Evaluating young children recently exposed to airborne Mycobacterium tuberculosis is a public health priority. If infected, children aged <2 years are at high risk for severe tuberculosis (TB) disease (e.g., TB meningitis).1 In December 2003, infectious pulmonary TB disease was diagnosed in a foreign-born nurse working in the newborn nursery and maternity ward of a New York City hospital (hospital A); the nurse had declined treatment for latent TB infection (LTBI) after testing positive 11 years earlier. An investigation including medical evaluation of contacts in the nursery and maternity ward was conducted by the Bureau of TB Control (BTBC) at the New York City Department of Health and Mental Hygiene, hospital A, and CDC. This report summarizes the results of that investigation, which determined that approximately 1,500 patients had been exposed to the nurse but the majority could not be located for evaluation. Among those who were tested, four infants had positive tuberculin skin test (TST) results, likely attributable to recent transmission of M. tuberculosis. The findings emphasize the difficulty of conducting contact investigations in certain settings and the importance of effective LTBI testing and treatment programs for health-care workers (HCWs) to prevent TB disease and subsequent health-care–associated transmission.
In December 2003, a female nurse (nurse A) working in the newborn nursery and maternity ward at hospital A received a diagnosis of acid-fast bacilli (AFB) sputum smear–positive, noncavitary pulmonary TB disease. Eleven years earlier, nurse A had LTBI diagnosed with a TST result of 15 mm induration during screening for employment at hospital A, after emigrating from the Philippines. She had elected not to take the isoniazid prescribed for treatment. The reason nurse A gave for declining treatment was that most adults from the Philippines, where TB is endemic, have positive TST results and generally do not take treatment for LTBI. She also stated that the positive TST result might have been caused by her bacille Calmette-Guérin (BCG) vaccination for TB disease at birth or potential exposures while she was employed as a nurse in the Philippines. Nurse A had an annual TB symptom screen on eight other occasions and had one other chest radiograph (when she began work in a different area of the hospital) without evidence of TB disease.
Nurse A's symptoms began in September 2003 as a productive cough, wheezing, and shortness of breath. Her initial chest radiograph was interpreted as “normal heart and lungs” by a radiologist at hospital A. She was symptomatically treated for asthma with inhaled beta-agonists, inhaled steroids, oral steroids, antihistamines, and a cough suppressant. After her symptoms persisted for approximately 8 weeks, she underwent a chest computed tomography scan (CT) and, approximately 1 week later, bronchoscopy. The CT revealed bilateral upper-lobe disease with volume loss and calcified mediastinal lymph nodes. The leading diagnosis at the time was hypersensitivity pneumonitis. Specimens from a transbronchial biopsy, routinely sent for microscopic examination, revealed rare AFB; culture of bronchial alveolar lavage subsequently yielded M. tuberculosis that was susceptible to the four first-line anti-TB drugs. Genotyping of the M. tuberculosis isolate did not match any pattern in the New York City or national databases. Nurse A subsequently was screened for human immunodeficiency virus (HIV) and had a negative HIV test result.
On the basis of nurse A's AFB smear status at start of treatment, her infectious period was defined as September 1–November 29, 2003. Work schedules and hospital records for all coworkers and patients in the newborn nursery and maternity ward who were contacts of nurse A during this period were reviewed to identify and prioritize contacts and to assess risk factors for transmission. During her infectious period, nurse A worked 60 night shifts at hospital A and potentially exposed 32 coworkers, 613 infants in the newborn nursery, and 900 patients in the maternity ward. During a 7-month period, hospital A and BTBC took the following measures to notify contacts: (1) mailing certified letters, making telephone calls, and attempting home visits to hospital patients and to mothers and guardians of all infants; (2) faxing notifications to all pediatric providers in the area; and (3) cross-matching the list of exposed infants with names in the city's immunization registry. All contacts were offered a free medical evaluation, including a TST; if indicated, contacts also were offered chest radiography and sputum specimen collection to exclude a diagnosis of TB disease. Results were reviewed to estimate the extent of transmission.
Of the 32 potentially exposed coworkers, 25 (78%) had a previously documented positive TST baseline result, and none had taken treatment for LTBI. On screening, none of these 25 persons had symptoms for TB; they were offered LTBI treatment, but all 25 declined. TSTs were administered to the remaining seven coworkers, all with negative results.
The majority of patients in the maternity ward had received TSTs and HIV screening during the prenatal period. Extensive outreach by the hospital and city health department workers resulted in medical evaluation of 227 (37%) of the 613 infant contacts and 216 (24%) of the 900 female contacts. None of these contacts were determined to have TB disease. TST results were positive (≥5 mm) for five (2%) of 227 infants, including one who had received BCG vaccination during a family trip to the Dominican Republic. A positive TST result among infants was determined to be associated with cesarean delivery (relative risk [RR] = 11.8, 95% confidence interval [CI] = 1.3-103.1). TST results of 19 (9%) of 216 women with a prior negative test changed to positive (≥5 mm). Change in TST result was associated with foreign birth among women (RR = 5.9, CI = 1.4-24.5). No association was evident between a positive TST result or change in TST result and duration of contact (e.g., estimated time in the hospital while nurse A was working) or type of contact (e.g., receiving direct care) with nurse A.
Of the 900 patients admitted to the maternity ward during nurse A's infectious period, 807 were admitted for postpartum care and 93 for gynecologic indications or complications during pregnancy. Documentation of HIV test results were available for 806 of the 807 postpartum patients; 16 (2%) tested positive for HIV infection. Of these HIV-infected females, 13 delivered infants admitted to the newborn nursery (12 single infants and one twin birth). These 16 women and 14 infants were assigned the highest priority for follow-up testing. Three of the women and seven of the infants were located and tested for TB; none had evidence of LTBI or TB disease.
