- Email: [email protected]
Safety of injectable opioid maintenance treatment for heroin dependence
Safety of Injectable Opioid Maintenance Treatment for Heroin Dependence Robert Stoermer, Juergen Drewe, Kenneth M. Dursteler-Mac Farland, Christoph Hock, Franz Mueller-Spahn, Dieter Ladewig, Rudolf Stohler, and Ralph Mager Background: There is a growing debate about injectable opioid treatment programs in many Western countries. This is the first placebo-controlled study of the safety of injectable opioids in a controlled treatment setting. Methods: Twenty-five opioid-dependent patients on intravenous (IV) heroin or IV methadone maintenance treatment were randomly assigned to either their individual prescribed IV maintenance dose or placebo. Acute drug effects were recorded, focusing on electrocardiography, respiratory movements, arterial blood oxygen saturation, and electroencephalography (EEG). Results: After heroin injection, marked respiratory depression progressing to a Cheyne–Stokes pattern occurred. Peripheral arterial blood oxygenation decreased to 78.9 ⫾ 8.7% (mean ⫾ SD) ranging from 52%–90%. During hypoxia, 7 of the 16 subjects experienced intermittent and somewhat severe bradycardia. Five subjects exhibited paroxysmal EEG patterns. After methadone injection, respiratory depression was less pronounced than after heroin injection. No relevant bradycardia was noted. Conclusions: Opioid doses commonly prescribed in IV opioid treatment induce marked respiratory and circulatory depression, as well as occasionally irregular paroxysmal EEG activity. Further studies are needed to optimize the clinical practice of IV opioid treatment to prevent serious complications. Moreover, the extent of the observed effects raises questions about the appropriateness of IV opioid treatment in the present form. Biol Psychiatry 2003;54:854 – 861 © 2003 Society of Biological Psychiatry Key Words: Heroin, methadone, hypoxia, bradycardia, EEG, adverse effects
From the Center of Applied Technologies in Neuroscience-Basel, Psychiatric University Clinic of Basel (RS, FM-S, RM), Department of Clinical Pharmacology and Toxicology, University Hospital (JD), and Division of Substance Use Disorders, Psychiatric University Clinic of Basel (KMD-M, DL), Basel; Division of Psychiatry Research, University of Zurich (CH), Zurich; and Psychiatric University Hospital (RS), Zurich-West, Switzerland. Address reprint requests to Ralph Mager, M.D., Center of Applied Technologies in Neuroscience-Basel, Psychiatric University Clinic of Basel, Wilhelm KleinStrasse 27, 4025 Basel, Switzerland. Received November 5, 2002; revised February 27, 2003; accepted March 4, 2003.
© 2003 Society of Biological Psychiatry
Introduction
I
n 1994, the Swiss program for medical prescription of narcotics (PROVE) started to treat opioid dependent patients with injectable opioid treatment (IOT) using either intravenous (IV) heroin-assisted treatment (HAT) or IV methadone maintenance treatment (MMT). Following initial controversy (Bammer et al 1999; Farrell and Hall 1998; Perneger et al 1998; Rusche 1999; Satel and Aeschbach 1999; Uchtenhagen et al 1997; World Health Organization 1999), IOT now (Drucker 2001; Rehm et al 2001) seems to fulfil the expectations in terms of retaining dependent patients in treatment and lowering criminal activities as well as high-risk sexual behavior and is presently available in the Netherlands (Copeman 2002; Sheldon 2002; van Kolfschooten 2002), Germany, and the United Kingdom. There is growing debate about initiating heroin trials in other countries, however (e.g., Italy, Spain, and the United States; Kuo et al 2000). The Swiss IOT program provides a high level of structure to the addicts (mandatory day-to-day clinical visits) and minimizes impulsive, uncontrolled aspects of addiction. This probably contributes to improvements in medical and social variables as reported on the Swiss and Dutch IOT programs (Rehm et al 2001; van Kolfschooten 2002). Within the evolving IOT setting, only a few reports deal with the safety of injected opioids (White and Irvine 1999; Zador 2001). Changes of cortical blood oxygenation determined by near-infrared spectroscopy were reported as preliminary data without appropriate monitoring of respiratory or cardiovascular function (Hock et al 1999; Stohler et al 1999). Peripheral blood oxygenation data generated in another study suggested a high incidence of systemic hypoxemia after IV heroin (Dursteler-Mac Farland et al 2000). To our knowledge, no systematic studies analyzing acute cardiac, respiratory, and neurophysiologic effects induced by IV heroin or methadone have been published so far. We report the findings of a randomized, placebocontrolled study of the effects of injectable opioids in 0006-3223/03/$30.00 doi:10.1016/S0006-3223(03)00290-7
Safety of Injectable Opioid Maintenance Treatment
dependent patients in stable IOT under controlled laboratory conditions. We focused on standard assessment methods, such as peripheral arterial pulse oxymetry (SaO2), electrocardiography (ECG), and electroencephalography (EEG).
Methods and Materials Participants Twenty-five subjects matching DSM-IV criteria for opioid dependence participated in this study. Exclusion criteria were other major mental or physical diseases, anticonvulsive medication, and extremely poor condition of veins. The maintenance dose and medication had to have been constant for at least 7 days before the study without any signs of severe, unexpected clinical symptoms of withdrawal or overdose toxicity. The dose could not be changed during the study. Subjects were recruited from the local IV HAT project (n ⫽ 16, 2 women, 14 men; mean age 34.4 years; range 20 – 42 years) and from the IV MMT program at the Psychiatric University Clinic of Basel (n ⫽ 9, 1 woman, 8 men; mean age 32.7 years, range 22– 45 years). All participants received their prescribed dose of heroin (two IV injections per day) or methadone (one IV injection per day) in an outpatient setting. All subjects underwent preliminary screening including a physical examination and completion of a medical questionnaire. Participants gave written informed consent after the procedures had been fully explained. The study complied with the principles of the declaration of Helsinki and was approved by the Ethics Committee of the University of Basel, and IOT is an approved treatment in Switzerland. Twenty-three subjects completed all test sessions. One subject in the heroin group and one subject in the methadone group did not attend their placebo session.
Monitoring For vital sign assessments, we used a standard multiparameter bedside monitoring device (ESCORT II⫹, Medical Data Electronics, Arleta CA). ECG: A single-lead ECG was recorded with chest electrodes placed at standard positions Ra, La, and Ll. Impedance pneumography: The respiratory movements were detected by measuring impedance between the ECG electrodes Ra and La (Stein and Luparello 1967). Pulse oxymetry: Peripheral SaO2 was measured by continuous pulse oxymetry. The sensor was attached to the right middle finger. EEG: Frontal, central, and parietal leads (F3/C3/P3; F4/C4/P4) were placed according to the international 10 –20 system. Six amplifiers (Grass Model KP57C8A; Grass Astro-med, Inc, West Warwick, Rhode Island) were used, and electrodes were referenced to Cz. All devices were linked to a personal computer– based data acquisition system (custom made using TurboLab Online II package by Bu¨ hrer & Partner GbR, Stockdorf, Germany) with a digitalization rate of 256 Hz.
Experimental Design and Procedures Subjects underwent two test sessions on different days within 2 weeks. They received their prescribed opioid (heroin or metha-
BIOL PSYCHIATRY 2003;54:854 – 861
855
done) or placebo (saline) in a randomized manner according to a randomization table generated before the study. The opioid dosage consisted of either the full amount of the once daily prescribed methadone or of half of the twice daily prescribed heroin according to their regular maintenance treatment. Prescribed single opioid doses ranged from 60 –300 mg of heroin or 80 –300 mg of methadone in accordance with the commonly used dose ranges (Perneger et al 2000). Heroin (diacetylmorphine as HCl salt) was provided by the Swiss Federal Office of Public Health. Methadone was supplied as a racemate by the local Institute of Pharmacy. During the test sessions, subjects remained in a supine position. After maintaining stable baseline values for 5 min, subjects received either their regular dose of opioid or placebo by bolus injection through a cannula within approximately 30 sec. The monitoring was continued for about 25 min thereafter.
