohort study of all AYA (ages 10–24 years at registration) with TB registered at study sites from January 2008 to December 2014 for whom the treatment outcome was classified as LTFU. Cases were identified from paper TB treatment registers at nine public primary care clinics in Gaborone, Botswana. These sites treat 75% of TB patients in Gaborone. Under Botswana National TB Program’s (BNTP) reporting system, standardized registers are used in each clinic. Registers were reviewed only for years 2011–2014 at Lesirane Clinic, and 2012–2014 at Nkoyaphiri Clinic, as earlier registers were not available. Registers at other sites were reviewed for the full study period. Data were abstracted in June 2016 from paper TB treatment registers and entered into a REDCap database.7
Subsequent registrations for LTFU AYA were sought through systematic searches of clinic paper TB registers and the BNTP Electronic TB Register (ETR). Botswana’s ETR was implemented in 1995, in partnership with the US Centers for Disease Control.8 District TB Coordinators enter data from paper registers into the ETR; these generate national notification reports and analyses.8 The ETR includes all TB registrations within BNTP clinics; however, it may miss some patients treated at private clinics or tertiary care sites.9 The ETR was queried using name searches (any order), confirmed by identifiers, such as birthdate, National Identity (Omang) number, or TB treatment registration number. Botswana citizens are required to apply for Omang at age 16; younger Batswana or non-citizens do not have Omang. TB focal persons and staff were asked if they recalled subsequent outcomes for individual AYA who were LTFU from their clinic. For patients who remained with undetermined outcome, the Botswana National Death Register was queried.
Treatment outcomes were as defined by WHO.10 LTFU indicated interruption of treatment for two months or longer. TB clinic staff attempt to re-engage absent patients through multiple calls, and home visits, if able. Deaths included all-cause mortality during treatment, and subsequent deaths identified through the Death Register.
Clinical characteristics and outcomes were analyzed using descriptive statistics and chi-square tests. Time from registration to LTFU date was calculated. Data were analyzed using Stata 13.1 (StataCorp, College Station, Texas, US).
This study was reviewed and approved by the Institutional Review Boards (IRB) of the Botswana Ministry of Health and the University of Pennsylvania; informed consent was waived.
RESULTS
We identified 68 AYA LTFU from TB treatment, representing approximately 8.8% of this age cohort (Table).4 Most (66.2%) were 20–24 years old. The majority (79.4%) had pulmonary TB, and of those with smear results, 88.2% were smear-positive. Of these, 63.3% did not have demonstrated conversion to smear-negative status. Ten (14.7%) of these originating LTFU registrations were retreatment cases. No patients were documented with MDR TB on initial or subsequent registrations.
Table.
Characteristics of youth (ages 10 to 24) who were lost to follow-up (LTFU) from TB treatment in Gaborone from 2008–2014.a,b
| Patient variable | Youth n (%) |
|---|---|
| Total | N = 68 |
| Age | |
| Median, y (IQR) | 21 (18–23) |
| 10 to 14 | 7 (10.3) |
| 15 to 19 | 16 (23.5) |
| 20 to 24 | 45 (66.2) |
| Sex | |
| Female | 32 (47.1) |
| Site of Disease | |
| Pulmonary | 54 (79.4) |
| Extrapulmonary or Both | 14 (20.6) |
| Site of Extrapulmonary Disease | N = 14 |
| Lymph nodes | 6 (42.9) |
| Pleura | 5 (35.7) |
| Bones / Joints | 1 (7.1) |
| Not documented | 2 (14.3) |
| Documented smear results at start of therapy | 34 (50.0) |
| Smear-positive | 30/34 (88.2) |
| Converted to smear-negative on therapy | 11/30 (36.7) |
| Treatment groupc | |
| New patient | 58 (85.3) |
| Previous LTFU | 2 (2.9) |
| Previous failure | 1 (1.5) |
| Relapse | 7 (10.3) |
| DOT settingd | |
| Community-based care | 16 (23.5) |
| Facility-based care | 40 (58.8) |
| Not documented | 12 (17.6) |
| HIV testing | |
| Not documented | 29 (42.7) |
| Documented | 39 (57.4) |
| Positive | 15/39 (38.5) |
| Documented HIV-positive | N = 15 |
| New or recent diagnosis of HIVe | 5 (45.5) |
| ART status | |
| Not documented | 7 (46.7) |
| Documented | 8 (53.3) |
| Initiated on ART | 7/8 (87.5) |
| Initiated on cotrimoxazole | 6/8 (75.0) |
| Time to LTFU | |
| Median, months (IQR) | 4.3 (2.4–5.4) |
| <1 month, n (%) | 5 (7.4) |
| 1 to <2 months | 5 (7.4) |
| 2 to <3 months | 12 (17.6) |
| >= 3 months | 46 (67.6) |
| Year of initiation of treatmentf | |
| 2008 | 15 (22.1) |
| 2009 | 16 (23.5) |
| 2010 | 10 (14.7) |
| 2011 | 7 (10.3) |
| 2012 | 8 (11.8) |
| 2013 | 9 (13.2) |
| 2014 | 3 (4.4) |
Abbreviations: y, years; ART, antiretroviral therapy; DOT, directly-observed therapy; HIV, human immunodeficiency virus; IQR, interquartile range; LTFU, loss to follow-up.
