INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Immigrants from countries where tuberculosis is endemic account for a significant proportion of cases of tuberculosis in industrialized countries (1). This proportion continues to increase and is a major cause of the overall resurgence of tuberculosis in Western Europe and the United States (1).

Factors influencing the risk of tuberculosis among immigrants include the prevalence of tuberculosis in the country of origin, the duration of residence in the host country, and the effectiveness and quality of screening processes (5). Most cases of tuberculosis result from reactivation of latent infection (6).

The efficacy and tolerability of isoniazid-preventive therapy (IPT) in subjects with latent tuberculosis infection have been demonstrated (7), but poor adherence to treatment limits the overall efficacy to 30 to 70% (8). Data on the rate of completion of preventive therapy among immigrants in general are scarce (3), and virtually unavailable for illegal immigrants. Cultural and linguistic barriers and social outcasting are likely to reduce adherence. Although directly observed therapy for patients with tuberculosis requires more human and financial resources, it significantly increases the rates of completion of treatment (9) and reduces the frequency of primary and acquired drug resistance and relapse (10). Directly observed delivery of preventive therapy (DOPT) has also been proposed, but the effect of direct supervision on completion rates has not been elucidated.

We have conducted a comparative prospective study to assess adherence to one supervised, medical service-based, twice-weekly regimen of isoniazid in illegal migrants in Northern Italy.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Design

A prospective, randomized, unblinded design was used. One health care unit for immigrants in Brescia and one tuberculosis clinic in Turin (Northern Italy) served as recruitment sites. The site in Brescia provides first-level medical services, whereas the one in Turin serves as a tuberculosis screening site for contacts and people applying to enter dormitories. The study population comprised illegal immigrants, defined as foreign-born persons with no residence permit and limited access (emergency interventions only) to public medical care services. As part of the protocol, participants were informed of the nature and aims of the study. Subjects who were eligible for screening were requested to provide written consent before enrollment. Written consent was also required to randomize subjects eligible for preventive therapy. Failure to provide written consent led to exclusion from randomization. Subjects were eligible for the preventive therapy trial if they came from countries with an estimated tuberculosis incidence of 50/100,000 or more, history of immigration of less than 5 yr, and development of a skin induration  10 mm 72 h after intradermal injection of 5 international units of PPD (Biocine Sclavo) (11). Exclusion criteria included pregnancy, age older than 35 yr, and liver enzymes (AST, ALT) five times or more than the upper normal values. Prior to randomization, active tuberculosis was ruled out by means of standard chest radiography and a thorough medical examination.

Treatment Regimens

Eligible subjects were randomly allocated to receive supervised IPT at a dose of 900 mg twice weekly for 6 mo (Regimen A), unsupervised IPT 900 mg twice weekly for 6 mo (Regimen B), or an unsupervised IPT regimen of 300 mg daily for 6 mo (Regimen C) as a reference standard (12). The randomization list was prepared in order to give to each subject an equal probability to be assigned to one of the three arms. Identical randomization lists were used in the two sites. Subjects receiving supervised IPT were invited to report twice weekly to the clinical service sites (either the tuberculosis clinic or the clinic for migrants) to collect the drugs. All patients were provided with free drugs, but no incentives or enablers were used.

Tolerability and Adherence

Adverse effects were checked monthly (Regimens B and C) or twice weekly (Regimen A) during treatment. A complete blood count and serum transaminases were checked at 1, 2, 4, and 6 mo. The investigators had to determine whether each possible side effect might be drug-related. To monitor adherence, regular reporting to the clinical sites was recorded, and the number of returned pills was counted. For subjects in Regimens B and C, isoniazid intake was checked by measuring drug metabolites in urine with the dipstick method (Mycodyn uritek; DynaGen, Inc., Cambridge, MA) at each visit. Treatment was considered to be completed if an estimated 80% or more of the prescribed medications was taken.

Statistical Analysis

The original study design called for 411 evaluable subjects to show a 15% difference in adherence between arms A and C. However, the trial was terminated early because of a larger than expected difference in adherence within the treatment arms.

