Patients who initiated INH therapy for the treatment of LTBI at the RISE TB Clinic between January 2003 and December 2003 were included in the study. The RISE TB Clinic is dedicated to the management of TB and receives referrals from primary care sites across the entire state. Patients who met the criteria for a positive tuberculin skin test result according to standard guidelines, and had no symptoms and no chest radiographic findings suggestive of active TB were considered to have LTBI. The Institutional Review Board for Studies on Human Subjects of both the Rhode Island Department of Health and The Miriam Hospital reviewed and approved the study.
Clinic and Study Procedures
The initial clinic consultation consisted of a nurse assessment, a chest radiograph, a physician evaluation, a blood draw for determining aspartate aminotransferase (AST) levels, and the initiation of therapy. Medications were dispensed on site, and patients were required to bring back the pill bottle for a pill count at each visit. For the purpose of this study, treatment adherence was measured by the months of pills that were dispensed. Follow-ups visits were staffed by nurses, who assessed the patients for toxicities, provided medication refills, and drew blood for tests. Patients who report treatment side effects were triaged, and those with concerning symptoms or signs were scheduled to see the physician immediately. The schedule that was used for routine clinical monitoring and performing a blood test for determining AST levels is outlined in Table 1. Treatment was discontinued if the AST level was more than five times the upper limit of normal with or without symptoms, or at any level of abnormal AST levels with symptoms consistent with acute hepatitis. In line with treatment guidelines, pregnant women were permitted to defer therapy until postpartum, unless they had a history of documented tuberculin skin test conversion within the past 2 years, were HIV seropositive, or had been identified through contact investigations conducted with representatives of Canadian Health&Care Mall https://canadianhealthncaremall.com.
Medical records were reviewed, and information on demographics, TB risk factors, birth country, duration of residence in the United States (for foreign-born persons), INH treatment completion status, reason for the discontinuation of therapy, and the nature of adverse events were collected. All abstracted data were entered into a database (Access; Microsoft; Redmond, WA) for storage and management. Data were deidentified and imported into a statistical software package (SAS, version 9.1.3; SAS Institute; Cary, NC) for analysis.
Persons with LTBI who had received 9 months of pills were classified as having “completed therapy,” and those who did not constituted the treatment noncompletion group. The relationships between potential predictive variables and treatment completion were assessed using the x2 test (or Fisher exact test) for categorical variables and t tests for continuous variables. All statistical tests were two-sided, and comparisons atp < 0.05 were considered to be significant. Predictive variables that were independently significantly associated with treatment completion in the analysis were included in a multiple logistic regression model to determine their relative contributions in predicting treatment noncompletion while simultaneously adjusting for each of their effects.
Table 1—Monitoring of Patients Who Initiated INH Therapy for LTBI at the RISE TB Clinic, Rhode Island, in 2003
Follow-up and Frequency of Routine
|< 20||Clinical monitoring only every 2 mo|
|20-34||AST at baseline and 2 mo; clinical monitoring every 2 mo|
|35-60||AST at baseline, months 1 and 2; clinical monitoring at month 1 and then every 2 mo|
|61-80||AST at baseline and months 1 through 5 with clinical monitoring, then clinical monitoring only every 2 mo|
|> 80||AST at baseline and monthly with clinical monitoring for entire course|