Initial drug regimen for active tuberculosis cases in Montreal , 1995 to 1998

Direction de la santé publique, Régie régionale de la santé et des services sociaux de Montréal-Centre; Joint Departments of Occupational Health and Epidemiology & Biostatistics, McGill University, Montreal, Québec Correspondence: Dr Paul Rivest, Direction de santé publique de Montréal, 1301, rue Sherbrooke Est, Montréal, Québec H2L 1M3. Telephone 514-528-2400 ext 3678, fax 514-528-2452, e-mail privest@santepub-mtl.qc.ca Received for publication May 31, 1999. Accepted March 18, 2000 P Rivest, TN Tannenbaum. Initial drug regimen for active tuberculosis cases in Montreal, 1995 to 1998. Can J Infect Dis 2001;12(2):89-92.

Although Canada did not experience the same rate increase as the United States, the Canadian Thoracic Society, in the Canadian Tuberculosis Standards, 4th edition, made similar recommendations (3).These recommendations included the use of a four-drug regimen of isoniazid (INH), rifampin (RIF), pyrazinamide (PZA) and ethambutol (EMB) or streptomycin (SM) for the initial treatment of active TB.In communities where overall INH resistance exceeds 4%, all patients with active TB should be treated initially with INH, RIF, PZA and EMB or SM.When susceptibility to INH and RIF has been demonstrated, EMB or SM can be stopped, and PZA may be discontinued after the initial eight weeks of treatment (1)(2)(3).
In Montreal, between 1992 and 1995, the proportion of TB cases resistant to INH was greater than 4% in almost all age groups among both foreign-born and Canadian-born cases (4).This article presents data on the initial drug regimen prescribed by clinicians between 1995 and 1998 and on factors that may have influenced the choice of initial treatment.

DATA AND METHODS
Throughout Canada, active TB is a reportable disease.In Montreal, all cases of active TB are reported to the Department of Public Health (DPH).Each case is investigated by a public health nurse who obtains clinical and epidemiological information and ensures that appropriate treatment is initiated, that the patient adheres to the treatment regimen, and that close contacts of the patient are tested and referred for further evaluation when necessary.
The DPH maintains a nominal database of all reported patients having active TB who reside in the region at the time of diagnosis.This database contains epidemiological data and the results of drug susceptibility testing.
Since 1995, the public health nurses have been systematically identifying all medications prescribed at the beginning of therapy.This allowed the authors to assess clinicians' adherence to the recommended four-drug regimen.
Logistical regression was performed to identify predictors of the use of a four-drug regimen.The SAS statistical software program (SAS Institute Inc, USA) was used for statistical analyses.Odds ratios (ORs) with 95% CIs were calculated.

RESULTS
During the study period, 741 cases of active TB were reported to the DPH (with an average annual rate of 10 cases/100,000 people).Of these, 672 (90.7%) were confirmed cases: 670 by culture and two by polymerase chain reaction only.All cultureconfirmed isolates of M tuberculosis were tested for susceptibility to the first-line drugs by the Quebec Public Health Laboratory, using the BACTEC 460 (Becton Dickinson, Canada) radiometric method.
The analysis excluded 15 cases that were diagnosed postmortem, 10 cases whose initial therapy began after the receipt of results of susceptibility testing and five cases that did not have adequate information on the initial drug regimen.Also excluded were 24 cases in children under age 12 years, because the use of ethambutol in young children is somewhat controversial.
In the present review, of the 670 culture-positive cases, there were 60 cases (9.0%) with resistance to at least one of the four first-line drugs -51 (7.6%) with INH resistance and six (0.9%) with multidrug-resistant TB.For 20 (33.3%) of the 60 patients, the initial treatment did not include the recommended four-drug regimen (Table 3).Thirteen were treated with INH-RIF-PZA, five with INH-RIF-EMB and two with INH-RIF.