Data Reduction and Analysis Data were analyzed in an intention-to-treat analysis. Subjects were included if they received at least one administration. The data obtained were visually inspected. Placebo injections did not alter the stable baseline values. Methadone or heroin application resulted in respiratory depression that differed in onset (latency), duration, and extent. Further data analysis included two approaches. In the first, treatment-induced hypoxemia was rated in accordance with a standard clinical assessment scale (Moller et al 1991) based on three ranges. Thus, SaO2 values of 86%–90% were termed mild hypoxemia, SaO2 of 81%– 85% moderate hypoxemia, and SaO2 ⬍81% severe hypoxemia. The individual duration of the three hypoxemic ranges was calculated for each subject, and group averages were then generated. In the second approach, two evaluation time windows were defined including the main effects of each compound as derived from the first approach. Consequently, the evaluation time windows were defined to be between 3 and 16 min after methadone injections and between 6 and 21 min after heroin injections (see Figure 1) in accordance with the different pharmacologic profiles. The significance of drug effects was tested by comparing baseline data with data in the evaluation time window after drug application. Therefore, temporal mean values and extreme values were calculated separately for both time periods in each subject. Group averages were then generated. Although two subjects participated in only a single session, their data were included in the intention-to-treat analysis. A visual estimate of the paroxysmal EEG patterns was performed by an experienced neurologist. For calculating compressed spectral arrays (Figure 2), artifact-free 2-sec EEG epochs were submitted to spectral analysis using the fast Fourier transform that computed the absolute power (in V2/Hz) of the EEG in .5-Hz steps. The power spectra were averaged for time periods of 15 sec. Relative spectral distributions were expressed as a portion of total power across the frequency range from 1–32 Hz, resulting in relative power values. Subsequently, spectral arrays were generated, giving the relative power for different frequencies in the course of time.
856
R. Stoermer et al
BIOL PSYCHIATRY 2003;54:854 – 861
Statistical Analyses Cross-tabulated data were analyzed using chi-squared or Fisher’s Exact Test, as appropriate. Continuously distributed data were compared by analysis of variance. If baseline values of parameters differed significantly, analysis was performed using baseline values as covariates. Level of statistical significance was p ⫽ .05. All comparisons were performed using SPSS for Windows software (version 10.0).
Results
Figure 1. Typical time course of respiration movements, SaO2, calculated respiratory rate, and heart rate in a 29-year-old male subject. The time point of bolus injection of heroin (150 mg) is indicated on the time axis. The horizontal bar indicates the time window of maximal action used in the calculation of mean values. Corresponding electroencephalogram data are shown in Figure 5.
All data are given as means ⫾ SD (medians; 25%–75% percentiles). There were no significant differences between the treatment groups with respect to gender and age. Whereas placebo administration did not result in significant changes in baseline values for any of the parameters studied (data not shown), both heroin and methadone significantly affected respiratory function and arterial blood oxygenation. Cardiovascular effects were seen after heroin injection, but not following methadone. Figure 1 shows the characteristic time courses of respiratory and
Figure 2. Spectral electroencephalogram changes caused by intravenous application of 150 mg of heroin (same subject as in Figures 1 and 5). The compressed spectral array gives the relative power values (in percent) for .5-Hz frequency intervals plotted against time. Heroin was applied at minute 5 (electrode P3). Note the marked spectral changes starting around 5 min after injection (minute 10 of the recording).
Safety of Injectable Opioid Maintenance Treatment
BIOL PSYCHIATRY 2003;54:854 – 861
857
cardiovascular parameters measured after heroin administration in a 29-year-old male subject.
Respiratory System In the heroin group, baseline respiratory rate was 14.4 ⫾ 3.1 breaths/min (13.6; 12.4 –16.4 breaths/min), and peripheral arterial blood oxygenation was 94.3 ⫾ 2.3% (94.6; 93.9 –95.8%). In the methadone group, baseline respiratory rate was 13.6 ⫾ 3.1 breaths/min (12.7; 11.3– 15.9 breaths/min), and arterial oxygen saturation was 93.2 ⫾ 2.8% (93.3; 90.9 - 95.4%). These differences were not significant. Heroin administration was associated with a Cheyne– Stokes breathing pattern with apnea episodes and periods of dead space ventilation in each case. In most cases, this pattern lasted about 15 min and subsequently changed to bradypnea. The respiratory depression resulted in a drop of arterial oxygen saturation below a saturation threshold of 90% with a latency of 5.1 ⫾ 3.1 min (4.8; 3.6 –5.7 min). Without oxygen supply, constant oxygenation values above 90% were reached again after 15.4 ⫾ 5.1 min (14.9; 9.7–18.4 min). The mean respiratory rate during this period was 8.7 ⫾ 1.8 breaths/min (8.6; 7.9 –9.5 breaths/ min). This was accompanied by a reduction of mean SaO2 to 88.9 ⫾ 3.0% (88.7; 90.8 – 89.7%), which was significantly lower than baseline values (p ⬍ .001). Although the disturbances of respiratory rhythm persisted until the end of the test session, none was judged to be clinically relevant at this time point. In the methadone group, significantly fewer subjects (5 of 9 [56%]; p ⫽ .01) than in the heroin group showed a Cheyne–Stokes breathing pattern, leading to a substantial decrease in SaO2. In these cases, oxygenation values below 90% occurred 2.3 ⫾ 1.2 min (2.2; 1.4 –3.1 min) after methadone bolus injections, which was significantly earlier than with heroin (p ⫽ .028). Values below 90% were found for 13.7 ⫾ 7.9 min (11.1; 7.2–19.2 min). Mean respiration rate during this period was 10.9 ⫾ 2.0 breaths/ min (10.7; 9.0 –11.9 breaths/min), and mean values of SaO2 were 90.3 ⫾ 3.1% (88.5; 92.4 –90.2%) and thus significantly lower than baseline values (p ⫽ .041). In contrast to heroin-treated subjects, all methadone-treated subjects showed a normal respiration pattern at the end of the test session. Heroin injections resulted in minimal SaO2 values of 78.9 ⫾ 8.7% (76; 85–79%), whereas minimal SaO2 values after methadone were 84.4 ⫾ 6.6% (80; 88%– 86%). These minimal SaO2 values were significantly lower after heroin than after methadone injections (p ⫽ .008). Duration and severity of hypoxemia as surveyed with the clinical rating scale, after the two opioids, are shown in Figure 3. Mild hypoxemia was found in all (100%)
Figure 3. Incidence, severity, and duration of hypoxemia after opioid injection. Black diamonds represent hypoxemia after heroin injection in 16 subjects, gray circles represent hypoxemia after methadone injection in 9 subjects. Subjects contribute up to three data points (one for each range of hypoxemia severity).
subjects receiving heroin but in only 7 of 9 (78%) subjects receiving methadone (p ⫽ .12) and lasted for 5.5 ⫾ 2.5 min (5.5; 4.3–7.3 min) and 6.3 ⫾ 6.5 min (3.3; 1.0 –10.2 min), respectively. Overall, 13 of 16 (81%) heroin-treated subjects and 4 of 9 (44%) methadone-treated subjects developed moderate hypoxemia (p ⫽ .087) lasting 2.2 ⫾ 2.0 min (1.7; .9 –2.5 min) and 2.7 ⫾ 2.3 min (.8; 1.8 –3.8 min), respectively. Severe hypoxemia developed in 8 of 16 (50%) heroin-treated and in 3 of 9 (33%) methadonetreated subjects (p ⫽ .68). Severe hypoxemia lasted for a period of 1.5 ⫾ 2.5 min (.4; .3–1.1 min) and .7 ⫾ .9 min (.2; .5–1.7 min), respectively.
