Use of Therapeutic Drug Monitoring for Multidrug-Resistant Tuberculosis Patients* Jiehui Li, MBBS, MS; Joseph N. Burzynski, MD, MPH; Yi-An Lee, MPH; Debra Berg, MD; Cynthia R. Driver, RN, MPH; Renee Ridzon, MD; and Sonal S. Munsiff, MD Study objectives: Therapeutic drug monitoring (TDM) is the process of obtaining the serum concentration of a medication and modifying the dose based on the results. Little is known about the application of TDM in the treatment of patients with multidrug-resistant (MDR) tuberculosis (TB) in clinical practice. This study characterized how TDM was applied in the management of MDR TB patients, and examined the clinical indications for ordering TDM, the process for obtaining drug concentrations, and the clinician response to the drug concentrations. Design: In a retrospective study, we compared the clinical and demographic characteristics of MDR TB patients who received TDM with those who did not. The clinical application of TDM also was described in patients who received TDM. Setting: A municipal TB control program. Patients or participants: Patients in whom TB was diagnosed that was caused by an isolate resistant to at least isoniazid and rifampin, and who received treatment for TB in one of the health department chest clinics between July 1, 1993, and August 31, 1997, were studied. Results: Forty-nine patients receiving TDM had a longer time to culture conversion and treatment duration, more pulmonary TB in combination with an extrapulmonary site, drug resistance, and visits to the health department clinics (p < 0.05) than the 60 patients without TDM. Of the 49 patients who had initial TDM, 73.5% of them had the reason for being tested specified. A total of 85.7% of initial TDM results were collected at the appropriate time of blood sampling. Clinician response to TDM results varied with the drug that was being tested. Conclusions: The use of TDM depended largely on the patient s clinical presentation. Sitespecific guidelines on the use of TDM for managing TB patients may maximize the benefit of TDM. (CHEST 2004; 126:1770 1776) Key words: drug monitoring; multidrug-resistant tuberculosis Abbreviations: BMI body mass index; BTBC Bureau of Tuberculosis Control, New York City Department of Health and Mental Hygiene; CI confidence interval; MDR multidrug-resistant; PAS para-aminosalicylic acid; TB tuberculosis; TDM therapeutic drug monitoring Therapeutic drug monitoring (TDM) is the process of obtaining the serum concentration of a medication and modifying the dose based on the results to optimize its therapeutic benefits, while *From the New York City Department of Health and Mental Hygiene (Drs. Li, Burzynski, Berg, and Munsiff, Ms. Lee, and Ms. Driver), Bureau of Tuberculosis Control, New York, NY; and the Division of Tuberculosis Elimination (Dr. Ridzon), Centers for Disease Control and Prevention, Atlanta, GA. Manuscript received January 9, 2004; revision accepted July 6, 2004. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (e-mail: permissions@chestnet.org). Correspondence to: Jiehui Li, MBBS, MS, Bureau of Tuberculosis Control, New York City Department of Health and Mental Hygiene, 225 Broadway, 22nd floor, New York, NY 10007; e-mail: jli1@health.nyc.gov minimizing its risk for side effects or toxicity. 1 Drug concentration monitoring has been accepted with widespread use for medications such as digoxin, For editorial comment see page 1722 lithium, theophylline, and aminoglycosides. However, TDM has not been used routinely in the management of patients with tuberculosis (TB) because most TB patients are treated with standard doses of the first-line medications and have high cure rates with infrequent toxicity. 2,3 It has been shown, 4 however, that among a small group of patients with a poor response to therapy, some have low serum drug concentrations, suggesting that these low concentra- 1770 Clinical Investigations
