Temporal Trends and Transmission Patterns During the Emergence of Multidrug-Resistant Tuberculosis in New York City: A Molecular Epidemiological Assessment Running Head: Emergence of MDR TB in New York City Authors: Robert W. Shafer, M.D.(1) Peter M. Small, M.D.(2) Christina Larkin, M.P.A.(3) Samir P. Singh, B.A.(2) Patricia Kelly, M.P.H., F.N.P.(4) Marcellino F. Sierra, Ph.D.(4) Gary Schoolnik, M.D.(2) Keith D. Chirgwin, M.D.(4) 1Center for AIDS Research, Stanford University, Stanford, CA; 2Department of Medicine, Division of Infectious Diseases, Stanford University, Stanford, CA; 3Division of Tuberculosis Control, New York City Department of Health; 4Division of Infectious Diseases, SUNY - Health Science Center at Brooklyn. 1) This research was presented in part at the 33rd Interscience Conference on Antimicrobial Agents and Chemotherapy, October 17-20, 1993, New Orleans, LA; Abstract 605:229. 2) The human experimentation guidelines of the U.S. Department of Health and Human Services and those of the authors' institutions were followed in the conduct of the clinical research. 3) No author has a commercial or other association that might pose a conflict of interest. 4) This work was funded in part by grants from the National Institutes of Health (AI-27762 and AI-27666). Requests for reprints should be addressed to: Robert W. Shafer, MD, Center for AIDS Research, Room S-156, Stanford University Medical Center, Stanford, CA 94305. Tel: 415-725-2946. Fax: 415-725-2395 Abstract To ascertain the role of the human immunodeficiency virus (HIV) and Mycobacterium tuberculosis transmission on multidrug-resistant (MDR) tuberculosis (TB) emergence in New York City, medical records, drug susceptibilities, and restriction-fragment-length-polymorphisms (RFLPs) of TB cases at a city hospital between two nine-month periods (1987/1988 and 1990/1991) were reviewed. The proportion of TB patients with MDR increased from 10% (27/267) to 17% (38/222; P = .03). Among MDR TB patients of known HIV status, the proportion with HIV increased from 16% (3/19) to 58% (22/38; P = .006). HIV-infected MDR TB patients were more likely than HIV-seronegative MDR TB patients to have initial MDR (88% vs 56%; P = .03). Among 56 MDR cases with RFLP results, 12 had unique patterns; 44 belonged to one of six clusters. 75% (27/36) of MDR TB patients during the 1990/1991 period were infected with strains cultured from HIV-seronegative patients during the 1987/1988 period. The increase in MDR TB caused by TB transmission from immunocompetent to immunocompromised persons underscores the urgency of TB control in populations with increasing HIV prevalence. Introduction The incidence of tuberculosis (TB) in the United States is increasing, and perhaps the most serious aspect of the problem is the increase in TB caused by strains resistant to both isoniazid and rifampin (multidrug-resistant (MDR) TB) [1]. The treatment of MDR TB is often unsuccessful, rendering patients chronically ill and persistently infectious [2]. The proportion of TB patients in the United States with MDR TB has increased from approximately 1% in the early 1980s to 3.5% in the first quarter of 1991 [1, 3]. In the first three months of 1991, 61% of cases of MDR TB in the United States were reported from New York City [3]. In a laboratory survey in April 1991, 19% of patients in New York City with positive cultures for Mycobacterium tuberculosis had MDR TB [4]. Drug-resistant TB occurs in individuals by two mechanisms. Patients initially infected with drug-susceptible strains of M. tuberculosis, may acquire drug resistance due to inadequate or inappropriately administered therapy (acquired drug resistance) [5]. Alternatively, patients may become infected with drug-resistant strains following exposure to a patient with infectious drug-resistant TB (initial or transmitted drug resistance). Recent institutional outbreaks have demonstrated the extraordinary risk of progression to clinical disease faced by HIV-infected persons who become newly infected with MDR strains of M. tuberculosis [6, 7]. However, the dynamics by which MDR strains of M. tuberculosis emerge within a population and the role of human immunodeficiency virus (HIV) infection in this process have not been examined. To address these questions, we assessed the prevalence of MDR TB among patients at an inner city public hospital in New York City which reported more than 10% of city-wide TB cases between 1987-1991 [8]. Patients with MDR TB were characterized according to their previous history of TB and according to the presence or absence