Epidemiology & Social
Predictors of tuberculosis transmission in prisons: an analysis using conventional and molecular methods
March, Francescaa,b; Coll, Perea,b,c; Guerrero, Rafael A.b,d; Busquets, Estherd; Caylà, Joan A.b,e; Prats, Guillema,c
From the aDepartment of Microbiology, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain; bTuberculosis Investigation Unit of Barcelona, Barcelona, Spain; cDepartment of Genetics and Microbiology, Universitat Autònoma de Barcelona, Barcelona, Spain; dHealth Program, Department of Justice of Catalonia, Barcelona, Spain; eEpidemiology Service, Municipal Institute of Health, Barcelona, Spain; fMedical Services, Penitentiary Centers, Barcelona, Spain; gMedical Service, Hospital Penitentiary Center, Barcelona, Spain; hTuberculosis Program, Department of Health of Catalonia, Barcelona, Spain; and iGeneral Lab Laboratory, Barcelona, Spain.
Xavier Garrigaab, Antonio Da Silvaf, Nuria Fauriaf, M. Jesus Lealf, Enric Mercadéf, Cristina Nietof, N. Teixidóf, Inmaculada Vallsf, Josep Gualg, José Alcaideh, Gisela Schmidtg and Nuria Miserachsi
Sponsorship: This work was supported by a grant from the Fondo de Investigación Sanitaria (97-0041).
Correspondence to: Pere Coll, Servei de Microbiologia, Hospital Santa Creu i Sant Pau, Avenida Sant Antoni Maria Claret 167, 08025 Barcelona, Spain. Tel: +34 93 291 90 71; e-mail: firstname.lastname@example.org
Received: 10 December 1999; accepted: 22 December 1999.
Objective: To determine the tuberculosis (TB) transmission patterns within the prison system in Catalonia, conventional epidemiological techniques were combined with DNA fingerprinting of Mycobacterium tuberculosis.
Methods: IS 6110- and polymorphic GC-rich repeat sequence (PGRS)-based restriction fragment length polymorphism (RFLP) were combined with epidemiological studies to assess the relatedness of isolates from all patients with confirmed TB at five prisons in the province of Barcelona (Catalonia, Spain), between 1 July 1994 and 31 December 1996. Risk factors for transmission were analysed to a logistic regression. The extent of drug-resistant TB was also assessed.
Results: The incidence of TB during the study period was 2775 cases per 100 000 inmate years. Of the 247 culture-positive cases, 126 (51%) appeared to have active TB as a result of recent transmission. Using conventional epidemiological methods, 14 active chains of transmission were identified in prison involving 65 isolates (52% of clustered patients). A lengthy history of imprisonment [odds ratio (OR) 2.8, 95% confidence interval (CI) 1.52–5.11] and pulmonary TB (OR 2.36, 95% CI 1.17–4.75) were independently associated with clustering. Low rates of both initial (2.9%) and acquired drug resistance (5.8%) were identified and there was no evidence of the transmission of drug-resistant TB.
Conclusion: In the prison system studied, the recent transmission of TB contributes substantially to the overall incidence of the disease. Both lengthy incarcerations and delays in identifying inmates with pulmonary symptoms play a key role in this recent transmission. Directly observed therapy (DOT) is a critical control strategy for reducing the emergence of drug resistance and for avoiding the transmission of resistant organisms.
Prison inmates have a greater risk of contracting both Mycobacterium tuberculosis infection and active tuberculosis (TB) disease [1–3]. A disproportionate number of inmates have social and clinical risk factors for exposure to the organism or, once infected, for the development of active disease. These risk factors include HIV infection [4–6], substance abuse and the membership of lower socioeconomic groups [7–9]. A high prevalence of TB infection on admission to prison has thus been described . In addition, the crowded living conditions in prisons are often conducive to TB transmission [11–13]. In such a context, higher incidences of TB in prisons are not unexpected, and it is essential to delineate the patterns by which disease spreads in the population so as to ensure that disease control measures are accurately targeted.
Several outbreaks of TB have been recorded in prisons [12–14], some documenting spread outside of the prison system. The main factors contributing to these outbreaks include delays in identifying inmates who are suspected of having TB, the frequent transfers of inmates, and lapses both in infection control procedures and in therapy. Outbreaks of drug-resistant TB [15,16] emphasize the urgent need for the effective control of TB in prisons because it represents a health problem both in the penitentiary system and the community into which inmates are released [9,13,17].
