- Email: [email protected]
Opioid detoxification with buprenorphine, clonidine, or methadone in hospitalized heroin-dependent patients with HIV infection
Drug and Alcohol Dependence 69 (2003) 263 /272 www.elsevier.com/locate/drugalcdep
Opioid detoxification with buprenorphine, clonidine, or methadone in hospitalized heroin-dependent patients with HIV infection Annie Umbricht a,1, Donald R. Hoover b, Marvin J. Tucker a, Jo M. Leslie c, Richard E. Chaisson c, Kenzie L. Preston a,* b
a National Institute on Drug Abuse Intramural Research Program, 5500 Nathan Shock Drive, Baltimore, MD 21224, USA Department of Statistics, Institute for Health, Health Care Policy and Aging Research, Rutgers University, Piscataway, NJ 08854, USA c Johns Hopkins University School of Medicine, Division of Infectious Diseases, Baltimore, MD 21205, USA
Received 24 January 2002; received in revised form 14 September 2002; accepted 7 October 2002
Abstract With the growing role of intravenous drug use in the transmission of HIV infection, HIV-infected patients frequently present with comorbid opioid dependence. Yet, few empirical evaluations of the efficacy and consequences of opioid detoxification medications in medically ill HIV-infected patients have been reported. In a randomized, double-blind clinical trial, we evaluated the impact of three medications on the signs and symptoms of withdrawal and on the pain severity in heroin-dependent HIV-infected patients (N/55) hospitalized for medical reasons on an inpatient AIDS service. Patients received a 3-day pharmacologic taper with intramuscular buprenorphine (n /21), oral clonidine (n /16), or oral methadone (n /18), followed by a clonidine transdermal patch on the fourth day. Observed and self-reported measures of opioid withdrawal and pain were taken 1 /3 times daily for up to 4 days. Opiate administration used as medically indicated for pain was also recorded. Observer- and subject-rated opiate withdrawal scores decreased significantly following the first dose of medication and overall during treatment. Among all 55 subjects, selfreported and observer-reported pain decreased after treatment (on average observer-rated opioid withdrawal scale (OOWS) scores declined 5.6 units and short opioid withdrawal scale (SOWS) declined 4.8 units, P B/0.001, for both) with no indication of increased pain during medication taper. There were no significant differences of pain decline and other measures of withdrawal between the three treatment groups. During the intervention period, supplemental opiates were administered as medically indicated for pain to 45% of the patients; only 34% of men versus 62% of women received morphine (P B/0.05). These findings suggest buprenorphine, clonidine, and methadone regimens each decrease opioid withdrawal in medically ill HIV-infected patients. Published by Elsevier Science Ireland Ltd. Keywords: Opioid dependence; Pain management; HIV; Heroin; Withdrawal
1. Introduction Hospitalization of opioid-dependent individuals for acute (as well as chronic) illness often interrupts their heroin use and can result in withdrawal. Untreated opioid withdrawal, aside from being very unpleasant,
* Corresponding author. Tel.: /1-410-550-1639; fax: /1-410-5501528 E-mail address: [email protected] (K.L. Preston). 1 Present address: Sinai Hospital Addictions Recovery Program, Baltimore, MD 21215, USA.
can confound medical assessment and increase patients’ anxiety towards necessary medical procedures, resulting in conflicts with the healthcare team or in premature discharges against medical advice (with potentially harmful consequences). Although pharmacological treatments have been shown to minimize opioid withdrawal intensity in medically stable patients, their effects on withdrawal have not been evaluated in drug users hospitalized for an acute illness. Of particular interest are hospitalized drug users infected with HIV, as injection drug use is a major cause of HIV transmission in the United States (Center
03765-8716/02/$ - see front matter. Published by Elsevier Science Ireland Ltd. PII: S 0 3 7 6 - 8 7 1 6 ( 0 2 ) 0 0 3 2 5 - 3
264
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
for Disease Control, 1997; Karon et al., 2001). When HIV-infected drug users are hospitalized, they are brought in contact with healthcare professionals who could alleviate drug use withdrawal discomfort and possibly intervene in the cycle of drug use. Besides withdrawal-related discomfort during hospitalization, HIV-infected opioid users may experience additional pain from their underlying HIV-related illnesses (Breitbart et al., 1996, 1997; Rosenfeld et al., 1996). Neurologic changes associated with HIV progression may affect the response to opioid detoxification and perhaps even exacerbate pain from withdrawal-related conditions. For example, animal research has demonstrated that neuromediator changes underlying neuropathic pain are similar to those observed