Clinical Implications of the Relationship Between Naltrexone Plasma Levels and the Subjective Effects of Heroin in Humans

Abstract

Background:

Extended-Release Naltrexone (XR-NTX) is an opioid antagonist approved for relapse prevention following medical withdrawal. Its therapeutic effect is dependent on the naltrexone (NTX) plasma level, and as it decreases, patients may lack protection against relapse and overdose. Therefore, identifying the minimally effective NTX level needed to block opioid-induced subjective effects has important clinical implications.

Methods:

This secondary, individual-level analysis of data collected in a human laboratory study was conducted to evaluate the relationship between NTX levels and subjective effects of an intravenously (IV) administered 25 mg challenge dose of heroin in non-treatment-seeking participants with opioid use disorder (N=12). Subjective ratings of drug liking using a 100mm Visual Analog Scale (VAS) and NTX levels were measured across 6 weeks after participants received a single injection of either XR-NTX 192 mg (N=6) or 384 mg (N=6). Cubic spline mixed effects models were used to provide 95% prediction intervals for individual changes in liking scores as a function of NTX levels.

Results:

NTX levels above 2 ng/ml blocked nearly all VAS ratings of drug liking after IV heroin administration. Participants with NTX levels ≥2 ng/ml had minimal (≤20 mm) changes from placebo in VAS ratings of drug liking based on 95% prediction intervals. In contrast, NTX levels <2 ng/ml were associated with greater variability in individual-level subjective responses.

Conclusions:

In clinical practice, a plasma level range of 1–2 ng/ml is considered to be therapeutic in providing heroin blockade. The current findings suggest that a higher level (>2 ng/ml) may be needed to produce a consistent blockade.

Introduction

Opioid Use Disorder (OUD) is a devastating illness with a high mortality due to overdose ^1,2. In the United States, there are three currently approved medications for treating OUD: methadone, buprenorphine and naltrexone ³. Of the three, extended-release naltrexone (XR-NTX) is least utilized in the United States ⁴. Barriers to increased adoption are the induction hurdle ⁵, high cost and formulary restrictions ⁶, and high treatment dropout ^3,7,8. One reason for dropout may be risk of increased cravings and relapse due to naltrexone (NTX) plasma levels gradually decreasing over time after XR-NTX administration.

NTX is metabolized to 6-β-naltrexol, which also has mu opioid receptor antagonist properties, though it is less potent than the parent compound ⁹. Previous pharmacokinetic data show the average trough level of NTX at the end of the dosing period is 1.33 ng/ml after receiving an XR-NTX IM dose of 400 mg (Vivitrex^®, naltrexone for injectable suspension; Alkermes, Inc., Cambridge, MA), monthly for 4 months ¹⁰. However, there is wide variability in NTX blood levels with some individuals having low blood levels during the 4^th week after XR-NTX administration ¹¹. This high variability may explain why a wide range of “therapeutic” blood level has been proposed, from 1ng/ml to 10 ng/ml, based on outcomes such as reported heroin use or evidence of use in the community ^12–14, reported effects after agonist administration in the laboratory setting ¹⁵, or reported in a systematic review of randomized control trials ¹⁶.

One of the primary therapeutic mechanisms of NTX in OUD treatment relates to its ability to block the subjective effects of opioids. This is typically attained by once monthly maintenance injections after medical withdrawal ¹⁷. Therefore, maintaining the sufficient “blocking” blood level of NTX is often necessary for its clinical effectiveness ¹⁸. This raises the question of what is the minimally effective NTX plasma level needed to produce its therapeutic effect, including blockade of opioid effects. While this could be complicated by the non-linear relationship between plasma levels to mu receptor binding, especially with oral NTX administration, we do not expect it to be significant with a sustained-release formulation of NTX ^19,20. Previous studies have shown blockade of the effects of moderate doses of opioids administered in the human laboratory for most individuals for 28 days after XR-NTX administration at doses of 75mg, 150mg or 300 mg IM ²¹ and 192mg and 384 mg SC ²², whereas clinical experience suggests that some patients begin to experience craving and possibly opioid effects around 21 days after XR-NTX 380 mg IM administration ²³. These studies combined demonstrate that the ability of NTX to block opioid agonist effects is dependent on the NTX dose administered, time since NTX dose administration, and dose of opioid agonist used.

