The World Health Organization (WHO) declared tuberculosis (TB) a global public health emergency since 1993, and in 2018, in Brazil 82,409 cases were reported (new and relapses) which corresponds to an incidence rate of 45 cases per 100,000 inhabitants.1 The state of Rio de Janeiro, has the second highest incidence rate in Brazil with an incidence rate of 66.3/100,000 inhabitants, which highlights the importance of actions for TB control and strategies to interrupt the transmission chain.2

Contact tracing of patients with pulmonary TB (PTB) is one of the most important strategies for interrupting the transmission and subsequent development of TB.3 A systematic review found that the prevalence of TB in contacts is 1.4% for high income countries and 3.1% for low/middle income countries.4 In addition, the screening of contacts helps identify contacts with latent Mycobacterium tuberculosis infection (LTBI), through a positive tuberculin skin test (TST) or by interferon gamma release assays (IGRA). Subjects with LTBI have a higher risk of developing TB disease and benefit from the LTBI Treatment (LTBIT). The use of isoniazid (INH) for 6–9 months is recommended for LTBIT with a 60–90% protective effect against active TB.5

The protective effect of INH for LTBIT varies depending on differences in treatment duration and adherence to the scheme.6 Adherence to LTBIT is strongly influenced by clinical, social, financial, and behavioral factors. Studies evaluating factors associated with low adherence to LTBIT under routine conditions are scarce. In high income countries, the proportions of completion of LTBIT vary from 50% to 78%,7–9 and in low/middle income countries, from 54% to 78%.10–12

According to the National TB program (NTP) guidelines, the Hospital TB Control Program in our hospital recommends treatment for LTBI in close contact with PTB patients once active TB is excluded.6,7 In this context, this study aimed to analyze the factors associated with non-completion of LTBIT in contacts with patients with PTB treated at our outpatient clinic.

According to inclusion and exclusion criteria, in the first period and in the second period, 11.6% (174/1495) and 46.1% (1464/3173) had LTBIT recommended, respectively. Among 174 contacts with LTBIT indication in the first period, 72 (41.4%) started the treatment. In the second period, among 1464 contacts with LTBIT indication, the following were excluded: 33 (2.3%) due to transfer to other health units at the physician's discretion or due to request by the contact and 998 (68.1%) were referred to a clinical trial linked to TB Trials Consortium, but among them, 250 (17.1%) came back to the Hospital Tuberculosis Program. After the exclusions, 683 remaining contacts with LTBIT indication, 438 (64.2%) started the treatment. By medical decision, another 25 contacts with LTBI in the first period and 1 contact in the second period started on LTBIT although they did not have criteria for this according to NTP, totaling 536 contacts (Fig. 1).

Figure 1.

Tuberculosis contacts description from 1999 to 2009. * 781 were excluded. After the exclusions, 683 remaining contacts with LTBIT indication. ILTBIT: latent tuberculosis infection treatment.

(0.15MB).

We enrolled 114 cases and 404 controls in the study; for 18 contacts we were unable to retrieve information from Hospital TB Control Program forms. The proportion of contacts that did not complete LTBIT increased from 10.3% (10/97) in the first period to 23.7% (104/439) in the second period (Table 1). In the first period, among cases and controls, females and household contacts were, respectively, 80% (8/10) and 56% (44/79), and 80% (8/10) and 81% (64/79). Median age was 12 years among cases and 10 years among controls. Ninety percent (9/10) of cases and 78% (62/79) of controls had 0 to 8 years of schooling.

In the analysis of clinical and epidemiological factors associated with non-completion of LTBIT we found that individuals who use illicit drugs were more likely to treatment default (OR 23.33, 95% CI 1.83–296.1, p=0.05), which maintained after fitting the data through multivariate analysis (adjusted OR 0.04, 95% CI 0.00–0.54). However, the small number of drug users among contacts (only 3) produced an imprecise estimate with a large confidence interval (Table 2).

