0
Clinical Investigations: TUBERCULOSIS |

Increased Incidence of Multidrug-Resistant Tuberculosis in Diabetic Patients on the Bellevue Chest Service, 1987 to 1997* FREE TO VIEW

Mona Bashar, MD; Phil Alcabes, PhD; William N. Rom, MD, MPH, FCCP; Rany Condos, MD
Author and Funding Information

*From Bellevue Chest Service (Drs. Bashar, Rom, and Condos), Division of Pulmonary and Critical Care Medicine, Department of Medicine and Environmental Medicine, New York University School of Medicine, New York; and Program in Urban Public Health (Dr. Alcabes), Hunter College School of Health Sciences, City University of New York, New York, NY.

Correspondence to: Rany Condos, MD, Division of Pulmonary and Critical Care Medicine, Department of Medicine, New York University School of Medicine, 550 First Ave, NB 7N24, New York, NY 10016; e-mail: condor01@gcrc.med.nyu.edu



Chest. 2001;120(5):1514-1519. doi:10.1378/chest.120.5.1514
Text Size: A A A
Published online

Study objectives: To investigate the characteristics of tuberculosis infection in diabetic patients at Bellevue Hospital.

Design: We conducted a case-control study retrospectively reviewing the records of patients at Bellevue Hospital Center from 1987 to 1997 with a discharge diagnosis of tuberculosis and diabetes mellitus.

Setting: Bellevue Hospital Center is a 1,200-bed, inner-city municipal hospital located in the Lower East Side of New York City.

Patients: Fifty-three identified patients had verified tuberculosis infection and diabetes; of these, 50 charts were available for review. One hundred five control cases were selected from nondiabetic patients with a discharge diagnosis of tuberculosis during the same time period.

Measurements and results: Thirty-six percent (18 cases) of the patients with diabetes and tuberculosis had multidrug-resistant tuberculosis (MDR-TB) compared to only 10% (10 cases) in the control group (p < 0.01) Controlling for homelessness, HIV status, and directly observed therapy, the relative risk of MDR-TB was calculated to be 8.6 (confidence interval, 3.1 to 23.6) in the diabetic group compared to the control group.

Conclusions: There was a significant association between diabetes and MDR-TB. Diabetes continues to be a risk factor for tuberculosis and was associated with MDR-TB in our patients.

Figures in this Article

The New York City tuberculosis rate of 19.9 per 100,000 persons is more than three times the national rate, although the reported 1,460 new cases in 1999 declined 61.7% from the peak of 3,811 cases in 1992 (New York City Department of Health; unpublished data; May 2001).1 There were 31 tuberculosis patients with resistance to isoniazid and rifampin, a 93% decline from the 441 cases reported in 1992.1The tuberculosis epidemic in New York City was fueled by HIV-1 infection, drug abuse, and homelessness.2

Mortality associated with multidrug-resistant tuberculosis (MDR-TB) is extremely high and in many cases is worse than the mortality from tuberculosis in the prechemotherapy era; this is more true in patients with HIV.3Though infection-control measures have effectively limited the spread of MDR-TB in New York City, a large reservoir of infection has been created as a result of the many well-documented outbreaks.4

Patients with tuberculosis infection are prone to reactivation; one of the conditions that may predispose to reactivation is diabetes mellitus.5 The relative risk of developing bacteriologically confirmed pulmonary tuberculosis is up to five times higher in diabetics. Convincing data regarding tuberculosis in diabetics are lacking, and information on diabetic status is not routinely obtained on reported cases in New York City. In order to better understand the association of tuberculosis and diabetes in our patients, we conducted a retrospective review of patients with diabetes mellitus and tuberculosis from 1987 to 1997 and compared their clinical features with those of nondiabetic tuberculosis patients. We were particularly interested in the cases of MDR-TB and HIV status in both groups.

