|
|
| |||||||||||||||||||||||||||||||
Background Tuberculosis has reemerged as an important public health problem, and the frequency of drug resistance is increasing. A major reason for the development of resistant infections and relapse is poor compliance with medical regimens. In Tarrant County, Texas, we initiated a program of universal directly observed treatment for tuberculosis. We report the effect of the program on the rates of primary and acquired drug resistance and relapse among patients with tuberculosis.
Methods We collected information on all patients with positive cultures for Mycobacterium tuberculosis in Tarrant County from January 1, 1980, through December 31, 1992. Through October 1986, patients received a traditional, unsupervised drug regimen. Beginning in November 1986, nearly all patients received therapy under direct observation by health care personnel.
Results A total of 407 episodes in which patients received traditional treatment for tuberculosis (January 1980 through October 1986) were compared with 581 episodes in which therapy was directly observed (November 1986 through December 1992). Despite higher rates of intravenous drug use and homelessness and an increasing rate of tuberculosis during this 13-year period, the frequency of primary drug resistance decreased from 13.0 percent to 6.7 percent (P<0.001) after the institution of direct observation of therapy, and the frequency of acquired resistance declined from 14.0 percent to 2.1 percent (P<0.001). The relapse rate decreased from 20.9 percent to 5.5 percent (P<0.001), and the number of relapses with multidrug-resistant organisms decreased from 25 to 5 (P<0.001).
Conclusions The administration of therapy for M. tuberculosis infection under direct observation leads to significant reductions in the frequency of primary drug resistance, acquired drug resistance, and relapse.
Nonadherence to medical regimens is underestimated by physicians and is difficult to predict7,8,9,10. Demographic factors such as age, sex, race, marital status, education level, and socioeconomic status are not correlated with the degree of compliance with the medication regimen. By the time noncompliance is detected, resistance, relapse, or further spread of the infection may already have occurred11.
A review of cases of M. tuberculosis in Tarrant County, Texas, before 1986 revealed frequent relapse and acquired resistance, and noncompliance contributed substantially to these trends. To address this problem, we instituted a program in which all patients were observed while taking their prescribed medications (directly observed therapy). We then compared the outcome of standard therapy with that of directly observed therapy.
Methods
We reviewed the charts of all patients with positive cultures for M. tuberculosis in Tarrant County from January 1, 1980, through December 31, 1992, whose cases were reported by April 1993. In addition, we included children identified during investigation of contacts if they had positive skin tests with purified protein derivative and abnormal chest films suggestive of active infection. All isolates of M. tuberculosis obtained at the Tarrant County Health Department were submitted to a regional laboratory for sensitivity testing. When the initial diagnosis and cultures were performed at another institution, sensitivity patterns from that facility were used. From January 1980 through October 1986, therapy was unsupervised unless patients acquired resistance or had a relapse while receiving therapy. The regimen usually included isoniazid, rifampin, and ethambutol until information on the susceptibility of the isolate was available. Supervision was usually discontinued when the cultures became negative or after one to two months of supervision if the patient promised future adherence to the treatment regimen.
Beginning in November 1986, nearly all patients referred to the Tarrant County Health Department for treatment of M. tuberculosis infection were treated with directly observed therapy (90.5 percent). The only patients not so supervised were those who lived outside Tarrant County and those under the care of practitioners who did not wish their patients to undergo direct observation. Patients who were not directly observed when taking their medications (9.5 percent) underwent random urine screening for isoniazid, and their patterns of drug refills were monitored for adherence. If the urine sample was negative for isoniazid or the patient was irregular in picking up medication, he or she was switched to directly observed therapy.
The team administering directly observed therapy was organized within the staff of the Tarrant County Health Department and consisted of community-service aides, public health investigators, and nurses. The principal duties of the community-service aides were to transport patients to clinics, obtain information about the patients, deliver medication, and observe the ingestion of the medication by the patients. These aides had a high-school education and were trained to question the patients about drug toxicity. The public health investigators performed contact investigations and delivered medications to noncompliant patients. They also received training in drug toxicity and data collection. Nursing personnel (registered nurses and licensed vocational nurses) administered medications to patients who were receiving injectable drugs and supervised the community-service aides and public health investigators.