of HIV infection. In addition, to identify the routes of MDR TB transmission between these patients, the M. tuberculosis isolates were also characterized by DNA restriction-fragment-length-polymorphism (RFLP) analysis. Our results indicate that MDR TB was first detected in HIV-seronegative patients and from this population spread to HIV-infected patients, facilitating the MDR TB epidemic. Methods We reviewed the medical records and drug susceptibility results of all patients at Kings County Hospital (KCH) from whom M. tuberculosis was isolated between October 1, 1987 - June 30, 1988 and between October 1, 1990 - June 30, 1991. The first study period was chosen because epidemiological and HIV-serological data were available on many of the patients as a result of an earlier HIV-seroprevalence study [9]. The second study period was chosen to be of the same duration (nine months) and concurrent with the increased prevalence of MDR TB throughout New York City. Patients with M. tuberculosis isolates resistant to both isoniazid and rifampin (with or without resistance to other drugs) were considered to have MDR TB. In these patients, all available drug susceptibility results from previous episodes of TB were obtained from KCH, the New York City Tuberculosis Registry, and from all other known clinics and hospitals where patients had previously been treated. Drug susceptibility testing of M. tuberculosis isolates from patients at KCH was performed at the New York City Department of Health Mycobacteriology Laboratory using a radiometric broth method (BACTEC System; Beckton Dickinson Diagnostic Instruments Systems, Towson, MD). During the first nine-month study period, the concentrations of the antituberculosis drugs tested were as follows: isoniazid 0.2 ug/ml, rifampin 2.0 ug/ml, streptomycin 6.0 ug/ml, and ethambutol 7.5 ug/ml. During the second nine-month study period, the concentration of isoniazid was 0.1 ug/ml and pyrazinamide was tested using BACTEC PZA Test Medium (Middlebrook 7H12, pH 6.0) at a concentration of 100 ug/ml. RFLP analysis was performed on MDR isolates of M. tuberculosis using a standardized protocol [10]. In brief, M. tuberculosis was harvested from Lowenstein-Jensen slants, and bacterial cell walls were digested with lysozyme, proteinase K, and SDS. Genomic DNA was extracted with cetyl-trimethylammonium bromide, chloroform/isoamyl alcohol and isopropanol. Extracted DNA was then digested with the restriction endonuclease, PvuII. The resulting fragments were electrophoretically separated and transferred onto nylon membranes. The DNA fragments were then probed with a 245 bp fragment of the insertion sequence, IS6110. Hybridization was detected with chemiluminescence (ECL, Amersham). RFLP patterns were compared visually and with the assistance of a computerized system, the Bioimage Whole Band Analyzer (Millipore Corporation, Ann Arbor, MI). M. tuberculosis strains with identical RFLP patterns or with RFLP patterns which differed by only a single band, were considered to be the same strain [11, 12]. MDR M. tuberculosis isolates were considered to be laboratory contaminants if the patient had only one positive culture for M. tuberculosis and if the RFLP pattern of this isolate was identical to that of an isolate from another patient processed in the clinical laboratory on the same day. None of the patients with M. tuberculosis isolates meeting these criteria for laboratory contamination had acid fast bacilli on examination of their clinical specimen and none had clinical evidence of drug-resistant TB. In contrast, 92% (57/62) of patients whose M. tuberculosis isolates were not considered to be laboratory contaminants had at least two MDR isolates cultured on different days. Patients with MDR TB were considered to have initial MDR TB if they had no previous history of TB or if the M. tuberculosis isolates from their initial episode of TB were resistant to both isoniazid and rifampin. Patients with MDR TB who previously had positive cultures for M. tuberculosis which were susceptible to either isoniazid and/or rifampin were considered to have acquired MDR TB. Patients with MDR TB with a history of previously active TB, whose initial M. tuberculosis isolates were not submitted for susceptibility testing, were considered to have MDR TB of uncertain origin. Sequential M. tuberculosis isolates were available from several patients with MDR TB whose initial isolates were not MDR. If the RFLP pattern of two or more initial non-MDR isolates differed from the RFLP of two or more MDR isolates, the patient was considered