Charts of patients from the time period from 1987 to 1997 with the combined discharge diagnosis of tuberculosis and diabetes mellitus were reviewed for the following information: demographic data, details of tuberculosis disease with site of disease, date of initial diagnosis and drug sensitivity patterns, prior history of tuberculosis disease, purified protein derivative status, HIV status (as determined by enzyme-linked immunosorbent assay and Western blot), CD4 count, details of diabetes treatment, and written report of chest radiographs interpreted by radiologists. Five-year follow-up data were obtained from medical records of inpatient and outpatient visits, health department records, and autopsy results, when available. MDR-TB was defined as any case of tuberculosis that was resistant to two or more first-line tuberculosis medications, including at least isoniazid and rifampin.

Control cases were randomly selected from nondiabetic patients discharged from Bellevue Hospital between from 1987 to 1997 with a discharge diagnosis of tuberculosis. All cases had bacteriologically proven evidence of tuberculosis.

Statistical analysis was done by calculating the relative risk within 95% confidence intervals (CIs) using the Mantel-Haenszel estimate. Logistic regression analysis was done using statistical software (SAS version 6.12; SAS Institute; Cary, NC). As homelessness and HIV status were believed to be risk factors for acquiring tuberculosis, adjusted relative risks were calculated controlling for these two variables. The data were further adjusted to account for the possible effects of directly observed therapy (DOT).

Sixty-nine patients were identified through the review of discharge diagnoses at Bellevue Hospital during the 11-year period from 1987 to 1997 as having both tuberculosis and diabetes mellitus. Sixteen of these cases were considered noncountable by the Department of Health because the sputum cultures either grew nontuberculous acid-fast bacilli or failed to grow and treatment was discontinued. Of the 53 confirmed cases, 50 had medical records that were available for review; three charts were not found. One hundred five nondiabetic tuberculosis patients were included as control subjects. The peak number of cases was seen in 1994, and the number of MDR-TB and diabetics with tuberculosis were constant over the 11 years (Fig 1 ).

The two groups were comparable demographically and clinically (Table 1 ). The male to female ratio in both groups was approximately five to one. The age distribution of when tuberculosis was initially diagnosed was similar in both groups. Most patients (50% of cases and 58% of controls) were in the 30- to 50-year-old range when disease was first diagnosed. The racial distribution of both groups was similar, with> 75% being drawn from minority populations (black, Hispanic, and Asian). The patients tended to be of low socioeconomic status, as represented by the high proportion of patients who were reported as homeless (54% of cases and 65% of controls). There was no significant difference in the rates of homelessness in the two groups. Although HIV-infection status was not available for many of the patients because they refused testing, there was a higher percentage of HIV-positive patients included in the control group (36%) as compared to the diabetic group (14%). This, however, was not statistically different. The lungs were the predominant sites of disease in both groups. Alcohol, cigarette smoking, and illicit drug use were also common in both groups.

Fifty-four percent of the study subjects had type 1 diabetes, while 40% had type 2. Sixty-eight percent of the patients used insulin, 22% used oral agents, and 6% were diet controlled. Twenty percent exhibited evidence of renal insufficiency (serum creatinine > 1.5 mg/dL) or renal failure. The duration of diabetes at the time tuberculosis was diagnosed varied from < 1 year to long-standing disease of > 10 years (average, 6.4 years). Overall, the patients had moderate glycemic control with average fasting serum glucose levels between 200 mg/dL and 300 mg/dL and average glycosylated hemoglobin levels of 10.4%. There were no noted incidents of diabetic ketoacidosis during the tuberculosis treatment period.

There was a significantly greater number of MDR-TB cases in the diabetic group (18 cases, 36%) as compared to the control group (10 cases, 10%). Table 2 gives the odds ratio for having MDR-TB in diabetic patients as compared to nondiabetic control subjects. The crude odds ratio was calculated to be 5.1 (CI, 2.1 to 12.5). As the prevalence of MDR-TB is known to be increased among homeless and HIV-positive individuals, the data were analyzed to control for these two possible confounding factors. Within the subgroup of homeless individuals, the odds of having MDR-TB was 2.7 (CI, 0.8 to 47.1) times greater in the diabetic patients. In domiciled individuals, the odds of having MDR-TB was 14.2 (CI, 2.7 to 73.8) times greater in the diabetic patients. The adjusted odds ratio controlling for homelessness was 5.1 (CI, 2.0 to 13.4). The adjusted relative risk controlling for HIV status was 5.5 (CI, 1.6 to 19.5). The overall adjusted odds ratio controlling for both homelessness and HIV status was 5.3 (CI, 1.9 to 14.7).

