A subset of mothers participating in KiBS with a CD4 count of >250 cells/μl received nelfinavir mesylate (1,250 mg twice daily; Viracept, Hoffman-La Roche Ltd, Germany) and zidovudine-lamivudine (300 mg/150 mg twice daily; Combivir, GlaxoSmithKline, United Kingdom) (11
). Maternal plasma, breast milk, and infant dried blood spots were collected at delivery and postnatal weeks 2, 6, 14, and 24 from a nonrandom subset of sequentially enrolled subjects participating in a breast milk substudy (7
). Nelfinavir was dispensed in pill bottles with MEMS (medication event monitoring system) caps (Aardex, Ltd., Union City, CA), which were used to determine the dosing times for the last maternal antiretroviral dose prior to sampling; timing was confirmed by pharmacy staff. Prior to analysis, plasma and breast milk specimens were stored at −70°C and dried blood spots were stored at −20°C. Included in this analysis were mother-infant pairs where maternal plasma, breast milk, and infant dried blood spot specimens were available.
Nelfinavir and M8 from plasma and breast milk (100 μl) were extracted with 370 μl of acetonitrile containing 270 ng/ml of internal standard (nelfinavir-d3; Toronto Research Chemicals Inc., Toronto, Canada). After a brief centrifugation to precipitate suspended proteins, the liquid was transferred to a 96-well plate and evaporated to almost dryness in a vacuum concentrator, after which 100 μl of buffer A (0.57 ml glacial acetic acid–0.40 ml formic acid–0.40 ml ammonium hydroxide in 1 liter of water) was then added. The solvent was again evaporated to almost dryness in a vacuum concentrator, and another aliquot of 100 μl of buffer A was added. The sample plate was then sealed and stored at −20°C for future analysis by liquid chromatography-mass spectrometry. Six-millimeter-diameter punches from infant dried blood spot samples were extracted with 370 μl of acetonitrile containing 270 ng/ml of internal standard and sonicated for 30 min at room temperature, and the liquid was removed to a 96-well plate that was then treated in like manner. Standard samples of known concentrations for both nelfinavir and M8 (Toronto Research Chemicals Inc.) were spiked into the appropriate matrices and treated in like manner for the construction of the standard curve used in quantification. Standard and quality control samples for blood spot analysis were created as described by Mei et al. (6
). To construct the standard curves, 6-mm-diameter punches from the standard dried blood spot samples were extracted with 370 μl of acetonitrile containing 270 ng/ml of internal standard and sonicated for 30 min at room temperature, and the liquid was removed to a 96-well plate also containing unknown samples and treated as described above. Nelfinavir and M8 are fairly insoluble in aqueous buffers, so recovery from breast milk and plasma was approximately twice as efficient (with 25% variability) as that from purely aqueous medium. Recovery from dried blood spots was enhanced by approximately 50% when acetonitrile was used as the elution buffer, while matrix effects were somewhat less.
Five microliters of the final processed sample was injected onto a 2- by 30-mm Ascentis C8 reverse-phase column (Supelco Inc., Bellefonte, PA) connected to a Shimadzu (Durham, NC) Prominence high-performance liquid chromatography (HPLC) system. An HPLC method consisting of a 3-min gradient from 5% to 90% buffer B (0.57 ml glacial acetic acid–0.40 ml formic acid–0.40 ml ammonium hydroxide in 1 liter of acetonitrile), followed by isocratic runs of 1 min at 90% buffer B and then 3 min at 5% buffer B at a flow rate of 0.5 ml/min, was used. A model 3200 QTrap mass spectrometer (AB Sciex, Foster City, CA) working in the positive multiple-reaction monitoring mode was used to monitor the transitions for the compounds of interest. For nelfinavir, m/z 568.3 → 135.1 and 568.3 → 330.2 were monitored, and for M8, m/z 584.2 → 135.0 and 584.2 → 330.2 were monitored. The limit of quantification (LOQ) for plasma and breast milk was 10 ng/ml, and that for dried blood spots was 5 ng/ml. Assay of infant dried blood spots was discontinued after 28 specimens due to the predominance of samples at less than the LOQ. Inter- and intraday variability for both analytes and each matrix was less than 10% at the LOQ. The linear range for this assay was 0.01 ng/ml to 3,000 ng/ml.
The median nelfinavir and M8 concentrations in plasma and breast milk were calculated for each study visit. The ratio between matched maternal breast milk and plasma concentrations (breast milk/plasma ratio) was calculated for each period. We excluded from the calculation of the breast milk/plasma ratio specimens where the maternal breast milk or plasma concentrations were less than the LOQ. Linear regression and Spearman-rank correlation statistics for trends were used to estimate the change in nelfinavir and M8 concentrations in maternal plasma and breast milk after the most recent nelfinavir dose. Analyses were performed using the SAS software program, version 9.2 (SAS Institute, Cary, NC).