BTBC recommended LTBI treatment with isoniazid daily for 9 months for all contacts with a positive TST result, after TB disease was excluded. BTBC also recommended LTBI treatment for all HIV-infected persons exposed to nurse A and infants whose mothers had known HIV infection, regardless of their TST results, after TB disease was excluded.2
F Fitzpatrick, MPH, M Purswani, MD, B Fazal, MD, A Burrowes, MD, hospital A; K Granville, MPH, C Driver, DrPH, C Clark, MPH, S Munsiff, MD, New York City Dept of Health and Mental Hygiene, New York. D Ruggiero, MPH, K Ijaz, MD, J Jereb, MD, M Haddad, MSN, Div of TB Elimination, National Center for HIV, STD, and TB Prevention; B Heyman, DVM, Office of Workforce and Career Development; A Finlay, MD, EIS Officer, CDC.
The findings in this report underscore the difficulty and substantial resources required to conduct contact investigations and provide appropriate follow-up for patients exposed to M. tuberculosis in health-care settings. Despite extensive outreach efforts, approximately 70% of nurse A's patient contacts could not be traced. Hospital A was located in an economically depressed community. Hospital records of telephone numbers and addresses for many of the patients were incorrect. Nonetheless, evidence indicated that limited transmission of M. tuberculosis had occurred in hospital A. The strongest evidence of transmission was that four infants had positive TST results (a fifth infant tested positive but had recently received BCG vaccination), which in children is a sentinel indicator for recent transmission of M. tuberculosis.
In this investigation, the only risk factor significantly associated with M. tuberculosis transmission to the infants was cesarean delivery. Post-cesarean infants might have required more nursing care, thus resulting in more exposure. A major limitation of this investigation was the incomplete follow-up of all exposed patients. In addition, the extent of M. tuberculosis transmission to the most heavily exposed group, nurse A's coworkers, was difficult to ascertain because 78% had positive TST baseline results.
Nurse A underwent bronchoscopy before TB disease was clinically suspected. Because bronchoscopy is a cough-inducing procedure that can result in increased transmission of M. tuberculosis, diagnosis of TB disease and microscopic examination of sputum for AFB should be considered before bronchoscopy.3 CDC recommends avoiding bronchoscopy if possible for patients with suspected or confirmed TB disease or postponing the procedure until the patient is determined to be noninfectious by confirmation of three negative AFB sputum smear results. If the patient cannot produce sputum, CDC recommends considering sputum induction before bronchoscopy.3
Eleven years after her LTBI was detected, nurse A had infectious pulmonary TB disease diagnosed. An opportunity to prevent TB disease was missed when she did not complete treatment for LTBI. In light of the investigation described in this report, hospital A began exploring ways to promote LTBI treatment for employees with positive TST results during annual screenings for TB. Although the nurse did not have HIV infection, it is the greatest risk factor for progression from LTBI to TB disease.2 Therefore, voluntary HIV counseling, testing, and referral should be routinely offered to all persons at risk for LTBI. Health-care settings should be particularly aware of the need to prevent transmission of M. tuberculosis in settings where persons infected with HIV might be encountered or might work.
In 2002, the incidence of TB disease among foreign-born HCWs in the state of New York was 17.5 per 100,000, compared with 2.0 among U.S.-born HCWs.4 During 1998-2002, among 297 HCWs (employed in hospitals, home health care, nursing homes, and ambulatory care facilities) who were reported to have TB disease, 221 (74%) had had LTBI diagnosed previously. Of these, 111 (50%) had met criteria for treatment for LTBI, but only 26 (23%) of these received treatment.4 Those data and the circumstances described in this report support the need for effective LTBI testing and treatment programs among HCWs, particularly those born outside the United States.
Studies have demonstrated poor adherence to LTBI treatment among HCWs.5 HCWs might attribute a positive TST result to BCG vaccination.6 Compared with U.S.-born physicians, foreign-born physicians in one U.S. medical residency program were less likely to recommend LTBI treatment for themselves, their family members, or recent immigrants if they had received BCG vaccination.7 However, in the absence of M. tuberculosis infection, tuberculin reactivity caused by BCG vaccination wanes over time and is unlikely to persist >10 years after vaccination.8 Current guidelines recommend considering treatment for HCWs who have a TST result of ≥10 mm, especially if they emigrated from a country with high TB prevalence during the preceding 5 years.3 A history of vaccination with BCG should not influence the decision to treat LTBI.
The proportion of HCWs in the United States who were born outside the country is growing.9,10 Approximately 25% of all U.S. practicing physicians graduated from medical schools outside of the United States.9 Moreover, the shortage of registered nurses in the United States is anticipated to increase from 6% in 2000 to 29% by 2020, and foreign-born nurses likely will increasingly be sought to fill this gap.10 All HCWs in the United States, particularly those foreign-born or foreign-trained, should be encouraged to follow U.S. guidelines for LTBI treatment. Guidelines for preventing transmission of M. tuberculosis in health-care settings, including baseline and periodic TB screening and effective LTBI treatment programs for HCWs in high-risk settings, should be followed.3 In addition, infection-control programs in health-care settings should implement interventions to increase adherence to treatment for infected HCWs working in high-risk settings. On-site, directly observed preventive therapy is one such option.
The findings in this report are based, in part, on contributions by R Kairam, M Mikhail, G Weinberg, E Tulia, I Cheer, L White, M Cherian, Employee Health Svcs, hospital A, New York City.
Mycobacterium tuberculosis. JAMA. 2006;295(4):380-382. doi:10.1001/jama.295.4.380