Pulmonary Tuberculosis and SARS, China

To the Editor: As part of a cohort study of 83 patients with severe acute respiratory syndrome (SARS) in Beijing, China, we conducted a follow-up study of all the patients by routine medical examination. During the process, 3 patients with chest radiographs consistent with active disease were identified as having pulmonary tuberculosis (TB). Here we describe the 1-year clinical outcome and immune response in these patients. 
 
Demographic details and coexisting conditions are shown in the Table. Patient 1 was a healthcare worker who became infected with SARS-associated coronavirus (CoV) while on duty with SARS patients. After he was transferred to a hospital dedicated to SARS management, pulmonary TB was diagnosed (positive acid-fast bacilli smear on sputum samples). Patients 2 and 3 were known to have cases of pulmonary TB and became infected with SARS-CoV after contact with other patients hospitalized for SARS. These 2 patients were sputum smear–negative for acid-fast bacilli, and diagnosis was made on the basis of previous exposure to TB, relevant symptoms of typical pulmonary TB, chest radiographs consistent with active disease, a positive tuberculin skin test result, and the finding of cavity regression on chest radiographs after anti-TB treatment was initiated. No cultures were obtained for isolation and comparison of Mycobacterium tuberculosis strains (1). All 3 patients had confirmed SARS based on amplification of SARS-CoV RNA by reverse transcriptase–polymerase chain reaction (RT-PCR) from sputum and stool specimens (2). Patients 2 and 3 recovered without complications; patient 1 had the most severe disease and required mechanical ventilation in an intensive care unit before recovering. 
 
 
 
Table 
 
Demographic and clinical information on severe acute respiratory syndrome (SARS) patients with pulmonary tuberculosis (TB), Beijing, China, 2003 
 
 
 
Both cellular and humoral immunity were evaluated during the follow-up of these patients. T-lymphocyte subsets were measured 6 months after disease onset by flow cytometry using fluorescein isothiocyanate–labeled specific monoclonal antibodies. Compared to other SARS patients (n = 47), the 3 patients with TB had lower mean CD4+ T cells (368.4/μL vs. 656.6/μL, respectively; p = 0.05) and lower mean CD8+ T cells (371.0/μL vs. 490.1/μL, respectively; p = 0.39). SARS-CoV immunoglobulin G (IgG) antibody titers were measured by enzyme-linked immunosorbent assay kit (Huada Company, Beijing, China) at months 1, 2, 3, 4, 7, 10, and 16 after disease onset. (Titers were not measured for the 3 TB patients at month l.) Compared to most (26 [78.8%] of 33) other SARS patients whose antibodies remained detectable throughout follow-up, 2 of the 3 TB patients (patients 1 and 3) had undetectable antibody titers as of months 7 and 16, respectively. In patient 1, antibody titers, when detectable, were unusually low (40). Both patients 1 and 3 had prolonged viral excretion in stools, sputum, or both. While the median (range) duration of virus excretion in stools and sputa for the entire measurable cohort (n = 56) was 27 (16–127) and 21 (14–52) days, respectively (3), it was 125 and 16 days for patient 1, and 109 and 52 days for patient 3 (viral excretion data could not be obtained from patient 2 because sequential specimens for detection were unavailable). 
 
TB in SARS patients has been reported on rare occasions (4,5). In a cohort of 236 patients in Singapore, it was diagnosed in 2 patients after recovery from SARS (4). As with patient 1 in this study, TB had developed after the patient acquired SARS, most likely as the result of reactivation of past infection or new infection with M. tuberculosis, while temporarily immunosuppressed because of SARS (6) and corticoid therapy. Such phenomena have been described with other viral infections such as measles and HIV (7,8). By contrast, patients 2 and 3 were known TB patients who acquired SARS through exposure to SARS patients in the same hospital wards. Both diseases are known to be transiently immunosuppressive (6,9), and their combined effect resulted in more pronounced CD4+ cell decreases in coinfected SARS patients than others. Such immunosuppression also resulted in poorer IgG antibody response in coinfected SARS patients than in others and delayed viral clearance, as shown by longer viral excretion in sputum and stools. While viral excretion could be prolonged in coinfected patients, no virus could be isolated from any RT-PCR–positive specimen collected after 6 weeks of illness, which suggests that excreted viruses were no longer infectious (3). 
 
These case reports remind us of the importance of strict isolation of SARS patients, careful use of steroids for their case management, and the possibility of coinfection with TB in SARS patients with incomplete recovery.