tions may contribute to the 2 to 5% of patients with adverse outcomes such as clinical failure or relapse. TB disease with organisms that are resistant, at least to isoniazid and rifampin (ie, multidrug resistance) was a major contributor to the recent TB epidemic in New York City. 5 In both the 1991 and the 1992 national surveys, 61% of cases with multidrug-resistant (MDR) TB were reported from New York City. 6 The treatment of MDR TB requires the use of second-line medications that are less effective, have more frequent side effects, have a more narrow therapeutic/toxic effect ratio, and require longer durations of treatment than disease caused by drug-susceptible organisms. 7 In 1994, TDM became available for the management of MDR TB patients in the New York City Department of Health Chest Clinics. While TDM has been suggested as a useful tool in the management of patients with MDR TB, 1,8 10 little has been published on how this tool has been applied in clinical practice. Studies in patients with TB caused by an organism that was susceptible to all first-line anti-tb agents have shown a dose-response relationship between the dosage administered and the response to treatment and relapse rate after treatment completion. 3 Therefore, monitoring serum drug concentrations presumably could be important in the treatment of patients with MDR TB. The present study was performed to describe the actual use of TDM, not its pharmacokinetic value, in the management of MDR TB patients in the setting of a municipal TB control program. Specifically, we compared the clinical and demographic characteristics of MDR TB patients who received TDM with those who did not, and described the clinical indications for ordering TDM, the processes for obtaining drug concentrations, and the clinician response to the drug concentrations. Study Population Materials and Methods This was a descriptive study to describe the use of TDM among MDR TB patients who were treated in 1 of 10 chest clinics in the Bureau of Tuberculosis Control, New York City Department of Health and Mental Hygiene (BTBC) between July 1, 1993, and August 31, 1997. In order to ensure that patients had sufficient opportunity for TDM, patients with MDR TB were included if they had at least five medical visits to a BTBC clinic within 12 months of the first positive culture result. Patients who ever received TDM in a BTBC chest clinic during TB treatment were classified as TDM patients, and those who never received TDM were classified as non-tdm patients. Measurement of TDM The procedures for obtaining TDM were the same in all chest clinics. Patients were instructed to arrive at the clinic in the morning. They were not given specific instructions for eating requirements or restrictions. The prescribed anti-tb medications were taken under direct observation. Patients were asked to stay at the clinic until the process was completed, although this was not enforced. An appropriate time of blood sampling was defined if blood was drawn 2 h after the observed dose for oral and injected medications, except for para-aminosalicylic acid (PAS), which was administered in a granular form, and a postdose concentration was obtained 4 to 6 h after administration of the drug. All blood samples were centrifuged within 60 min. Serum was placed in a labeled polypropylene tube (Cryovial; Federal Industries Corp; Plymouth, MN), which was frozen and sent the following day on ice to the New York City Bureau of Laboratories, Mycobacteriology Laboratory, and was immediately placed in a 70 C freezer. The sample was frozen overnight, then was packed in dry ice and shipped to Infectious Disease Pharmacokinetics Laboratory, at National Jewish Medical and Research Center (Denver, CO). Results of the TDM were faxed and mailed back to the originating chest clinic. Data Collection Data were obtained through reviews of the clinic medical record and also from the TB case registry. The results were transferred to a standard questionnaire that had been designed for the study. The data extracted included demographic, clinical, bacteriologic, and pharmacologic information. The data included the patients height and weight, the number of clinical visits, the site of TB disease, symptoms at the time of presentation, other prescribed medications, other medical conditions, HIV status, tobacco and alcohol history, radiograph results, clinical specimen smear and culture result pattern, drug susceptibility, TB drug regimen at the time of testing, the actual time that the dose was received, the actual times