Use of an Acetyl Derivative to Improve GC–MS Determination of Norbuprenorphine in the Presence of High Concentrations of Buprenorphine in Urine

Abstract

Certain patients being treated with Suboxone™ or Subutex™ can exhibit very high buprenorphine and low norbuprenorphine concentrations in urine. Very high buprenorphine can interfere with buprenorphine-D4 used as an internal standard, causing errors in norbuprenorphine determination by gas chromatography–mass spectrometry (GC–MS). We used a modified method of Wu et al. to introduce norbuprenorphine-D3 as a separate internal standard for norbuprenorphine. This allowed us to accurately measure norbuprenorphine in neat urine specimens when buprenorphine is present in extremely high concentrations. Laboratories measuring buprenorphine and metabolite by GC–MS may face this problem if their clientele includes patients being treated with other medications that interfere with the cytochrome p450 CYP 3A4-mediated conversion of buprenorphine to norbuprenorphine.

Introduction

Buprenorphine, an active constituent of Suboxone^® and Subutex^®, has gained widespread recognition as an effective drug in the treatment of opioid addiction (1, 2). Urine drug testing is recommended to support the clinical assessment and treatment plan for patients who are prescribed buprenorphine (3). In urine, large percentages of total drug eliminated typically exist as the metabolite norbuprenorphine, as well as the parent buprenorphine or their glucuronides (4, 5).

Immunoassays exist for screening purposes and at least one (Cloned Enzyme Donor Immunoassay, or CEDIA) detects only parent drug and not metabolite (6). However, given the consequences of a result indicating unexpected non-compliance, it is often necessary to confirm screen results by liquid chromatography–tandem mass spectrometry (LC–MS-MS) or gas chromatography–mass spectrometry (GC–MS).

GC–MS methods have been developed for the analysis of buprenorphine and norbuprenorphine in whole blood (7), plasma (8) and urine typically using deuterated buprenorphine (buprenorphine-D4) as an internal standard (9, 10) though some workers have included norbuprenorphine-D3 (11, 12). One of the challenges for laboratories performing buprenorphine analysis by GC–MS method is the case of the unusual patient urine specimen with very high buprenorphine concentrations (>25,000 ng/mL) and low but detectable concentrations of norbuprenorphine (<10 ng/mL). Owing to the limited dynamic range of the GC–MS assay, such specimens must be diluted extensively in order to measure buprenorphine, then analyzed undiluted (neat) to measure norbuprenorphine. However, the very high concentrations of buprenorphine present in the neat samples interfere with the buprenorphine-D4 internal standard, increasing the apparent response from 2-fold to 10-fold. As drug concentration is based on a ratio of drug to internal standard, this interference results in an erroneous measurement of norbuprenorphine.

This problem could be avoided by the use of a second internal standard for norbuprenorphine. Ideally, this would be a deuterated norbuprenorphine (norbuprenorphine-D3). However, the typically used trimethylsilane (TMS) derivatives produce a base peak at 468 m/z for both norbuprenorphine and norbuprenorphine-D3, making them indistinguishable (13, Figure 3). The use of other mass fragments is limited by the analytical sensitivity required for this assay.

Figure 3.

Mass spectra of derivatized, single compound standards of norbuprenorphine from methanolic stocks. Norbuprenorphine-TMS (A), norbuprenorphine-D3-TMS (B), norbuprenorphine-acetyl (C) and norbuprenorphine-D3-acetyl (D) are shown. The base peak for both TMS derivatives is 468 m/z. The base peak for the acetyl derivatives is 440 m/z for norbuprenorphine and 443 m/z for norbuprenorphine-D3.

Wu et al. (14) evaluated several different derivatization schemes for buprenorphine and norbuprenorphine. They showed that an acetyl derivative of these compounds allowed for the use of norbuprenorphine-D3 as an internal standard. We report here on an analytical method for the determination of buprenorphine and norbuprenorphine using an acetyl derivative, in which both buprenorphine-D4 and norbuprenorphine-D3 are present as internal standards. We demonstrate the utility of this approach in cases of high buprenorphine, low norbuprenorphine as described above.