to be exogenously reinfected with the MDR strain. This assessment is based on the well-documented stability of IS6110-based RFLP patterns over short periods of time (several months to several years) [13]. For purposes of analysis, patients with exogenous reinfection were considered to have initial MDR TB. Patients were considered to be infected with HIV if they had an AIDS-defining diagnosis, oral candidiasis, or a positive HIV serology by enzyme immunoassay and western blot. During the 1987/1988 study period, 178 hospitalized patients aged 18-65 with a new diagnosis of TB were offered HIV serological testing and T-lymphocyte subset measurements as part of a prospective HIV seroprevalence study [9]. During the 1990/1991 study period, HIV serological testing and T-lymphocyte subset measurements were offered to patients with TB at the discretion of the patient's hospital or clinic physicians, and were frequently not performed if a patient did not have a risk factor for HIV infection. Nonetheless, during the 1990/1991 study period, HIV status was determined on each of the patients with MDR TB. Results Demographic characteristics In the first and second study periods, there were 272 and 233 patients, respectively, with positive cultures for M. tuberculosis. The median ages of patients in both study periods were similar (35 vs 37 years-old); however, the proportion of patients less than 18 years-old increased from 2% to 7% between the two periods (P = .009; Table 1). At least 96% of the patients during both study periods were African American, Hispanic, or Haitian. Previous active tuberculosis (23% vs 30%; p=0.06) and underlying HIV infection (41% vs 50%; P = .10) were common during both study periods. Substance abuse, incarceration, and/or homelessness were present in 62% of patients in the 1987/1988 period and in 64% of patients in the 1990/1991 period (Table 1). Frequency of MDR TB Antituberculosis drug susceptibilities were available on M. tuberculosis isolates from 98% (267/272) of patients during the first study period and 95% (222/233) of patients during the second study period. The proportion of patients with MDR TB increased from 10% (27/267) to 17% (38/222) from the first to the second study periods (P = .03) (Table 1). Three patients had positive cultures for MDR TB at KCH during both periods. Ninety-two percent (57/62) of the patients with MDR TB had at least two MDR M. tuberculosis isolates cultured on different days. Patients A16, A23, A27, B9, and B37 each had only one positive culture for M. tuberculosis but there was no evidence that their positive culture resulted from laboratory cross contamination. Previous drug susceptibilities in patients with MDR TB To classify the 62 patients with MDR TB as having either initial or acquired MDR TB, their TB history and the drug-resistance patterns of their infecting strains were analyzed (Figure 1). From this analysis, five patient categories were identified: (1) 26 patients had no previous episodes of TB (initial MDR TB); (2) 13 patients had a history of TB before their presentation during the study (median of 14 months; range, 5 - 108 months), but their initial M. tuberculosis isolate was already MDR (initial MDR TB); and (3) 4 patients (B15, B21, B22, B23) were shown to be exogenously reinfected with an MDR TB strain and thus were also considered to have initial MDR. The courses of three of these patients have recently been described [13]. (4) In contrast, 16 patients with MDR TB had a previous history of TB (median of 36 months; range, 2-180 months) with an isolate that was either fully susceptible or resistant to only a single drug and they were classified as having acquired MDR TB. (5) Finally, for three patients (A18, B4, B9), initial drug susceptibilities were not available and these patients were considered to have MDR TB of uncertain origin. In summary, 43 patients (69%) had initial MDR TB, 16 patients (26%) had acquired MDR TB, and 3 patients (5%) had MDR TB of uncertain origin. MDR TB and HIV infection The possible contribution of HIV infection to the increase in MDR TB between the first and second study periods was assessed by correlating the HIV infection status of the MDR TB patients with their time of presentation. During the 1987/1988 period, three patients (11%) with MDR TB were HIV-infected; 16 patients (59%) were HIV-seronegative; and eight patients (30%) could not be assessed for the presence of HIV infection (Table 2, Figure 1A). During the 1990/1991 period, 22 (58%) of the MDR TB patients were HIV-infected and 16 (42%) patients were