in opioid tolerance (Mao et al., 1995). This finding raises the question of whether intensity of HIV-related pain could be increased by opioid withdrawal and, thus, promote relapse to heroin use upon discharge. It is, therefore, important to assess the potential impact of opioid detoxification on pain syndromes and to develop standards of care that minimize suffering. This study was designed to evaluate responses to three different opioid detoxification medications (clonidine, methadone, and buprenorphine) on withdrawal and pain measures while HIV-infected patients received treatment for their acute medical condition. Clonidine, an a2-noradrenergic agent, is a presynaptic inhibitor of norepinephrine release in the locus coeruleus with weak analgesic activity at the spinal level (Aghajanian, 1978; Gold et al., 1978, 1980; Guthrie, 1990; Striebel et al., 1993; Bischoff and Kochs, 1993), though the importance of this activity in opioid-dependent human is not known. Methadone, a m-opioid agonist with pharmacological properties qualitatively similar to those of morphine, has been recommended for managing opioid dependence in the hospital setting (Perkins, 1970; Stimmel, 1997). Buprenorphine is a partial m-opioid agonist analgesic that is 25/50 times more potent than morphine (Jaffe and Martin, 1990; Jasinski et al., 1978). All three medications have been tested in clinical trials for opioid detoxification and shown to be effective in suppressing opioid withdrawal in healthy opioid-dependent volunteers seeking treatment (Umbricht et al., 1999; Jasinski et al., 1985; Agren, 1986; Buccafusco, 1992; Cheskin et al., 1994; Nigam et al., 1993; O’Connor et al., 1997; Bearn et al., 1998; Bickel et al., 1988). The outcome of opioid detoxification in HIV-infected individuals hospitalized for an acute medical condition has not been evaluated. We hypothesized that, while all three regimens would be effective among hospitalized HIV-infected persons, buprenorphine would be superior to clonidine and methadone because it has high receptor affinity, a long duration of action, and low intensity of withdrawal symptoms following abrupt discontinuation (Jasinski et al., 1978). Because of its partial agonist
activity, buprenorphine causes little or no cardiorespiratory depression in opioid-tolerant individuals, even at high doses (Walsh et al., 1994; Preston et al., 2001) and, thus, may be the safest in medically compromised patients. The objectives of this study were to compare the effectiveness of these three opioid detoxification methods on withdrawal and pain scores in HIV/AIDS patients during hospitalization for an acute medical condition. A subsidiary objective was to evaluate patterns of medical indication for opiate (morphine) administration during the detoxification and associated treatment outcome.
2. Methods 2.1. Patients and setting Patients were admitted to the inpatient AIDS service of the Johns Hopkins Hospital, either through the emergency department or via the outpatient AIDS clinic. Within 18 h of admission, subjects were screened by questionnaire and urine toxicology for current heroin use and for evidence of physical dependence. Inclusion criteria were HIV seropositivity, age of at least 18, hospitalization for an acute medical illness, and current opioid dependence by self-report and physical examination. Exclusion criteria were concurrent alcohol dependence, inability to give informed consent, acute psychosis or AIDS dementia (measured by the AIDS dementia scale; Power et al., 1995), hypotension, bradycardia, coagulopathy or severe thrombocytopenia precluding intramuscular injections, and enrollment in methadone maintenance treatment. Eligibility was determined in a 45 min screening procedure. All volunteers gave informed written consent prior to participation and were paid $10 for the screening procedure, and $10 a day for the 4-day study. Each patient’s physician was informed about his or her participation in the study. The study was approved by the National Institute on Drug Abuse and the Johns Hopkins University School of Medicine Institutional Review Boards. 2.2. Medications Study medications were administerd double-blind, triple-dummy in tapering doses over 3 days. Buprenorphine (Buprenex 0.3 mg injection, Reckitt & Colman Pharmaceuticals, Inc., Richmond, VA) was administered intramuscularly: 0.6 mg every 4 h on day 1, 0.6 mg every 6 h on day 2, and 0.6 mg every 8 h on day 3. Clonidine (Clonidine HCl 0.1 mg tablet, Schein Pharmaceutical, Inc., Florham Park, NJ) was administered orally as a loading dose of 0.2 mg, followed by 0.1 mg
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