that the blood was drawn for TDM, the names of the drugs tested, and the daily doses for tested drugs. Weight and height were abstracted from the patient s medical chart, and body mass index (BMI) was calculated as weight/height squared (in kilograms per square meter). The TDM results were abstracted from a report form that had been received from the National Jewish Medical and Research Center. If a TDM report could not be found in the patient s record, the result was obtained directly from National Jewish Medical and Research Center. Definitions Drug susceptibility testing for most cases was performed by standard methodology at the mycobacteriology laboratory of the health department. 11 The duration of appropriate treatment was defined as the time from the start of therapy with at least two drugs to which the organism was susceptible until the last treatment date. Gaps in treatment of 1 month were excluded from the calculation of the duration of treatment. Culture conversion was defined as two negative culture results 1 month apart with no subsequent positive culture results for 6 months. Time to culture conversion, therefore, was defined as the period from the date that appropriate treatment started to the date that culture conversion occurred. Statistical Analysis Demographic and clinical characteristics of patients who ever received TDM were compared to those who never received TDM during the course of treatment. In univariate analysis, the Pearson 2 test was used for the comparison of proportions. The Fisher exact test was used when one cell had an expected frequency of less than five. The Student t test was used to compare means, and the Wilcoxon rank-sum test was used to compare medians. A two-sided p value of 0.05 was regarded as www.chestjournal.org CHEST / 126 / 6/ DECEMBER, 2004 1771
being statistically significant. A statistical software package (SAS, release 7; SAS Institute; Cary, NC) was used for data analysis. This study was approved by the Institutional Review Board at the New York City Department of Health and Mental Hygiene, and the Centers for Disease Control and Prevention. Results One hundred twenty-three patients met the study criteria, and 109 patients (89%) had medical charts available for review. There were no significant differences in demographic characteristics between patients with charts available for review and those that were not reviewed except for ethnicity. Patients whose charts were not available were more likely to be Hispanic. Forty-nine of 109 patients (45%) received TDM during the course of treatment in Department of Health chest clinics. Compared with 60 non-tdm patients (Table 1), there were no significant differences in age, sex, country of origin, or race, but TDM patients had a greater BMI. In a univariate analysis of clinical characteristics (Table 2), TDM patients had a greater number of isolates with first-line drug resistance (4.1 vs 3.6 drugs, respectively; p 0.01), a longer mean duration of treatment (24 vs 22 months, respectively; p 0.04), more months receiving an injectable medication (9 vs 7 months, respectively; p 0.04), and longer time to culture conversion (1.7 vs 1.0 months, respectively; p 0.04) than did non- TDM patients. TDM patients were more likely to have both pulmonary and extrapulmonary disease (20% vs 2%, respectively; p 0.01), and to have undergone a surgical procedure as part of the TB management (12% vs 2%, respectively; p 0.04). Having TDM was not associated with HIV status, Table 1 Demographic Characteristics of Study Subjects by TDM Status* Variable TDM Group (n 49) Non-TDM Group (n 60) p Value Age, yr 38 (20 78) 36 (0 77) 0.199 BMI 23 (14 44) 21 (14 30) 0.004 Sex, Male 27 (55) 34 (57) 0.870 Female 22 (45) 26 (43) Country of birth United States 27 (55) 33 (55) 0.991 Non-United States 22 (46) 27 (45) Race White 24 (49) 25 (42) 0.521 Black 19 (39) 23 (38) Asian 6 (12) 12 (20) *Values given as median (range) or No. (%), unless otherwise indicated. Only 37 patients (76%) in the TDM group and 35 patients (58%) in the non-tdm group had BMI measurements available. receiving directly observed therapy, the occurrence of adverse effects, or having cavitary lesions on a chest radiograph. TDM use also was not