HIV-seronegative (Table 2, Figure 1B). Therefore, among patients of known HIV status, the proportion of patients with MDR TB who were HIV-infected increased from 16% (3/19) to 58% (22/38) (P = .006). Three of the eight patients with MDR TB during the 1987/1988 period who could not be assessed for the presence of HIV infection had a risk factor for HIV infection. Each of the remaining five unassessable patients had a risk factor for TB (i.e. alcohol abuse, malignancy, or diabetes) and were followed for a median of six years without developing signs of HIV infection. Among the 24 HIV-infected patients with MDR TB, 21 (88%) had initial MDR and 3 (12%) had acquired MDR. Among the 30 HIV-seronegative patients with MDR TB, 17 patients (57%) had initial MDR, 11 patients (37%) had acquired MDR, and two patients had MDR of uncertain origin. HIV-infected MDR TB patients were therefore more likely than HIV-seronegative MDR TB patients to have initial MDR TB (21/24 (88%) vs 17/30 (57%); P = .03). To determine if patient HIV status correlated with characteristics associated with the infectiousness of MDR TB, we assessed the duration of patient survival with MDR TB, the proportion of patients with positive sputum acid-fast smears, and the proportion of patients with pulmonary cavitation on chest radiograph. HIV-seronegative patients were more likely than HIV-infected patients to survive their initial hospitalization with MDR TB (29/30 (97%) vs 13/24 (54%); P < .001). A higher proportion of HIV-seronegative patients had positive acid fast sputum smears (24/30 (80%) vs 14/24 (58%)), but this difference was not significant (P = .15). However, HIV-seronegative patients were more likely than HIV-infected patients to have cavitary lung disease on chest radiograph (22/30 (73%) vs 6/24 (25%); P = .001). RFLP analysis of MDR TB isolates RFLP analysis was performed on M. tuberculosis isolates that were available from 56 of the 62 patients with MDR TB. Each of the resulting molecular fingerprints was then compared with the other 55 fingerprints and isolates with matched fingerprints were considered to comprise a cluster of epidemiologically associated strains. Twelve patients were infected with unique strains of M. tuberculosis and 44 patients were infected with M. tuberculosis strains belonging to one of six clusters. The largest cluster consisted of 26 patients infected with a strain designated "cluster strain #1"; 18 other patients were infected with one of five other strains ("cluster strains #2-6") (Figure 1). A computerized simulation of each of the RFLP patterns is shown in Figure 2. Five of the six cluster strains were present among patients during both study periods. Among patients presenting during the 1990/1991 study period from whom M. tuberculosis isolates were available for RFLP analysis, 75% of both HIV-infected (15/20) and HIV-seronegative (12/16) patients were infected by strains of M. tuberculosis isolated earlier from HIV-seronegative patients. Specifically, 14 HIV-infected patients presenting during the 1990/1991 period were infected by M. tuberculosis strains that had been isolated from HIV-seronegative patients during the 1987/1988 study period; and, one HIV-infected patient (B34) had the same strain as an HIV-seronegative patient (B3) who also presented during the 1990/1991 study period, but who had acquired MDR during the preceding five years. Thus many of the same MDR strains of M. tuberculosis were first detected among HIV-seronegative patients during the 1987/1988 period and then among HIV-seronegative and HIV-infected patients during the 1990/1991 period. Epidemiologic links between patients with MDR TB Outbreak investigations were not undertaken because the RFLP clusters were recognized months to years after patients had been hospitalized. Nonetheless, in several instances, epidemiologic links between patients with the same strain of M. tuberculosis were apparent. Among the ten HIV-infected patients with "cluster strain #1" during the 1990/1991 study period (Figure 1b), eight had been hospitalized at KCH two to nine months before presenting with MDR TB. Seven of these patients had been in the same room or ward as at least one other patient with "cluster strain #1". Among the eight HIV-seronegative patients with "cluster strain #1" during the 1990/1991 study period (Figure 1B), seven had close community contact with each other: three patients lived in the same apartment (B5, B7, B11); three other patients (B2, B9, B13) were close friends and frequent visitors of the three apartment dwellers; and, one patient (B14) was