Tuberculosis Receiving Treatment for Pulmonary ...

Time to Detection of Mycobacterium Tuberculosis in Sputum Culture Correlates With Outcome in Patients Receiving Treatment for Pulmonary Tuberculosis Matthew D. Epstein, Neil W. Schluger, Amy L. Davidow, Stanley Bonk, William N. Rom and Bruce Hanna Chest 1998;113;379-386 DOI 10.1378/chest.113.2.379 The online version of this article, along with updated information and services can be found online on the World Wide Web at: http://chestjournal.chestpubs.org/content/113/2/379

Chest is the official journal of the American College of Chest Physicians. It has been published monthly since 1935. Copyright1998by the American College of Chest Physicians, 3300 Dundee Road, Northbrook, IL 60062. All rights reserved. No part of this article or PDF may be reproduced or distributed without the prior written permission of the copyright holder. (http://chestjournal.chestpubs.org/site/misc/reprints.xhtml) ISSN:0012-3692

Downloaded from chestjournal.chestpubs.org by guest on July 21, 2011 1998 by the American College of Chest Physicians

Time to Detection of Mycobacterium Tuberculosis in Sputum Culture Correlates With Outcome in Patients Receiving Treatment for Pulmonary Tuberculosis* Matthew D.

Epstein, MD; Neil W. Schluger, MD, FCCP; Amy L. Davidow, PhD;

Stanley Bonk, BS; William N. Rom, MD, MPH, FCCP; and Bruce Hanna, PhD

Study objective: The purpose of this study was to determine whether the time to detection (TTD) of Mycobacterium tuberculosis in sputum culture correlates with the response to antituberculous treatment in patients with pulmonary tuberculosis. Study design: Twenty-six consecutive patients were studied who had active pulmonary tubercu¬ losis and sufficient sputum cultures and clinical follow-up to allow adequate assessment. Results: Following initiation of antituberculous therapy, 13 patients (group 1, responders) had a the TTD of M tuberculosis using the mycobacterial growth complete response to treatment, andThe indicator tube increased steadily. remaining 13 patients (group 2, nonresponders) had and demonstrated little or no increase in the TTD with of active disease evidence persistent treatment unless an additional therapeutic intervention was implemented (surgery, improved compliance with medications, or a change in medications). The presence of HIV infection, intravenous drug use, multidrug resistance, treatment with second-line therapy, extensive radiographic involvement, and cavitary disease were associated with a delayed increase in the TTD.

Conclusions: The TTD was superior to clinical, radiographic, or conventional bacteriologic evaluation in determining treatment outcome. The TTD closely correlates with the overall response to treatment for pulmonary tuberculosis and may represent a useful adjunct to predict outcome in these patients. 1998; 113:379-86) (CHEST

Key words: Mycobacterium tuberculosis; outcome; treatment Abbreviations: AFB=acid-fast bacilli;

PCR=polymerase chain reaction; TTD=time to detection

/^ onsequent to the recent resurgence of tubercu^^ losis, the rapid diagnosis of patients with active disease has become a focus of interest.1-2 Of equal importance to the control of tuberculosis is the assurance of effective treatment,3 yet over the past two decades, there has been little progress in the ability to predict a patient's response to antitubercu¬ lous therapy. Assessment or prediction of treatment response is especially important in patients at high *From the Departments of Medicine, Pathology and Environ¬ mental Medicine, Division of Pulmonary & Critical Care Med¬ icine, Bellevue Chest Service, NYU Medical Center, New York. Supported by grants MOl RR00096 and HL-51494. Dr. Schluger is supported by a Tuberculosis Academic Award (K07 HL03030) from the National Institutes of Health. Dr. Davidow is an Aaron Diamond Foundation Fellow, and this work was supported in part by a grant from the Aaron Diamond Foundation. Manuscript received March 20,1997; revision accepted July 3,1997. Reprints requests: Bruce A. Hanna, Dept of Pathology, Bellevue Hospital, New York, NY 10016

risk for treatment failure (such as those with multidrug-resistant tuberculosis or HIV infection or both) but remains essentially limited to clinical, radiographic, and conventional bacteriologic evaluation despite the inherent inaccuracies of these meth¬ ods.24 For example, although improving clinical status may indicate a response to treatment, patient heterogeneity and the presence of other concurrent illnesses reduce the certainty of clinical evaluation. Radiographic resolution as a marker of treatment response is inadequate because it lags significantly behind clinical improvement.56 Various serologic markers have been evaluated as indicators of disease activity, but their clinical usefulness remains uncer¬ tain.79 The most widely accepted measure of treat¬ ment response in patients with pulmonary tubercu¬ losis is the disappearance of acid-fast bacilli (AFB) from sputum, as assessed by microscopic examina-