Therapy was dispensed and ingestion was observed in the clinic, residence, place of work, or wherever the patient preferred. Patients were treated with standard antituberculous regimens7,12,13,14,15 as prescribed by the physician at the Tarrant County Health Department. The regimens usually included isoniazid, rifampin, and pyrazinamide; most patients also received a fourth drug -- ethambutol, streptomycin, or capreomycin -- until their isolates were confirmed to be fully susceptible to the other drugs. Medications were administered under observation Monday through Friday for the first two to four weeks of therapy; weekend doses were issued each Friday but the patients' ingestion of these doses was not observed. The regimen was continued with medications given twice weekly after the initial two to four weeks. Patients with drug-sensitive M. tuberculosis underwent six months of directly observed therapy. Therapy initiated during a hospitalization was considered part of the therapeutic regimen if directly observed therapy was started immediately after discharge. Unsupervised therapy administered before the directly observed therapy or therapy that included an unsupervised period was not counted as part of the six months of observed therapy. The number of missed doses was doubled and added at the end of therapy. When patients who did not receive therapy under direct observation were identified as noncompliant, new cultures were obtained and the patients were treated with six months of directly observed therapy. Noncompliant patients were counseled by the physician and were issued a warning letter that outlined the consequences of continued nonadherence to the treatment regimen. If compliance remained poor, they were hospitalized or quarantined. Records were requested from other providers for patients who received part of their therapy outside Tarrant County. All requests for medical records were investigated to determine whether a patient had had a relapse elsewhere. If a relapse occurred, these records were obtained and included in the study data.
Definitions
A relapse was defined as a positive culture for M. tuberculosis in a patient who had previously completed an adequate course of therapy. A relapse during therapy was defined as a positive culture in a patient currently under treatment who had been culture-negative for a minimum of four months, with at least three negative cultures. A multidrug-resistant organism was defined as an isolate demonstrating resistance to at least two of the following drugs: isoniazid, rifampin, ethambutol, and streptomycin. A multidrug-resistant relapse was defined as a positive culture for multidrug-resistant M. tuberculosis in a patient in whom the definition of a relapse was met. Primary resistance was defined as resistance to one or more drugs present on the first sensitivity test for which results were available. Acquired resistance was defined as resistance that was not present on the initial sensitivity testing and that developed during or after treatment.
All clinic services, x-ray films, laboratory tests, transportation, and subsequent hospital services were free to the patients in this study. Apart from those described above, no other changes occurred in the care provided to patients with tuberculosis from 1980 through 1992.
Statistical Analysis
The data were initially analyzed by two-by-k chi-square contingency analysis, with k equal to the number of groups. If statistical significance was demonstrated, each group was analyzed with two-by-two chi-square contingency analysis with a continuity correction. To minimize time-dependent effects, data were entered according to the date of occurrence (Table 1). Episodes of relapse or recurrence were included in the year during which they were identified. The alpha level was set at 0.05 in this analysis.
|
Between January 1, 1980, and December 31, 1992, there were 988 cases of M. tuberculosis. Fifteen occurred in children with active disease who were identified on the basis of clinical presentation alone during the tracing of patients' contacts. During this period there were 25 relapses during therapy. A total of 407 cases occurred in 379 patients between January 1, 1980, and October 31, 1986, and 581 cases in 578 patients between November 1, 1986, and December 31, 1992. In the latter period, 9.5 percent of the patients were receiving unsupervised therapy. The rates of multidrug-resistant tuberculosis, relapse, primary resistance, acquired resistance, and M. tuberculosis per 100,000 people in Tarrant County are shown in Figure 1.
|
|
The proportion of cases with primary resistance decreased from 13.0 percent to 6.7 percent (P<0.001) after the program of directly observed therapy was implemented (Table 3). Primary resistance occurred between 1980 and 1986 in 10 persons who were in close contact with noncompliant patients with acquired drug-resistant M. tuberculosis. When these contacts were excluded from consideration because of their exposure, the rate of primary resistance was 10.8 percent with traditional therapy. This figure is still significantly different from the 6.7 percent rate of primary resistance in the directly observed group (P = 0.041). In the directly observed group there were no cases of primary resistance resulting from close contact with a patient with acquired resistance.