IS 6110 DNA restriction fragment length polymorphism (RFLP), the most common DNA fingerprinting method used, has improved our understanding of the epidemiology of TB, facilitating the identification of previously unsuspected chains of transmission [18–22]. This technique, combined with detailed epidemiological study, is a very useful genetic marker for answering key epidemiological questions in the context of populations (and control programmes) in which TB is most prevalent .
The aim of this study was to evaluate the relative contribution of recent transmission to incident cases of TB in the penitentiary system of Barcelona (Catalonia, Spain) over a 30 month period (1994–1996), identifying the risk factors associated with transmission. Systematic IS 6110- and polymorphic GC-rich repeat sequence (PGRS)-based DNA fingerprinting and univariate and multivariate analyses were used. To validate the use of these techniques in the molecular epidemiology of TB, the results of DNA fingerprinting were compared with epidemiological data. Susceptibility against first-line drugs was also studied.
Materials and methods
The study was carried out at the five main centres of Catalonia (Spain), all of which are located in the province of Barcelona (capacity for 3927 inmates). Most inmates are from Barcelona and the surrounding areas.
The public health authorities in Catalonia instituted the following essential TB control measures throughout the correctional health care system in 1992: (i) systematic tuberculin skin test screening and roentgenograms in all new inmates on entry and in all inmates with symptoms suggesting TB (e.g. cough, weight loss, fever); (ii) sputum smear and culture examinations in all inmates with abnormal chest radiographs; inmates with known or suspected HIV infection (including those with normal chest radiographs) and symptoms compatible with TB routinely receive a bacteriological analysis as part of the initial screening; and (iii) directly observed therapy (DOT) for the treatment of active TB cases with a regimen of isoniazid, rifampin, pyrazinamide and ethambutol for 6 months in HIV-negative patients and 12 months in HIV-infected patients. The proportion of patients receiving DOT has increased substantially since 1992, being extended to all new cases in 1994.
Between 1 July 1994 and 31 December 1996 inmates confirmed as having TB (i.e. those who met the Centers for Disease Control and Prevention case definition for TB)  while in prison or on entry to the penitentiary were included in the study. For all patients who had at least one positive culture of M. tuberculosis, one initial isolate was selected for further analysis.
The incidence of TB was calculated using the total number of days inmates were incarcerated during the study period, and is expressed as cases per inmate-years. Only cases of TB newly diagnosed while the patient was an inmate were included.
Antimicrobial susceptibility testing
Susceptibility testing against first-line drugs was carried out using the BACTEC radiometric method (Bactec System, Becton Dickinson, Sparks, MD, USA) and the drug resistances were confirmed by the modified-proportion-agar-dilution technique . Rates of primary and acquired resistance were calculated according to the WHO recommended definitions .
Restriction fragment length polymorphism analysis
The IS 6110 RFLP patterns were determined in accordance with internationally standardized procedures . The resulting fragments, once hybridized with a 245 base-pair fragment of the insertion sequence IS 6110, were detected by chemiluminescence (ECL kit, Amersham, Arlington Heights, IL, USA). Patients were grouped in the same cluster if their isolates were identical with respect to the number and size of hybridizing bands. Isolates with fingerprints that differed by only one or two bands were considered as being highly related but non-identical.
In order to confirm the IS 6110 fingerprinting results, all isolates that had either identical or highly related patterns were analysed with a second genetic marker based on PGRS . This supplementary analysis was performed in accordance with described procedures , and DNA was restricted with Alu I. The labelling of the probe, hybridization and identification of clusters were carried out as described previously for IS 6110.
Fingerprint patterns of the isolates were analysed and compared by computer-assisted analyses using Whole Band Analyzer software (Bioimage, Ann Arbor, MI, USA). The relationships between isolates were assessed by the unweighted pair group method of averages  using the Dice coefficient.
For patients whose isolates were within IS 6110-defined clusters, as well as those whose isolates had highly related patterns, information concerning their transfer histories, cell locations and prison activities was obtained from the Tuberculosis Prevention and Control Program database at the Department of Correctional Services. From January 1994 to TB diagnosis, a graphic representation detailing the precise location (prison, cell block, activities) within the penitentiary system was used to identify the epidemiological links between these patients.