Methods

Samples

Samples for analysis were either patient samples (as indicated) or negative urine samples (Certified Drug-Free Urine from Utak Laboratories, Inc., Valencia, CA, USA) fortified with commercially available stocks of buprenorphine, buprenorphine-glucuronide and norbuprenorphine (Cerilliant Corporation, Round Rock, TX, USA). Patients were from addiction management clinics. All patient specimen identifiers were removed prior to the study.

Solid-phase extraction

About 5.0 mL of urine was fortified with internal standard (a methanolic solution of 2.0 µg/mL each burpenorphine-D4 and norbuprenorphine-D3 (Cerilliant) to a final concentration of 25 ng/mL). pH was adjusted by the addition of 0.7 mL of 0.2 M sodium acetate buffer, pH 5.0. Samples were hydrolyzed by the addition of 50 µL 100,000 U/mL β-glucuronidase (Campbell Science, Rockford, IL, USA) and incubated for 2 h at 60°C. pH was then adjusted to 2–3 by the addition of 0.5 mL 10% HCl. Hydrolyzed samples were extracted on a GV-65C solid-phase extraction column (Biochemical Diagnostics, Inc., Edgewood, NY, USA) that had been pretreated with 1.0 mL of methanol. After the samples passed through the column, the column was washed with 3.0 mL of 0.01% HCl, followed by 2.0 mL of methanol, then 1.0 mL of ethyl acetate. Drug was eluted by the addition of 2.5 mL elution cocktail (mixture of 80% n-butyl chloride and 20% ethyl acetate : triethylamine, 96 : 4). Extracts were evaporated to dryness at 45°C under a stream of nitrogen. Dried extracts were resuspended in 500 µL of acetic anhydride and 200 µL of pyridine for derivatization. Extracts were incubated at 70°C for 30 min, then evaporated to dryness at 45°C under a stream of nitrogen. Evaporated, derivatized extracts were resuspended in 200 µL of ethyl acetate, and transferred to GC vials containing glass inserts.

Extraction of TMS derivatives was identical, except for the following: (i) norbuprenorphine-D3 was excluded from the internal standard; (ii) after evaporation of extracts, derivatives were formed by the addition of 100 µL of N,O-bis(trimethylsilyl)trifluoroacetamide + 1% trimethylchlorosilane (Campbell Bioscience) and 100 µL of ethyl acetate, and incubated at 70°C for 30 min before injection into the GC–MS.

GC–MS analysis

For acetyl derivatives, 2 µL of sample was injected (splitless mode, injection port temperature = 250°C). The gas chromatograph was a model 7890 (Agilent Corporation, Santa Clara, CA, USA) and the mass spectrometer was a model 5975 (Agilent). The oven program was as follows: initial temperature = 150°C, initial time = 0.50 min; ramp = 30°C/min; final temperature = 320°C; final time= 3.00 min. The chromatography column was a ZB-5MS (Phenomenex, Inc., Torrence, CA, USA) with dimensions of 15 m in length, 250 µm in diameter, with a film thickness of 0.25 µm. The carrier gas was He, with a constant flow of 1.0 mL/min. The mass spectrometer was run in SIM mode, and the following ions (m/z) were used for quantitation (underlined) and identification: buprenorphine, 420, 452, 394; buprenorphine-D4, 424, 456, 398; norbuprenorphine, 440, 380, 422 and norbuprenorphine-D3, 443, 444, 425.

Quantitation was based on a five-point calibration curve extrapolated to the y-intercept. Calibration was considered acceptable if the correlation coefficient was ≥0.990. Calibrators containing buprenorphine and norbuprenorphine at 5, 10, 20, 50 and 100 ng/mL were run with each batch of samples. GC–MS method control, data compilation and calculations were performed using ChemStation^® software (Agilent Corporation).

For TMS derivatives, conditions were identical except that the following ions (m/z) were used for quantitation (underlined) and identification: buprenorphine, 450, 482, 506; buprenorphine-D4, 454, 486, 510 and norbuprenorphine, 468, 524, 510.

Assessment of results

Precision and limit of quantitation were considered acceptable if the standard deviation (SD, measured in ng/mL) was ≤1/3 allowable error (allowable error = 20%). An assay was considered linear if the square of the correlation coefficient was ≥0.990 (15). The same standard was used to assess comparison studies.