Included clinics: Bontleng, Broadhurst 3, Extension 2, Gaborone West, Lesirane, Mogoditshane, Nkoyaphiri, Old Naledi, and Phase 2.
According to Botswana National TB Programme and WHO definitions.
For patients registered for community-based care, a trained and supervised family member, healthcare worker, or other treatment partner directly observes medication administration and documents adherence. Under facility-based care, this is done by TB clinic staff.
HIV test date within 6 months of start of TB treatment. Among 11 patients with documented test dates.
For the index registration resulting in LTFU.
Many AYA (42.7%) had no documentation of HIV status in the treatment register. Of those with documented status, 38.5% were HIV-positive. Among those with test dates, 5/11 (45.5%) had recent diagnosis of HIV.
LTFU occurred a median of 4.3 months (IQR 2.4–5.4 months) after initiating treatment. Median time from LTFU until study data collection was 72.8 months (IQR 43.7–85.6 months).
Subsequent TB registrations were found for 16 AYA. Re-registration was more common among 15–19 (50.0%) and 10–14 year-olds (28.6%) than among 20–24 year-olds (13.3%, p=0.012); was associated with HIV-positive status (46.7% vs. 16.7%, p=0.043); and was more common among males, though not statistically significant (30.6% vs 15.6%, p=0.15). There was no association with clinic site, disease classification, or smear-positivity. Six re-registered AYA were again LTFU. One patient completed retreatment but relapsed again and died within a month of restarting treatment. Eight others completed treatment. An additional patient was reported by clinic staff to have restarted and completed treatment at a local hospital. In total, 9/68 (13.2%) of the LTFU AYA had subsequent successful treatment outcomes.
Death Register query found that three additional patients had died: at 2.1, 9.9, and 36 months after LTFU. Three of the four (75%) known deaths were among HIV-positive AYA, representing a minimum mortality among confirmed HIV-positive AYA LTFU from TB treatment of 3/15 (20%).
DISCUSSION
In a cohort of AYA LTFU from TB care in Gaborone, we identified later successful treatment outcomes for a minority (13.2%) of patients. This study raises questions as to the reasons for LTFU among AYA TB patients, problems integrating TB/HIV care for AYA, and systems challenges tracking TB patients.
We previously reported that adolescents with TB have double the risk of LTFU compared with adults.4 Here we describe that after multiple systematic inquiries, subsequent successful TB outcomes could only be found for a minority of LTFU AYA. Re-registration and treatment completion does not undo the first problem of disengagement from care and its potential for development of resistant TB and continued transmission in the community.
Reasons for increased LTFU among AYA with TB and approaches to retain them have not been elucidated. AYA may require more intensive follow-up or counseling after TB diagnosis, access to youth-friendly services, and/or more flexible services given their mobility and needs to attend school or work.
Knowledge of HIV status is an important driver of care. In Gaborone the prevalence of undocumented HIV status in TB registrations is 9.2% among ages 10–19, 4.8% for ages 20–24, and 1.8% among adults.4 In this cohort of LTFU AYA, 42.7% did not have documentation of HIV status. All TB patients should be tested for HIV, particularly in Botswana, where most TB cases are HIV-positive. This represents a critical missed opportunity for HIV testing among AYA. Reasons for undocumented HIV status may relate to deferred testing on diagnosis of TB, with the LTFU group ultimately not tested. It is possible, but not clearly known, whether risk factors for LTFU from TB treatment overlap with factors influencing deferred HIV testing, such as experience of intense stigma, family-level factors including orphan status, or barriers related to school or work.6 Within AYA TB care, increased focus is needed on timely HIV testing and linkage to HIV care, including through youth-friendly services.
At least four AYA died after LTFU, including 3/15 (20%) of HIV-positive AYA. Adolescents have been the only age group with increasing HIV-associated mortality in the ART era.6 LTFU complicates HIV care and contributes to mortality. Attention to LTFU from TB treatment could represent an important target to prevent adolescent TB/HIV deaths.
Most LTFU AYA had pulmonary TB, with most sputum smears positive. LTFU smear-positive AYA may be a source of ongoing TB transmission, highlighting the urgent public health need to address AYA TB outcomes. Patients LTFU a median four months from initiation might not be infectious; however most did not have repeat smears, and we cannot exclude that the duration on effective treatment was shorter, if poor adherence or discontinuation of therapy occurred prior to the documented LTFU date.