Cardiovascular System Baseline values of heart rate were comparable (p ⫽ .19) in the heroin and methadone groups (74.4 ⫾ 11.8 beats/min [72.0; 66.7– 84.4 beats/min] and 82.2 ⫾ 16.9 beats/min [85.9; 66.4 –95.7 beats/min], respectively). In general, we found increased heart rates during and shortly after opioid injections. Simultaneously with the onset of respiratory depression, the heart rate decreased in all heroin-treated subjects (Figure 4). We found a characteristic heart rate pattern with episodes of bradycardia during apnea and heart rate increases within the ensuing inspiration periods. We observed heart rate variations of up to 22 beats/min
858
R. Stoermer et al
BIOL PSYCHIATRY 2003;54:854 – 861
Figure 5. Electroencephalogram (EEG) segment of 10 sec duration shows the onset of the paroxysmal activity 5.5 min after bolus injection of heroin (150 mg). Corresponding respiratory and circulatory data are given in Figure 1.
Figure 4. Incidence, heart rate, and duration of bradycardia episodes after heroin injection in 16 subjects. Symbols represent the incidence of bradycardia in the specified range.
during a respiration cycle. In two subjects, the heart rate fell below 30 beats/min. One of these subjects showed recurrent sinus arrests and idioventricular rhythms at a heart rate below 30 beats/min lasting for 1.2 min. In 7 of 16 (44%) heroin-treated subjects, heart rates between 40 and 50 beats/min with a mean duration of 3.5 ⫾ 4.3 min (.4; .1–9.0 min) were observed. After methadone injection, the heart rate changes were negligible, and bradycardia was not observed.
EEG Because of the difficult measurement circumstances (withdrawal symptoms, sweating, excitation, lack of compliance), usable EEG recordings were obtained from only 9 of 16 heroin-treated subjects and from 8 of 9 methadonetreated subjects. In five EEG recordings among the heroin group, irregular paroxysmal activity was noted, with a time delay of 5.6 ⫾ .3 min after the injection. The pattern of paroxysmal activity varied between the subjects, although nonspecific intermittent irregular high-voltage theta– delta slowing occurred on the background of normal EEG activity in all five subjects. In one heroin-treated subject, spike and wave activity occurred, whereas in two other heroin-treated subjects, sharp waves were noted. Figure 5 provides an example of an EEG in which a more regular 3/sec paroxysmal delta activity developed 5.5 min after heroin injection (same subject as in Figure 1).
To illustrate the complex spectral EEG changes after drug application, a corresponding compressed spectral array is shown in Figure 2, which displays the time course of relative power values in .5-Hz intervals. About 5 min after injection, an initial increase in alpha power (8 –12Hz) could be observed, followed by a marked decrease in alpha power and an increase in delta power (1–3.5Hz). The delta power increase was largely caused by the paroxysmal 3/sec delta activity as shown in Figure 5, but an increase in the slow delta range occurred as well. Spectral changes were closely linked to the development of the hypoxemic state as shown in Figure 1 and restored during increase of the oxygen saturation approximately 7–10 min after injection. All irregular EEG recordings were obtained from subjects showing moderate to severe hypoxemia. Three of them showed intermittent bradycardia. In four EEG recordings, visual inspection did not reveal any signs of enhanced excitability. In contrast, only one subject in the methadone group exhibited a similar hypersynchronous pattern after injection. No clinical signs of seizures were observed in any subject.
Adverse Effects and Interventions In four sessions involving heroin, breathing instructions and oxygen supply were needed because of severe hypoxia and bradycardia. Here the oxygen supply was started between minute 5 and 8 following drug administration. The decision for intervening with oxygen and breathing instructions depended on the clinical status of the subject. Because none of the subjects lost consciousness and all could respond to verbal instructions, these interventions
Safety of Injectable Opioid Maintenance Treatment
resulted in an immediate improvement of cardiorespiratory functions.
Discussion Currently, oral MMT is the most widely used pharmacologic treatment of opioid dependence, but it is not accepted by all opioid-dependent patients (Bammer et al 1999). Consequently, IV MMT and IV HAT have been introduced in Switzerland, and in the Netherlands the government has proposed the introduction of IV or inhalable heroin as a coprescription to MMT. In the United Kingdom, IOT has been an established treatment for many decades; however, the level of medical supervision of the latter is low compared with the Swiss IOT program. Our study assessed the acute respiratory, circulatory, and electrocortical effects within the first 25 min after IOT injection in opioid-dependent subjects. There are some general limitations regarding research with illicit drugs in this group of severely drug-dependent patients that are related to the reduced ability to cooperate. To get the most reliable data with respect to drug safety, our study design was adapted as closely as possible to the outpatient setting. Accordingly, the actual doses of opioids varied greatly within the treatment groups. This reflects experience showing that, for pharmacologic effect, individualized opioid dose (adapted to the individual level of receptor downregulation) is more important than absolute dose. Therefore, each subject was given the dose he or she received for at least 1 week before the start of the study without significant subjective adverse effect. Opioids are known to depress respiration by affecting both central respiratory centers and peripheral chemoreceptors (Reisine and Pasternak 1996; Yeadon and Kitchen 1989). Furthermore, opioids may decrease hemodynamic performance by direct vascular effects, reducing sympathetic and enhancing vagal tone (Bailey 1993; Ghuran and Nolan 2000). In our study, opioid administration resulted in a stereotyped response pattern of respiratory parameters. Cardiovascular effects were seen only with heroin. In all cases, heroin caused respiration rates matching, at least temporarily, the clinical criteria of heroin intoxication (Sporer 1999) and resulted in a marked decrease in SaO2 for relevant time periods. In parallel with the onset of respiratory depression, the heart rate decreased in herointreated subjects, with some subjects presenting severe bradycardia. Irregular paroxysmal EEG activities coinciding with the time course of bradycardia and hypoxemia were found. The heroin-induced EEG patterns were clearly distinguishable from the well-known major slowing of EEG frequency corresponding to opioid-induced anesthesia. The EEG changes associated with high-dose
BIOL PSYCHIATRY 2003;54:854 – 861
859
opioid anesthesia in humans are well characterized and typically do not reveal paroxysmal activity (Smith et al 1989). The hypoxemic state in our patients might facilitate an increase of excitability. Similar effects on electrocortical activity after IV heroin application have been reported (Roubicek et al 1969; Volavka et al 1971), but a correlation with cardiopulmonary parameters has not been described previously. To assess the clinical implications of our findings, we note that our study conditions differed in some aspects from the common IOT setting. First, laboratory conditions provide different external stimuli than the regular treatment situation. Moreover, whereas the opioid is selfadministrated (active administration) in the IOT settings, the opioids (or placebo) in this study were injected in a standardized way as a bolus injection over 30 sec (passive administration) to minimize between-subject variability. This procedure, which lacked typical drug-paired cues, may have caused a reduced (conditioned) tolerance, resulting in an enhanced opioid effect. There is growing evidence that these difference in the mode of administration may significantly influence the effects of opioids (Siegel 1999; Siegel et al 2000). On the other hand, drug self-administration results in much higher rates of drug injection, leading most probably to even more pronounced respiratory depression and central vagal stimulation. Furthermore, the severity of effects seen in our study is underestimated because four subjects required intervention including oxygen supply after heroin application. Our findings have shown that common treatment doses of IV heroin and, to a lesser extent, IV methadone induce clinically relevant hypoxemia. Heroin-induced hypoxemia, often combined with bradycardia, may result in a reduced cerebral blood supply, thus putting opioid users at risk for the development of cerebral hypoxia or ischemia. The observed physiologic responses are also likely to occur in the nontreatment setting. Reports of cardiovascular failure (Ghuran and Nolan 2000), nonobstructive stroke (Brust and Richter 1976; Buttner et al 2000; Jensen et al 1990) with permanent or transient neurologic deficits (Neiman et al 2000), and possibly heroin-associated seizures (Alldredge et al 1989; Ng et al 1990; Richter et al 1973) in opioid-dependent subjects under nontreatment conditions possibly reflect general risks of opioid use. Our findings raise concerns about potential side effects of IV opioids and the possibility of accidental death even in a controlled treatment setting. Data about such deaths or other serious events related to IOT in Switzerland are not available. Apart from the risks of overdose, contamination, multiple drug use, infection, or accidents, long-term complications such as brain lesions induced by frequent hypoxic states (Andersen and Skullerud 1999; Oehmichen et al 1996) and consequent cognitive impairment (Darke et
860
BIOL PSYCHIATRY 2003;54:854 – 861
al 2000) appear to be possible not only with illicit opioid use but with therapeutic use as well. Loss of opioid tolerance, multiple drug use, and systemic disease have been identified as risk factors of illicit opioid overdose (Warner Smith et al 2001). We hypothesize that these factors may also be relevant for safety in the IOT setting. A better understanding of tolerance-modifying mechanisms, such as environmental factors, comedication, and multiple drug use is clearly needed. Our results suggest that the clinical practice of IV MMT and IV HAT could be improved with respect to safety. We propose basic monitoring following opioid injection could be helpful to identify acute hazardous situations. Intervention strategies should be developed to prevent acute hazards or permanent damage. The criteria for the prescription and therapeutic use of IV MMT or IV HAT should be critically revised. The objective of this study was to investigate safety of IOT. A judgment of overall benefit and risks of this treatment must balance these risks with those of untreated opioid addiction. Therefore, additional studies of the safety of IOT are required.