The minimum plasma levels of NTX and 6-β-naltrexol that block subjective effects of opioids has not been directly studied with parenteral XR-NTX formulations in humans. We therefore conducted a secondary analysis of a previously published pharmacokinetic and pharmacodynamic study of XR-NTX ²² where plasma levels of NTX and 6-β-naltrexol and the subjective effects of IV heroin were measured at multiple time points, allowing a direct examination of the relationship between plasma level and subjective effects of heroin.

Methods

Participants

Twelve heroin-dependent men (eight non-Hispanic Caucasian, three Hispanic, and one African American) with a mean age of 31.5 years (SEM 2.1) who were not seeking treatment for their opioid use and who were not physically dependent on other substances completed an inpatient 8-week protocol including medically supervised withdrawal and induction onto XR-NTX. Participants met criteria for opioid dependence (DSM-IV) and demonstrated opioid physical dependence through a naloxone challenge test (Wang 1974). Six participants received XR-NTX 192 mg, and six received XR-NTX 384 mg as a single subcutaneous injection. This subcutaneous formulation of XR-NTX was never marketed but showed an overall similar pharmacokinetic profile to the marketed XR-NTX (Vivitrol ^®). Participants then underwent daily challenge sessions of double-blind heroin administration in the human laboratory setting, including placebo, 6.75mg, 12.5mg, and 18.75 mg on separate, individual days before the 25 mg dose. The present secondary analysis only considered the heroin 25 mg dose, which was administered six times overall. For additional details about the study protocol please see the original paper (Comer et al., 2002).

Informed consent detailing potential study risks and benefits was obtained from all participants for the initial study (registered with ClinicalTrials.gov, NCT03205423), and this secondary analysis on de-identified data was determined by the NYSPI Institutional Review Board to be non-human subject research.

Subjective Measures

In the human laboratory setting, participants received an intravenous (IV) challenge dose of heroin 25 mg and subjective effects of the challenge dose were assessed at 4, 40, and 90 min after the dose. Participants received six challenge doses of heroin 25 mg IV across six weeks after the XR-NTX injection. Ratings of Drug Liking were used to assess the opioid-induced subjective effects. The responses were recorded on a visual analog scale (VAS; modified from Foltin and Fischman 1995). Participants rated each item on the 100 mm scale from “Not at all” (0 mm) to “Extremely” (100 mm). As has been done in prior work to increase reliability, the calculated average of the 4-, 40-, and 90-min ratings was used for each participant. Moreover, the difference between the average liking of heroin 25 mg and the average liking of placebo each week serve as the primary outcome.

Plasma Levels

Blood was drawn to measure NTX and metabolites at baseline (Day 0), 1, 2, 3, 4, 5, 6, 8, 11, 15, 18, 22, 25, 29, 32, 36, and 39 days after XR-NTX administration, and sent for analyses of NTX and 6-β-naltrexol (Center for Human Toxicology, University of Utah, Salt Lake City, Utah, USA). Analyses were performed by solid phase extraction and negative ion chemical ionization gas chromatography/mass spectrometry, as described by Huang and colleagues (1997). The lower limit of detectability for both analytes was 0.1 ng/ml.

Drugs and Dosing Procedure

Extended-release naltrexone (Depotrex^®) was manufactured by Biotek Inc. (Woburn, Mass., USA) and provided by the National Institute on Drug Abuse (NIDA). Heroin was provided by the NIDA pharmacy, with doses administered intravenously in a double-blind fashion. Heroin 25 mg IV was administered the same day NTX blood levels were drawn from participants for 59 of the 72 (82%) data points collected. The remaining 13 (18%) NTX levels were captured one day before or after the heroin 25 mg IV challenge. For details on dosing and blood draws see original paper (Comer et al., 2002).