Difficulties finding subsequent outcomes in the ETR and Death Register may result from discrepant identifiers. If LTFU AYA were re-treated for TB at a tertiary care site or private clinic (estimated <10% of cases), these repeat courses might not be captured in the ETR.9 It is possible more LTFU AYA may have died but were either not matched or not captured in the Death Register. It may be that no subsequent outcomes were found for some LTFU AYA because they either fully recovered or remained chronically sick without re-registering in TB treatment. We were not able to incorporate calls or home visits to LTFU patients, which may have yielded outcomes for more AYA.
In conclusion, LTFU from TB treatment is a significant challenge among AYA, resulting in poor clinical outcomes, mortality and likely on-going TB transmission. In this study, re-registration and successful TB outcomes were only found for a minority of patients. Decreased HIV testing among AYA TB cases represents a critical missed opportunity. Further research is needed to investigate factors contributing to LTFU among AYA and integrated approaches to AYA TB and HIV diagnosis and care.
Acknowledgements
We thank the Botswana National Tuberculosis Programme for their support of this work, and the healthcare workers for their participation in this study and for their work to fight TB. We acknowledge the patients around the world who are affected by TB.
This work is dedicated to the memory of Dr. Melissa Ketunuti. As a compassionate pediatrician and researcher she worked to improve health for vulnerable children in Botswana and around the world, and she continues to inspire our work.
Dr. Laura Smallcomb contributed to the data entry for this cohort.
Funding: The authors have no conflicts of interest. The project was made possible through core services & support from the Penn Center for AIDS Research (CFAR), an NIH-funded program (P30 AI 045008). Drs. Arscott-Mills, Lowenthal, and Steenhoff are supported by the CHOP Global Health Center. Ms. Eby was supported by the Benjamin H. Keane Travel Fellowship in Tropical Medicine, of the American Society of Tropical Medicine and Hygiene. Dr. Enane is supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development of the National Institutes of Health (K23HD095778) and by Thrasher Research Fund. The content of this manuscript is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
References
- 1.Mokdad AH, Forouzanfar MH, Daoud F, et al. Global burden of diseases, injuries, and risk factors for young people’s health during 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. The Lancet. 2016;387(10036):2383–2401. doi: 10.1016/s0140-6736(16)00648-6. [DOI] [PubMed] [Google Scholar]
- 2.Gupta RK, Lucas SB, Fielding KL, Lawn SD. Prevalence of tuberculosis in post-mortem studies of HIV-infected adults and children in resource-limited settings. AIDS. 2015;29(15):1987–2002. doi: 10.1097/QAD.0000000000000802. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Snow KJ, Sismanidis C, Denholm J, Sawyer SM, Graham SM. The incidence of tuberculosis among adolescents and young adults: a global estimate. Eur Respir J. 2018;51(2):1702352. doi: 10.1183/13993003.02352-2017. [DOI] [PubMed] [Google Scholar]
- 4.Enane LA, Lowenthal ED, Arscott-Mills T, et al. Loss to follow-up among adolescents with tuberculosis in Gaborone, Botswana. Int J Tuberc Lung Dis. 2016;20(10):1320–1325. doi: 10.5588/ijtld.16.0060. [DOI] [PubMed] [Google Scholar]
- 5.Snow K, Hesseling AC, Naidoo P, Graham SM, Denholm J, Preez Du K. Tuberculosis in adolescents and young adults: epidemiology and treatment outcomes in the Western Cape. Int J Tuberc Lung Dis. 2017;21(6):651–657. doi: 10.5588/ijtld.16.0866. [DOI] [PubMed] [Google Scholar]
- 6.Enane LA, Davies M-A, Leroy V, et al. Traversing the cascade: urgent research priorities for implementing the “treat all” strategy for children and adolescents living with HIV in sub-Saharan Africa. J Virus Erad. 2018;4(Suppl 2):40–46. [PMC free article] [PubMed] [Google Scholar]
- 7.Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)—A metadata-driven methodology and workflow process for providing translational research informatics support. Journal of Biomedical Informatics. 2009;42(2):377–381. doi: 10.1016/j.jbi.2008.08.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Alpers L, Chrouser K, Halabi S, et al. Validation of the surveillance system for tuberculosis in Botswana. Int J Tuberc Lung Dis. 2000;4(8):737–743. [PubMed] [Google Scholar]
- 9.Arscott-Mills T, Ejelonu A, Mokalane K, et al. Over a third of child tuberculosis patients diagnosed in a tertiary centre in Botswana are missing from the national Electronic Tuberculosis Register. In: The Hague, Netherlands; 2018. https://thehague.worldlunghealth.org/programme/abstract-book/. [Google Scholar]
- 10.World Health Organization. Guidance for National Tuberculosis Programmes on the Management of Tuberculosis in Children. Second edition Geneva, Switzerland: World Health Organization; 2014:1–146. [PubMed] [Google Scholar]