Scanning analysis of derivatized buprenorphine, norbuprenorphine and their deuterated forms

Negative drug-free urine was fortified to 1.0 µg/mL with one compound (buprenorphine, norbuprenorphine, buprenorphine-D4 or norbuprenorphine-D3). Samples were extracted according to the methods described above. Each sample was analyzed by the identical GC program described, except that the MSD was run in a scan mode.

Results

Interference with buprenorphine-D4 from very high buprenorphine

Figure 1 shows the response of buprenorphine-D4 in the presence or absence of very high buprenorphine. Using the TMS derivative, buprenorphine-D4 was substantially increased in the presence of high buprenorphine in the patient samples. The same was evident when using the acetyl derivative. However, norbuprenorphine-D3 was unaffected by high buprenorphine. The practical effect of this is seen in Figure 2: norbuprenorphine concentrations were substantially higher when using the acetyl derivatization scheme compared with the TMS derivative. This finding can be explained by the presence of norbuprenorphine-D3 as the internal standard for norbuprenorphine, which was unaffected by very high concentrations of buprenorphine.

Figure 1.

Mass spectrometer responses for deuterated internal standards under different methods. Response was of the Q ion in each case. ‘Normal’ responses refer to the mean of the five standards and three quality control samples. High buprenorphine patient responses are mean buprenorphine-D4 or norbuprenorphine-D3 responses from five patients with very high buprenorphine and low norbuprenorphine. All analyses are on neat samples.

Figure 2.

Analysis of urine from five patients with very high buprenorphine and low norbuprenorphine by both TMS and acetyl derivatization methods for norbuprenorphine. The acetyl derivatization method resulted in higher concentrations of norbuprenorphine, likely due to the lack of interference with the internal standard (Figure 1).

Acetyl derivative versus TMS derivative of norbuprenorphine

Mass spectra of TMS derivatives are shown in Figure 3A and B. The data argue that norbuprenorphine cannot be distinguished from norbuprenorphine-D3 when derivatized in this way, using the base peak of 468 m/z. Figure 3C and D shows mass spectra of acetyl derivatives norbuprenorphine and norbuprenorphine-D3. With some minor exceptions, these data agree with that of Wu et al. (14). With this derivatization scheme, norbuprenorphine-D3 base peak differed from norbuprenorphine base peak and could be distinguished from norbuprenorphine. It was therefore suitable as an internal standard for norbuprenorphine.

Precision

Using the acetyl derivatization method, two levels of control were measured in replicates of five separate extractions on each of 5 separate days. Table I summarizes these data. In every case, the precision for the assay meets the requirement of SD ≤ allowable error (20%).

Table I.

Day-to-Day Reproducibility of the GC–MS Method for Bup and Nor-Bup

Summary	Buprenorphine		Norbuprenorphine
	Level 1	Level 2	Level 1	Level 2
N	25	24a	25	24a
Mean (ng/mL)	4.0	17.5	4.8	24.2
SD (ng/mL)	0.21	0.49	0.29	0.69
CV (%)	5.35	2.82	5.99	2.86
SD ≤ Ea/3?	Yes	Yes	Yes	Yes

SD, standard deviation; CV, coefficient of variation; Ea, allowable error (20%); GC–MS, gas chromatography–mass spectrometry.

^aOne outlier removed due to pipetting error; same tube for both analytes.

Linearity/recovery

Serial dilutions were performed from a 100-ng/mL standard. Quantitative results are graphed in Figure 4. Squares of correlation coefficients (R²) and the equations for the lines are shown for buprenorphine (Figure 3A) and norbuprenorphine (Figure 3B). R² meets a recommended target of >0.990 (15).

Figure 4.

Linearity/recovery of serially diluted duplicate urine specimens fortified with buprenorphine (A) and norbuprenorphine (B).

Limit of detection/limit of quantitation

Limit of detection (LOD) was defined as the lowest concentration where both detection and identification criteria are met. We found that this was very close to the limit of quantitation (LOQ) since we were not able to consistently detect buprenorphine or norbuprenorphine at <LOQ without qualifier ions exceeding tolerance limits. Therefore, we conclude that LOD = LOQ. The LOQ was defined as the lowest concentration at which total SD ≤ allowable error/3. Allowable error for this assay was 20%. LOQ data are summarized in Table II. It was determined by measuring successively lower concentrations of fortified urine, in replicates of five, on 2 different days.