Statistical Analyses

Descriptive plots of individual level liking scores (difference in liking between 25 mg heroin and placebo challenge each week) and individual NTX plasma levels by week were shown stratified by two doses of XR-NTX (192 and 394 mg). Both doses of XR-NTX were considered for the present analysis as they both produced blood NTX levels with antagonist effects. While the lower 192 mg dose did not result in peak values of NTX > 2.3 ng/ml, these data still captured the loss of opioid blockade as evidenced by increases in VAS ratings of liking. Individual-level heroin liking scores on a visual analog scale (VAS) were then modeled as a function of contemporaneous NTX level using cubic spline mixed effects models including a random effect for individual to account for repeated measures. A difference less than 20 mm in liking scores between heroin and placebo on this unipolar VAS scale was considered to be not clinically meaningful (i.e. this difference represents complete opioid blockade). To determine the threshold plasma levels of NTX above which most people have liking scores <20 mm we examined the 95% prediction intervals from the cubic spline mixed effect model ²⁴. Prediction intervals provide regions where we expect 95% of scores to fall for future individuals in the population. Prediction intervals are wider than confidence intervals and were used in the present analysis because confidence intervals provide the likely region within which the mean of the population falls, while here we want to identify the region where most patient scores fall rather than the average score of patients. Additional analyses re-ran the mixed-effect model stratified by the two doses of XR-NTX (192 and 394 mg). A similar analysis was also repeated for 6-β-naltrexol plasma levels. The R package glmer was used to fit the models ²⁵.

Results

Subjective Effects

Naltrexone Plasma Levels and Subjective Response to Heroin

The raw individual-level drug liking scores and NTX plasma levels collected each week by the two dosage groups are shown in Figure 1. Eleven out of the 12 participants achieved blockade of liking (i.e. difference in liking between heroin and placebo < 20 mm) for the first three weeks with varying increases in liking after that until week 6 (Figure 1 right panel). NTX levels in the high-dose group were above 2 ng/ml for all participants for the first two weeks with levels declining after that (Figure 1 left panel).

Figure 1

Figure 2 shows that lower NTX levels were associated with increased drug liking. Results from the cubic spline mixed effects model show that the 95% prediction interval was fully above 20 mm for NTX values < 2.0 ng/ml indicating that 95% of the population would not have blockade effects on liking scores for NTX level values < 2.0 ng/ml. There was substantial person-to-person variability in liking scores for NTX levels between 1.0 and 2.0 ng/ml such that some individuals had complete blockade effects in that region (i.e. liking change scores = 0) while others had liking change scores larger than 20 mm (95% prediction interval is both above and below 20). While all but one individual in the present study had liking change scores less than 20 mm when their NTX levels were between 1–2 ng/ml, the 95% prediction interval reached outside the cut-off of 20 mm indicating future individuals may have liking values larger than the clinically meaningful value of 20mm when their NTX level reaches that range.

Figure 2

6-β-Naltrexol Plasma Levels and Subjective Responses to Heroin

6-β-naltrexol plasma levels greater than 3.0 ng/ml were associated with a significant reduction in subjective ratings of drug liking for both doses of XR-NTX. Similar to the findings with NTX, drug liking generally increased as 6-β-naltrexol levels decreased, with the cubic spline mixed effects model yielding a 95% prediction interval fully above 20 mm for 6-β-naltrexol levels < 3.0 ng/ml (data not shown).