associated with TB symptoms at diagnosis, the duration of symptoms, prior history of liver, renal or GI disease, and the results of the sputum smear (data not shown). In a multivariate model that included the site of disease, the time to culture conversion, and the number of isolates with first-line drug resistance, patients with both pulmonary and extrapulmonary TB were 9.5 times more likely to have TDM performed than patients with pulmonary TB only (95% confidence interval [CI], 1.09 to 83.7). The time to culture conversion (adjusted odds ratio, 1.1; 95% CI, 0.9 to 1.2), and number of first-line drugs to which organisms were resistant (adjusted odds ratio, 1.4; 95% CI, 0.9 to 2.2) did not remain significantly associated with the use of TDM. The other variables that were statistically significant in univariate analysis were not included in the multivariate analysis because of collinearity with other variables that were included in the model. The most common reason for ordering the initial TDM was routine drug assessment (31%), followed by unstated/unknown reasons (27%), lack of clinical response to treatment (22%), and GI complaints or other adverse reaction (16%) [Table 3]. Among those patients receiving a second TDM, the most common reason was to follow-up on the first TDM concentrations or a new regimen (45%), followed by routine drug assessment (19%) and unstated/unknown reasons (16%). Ten percent of second TDMs were ordered because clinicians questioned the accuracy of the initial TDM. In the three second TDMs that were performed to verify the first TDM, five drugs were retested, and three of the five drugs were concordant with the repeat concentrations. The other two drugs, which had a lower than normal concentration in the first TDM, had normal concentrations on the second TDM. The reasons for ordering three or more TDMs were similar to the ones for ordering the second TDM. Among TDM patients, 42 (86%) had their initial drug concentrations obtained within the appropriate time of blood sampling (Table 4). Subsequent drug concentrations, however, demonstrated a wide range of appropriate blood sampling times (63 to 88%). The average number of drugs tested initially was 3.0, and subsequent TDMs averaged 1.6 to 2.5 drugs tested. The median number of weeks between consecutive TDMs was 9 weeks (range, 0 to 110.9 weeks). Of the 42 patients who had their first TDM specimen collected at the appropriate time for sampling, the most commonly tested drugs were cycloserine (23 patients), ethionamide (21 patients), the fluoroquinolones ofloxacin and ciprofloxacin (22 and 13 patients, respectively), ethambutol (12 pa- 1772 Clinical Investigations
Table 2 Clinical Characteristics and Treatment Outcomes of Study Subjects by TDM Status* Variable TDM Group (n 49) Non-TDM Group (n 60) p Value HIV-positive, cells/ L 12 (25) 14 (23) 0.460 CD4 count 200 8 7 0.391 CD4 count 200 1 1 CD4 count unknown 3 6 Previously treated for TB 18 (37) 18 (30) 0.457 Disease site Pulmonary only 36 (74) 55 (92) 0.011 Extrapulmonary only 3 (6) 4 (7) 1.000 Both sites 10 (20) 1 (2) 0.002 Cavitary lesion 15 (31) 12 (20) 0.546 First line drugs to which isolate was resistant 4.1 (1) 3.6 (1) 0.011 Use of injectable TB medication, mo 9 (1 31) 7 (1 32) 0.036 Surgical procedure as an adjunct to treatment 6 (12) 1 (2) 0.044 Time to culture conversion, mo 1.7 (0.03 29.4) 1.0 (0 30.2) 0.038 Died prior to treatment completion 6 (12) 5 (8) 0.500 Ever on DOT 47 (96) 59 (98) 0.443 Compliance with DOT, % 90 (0 100) 87 (0 100) 0.789 Occurrence of adverse effects 26 (53) 22 (37) 0.086 Duration of treatment among those who completed 20 mo 2 (8) 11 (30) 0.005 20 30 mo 17 (65) 25 (68) 30 mo 7 (27) 1 (3) *Values given as No. (%), unless otherwise indicated. DOT directly observed therapy. Only limited to patients with pulmonary TB. Values given as mean (SD). Values given as median (range). Only limited to those who completed treatment (n 63). tients), and pyrazinamide (12 patients) [Table 5]. Of these drugs, more than half of the drug concentrations for ofloxacin, ethionamide, and ethambutol were found to be below the normal range at 2-h postdose, and the clinician s response to these low concentrations was to increase the