every 4 h on day 1, 0.1 mg every 6 h on day 2, and 0.1 mg every 8 h on day 3. Methadone (Methadone HCl USP, 10 mg tablet, Roxane Laboratories, Inc., Columbus, OH) was administered orally once a day, 30 mg on day 1, 20 mg on day 2, and 10 mg on day 3. Methadone and its placebo were dispensed in blue capsules (size 0), and clonidine and its placebo were dispensed in yellow capsules (size 0). Placebo capsules were filled with dextrose anhydrous granular USP (Amend Drug & Chemical Co., Inc., Irvington, NJ). Equivalent volumes of normal saline given intramuscularly served as the buprenorphine placebo. In addition to these study medications, to prevent the development of opioid withdrawal between the time of admission to the hospital and initiation of the study, attending physicians were instructed to prescribe morphine (recommended dose: 10 mg intramuscularly every 4 h as needed) up to 6 h prior to enrollment. The use of morphine and other opiates, as adjunctive medications for pain during the study intervention, was also recorded as an outcome measure. At the end of the study, starting on day 4, a transdermal clonidine patch (Catapres-TTS-2 patch, Boehringer Mannheim Corp./Roche Diagnostic Systems, Inc., Somerville, NJ) was applied to all patients to minimize protracted withdrawal symptoms. 2.3. Randomization and blinding procedures After patients consented to participate, they were randomly assigned to one of the three treatment groups: clonidine, buprenorphine, or methadone. To obtain comparable treatment groups, patients were stratified on four characteristics: opioid withdrawal severity, presence of pain, cocaine use, and most recent CD4 cell count (above or below 200 mm 3). To maintain the blind, one active medication and two inactive medications were administered to all participants. At each medication time, all patients received a subcutaneous injection (either buprenorphine or a placebo) and capsules (active treatment or placebo) to swallow. As all three treatments have demonstrated effectiveness for decreasing symptoms of opioid withdrawal, the inclusion of a placebo control group (no active medications) was judged to be unethical, especially among a medically ill population. 2.4. Data collection Data were collected by trained interviewers with paper and pencil questionnaires. Baseline measures (T0) were collected prior to administration of the test medications, representing patient status on the first hospital day. Thereafter, as described below, measures were taken 1 /3 times daily depending on the measure.
265
Measures taken once a day were collected in the late afternoon, measures taken twice a day were collected in the morning and late afternoon, and measures taken three times a day were collected at the beginning of each 8 h nursing shift. Patients rated opioid withdrawal symptoms on the short opioid withdrawal scale (SOWS; total score: 0 /30; Gossop, 1990) and three visual analog scales (VASs; scores 0/100 mm). Patients completed SOWS twice a day, rating the severity (not at all, a little, quite a bit, or extremely) of 10 symptoms (feeling sick, stomach cramps, muscle spasms/twitching, feeling of coldness, heart pounding, muscular tension, aches and pains, yawning, runny eyes, and insomnia). Patients completed VAS once a day in the afternoon, rating ‘‘crave heroin’’, ‘‘need heroin’’, and ‘‘heroin sickness’’ on 100 mm lines with the left extremity anchored ‘‘not at all’’ and the other ‘‘most ever’’. Pain was assessed twice a day with Pain Faces cards and with two items from VAS. The Pain Faces cards, a set of eight cards representing faces with a range of emotions from happiness to increasing degree of suffering (score 0 /8; Frank et al., 1982), were presented in random order. Patients selected the face that best represented how they felt. The two VAS items were ‘‘How much pain do you have now?’’ and ‘‘In the last 24 hours, how much pain did you have?’’ Patients completed these as described above. An observer-rated opioid withdrawal scale (OOWS, total score: 0 /30; modified from the Clinical Institute Narcotic Assessment; Peachey and Lei, 1988) was completed by nurses three times daily. The 11 items on OOWS were nausea/vomiting, abdominal complaints/ diarrhea, subjective changes in temperature, myalgia, gooseflesh, sweating, restlessness, tremor, lacrimation, rhinorrhea, and yawning. Pupil diameter was evaluated by monocular infrared television pupillometry (Eye Dynamics, Inc., Torrance, CA). Blood pressure and heart rate were measured by nursing staff at change of shift. Demographic characteristics, most recent CD4 cell count, presumptive admission diagnoses, and morphine and other opiate administration data were collected by chart review. Use of adjunctive opioid medications during the study intervention was ascertained through review of medication administration records. Doses of non-morphine opioid drugs were converted to morphine equivalents (Stimmel, 1997). The time of discharge was decided by the medical team, based on the patient’s medical status. A number of patients were discharged prior to the scheduled end of the study. Because the study was not a primary determinant of discharge, time of discharge was not included as an outcome measure.