Because the study period spanned a time when DOT was not the standard of care, there was a proportion of diabetic patients who were not followed up by DOT (Table 3 ). The data were further analyzed to account for any effect from DOT (Table 3). Adjusting for DOT gave an odds ratio of 8.8 (CI, 3.2 to 24.0). Adjusting for both DOT and homelessness gave an odds ratio of 8.6 (CI, 3.1 to 23.6). This means that a diabetic tuberculosis patient is almost nine times more likely to have MDR-TB than a nondiabetic tuberculosis patient. Because DOT is associated with a fivefold to sevenfold increase in MDR-TB odds in our data, a tuberculosis patient who is both diabetic and receiving DOT has a 49-fold greater chance of having MDR-TB.

There were 41 chest radiograph reports available for review in the study group and 99 reports available in the control group. The remaining reports could not be retrieved from the radiology archives. Both the study and control groups had similar findings, with upper-lobe cavitary or cystic disease being the predominant feature. In the diabetics, there were 21 cases with cavitary or cystic disease, and 20 cases (95%) involved the upper lobes. In the control subjects, there were 34 cases with cavitary or cystic disease, and 32 cases (94%) involved the upper lobes.

Five-year follow-up data were available for most patients (Table 4 ). There was a greater number of patients in the diabetic group who died with active tuberculosis (7 patients, 14%) as compared to the control group (1 patient, 1%). Other causes of death in both groups were complications related to AIDS noted in two patients from the diabetic group and one patient from the control group. The diabetic group also had three deaths from renal failure, one death from mucormycosis infection, and one death from cardiac arrest.

More patients in the control group were able to complete therapy (57%) than in the study group (42%). Completion of therapy refers to the patient’s ability to comply with the standard length of treatment and to have bacteriologic clearance of sputum cultures. There were 10 cases (20%) of relapse/recurrent tuberculosis noted in the study group; five patients remained pansensitive, one patient became isoniazid resistant, and four patients became multidrug resistant (Table 4). Four cases (4%) of relapse/recurrent tuberculosis were noted in the control group; three patients remained pansensitive and one patient developed multidrug resistance. Relapse/recurrence was defined as emergence of tuberculosis within 2 years of initial diagnosis and treatment. The study group also included six MDR-TB patients who underwent lung resection for tuberculosis (Table 4). This represented 33% of the MDR-TB diabetic cases. These patients all had evidence of persistent infection despite appropriate therapy. There were no lung resections noted in the control group.

Our study revealed that the proportion of cases of MDR-TB was significantly higher in patients with coexisting diabetes mellitus. After controlling for homelessness and HIV status, both risk factors for the development of drug-resistant tuberculosis, we found that diabetic patients were at increased risk for MDR-TB, with an adjusted odds ratio of 5.3 (CI, 1.9 to 14.6). Frieden and colleagues6 reported the emergence of drug-resistant tuberculosis in New York City in 1991. They reported that the leading cause of drug resistance was noncompliance or incomplete medical treatment. They also noted that in the patients who had never been treated for tuberculosis, the percentage of patients who were infected with a strain of tuberculosis resistant to one or more drugs was 23%.6 Our findings were similar to this observation. In our analysis, there were 10 diabetic patients with a known recurrence of tuberculosis; 4 patients had a resistant strain that likely emerged from incomplete treatment. The majority of patients, however, had primary infection with a drug-resistant strain.

