Yield of repeat tuberculin skin testing for people living with HIV in Brazil

Abstract

Objectives:

In Brazil, annual tuberculin skin tests (TST) are recommended for people living with HIV (PLWH) with CD4>350, with tuberculosis preventive therapy (TPT) provided upon test conversion. We aimed to determine the yield of repeat TST for PLWH.

Design:

Secondary analysis of the stepped-wedge, cluster-randomized THRio trial for isoniazid preventive therapy (IPT) to prevent tuberculosis (TB).

Methods:

We analyzed data from newly registered PLWH with negative baseline TST results. We calculated the number of TST conversions after 1 and/or 2 years among patients eligible for follow-up TSTs; the proportion of converters initiating IPT; and incidence of TB/death.

Results:

Among 1,770 PLWH with a negative baseline TST, 679 (38%) were female and median age was 36 years (IQR 29–43). Eighty-six (5%) developed TB or died within 1 year. Among 1,684 eligible for a follow-up 1-year TST, 582 (35%) were tested and 53 (9%) were positive. Forty-nine (92%) converters started IPT. Of 529 patients with a negative 1-year TST, 7 (1%) developed TB or died over the following year. Of 522 patients eligible for a 2-year TST, 158 (30%) were tested and 13 (8%) were positive. Ten (77%) converters started IPT. Of 1,102 patients who did not receive a 1-year TST, 33 (3%) developed TB or died. Of the 1,069 patients eligible for a 2-year TST, 259 (24%) were tested and 34 (13%) were positive. Thirty (88%) converters started IPT.

Conclusions:

In this cohort of PLWH in Brazil, TST conversion was high among those re-tested, but only 48% received a follow-up TST within 2 years.

INTRODUCTION

Brazilian guidelines recommend tuberculosis (TB) preventive therapy (TPT) for all people living with HIV (PLWH) with CD4≤350, and those with CD4>350 who have a positive tuberculin skin test (TST) or interferon gamma release assay (IGRA).¹ PLWH with CD4>350 and a negative baseline TST are recommended to receive annual TSTs, with TPT provided upon test conversion. However, the effectiveness of annual latent TB infection (LTBI) testing to guide TPT for PLWH is unclear. We aimed to determine the yield of repeat TST for PLWH in this moderate TB/HIV burden setting.

METHODS

Study design

We performed an individual patient analysis of participants in the TB/HIV in Rio (THRio) study, a cluster-randomized trial that evaluated an intervention to increase TST and isoniazid preventive therapy (IPT) for PLWH in Rio de Janeiro, Brazil.^2–5 We included adults (≥18 years) newly registered at the 29 participating HIV clinics with a TST from September 1, 2005-August 31, 2008. We excluded patients already registered at participating clinics at the start of THRio, diagnosed with active TB within 30 days of registration, or whose first TST was positive. The THRio data collection period was from September 1, 2005-August 31, 2010; additional TB diagnoses and deaths were obtained through linkage with the Rio TB and mortality registries through October 31, 2012.

TB evaluation

At the time of the study, TPT was recommended for all PLWH with a positive TST and active TB ruled-out. The THRio intervention included training for physicians and nurses to perform comprehensive TB evaluation based on national guidelines at the time of the study, including TST for all patients who had not previously been diagnosed with TB, prescribed IPT, or had a positive TST. TST was performed with purified protein derivative RT23 (Statens Serum Institut, Copenhagen, Denmark), with results read within 2–4 days. Those with a positive TST (≥ 5mm) and active TB ruled-out (based on clinical history and chest radiography) were prescribed 6-months of daily isoniazid 300mg with pyridoxine 25mg. Physicians and nurses were trained to repeat TST annually for those with negative results. Active TB was diagnosed according to national guidelines (≥1 positive culture for Mycobacterium tuberculosis, positive acid-fast bacilli smear, or clinical and radiographic presentation consistent with TB and response to treatment).⁶ Antiretroviral therapy (ART) was recommended for patients with CD4<200 from 2005–2007, <350 from 2008–2009, and <500 starting in 2010.

Analysis

We compared characteristics of patients who were eligible for and received a baseline TST with those who were eligible for but did not receive any TST. Among those with a baseline TST, we determined the number eligible for a 1-year follow-up TST (alive and not diagnosed with active TB within 18 months of baseline TST) and 2-year follow-up TST (alive, not diagnosed with active TB within 30 months of baseline TST, with a negative or missing 1-year TST); and calculated the number of eligible patients who received 1-year (7–18 months) and 2-year (19–30 months) follow-up TSTs (these ranges were selected to ensure follow-up TSTs were counted for patients who did not complete a clinic visit at exactly 12 or 24 months).