dose, proportionately more so compared to other drugs with low concentrations. Less than half of the drug concentrations for cycloserine and ciprofloxacin were below the normal range at 2-h postdose, but the overwhelming clinician response to low concentrations of cycloserine was to order a second TDM, rather than increasing the dose or discontinuing the drug. Only two concentrations were obtained for streptomycin, and both were below the normal range. Only one concentration each was taken for all other anti-tb drugs, which are not listed in Table 5. All were within their respective normal ranges. Relatively fewer drugs had higher than normal drug concentrations, with cycloserine having the most higher-thannormal drug concentrations (six concentrations) followed by ofloxacin (three concentrations). Of these concentrations, half of the cycloserine dosages and two thirds of the ofloxacin dosages were adjusted by lowering the doses by treating physicians. The time between the blood collection date of Table 3 Reasons for Ordering TDM by Treating Physicians* First TDM (n 49) Second TDM (n 31) Third or Greater TDMs* (n 38) Reason for Ordering TDM No. % No. % No. % Drug concentrations routinely evaluated 15 31 6 19 8 21 Lack of clinical response to treatment 11 22 1 3 0 0 GI complaints or other adverse reactions 8 16 2 7 2 5 Renal failure 2 4 0 0 0 0 Not stated (unknown) 13 27 5 16 11 29 Verify first TDM results NA NA 3 10 0 0 Follow-up on the change made from previous TDM results NA NA 14 45 17 45 *Sixteen patients underwent three or more TDMs, which resulted in 38 TDMs. NA not available. www.chestjournal.org CHEST / 126 / 6/ DECEMBER, 2004 1773
Variable Table 4 Number of Times TDM Was Performed in Patients TDM First Second Third Fourth Fifth or More Patients 49 31 16 8 5* Drugs tested Average 3.0 2.2 2.5 2.1 1.6 Range 1 5 1 4 1 4 1 3 1 3 TDM was obtained at the correct time point after dose Yes 42 (86) 23 (74) 14 (88) 5 (63) 10 (63) No 7 (14) 8 (26) 2 (13) 3 (38) 6 (38) Time between TDMs, wk Median 9.6 9.0 9.7 12.1 Range 0 110.9 1.0 22.4 5.3 31.1 4.1 31.4 *There were five patients who underwent five to nine TDMs, which resulted in 16 blood drawings. Calculated as: ([No. of patients whose time at which blood was drawn is correct or not/no. of patients having TDM] 100). Values given as No. (%). the first TDM and the receipt of the results was 8 to 167 days (median, 19 days). This turnaround time was not different between those patients who had their dose modified and those who did not (median, 19 vs 20 days, respectively; p 0.639). HIV-positive patients were more likely to have a higher proportion of concentrations for ofloxacin and ethionamide that were lower than the normal range of serum concentrations, compared to HIV-negative patients (ofloxacin, 86% vs 39%, respectively [p 0.07]; and ethionamide, 83% vs 43%, respectively [p 0.15]). However, the differences were not statistically significant. Discussion The use of TDM for managing patients with MDR TB has been recommended in our program for all MDR TB patients, but only 45% of the patients in this study had TDM performed in the course of treatment during the study period. For the most part, the reasons for or against performing TDM were not recorded in the medical charts. Possible reasons for clinicians to not use TDM for all MDR TB patients include the following: (1) the patient already was receiving the highest dose that the clinician was comfortable prescribing, and the drugs to which the isolate was susceptible were limited, thus, low serum drug concentration would not be helpful in guiding a decision on increasing medication dosage; (2) the patient was responding to therapy, and the culture remained negative; and (3) there was limited exposure to TDM management and the interpretation of results. Thus, TDM was used for patient management on a case-by-case basis. However, one cannot assume that the ordering of TDM Table 5 Physician s Adjustment of Treatment Based on TDM Results by Anti-TB Drugs Tested* Treating Physician Adjustment of Low Concentration Anti-TB Drug Tests, No. Lower Than Normal Range Increased Dose Discontinued the Drug Ordered 2nd TDM No Change No 2nd TDM Cycloserine 23 10 (44) 2 (20) 0 (0) 7 (70) 1 (10) Ofloxacin 22 12 (55) 6 (50) 1 (8) 4 (33) 1 (8) Ethionamide 21 11 (52) 4 (36) 0 (0) 6 (55) 1 (9) Ciprofloxacin 13 5 (39) 2 (40) 1 (20) 2 (40) 0 (0) Ethambutol 12 7 (58) 4 (57) 0 (0) 3 (43) 0 (0) Pyrazinamide 12 1 (8) 0 (0) 0 (0) 1 (100) 0 (0) PAS 4 2 (50) 1 (50) 0 (0) 0 (0) 1 (50) Streptomycin 2 2 (100) 1 (50) 0 (0) 0 (0) 1 (50) *Data are presented as No. (%) unless otherwise indicated. First TDM concentration was collected the right time for all anti-tb drugs listed at 2 h after administration, except where noted. All but one patient were receiving the normal daily dose. The correct collection time was 4 to 6 h postdose. 