the two month-old child of patient B13. Discussion Over the past two decades, fiscal constraints have led to cutbacks in many TB control programs [14, 15]. At the same time, the overlapping problems of HIV infection, homelessness, and substance abuse, have increased in populations disproportionately affected by TB, interacting to impede TB control in these groups. In some urban areas in the United States, fewer than 60% of patients complete a recommended course of anti-TB therapy [14, 15]. Incomplete treatment leads to continued TB transmission and an increased prevalence of drug resistant M. tuberculosis strains. In the present study, 10% of TB patients in the 1987/1988 period and 17% of TB patients in the 1990/1991 period had MDR TB. The 10% prevalence of MDR TB in 1987/1988 occurred when MDR TB was not yet recognized as a problem in New York City. The 17% prevalence of MDR TB in 1990/1991 is similar to the 19% prevalence reported in a laboratory-based survey in New York City in April 1991 [4]. Although, the cutoff for INH susceptibility was decreased from 0.2 ug/ml to 0.1 ug/ml between the two study periods, this is unlikely to be responsible for the increase in MDR TB since the 0.2 ug/ml cut-off is ³ 97% sensitive at detecting INH resistance [16] and since the changed cutoff does not explain the increase in rifampin resistance. The high rate of MDR TB during the first study period is not unexpected in light of the fact that drug-resistant TB had been endemic in Brooklyn for several years. In 1961-1980, 10% of children with TB at KCH were infected with INH-resistant strains of M. tuberculosis and, in 1981-1984, 3/19 children were infected with rifampin-resistant strains of M. tuberculosis [17, 18]. DNA RFLP analysis and conventional epidemiology support the hypothesis that most cases of MDR TB, in the present study, resulted from the transmission of M. tuberculosis strains that were already resistant to both isoniazid and rifampin. Among the 56 patients whose M. tuberculosis isolates were submitted for RFLP analysis, only 18 different strains were detected. Twelve strains infected individual patients whereas six strains were responsible for infection in the remaining 44 patients. Moreover, nearly 70% of patients with MDR TB were already infected with MDR strains of M. tuberculosis at their initial presentation with TB. These findings are therefore consistent with those of the recent city-wide survey in which more than 60% of New York City TB patients in April 1991 with drug-resistant TB had drug resistance at the time of their initial episode of TB [4]. DNA RFLP analysis and conventional epidemiology also suggest that most of the MDR TB among HIV-infected patients resulted from resistance that developed among HIV-seronegative patients. First, HIV-seronegative patients were more likely than HIV-infected patients to have a past history of TB and to have acquired drug resistance. Second, among the 20 HIV-infected patients during the 1990/1991 study period whose M. tuberculosis isolates were analyzed by DNA fingerprinting, 15 were infected by strains found earlier among HIV-seronegative patients. A greater likelihood for HIV-seronegative patients to develop acquired MDR TB and to transmit infection may be related to their increased frequency of cavitary pulmonary disease and to their increased probability of surviving inadequate therapy. Indeed, other studies have shown that cavitary pulmonary disease occurs more commonly among HIV-seronegative patients with TB [19] and is a risk factor for both the acquisition of drug resistance and for TB transmission [20, 21]. The environments in which MDR TB transmission occurred could not be determined for most patients, however, both community and nosocomial transmission were implicated. Seven HIV-seronegative patients infected with the "cluster strain #1" during the 1990/1991 study period had close community contact, and, as in previously reported outbreaks of MDR TB among patients without HIV infection, the development of clinical TB in these patients was staggered over several years [22, 23]. Seven HIV-infected patients during the 1990/1991 study period who were also infected with "cluster strain #1" appeared to be infected nosocomially. As in previously reported outbreaks of MDR TB among HIV-infected patients, each of these patients became ill within several weeks to months following exposure to a patient with the same M. tuberculosis strain [6, 7]. Drug resistant TB and increasing HIV prevalence are problems in many developing countries and in some urban areas in the