Data were entered in Access for Windows database computer software. Statistical analysis was performed with Epi Info (Centers for Disease Control and Prevention) and SPSS for Windows software (SPSS Inc., Chicago, IL). Continuous data were compared using Student's t test; categorical data were analyzed by means of the chi-square test with Mantel-Haenszel stratified analysis. Probabilities of completing treatment were compared by means of the Kaplan-Meier plots and the Mantel-Haenszel Log-rank test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Between April 1996 and October 1997, when the study was terminated, 208 patients were randomized, using the same randomization list at both study sites. For these patients, the randomization list selected more patients to receive Regimens A and B than Regimen C, resulting in the assignment of 82 patients to Regimen A, 73 patients to Regimen B, and 53 patients to Regimen C. The demographic and baseline characteristics of the subjects in the three regimens are shown in Table 1. The three groups were similar for all parameters. The proportions of subjects completing the 26-wk regimen were 7.3, 26, and 41% in arms A, B, and C, respectively. In a logistic regression analysis model, adherence was not associated with study site, patient's sex or age, country of origin, alcohol/drug use, marital status, employment status, or religion (data not shown). Of the 156 patients who did not complete treatment, 127 were lost to follow-up, 21 decided to stop treatment, six moved away from the study areas, one became pregnant, one was imprisoned, and five (2.4%) developed adverse events. In the last group, two subjects developed skin rash, one increased ALT values, one gastrointestinal symptoms, and one presented with seizures (the patient, belonging to arm B, denied having taken an isoniazid overdose). The rate of adverse events was similar in the three treatment arms. It was impossible to determine the reasons for dropout in subjects lost to follow-up as interviews could not be carried out. The Kaplan-Meier curves of the number of subjects continuing treatment throughout the study period are shown in Figure 1. Subjects in Regimen A had a significantly lower probability of completing treatment than did subjects in Regimen C (p = 0.001, Log- rank test) and subjects in Regimen B (p = 0.006, Log-rank test). The mean time to dropout from treatment was 5 wk (range, 1 to 19 wk), being significantly shorter for subjects in Regimen A (3.8 wk) compared with subjects in Regimen B (6 wk) and Regimen C (6.2 wk) (p = 0.003).

View larger version (14K): [in this window] [in a new window]   Figure 1.   Kaplan-Meier curve of subjects continuing treatment during the 26-wk study period.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Screening for and treatment of latent infections in immigrants from areas where tuberculosis is endemic are considered effective means for reducing the incidence of tuberculosis in Western countries, and these procedures are recommended by national and international bodies (4, 13). Although adherence to IPT plays a key role in the determination of effectiveness, little information is available on adherence to IPT among illegal immigrants.

The completion rate of 41% for standard IPT in our trial was low when compared with rates of 63 to 74% in previous studies of immigrants (14, 15). One such study evaluated close contacts of immigrants with contagious tuberculosis who were probably more motivated than the subjects in our series (14). Another study evaluated persons who entered the country legally and were referred to the health authorities for medical evaluation (15). We speculate that the lower IPT completion rates in illegal immigrants are due to severe socioeconomic outcasting, unstable working and living conditions, high mobility and, perhaps, the perception of tuberculosis-preventive therapy as much less important than more immediate needs. Fear of deportation may also contribute to the problem, but this probably did not apply to the present study since the recruitment sites are well known to and considered reliable by illegal immigrants.

At least two studies have suggested a benefit for DOPT in difficult populations, including homeless men with a high prevalence of alcoholism (16), and aboriginal peoples in British Columbia (17). Application of DOPT through an outreach system of drug delivery within the community is a labor-intensive strategy that requires political will and adequate resources. As this was not available in Italy when the study was started, we studied the effect of clinic-based DOPT on adherence in a population of illegal immigrants. The completion rate we observed was unacceptably low (7%). In particular, it was significantly lower than the completion rate for standard IPT. We believe that the difference in adherence between the two regimens was due to the difficulty illegal immigrants have in reporting regularly to health services to collect their drug doses.

The mean time to dropout of less than 2 mo in our DOPT regimen clearly shows that short regimens, in addition to an out-reach drug delivery system, are required to ensure adherence. Short-course multidrug regimens for preventive therapy of tuberculosis have proved effective and well tolerated (18).

Tolerability of IPT was high, in both the daily and the intermittent regimens: a 2.4% discontinuation rate caused by adverse events and no serious clinical side effects were recorded. Although the small number of patients in this study is a limitation, these results are consistent with previous reports demonstrating a high tolerability of IPT (8, 14). The young age and the absence of chronic health problems in this population are likely to be the reasons for the lack of serious adverse events.

Management of active cases is the highest priority for tuberculosis control measures in industrialized countries. However, without screening for latent infection and subsequent preventive therapy, notification rates among immigrants are likely to remain high (2, 19). Early detection and treatment of active cases by directly observed therapy reduces disease transmission but will not substantially reduce cases of tuberculosis among immigrants, as most cases of disease are due to reactivation of infections acquired in the country of origin (19). Low completion rates of preventive therapy reduce the cost-effectiveness of this approach, making it unattractive to policymakers. In order to use IPT effectively among illegal immigrants, it is necessary to identify means for achieving higher completion rates. Future studies should evaluate the efficacy of short-term multidrug regimens delivered through outreach DOPT to illegal immigrants in industrialized countries.