Published ahead of print on November 21, 2002, doi:10.1164/rccm.200208-873BC
© 2003 American Thoracic Society
Factors Associated with Tuberculin Conversion in Canadian Microbiology and Pathology WorkersDepartment of Medicine, University of British Columbia, Vancouver, British Columbia; Division of Infectious Diseases, Department of Medicine, University of Alberta, Edmonton, Alberta; Department of Public Health Services, University of Toronto, Ontario; Departments of Medicine and Epidemiology and Biostatistics, Montreal Chest Institute, McGill University, Montreal, Quebec, Canada Correspondence and requests for reprints should be addressed to Dick Menzies, M.Sc., M.D., Montreal Chest Institute, 3650, St. Urbain Street, Montreal, QC, H2X 2P4 Canada. E-mail: Dick.menzies{at}mcgill.ca
The risk of occupational tuberculosis (TB) infection and associated factors was estimated among all microbiology and pathology technicians and compared with a sample of nonclinical personnel in 17 Canadian acute care hospitals. Participants underwent tuberculin skin testing and completed questionnaires. Prior skin tests and vaccinations and all patients with TB hospitalized in the preceding 3 years were reviewed. Of the work areas where direction of air flow and air changes per hour were measured, only 51% were adequately ventilated. Among participating lab workers the average annual risk of tuberculin conversion was 1.0%. This was associated with lower hourly air exchange rates (16.7 versus 32.5 in workers with no conversion, p < 0.001) work in pathology (adjusted odds ratio [OR]: 5.4; [95% confidence interval: 1.3, 22], higher proportion of patients with missed diagnosis in the first 24 hours (per 20% increase—OR: 2.0; [1.3, 3.2], treatment delayed 1 week or more (per 20% increase—OR: 2.0; [3.2, 3.2]), and higher mortality (per 20% increase—OR: 2.5; [1.1, 5.6]). We conclude that laboratory workers, with no direct patient contact, have increased risk of tuberculin conversion in hospitals where a greater proportion of patients with TB die, or have delayed, or missed diagnosis, although this may be modified by workplace ventilation.
Key Words: tuberculosis • nosocomial transmission • occupational tuberculosis infection • autopsy • pathology and microbiology The risk of occupational tuberculosis (TB) among microbiology technicians was recognized (1, 2) and resulted in development of engineering controls and infection control procedures. These measures resulted in substantial reduction in morbidity (3–6), although transmission continues to occur (7). However, case reports (8, 9) and population-based studies (10–12) have demonstrated that pathology workers, particularly those involved in autopsies (8, 9), continue to have significantly increased risk (13). An illustrative case report is described in the online supplement. Is this because of greater exposure or less effective infection control procedures for pathology workers? Among all pathology and microbiology technicians and a sample of nonclinical personnel in 17 acute care hospitals in 4 Canadian cities we have estimated the association of tuberculin skin test conversion with work-site ventilation, and occupational TB exposure.
Detailed methods are described in the online supplement; an abbreviated description follows. A total of 17 acute care hospitals in 4 Canadian cities were selected: 15 classified as moderate to high risk for nosocomial TB transmission because they had at least six patients with TB admitted annually (14, 15), and 2 low-risk hospitals with fewer than two TB admissions annually. Within these hospitals, nursing personnel, respiratory therapists, physiotherapists, microbiology, and pathology technicians were studied. A sample of nonclinical workers served as a reference group for clinical and laboratory personnel. Because the laboratory technicians had very different working conditions and potential exposures, their results were analyzed separately from the clinical personnel with direct patient contact as reported elsewhere (16, 17). Signed informed consent was obtained from the participants. This study was approved by ethics committees in all the participating institutions. In microbiology laboratories, ventilation was assessed where specimens were initially processed and where TB specimens were handled. In pathology departments, the cutting rooms and autopsy suites were studied. Each room was inspected and temperature, humidity, air movement, area, and volume were measured. Smoke tubes were used to measure direction of airflow, and air changes per hour were estimated using pure carbon dioxide as a tracer gas (18), with doors and windows closed and ventilation systems in operation. Workers completed self-administered questionnaires regarding demographic information, training, work, and recognized episodes of TB exposure. Prior tuberculin testing and bacillus Calmette-Guérin (BCG) vaccination was verified from employee health records. For Quebec-born participants, information regarding BCG vaccinations was obtained from a previously validated (19) central registry. Charts were reviewed of all patients with newly diagnosed confirmed active pulmonary TB admitted to the study hospitals during the 3 years preceding the study. Workers with prior documented positive tuberculin tests and/or history of prior anti-TB therapy were not retested. Five tuberculin units of PPD-T, bioequivalent to PPD-S (Tubersol; Connaught Laboratories, Toronto, ON, Canada) was administered intradermally to the volar aspect of the forearm using the Mantoux technique to all other participants. The transverse diameter of induration was demarcated using the ballpoint technique 48 to 72 hours later, measured using machinists calipers and recorded in millimeters. Participants with initial tuberculin reactions less than 10 mm underwent repeat testing 1 to 4 weeks later.