To determine the yield of repeat TST, we calculated the 1) number of TST conversions after 1 year among patients who were eligible for and received a 1-year TST, 2) number of conversions after 2 years among patients who were eligible for and received a 2-year TST after a negative 1-year TST, and 3) number of conversions after 2 years among patients who were eligible for and received a 2-year TST who did not receive a 1-year TST. We calculated the proportion receiving ART and the proportion of converters who received IPT, and the incidence of TB or death over each year of follow-up. We stratified analyses by CD4 count (measured within 90 days of TST).

THRio was approved by the Johns Hopkins Medical Institutions and Municipal Health Secretariat of Rio de Janeiro institutional review boards.

RESULTS

Among 4,607 PLWH entering care, 1,617 (35%) never had a TST. Those without a TST were younger (median 34 vs. 36 years, p<0.001), less likely to be on ART (46% vs. 54%, p<0.001), and had lower baseline CD4s (median 321 vs. 346 cells/μL, p=0.02) than those with ≥1 TST. Baseline TSTs were performed for 2,141 (46%) individuals before August 31, 2008, of whom 371 (17%) were positive. Of the 1,770 with a negative first test, 679 (38%) were female and median age was 36 years (IQR 29–43). At baseline TST, 775 (44%) patients were on ART and median CD4 was 339 cells/μL (IQR 181–515).

1-year TST

Of the 1,770 PLWH with a negative first TST, 30 (2%) developed TB and 56 (3%) died (of whom 5 had TB) and were ineligible for repeat testing (Figure); of these, 71 (83%) received ART. Among 1,684 patients eligible for a 1-year TST, 582 (35%) were tested and 53 (9%) were positive.

Figure.

Yield of repeat TST among PLWH in Rio de Janeiro, Brazil

Abbreviations: TST, tuberculin skin test; PLWH, people living with HIV; TB, tuberculosis; IPT, isoniazid preventive therapy

Among 529 non-converters, CD4 at 1-year follow-up was ≤350 for 179 (34%), >350 for 283 (54%), and unknown for 67 (13%); among 53 converters, CD4 was ≤350 for 16 (30%), >350 for 33 (62%), and unknown for 4 (8%, Table). Forty-nine (92%) converters started IPT in a median of 30 days (IQR 7–71). Two (4%) converters developed TB and 1 (2%) died over the following year: of these, all had CD4>350, 1 (33%) received IPT, and none received ART.

Table.

CD4 count and TST conversion among PLWH in Rio de Janeiro, Brazil

	Received TST	CD4≤350			CD4>350			CD4 unknown
		Total	TST+	TST−	Total	TST+	TST−	Total	TST+	TST−
Baseline	1770	804	—	804 (100%)	750	—	750 (100%)	216	—	216 (100%)
1-year	582	195	16 (8%)	179 (92%)	316	33 (10%)	283 (90%)	71	4 (6%)	67 (94%)
2-year	417	108	9 (8%)	99 (92%)	247	29 (12%)	218 (88%)	62	9 (15%)	53 (85%)

Abbreviations: TST, tuberculin skin test; PLWH, people living with HIV

Percentages reflect the percent of participants testing TST positive or negative, stratified by follow-up year and CD4 count (row percentages); Persons with a missing TST not included; CD4 count measured within 90 days of TST

2-year TST

Of 529 patients with a negative 1-year TST, 1 (0.2%) developed TB and 6 (1%) died (of whom 1 had TB) and were ineligible for additional repeat testing; of these, 6 (86%) received ART. Of the 522 patients eligible for a 2-year TST, 158 (30%) were tested and 13 (8%) were positive. Among 145 non-converters, CD4 at 2-year follow-up was ≤350 for 39 (27%), >350 for 89 (61%), and unknown for 17 (12%). One (1%) non-converter with CD4≤350 who received ART died over the following year. Among 13 converters, CD4 was ≤350 for 2 (15%), >350 for 7 (54%), and unknown for 4 (31%). Ten (77%) converters started IPT in a median of 45 days (IQR 7–71). No converters developed TB or died over the following year. Among 364 who did not receive a 2-year TST, none developed TB or died.