1774 Clinical Investigations
was solely dependent on patient factors alone and not on physicians preference or practice. In this study, the use of TDM was more common in patients with a longer length of treatment and time to culture conversion, more severe disease, as reflected in the greater number of first-line drugs to which the Mycobacterium tuberculosis isolates were resistant, concomitant pulmonary and extrapulmonary disease, and undergoing surgical procedures. These findings are not consistent with those of a metaanalysis 12 of 247 studies conducted between 1974 and 1994 on TDM intervention for conditions other than TB, and outcome measures revealed that TDM reduced adverse reactions, mortality rates, and length of stay when compared to those of non-tdm patients. However, a study 13 on using gentamicin TDM for burn patients with Gram-negative septicemia revealed a longer mean length of infection and length of hospital stay in those with TDM compared to those without. It was theorized that TDM increased the survival of patients with severe burns. 13,14 It appears from our data that TDM was used in the patients with more complicated and advanced TB disease, which was probably appropriate. Another possibility is that the duration of therapy for the TDM group might have been prolonged by the lack of dosage modification or the lack of using TDM to monitor the optimal serum concentrations of TB drugs by treating physicians. Most initial TDM was ordered with a specific reason or as a routine part of the care, and blood samples were collected within a required time frame in this study, indicating an appropriate application of TDM among patients with MDR TB in the TB control setting. However, while 86% of initial TDMs were collected at the appropriate time of blood sampling, subsequent collections showed a great degree of time variability (63 to 88% of the blood sampling times were appropriate). This has a direct effect on the clinician s ability to interpret the test accurately and contributes to increased costs due to the need for retesting. While the reasons why certain blood samples were drawn outside the appropriate time frame are unknown, possible reasons include inaccurate documentation of the time of drug administration and serum concentrations, and patients not returning to the clinic at the appropriate time for drug concentration determination, especially for drug concentrations collected 6 h after the dose was ingested. Furthermore, lack of clinician knowledge on TDM procedures also may have played a crucial role in the inappropriate time of collection. For example, 13 of 17 PAS drug concentrations were collected at 2 h after the dose was administered rather than at the recommended time of 4 to 6 h after the dose was administered for the granular form of the drug, which has a peak serum concentration at 4 to 6 h compared to that of the tablet form (peak concentration, 2 h). 10 Clinicians may have been unclear on the distinction between the pharmacokinetic properties of the two drug formulations and unwittingly ordered serum drug concentrations at the incorrect time. Consistent with the findings of a recent study by Ray et al, 15 not all clinicians in this study responded to low serum drug concentrations by increasing the dosages. Of all the drugs with low concentrations, clinicians tended to increase the dosages of ethambutol and ofloxacin, but not those of cycloserine and ethionamide. While the reasons for adjusting a particular drug dose often were not documented, the possible reasons for increasing the dosage of ofloxacin include superior activity at higher doses (ie, dose-related response), superior response in the treatment of MDR TB, and fewer significant adverse reaction. 