United States. This may lead to community and institutional settings in which patients with drug-resistant TB will be in close contact with immunocompromised individuals. The dramatic increase in disease that may result from the transmission of MDR TB from immunocompetent to immunocompromised persons highlights the urgency of instituting effective TB control in populations with increasing HIV prevalence. Recent recommendations to expand the use of directly observed treatment for TB are designed to prevent patients with drug-susceptible TB from developing drug-resistant TB [24, 25]. In addition, the transmission patterns we have described underscore the need for the complete and effective treatment of all cases of TB, especially MDR TB. References 1. Centers for Disease Control. National action plan to combat multidrug-resistant tuberculosis. MMWR Morb Mortal Wkly Rep 1992;41(RR-11):5-8. 2. Goble M, Iseman MD, Madsen LA, Waite D, Ackerson L, Horsburgh R. Treatment of 171 patients with pulmonary tuberculosis resistant to isoniazid and rifampin. N Engl J Med 1993;328:527-32. 3. Bloch AB, Cauthen GM, Onorato IM, et al. Nationwide survey of drug-resistant tuberculosis in the United States. JAMA 1994; 271:665-71. 4. Frieden TR, Sterling T, Pablos-Mendez A, Kilburn JO, Cauthen GM, Dooley SW. The emergence of drug-resistant tuberculosis in New York City. N Engl J Med 1993;328:521-6. 5. Mahmoudi A, Iseman MD. Pitfalls in the care of patients with tuberculosis: common errors and their association with drug resistance. JAMA 1993;270:65-8. 6. Edlin BR, Tokars JI, Grieco MH, et al. An outbreak of multidrug- resistant tuberculosis among hospitalized patients with the acquired immunodeficiency syndrome. N Engl J Med 1992;326:1514-21. 7. Dooley SW, Jarvis WR, Martone WJ, Snider DE. Multidrug- resistant tuberculosis [editorial]. Ann Intern Med 1992;117:257-9. 8. Bureau of Tuberculosis Control, New York City Department of Health. Tuberculosis in New York City, 1991: information summary. 1992. 9. Shafer RW, Chirgwin KD, Glatt AE, Dahdouh MA, Landesman SH, Suster B. HIV prevalence, immunosuppression, and drug resistance in patients with tuberculosis in an area endemic for AIDS. AIDS 1991;5:399-405. 10. van Embden JDA, Cave D, Crawford JT, et al. Strain identification of Mycobacterium tuberculosis by DNA fingerprinting: Recommendations for a standardized methodology. J Clin Microbiol 1993;31:406-9. 11. Small PM, Hopewell PC, Singh SP, et al. The epidemiology of tuberculosis in San Francisco: A population-based study using conventional and molecular methods. N Engl J Med 1994;330:1703-9. 12. Alland D, Kalkut G, Moss AR, et al Transmission of tuberculosis in New York City: An analysis by DNA fingerprinting and conventional epidemiological methods. N Engl J Med 1994;330:1710-6. 13. Small PM, Shafer RW, Hopewell PC, et al. Exogenous reinfection with multidrug-resistant tuberculosis in patients with advanced HIV infection. N Engl J Med 1993;328:1137-44. 14. Bloom BR, Murray CJL. Tuberculosis: commentary on a reemergent killer [editorial]. Science 1992;146:1623-33. 15. Brudney K, Dobkin J. Resurgent tuberculosis in New York City. Human immunodeficiency virus, homelessness, and the decline of tuberculosis control programs. Am Rev Respir Dis 1991;144:745-9. 16. Siddiqi SH, Hawkins JE, Laszlo A. Interlaboratory drug susceptibility testing of Mycobacterium tuberculosis by a radiometric procedure and two conventional methods. J Clin Microbiol 1985;22:919-23. 17. Steiner P, Rao M, Victoria MS, Hunt J, Steiner M. A continuing study of primary drug-resistant tuberculosis among children observed at the Kings County Hospital Center between 1961 and 1980. Am Rev Respir Dis 1983;128:425-8. 18. Steiner P, Rao M, Mitchell M, Steiner M. Primary drug-resistant tuberculosis in children: Emergence of primary drug-resistant strains of M. tuberculosis to rifampin. Am Rev Respir Dis 1986;134:446-8. 19. Pitchenik AE, Rubison HA. The radiographic appearance of tuberculosis in patients with the acquired immunodeficiency syndrome (AIDS) and pre-AIDS. Am Rev Respir Dis 1985;131:393-6. 20. American Thoracic Society. Control of tuberculosis in the United States. Am Rev Respir Dis 1992;146:1623-33. 21. Iseman MD. Treatment of multidrug-resistant tuberculosis. N Engl J Med 1993;329:784-91. 22. Centers for Disease Control. Outbreak of multidrug-resistant tuberculosis - Texas, California, and Pennsylvania. MMWR 1990;39:369-72. 23. Villarino ME, Geiter LJ, Simone PM. The multidrug-resistant tuberculosis challenge to public health efforts to control tuberculosis. Public Health Rep 1992;107:616-25. 