Liquid chromatography-mass spectrometry measurement of tacrolimus in finger-prick samples compared with venous whole blood samples

Abstract

Background

Finger-prick sampling is an alternative strategy for monitoring immunosuppressive drug concentrations that could be useful in reducing outpatient visits. We investigated the correlation between venous and finger-prick samples in a group of adult thoracic transplant patients.

Methods

Blood samples (n = 65) for the measurement of whole blood tacrolimus were collected from adult heart (n = 18) and lung transplant (n = 20) recipients receiving tacrolimus-based immunosuppressive therapy and a finger-prick sample was taken at the same time.

Results

Between-batch assay imprecision (coefficient of variation [CV], %) for the last 12 months (n = 270) at concentrations of 3.5, 6.9 and 13.9 μg/L was 8.0%, 5.4% and 5.2%, respectively. Passing and Bablok regression analysis between finger-prick and venous blood showed finger-prick tacrolimus = 1.02 (venous blood tacrolimus) −0.06. Bland–Altman analysis showed good agreement with a bias of 0.114 μg/L and 95% limits of agreement from −1.0 to 1.2 μg/L.

Conclusions

The liquid chromatography mass spectrometry methodology that we have developed has the potential to allow patients or their carers to collect finger-prick blood samples at home and send them to the laboratory using the routine mail service. We believe that finger-prick blood sampling has an important role to play in the care of transplant patients receiving immunosuppressive drugs, including tacrolimus.

Introduction

Regular monitoring of the immunosuppressive drug tacrolimus is necessary to tailor the therapeutic regimen in transplant recipients maintained on this drug. Blood sampling is usually performed at the time of routine clinical visits and the recommended sample type is whole blood anticoagulated with ethylenediaminetetraacetic acid (EDTA). Alternative strategies for blood sampling, such as finger-prick collection, have also been investigated for ciclosporin and tacrolimus with some success. Ciclosporin measured in a finger-prick blood sample has been shown to correlate well with venous whole blood and this approach offers the possibility of using the postal service to send samples from home to the central laboratory for analysis. ^1–3 This can significantly reduce the number of clinical visits that patients and their families would normally have to make in order for drug concentrations to be monitored. In our transplant clinic, we have seen a fall in outpatient appointments of approximately 10–15% since we introduced finger-prick sampling for ciclosporin.

We have previously shown that tacrolimus in finger-prick blood correlates well with tacrolimus in venous blood taken at the same time in paediatric renal transplant patients, but we wanted to extend this work to test the correlation in adult heart and lung transplant recipients. ⁴

Methods

Blood samples (n = 65) for the measurement of whole blood tacrolimus were collected from adult heart (n = 18) and lung transplant (n = 20) recipients receiving tacrolimus-based immunosuppressive therapy and a finger-prick sample was taken at the same time. Sampling was performed at the time of their outpatient review or while they were an inpatient on the ward. Fourteen patients had more than one sample taken on separate occasions. Median age of the patients was 59 y (range 19–71 y). Finger-prick blood samples were collected using the Softclix^® Pro lancing device (Roche Diagnostics Ltd, Lewes, UK). Samples were collected from the fingertip, which was cleansed with an isopropanol 70% v/v impregnated swab and thoroughly dried. The use of petroleum jelly as an aid to sampling was avoided because of concern that this might absorb tacrolimus and thus artefactually reduce finger-prick blood tacrolimus concentrations. Venous blood samples were collected by standard venepuncture into K₂EDTA 13 × 75 BDVacutainer^® tubes. A tourniquet was tied to the arm before venepuncture and released prior to withdrawal of the needle. Finger-prick samples were collected into K₂EDTA BD Microtainer^® tubes with Microgard (BD, Oxford, UK).

For the liquid chromatography mass spectrometry (LC-MS/MS) assay, samples were prepared in a 96-deep well microtitre plate by adding 10 μL of blood to 40 μL of 0.1 mol/L zinc sulphate solution (VWR international, Lutterworth, UK). ⁵ Proteins were precipitated by adding 100 μL acetonitrile containing ascomycin as internal standard (Sigma–Aldrich, Poole, UK). After vigorous mixing and centrifugation (10,000 rpm for 3 min), 20 μL of the supernatant was injected into the LC-MS/MS system. A Waters MasstrakTDM™ cartridge (C18 2.1 mm × 10 mm) was eluted with a step-gradient of 50–100% methanol containing 2 mmol/L ammonium acetate (Sigma–Aldrich) and 0.1% (v/v) formic acid (Sigma–Aldrich), at 0.6 mL/min. The column was maintained at 55°C. The analytes were monitored using a Quattro micro™ mass spectrometer (Waters Corporation, Manchester, UK) operated in multiple reaction monitoring mode using the following transitions: m/z 821 > 768 (tacrolimus) and m/z 809 > 756 (ascomycin).

Results

The limit of quantitation, i.e. repeated measurement of the lowest control was 0.5 μg/L and the assay was linear up to 30 μg/L. ⁵ This method has been in routine use in our department for more than 5 y, and despite the perceived problems of handling such a small sample volume, it has proved to be very precise in our hands. Between-batch imprecision (CV, %) for the last 12 months (n = 270) at concentrations of 3.5, 6.9 and 13.9 μg/L was 8.0%, 5.4% and 5.2% respectively. Passing and Bablok regression analysis between finger-prick and venous blood showed finger-prick tacrolimus = 1.02 (venous blood tacrolimus) −0.06. Bland–Altman analysis showed good agreement with a mean bias of 0.114 μg/L (95% confidence interval [CI] −1.005 to 1.232), standard error of mean 0.0797 (95% CI −0.028 to 0.255) and 95% limits of agreement from −1.0 μg/L (95% CI −1.247 to −0.762) to 1.2 μg/L (95% CI 0.989–1.475) (Figure 1). There was no significant difference between the results in the finger-prick and venous samples using a Wilcoxon-paired ranks test (P = 0.1543).

Figure 1

Bland–Altman plot of finger-prick tacrolimus – venous blood tacrolimus. n = 65, bias (±2 standard deviations) =0.114 (−1.005 to 1.232). Dashed lines indicate the 95% limits of agreement and solid line indicates bias

Discussion

Tacrolimus measurement using LC-MS/MS is becoming increasingly popular and it is now performed by approximately 22% of laboratories participating in the International Proficiency Testing Scheme (http://www.bioanalytics.co.uk) compared with <6% 5 y ago. LC-MS/MS analysis of tacrolimus can be performed using only a 10 μL sample and it is this ability to scale-down the assay, which makes it suitable for the frequently small sample volumes obtained by the finger-prick technique. Using appropriate positive displacement micropipettes, we can handle these small blood volumes with excellent accuracy and precision.

The LC-MS/MS methodology, which we have developed, has the potential to allow patients or their carers to collect finger-prick blood samples at home and to send them to the laboratory using the routine mail service. We have tested this for ciclosporin but this still needs to be validated for tacrolimus. Home finger-prick blood sampling is already routinely performed by children and adults with diabetes mellitus. It is an easy technique to teach and uses inexpensive equipment that is widely available. We believe that finger-prick blood sampling has an important role to play in the care of transplant patients receiving immunosuppressive drugs, including tacrolimus.