In the second period, females were 56% (58/104) of cases vs. 57% (187/325) among controls and household contacts were %, and 56% (58/104) and 60% (196/325). Median age was 13 and 17 years among cases and controls, respectively. Years of school were ≤8 for 71% (74/104) of cases and 67% (318/325) of controls. Using bivariate analysis, the following clinical and epidemiological factors were associated with non-completion of treatment: co-morbidities (OR 2.09, 95% CI 1.13-3.89, p=0.01), use of other medications (OR 3.54, 95% CI 1.48-8.48, p=0.002), absence of drug intolerance (OR 2.46, 95% CI 1.39-4.35, p=0.001) and TB treatment default by IC (OR 16.97, 95% CI 3.63-79.24, p=0.00005). However, after fitting the data through multivariate analysis, only the absence of drug intolerance (adjusted OR 2.46, 95% CI 1.39-4.35) and TB treatment default by IC (OR 9.37 95% CI 2.24-39.13) were independently associated with non-completion of LTBIT (Table 3).

Two years after LTBIT recommendation, active TB occurred in 1.0% (1/102) and 0.4% (1/245) of contacts who did not start treatment in the first and the second period, respectively. In the first period, TB disease was not observed among contacts who completed and did not complete the LTBIT. In the second period, no active TB occurred among contacts who completed the LTBIT, but active TB did occur in 1.9% (2/104) of contacts who started but did not complete the LTBIT.

After LTBIT indication in both periods, active TB occurred in 0.6% (2/347) of the contacts who were not treated, there was no active TB among contacts who completed the LTBIT and active TB occurred in 1.8% (2/114) of contacts with incomplete LTBIT (Table 4).

In the first period, 112 LTBIT were required to prevent one active TB case. In the second period, 117 LTBIT were required to prevent one active TB case. In both periods, 116 LTBIT were required to prevent one active TB case (Table 4).

Early screening and treatment of LTBI in contacts of pulmonary TB cases are among the priorities launched by the End TB strategy.15 And studies that evaluate the factors associated with completion of LTBIT in high TB burden countries become even more relevant.16,17

The proportion of completed LTBIT is variable between authors, Sharma et al. evidenced in a clinical trials conducted in research centers, this value is high (˃90%),18 but in a recent meta-analysis, Alsdurf et al. after evaluating 70 distinct cohorts pointed out several gaps in the diagnostic and treatment cascades; the general estimate is around 50% of people with medical indications completed LTBIT.19

However, under routine conditions, LTBIT completion rates range from 50% to 78%.8–12 LTBIT, for both children and adults contacts of active pulmonary TB patients, has been recommended by PNCT since 2010.20 However, in the IDT-HUCFF hospital complex of UFRJ, LTBIT became a priority target in 1999, after Hospital TB Control Program implementation. Therefore, it was possible to evaluate the impact of LTBIT under field conditions in a high burden urban area. Non-completion LTBIT rates (10.3% in the first period and 23.7% in the second period) were lower than those described in other series.

Evaluating the indication of LTBI treatment, we observed that there was a four times increase in the number of contacts with treatment indication in the second study period. This increase can be explained by the difference of the LTBI criteria used in each study period. The criteria changes included the treatment of adult contacts and TST positivity cut off decrease from 10mm to 5mm since April 2003. Using this TST positivity cutoff, the Hospital TB Program identified a higher number of infected contacts, which could be subjected to LTBIT.

We observed a significant difference in the proportion of contacts who started LTBIT between the first and second period. The increasing number of contacts who started LTBIT in the second study period could be attributed to the beginning of a clinical trial linked to TB Trials Consortium, in which contacts received incentives and directly observed treatment. These incentives may have had a positive influence on adherence to treatment onset in TB Program. Nevertheless, a majority of contacts did not start LTBIT in either cohort, and the main reason was the refusal of treatment.