Table II.

LOQ Achieved with the GC–MS Method for Bup and Nor-Bup

Summary	Buprenorphine LOQ	Norbuprenorphine LOQ
N	10	10
Mean (ng/mL)	2.3	2.3
SD (ng/mL)	0.07	0.04
SD ≤ Ea/3?	Yes	Yes

LOQ, limit of quantitation; GC–MS, gas chromatography–mass spectrometry; SD, standard deviation.

Specificity

The acetyl derivative method showed no false positives or interfering peaks when used to assay a commercial high concentration multidrug quality control sample (Liquicheck Urine Toxicology Control, level C4; Bio-Rad, Hercules, CA, USA) or when used to assay urine fortified with a mix containing acetaminophen, salicylate, ephedrine, naproxen, ibuprofen, dextromethorphan (each at a final concentration of 10 µg/mL) and caffeine (at a final concentration of 7.4 µg/mL). Figure 5 shows correlation of acetyl derivatives with the TMS derivatives, over a range of commonly encountered values. Correlation between the two methods was excellent, indicating that the acetyl method is applicable to all sample types, not just those with very high buprenorphine and low norbuprenorphine.

Figure 5.

Comparison of buprenorphine (A) and norbuprenorphine (B) values in the normally expected range of values. Twenty patient samples were split and extracted separately by each method as described. Equation for the line and correlation coefficient are shown.

Discussion

In practice, our laboratory dilutes specimens that screen positive for buprenorphine by immunoassay based on the absorbance rate value reported by the analyzer. Therefore, high buprenorphine samples like the ones described here would be extensively diluted (50×). However, this extensive dilution resulted in a norbuprenorphine concentration less than the LOQ for the assay. When run neat, the very high buprenorphine interfered with the internal standard (buprenorphine-D4), causing a higher response and therefore an erroneously low norbuprenorphine measurement. Intermediate dilutions were not successful in resolving the issue. Therefore, use of a second internal standard specific to norbuprenorphine was needed. The method of Wu et al. (14), slightly modified, was applied to this problem, relating to a specific subset of Suboxone^® patients.

Wu et al. (14) did not recommend the use of an acetyl derivative in practice where an LOQ of ≤10 ng/mL is needed. However, in our hands, the assay achieved an LOQ of ∼2.0 ng/mL. This is less than the common screening cutoff of 5 ng/mL and makes the assay eminently useful for confirmation of buprenorphine and/or metabolite. The fact that we used 5 mL of urine as opposed to 2 mL (14) may explain our lower LOQ. Another factor may be the use of solid-phase extraction as opposed to the liquid–liquid approach used by Wu et al. (14). Precision, linearity and recovery experiments argue that the acetyl method performs well and meets normal laboratory criteria for acceptance. That the acetyl method yields essentially equivalent results to the TMS method in a comparison study argues that laboratories may convert to this scheme for all buprenorphine/norbuprenorphine analysis by GC–MS.

The problem of very high buprenorphine/very low norbuprenorphine is unusual. A laboratory may face significant numbers of these cases or very few depending on the location and clientele. It has been estimated that in 2008, there were approximately 80,000–100,000 individuals living with AIDS who acquired HIV from intravenous drug use (16). It is reasonable to assume that a significant portion of those individuals would be undergoing both AIDS therapy and opioid addiction therapy. If AIDS therapy includes drugs of the protease inhibitor class, these patients could exhibit the very high buprenorphine/low norbuprenorphine urine profile described here.

Other possible causes of the very high buprenorphine/low norbuprenorphine profile in urine have been investigated (17). Buprenorphine is metabolized principally to norbuprenorphine by cytochrome p450 CYP 3A4 (18, 19). It is possible that any inhibitor of CYP3A4 may inhibit the conversion of buprenorphine to norbuprenorphine and result in a very high buprenorphine/low norbuprenorphine concentration in urine (13, 20–22). Therefore, laboratories performing buprenorphine and norbuprenorphine analysis by GC–MS must be prepared to encounter these types of specimens and use methods that can report accurate values for buprenorphine and norbuprenorphine. The use of the acetyl derivatization method can accomplish this.