Homelessness has been associated with the emergence of MDR-TB. Homeless populations are less likely to have access to medical care, and the close contact, overcrowding, and lack of sufficient ventilation that may exist in homeless shelters would foster the spread of tuberculosis infection and the emergence of drug-resistant strains.2 Fifty-four percent of our diabetic patients reported being homeless. However, in our controlled analysis, homelessness was not found to be a confounding variable. Infection with the HIV virus has also been linked with the incidence of MDR-TB.7 In our study, the HIV status for a large percentage of the cases and controls was not available. With the available data, HIV status was also found not to be a confounding variable.

The availability of DOT impacts on the rates of MDR-TB. The lack of DOT can lead to poor compliance and the emergence of resistant organisms. However, patients with an established diagnosis of resistant tuberculosis are more likely to be enrolled in DOT to better supervise their therapy. In New York City, DOT became standard therapy starting in 1992.8 Our study included some patients who received a diagnosis prior to the time DOT became the standard of care at Bellevue Hospital. We tried to account for these possible interactions and adjusted our data, finding no effect. There were 18 cases of MDR-TB in the diabetic patients, and 13 patients were enrolled in DOT and successfully treated. Four patients received a diagnosis prior to DOT and had previous treatment failures. Only 4 of 10 of the MDR-TB patients in the control group were receiving DOT, resulting in the greater odds ratio for MDR-TB in diabetics and DOT. Since diabetes is a risk factor for increased mortality with tuberculosis as well as MDR-TB, DOT should be the required standard of care for these patients.

We evaluated the radiographic presentation of tuberculosis in diabetic patients and found no significant difference from control cases. Both groups had predominantly upper-lobe cystic or cavitary disease. This is in contrast to older studies by Sosman and Steidel,9who reported that diabetic patients presented more often with lower lobe disease. A more recent study by Perez-Guzman et al,10who reviewed the radiographs of 192 diabetic patients with tuberculosis, also reported this finding. However, other studies11 suggest that multilobe involvement is the predominant radiographic finding in both diabetic and nondiabetic patients with pulmonary tuberculosis. Upper-lobe disease predominated, and lower-lobe involvement was infrequent, occurring in < 7% of patients.11

It appears that diabetic patients as a group are more susceptible to having a more aggressive course of tuberculosis disease.12 Despite this observation having been made in the past, there has been no clear explanation as to why they are more susceptible. It is possible that diabetic patients have some degree of impaired GI drug absorption even in the absence of clinical gastroparesis. The hyperglycemic state may additionally interfere with achieving adequate tissue levels of the medications, or interfere with alveolar macrophage or CD4+ cell function. This could be further evaluated by measuring serum levels of antituberculosis medications in diabetic patients as compared to nondiabetic patients.

This study was limited by its small number of subjects and by retrospective data collection. However, it is evident that the pattern of tuberculosis in diabetic patients in our population differed significantly from nondiabetic patients. The diabetic patients were more than five times as likely to have infection with a multidrug resistant strain of tuberculosis. This has important clinical and treatment implications. A greater suspicion for MDR-TB should be entertained in diabetic patients, with treatment consisting of at least four antituberculosis medications until sensitivity patterns are available. If resistance to isoniazid and rifampin is detected, an antituberculosis regimen with at least two new drugs including an injectable drug tailored to the resistance pattern is recommended.13

Abbreviations: CI = confidence interval; DOT = directly observed therapy; MDR-TB = multidrug-resistant tuberculosis

This research was supported in part by a General Clinical Research Center grant from the National Institutes of Health, National Center for Research Resources (M01RR00096) awarded to the New York University School of Medicine, New York, NY.

Figure Jump LinkFigure 1. Rates of MDR-TB cases reported per year of study in diabetic and control cases. Squares = total No. of cases; diamonds = No. of diabetic cases; circles = No. of pansensitive diabetic cases; triangles = No. of MDR-TB diabetic cases.Grahic Jump Location
Table Graphic Jump Location
Table 1. Demographic and Clinical Characteristics of Study and Control Subjects*
* 

Data are presented as No. (%). TB = tuberculosis; DM = diabetes mellitus; PPD = purified protein derivative; NA = not available.