24. Centers for Disease Control. Initial therapy for tuberculosis in the era of multidrug resistance: recommendations of the advisory council for the elimination of tuberculosis. MMWR Morb Mortal Wkly Rep 1993;42(RR-7). 25. Weis SE, Slocum PC, Blais FX, et al. The effect of directly observed therapy on the rates of drug resistance and relapse in tuberculosis. N Engl J Med 1994;330:1179-84. Legend for Figure 1 Duration of tuberculosis (TB), previous drug susceptibilities, and restriction-fragment-length-polymorphism (RFLP) patterns of patients with multidrug-resistant (MDR) TB between October 1987 - June 1988 (Figure 1A) and October 1990 - June 1991 (Figure 1B) at Kings County Hospital (KCH). Patients are grouped by human immunodeficiency virus (HIV) status and each bar represents the course of one patient. The rectangle fill pattern reflects the drug-susceptibility pattern of patients' Mycobacterium tuberculosis isolates. Patients with rectangles containing a "?" did not have drug susceptibilities at the time of their earliest positive culture for M. tuberculosis. The RFLP pattern column contains one of six symbols for patients with M. tuberculosis strains isolated from a cluster of patients, the designation "unique" for patients with strains having unique RFLP patterns, and the designation "NA" for patients with strains not available for RFLP analysis. Patients A5/B1, A15/B6, and A25/B17 had positive cultures for MDR TB at KCH during both study periods. Legend for Figure 2 Computerized simulations of the restriction-fragment-length polymorphism patterns of the six Mycobacterium tuberculosis strains isolated from patient clusters. The molecular weight standards are shown on the left and the Kings County Hospital (KCH) cluster strain designation is shown above each of the simulations. Three of the 26 patients infected with KCH cluster strain #1 had strains with a single additional band of either 2.2 or 2.3 kb. Table 1. Demographic Characteristics, TB History, Human Immunodeficiency Virus (HIV) Status and Drug Susceptibilities of TB Patients During Two Nine-Month Study Periods _____________________________________________________________________ Oct 87 - June 88 Oct 90 - June 91 P (N = 272) (N = 233) (P < .05) % % _____________________________________________________________________ Age <18: 2 7 .009 18-49: 86 78 >49: 12 15 Sex Male: 75 67 Female: 25 33 Race African American: 72 67 Hispanic: 14 18 Haitian: 11 11 White: 2 3 Asian: 1 1 Social History Substance abuse: 51 62 .02 Prison inmate: 13 19 Homeless: 8 16 .004 None of the above: 38 36 HIV Infection* HIV-infected: 41 50 HIV-seronegative: 27 16 Unknown: 32 34 Past TB History 23 30 Drug resistance† ³ 1 drug: 22 28 ³isoniazid + rifampin: 10 17 .03 Total isoniazid: 17 25 .03 Total rifampin: 12 18 Total pyrazinamide: NT 14 Total streptomycin: 7 4 Total ethambutol: 2 3 _________________________________________________________________________ * During the 1987/1988 period, an HIV-seroprevalence study was conducted in which hospitalized TB patients were offered HIV serological testing. During the 1990/1991 study period, serological testing was performed at the discretion of the patients' physicians and was frequently not performed if a patient did not have a risk factor for HIV infection. † Drug susceptibility results were available on 267 patients during the 1st study period and 222 patients during the 2nd study period. Table 2. Comparison of Patients with Multidrug-Resistant (MDR) TB During the First and Second Nine Month Study Periods _________________________________________________________________________ Oct 87 - June 88 Oct 90 - June 91 P (N = 27) (N = 38) (P < .05) % % _________________________________________________________________________ Age <18: 3 3 18-49: 67 92 .02 >49: 30 5 Sex Male: 67 68 Female: 33 32 Social History Substance abuse: 59 68 alcohol 59 45 IV drugs 19 24 cocaine 11 32 Prison inmate: 0 13 Homeless: 7 16 None of the above: 41 24 Human immunodeficiency virus (HIV) HIV-infected: 11 58 .006* HIV-seronegative: 59 42 Unknown: 30 0 Development of Resistance Initial MDR: 63 74 Acquired MDR: 33 21 Unknown: 4 5 ________________________________________________________________________ * The eight patients with unknown HIV status were not included in this comparison. Three of these eight patients had risk factors for HIV infection. Each of the remaining five unassessable patients had a risk factor for TB (i.e. alcohol abuse, malignancy, or diabetes) and were followed for a median of six years without developing signs of HIV infection.