As a cell block houses persons who sleep, live, work or otherwise share a common ventilation system, those inmates occupying a cell block with a prisoner with TB may be considered to be at the greatest risk of infection. Therefore, it was assumed that there was clear evidence of a transmission link if a patient potentially susceptible to infection (i.e. any potential contact) was living in the same cell block at a time when one or more of the other patients included in the same IS 6110-defined cluster were potentially infectious.
For study purposes, a case of pulmonary TB was considered potentially infectious from the date of the onset of symptoms until they were identified and isolated (e.g. the date of TB diagnosis) . In addition, inmates were considered to be susceptible to infection 30 days or more before the onset of disease .
Chi-square tests were performed to test the association of clustering with categorical predictor variables. Distributions of the continuous variables were compared by t-tests. For the multivariate analysis, variables showing univariate significance levels of 0.15 or less were included in a logistic regression model to determine independent risk factors for clustering.
The demographic characteristics of the five prisons in the province of Barcelona are shown in Table 1. Two of them (C and E) were operating according to their rated capacities. The other prisons, A, B and D, were operating at 67, 37 and 17% above design capacity, respectively. The monthly turnover for prisons A, B and C was greater than for prisons D and E.
During the study period, there were 416 cases of TB, 61 (15%) of which were undergoing chemotherapy on admission. The incidence of TB was 2775 cases per 100 000 inmate years. Of 355 patients who were diagnosed during incarceration, 267 (75%) were confirmed by the isolation of M. tuberculosis. Isolates were available for 93% (247 out of 267) of these culture-positive cases to allow for RFLP fingerprinting and susceptibility testing. The remaining 20 isolates (7%) were non-viable, contaminated or lost.
Of the 355 inmates diagnosed with TB while in prison, 217 (61%) were confined in the two prisons that only house sentenced inmates (D and E). A total of 273 (77%) patients were known to be dually infected with M. tuberculosis and HIV, and 288 (81%) had a history of injecting drug use, 72% being both HIV positive and injecting drug users (IDU).
As shown in Table 2, the proportions of patients with primary and acquired drug resistance were 2.9 and 5.8%, respectively. Four out of five strains with primary drug resistance (all of them isolated from HIV-positive patients) were resistant to only one antituberculous drug; only one strain from an HIV-negative patient displayed resistance to three drugs. Resistance to isoniazid was seen in all three cases of acquired resistance, associated with rifampin resistance in two.
DNA fingerprinting analysis
Of the 247 culture-positive cases, 121 (49%) had strains of M. tuberculosis with unique IS 6110 RFLP fingerprints and 126 (51%) had strains whose RFLP patterns were identical to those of one to 25 isolates from other patients, which were classified into 22 clusters (Fig. 1, Table 3). The strains with identical fingerprints were isolated at different times (interval of several days), so cross-contamination is highly unlikely. Nine other isolates had highly related but non-identical IS 6110 fingerprint patterns. They differed in the presence or position of a single band (three isolates) or two bands (six isolates). The vast majority (96%) of the strains had more than six copies (range two to 19).
None of these 135 clustered or highly related strains had unique banding PGRS patterns (Table 3). In contrast, the performance of this supplementary analysis reduced the number of clusters from 22 to 14. Isolates included in IS 6110-defined clusters with more than 65% of similarity were more likely to have identical PGRS patterns.
A two-dimensional matrix of pattern similarities (Fig. 1) shows the degree of relatedness of each IS 6110 banding pattern with each other; most isolates (66%) could be grouped into three families with highly related IS 6110 banding patterns, comprising 95, 38 and 31 strains, respectively.
Identification of risk factors for clustering
The risk factors for clustering are shown in Table 4. The patients with clustered isolates were more likely to have been diagnosed at prisons D and E. Lengthy incarcerations and a history of imprisonment in the same centre in the year before TB diagnosis were also significantly associated with having a clustered isolate. Conversely, inmate movement, either between different prisons or between prison and the community, was not associated with clustering.
In the multivariate model, the odds of clustering were significantly increased among patients who had lived in prison for more than 1 year before TB diagnosis, compared with those patients who had spent less than 1 year there (adjusted OR 2.8, 95% CI 1.52–5.11). Having a pulmonary form of disease was also significant and was an independent risk factor for clustering (OR 2.36, 95% CI 1.17–4.75).