Table Graphic Jump Location
Table 2. Relative Risk of MDR-TB in Diabetic Patients vs Control Patients
* 

Odds ratio.

Table Graphic Jump Location
Table 3. Logistic Regression Analysis Odds of MDR-TB Adjusted for DOT
Table Graphic Jump Location
Table 4. Summary of MDR-TB Cases in Diabetics and Control Subjects*
* 

IRP = isoniazid, rifampin, pyrazinamide; IRE = isoniazid, rifampin, ethambutol; IRS = isoniazid, rifampin, streptomycin; IR = isoniazid, rifampin; IRPES = isoniazid, rifampin, pyrazinamide, ethambutol, streptomycin; IRRifabutin = isoniazid, rifampin, rifabutin; IRPE = isoniazid, rifampin, pyrazinomide, ethambutol; IRPEthio = isoniazid, rifampin, pyrazinamide, ethionamide; IRPESCipro = isoniazid, rifampin, pyrazinamide, ethambutol, streptomycin, ciprofloxacin; rx = treatment; NA = not available.

 

Died with evidence of active tuberculosis disease.

Neville, K, Bromberg, A, Bromley, R, et al (1994) The third epidemic: multi-drug resistant tuberculosis.Chest105,45-48. [PubMed] [CrossRef]
 
Concato, J, Rom, WN Endemic tuberculosis among homeless men in New York City.Arch Intern Med1994;154,2069-2073. [PubMed]
 
Park, MM, Davis, AL, Schluger, NW, et al Outcome of MDR-TB patients, 1983–1993: prolonged survival with appropriate therapy.Am J Respir Crit Care Med1996;153,317-324. [PubMed]
 
Schluger, NW, Huberman, R, Holzman, R, et al Screening for infection and disease as a tuberculosis control measure among indigents in New York City, 1994–1997.Int J Tuberc Lung Dis1999;3,281-286. [PubMed]
 
Kim, SH, Hong, YP, Lew, WJ, et al Incidence of pulmonary tuberculosis among diabetics.Tuberc Lung Dis1995;76,529-523
 
Frieden, T, Sterling, T, Pablos-Mendez, A, et al The emergence of drug-resistant tuberculosis in New York City.N Engl J Med1993;328,521-526. [PubMed]
 
Brudney, K Resurgent tuberculosis in New York City: HIV, homelessness, and the decline of TB control programs.Am Rev Respir Dis1991;144,745-749. [PubMed]
 
Schluger, N, Ciotoli, C, Cohen, D, et al Comprehensive tuberculosis control for patients at high risk for noncompliance.Am J Respir Crit Care Med1995;151,1486-1490. [PubMed]
 
Sosman, M, Steidel, J Diabetic tuberculosis.AJR Am J Roentgenol1927;17,625-631
 
Perez-Guzman, C, Torres-Cruz, A, Villarreal-Velarde, H, et al Atypical radiological images of pulmonary tuberculosis in 192 diabetic patients: a comparative study.Int J Tuberc Lung Dis2001;5,455-461. [PubMed]
 
Morris, J, Seaworth, B, Mc Allister, C Pulmonary tuberculosis in diabetics.Chest1992;102,539-541. [PubMed]
 
Pablos-Mendez, A, Blustein, J, Knirsch, C The role of diabetes mellitus in the higher prevalence of tuberculosis among Hispanics.Am J Public Health1997;87,574-579. [PubMed]
 
Harkin, TJ, Harris, HW Treatment of multidrug resistant tuberculosis. Rom, WN Garay, S eds.Tuberculosis1996,843-850 Little, Brown,. Boston, MA::
 

Figures

Figure Jump LinkFigure 1. Rates of MDR-TB cases reported per year of study in diabetic and control cases. Squares = total No. of cases; diamonds = No. of diabetic cases; circles = No. of pansensitive diabetic cases; triangles = No. of MDR-TB diabetic cases.Grahic Jump Location

Tables

Table Graphic Jump Location
Table 1. Demographic and Clinical Characteristics of Study and Control Subjects*
* 

Data are presented as No. (%). TB = tuberculosis; DM = diabetes mellitus; PPD = purified protein derivative; NA = not available.