DISCUSSION
The present study was limited to information that was recorded routinely on the TB surveillance database.All susceptibility testing was performed at the provincial reference laboratory, and the reference laboratory staff was able to verify that the information in the provincial database coincides with laboratory results.Therefore, the authors are confident that the information on drug sensitivities is valid.Information on initial drug treatment was also collected systematically by public health nurses who are dedicated to TB control.This information can be obtained from several sources -the hospital chart, the patient and/or the pharmacy.Therefore, the authors feel that this information is also valid.
In Montreal, between 1995 and 1998, 40% of TB patients were treated initially with fewer than four drugs.The age, site of TB and country of origin were associated with the use of four-drug regimens.However, even when one of these factors was present, between 30% and 39% of patients received three drugs or less.This implies that clinicians in Montreal were unaware of the recommendations, were unaware of the extent of INH resistance in Montreal or did not agree with the treatment recommendations.
Surveillance data provide only limited information on treatment decisions; individual patient factors that may have led the treating physicians to choose the specific drug regimen cannot be fully appreciated.It is possible that, although the physicians were aware of the recommendations, some chose to exclude one of the medications because of anticipated side effects, the presence of minimal disease or possible interaction with other medications that the patient was taking.However, many surveys in the United States have revealed poor adherence to recommended treatment guidelines (5)(6)(7).In Toronto, of 141 culture-confirmed pulmonary TB cases diagnosed between 1992 and 1993, only 32 (22.7%) received an initial four-drug treatment, and 90 (63.8%) received a regimen of three drugs (8).According to the authors, in 25% of cases, the initial treatment was a regimen not considered standard by the 1993 current guidelines (INH-RIF-PZA with or without EMB), but the study looked at patients who were started on therapy before the four-drug guidelines of 1993 (1,2).From 1994 to 1995 in New Jersey, where almost all TB patients (98%) were reported from counties with an INH-resistant proportion of 4% or more, 35.5% of the 1230 culture-positive cases were not initially treated with the four-drug regimen.Nonhispanic white patients were more likely to be treated with fewer than four drugs than nonhispanic black patients.Private practitioners and physicians at chest clinics were approximately five times more likely to prescribe fewer than four drugs initially than physicians at the hospital where a national TB centre is located (9).

Can J Infect
The fact that younger age and being foreign-born were associated with initial treatment with four drugs suggests that physicians were less likely to suspect resistance in Canadianborn patients and perhaps were less likely to fear secondary effects in young patients.In fact, of the 20 patients with resistance to at least one of the first-line drugs who did not receive four drugs initially, 50% were either Canadian-born or elderly.
Inadequate treatment strategies may contribute to poor outcomes for TB patients and to the emergence of drug resistance.Unfortunately, the TB surveillance database presently does not contain information that would allow the comparison of outcomes among cases having initial four-drug regimens with cases having nonstandard regimens.These outcomes include the development of secondary resistance or side effects to the medication.
However, among the 406 patients treated with an initial four-drug regimen, no relapse was identified in the surveillance database as of January 2000.A relapse case is defined as a new episode of active TB in an individual who received a complete course of treatment, was free of disease for six months or was lost to care for more than 12 months.Such a case would be entered as a new episode of active TB in the database.Among the other 281 patients, one patient who was initially treated with INH-RIF for nine months had a relapse eight months after the completion of the initial treatment.
Although information is limited, the public health files did not identify any deaths, relapses or development of secondary resistance among the 20 patients who were not started on four drugs and whose strain was resistant to at least one of the four first-line drugs.
In Montreal, from 1995 to 1998, the proportion of INH resistance was less than 4% among Canadian-born patients.But because the annual number and the total number of INHresistant cases among Canadian-born patients are small, the proportion of INH resistance may fluctuate from year to year.From 1992 to 1995, the percentage of INH-resistant cases among Canadian-born patients was 8.7% (4), and over the seven-year period of 1992 to 1998, the percentage of INHresistant cases was 5.9% and was greater than 4% among all age groups.
In Quebec, any physician may treat patients with TB, and since 1997, medication can be dispensed by any community pharmacy.Public health departments, therefore, do not become aware of individual treatment regimens before the case is declared, which may occur several weeks after treatment has been initiated.In this type of health care system, it is necessary for public health departments to ensure that clinicians are aware of resistance levels in their community and treatment guidelines.It is also necessary for public health departments to monitor treatment regimens and ensure that the regimens used are appropriate for the susceptibility patterns of the individual patient and that they are in accordance with guidelines.Since 1995, several guidelines have been published and disseminated by the public health department and other bodies.In addition, a report on the epidemiology of TB in Montreal has been disseminated, and public health physicians have met with various groups of physicians.It is possible that these efforts contributed to the increase in the use of the initial four-drug regimen in Montreal.However, the results of the present study demonstrate that physicians have not adopted the optimum treatment regimen and that efforts to increase the use of these regimens must continue.
EMB Ethambutol; INH Isoniazid; PZA Pyrazinamide; RIF Rifampin pulmonary (including miliary) disease, compared with 32.7% of cases with peripheral adenitis and 42.2% of cases with other sites of disease.Of the 456 cases of pulmonary disease, those with a positive smear examination were more likely to receive a four-drug regimen (181 of 249 patients [72.7%]) than were those with a negative smear examination (138 of 207 patients [66.7%]).

TABLE 2 Number and percentage of tuberculosis cases on a four-drug regimen in Montreal from 1995 to 1998
*Univariate analysis; † Multivariate analysis