This study was supported by Grant No. 316.98.8104 from the Swiss Federal Office of Public Health. We thank all participants, without whom this study would not have been possible. We gratefully acknowledge the valuable support of M. Kuntze and A. Bullinger. In addition, our thanks go to M. Schmidlin and M. v. Arx for their excellent technical laboratory support and J. Coutino and S. Begre´ for medical assistance during the test sessions.
References Alldredge BK, Lowenstein DH, Simon RP (1989): Seizures associated with recreational drug abuse. Neurology 39:1037– 1039. Andersen SN, Skullerud K (1999): Hypoxic/ischaemic brain damage, especially pallidal lesions, in heroin addicts. Forensic Sci Int 102:51–59. Bailey PL (1993): Opioids in operative anesthesia. In: Herz A, editor. Opioids, Volume II. Berlin: Springer Verlag. Bammer G, Dobler-Mikola A, Fleming PM, Strang J, Uchtenhagen A (1999): The heroin prescribing debate: Integrating science and politics. Science 284:1277–1278. Brust JC, Richter RW (1976): Stroke associated with addiction to heroin. J Neurol Neurosurg Psychiatry 39:194 –199. Buttner A, Mall G, Penning R, Weis S (2000): The neuropathology of heroin abuse. Forensic Sci Int 113:435–442. Copeman M (2002): Heroin prescription for opioid addicts. Lancet 359:889 –890. Darke S, Sims J, McDonald S, Wickes W (2000): Cognitive impairment among methadone maintenance patients. Addiction 95:687–695. Drucker E (2001): Injectable heroin substitution treatment for opioid dependency. Lancet 358:1385.
R. Stoermer et al
Dursteler-Mac Farland KM, Stormer R, Seifritz E, Hug I, Muller-Spahn F, Ladewig D, Stohler R (2000): Opioidassociated effects on oxygen saturation. Addiction 95:285– 287. Farrell M, Hall W (1998): The Swiss heroin trials: Testing alternative approaches. Bmj 316:639. Ghuran A, Nolan J (2000): Recreational drug misuse: Issues for the cardiologist. Heart 83:627–633. Hock C, Stormer R, Dursteler K, Muller-Spahn F, Ladewig D, Stohler R (1999): Heroin-induced deoxygenation of cerebral hemoglobin: Monitoring by means of near-infrared spectroscopy. Ann Neurol 45:134 –135. Jensen R, Olsen TS, Winther BB (1990): Severe non-occlusive ischemic stroke in young heroin addicts. Acta Neurol Scand 81:354 –357. Kuo II, Fischer B, Vlahov D (2000): Consideration of a North American heroin-assisted clinical trial for the treatment of opiate-dependent individuals. Int J Drug Policy 11:357–370. Moller JT, Johannessen NW, Berg H, Espersen K, Larsen LE (1991): Hypoxaemia during anaesthesia—an observer study. Br J Anaesth 66:437–444. Neiman J, Haapaniemi HM, Hillbom M (2000): Neurological complications of drug abuse: Pathophysiological mechanisms. Eur J Neurol 7:595–606. Ng SK, Brust JC, Hauser WA, Susser M (1990): Illicit drug use and the risk of new-onset seizures. Am J Epidemiol 132:47– 57. Oehmichen M, Meissner C, Reiter A, Birkholz M (1996): Neuropathology in non-human immunodeficiency virus-infected drug addicts: Hypoxic brain damage after chronic intravenous drug abuse. Acta Neuropathol (Berl) 91:642–646. Perneger TV, Giner F, del Rio M, Mino A (1998): Randomised trial of heroin maintenance programme for addicts who fail in conventional drug treatments. Bmj 317:13–18. Perneger TV, Mino A, Giner F, Broers B (2000): Patterns of opiate use in a heroin maintenance programme. Psychopharmacology (Berl) 152:7–13. Rehm J, Gschwend P, Steffen T, Gutzwiller F, Dobler-Mikola A, Uchtenhagen A (2001): Feasibility, safety, and efficacy of injectable heroin prescription for refractory opioid addicts: A follow-up study. Lancet 358:1417–2420. Reisine T, Pasternak G (1996): Opioid analgesics and antagonists. In: Goodman LS, Perry M, Gilman AG, Hardman JG, editors. Goodman and Gilman’s The Pharmacological Basis of Therapeutics. New York: McGraw-Hill. Richter RW, Pearson J, Bruun B, Challenor YB, Brust JC, Baden MM (1973): Neurological complications of addiction to heroin. Bull N Y Acad Med 49:3–21. Roubicek J, Zaks A, Freedman AM (1969): EEG changes produced by heroin and methadone. Electroencephalogr Clin Neurophysiol 27:667–668. Rusche S (1999): Prescribing heroin. Science 285:53; discussion: 533. Satel SL, Aeschbach E (1999): The Swiss heroin trials. Scientifically sound? J Subst Abuse Treat 17:331–335. Sheldon T (2002): Netherlands considers prescribing heroin to addicts. BMJ 324:385.
Safety of Injectable Opioid Maintenance Treatment
Siegel S (1999): Drug anticipation and drug addiction. The 1998 H. David Archibald Lecture. Addiction 94:1113–1124. Siegel S, Baptista MA, Kim JA, McDonald RV, Weise-Kelly L (2000): Pavlovian psychopharmacology: The associative basis of tolerance. Exp Clin Psychopharmacol 8:276 –293. Smith NT, Benthuysen JL, Bickford RG, Sanford TJ, Blasco T, Duke PC, et al (1989): Seizures during opioid anesthetic induction–are they opioid-induced rigidity? Anesthesiology 71:852–862. Sporer KA (1999): Acute heroin overdose. Ann Intern Med 130:584 –590. Stein M, Luparello TJ (1967): Measurement of respiration. In: Brown CC, editor. Methods in Psychophysiology. Baltimore: Williams & Wilkins, 75–94. Stohler R, Dursteler KM, Stormer R, Seifritz E, Hug I, SattlerMayr J, et al (1999): Rapid cortical hemoglobin deoxygenation after heroin and methadone injection in humans: A preliminary report. Drug Alcohol Depend 57:23–28. Uchtenhagen A, Gutswiller F, Dobler-Mikola A (1997): Programme for a medical prescription of narcotics: Final report of the research representatives. Summary of the synthesis
BIOL PSYCHIATRY 2003;54:854 – 861
861
report. Zurich: University of Zurich. Available at http:// www.cfdp.ca/switz.html. van Kolfschooten F (2002): Dutch investigators recommend prescription of heroin to addicts. Lancet 359:590. Volavka J, Zaks A, Roubicek J, Fink M (1971): Acute EEG effects of heroin and naloxone. Electroencephalogr Clin Neurophysiol 30:165. Warner Smith M, Darke S, Lynskey M, Hall W (2001): Heroin overdose: Causes and consequences. Addiction 96:1113– 1125. White JM, Irvine RJ (1999): Mechanisms of fatal opioid overdose. Addiction 94:961–972. World Health Organization (1999): Report of the External Panel on the Evaluation of the Swiss Scientific Studies of Medically Prescribed Narcotics to Drug Addicts. Geneva: World Health Organization. Available at: http://www.emory.edu/ NFIA/legal/heroin/whoheroin.html. Yeadon M, Kitchen I (1989): Opioids and respiration. Prog Neurobiol 33:1–16. Zador D (2001): Injectable opiate maintenance in the UK: Is it good clinical practice? Addiction 96:547–553.
