Letter to the Editor

Clinical Chemistry 60:3 559–560 (2014)

Practice-Oriented Quality Specifications for Therapeutic Drug Monitoring To the Editor: The introduction of biological criteria has transformed quality control from descriptive process statistics to control of fulfillment of quality requirements on the basis of normally observed biological variation of analytes (1 ). Because drugs are xenobiotic, it is difficult to determine “biological” variation. Quality requirements remain contingent on achievable instrument performance, in the most optimistic case based on best practice criteria, or on arbitrary rules. An article by Fraser, published in 1987 (2 ), tried to address this gap (3, 4 ). Fraser estimated the biological variation for the drug concentration on the basis of fundamental pharmacokinetic theory, assuming a first order kinetics of elimination, rapid absorption as compared to elimination, a single compartmental system, and constant half-life and dosing regimens. For drugs with ⬎2 doses a day and/or long half-lives, these requirements were much too stringent for

instrument performance (desirable CV phenobarbital ⱕ1%). Even with current analytical systems, these CVs are often unrealistic and contrast with variations in consecutive drug levels observed in clinical practice. Moreover, Fraser took the range from the steady-state maximum plasma concentration to the steady-state minimum plasma concentration as biological variation. In clinical practice, however, the drug level is measured at a specific time point (e.g., trough concentrations). The pharmacological variability at this time point is probably different. To overcome the abovementioned limitations, we chose a realistic and practice-oriented approach. For several drugs that are routinely monitored in our lab, we retrospectively checked our laboratory information system for the past 7 months (January 1, 2013, to July 31, 2013) to find patients with 2 consecutive drug level determinations. To be included, the results for both determinations had to be within the therapeutic interval reported for the drug and/or the interpretative commentary on the report had to suggest continuation of the same dose for both determinations. As for most anti-epileptic

drugs, a personalized seizure-free target concentration rather than a therapeutic interval is clinically used, and we report only a cutoff for toxicity. Measured concentrations below the limit of quantification were not taken into account, and differences of ⬎75% between 2 consecutive measurements, unlikely to be caused by biological variation, were omitted from analysis. Both determinations were made by the same method on the same instrument. Except for lamotrigine, all measurements were performed on a Cobas 8000 c702 module (Roche Diagnostics). Desirable CV, bias, and allowable total error were calculated (1 ). Within- and between-patient variability was calculated using standard protocols (5 ). Analytical variability (CVa) was determined by the monthly instrument CV on lowquality control and subtracted from the total within-person (CVI) and between-person (CVG) variation (to2 2 tal CVI/G – CV2a ⫽ CVI/G ) (Table 1). Our approach has some shortcomings. First, this method determines not only biological intraand interindividual variability, but also includes elements of preanalytical variation (e.g., sampling

Table 1. Quality specifications for desirable measurement imprecision (I), bias (B) and total error allowable (TEa) on the basis of observed within-person (CVI) and between-person variation (CVG), corrected for analytical variation (CVa). Observed variation

Mean concentration (SD)

Analyte

b

CVa (%)

Interpretative comment or therapeutic interval

First measurement

Second measurement

Median time interval (range)

CVI (%)

CVG (%)

I B (%) (%)

TEa (%)

Phenobarbital

20

3.7

Toxic trough ⬎50 mg/L

25.5 (8) mg/L

26.8 (9.4) mg/L

8 days (1–181)

19.8

26.4

9.9

6.6

22.9

Phenytoin

56

6.2

Toxic trough ⬎30 mg/L

12.8 (5.3) mg/L

14.2 (5.6) mg/L

1 day (1–14)

19.0

34.6

9.5

8.7

24.4

36

5.3

Toxic trough ⬎15 mg/L

7.4 (2.6) mg/L

8.5 (2.6) mg/L

14 days (1–189)

25.5

20.7

12.8

5.2

26.3

Valproic acid

178

9.0

Toxic trough ⬎150 mg/L

64.5 (22.2) mg/L

63.5 (21.9) mg/L

37 days (1–189)

14.9

28.4

7.4

7.1

19.3

Lamotriginea

19

5.2

Therapeutic: 3–14 mg/L

8.1 (2.9) mg/L

50 days (4–146)

16.2

29.8

8.1

7.5

20.9

Lithium

29

6.7

Therapeuticb: 0.60–1.20 mmol/L

0.82 (0.14) mmol/L

0.86 (0.13) mmol/L

30 days (1–104)

10.4

5.2

2.1

10.7

Digoxin

52

5.2

Therapeutic: 0.80–2.00 ␮g/L

1.21 (0.33) ␮g/L

1.16 (0.25) ␮g/L

18 days (1–182)

17.4

8.7

4.0

18.4

Carbamazepine

a

n

Desirable specifications

8.7 (2.9) mg/L

8.22 16.0

Measured by liquid chromatography–tandem mass spectrometry. Therapeutic in the acute phase ⫽ 0.80 –1.20 mmol/L and in the chronic phase ⫽ 0.60 – 0.80 mmol/L; toxicity possible from ⬎1.50 mmol/L.

559

Letter to the Editor time). However, this result reflects real-life clinical practice variation at our hospital. Second, we assumed that no changes in therapeutic regimen occurred between 2 good plasma concentration measurements (coadministration of inducers/inhibitors, change in dose regimen), and patient variables did not substantially change in the period between the 2 measurements (liver or kidney function). Third, we did not distinguish dosing regimens. Some dosing regimens might yield lower variability than others. In conclusion, we determined realistic, practice-oriented criteria for bias, precision, and allowable total error for therapeutic drug monitoring. Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the

560 Clinical Chemistry 60:3 (2014)

Steven Pauwels1,2* Pieter Vermeersch1,2 Koen Desmet1,2 Florent Vanstapel1,3

following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article. 1

Authors’ Disclosures or Potential Conflicts of Interest: No authors declared any potential conflicts of interest.

References 1. Ricos C, Alvarez V, Cava F, Garcia-Lario JV, Hernandez A, Jimenez CV, et al. Current databases on biologic variation: pros, cons and progress. Scand J Clin Lab Invest 1999;59:491–500. 2. Fraser CG. Desirable standards of performance for therapeutic drug monitoring. Clin Chem 1987;33:387–9. 3. Jenny RW. Analytical goals for determination of theophylline concentration in serum. Clin Chem 1991;37:154 – 8. 4. Harris EK, Boyd JC. Statistical bases of reference values in laboratory medicine. New York: Marcel Dekker; 1995. Chapter 9, Analytical goals in therapeutic drug monitoring; p 299 –308. 5. CLSI. Evaluation of precision performance of quantitative measurement method; approved guideline— second edition. Wayne (PA): CLSI; 2004.

Clinical Department of Laboratory Medicine University Hospitals Leuven, Leuven, Belgium 2 Department of Cardiovascular Sciences KU Leuven, Leuven, Belgium 3 Department of Public Health, KU Leuven Leuven Belgium

* Address correspondence to this author at: University Hospitals Leuven, Laboratory Medicine Herestraat 49, B-3000 Leuven, Belgium Fax ⫹32-16-34-79-31 E-mail [email protected] Previously published online at DOI: 10.1373/clinchem.2013.217125

Practice-oriented quality specifications for therapeutic drug monitoring.

Practice-oriented quality specifications for therapeutic drug monitoring. - PDF Download Free
52KB Sizes 0 Downloads 0 Views