Analysis of clustered cases according to epidemiological data
Twelve IS 6110-RFLP patterns involving 65 (52%) clustered isolates were identified as being related to 14 active chains of TB transmission in prison (Fig. 2). Only one of the nine patients who had isolates with highly related patterns had a clear epidemiological connection to their IS 6110 cluster. Delays in identifying and isolating inmates who were symptomatic or suspected of having TB were identified as the source of all these transmission chains. For example, the apparent index patient B in cluster 1 was an HIV-negative man who had had active TB with cough for more than 3 months. For 84 patients not in clusters, the median delay in the suspicion of TB (time from the onset of symptoms to treatment initiation) was lower (27.2 days, SD = 22.6) than for 11 source cases (66.2 days, SD = 76.6), but there was no significant difference (P = 0.12). However, the available sample size was small. No other factors contributing to this transmission, such as transfers of inmates and lapses in therapy, were identified during the study period.
No evidence of contact in the same cell block during the study period was identified for the remaining 61 clustered patients. However, 12 of them (20%) had been held in the same cell block as another patient with identical RFLP pattern in late 1993 and early 1994, suggesting that exposure to active TB occurred before the study. In addition, although no common work or social activity within a prison could be identified, a further eight (13%) of these 61 patients were incarcerated in the same prison as another patient with potentially infectious TB and the same RFLP pattern (Fig. 2). Overall, the percentage of patients in clusters having an epidemiological connection increased from 52% (65 out of 126) to 67% (85 out of 126) when possibly related cases were added to those with clear epidemiological connections.
If one eliminates the source case in each transmission chain, 51 patients were epidemiologically linked to another in their respective clusters. In terms of penal category, the prisons housing sentenced inmates (D and E) were responsible for approximately 61% of active transmission (45 and 16%, respectively). The remaining 39% of this transmission was documented to have occurred in prison A; however, most of these cases (60%) were transferred to prisons housing sentenced inmates before developing active TB.
A longer survey time was associated with a lower number of cases without epidemiological connections. As shown in Fig. 3, no epidemiological links could be identified among 65% (20 out of 31), 68% (32 out of 47) and 48% (23 out of 48) of clustered cases diagnosed in 1994, 1995 and 1996, respectively.
For the clustered patients who showed no evidence of active transmission during the study period, the median time spent in prison in the 2 years before diagnosis was not significantly lower than that in the group of patients included in transmission chains (618 versus 717 days, P = 0.62). In addition, all of these patients (with the exception of three cases) had been incarcerated before July 1994, when it was not possible to perform RFLP analysis and to determine who had transmitted to who in each case. With regard to the possible exposure to TB within the community during the survey period, 68% of clustered patients without epidemiological connections were released on one or more occasions, as opposed to 57% of clustered patients included in transmission chains (P = 0.28). The remaining patients spent all this time in prison.
Prisons remain a reservoir of TB for society as a whole, with high prevalences of infection with M. tuberculosis and active TB among inmates [1–3,10,11,16]. This situation appears to create a health problem for the institutions, as well as for the community into which inmates are released [9,13,17]. In the prison system studied, the incidence of TB was very high (2775 per 100 000), almost 50 times the incidence in Barcelona for the general population (55.7 per 100 000) . These results support the idea that the over-representation per se of multiple high-risk groups is a significant factor in explaining this markedly higher incidence [8,17]. Inmates diagnosed with TB were predominantly IDU (81%), HIV positive (77%), and both IDU and HIV positive (72%); not surprisingly, a large number of inmate cases (15%) were diagnosed before their incarceration. In 1989–1990, of 729 prisoners with a previous history of prison examined in a Barcelona penitentiary, 56% were tuberculin positive at the time of admission .
Several TB outbreaks involving both inmates and employees have been recognized in correctional facilities [12–16]. After applying IS 6110 DNA fingerprinting systematically to identify interstrain relations combined with conventional epidemiological data, 14 unsuspected active chains of TB transmission were found involving 65 patients (26% of the total culture-positive cases diagnosed while in prison); of which 82% were HIV positive. However, it is not yet possible to determine the full extent of this prison-associated transmission because the case patients' incubation periods were artificially limited by the time of observation.