The low percentages of non-completion of LTBIT observed in our study were similar to those reported by Codecasa et al., in Italy, and by Mendonça, in Rio de Janeiro and lower than the rates reported in the USA, Canada and in the state of Bahia, Brazil.8,10–12 In Italy, where the study was performed under routine conditions, involving adults from risk groups for active TB, only 21.6% of contacts did not complete LTBIT.9 Our study also found illicit drug use to be an important risk factor for non-completion of LTBIT, and our data are similar to a study conducted in Portugal that highlights the abuse of illicit drugs as well as alcohol for increased risk failure to complete LTBIT.21 Mendonça observed a non-completion rate of 25% among subjects aged under 15 years of age who were treated in a reference unit from 2002 to 2009.12 In large series in the USA and Canada, higher rates of non-completion of LTBIT were reported (40–52%).9,22 In Salvador, Brazil, 46.5% of children and adults contacts did not complete LTBIT.10

Several studies have reported risk factors associated with non-completion of LTBIT. In the USA and Canada, Pettit et al. observed that among adults who initiated INH, female sex and alcohol use were independently associated with LTBIT discontinuation due to drug intolerance.23 In another study, Horsburgh et al. identified the following as risk factors for non-completion of LTBIT: nine months treatment with INH, living in a nursing home, shelter, or prison; illicit drug use, age ≥15 years, and being a health care professional.9 Our study also found illicit drug use to be an important risk factor for non-completion of LTBIT, despite the small sample analyzed. Machado et al. identified the presence of drug intolerance and the need to take two buses to reach the hospital as risk factors for non-completion of LTBIT.10

In contrast to the observations of Machado et al., we found that the absence of drug intolerance was a risk factor for a non-completion LTBIT.10 A possible reason for this finding could be the differentiated care offered by the Hospital TB Program's multidisciplinary team in the presence of drug intolerance, positively influencing the adherence to LTBIT and encouraging contacts not to abandon the treatment. However, this association may be biased as drug intolerance prevalence was low and all cases fulfilled the minor criteria, which do not usually lead to non-completion of LTBIT. Maciel et al., when analyzing different risk groups who had undergone LTBIT, observed the following factors to be associated with LTBIT default: being a health care professional, HIV positive, and a contact of a TB patient. Being a contact of TB patient increased the odds of non-completion of LTBIT by 2.65-fold and, possibly, the concentration of family efforts around the care of patient with TB disease leads to a lower priority given to LTBI.11

In our study, non-completion of TB treatment by the IC was associated with non-completion of LTBIT in the second study period. In this period, the contacts who underwent LTBIT were referred from other Health Units where IC were treated. Default of treatment by the IC possibly had a negative effect on the priority given by their contacts to LTBIT.

In neither period, two years after enrollment in the study, were there any cases of TB disease among contacts who completed LTBIT. On the other hand, TB disease was found in 1.8% of contacts who did not start LTBIT and 0.6% of those who started but did not complete it. In New York city, Anger et al. reported similar results, with higher TB disease rates among contacts who did not start LTBIT (1.5%) than among contacts who started LTBIT (0.4%).24 In Brazil, in other series involving contacts with LTBIT indication, slightly higher rates of TB disease were described, ranging from 2.3% to 3.2%.14,25

The low rates of TB disease found in our study may be due to reasons other than the efficacy of INH therapy itself. Among these, we can cite: (a) hospital setting in a metropolitan area from middle income country, (b) contacts living in a urban area have lower risk for TB development (2.1–2.8%), as cited by Blok26, (c) shorter follow-up period (2 years after the LTBIT recommendation), compared to the other series that followed-up contacts for 5 years.14,25

Overall, the NNT was similar in both periods. Therefore, it was not possible to state in which of the two cohorts the recommendation for LTBIT was more appropriate. Anger et al. observed results slight lower than to the ones found in our study, with 88 LTBITs required to prevent one TB case.24

The main limitations of the present study include the differences in the screening cascade of contacts in the two study periods, the limited sample size (particularly in the first study period) that could have precluded the emergence of other significant associations, the use of secondary data to diagnose TB disease among contacts, and no qualitative or economic analysis carried out to evaluate the barriers for implementation and impact of LTBIT.

When analyzing the variables associated with non-completion of LTBIT, we found that illicit drug use and non-completion of TB treatment by the IC were the main risk factors. TB disease rates among contacts who did not start LTBIT or who started and did not complete it were lower than described in the literature and treatment completion of LTBIT may be a protective factor against the development of active TB. The high NNT observed in both periods among contacts who attended a university hospital from urban area suggests that LTBIT may have lower relevance than expected in preventing a TB case.

The authors have no conflicts of interest to declare.