From the Center of Applied Technologies in Neuroscience-Basel, Psychiatric University Clinic of Basel (RS, FM-S, RM), Department of Clinical Pharmacology and Toxicology, University Hospital (JD), and Division of Substance Use Disorders, Psychiatric University Clinic of Basel (KMD-M, DL), Basel; Division of Psychiatry Research, University of Zurich (CH), Zurich; and Psychiatric University Hospital (RS), Zurich-West, Switzerland. Address reprint requests to Ralph Mager, M.D., Center of Applied Technologies in Neuroscience-Basel, Psychiatric University Clinic of Basel, Wilhelm KleinStrasse 27, 4025 Basel, Switzerland. Received November 5, 2002; revised February 27, 2003; accepted March 4, 2003.
© 2003 Society of Biological Psychiatry
Introduction
I
n 1994, the Swiss program for medical prescription of narcotics (PROVE) started to treat opioid dependent patients with injectable opioid treatment (IOT) using either intravenous (IV) heroin-assisted treatment (HAT) or IV methadone maintenance treatment (MMT). Following initial controversy (Bammer et al 1999; Farrell and Hall 1998; Perneger et al 1998; Rusche 1999; Satel and Aeschbach 1999; Uchtenhagen et al 1997; World Health Organization 1999), IOT now (Drucker 2001; Rehm et al 2001) seems to fulfil the expectations in terms of retaining dependent patients in treatment and lowering criminal activities as well as high-risk sexual behavior and is presently available in the Netherlands (Copeman 2002; Sheldon 2002; van Kolfschooten 2002), Germany, and the United Kingdom. There is growing debate about initiating heroin trials in other countries, however (e.g., Italy, Spain, and the United States; Kuo et al 2000). The Swiss IOT program provides a high level of structure to the addicts (mandatory day-to-day clinical visits) and minimizes impulsive, uncontrolled aspects of addiction. This probably contributes to improvements in medical and social variables as reported on the Swiss and Dutch IOT programs (Rehm et al 2001; van Kolfschooten 2002). Within the evolving IOT setting, only a few reports deal with the safety of injected opioids (White and Irvine 1999; Zador 2001). Changes of cortical blood oxygenation determined by near-infrared spectroscopy were reported as preliminary data without appropriate monitoring of respiratory or cardiovascular function (Hock et al 1999; Stohler et al 1999). Peripheral blood oxygenation data generated in another study suggested a high incidence of systemic hypoxemia after IV heroin (Dursteler-Mac Farland et al 2000). To our knowledge, no systematic studies analyzing acute cardiac, respiratory, and neurophysiologic effects induced by IV heroin or methadone have been published so far. We report the findings of a randomized, placebocontrolled study of the effects of injectable opioids in 0006-3223/03/$30.00 doi:10.1016/S0006-3223(03)00290-7
Safety of Injectable Opioid Maintenance Treatment
dependent patients in stable IOT under controlled laboratory conditions. We focused on standard assessment methods, such as peripheral arterial pulse oxymetry (SaO2), electrocardiography (ECG), and electroencephalography (EEG).
Methods and Materials Participants Twenty-five subjects matching DSM-IV criteria for opioid dependence participated in this study. Exclusion criteria were other major mental or physical diseases, anticonvulsive medication, and extremely poor condition of veins. The maintenance dose and medication had to have been constant for at least 7 days before the study without any signs of severe, unexpected clinical symptoms of withdrawal or overdose toxicity. The dose could not be changed during the study. Subjects were recruited from the local IV HAT project (n ⫽ 16, 2 women, 14 men; mean age 34.4 years; range 20 – 42 years) and from the IV MMT program at the Psychiatric University Clinic of Basel (n ⫽ 9, 1 woman, 8 men; mean age 32.7 years, range 22– 45 years). All participants received their prescribed dose of heroin (two IV injections per day) or methadone (one IV injection per day) in an outpatient setting. All subjects underwent preliminary screening including a physical examination and completion of a medical questionnaire. Participants gave written informed consent after the procedures had been fully explained. The study complied with the principles of the declaration of Helsinki and was approved by the Ethics Committee of the University of Basel, and IOT is an approved treatment in Switzerland. Twenty-three subjects completed all test sessions. One subject in the heroin group and one subject in the methadone group did not attend their placebo session.
Monitoring For vital sign assessments, we used a standard multiparameter bedside monitoring device (ESCORT II⫹, Medical Data Electronics, Arleta CA). ECG: A single-lead ECG was recorded with chest electrodes placed at standard positions Ra, La, and Ll. Impedance pneumography: The respiratory movements were detected by measuring impedance between the ECG electrodes Ra and La (Stein and Luparello 1967). Pulse oxymetry: Peripheral SaO2 was measured by continuous pulse oxymetry. The sensor was attached to the right middle finger. EEG: Frontal, central, and parietal leads (F3/C3/P3; F4/C4/P4) were placed according to the international 10 –20 system. Six amplifiers (Grass Model KP57C8A; Grass Astro-med, Inc, West Warwick, Rhode Island) were used, and electrodes were referenced to Cz. All devices were linked to a personal computer– based data acquisition system (custom made using TurboLab Online II package by Bu¨ hrer & Partner GbR, Stockdorf, Germany) with a digitalization rate of 256 Hz.
Experimental Design and Procedures Subjects underwent two test sessions on different days within 2 weeks. They received their prescribed opioid (heroin or metha-
BIOL PSYCHIATRY 2003;54:854 – 861
855
done) or placebo (saline) in a randomized manner according to a randomization table generated before the study. The opioid dosage consisted of either the full amount of the once daily prescribed methadone or of half of the twice daily prescribed heroin according to their regular maintenance treatment. Prescribed single opioid doses ranged from 60 –300 mg of heroin or 80 –300 mg of methadone in accordance with the commonly used dose ranges (Perneger et al 2000). Heroin (diacetylmorphine as HCl salt) was provided by the Swiss Federal Office of Public Health. Methadone was supplied as a racemate by the local Institute of Pharmacy. During the test sessions, subjects remained in a supine position. After maintaining stable baseline values for 5 min, subjects received either their regular dose of opioid or placebo by bolus injection through a cannula within approximately 30 sec. The monitoring was continued for about 25 min thereafter.