|
The proportion of patients with acquired resistance decreased from 10.3 percent to 1.4 percent (P<0.001) after directly observed therapy was instituted (Table 3), and the number of episodes of acquired resistance fell from 57 to 12 (P<0.001). No patient treated with directly observed therapy from the beginning of treatment to its completion acquired resistance. Twelve episodes of acquired resistance occurred in eight patients between November 1986 and December 1992; three occurred in two patients treated from 1980 through 1986 who had relapses outside Tarrant County, two in a patient treated with unsupervised therapy in a corrections facility, two in two patients whose initial therapy was unsupervised and given outside our county, and two in a patient receiving unsupervised care at a military facility. Another two episodes occurred in a patient treated with traditional therapy outside Tarrant County in whom multidrug resistance developed and who had a relapse of disease that was resistant to two additional drugs. Finally, one episode occurred in a patient treated in another county under direct observation but with an inadequate regimen.
Relapse
The rate of relapse decreased from 20.9 percent to 5.5 percent (P<0.001) with directly observed therapy (Table 3). Only two patients who completed a full course of directly observed therapy had relapses, both more than five years after the completion of therapy. Both were homeless and alcoholic, lived in night shelters, and were at high risk for reinfection. The relation of the organisms causing the relapse to the original strain could not be determined by analysis of restriction-fragment-length polymorphisms since the original cultures were no longer available. Three patients had relapses after incomplete courses of directly observed therapy. All patients treated with directly observed therapy who relapsed had drug-susceptible organisms on repeat culture.
Multidrug-Resistant Relapse
The rate of multidrug-resistant relapse decreased from 6.1 percent to 0.9 percent (P<0.001) (Table 3). All five patients who had relapses of multidrug-resistant tuberculosis between November 1986 and December 1992 had previously received unsupervised therapy. No patient treated with directly observed therapy from the beginning to the completion of therapy had a multidrug-resistant relapse.
Discussion
The resurgence of M. tuberculosis in the United States since 1988 is well documented. In addition to the increased numbers of cases, the proportion of patients with resistant organisms has also increased,2,3 with many outbreaks of multidrug-resistant disease16,17,18. Drug resistance develops primarily as a result of patients' noncompliance with treatment19. It occasionally results from malabsorption of medication or an inadequate regimen7,11. Tuberculosis is unusual in that the failure to take medication as prescribed can change an easily treated infection to an untreatable disease in the patient and his or her contacts. The treatment of multidrug-resistant tuberculosis is long, complicated, and often unsuccessful even when patients receive care as inpatients at referral centers. In one recent review of the experience at a referral center, 35 percent of the patients with multidrug-resistant tuberculosis had no response to treatment after a median of 7.8 months in the hospital19.
Increasing numbers of patients are not completing therapy and have either ongoing or recurrent disease2,3. Eight percent of patients treated for M. tuberculosis have multiple relapses20. Such patients typically remain infectious for long periods, contribute to the increase in the incidence of M. tuberculosis, and are more likely than others to have drug-resistant isolates. Treatment of relapses in such patients is costly11.
Noncompliance on the part of patients is the most serious remaining problem in the control of tuberculosis21 and the chief cause of relapse and drug resistance6,7. Compliance with drug treatment for M. tuberculosis has been extensively studied since the 1950s10,22. Age, sex, religion, education, race, and socioeconomic status do not predict compliance23. Psychiatric disease, alcoholism, drug addiction, substance abuse, and homelessness do predict noncompliance24,25. Nonadherence to a treatment regimen can be difficult to detect. Predictions based on interviews with patients correlate poorly with more direct measures of compliance such as pill counts and screening of urine and serum for antituberculous drugs26. Medical personnel predict compliance poorly9,10. It has been suggested that physicians should expect a 35 percent rate of noncompliance with prescribed therapy27,28.
Problems associated with the self-administration of antituberculous medications led to studies of the efficacy of therapy given intermittently12,13,14,29. Daily therapy for two to four weeks followed by twice-weekly therapy for a total of six months is as effective as daily therapy. Regimens without the initial two to four weeks of daily therapy, in which medications are given thrice weekly for the entire treatment period, are equally effective15. Directly observed therapy on an outpatient basis was successful in a group of patients who abused alcohol and used illicit drugs in whom unsupervised ambulatory therapy had already failed30. Despite this evidence, directly observed therapy is not widely used. Estimates from the United States suggest that at present only 10 to 12 percent of patients are treated with directly observed therapy31; worldwide, the practice is even less common. The underuse of this strategy is surprising in view of the well-known frequency of noncompliance among patients infected with M. tuberculosis. In a recent report from New York, 53 percent of patients were alcoholic, 55 percent used injection drugs, and 45 percent were homeless2; directly observed therapy is already recommended for all these groups.