Conclusions

In the present study, the antagonist effects of XR-NTX diminished as the plasma levels of NTX and 6-β-naltrexol decreased. Above NTX 2 ng/ml, heroin administration resulted in subjective ratings of drug liking that were similar to what was observed after placebo challenges, suggesting consistent blockade. At NTX plasma levels below 2 ng/ml, subjective responses to a heroin challenge dose resembled an “unblocked” state. That is, below these plasma levels, responses to heroin were greater than the placebo threshold. Below 1 ng/ml NTX, heroin administration produced subjective ratings above what would be expected for placebo, confirming previous assumptions that NTX levels less than 1 ng/ml are not clinically adequate. Importantly, we observed that in the range of plasma levels between 1 and 2 ng/ml, heroin also yielded responses above what would be expected on placebo. This suggests that to be confident of blockade, clinicians should aim to achieve NTX levels above the 2 ng/ml threshold, rather than the previously suggested minimum range of 1 to 2 ng/ml. This in turn can influence the suggested frequency of XR-NTX administration, which in case of Vivitrol at day 28 post injection produces average NTX levels around 1.33 ng/ml but has significant variability ¹⁰. Therefore, a large number of patients may benefit from a shorter re-dose interval of Vivitrol such as every 3 weeks to maximize therapeutic benefits. In turn, variability in NTX levels across participants may have been due to differences in absorption, metabolism and clearance ^11,26, with some individuals having detectable levels at day 42 after the initial injection ²⁷.

At plasma levels below 3 ng/ml for 6-β-naltrexol, subjective responses to a heroin challenge dose also resembled an “unblocked” state. Our findings are consistent with a similar study of the subjective effects of alcohol ingestion after NTX administration in heavy drinkers ²⁸, which described a negative relationship between 6-β-naltrexol levels and positive subjective ratings of alcohol. A 6-β-naltrexol threshold level for reducing the subjective effects of alcohol was not reported by McCaul and colleagues (2000), however. It is important to note that due to the first-pass liver metabolism, 6-β-naltrexol levels are proportionately higher in oral vs. parenteral administration of XR-NTX ¹¹. This is the first study in humans which specifically considers the threshold NTX and/or 6-β-naltrexol levels needed to block a response to the subjective effects of heroin.

The study has a number of possible limitations. NTX plasma levels were drawn at predetermined times that didn’t always coincide with the exact day that 25 mg heroin was administered, however blood levels from the XR-NTX administration were relatively stable within the 1-day window between assessment of plasma NTX levels and subjective ratings of 25 mg heroin. The study also had a small sample size with 12 non-treatment-seeking men, which may limit generalizability to the clinical, treatment-seeking population. As the two doses of XR-NTX resulted in different peak NTX levels, it is conceivable that participants in the two groups experienced different effects on heroin administration days. That is those who received the lower dose may have had more priming of the subjective effects due to an earlier loss of the antagonist effect. Of note, sex differences have been observed, with females having 30% lower maximum plasma concentrations of naltrexone and 6-beta-naltrexol than males after receiving both 190 mg or 380 mg of XR-NTX ¹¹. Another limitation is that only a single dependent measure was analyzed in this study. Ideally, a clinically relevant objective endpoint such as oxygen saturation or respiratory rate also would have been included. However, in the parent study, supplemental oxygen was provided as a safety measure during experimental sessions. Because the respiration data were likely to be affected by this intervention, they were not analyzed. Likewise, pupil diameter is an objective response to heroin administration but it is not clear what level of pupil constriction corresponds with effective treatment response to NTX. Therefore, we focused the present analysis on drug liking as an outcome measure, consistent with FDA guidelines for abuse liability assessment of a substance. And finally, this analysis does not address the ability of NTX to antagonize the effects of fentanyl, which has recently been commonly found in the illicit opioid supply. Although fentanyl was not tested in our study, previous preclinical studies using pA₂ analyses have demonstrated that the antagonist potency of NTX is the same against fentanyl and morphine ²⁹.

Our findings suggest that the plasma level of NTX needed to almost completely block the effects of heroin is 2 ng/ml or greater. Below this threshold, patients may be susceptible to the risks of being “unblocked” with increases in craving, escalation of illicit opioid use ³⁰, and negative clinical outcomes such as return to illicit use and opioid overdose. Levels between 1–2ng/ml may offer blocking effects sufficient for highly motivated patients; however, clinicians should consider that patients who report recurrence of opioid cravings and/or have episodic lapses may have lower NTX levels and therefore may respond well to XR-NTX dosing at a greater frequency.

Abbreviations:

NIDA

National Institute on Drug Abuse

NTX

Naltrexone

NYSPI

New York State Psychiatric Institute

OUD

Opioid Use Disorder

XR-NTX

Extended-release Naltrexone