A Retrospective Evaluation of Completion Rates, Total Cost, and Adverse Effects for Treatment of Latent Tuberculosis Infection in a Public Health Clinic in Central Massachusetts

Tuberculosis continues to be a major public health concern. In 2006, there were 13,779 cases of active tuberculosis diagnosed in the United States, 259 in Massachusetts, and 19 in Worcester, the largest city in central Massachusetts [1]. Latent tuberculosis infection (LTBI) is the most common precursor to active disease in the United States, and the long-term tuberculosis control strategy in the United States has focused on the treatment of persons with LTBI [2].

The Centers for Disease Control and Prevention (CDC) and the American Thoracic Society (ATS) recommend at the A(II) level (ie, preferred regimen based on data from nonrandomized trials) 9 months of daily isoniazid for the treatment of LTBI [3]. Although this regimen is ∼70% effective in preventing active tuberculosis, nonadherence is reported in 24%-47.4% of patients [4–10]. The CDC and ATS also endorse rifampin for 4 months at the B(II) level (ie, acceptable alternative based on data from nonrandomized clinical trials) as an alternative regimen for LTBI [3, 11]. To date, few studies have compared completion rates and adverse effects for these medications and just 1 study has addressed costs [6–10, 12]. To our knowledge, our study is the largest study published to date that compares the triad of completion rates, adverse effects, and clinical charges for patients receiving 4 months of rifampin and 9 months of isoniazid in the United States.

Methods. A retrospective medical record review was used to obtain data from patients ⩾18 years old who began therapy for LTBI at a public health clinic in Worcester County, Massachusetts, on or after 15 May 2003 and either completed or discontinued treatment before 15 March 2007. The institutional review board at the University of Massachusetts Medical School approved this study.

A total of 777 patients were included in the study. The following patients were excluded from the study: (1) patients with unavailable medical records, (2) patients <18 years old, (3) prisoners, (4) patients who declined treatment or who were lost to follow-up after <1 month, (5) patients who were previously treated for either active tuberculosis or LTBI, (6) patients who began treatment for LTBI at another facility but transferred care to the study clinic, and (7) patients who had active tuberculosis.

Local medical professionals and community programs referred patients to the study clinic for a positive tuberculin skin test result on the basis of the patient's risk group (⩾5 mm of induration for immunocompromised patients or contacts of an active case and ⩾10 mm of induration for standard high-risk patients). At the initial clinic visit, the nurses and physicians obtained demographic information, medical history (including a self-reported history of hepatitis), and data regarding previous tuberculin skin tests. Chest radiography was performed if it had not been done in the previous 2 months. If the patient met criteria for LTBI and agreed to treatment, the attending physician prescribed either 9 months of isoniazid or 4 months of rifampin at his or her discretion.

Patients returned to the study clinic for monthly nursing visits. The nurse assessed the patient's adherence to the medication and screened for adverse effects. If adverse effects were reported, the nurse obtained blood work and/or referred the patient to the attending physician. Alanine transaminase (ALT) monitoring was based on clinic protocols and physician assessment of the presence of risk factors that could lead to hepatotoxicity. Patients were given a 1- or 2-month supply of medication by the nurse at each follow-up appointment. Patients were discharged from the clinic after they completed the protocol (ie, 9 months of isoniazid within 12 months or 4 months of rifampin within 6 months).

The mean cost of therapy was defined as the sum of the charges for ALT monitoring ($15.50 per ALT level measurement), the charges (state-approved payment rate) for a “doctor visit” ($57.34) or a “doctor visit plus radiograph” ($83.97) or a “nurse visit” ($17.46), and the wholesale cost of the medications to the Massachusetts Department of Public Health ($2.38 per month for isoniazid or $40.49 per month for rifampin). Additional costs (eg, staff salary, additional laboratory data, and specialty referrals) were not included. Total actual charges were summed for each patient, and the mean charge for isoniazid was compared with that for rifampin.

Before beginning the study, a power calculation was performed to guide the sample size. At least 107 patients were needed for each arm of the study to have 90% power to detect a 20% difference in the completion rates between 9 months of isoniazid and 4 months of rifampin. The data were processed using SPSS, version 15.0 (SPSS). Continuous variables were compared by the mean value and the 95% confidence interval. Discreet variables were compared with the Pearson χ²or Fisher exact tests. Univariate and multivariate logistic regressions were used to control for confounding variables.

Results. There were 639 patients in the group receiving 9 months of isoniazid and 138 patients in the group receiving 4 months of rifampin. Demographic information and medical histories were similar. Most patients had normal radiography results (90.5% in the isoniazid group vs 91.3% in the rifampin group; P=.820), and 35 patients had radiological evidence of old tuberculosis (4.7% in the isoniazid group vs 3.6% in the rifampin group; P=.820). All patients who were using hormonal contraception (34 patients) or taking antiretroviral therapy for human immunodeficiency virus (3 patients) received 9 months of isoniazid because of the possibility of medication interaction with rifampin.