/ 113 / 2 / FEBRUARY, 1 £ Downloaded from chestjournal.chestpubs.org by guest on JulyCHEST 21, 2011 1998 by the American College of Chest Physicians

379

smear sensitivity Unfortunately, the depends patient, type and degree of pulmonary parenchymal involvement, and the skill of the microscopist, and it ranges from 22 to 80%.10 Specificity is reduced by the presence of nontuberculous mycobacteria or dead organ¬ isms.24 Sputum culture is superior to direct exam¬ ination in detecting AFB, but because Mycobac¬ terium tuberculosis is relatively slow growing, cultivation is an

tion and culture.6 on

the

serial sputum

inefficient means to

response.4 In 1993, a preclinical study of the mycobacterial growth indicator tube (MGIT; Becton-Dickinson; assess

treatment

Cockeysville, Md) was conducted and found that this system can accurately detect the presence of M tuberculosis in as short a time as 2 days.11 This that the time to culture detection study observed M in serial samples steadily tuberculosis of (TTD) increased in most patients receiving treatment for tuberculosis. Another observation was pulmonary that in those patients in whom the TTD did not increase, there was a poor response to antituber¬ culous therapy.12 In order to further evaluate the of M tuberculosis in relationship between the TTD with active pulmonary the sputum of patients tuberculosis and their response to treatment, a systematic analysis was executed on patients who had received treatment for pulmonary tuberculo¬ sis and a comparison of their TTD data was made between these and clinical, radiographic, and microbiologic characteristics. Methods Patient

defined as three or more consecutive, culture-negative specimens with no subsequent positive cultures. Clinical, bacteriologic, and radiographic follow-up was obtained for at least 6 months after hospital admission.

Specimen Processing for MGIT Cultures Sputum induction was performed by having patients

rinse

their mouths with sterile water and then inhale nebulized 3% saline solution for 10 to 15 min or until the onset of sputum expectoration. All patient material was collected in sterile containers for processing. All specimens were digested and decontaminated with 2% NaOH, 2.9% sodium citrate, and 5% n-acetylcysteine for 15 min, and the reaction was terminated by addition of an equal volume of 0.9% saline solution. Specimens were centrifuged at 5,000 g for 15 min at 18°C and decanted, and the sediment was used to inoculate culture media and prepare smears using an auramine acid-fast stain. All cultures were inoculated to solid media and to a MGIT system. The MGIT is a tube containing Middlebrook and Cohn 7H9 broth supplemented with oleic acid, albumin, dextrose, catalase, polymyxin, amphotericin B, nalidixic acid, trimethoprim, and azlocillin sodium. An indicator at the bottom ofthe tube fluoresces as the level of oxygen decreases, signaling microbial growth. Tubes were examined daily on an ultraviolet transilluminator to detect growth, and a positive result w7as verified by examining Kinyoun- and Gram-stained smears. Specimens without growth after 60 days in culture were considered negative. If present, mycobacteria were then identified by DNA-RNA hybridization with the M tuberculosis culture confirmation system (AccuProbe Culture Confirma¬

System; Gen-Probe; San Diego, Calif). Susceptibility testing was performed using a modification of the method of proportions.13 M tuberculosis isolates were tested for suscep¬ tibility to isoniazid, rifampin, pyrazinamide, ethambutol hy¬ drochloride, streptomycin, kanamycin sulfate, ethionamide, rifabutin, ciprofloxacin hydrochloride, and amikacin sulfate. tion

Statistical Analysis Fisher's Exact Test was used to compare patient characteris¬ relationship between the TTD and treatment duration was studied using a linear mixed-effects model. Two such models were considered: the first (model 1) had a random baseline value of TTD for each patient and a fixed effect for treatment duration, ie, a common rate of change (slope) of TTD. The second model (model 2) included a mixed term formed from a group indicator and treatment duration, ie, distinct slopes of TTD for the two groups. To establish whether the change in TTD as a linear function of treatment duration differed by treatment response group, a likelihood ratio test of model 2 against model 1 was tics. The

Study Group

A review of data was done on all patients with a diagnosis of pulmonary tuberculosis who were admitted to the Chest Service of Bellevue Hospital Center in New York City between January 1, 1995, and June 30, 1995. Patients were included in this study if they had adequate clinical, radiographic, and microbiologic records; 6 or more positive sputum cultures over a period of at least 2 weeks (to allow for adequate data analysis); and adequate follow-up data. Medical records were reviewed to obtain the following information: age, gender, race, usage of tobacco, alcohol and illicit drug usage, domiciliary status, and HIV

performed.

serostatus.