There are no easy solutions to the problems of drug addiction, alcoholism, homelessness, psychiatric illness, or indifference on the part of patients, but immediate solutions to the problems of drug resistance and relapse are needed. Directly observed therapy is the only approach that guarantees effective treatment of M. tuberculosis in these difficult circumstances. Some suggest that the universal use of directly observed therapy is impractical, inefficient, and wasteful and that it undercuts a legitimate desire to individualize treatment regimens32. Others suggest that the universal application of directly observed therapy would be restrictive and intrusive and, although politically correct, is unnecessary for patients who are motivated to comply with treatment32,33.
Our data show that directly observed therapy is more effective than traditional unsupervised therapy. Among the patients we studied, implementation of a policy of using directly observed therapy for virtually all patients resulted in a decrease in the frequency of primary resistance from 13.0 percent to 6.7 percent of patients (P = 0.001) and a decrease in the proportion with acquired resistance from 10.3 percent to 1.4 percent (P<0.001). The rate of relapse decreased from 20.9 percent to 5.5 percent (P<0.001), and the rate of multidrug-resistant relapse decreased from 6.1 percent to 0.9 percent (P<0.001). These improvements occurred despite an increased frequency of intravenous drug use, homelessness, and seropositivity for the human immunodeficiency virus. We also found that when treatment with directly observed therapy was unsuccessful and patients relapsed, relapse was due to drug-sensitive organisms34 that had an excellent chance of successful retreatment.
The cost of directly observed therapy has been cited as a reason for not implementing these programs widely. Given the reductions in the rates of relapse, multidrug-resistant relapse, and acquired resistance (Table 1), our data suggest that universal directly observed therapy is not only successful but also cost effective. The cost of treating multidrug-resistant tuberculosis is high, with median hospital stays of 7.3 months in successfully treated cases and 7.8 months in cases in which treatment fails19. Hospital expenses alone may range from $100,000 to $180,000 per case,11,16 and hospital costs for a single outbreak approach $1 million16. Finally, the Tarrant County program of directly observed therapy was implemented with the same staff and support services as were present in 1980 through 1986.
To be successful, directly observed therapy must be individualized and must not be intrusive. Our patients, from all socioeconomic groups, accepted directly observed therapy after the benefits to them and to society were explained to them. Many freely admitted that during a previous infection they took medication irregularly or only until their symptoms resolved. Patients and staff members frequently developed close personal bonds. Some patients even expressed regret that they would no longer be seeing their case workers when their therapy was completed. Now that the program of directly observed therapy has been in place for more than seven years, new patients frequently ask about directly observed therapy at the initial visit.
M. tuberculosis disproportionately affects socioeconomically disadvantaged groups. Compliance with therapy is difficult to predict even though noncompliance is expected among alcoholics, drug users, the homeless, and those with psychiatric illnesses. The complications of erratic therapy are serious and affect all groups of patients. This study confirms that the universal administration of treatment under direct observation by a local health department at no cost to the patient is effective and feasible; such a program can significantly reduce the rates of primary drug resistance, acquired drug resistance, and relapse.
Supported in part by a grant (U52/CCU 600 497-11) from the Centers for Disease Control and Prevention.
We are indebted to Dr. Thomas Beitinger of the University of North Texas for performing the statistical analysis; to Dr. Michael Iseman and Dr. John Sbarbaro for serving as consultants on universal directly observed therapy; and to Thelma McCormick for administrative assistance and assistance in the preparation of the manuscript.
Source Information
From the Research Division (S.E.W., M.N.), the Division of Pulmonary-Critical Care (P.C.S., B.H.F.), and the Division of Infectious Disease (F.X.B.), Department of Medicine, University of North Texas Health Science Center at Fort Worth; the Fort Worth-Tarrant County Health Department (S.E.W., B.K., G.B.M.); and John Peter Smith Hospital (E.G.) -- all in Fort Worth, Tex.
Address reprint requests to Dr. Weis at the Research Division, Department of Medicine, University of North Texas Health Science Center at Fort Worth, Suite 232-ME2, 3500 Camp Bowie, Fort Worth, TX 76107.
References
| |||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
