| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
Dr. Villarino and Ms. Clairy are members of the Division of Tuberculosis Elimination, Centers for Disease Control and Prevention.
Correspondence to: Margarita E. Villarino, MD, MPH, Division of Tuberculosis Elimination, Centers for Disease Control and Prevention, NCHSTV MS-E10, Atlanta, GA 30333
In most studies on the subject, persons with diabetes mellitus (DM) have a twofold to fourfold higher incidence of active tuberculosis (TB) than nondiabetic persons.1 Furthermore, some studies report that diabetics may present with more advanced disease at the time of TB diagnosis and may have an increased mortality rate.2 The predisposition of diabetics to infections that are normally controlled by cell-mediated immunity (eg, infection with Mycobacterium tuberculosis) may result from one or more defects of pulmonary host defense, including conditions that interfere with normal clearance mechanisms or that impair pulmonary immune cell function. The greatest difficulty in studying DM as an independent risk factor for the development of TB is the presence of potential confounding variables. These variables include other coexistent medical conditions (eg, malnutrition, chronic renal disease) and personal behaviors, such as smoking and alcohol abuse, that may further weaken host defenses.2 Alternatively, some of the differences in the risk for TB disease may be more attributable to exposure differences. Hendy and Stableforth3 described findings from a group of diabetic TB patients living in the United Kingdom. Asian diabetic TB patients presented with a higher incidence of cavitary disease and sputum-positive status than did a control group of non-Asian diabetic TB patients. This was presumably due to higher dose exposure to M tuberculosis. This higher exposure risk of the Asians appeared to be associated with their living in settings that replicate and concentrate the social behaviors and conditions of their birth communities. Even though DM patients were shown to be powerfully predisposed to TB in the era before insulin therapy, some experts now consider this risk to be rather modest4 ; in studies of diabetes as a risk factor for TB, investigators need to consider other risks, such as exposure, when comparing differences between populations. In this context of uncertainty, what might account for a specific type of TB disease presenting in diabetic patients at a much higher rate than in nondiabetic TB patients?
In this issue of CHEST (see page 1514), Bashar et al add to the body of studies that have found an association between DM and TB. In this study, they reviewed records of 50 case patients hospitalized between 1987 and 1997 at Bellevue Hospital Center (New York City) with a discharge diagnosis of TB and diabetes. They compared these to 105 nondiabetic control patients with a discharge diagnosis of TB from the same hospital during the same time period. These control patients were randomly selected (methodology not described) but were not matched by either time of hospital admission or discharge. Case patients and control patients had similar high frequencies of characteristics commonly associated with "urban" TB cases: young age (20 to 50 years), male sex, nonwhite race, and high rates of homelessness, alcohol use, and cigarette use. Notably, the investigators report that diabetic patients were more than five times as likely to be infected with a multidrug-resistant (MDR) strain of M tuberculosis (ie, resistant to at least isoniazid and rifampin). There are two types of TB drug resistance: primary and secondary (or acquired).5 Primary resistance develops in persons who are initially infected with resistant organisms. Secondary resistance develops during therapy in patients initially infected with susceptible organisms, either because the patients were treated inadequately or because they did not follow the prescribed regimen appropriately. In this study, 22% (4 of 18 patients) of those with diabetes and multidrug resistant (MDRTB) had a history of previous therapy for TB; thus, it appears that the majority (78%) of patients with MDRTB and DM included in the analysis represent instances of primary resistance.