Clinical presentation, symptoms, treatment regimens, and patient compliance were reviewed. Microbiologic data included AFB smear analysis, drug susceptibility patterns, and the TTD of M tuberculosis in culture using the MGIT system (described later). Chest radiographs were examined to determine initial

radiographic manifestations and changes with treatment. Response to treatment was determined (blinded to TTD results) by review of inpatient and outpatient medical records of Bellevue Hospital Center and the New York City Department of Health. Clinical improvement was defined as defervescence, weight gain, decreased cough or hemoptysis, and increased appetite. Bacteriologic improvement (culture conversion) was 380

Results

Demographic, Clinical, Bacteriologic, and Radiographic Characteristics During the study period, 65 patients were admit¬ ted to the Bellevue Chest Service with active pulmo¬ nary tuberculosis. Twenty-six of these patients had 6 or more positive sputum cultures processed by the MGIT system during a period of at least 2 weeks and had sufficient clinical, radiographic, and bacterio-

Downloaded from chestjournal.chestpubs.org by guest on July 21, 2011 1998 by the American College of Chest Physicians

Clinical

Investigations

The remaining logic follow-up to assessfromoutcome. this were excluded study. patients Characteristics of the study population are sum¬ marized in Tables 1 and 2; groups 1 and 2 represent patients who were differentiated by their response to treatment and are further described later. At the time of hospitalization, 81% of patients had consti¬ tutional symptoms (fever, sweats, weight loss, mal¬ aise) and 88% had respiratory symptoms (cough, pain). In 12 patients, dyspnea, pleuritic hemoptysis, of tuberculosis were either and

signs

symptoms

indiscernible from other concomitant ill¬ nesses, which included anemia, end-stage HIV infec¬ tion, psychosis, COPD, hepatitis, lung cancer, and chronic pancreatitis. absent

or

Clinical Outcome and Time to Detection Sixteen patients received standard therapy, which included both isoniazid and rifampin in addition to for the first 2 months of treatment. pyrazinamide Ethambutol also was given until drug susceptibilities were known. Two patients were not initially treated with both isoniazid and rifampin because of abnor¬ mal liver function. An additional eight patients re¬ ceived a second-line regimen that did not contain both isoniazid and rifampin because of drug resis¬ tance.

therapy, 13 Following initiation of antituberculous had clinical improve¬ patients (group 1, responders) ment (defervescence, weight gain, decreased cough or hemoptysis, and increased appetite) and a pro¬ while the rise in the TTD with gressive

treatment,

Table

1.Demographic Characteristics Total

(n=26)

Characteristics

Group 2

51 33-92

41 30-55

11 2

8 5

6 3 2 2

7 4 2 0

10 7 0 1

13 4 4 1

(n=13)

(n=13)

Age

Mean, yr 46 30-92

Range

Group 1

Gender

M 19 7

F Race

Black 13

Hispanic

7

Caucasian 4 Asian 2

4

Homeless 2 HIV infection HIV+ HIV-

(CD4 range)

Unknown 5

Total

Clinical Data

Signs and symptoms* Constitutional

Respiratory Drug susceptibility Pan-s

Isoniazid-R

Rifampin-R MDR Radiographic findings

Average No. involved lobes Average No. cavities

Military disease

Treatment' First line

Group 1

Group 2

21 23

11 12

10 11

18 1 1

10 1 1 1

(n=26)

2.3 1 1 16

(n=13)

1.8 0.7 1

(n=13)

2.7 1.4 0

10

10 3 Second line * Constitutional symptoms include fever, sweats, weight loss, and malaise. Respiratory symptoms include cough, hemoptysis, dyspnea, and pleuritic pain.

R=resistant; MDR=multidrug-resistant. +Pan-s=pan-susceptible; * second-line both isoniazid First-line therapy included and rifampin; therapy did not include both isoniazid and rifampin.

remaining 13 patients (group 2, nonresponders) had persistent, active disease and demonstrated little or no increase in TTD during treatment (Fig 1). Tables 1 and 2 compare characteristics between patients in groups 1 and 2. No statistically significant differences (between groups) were noted for any of the charac¬ teristics listed. In a test of model 2

(change in TTD related to clinical improvement) against model 1 (change in TTD unrelated to clinical improvement), the re¬ sulting change in (restricted maximal) log likeli¬ hood was 70.5. Comparing this with a x2 distribu¬ tion with 1 degree of freedom gives a probability value < Q

Z~