16 For ethambutol, clinicians may have decided to increase the dosage, in the absence of ocular toxicity, since it is a relatively weak anti-tb drug. 10 Also, there is some evidence of higher activity at higher doses of ethambutol. 17 In not increasing the dosages of cycloserine and ethionamide immediately, clinicians may have been more concerned about adverse side effects and tolerance of these drugs. 10 While knowing the clinical benefit of TDM on patients, the cost-effectiveness of TDM for TB patients, especially for patients with MDR TB, is also important to understand. Inappropriate indications for ordering TDM can be costly. 18 Reviews of studies on TDM cost-effectiveness 14,19,20 generally have shown TDM to be cost-effective, but these studies were limited by insufficient sample size and duration, lack of patient-centered outcomes (ie, cure rate and adverse reaction rate) or economic outcomes, and lack of adequate controls. In general, TDM was suggested to be the most cost-effective when it was used selectively on patients who meet certain criteria, on drugs with specific and costly toxicities or narrow therapeutic ranges, 17,21,22 and in cooperation with an established pharmacokinetic service. 14,22 There were a number of limitations in our study. First, the nature of the retrospective study design did not enable us to examine why certain individuals did not get TDM although they were receiving the same drugs or had similar clinical conditions as those who received TDM. Second, we could not assess whether low serum drug concentrations were associated with worse clinical outcomes because of the nature of the study design and the small number of patients who had received TDM. Third, we used concentrations that were obtained at 2 h after ingestion as an approximation of the maximum concentration based on the proposed normal range. 8 If the time to reach maximum concentration was significantly longer in patients with MDR TB than in patients with suscepwww.chestjournal.org CHEST / 126 / 6/ DECEMBER, 2004 1775
tible strains, our concentrations could have been underestimated. National Jewish Medical and Research Center recommends collecting a second sample 4 h after the first, to rule out delayed absorption. Fourth, when conducting TDM, patients were not given specific instructions for eating requirements or restrictions, because we wanted to determine drug concentrations under routine clinical conditions. The potential impact of food (specifically, the diminished absorption of some drugs) for measuring specific drug concentrations may have occurred. It is known that isoniazid and rifampin may have drug concentrations lowered by food ingestion prior to drug ingestion, 23,24 and the second-line medications have not been studied as closely. However, it was expected that the impact would be insignificant clinically. Finally, the information on the rationale behind ordering tests and the response to the drug concentrations by physicians was limited by lack of specific documentation. While the current guidelines for treatment of TB by the American Thoracic Society, the Centers for Disease Control and Prevention, and the Infectious Diseases Society of America 25 recommend the use of TDM in the management of patients with MDR TB with second-line drugs, our experience with TDM in MDR TB patients reflects the way in which TDM was actually used in a TB control program setting and may be useful for other programs making decisions about its use. The results of our study demonstrate areas in the TDM process where changes might be made to make its use cost-effective. Several quality-management areas have been suggested by others, 18 such as clinical indications for testing, the timing and collection of specimens, the performance of analysis, the interpretation of the results, the initiation of appropriate action, and the impact of patient care. Based on our experience, TDM for patients with TB should be used when there is clinical suspicion of toxicity from potentially high doses, when there is suspicion of treatment failure from doses that may be too low, or when results