Data Reduction and Analysis Data were analyzed in an intention-to-treat analysis. Subjects were included if they received at least one administration. The data obtained were visually inspected. Placebo injections did not alter the stable baseline values. Methadone or heroin application resulted in respiratory depression that differed in onset (latency), duration, and extent. Further data analysis included two approaches. In the first, treatment-induced hypoxemia was rated in accordance with a standard clinical assessment scale (Moller et al 1991) based on three ranges. Thus, SaO2 values of 86%–90% were termed mild hypoxemia, SaO2 of 81%– 85% moderate hypoxemia, and SaO2 ⬍81% severe hypoxemia. The individual duration of the three hypoxemic ranges was calculated for each subject, and group averages were then generated. In the second approach, two evaluation time windows were defined including the main effects of each compound as derived from the first approach. Consequently, the evaluation time windows were defined to be between 3 and 16 min after methadone injections and between 6 and 21 min after heroin injections (see Figure 1) in accordance with the different pharmacologic profiles. The significance of drug effects was tested by comparing baseline data with data in the evaluation time window after drug application. Therefore, temporal mean values and extreme values were calculated separately for both time periods in each subject. Group averages were then generated. Although two subjects participated in only a single session, their data were included in the intention-to-treat analysis. A visual estimate of the paroxysmal EEG patterns was performed by an experienced neurologist. For calculating compressed spectral arrays (Figure 2), artifact-free 2-sec EEG epochs were submitted to spectral analysis using the fast Fourier transform that computed the absolute power (in V2/Hz) of the EEG in .5-Hz steps. The power spectra were averaged for time periods of 15 sec. Relative spectral distributions were expressed as a portion of total power across the frequency range from 1–32 Hz, resulting in relative power values. Subsequently, spectral arrays were generated, giving the relative power for different frequencies in the course of time.
856
R. Stoermer et al
BIOL PSYCHIATRY 2003;54:854 – 861
Statistical Analyses Cross-tabulated data were analyzed using chi-squared or Fisher’s Exact Test, as appropriate. Continuously distributed data were compared by analysis of variance. If baseline values of parameters differed significantly, analysis was performed using baseline values as covariates. Level of statistical significance was p ⫽ .05. All comparisons were performed using SPSS for Windows software (version 10.0).
Results
Figure 1. Typical time course of respiration movements, SaO2, calculated respiratory rate, and heart rate in a 29-year-old male subject. The time point of bolus injection of heroin (150 mg) is indicated on the time axis. The horizontal bar indicates the time window of maximal action used in the calculation of mean values. Corresponding electroencephalogram data are shown in Figure 5.
All data are given as means ⫾ SD (medians; 25%–75% percentiles). There were no significant differences between the treatment groups with respect to gender and age. Whereas placebo administration did not result in significant changes in baseline values for any of the parameters studied (data not shown), both heroin and methadone significantly affected respiratory function and arterial blood oxygenation. Cardiovascular effects were seen after heroin injection, but not following methadone. Figure 1 shows the characteristic time courses of respiratory and
Figure 2. Spectral electroencephalogram changes caused by intravenous application of 150 mg of heroin (same subject as in Figures 1 and 5). The compressed spectral array gives the relative power values (in percent) for .5-Hz frequency intervals plotted against time. Heroin was applied at minute 5 (electrode P3). Note the marked spectral changes starting around 5 min after injection (minute 10 of the recording).
Safety of Injectable Opioid Maintenance Treatment
BIOL PSYCHIATRY 2003;54:854 – 861
857
cardiovascular parameters measured after heroin administration in a 29-year-old male subject.
Respiratory System In the heroin group, baseline respiratory rate was 14.4 ⫾ 3.1 breaths/min (13.6; 12.4 –16.4 breaths/min), and peripheral arterial blood oxygenation was 94.3 ⫾ 2.3% (94.6; 93.9 –95.8%). In the methadone group, baseline respiratory rate was 13.6 ⫾ 3.1 breaths/min (12.7; 11.3– 15.9 breaths/min), and arterial oxygen saturation was 93.2 ⫾ 2.8% (93.3; 90.9 - 95.4%). These differences were not significant. Heroin administration was associated with a Cheyne– Stokes breathing pattern with apnea episodes and periods of dead space ventilation in each case. In most cases, this pattern lasted about 15 min and subsequently changed to bradypnea. The respiratory depression resulted in a drop of arterial oxygen saturation below a saturation threshold of 90% with a latency of 5.1 ⫾ 3.1 min (4.8; 3.6 –5.7 min). Without oxygen supply, constant oxygenation values above 90% were reached again after 15.4 ⫾ 5.1 min (14.9; 9.7–18.4 min). The mean respiratory rate during this period was 8.7 ⫾ 1.8 breaths/min (8.6; 7.9 –9.5 breaths/ min). This was accompanied by a reduction of mean SaO2 to 88.9 ⫾ 3.0% (88.7; 90.8 – 89.7%), which was significantly lower than baseline values (p ⬍ .001). Although the disturbances of respiratory rhythm persisted until the end of the test session, none was judged to be clinically relevant at this time point. In the methadone group, significantly fewer subjects (5 of 9 [56%]; p ⫽ .01) than in the heroin group showed a Cheyne–Stokes breathing pattern, leading to a substantial decrease in SaO2. In these cases, oxygenation values below 90% occurred 2.3 ⫾ 1.2 min (2.2; 1.4 –3.1 min) after methadone bolus injections, which was significantly earlier than with heroin (p ⫽ .028). Values below 90% were found for 13.7 ⫾ 7.9 min (11.1; 7.2–19.2 min). Mean respiration rate during this period was 10.9 ⫾ 2.0 breaths/ min (10.7; 9.0 –11.9 breaths/min), and mean values of SaO2 were 90.3 ⫾ 3.1% (88.5; 92.4 –90.2%) and thus significantly lower than baseline values (p ⫽ .041). In contrast to heroin-treated subjects, all methadone-treated subjects showed a normal respiration pattern at the end of the test session. Heroin injections resulted in minimal SaO2 values of 78.9 ⫾ 8.7% (76; 85–79%), whereas minimal SaO2 values after methadone were 84.4 ⫾ 6.6% (80; 88%– 86%). These minimal SaO2 values were significantly lower after heroin than after methadone injections (p ⫽ .008). Duration and severity of hypoxemia as surveyed with the clinical rating scale, after the two opioids, are shown in Figure 3. Mild hypoxemia was found in all (100%)
Figure 3. Incidence, severity, and duration of hypoxemia after opioid injection. Black diamonds represent hypoxemia after heroin injection in 16 subjects, gray circles represent hypoxemia after methadone injection in 9 subjects. Subjects contribute up to three data points (one for each range of hypoxemia severity).
subjects receiving heroin but in only 7 of 9 (78%) subjects receiving methadone (p ⫽ .12) and lasted for 5.5 ⫾ 2.5 min (5.5; 4.3–7.3 min) and 6.3 ⫾ 6.5 min (3.3; 1.0 –10.2 min), respectively. Overall, 13 of 16 (81%) heroin-treated subjects and 4 of 9 (44%) methadone-treated subjects developed moderate hypoxemia (p ⫽ .087) lasting 2.2 ⫾ 2.0 min (1.7; .9 –2.5 min) and 2.7 ⫾ 2.3 min (.8; 1.8 –3.8 min), respectively. Severe hypoxemia developed in 8 of 16 (50%) heroin-treated and in 3 of 9 (33%) methadonetreated subjects (p ⫽ .68). Severe hypoxemia lasted for a period of 1.5 ⫾ 2.5 min (.4; .3–1.1 min) and .7 ⫾ .9 min (.2; .5–1.7 min), respectively.
Cardiovascular System Baseline values of heart rate were comparable (p ⫽ .19) in the heroin and methadone groups (74.4 ⫾ 11.8 beats/min [72.0; 66.7– 84.4 beats/min] and 82.2 ⫾ 16.9 beats/min [85.9; 66.4 –95.7 beats/min], respectively). In general, we found increased heart rates during and shortly after opioid injections. Simultaneously with the onset of respiratory depression, the heart rate decreased in all heroin-treated subjects (Figure 4). We found a characteristic heart rate pattern with episodes of bradycardia during apnea and heart rate increases within the ensuing inspiration periods. We observed heart rate variations of up to 22 beats/min
858
R. Stoermer et al
BIOL PSYCHIATRY 2003;54:854 – 861
Figure 5. Electroencephalogram (EEG) segment of 10 sec duration shows the onset of the paroxysmal activity 5.5 min after bolus injection of heroin (150 mg). Corresponding respiratory and circulatory data are given in Figure 1.