For nearly half of those inmates included in transmission chains, active transmission was documented to have occurred at a prison that was operating at 17% above design capacity, and where a large proportion of inmates are convicted to terms of more than 1 year. A further 39% of cases occurred at a prison that was characterized both by crowding (occupancy index, 1.67) and an extremely mobile population. Delayed diagnosis of active pulmonary TB cases (source cases) resulted in the transmission of M. tuberculosis. In contrast to a comparable study , lapses in treatment were not detected as another factor contributing to transmission because all patients with TB received DOT.
It has long been assumed that clonality defined by IS 6110 fingerprints implies recent transmission in thorough population-based studies [18,19,22,34]. However, recent reports [35–37] have suggested caution in assuming this premise, because a high percentage of transmission links predicted by molecular techniques were not detected by conventional epidemiological investigation, particularly among isolates with fewer than five copies of IS 6110. Although clear epidemiological connections were only found for 52% of clustered patients, the high number of hybridizing bands, the characteristics of the population (crowded and shifting, with a constant introduction of new strains), and the use of a second fingerprinting technique strongly suggest that the extent of recent transmission may be similar to that predicted with IS 6110.
In this context, an estimation of overall recently transmitted infection (rather than demonstrated active transmission during the survey period) may be helpful in evaluating the true contribution of prison-associated transmission to the high incidence of TB in the penitentiary system. In this sense, it was found that recent transmission accounts for almost half of the incident cases in Barcelona prisons, because 51% of TB cases were caused by strains with identical IS 6110 RFLP fingerprints; a much lower value than that estimated in the prison population of Madrid (74%) . Moreover, 66% of the strains could be grouped into three genetically related families, suggesting that the majority of M. tuberculosis strains descended from a few recently expanded clones .
The genetic and epidemiological relationships among isolates having highly related but non-identical IS 6110 fingerprint patterns were also evaluated. As observed previously [39–41], with the exception of one isolate, all were found to have identical PGRS fingerprints, suggesting that highly related IS 6110 fingerprints may represent a single clone of M. tuberculosis. However, no evidence of epidemiological connections was found, so this assumption of clonality seems to be unclear, as demonstrated in a recent study .
None of the risk factors associated with clustering in previous community-based studies [18,19,22] were identified in this investigation, possibly because of the disproportionate number of high-risk group members in the penitentiary population. In contrast, the mean length of stay in prison before diagnosis was closely related with clustering, supporting the idea that lengthy incarceration increases the likelihood of infection [8,42]. In this sense, patients with clustered isolates were more likely to have been diagnosed at prisons housing convicted inmates, although the site of transmission did not necessarily involve these centres. On the other hand, although it has been suggested that frequent inmate transfers is conducive to TB transmission , in this study there was no association between clustering and inmate movement between prisons, partly because no transfers of source cases while symptomatic were documented. Moreover, inmates being transferred are more likely to be screened by medical staff, thus preventing delays in the identification of patients who were symptomatic or suspected to have TB.
A previous study  has shown a strong correlation in HIV-positive patients between pulmonary TB and a concurrent stay in prison, suggesting that rapid progression of exogenous infections in immunosuppressed patients may encourage the emergence of pulmonary forms of TB. Our finding that pulmonary TB is an independent risk factor for having a clustered isolate supports this hypothesis. It also agrees with the data from several outbreaks [5,16,44–46] involving HIV-infected persons who developed a pulmonary form of disease after exposure to the source case.
It has been hypothesized that the occurrence of transmission with minimal contact [47,48] is a major factor in explaining the low correlation between molecular and conventional epidemiological techniques. In this prison-based study, casual transmission during the survey period seems to have been unlikely among cases without identifiable epidemiological links because movements inside the prisons are controlled and contact between inmates from different cell blocks is extremely unlikely. Assuming that our inmate population move frequently between prisons and the community, with the equivalent of the entire population changing every 6 months, it is tempting to speculate that the excess of recently acquired infections and no definite recent contact would reflect the dimensions of the transmission that occurred in the inmates' home communities.
Conversely, in these data, several findings support the idea that prison is the most important locus of TB transmission among this population in the time preceding the survey period. First, no correlation was found between clustering and either the area of residence outside the prison or inmate movement between prison and the community. Second, the median time spent in prison by clustered patients with or without evidence of active transmission was similar, as was their possible exposure to TB in the community. In addition, 21% of clustered patients without identifiable epidemiological links lived in the same cell block as another patient with identical RFLP pattern before study, reflecting the possibility that exposure to active TB occurred there. Finally, there was a direct relationship between longer survey time and an increased number of patients with epidemiological links, suggesting that short periods of survey in population-based studies may partly explain the low correlation between the occurrence of identical IS 6110 patterns and classification into an active chain of transmission .