Table Graphic Jump Location
Table 2. Relative Risk of MDR-TB in Diabetic Patients vs Control Patients
* 

Odds ratio.

Table Graphic Jump Location
Table 3. Logistic Regression Analysis Odds of MDR-TB Adjusted for DOT
Table Graphic Jump Location
Table 4. Summary of MDR-TB Cases in Diabetics and Control Subjects*
* 

IRP = isoniazid, rifampin, pyrazinamide; IRE = isoniazid, rifampin, ethambutol; IRS = isoniazid, rifampin, streptomycin; IR = isoniazid, rifampin; IRPES = isoniazid, rifampin, pyrazinamide, ethambutol, streptomycin; IRRifabutin = isoniazid, rifampin, rifabutin; IRPE = isoniazid, rifampin, pyrazinomide, ethambutol; IRPEthio = isoniazid, rifampin, pyrazinamide, ethionamide; IRPESCipro = isoniazid, rifampin, pyrazinamide, ethambutol, streptomycin, ciprofloxacin; rx = treatment; NA = not available.

 

Died with evidence of active tuberculosis disease.

References

Neville, K, Bromberg, A, Bromley, R, et al (1994) The third epidemic: multi-drug resistant tuberculosis.Chest105,45-48. [PubMed] [CrossRef]
 
Concato, J, Rom, WN Endemic tuberculosis among homeless men in New York City.Arch Intern Med1994;154,2069-2073. [PubMed]
 
Park, MM, Davis, AL, Schluger, NW, et al Outcome of MDR-TB patients, 1983–1993: prolonged survival with appropriate therapy.Am J Respir Crit Care Med1996;153,317-324. [PubMed]
 
Schluger, NW, Huberman, R, Holzman, R, et al Screening for infection and disease as a tuberculosis control measure among indigents in New York City, 1994–1997.Int J Tuberc Lung Dis1999;3,281-286. [PubMed]
 
Kim, SH, Hong, YP, Lew, WJ, et al Incidence of pulmonary tuberculosis among diabetics.Tuberc Lung Dis1995;76,529-523
 
Frieden, T, Sterling, T, Pablos-Mendez, A, et al The emergence of drug-resistant tuberculosis in New York City.N Engl J Med1993;328,521-526. [PubMed]
 
Brudney, K Resurgent tuberculosis in New York City: HIV, homelessness, and the decline of TB control programs.Am Rev Respir Dis1991;144,745-749. [PubMed]
 
Schluger, N, Ciotoli, C, Cohen, D, et al Comprehensive tuberculosis control for patients at high risk for noncompliance.Am J Respir Crit Care Med1995;151,1486-1490. [PubMed]
 
Sosman, M, Steidel, J Diabetic tuberculosis.AJR Am J Roentgenol1927;17,625-631
 
Perez-Guzman, C, Torres-Cruz, A, Villarreal-Velarde, H, et al Atypical radiological images of pulmonary tuberculosis in 192 diabetic patients: a comparative study.Int J Tuberc Lung Dis2001;5,455-461. [PubMed]
 
Morris, J, Seaworth, B, Mc Allister, C Pulmonary tuberculosis in diabetics.Chest1992;102,539-541. [PubMed]
 
Pablos-Mendez, A, Blustein, J, Knirsch, C The role of diabetes mellitus in the higher prevalence of tuberculosis among Hispanics.Am J Public Health1997;87,574-579. [PubMed]
 
Harkin, TJ, Harris, HW Treatment of multidrug resistant tuberculosis. Rom, WN Garay, S eds.Tuberculosis1996,843-850 Little, Brown,. Boston, MA::
 
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
CHEST Journal Articles
PubMed Articles
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543