Figure 4. Incidence, heart rate, and duration of bradycardia episodes after heroin injection in 16 subjects. Symbols represent the incidence of bradycardia in the specified range.
during a respiration cycle. In two subjects, the heart rate fell below 30 beats/min. One of these subjects showed recurrent sinus arrests and idioventricular rhythms at a heart rate below 30 beats/min lasting for 1.2 min. In 7 of 16 (44%) heroin-treated subjects, heart rates between 40 and 50 beats/min with a mean duration of 3.5 ⫾ 4.3 min (.4; .1–9.0 min) were observed. After methadone injection, the heart rate changes were negligible, and bradycardia was not observed.
EEG Because of the difficult measurement circumstances (withdrawal symptoms, sweating, excitation, lack of compliance), usable EEG recordings were obtained from only 9 of 16 heroin-treated subjects and from 8 of 9 methadonetreated subjects. In five EEG recordings among the heroin group, irregular paroxysmal activity was noted, with a time delay of 5.6 ⫾ .3 min after the injection. The pattern of paroxysmal activity varied between the subjects, although nonspecific intermittent irregular high-voltage theta– delta slowing occurred on the background of normal EEG activity in all five subjects. In one heroin-treated subject, spike and wave activity occurred, whereas in two other heroin-treated subjects, sharp waves were noted. Figure 5 provides an example of an EEG in which a more regular 3/sec paroxysmal delta activity developed 5.5 min after heroin injection (same subject as in Figure 1).
To illustrate the complex spectral EEG changes after drug application, a corresponding compressed spectral array is shown in Figure 2, which displays the time course of relative power values in .5-Hz intervals. About 5 min after injection, an initial increase in alpha power (8 –12Hz) could be observed, followed by a marked decrease in alpha power and an increase in delta power (1–3.5Hz). The delta power increase was largely caused by the paroxysmal 3/sec delta activity as shown in Figure 5, but an increase in the slow delta range occurred as well. Spectral changes were closely linked to the development of the hypoxemic state as shown in Figure 1 and restored during increase of the oxygen saturation approximately 7–10 min after injection. All irregular EEG recordings were obtained from subjects showing moderate to severe hypoxemia. Three of them showed intermittent bradycardia. In four EEG recordings, visual inspection did not reveal any signs of enhanced excitability. In contrast, only one subject in the methadone group exhibited a similar hypersynchronous pattern after injection. No clinical signs of seizures were observed in any subject.
Adverse Effects and Interventions In four sessions involving heroin, breathing instructions and oxygen supply were needed because of severe hypoxia and bradycardia. Here the oxygen supply was started between minute 5 and 8 following drug administration. The decision for intervening with oxygen and breathing instructions depended on the clinical status of the subject. Because none of the subjects lost consciousness and all could respond to verbal instructions, these interventions
Safety of Injectable Opioid Maintenance Treatment
resulted in an immediate improvement of cardiorespiratory functions.
Discussion Currently, oral MMT is the most widely used pharmacologic treatment of opioid dependence, but it is not accepted by all opioid-dependent patients (Bammer et al 1999). Consequently, IV MMT and IV HAT have been introduced in Switzerland, and in the Netherlands the government has proposed the introduction of IV or inhalable heroin as a coprescription to MMT. In the United Kingdom, IOT has been an established treatment for many decades; however, the level of medical supervision of the latter is low compared with the Swiss IOT program. Our study assessed the acute respiratory, circulatory, and electrocortical effects within the first 25 min after IOT injection in opioid-dependent subjects. There are some general limitations regarding research with illicit drugs in this group of severely drug-dependent patients that are related to the reduced ability to cooperate. To get the most reliable data with respect to drug safety, our study design was adapted as closely as possible to the outpatient setting. Accordingly, the actual doses of opioids varied greatly within the treatment groups. This reflects experience showing that, for pharmacologic effect, individualized opioid dose (adapted to the individual level of receptor downregulation) is more important than absolute dose. Therefore, each subject was given the dose he or she received for at least 1 week before the start of the study without significant subjective adverse effect. Opioids are known to depress respiration by affecting both central respiratory centers and peripheral chemoreceptors (Reisine and Pasternak 1996; Yeadon and Kitchen 1989). Furthermore, opioids may decrease hemodynamic performance by direct vascular effects, reducing sympathetic and enhancing vagal tone (Bailey 1993; Ghuran and Nolan 2000). In our study, opioid administration resulted in a stereotyped response pattern of respiratory parameters. Cardiovascular effects were seen only with heroin. In all cases, heroin caused respiration rates matching, at least temporarily, the clinical criteria of heroin intoxication (Sporer 1999) and resulted in a marked decrease in SaO2 for relevant time periods. In parallel with the onset of respiratory depression, the heart rate decreased in herointreated subjects, with some subjects presenting severe bradycardia. Irregular paroxysmal EEG activities coinciding with the time course of bradycardia and hypoxemia were found. The heroin-induced EEG patterns were clearly distinguishable from the well-known major slowing of EEG frequency corresponding to opioid-induced anesthesia. The EEG changes associated with high-dose
BIOL PSYCHIATRY 2003;54:854 – 861
859
opioid anesthesia in humans are well characterized and typically do not reveal paroxysmal activity (Smith et al 1989). The hypoxemic state in our patients might facilitate an increase of excitability. Similar effects on electrocortical activity after IV heroin application have been reported (Roubicek et al 1969; Volavka et al 1971), but a correlation with cardiopulmonary parameters has not been described previously. To assess the clinical implications of our findings, we note that our study conditions differed in some aspects from the common IOT setting. First, laboratory conditions provide different external stimuli than the regular treatment situation. Moreover, whereas the opioid is selfadministrated (active administration) in the IOT settings, the opioids (or placebo) in this study were injected in a standardized way as a bolus injection over 30 sec (passive administration) to minimize between-subject variability. This procedure, which lacked typical drug-paired cues, may have caused a reduced (conditioned) tolerance, resulting in an enhanced opioid effect. There is growing evidence that these difference in the mode of administration may significantly influence the effects of opioids (Siegel 1999; Siegel et al 2000). On the other hand, drug self-administration results in much higher rates of drug injection, leading most probably to even more pronounced respiratory depression and central vagal stimulation. Furthermore, the severity of effects seen in our study is underestimated because four subjects required intervention including oxygen supply after heroin application. Our findings have shown that common treatment doses of IV heroin and, to a lesser extent, IV methadone induce clinically relevant hypoxemia. Heroin-induced hypoxemia, often combined with bradycardia, may result in a reduced cerebral blood supply, thus putting opioid users at risk for the development of cerebral hypoxia or ischemia. The observed physiologic responses are also likely to occur in the nontreatment setting. Reports of cardiovascular failure (Ghuran and Nolan 2000), nonobstructive stroke (Brust and Richter 1976; Buttner et al 2000; Jensen et al 1990) with permanent or transient neurologic deficits (Neiman et al 2000), and possibly heroin-associated seizures (Alldredge et al 1989; Ng et al 1990; Richter et al 1973) in opioid-dependent subjects under nontreatment conditions possibly reflect general risks of opioid use. Our findings raise concerns about potential side effects of IV opioids and the possibility of accidental death even in a controlled treatment setting. Data about such deaths or other serious events related to IOT in Switzerland are not available. Apart from the risks of overdose, contamination, multiple drug use, infection, or accidents, long-term complications such as brain lesions induced by frequent hypoxic states (Andersen and Skullerud 1999; Oehmichen et al 1996) and consequent cognitive impairment (Darke et
860
BIOL PSYCHIATRY 2003;54:854 – 861
al 2000) appear to be possible not only with illicit opioid use but with therapeutic use as well. Loss of opioid tolerance, multiple drug use, and systemic disease have been identified as risk factors of illicit opioid overdose (Warner Smith et al 2001). We hypothesize that these factors may also be relevant for safety in the IOT setting. A better understanding of tolerance-modifying mechanisms, such as environmental factors, comedication, and multiple drug use is clearly needed. Our results suggest that the clinical practice of IV MMT and IV HAT could be improved with respect to safety. We propose basic monitoring following opioid injection could be helpful to identify acute hazardous situations. Intervention strategies should be developed to prevent acute hazards or permanent damage. The criteria for the prescription and therapeutic use of IV MMT or IV HAT should be critically revised. The objective of this study was to investigate safety of IOT. A judgment of overall benefit and risks of this treatment must balance these risks with those of untreated opioid addiction. Therefore, additional studies of the safety of IOT are required.