Pulmonary Tuberculosis and SARS, China
To the Editor: As part of a cohort study of 83 patients with severe acute respiratory syndrome (SARS) in Beijing, China, we conducted a follow-up study of all the patients by routine medical examination. During the process, 3 patients with chest radiographs consistent with active disease were identified as having pulmonary tuberculosis (TB). Here we describe the 1-year clinical outcome and immune response in these patients.
Demographic details and coexisting conditions are shown in the Table. Patient 1 was a healthcare worker who became infected with SARS-associated coronavirus (CoV) while on duty with SARS patients. After he was transferred to a hospital dedicated to SARS management, pulmonary TB was diagnosed (positive acid-fast bacilli smear on sputum samples). Patients 2 and 3 were known to have cases of pulmonary TB and became infected with SARS-CoV after contact with other patients hospitalized for SARS. These 2 patients were sputum smear-negative for acid-fast bacilli, and diagnosis was made on the basis of previous exposure to TB, relevant symptoms of typical pulmonary TB, chest radiographs consistent with active disease, a positive tuberculin skin test result, and the finding of cavity regression on chest radiographs after anti-TB treatment was initiated. No cultures were obtained for isolation and comparison of Mycobacterium tuberculosis strains (1). All 3 patients had confirmed SARS based on amplification of SARS-CoV RNA by reverse transcriptase-polymerase chain reaction (RT-PCR) from sputum and stool specimens (2). Patients 2 and 3 recovered without complications; patient 1 had the most severe disease and required mechanical ventilation in an intensive care unit before recovering.
Both cellular and humoral immunity were evaluated during the followup of these patients. T-lymphocyte subsets were measured 6 months after disease onset by flow cytometry using fluorescein isothiocyanate-labeled specific monoclonal antibodies. Compared to other SARS patients (n = 47), the 3 patients with TB had lower mean CD4+ T cells (368.4/µL vs. 656.6/µL, respectively; p = 0.05) and lower mean CD8+ T cells (371.0/µL vs. 490.1/µL, respectively; p = 0.39). SARS-CoV immunoglobu-lin G (IgG) antibody titers were measured by enzyme-linked immunosorbent assay kit (Huada Company, Beijing, China) at months 1, 2, 3, 4, 7, 10, and 16 after disease onset. (Titers were not measured for the 3 TB patients at month l.) Compared to most (26 [78.8%] of 33) other SARS patients whose antibodies remained detectable throughout follow-up, 2 of the 3 TB patients (patients 1 and 3) had undetectable antibody titers as of months 7 and 16, respectively. In patient 1, antibody titers, when detectable, were unusually low (40). Both patients 1 and 3 had prolonged viral excretion in stools, sputum, or both. While the median (range) duration of virus excretion in stools and sputa for the entire measurable cohort (n = 56) was 27 (16-127) and 21 (14-52) days, respectively (3), it was 125 and 16 days for patient 1, and 109 and 52 days for patient 3 (viral excretion data could not be obtained from patient 2 because sequential specimens for detection were unavailable).
TB in SARS patients has been reported on rare occasions (4,5). In a cohort of 236 patients in Singapore, it was diagnosed in 2 patients after recovery from SARS (4). As with patient 1 in this study, TB had developed after the patient acquired SARS, most likely as the result of reactivation of past infection or new infection with M. tuberculosis, while temporarily immunosuppressed because of SARS (6) and corticoid therapy. Such phenomena have been described with other viral infections such as measles and HIV (7,8). By contrast, patients 2 and 3 were known TB patients who acquired SARS through exposure to SARS patients in the same hospital wards. Both diseases are known to be transiently immunosuppressive (6,9), and their combined effect resulted in more pronounced CD4+ cell decreases in coinfected SARS patients than others. Such immunosuppression also resulted in poorer IgG antibody response in coinfected SARS patients than in others and delayed viral clearance, as shown by longer viral excretion in sputum and stools. While viral excretion could be prolonged in coinfected patients, no virus could be isolated from any RT-PCR-positive specimen collected after 6 weeks of illness, which suggests that excreted viruses were no longer infectious (3).
These case reports remind us of the importance of strict isolation of SARS patients, careful use of steroids for their case management, and the possibility of coinfection with TB in SARS patients with incomplete recovery. To the Editor: The epidemiology of tetanus in the United Kingdom changed in 2003 when a cluster of cases in injecting drug users (IDUs) occurred (1,2). Before 2003, the incidence of tetanus was low in the United Kingdom, with occasional cases predominantly in unvaccinated elderly persons (3). The situation contrasted with the United States where injecting drug use is commonly reported among persons with tetanus (4).
We investigated the UK cluster to identify the source of infection and opportunities for prevention. We ascertained cases through statutory and nonstatutory reporting to the Health Protection Agency and collected additional information on IDUs for all reported cases of tetanus since January 1, 2003, by adapting the existing enhanced tetanus surveillance. A case was defined as mild-tomoderate trismus and at least 1 of the following: spasticity, dysphagia, respiratory embarrassment, spasms, autonomic dysfunction, in a person who injected drugs in the month before symptom onset.
Twenty-five cases were reported from July 2003 to September 2004 ( Figure). Thirteen (50%) were women; the median age of male and female patients was 39 and 32 years of age, respectively (range 20-53, p = 0.1). Twenty patients were white, and 1 was Chinese (information was missing for 4). None reported travel overseas before becoming sick. Seventeen of 21 patients with information reported having injected heroin intramuscularly or subcutaneously (popping) or having missed veins. Most patients (16/25) came to the hospital with severe generalized tetanus. Injection site infections were common (17/19).
Two patients died (case fatality 8%). Of 23 survivors, 2 had mild disease and 21 required intensive treatment for a median of 40 days (range 24-65 in 15 cases with complete information). Tetanus immunization status available for 20 case-patients (based on medical records or patient and parental recall) indicated that only 1 patient (with severe disease) had received the 5 doses necessary for complete coverage. Nine patients were never vaccinated. Twelve of 14 patients tested for tetanus immunity on admission by a standard indirect enzyme-linked immunosorbent assay had antibody levels lower than the cutoff value for protection (<0.1 IU/mL). One patient with severe disease had a level just above the cutoff value and 1 patient with mild disease had a protective antibody level. Clostridium tetani was isolated from 2 patients; tetanus toxin was detected in serum from 1 and also from another patient. Other anaerobes, including C. novyi, C. histolyticum, and C. perfringens, were