Surveillance for drug resistance is an essential component of TB control programs at the national level, the local level, or within a hospital. Primary drug resistance should signal ongoing transmission via contact with infectious drug-resistant TB. The phenomenon of drug-resistant TB was described soon after the discovery of anti-TB drugs, but the first documented outbreak of drug-resistant TB was not reported until 1970.6 In the 20 years from 1970 to 1990, the literature shows only five relatively small TB outbreaks involving strains of M tuberculosis resistant to two or more drugs. From 1990 through August 1992, several large outbreaks of primary MDRTB occurred in seven hospitals (four in New York City, one in New York State, one in New Jersey, and one in Miami) and in the New York State correctional system. In the present study, 15 of the 18 diabetic patients with MDRTB received diagnoses in or after 1990; the 3 diabetic patients with MDRTB diagnosed before 1990 all had prior therapy as a risk factor for MDRTB (only 1 of the 15 patients with MDRTB diagnosed after 1990 had prior therapy as a risk factor). In contrast, of the 10 MDRTB patients without diabetes, 5 patients received diagnoses after 1990 and 5 patients received diagnoses in or before 1990. However, information about previous therapy is not provided for the control patients. The apparent clustering of patients with MDRTB and DM in the post-1990 era almost has statistical difference (Yates corrected p value, 0.07) when compared to the time distribution of MDRTB cases in nondiabetics. The timing of this clustering of patients at Bellevue Hospital coincides with the nosocomial MDRTB outbreaks identified in the other New York City hospitals previously mentioned.
The MDRTB nosocomial outbreaks of the early 1990s were all characterized by delayed diagnosis, leading to delayed initiation of isolation and treatment and thus to extended infectiousness. Because they primarily involved patients with HIV infection, the M tuberculosis infections rapidly developed into new cases of active disease that were sources of further transmission. Because the timing and location of the presently reported study corresponds to the peak time and the epicenter of the MDRTB epidemic in the United States,7 we believe that not analyzing differential exposure to M tuberculosis in hospitals and clinics represents a methodologic shortcoming of the analysis. It is well recognized that MDRTB outbreaks require two catalysts: ongoing transmission of drug-resistant microorganisms, and exposure of highly susceptible patients who are at increased risk for TB developing once infected with M tuberculosis. Patients with diabetes may be expected to have higher hospitalization and clinic visit rates, longer duration of stay if hospitalized, as well as increased risk for active TB disease. Although a biological factor or factors might predispose persons to both diabetes and MDRTB, we postulate that the more likely explanation for this association is an environmental factor unmeasured by this study. The retrospective nature of this analysis could have precluded the investigation of possible nosocomial transmission of MDRTB; however, the number of clinic visits and the total in-hospital days in the year prior to the diagnosis of TB might have easily been quantified.
The national surveillance system for TB has collected information on drug resistance since 1993. From 1993 through 1996, 1,457 MDRTB cases were reported from 42 states, New York City, and Washington, DC; however, 38% of these were reported from New York City.8 In 1993, of persons without a known prior TB episode, 9% of cases reported in New York City had resistance to both isoniazid and rifampin; prior to that, Frieden et al9 found a 7% rate of primary MDRTB in a survey of patients in New York City with culture-positive TB diagnosed during April 1991. Analysis of the 1993–1998 national surveillance data found that ongoing transmission of MDRTB strains primarily affected young adults aged 25 to 44 years in urban areas, particularly United States-born persons and those with HIV coinfection.10 In this study, there were more HIV-seropositive control patients (36%) than case patients (14%). However, 58% of the diabetic patients and 32% of the nondiabetic patients were not tested for HIV. With the available data in a multivariate analysis model, HIV status was not identified as a risk factor for MDRTB.
Overall, we agree with the authors’ recommendations aimed at preventing cases of acquired MDRTB. Drug susceptibility testing should be performed on all initial isolates of M tuberculosis. Additional isolates should be tested for patients with positive culture findings after 3 months of therapy. Initial therapy should include four drugs (isoniazid, rifampin, pyrazinamide, and either ethambutol or streptomycin) and should always be directly observed by a health-care provider. Furthermore, adherence to infection control policies is essential to prevent the transmission of M tuberculosis in institutional settings and the occurrence of cases of primary MDRTB.11 Of critical importance are the administrative procedures for identifying patients with suspected TB, assessing their potential infectiousness, and ensuring appropriate action (eg, isolation, prompt institution of therapy) to limit exposure of staff and patients in these settings.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