will be used to guide the dosing of the medications to achieve the recommended concentration because the therapeutic/toxic ratio is narrow for many drugs. Samples should be collected 2 h postingestion for most anti-tb medications, and all TDM results should be documented clearly in the medical records, including the interpretation of the results and the action taken for patient management. ACKNOWLEDGMENT: We thank Dr. William Burman for his thoughtful review and critical comments. References 1 Peloquin CA. Therapeutic drug monitoring: principles and applications in mycobacterial infections. Drug Ther 1992; 22:31 36 2 Hong Kong Chest Service/BMJ Research Council. Five-year follow-up of a controlled trail of five 6-month regimens of chemotherapy for pulmonary tuberculosis. Am Rev Respir Dis 1987; 136:1339 1342 3 Long MW, Snider DE, Farer LS. U.S. public health service cooperative trial of three rifampin-isoniazid regimens in treatment of pulmonary tuberculosis. Am Rev Respir Dis 1979; 119:879 894 4 Kimerling ME, Phillips P, Patterson P, et al. Low serum antimycobacterial drug concentrations in non-hiv infected tuberculosis patients. Chest 1998; 113:1178 1183 5 Frieden TR, Sterling T, Pablos-Mendez A, et al. The emergence of drug-resistant tuberculosis in New York City. N Engl J Med 1993; 328:521 526 6 Simone PM. Multidrug-resistant tuberculosis. In: Andriole VT, ed. Mediguide to infectious diseases. New York, NY: Lawrence DellaCorte Publications, 1994; 1 9 7 Iseman MD. Treatment of multidrug resistant tuberculosis. N Engl J Med 1993; 329:784 791 8 Peloquin CA. Using therapeutic drug monitoring to dose the antimycobacterial drugs. Clin Chest Med 1997; 18:79 87 9 Peloquin CA. Pharmacological issues in the treatment of tuberculosis. Ann N Y Acad Sci 2001; 953:157 164 10 Peloquin CA. Therapeutic drug monitoring in the treatment of tuberculosis. Drugs 2002; 62:2169 2183 11 Kent PT, Kubica GP. Public health mycobacteriology: a guide for the level III laboratory. Atlanta, GA: Division of Laboratory Training and Consultation, Laboratory Program Office, US Department of Health and Human Services, Public Health Service, Centers for Disease Control, 1985 12 Schumacher GE, Barr JT. Economic and outcome issues for therapeutic drug monitoring in medicine. Ther Drug Monit 1998; 20:539 542 13 Bootman JL, Wertheimer AI, Zaske D, et al. Individualizing gentamicin dosage regimens in burn patients with gram-negative septicemia: a cost-benefit analysis. J Pharm Sci 1979; 68:267 272 14 Gardner DM, Hardy BG. Cost-effectiveness of therapeutic drug monitoring. Can J Hosp Pharm 1990; 1:7 12, xii 15 Ray J, Gardiner I, Marriott D. Managing antituberculosis drug therapy by therapeutic drug monitoring of rifampicin and isoniazid. Intern Med J 2003; 33:229 234 16 Berning SE. The role of fluoroquinolones in tuberculosis today. Drugs 2001; 61:9 18 17 Albert RK, Sbarbaro JA, Hudson LD, et al. High-dose ethambutol: its role in intermittent chemotherapy; a six-year study. Am Rev Respir Dis 1976; 114:699 704 18 Schoenenberger RA, Tanasijevic MJ, Jha A, et al. Appropriateness of antiepileptic drug level monitoring. JAMA 1995; 274:1622 1626 19 Schumacher GE, Barr JT. Total testing process applied to therapeutic drug monitoring: impact on patients outcomes and economics. Clin Chem 1998; 44:370 374 20 Schumacher GE, Barr JT. Therapeutic drug monitoring: do the improved outcomes justify the costs? Clin Pharmacokinet 2001; 40:405 409 21 Destache CJ. Use of therapeutic drug monitoring in pharmacoeconomics. Ther Drug Monit 1993; 15:608 610 22 Bertrino JS, Rodvold KA, Destache CJ. Cost considerations in therapeutic drug monitoring of aminoglycosides. Clin Pharmacokinet 1994; 26:71 81 23 Peloquin CA, Namdar R, Singleton MD, et al. Pharmacokinetics of rifampin under fasting conditions, with food, and with antacids. Chest 1999; 115:12 18 24 Peloquin CA, Namdar R, Dodge AA, et al. Pharmacokinetics of isoniazid under fasting conditions, with food, and with antacids. Int J Tuberc Lung Dis 1999; 3:703 710 25 Centers for Disease Control and Prevention. Treatment of tuberculosis: American Thoracic Society, CDC, and Infectious Diseases Society of America. MMWR Morb Mortal Wkly Rep 2003; 52(RR-11):44 1776 Clinical Investigations