Statistical Analysis
Bivariate associations were tested for statistical significance, with unpaired t tests for continuous variables and
Of 517 eligible microbiology and pathology technicians, 445 (86%) completed questionnaires and 411 (80%) had TST. Of the 397 eligible nonclinical personnel, 345 (87%) completed questionnaires and 338 (85%) had TST. The most common reason for incomplete TST was that participants reported prior but undocumented positive TST. The only significant differences in personal or work characteristics between the different groups of workers were that nonclinical personnel had worked fewer years than lab technicians (9 versus 12 years; p < 0.01), and 27% of the pathology technicians were males compared with 12% of the microbiology technicians (p < 0.01). Of the 411 participating lab technicians, and 338 nonclinical personnel, 111 lab technicians, and 74 nonclinical personnel had documented prior negative TST. Of these 14 (13%) and 4 (5%), respectively, had conversion. The study of the 429 hospitalized patients with TB in the period reviewed is reported in detail elsewhere (18). In brief, the diagnosis was missed (and the patient was not isolated) in the first 24 hours in 45%, treatment was delayed by 1 week or more in 30%, and 12% died. In 20 microbiology laboratories air change rates averaged 31.1 per hour, although only 10 (50%) had more than 15 air changes per hour and negative pressure (inward airflow). In the 27 pathology work areas measured there was an average of 16.7 air exchanges per hour; only 14 (52%) had more than 15 air exchanges per hour and negative pressure. As shown in Table 1 , workers' tuberculin conversions were associated with BCG vaccination, work in pathology, lower ventilation, indicators of delayed diagnosis, and overall mortality of patients with TB admitted to the same hospital.
In multivariate analysis, the only significant demographic or work characteristics associated with TST conversion were work in pathology and BCG vaccination (Table 2) . Among the exposure indicators shown in Table 2, after adjustment for all worker characteristics, TST conversion was significantly associated with a number of indices of delayed diagnosis and worse patient outcomes, particularly mortality. Ventilation was no longer significant, after adjustment for department, but ventilation was strongly correlated with department of work.
In this study 13% of laboratory technicians had documented tuberculin conversion, after an average of 14.5 years of work, corresponding to an annual risk of TB infection of 1.0%. After adjusting for potentially confounding personal characteristics, tuberculin conversion was significantly associated with work in pathology and indicators of delayed diagnosis of patients with TB. Lower average air change rates in pathology work areas could have contributed to the greater risk in pathology workers. BCG vaccination was the only nonoccupational characteristic associated with conversion, which may have been due to the boosting effect of repeated TST associated with remote BCG (23). It is also possible that this occurred because a second mycobacterial infection (such as true TB infection) in persons with prior mycobacterial exposure (such as BCG) can elicit a greater tuberculin response (23). Strengths of the study include the involvement of 17 Canadian hospitals in which the number of TB admissions ranged from 1 to 135 in the period reviewed. Potentially confounding workers' characteristics were measured and standardized protocols for tuberculin testing, exposure definition, and ventilation measurement were used in all sites. This made it possible to estimate the effect of workplace ventilation and delayed TB diagnosis on tuberculin conversion, adjusted for potentially confounding occupational and nonoccupational factors. Nevertheless there were a number of important limitations. A large number of laboratory workers participated but relatively few had documented prior negative TST, potentially limiting power. Despite this, a number of significant relationships were detected. Indirect indicators of exposure were used rather than actual handling of specimens from patients with recognized or (at the time) unrecognized TB. Tuberculin conversion was inversely associated with the number of known TB admissions but strongly correlated with higher proportion of missed/delayed diagnosis, a more frequent phenomenon in hospitals with fewer TB admissions (17). This suggests that indices of exposure based on the number of known patients may not accurately reflect exposure to