This study was supported by Grant No. 316.98.8104 from the Swiss Federal Office of Public Health. We thank all participants, without whom this study would not have been possible. We gratefully acknowledge the valuable support of M. Kuntze and A. Bullinger. In addition, our thanks go to M. Schmidlin and M. v. Arx for their excellent technical laboratory support and J. Coutino and S. Begre´ for medical assistance during the test sessions.
References Alldredge BK, Lowenstein DH, Simon RP (1989): Seizures associated with recreational drug abuse. Neurology 39:1037– 1039. Andersen SN, Skullerud K (1999): Hypoxic/ischaemic brain damage, especially pallidal lesions, in heroin addicts. Forensic Sci Int 102:51–59. Bailey PL (1993): Opioids in operative anesthesia. In: Herz A, editor. Opioids, Volume II. Berlin: Springer Verlag. Bammer G, Dobler-Mikola A, Fleming PM, Strang J, Uchtenhagen A (1999): The heroin prescribing debate: Integrating science and politics. Science 284:1277–1278. Brust JC, Richter RW (1976): Stroke associated with addiction to heroin. J Neurol Neurosurg Psychiatry 39:194 –199. Buttner A, Mall G, Penning R, Weis S (2000): The neuropathology of heroin abuse. Forensic Sci Int 113:435–442. Copeman M (2002): Heroin prescription for opioid addicts. Lancet 359:889 –890. Darke S, Sims J, McDonald S, Wickes W (2000): Cognitive impairment among methadone maintenance patients. Addiction 95:687–695. Drucker E (2001): Injectable heroin substitution treatment for opioid dependency. Lancet 358:1385.
R. Stoermer et al
Dursteler-Mac Farland KM, Stormer R, Seifritz E, Hug I, Muller-Spahn F, Ladewig D, Stohler R (2000): Opioidassociated effects on oxygen saturation. Addiction 95:285– 287. Farrell M, Hall W (1998): The Swiss heroin trials: Testing alternative approaches. Bmj 316:639. Ghuran A, Nolan J (2000): Recreational drug misuse: Issues for the cardiologist. Heart 83:627–633. Hock C, Stormer R, Dursteler K, Muller-Spahn F, Ladewig D, Stohler R (1999): Heroin-induced deoxygenation of cerebral hemoglobin: Monitoring by means of near-infrared spectroscopy. Ann Neurol 45:134 –135. Jensen R, Olsen TS, Winther BB (1990): Severe non-occlusive ischemic stroke in young heroin addicts. Acta Neurol Scand 81:354 –357. Kuo II, Fischer B, Vlahov D (2000): Consideration of a North American heroin-assisted clinical trial for the treatment of opiate-dependent individuals. Int J Drug Policy 11:357–370. Moller JT, Johannessen NW, Berg H, Espersen K, Larsen LE (1991): Hypoxaemia during anaesthesia—an observer study. Br J Anaesth 66:437–444. Neiman J, Haapaniemi HM, Hillbom M (2000): Neurological complications of drug abuse: Pathophysiological mechanisms. Eur J Neurol 7:595–606. Ng SK, Brust JC, Hauser WA, Susser M (1990): Illicit drug use and the risk of new-onset seizures. Am J Epidemiol 132:47– 57. Oehmichen M, Meissner C, Reiter A, Birkholz M (1996): Neuropathology in non-human immunodeficiency virus-infected drug addicts: Hypoxic brain damage after chronic intravenous drug abuse. Acta Neuropathol (Berl) 91:642–646. Perneger TV, Giner F, del Rio M, Mino A (1998): Randomised trial of heroin maintenance programme for addicts who fail in conventional drug treatments. Bmj 317:13–18. Perneger TV, Mino A, Giner F, Broers B (2000): Patterns of opiate use in a heroin maintenance programme. Psychopharmacology (Berl) 152:7–13. Rehm J, Gschwend P, Steffen T, Gutzwiller F, Dobler-Mikola A, Uchtenhagen A (2001): Feasibility, safety, and efficacy of injectable heroin prescription for refractory opioid addicts: A follow-up study. Lancet 358:1417–2420. Reisine T, Pasternak G (1996): Opioid analgesics and antagonists. In: Goodman LS, Perry M, Gilman AG, Hardman JG, editors. Goodman and Gilman’s The Pharmacological Basis of Therapeutics. New York: McGraw-Hill. Richter RW, Pearson J, Bruun B, Challenor YB, Brust JC, Baden MM (1973): Neurological complications of addiction to heroin. Bull N Y Acad Med 49:3–21. Roubicek J, Zaks A, Freedman AM (1969): EEG changes produced by heroin and methadone. Electroencephalogr Clin Neurophysiol 27:667–668. Rusche S (1999): Prescribing heroin. Science 285:53; discussion: 533. Satel SL, Aeschbach E (1999): The Swiss heroin trials. Scientifically sound? J Subst Abuse Treat 17:331–335. Sheldon T (2002): Netherlands considers prescribing heroin to addicts. BMJ 324:385.
Safety of Injectable Opioid Maintenance Treatment
Siegel S (1999): Drug anticipation and drug addiction. The 1998 H. David Archibald Lecture. Addiction 94:1113–1124. Siegel S, Baptista MA, Kim JA, McDonald RV, Weise-Kelly L (2000): Pavlovian psychopharmacology: The associative basis of tolerance. Exp Clin Psychopharmacol 8:276 –293. Smith NT, Benthuysen JL, Bickford RG, Sanford TJ, Blasco T, Duke PC, et al (1989): Seizures during opioid anesthetic induction–are they opioid-induced rigidity? Anesthesiology 71:852–862. Sporer KA (1999): Acute heroin overdose. Ann Intern Med 130:584 –590. Stein M, Luparello TJ (1967): Measurement of respiration. In: Brown CC, editor. Methods in Psychophysiology. Baltimore: Williams & Wilkins, 75–94. Stohler R, Dursteler KM, Stormer R, Seifritz E, Hug I, SattlerMayr J, et al (1999): Rapid cortical hemoglobin deoxygenation after heroin and methadone injection in humans: A preliminary report. Drug Alcohol Depend 57:23–28. Uchtenhagen A, Gutswiller F, Dobler-Mikola A (1997): Programme for a medical prescription of narcotics: Final report of the research representatives. Summary of the synthesis
BIOL PSYCHIATRY 2003;54:854 – 861
861
report. Zurich: University of Zurich. Available at http:// www.cfdp.ca/switz.html. van Kolfschooten F (2002): Dutch investigators recommend prescription of heroin to addicts. Lancet 359:590. Volavka J, Zaks A, Roubicek J, Fink M (1971): Acute EEG effects of heroin and naloxone. Electroencephalogr Clin Neurophysiol 30:165. Warner Smith M, Darke S, Lynskey M, Hall W (2001): Heroin overdose: Causes and consequences. Addiction 96:1113– 1125. White JM, Irvine RJ (1999): Mechanisms of fatal opioid overdose. Addiction 94:961–972. World Health Organization (1999): Report of the External Panel on the Evaluation of the Swiss Scientific Studies of Medically Prescribed Narcotics to Drug Addicts. Geneva: World Health Organization. Available at: http://www.emory.edu/ NFIA/legal/heroin/whoheroin.html. Yeadon M, Kitchen I (1989): Opioids and respiration. Prog Neurobiol 33:1–16. Zador D (2001): Injectable opiate maintenance in the UK: Is it good clinical practice? Addiction 96:547–553.