Of 1,102 patients who did not receive a 1-year TST, 13 (1%) developed TB and 20 (2%) died (of whom 1 had TB) within 30 months of baseline; of these, 28 (85%) received ART. Of the 1,069 patients eligible for a 2-year TST, 259 (24%) were tested and 34 (13%) were positive. Among 225 non-converters, CD4 at 2-year follow-up was ≤350 for 60 (27%), >350 for 129 (57%), and unknown for 36 (16%); among 34 converters, CD4 was ≤350 for 7 (21%), >350 for 22 (65%), and unknown for 5 (15%). Thirty (88%) converters started IPT in a median of 20 days (IQR 7–44). No converters developed TB or died over the following year. Three (1%) non-converters died and 1 developed TB: 1 (25%) had CD4>350 and 3 (75%) received ART. Among the 810 who did not receive a 2-year TST, 19 (2%) developed TB and 11 (1%) died (of whom 4 had TB); of these, 20 (67%) received ART.

DISCUSSION

We found that initial and repeat testing for TB infection in this cohort of patients receiving HIV care in Brazil was less than desirable. Only 13% of those with a negative baseline TST received follow-up testing according to guidelines at the time of the study, and only 48% received any repeat testing within 2 years. Of those who did undergo repeat TST, 10% converted to positive and were eligible for TPT, including 11% of those with CD4>350. Our findings suggest that improving adherence to routine LTBI evaluation will identify a large number of PLWH with TB infection who are at high risk of progressing to active TB.

TPT is highly effective for PLWH, but implementation remains poor globally.^4,5,7–12 The THRio study demonstrated that a clinic-based training program was successful in increasing the proportion of patients who received an initial TST and subsequent TPT, but less than half of those eligible for retesting received a follow-up test to reassess TPT eligibility. To scale-up TPT for those at highest TB risk, in 2018 Brazil amended national guidelines to recommend TPT for all PLWH with CD4≤350, with no requirement for LTBI testing.¹ However, a positive TST/IGRA is still required to treat those with CD4>350; annual repeat TST is recommended for TST-negatives to re-evaluate TPT eligibility. Here, we found that TST conversion was high among those re-tested, though adherence to testing guidelines was poor: Among those with CD4>350 or with unknown CD4s (who would require a positive TST or IGRA to initiate TPT under current national guidelines), 11% converted to positive.

Many factors contribute to poor TST uptake, including global tuberculin shortages and logistical and economic challenges associated with testing. Here, we found that 35% of patients never had a TST. Patients without TSTs have been shown to be extremely high-risk, with 40% of THRio TB cases occurring among those not tested.¹³ Moreover, our results highlight the risk of TB for PLWH with negative TSTs: 2% of patients developed active TB within 1 year of a negative baseline TST, before they were eligible for a repeat test. PLWH with high CD4s and negative TST/IGRAs are at markedly elevated risk of TB compared with persons without HIV infection: previous THRio analyses showed that patients with CD4>350 and negative TSTs had a TB incidence rate 15-fold higher than HIV-negative persons in Brazil.¹³ Finally, loss to follow-up of patients receiving HIV care may contribute to sub-optimal repeat TST uptake. Importantly, we found that approximately 75% of patients had at least one interaction with the healthcare system after a missed follow-up TST, confirming that both provider- and patient-level factors contribute to poor TST uptake, and highlighting the importance of retaining HIV patients in care.

The THRio study was conducted from 2005–2010, potentially limiting our ability to make inferences in the current era of expanded TPT eligibility and universal provision of ART for all PLWH. Importantly, we have previously found that ART uptake for PLWH remains suboptimal in Rio de Janeiro, despite universal ART guidelines.¹⁴ In addition, ART is insufficient to eliminate TB risk in PLWH, and identifying strategies to improve TPT uptake remains essential for TB control. An additional limitation of our study is potential differential follow-up testing of those at lower TB risk, if PLWH at higher TB risk were more likely to be lost to follow-up. We obtained outcome data on active TB diagnoses and deaths through linkage with TB and mortality registries, and our data indeed suggest that those who did not receive any follow-up TST were at higher risk of TB and death (4% of PLWH who did not receive a follow-up TST developed TB or died). Ensuring that PLWH remain engaged in care is essential to reduce morbidity and mortality.

Novel strategies to routinize IGRA may contribute to increased LTBI diagnosis and TPT initiation. The Brazilian National Committee for Health Technology Incorporation has recommended that QuantiFERON-TB Gold Plus be made available for PLWH,¹⁵ and a strategy to link this test to other routine blood draws is under evaluation.¹⁶ Alternatively, given the proven benefits of TPT for PLWH with negative TST/IGRAs,^7,9 guidelines expanding TPT eligibility or eliminating the requirement for TST/IGRA could be considered in Brazil.