On the other hand, with the emergence of TB caused by multidrug-resistant (MDR) M. tuberculosis[49,50] and outbreaks of MDR TB in prisons [15,16], the potential for drug-resistant organisms to spread within prison has generated public and professional alarm. In this context, it was particularly reassuring to find low rates of both initial (2.9%) and acquired drug resistance (5.8%) in this study, as well as no evidence of transmission of drug-resistant TB. These lines of evidence run counter to the hypothesis that HIV infection per se is a risk factor for the development of drug resistance , and corroborate the efficacy of DOT, even in communities with high rates of HIV infection. This last finding accords with a recent experience in New York City, where the application of DOT has been associated with a 75% decrease in MDR TB . Whatever our results, one should not forget that the priority in high-risk groups such as prisoners (with high rates of TB and high rates of ongoing transmission) is to detect MDR-TB and to decide what drug combination is likely to be most effective. We suggest that systematic real-time susceptibility testing should be performed on the initial isolate from every patient, and the initial regimen including four drugs should be adjusted when susceptibility data are available.
This study has several implications for prison TB control. Despite the success of DOT, transmission plays an important role in the global incidence of TB within the prison system of Catalonia. Because this transmission is highly correlated with prison population density and the duration of incarceration, increased emphasis should be placed on the screening of long-term inmates for M. tuberculosis infection and disease, identifying contagious cases and candidates for preventative therapy. Periodical screening of inmates and correctional employees may be a cost-effective and strategically important TB control measure, meriting serious consideration to protect the community from further transmission of the disease. In addition, the monitoring of IS 6110 RFLP patterns appeared to be useful to learn more about the dynamics of transmission, and should be considered a high priority to evaluate the strength of TB control programmes in this high-risk population.
The authors are indebted to health workers of the prisons of the province of Barcelona, whose high quality of service and cooperation have made this work possible; and to Montse Garrigó, Carme Moreno and Elena Garcia of the Hospital Santa Creu i Sant Pau for diligent assistance with antimicrobial susceptibility testing.
1. Abeles H, Feibes H, Mandel E, Girard JA. The large city prison – a reservoir of tuberculosis.
Am Rev Respir Dis 1970, 101: 706 –709.
2. Anderson KM, Keith EP, Norsted SW. Tuberculosis screening in Washington state male correctional facilities.
Chest 1986, 89: 817 –822.
3. Coninx R, Eshaya-Chauvin B, Reyes H. Tuberculosis in prisons.
Lancet 1995, 346: 1238 –1239.
4. Barnes PF, Bloch AB, Davidson PT, Snider DE. Tuberculosis in patients with human immunodeficiency virus infection.
N Engl J Med 1991, 324: 1644 –1650.
5. Daley CL, Small PM, Schecter GF. et al
. An outbreak of tuberculosis with accelerated progression among persons infected with the human immunodeficiency virus.
N Engl J Med 1992, 326: 231 –235.
6. Schürmann D, Nightingale SD, Bergmann F, Ruf B. Tuberculosis and HIV infection: a review.
Infection 1997, 25: 274 –280.
7. Braun MM, Truman BI, Maguire B. et al
. Increasing incidence of tuberculosis in a prison inmate population: association with HIV infection.
JAMA 1989, 261: 393 –397.
8. Martin V, Alvarez-Guisasola F, Cayla JA, Alvarez JL. Predictive factors ofMycobacterium tuberculosisinfection and pulmonary tuberculosis in prisoners.
Int J Epidemiol 1995, 24: 630 –636.
9. Glaser JB, Greifinger RB. Correctional health care: a public health opportunity.
Ann Intern Med 1993, 118: 139 –145.
10. Martin V, Gonzalez P, Caylà JA. et al
. Case-finding of pulmonary tuberculosis on admission to a penitentiary centre.
Tubercl Lung Dis 1994, 74: 49 –53.
11. Chaves F, Dronda F, Cave MD. et al
. A longitudinal study of transmission of tuberculosis in a large prison population.
Am J Respir Crit Care Med 1997, 155: 719 –725.
12. King L, Geis G. Tuberculosis transmission in a large urban jail.
JAMA 1977, 237: 791 –792.