Patients receiving 4 months of rifampin had a significantly better completion rate than those receiving 9 months of isoniazid (90.6% vs 65%; P<.001). These patients also had a higher completion rate than patients who completed at least 6 months of isoniazid (90.6% vs 72.5%; P<.001), a regimen that is a B(I) recommendation (ie, acceptable alternative based on data from randomized clinical trials) of the CDC and ATS [3]. There was a trend for patients who were prescribed 9 months of isoniazid to be lost to follow-up more often than those who were prescribed 4 months of rifampin (19.3% vs 5.8%; P=.677).

Most adverse effects (eg, nausea, vomiting, abdominal pain, jaundice, fatigue, and rash) did not differ between the 2 groups. Neuropathy was statistically more common in patients receiving 9 months of isoniazid (4.5% vs 0%; P=.005), and patients receiving isoniazid were more likely to have symptomatic elevations in ALT levels. Fifteen patients experienced a symptomatic elevation in ALT level >2.5 times the upper limit of normal (ie, ⩾100 IU/L), 14 of whom were taking isoniazid. These 14 patients were more likely to report a history of hepatitis (35.7% vs 4.5%; P<.001) or to be treated for hyper-lipidemia (21.4% vs 3.4%; P=.024) than those taking isoniazid who did not have an elevated ALT level. Patients with a history of alcohol abuse also trended toward higher ALT values. Six of the patients, all of whom were taking isoniazid, had a symptomatic elevation in ALT level >5 times the upper limit of normal (ie, ⩾200 U/L). None of the patients experienced irreversible hepatotoxic effects, and the ALT level was measured at regular intervals until it returned to normal.

The mean total cost of a full course of 9 months of isoniazid was less than that of 4 months of rifampin ($259.92 vs $313.81; P<.001). Both the cost of screening ALT levels and the clinic charge for monthly visits were higher in patients taking isoniazid for 9 months than in patients taking rifampin for 4 months ($19.82 vs $15.13, P=.015, compared with $217.98 vs $135.42, P<.001). However, the medication cost of 4 months of rifampin was higher than that of 9 months of isoniazid ($163.26 vs $22.12; P<.001). Table 1gives the results of the multivariable logistic regression analysis.

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Table 1

Multivariable Analysis of Patients Who Finished Treatment Compared with Those Who Discontinued Treatment

Discussion. Patients with LTBI treated with rifampin have a higher completion rate than patients given 9 months or 6 months of isoniazid in Worcester County, Massachusetts. In fact, patients taking 4 months of rifampin had a higher completion rate than those taking 4 months of isoniazid (91.3% vs 77.2%; P=.002), which suggests that rifampin may be better tolerated than isoniazid in the early period. Consistent with previously published studies, the multivariable logistic regression (table 1) indicates that the therapy chosen is among the strongest predictors of whether a patient will complete a course of treatment (odds ratio [OR], 4.40; P<.001) [6–8, 10]. Other factors that predicted completion of treatment included female sex (OR, 1.60; P=.039), increasing age (OR, 1.02; P=.009), and Asian race (OR, 8.59; P=.002).

The incidence of elevated ALT levels attributable to isoniazid in this study (2.2%) is consistent with the rates from previous studies (0.1%-4%) [13, 14]. Patients who experienced hepatotoxicity in the isoniazid group were predominately male, had a higher self-reported history of hepatitis or alcohol abuse, and were more likely to be treated for hyperlipidemia. Although the numbers are small, these findings suggest that rifampin may be a better choice than isoniazid for this subset of patients.

Although our data suggest that 9 months of isoniazid is a more cost-effective medication regimen, it is important to realize that the costs of isoniazid and rifampin vary geographically; rifampin is ∼17 times more expensive than isoniazid for a monthly supply in Massachusetts. In other countries, rifampin and isoniazid have more comparable monthly costs because a generic formulation is widely available [6–8, 10]. Rifampin for 4 months allows public health clinics to use the same number of staff members to treat twice as many patients because there is a mean of 4 visits rather than 9 visits per patient. The public health benefit to having fewer partially treated patients with LTBI is difficult to quantify but must not be overlooked. Despite these benefits, the protocol at our public health clinic has not changed as a result of the data presented in this article.

If an active case of tuberculosis is missed and the patient is given 4 months of rifampin monotherapy, there may be rapid development of resistance to rifampin. Although a small previous study did not find this to be problematic, resistance to rifampin in patients with active tuberculosis requires a longer treatment course and is associated with a lower cure rate [15]. In addition, few studies have examined the efficacy of LTBI treatment with rifampin, and none have been large randomized controlled studies [10, 15, 16].

In summary, patients are more likely to complete 4 months of rifampin therapy, compared with 9 months of isoniazid therapy, and adverse effects are not significantly different. Before 4 months of rifampin can be endorsed as the preferred regimen for LTBI, larger prospective studies need to be performed to determine the effect of rifampin on overall tuberculosis resistance and its relative efficacy in the prevention of reactivation of tuberculosis.

• Received October 21, 2008.
• Accepted March 10, 2009.

• © 2009 by the Infectious Diseases Society of America