the unknown patients—the most important sources of transmission in case reports and outbreak investigations (13). Ventilation conditions were measured once and may have been different at other times or in the past. To minimize this problem, information on prior renovations was collected, and at the time of measurement, ventilation was controlled to be representative of normal working conditions. Indicators of exposure to patients with TB were based on the preceding 3 years only. Exposure could have been misclassified if incidence in that hospital had changed. However, this seems unlikely because overall incidence in the four Canadian cities did not change appreciably between 1980–1995 (24, 25). The most important interpretation of these results is that delayed or missed diagnosis of patients with TB, a common phenomenon in the hospitals studied (17) and elsewhere (26, 27), results in similar level of risk for laboratory workers as it does for clinical personnel. This is presumably because laboratory workers handle specimens from patients with unrecognized TB. Involvement in autopsies of undiagnosed patients may play an important role (28). Nine of the 52 patients who died of TB in this study were only diagnosed at autopsy, and the elevated risk of this procedure has been recognized in other studies (8, 9) (10–12) due to massive aerosolization of TB bacilli (9). Because delayed diagnosis is associated with fewer TB admissions (18, 27), delayed diagnosis and death from unrecognized TB will continue to occur despite the best efforts of health care providers (29). In view of this, laboratory technicians must use universal precautions (30) for handling microbiologic and pathologic specimens, particularly for autopsies (28). The introduction of (1) much stricter infection control procedures and engineering controls resulted in dramatic reduction in rates (2, 6). The effectiveness of the engineering controls may be inferred from recent reports of high rates of TB infection and disease in laboratory workers in resource-poor countries (31) where these measures are too costly to be implemented (32). The present findings demonstrate that risk for pathology workers remains elevated, consistent with several other studies (8–10, 12). It is unclear whether this is due to greater exposure or deficiencies of infection control procedures. The finding that the air exchange rates in pathology work areas were about half those in microbiology areas and that lower ventilation was associated with TST conversion suggests that re-evaluation of the current recommended (14, 15) minimum air exchange rates may be warranted, although inferences based on these findings are limited. In conclusion, risk of tuberculin conversion was elevated in laboratory workers, particularly in pathology, and was associated with lower levels of ventilation and indicators of delayed TB diagnosis. Implementation of strict universal precautions could reduce the risk of TB infection and subsequent disease among these laboratory workers.
The authors thank the administration, infection control, and employee health departments of the hospitals, and the participating workers for their help and collaboration.
Members of the Canadian Collaborative Group: Montreal: Dr. Gilles Blanchette, Hôpital Sacré-Coeur; Dr. Philippe Bolduc, Hôpital Notre-Dame; Dr. Serge Déry, Hôpital Ste-Croix; Dr. Mike Libman, St-Mary's Hospital; Dr. Mark Miller, Jewish General Hospital; Dr. Pierre Robillard, Unité de Santé Publique; Dr. KevinSchwartzman, Montreal Chest Institute and Royal Victoria Hospital; Dr. Terry Nan Tannenbaum, Unité de Santé Publique; Dr. Jacques Tremblay, Hôpital Maisonneuve-Rosemont. Toronto: Dr. Monika Avandano, Westpark Hospital; Dr. John Conly, Toronto Hospital, Dr. Andrew Simor, Sunnybrook Hospital, Dr. Monika Naus, Ontario Ministry of Health. Edmonton: Dr. Geoff Taylor, University Hospital; Dr. Manuel Ma, Dr. Edith Blondel-Hill, Royal Alexandra Hospital. Vancouver: Dr. Julio Montaner, St. Paul's Hospital. Supported by the National Health Research and Development Program of Health Canada grant 6605–4437–502 and l'Association Pulmonaire du Quebec; D.M. was supported by a Chercheur Boursier Clinicien award from the Fonds de la Recherche en Santé du Quebec (1993–1998) and a Medical Scientist award from the Medical Research Council of Canada (1998–2003). This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Received in original form August 15, 2002; accepted in final form November 6, 2002
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
2 tests for categoric variables (