13. Stead WW. Undetected tuberculosis in prison: source of infection for community at large.
JAMA 1978, 240: 2544 –2547.
14. Centers for Disease Control. Tuberculosis transmission in a state correctional institution – California, 1990–1991.
MMWR 1992, 41: 927 –929.
15. Valway SE, Richards SB, Kovacovich J, Greifinger RB, Crawford JT, Dooley SW. Outbreak of multi-drug-resistant tuberculosis in a New York State prison, 1991.
Am J Epidemiol 1994, 140: 113 –122.
16. Valway SE, Greifinger RB, Papania M. et al
. Multidrug-resistant tuberculosis in the New York State Prison System, 1990–1991.
J Infect Dis 1994, 170: 151 –156.
17. Drobniewski F. Tuberculosis in prisons – forgotten plague.
Lancet 1995, 346: 948 –949.
18. 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 –1709.
19. Genewein A, Telenti A, Bernasconi C. et al
. Molecular approach to identifying route of transmission of tuberculosis in the community.
Lancet 1993, 342: 841 –844.
20. Friedman CR, Stoeckle MY, Kreiswirth BN. et al
. Transmission of multidrug-resistant tuberculosis in a large urban setting.
Am J Respir Crit Care Med 1995, 152: 355 –359.
21. Bifani PJ, Plikaytis BB, Kapur V. et al
. Origin and interstate spread of a New York City multidrug-resistantMycobacterium tuberculosisclone family.
JAMA 1996, 275: 452 –457.
22. Alland D, Kalkut GE, Moss AR. et al
. Transmission of tuberculosis in New York City.
:An analysis by DNA fingerprinting and conventional epidemiologic methods.
N Engl J Med 1994, 330: 1710 –1716.
23. Cohn DL, O'Brien RJ. The use of restriction fragment length polymorphism (RFLP) analysis for epidemiological studies of tuberculosis in developing countries.
Int J Tubercl Lung Dis 1998, 2: 16 –26.
24. Centers for Disease Control. Case definitions for public health surveillance.
MMWR 1990, 39: 39 –40.
25. Hacek D, Siddiqi SH. Antimicrobial susceptibility testing.
In:Clinical microbiology procedures handbook.
H. Isenberg, J. Hindler (editors). Washington, DC: American Society for Microbiology; 1992.
26. World Health Organization/IUATLD. Guidelines for surveillance of drug resistance in tuberculosis.
:WHO Geneva/International Union Against Tuberculosis and Lung Disease (IUATLD) Paris.
Int J Tubercl Lung Dis 1998, 2: 72 –89.
27. van Embden JDA, Cave MD, Crawford JT. et al
. Strain identification ofMycobacterium tuberculosisby DNA fingerprinting: recommendations for a standardized methodology.
J Clin Microbiol 1993, 31: 406 –409.
28. Ross BC, Raios K, Jackson K, Dwyer B. Molecular cloning of a highly repetitive DNA element fromMycobacterium tuberculosisand its use as an epidemiologic tool.
J Clin Microbiol 1992, 30: 942 –946.
29. Chaves F, Yang Z, El Hajj H. et al
. Usefulness of the secondary probe pTBN12 in DNA fingerprinting ofMycobacterium tuberculosis.
J Clin Microbiol 1996, 34: 1118 –1123.
30. Li WH. Simple method for constructing phylogenetic trees from distance matrices.
Proc Natl Acad Sci USA 1981, 78: 1085 –1089.
31. Zaza S, Blumberg HM, Beck-Sagué C. et al
. Nosocomial transmission ofMycobacterium tuberculosis: role of health care workers in outbreak propagation.
J Infect Dis 1995, 172: 1542 –1549.
32. Frieden TR, Woodley CL, Crawford JT, Lew D, Dooley SM. The molecular epidemiology of tuberculosis in New York City: the importance of nosocomial transmission and laboratory error.
Tubercl Lung Dis 1996, 77: 407 –413.
33. Caylà JA, Galdós-Tangüis H, Jansà JM, Garcia de Olalla P, Brugal T, Pañella H. Evolución de la tuberculosis en Barcelona (1987–1995).
:Influencia del virus de la inmunodeficiencia humana y de las medidas de control.
Med Clin (Barc) 1998, 111: 608 –615.
34. Yang ZH, de Haas PEW, Wachmann CH, van Soolingen D, van Embden JDA, Andersen AB. Molecular epidemiology of tuberculosis in Denmark in 1992.
J Clin Microbiol 1995, 33: 2077 –2081.
35. Warren R, Richardson M, Sampson S. et al
. Genotyping ofMycobacterium tuberculosiswith additional markers enhances accuracy in epidemiological studies.
Clin Microbiol 1996, 34: 2219 –2224.
36. Braden CR, Templeton GL, Cave MD. et al
. Interpretation of restriction fragment length polymorphism analysis ofMycobacterium tuberculosisisolates from a state with a large rural population.
J Infect Dis 1997, 175: 1446 –1452.
37. Burman WJ, Reves RR, Hawkes AP. et al
. DNA fingerprinting with two probes decreases clustering ofMycobacterium tuberculosis.
Am J Respir Crit Care Med 1997, 155: 1140 –1146.
38. Hermans PWM, Messadi F, Guebrexabher H. et al
. Analysis of the population structure ofMycobacterium tuberculosisin Ethiopia, Tunisia and the Netherlands: usefulness of DNA typing for global tuberculosis epidemiology.
J Infect Dis 1995, 171: 1504 –1513.
39. van Soolingen D, Hermans PWM, de Haas PEW, Soll DR, van Embden JDA. Occurrence and stability of insertion sequences in Mycobacterium tuberculosis complex strains: evaluation of insertion sequence-dependent DNA polymorphism as a tool in the epidemiology of tuberculosis.
J Clin Microbiol 1991, 29: 2578 –2586.
40. Dwyer B, Jackson K, Raios K, Sievers A, Wilshire E, Ross B. DNA restriction fragment analysis to define an extended cluster of tuberculosis in homeless men and their associates.
J Infect Dis 1993, 167: 490 –494.
41. Cave MD, Eisenach KD, Templeton G. et al
. Stability of DNA fingerprint pattern produced with IS6110in strains ofMycobacterium tuberculosis.
J Clin Microbiol 1994, 32: 262 –266.
42. Bellin EY, Fletcher DD, Safyer SM. Association of tuberculosis infection with increased time in or admission to the New York City jail system.
JAMA 1993, 270: 940 –941.
43. Castilla J, Gutierrez A, Guerra L. et al
. Pulmonary and extrapulmonary tuberculosis at AIDS diagnosis in Spain: epidemiological differences and implications for control.
AIDS 1997, 11: 1583 –1588.
44. Di Perri G, Danzi MC, de Checchi G. et al
. Nosocomial epidemic of active tuberculosis among HIV-infected patients.
Lancet 1989, 30: 1502 –1504.
45. Frieden TR, Sherman LF, Maw KL. et al
. A multi-institutional outbreak of highly drug-resistant tuberculosis.
JAMA 1996, 276: 12229 –12235.
46. 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 –1521.
47. Riley RL, Mills CC, O'Grady F, Sultan LU, Wittstadt F, Shivpuri DN. Infectiousness of air from a tuberculosis ward.
:Ultraviolet irradiation of infected air: comparative infectiousness of different patients.
Am Rev Respir Dis 1962, 85: 511 –525.
48. Haley CE, McDonald RC, Rossi L, Jones WD, Haley RW, Luby JP. Tuberculosis epidemic among hospital personnel.
Infect Control Hosp Epidemiol 1989, 10: 204 –210.
49. 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 –526.
50. Cohn DL, Bustreo F, Raviglione MC. Drug-resistant tuberculosis: review of the worldwide situation and the WHO/IUATLD Global Surveillance Project.
Clin Infect Dis 1997, 24 (Suppl. 1) : S121 –130.
51. Bradford WZ, Martin JN, Reingold AL, Schecter GF, Hopewell PC, Small PM. The changing epidemiology of acquired drug-resistant tuberculosis in San Francisco, USA.
Lancet 1996, 348: 928 –931.
52. Kaye K, Frieden TR. Tuberculosis control: the relevance of classic principles in an era of acquired immunodeficiency syndrome and multidrug resistance.
Epidemiol Rev 1996, 18: 52 –63.
cluster analysis; DNA fingerprinting; drug resistance; epidemiological methods; multivariate analysis; prisons; risk factors; transmission; tuberculosis; univariate analysis
© 2000 Lippincott Williams & Wilkins, Inc.
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Highlight selected keywords in the article text.
Data is temporarily unavailable. Please try again soon.