266
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
2.5. Statistical analyses Treatment outcome was evaluated as time-averaged data, i.e., by averaging measures during treatment after initiation of the study medications (mean treatment score; Tmean /mean(T1, T2,. . ., TN ), where N is the number of measures taken for that patient, maximum N was 4/12 depending on how frequently study measures were taken. This summary value was used for the following reasons: (1) the study was completed during a patient’s hospital course, which was driven by the medical treatment; (2) the time of discharge was determined by the medical condition and not influenced by withdrawal symptoms and thus the treatment assignment; and (3) due to the competing needs of clinical care, the same measures could not be taken at the same time points for all patients. A cross-sectional data set was created: baseline scores (T0), scores after the first medication dose (T1), mean treatment scores (Tmean), standard deviation of all treatment scores (T1, T2,. . ., TN ), difference between T1 and baseline (T1/T0), and difference between mean treatment score and baseline (Tmean/T0). Increased or higher scores (or changes) indicate more severe withdrawal and lack of therapeutic efficacy while lower scores indicate treatment benefit. The main outcome measures were compared between the treatment groups and analyzed with respect to the following covariates of interest: sex, presence of pain, OOWS ]/5, CD4 cell count 5/200 mm 3, morphine administration (patients who received morphine during the study intervention were compared with those who never received morphine). As needed, other comparisons were made with t-tests and contingency tables (exact tests). Multivariate analyses were performed using linear models. One-sample t-tests were used to test the null hypothesis for the differences T1/T0 and Tmean/T0; 95% confidence intervals (from the t-distribution) were calculated for these differences on OOWS, SOWS, scores on visual analogue scales, and on Pain Faces cards, heart rate, systolic and diastolic blood pressures, and pupil sizes. Adjustment for baseline characteristics (sex, presence of pain, OOWS /5, CD4 cell countB/200 mm 3, or the admission diagnosis ‘‘fever/cellulitis’’) did not statistically alter the results. Therefore, data are presented as unadjusted means. As appropriate, statistical comparisons were made with paired t-tests, two-sample t-tests, one-way ANOVA, exact tests and two-sample t-tests. Only P -values 5/0.05 are reported.
3. Results Sixty-three eligible heroin-dependent patients were enrolled in the study, and 21 were assigned to each treatment group. Data were available for 55 patients:
three patients were dropped out after baseline before receiving any study medication, and five were excluded due to medication errors (e.g., staff administered openlabel buprenorphine instead of ‘‘study (blinded placebo or active) buprenorphine’’). Table 1 shows patients’ characteristics at enrollment. There were no other significant differences in intake characteristics among the groups except that patients randomized to the buprenorphine group were more likely to have been admitted for fever/cellulitis (i.e., 63% versus approximately 20% for the other two groups; P B/0.05). A post hoc analysis controlling for this diagnosis showed that it did not significantly change the results of the outcome measures. Self-reported heroin use histories were available for 43 participants. On average, participants had used heroin for 17.9 (S.D. 8.6; range 4/37) years and had used 13.0 (S.D. 10.7; range 0 /30) days out of the last 30. Twenty four of 55 (43%) participants completed the study; within the treatment groups, 7 of 21 (33.3%) in the buprenorphine group, 8 of 16 (50%) in the clonidine group, and 9 of 18 (50%) in the methadone group completed. Among the 31 who did not complete, 25 (81%) were discharged from hospital, transferred to another unit or dropped out for non-study-related medical reasons prior to completion; two (6%, both in the buprenorphine treatment group) left the hospital against medical advice; one (3%, in the buprenorphine treatment group) chose to discontinue their participation; one (3%, in the buprenorphine treatment group) was dropped from the study due to non-compliance with study procedures; and two (6%, in the clonidine group) dropped out due to side effects. Drop out from the study was not correlated with reason for admission, CD4 count on admission, sex, OOWS, nor treatment group. 3.1. Subject- and observer-rated opioid withdrawal Fig. 1 shows the mean withdrawal scores at baseline (T0), after the first dose of study medications (T1), and during treatment (Tmean) for all patients and by treatment group. Table 2 shows the mean changes in scores between baseline and during treatment (Tmean/T0) for all patients and by group with the 95% CI for the difference. Overall, both subject- and observer-rated opioid withdrawal decreased significantly after initiation of the study medications (P B/0.001 for each group) as indicated by SOWS and OOWS (Fig. 1; Table 2). After a marked first-dose effect, further improvements were minimal, and overall mean scores during treatment (Tmean) were not significantly lower than scores after one treatment dose (T1) (Fig. 1). The overall mean timeaveraged decreases ranged from /5.1 to /6.0 for OOWS and /3.3 to /4.7 for SOWS. Scores on subjective VAS for ‘‘crave heroin’’, ‘‘need heroin’’, and ‘‘heroin sickness’’ indicated more gradual
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
267
Table 1 Patient characteristics at intake Variables
All patients
Buprenorphine
Clonidine
Methadone
N evaluable Years of age, mean9S.D. Race: African /American (%) Sex: male (%) CD4 cell 5 200/mm3 (%) Presence of pain (%)
55 39.795.6 98 62 67 78
21 39.695.4 95 71 67 81
16 40.095.4 100 69 69 81
18 40.095.9 100 44 68 72
Opioid withdrawal OOWS, mean9SE OOWS] 5 (%)
7.590.6 60
7.791.1 62
7.791.1 63
7.390.9 56
Admission diagnosis (%) Fever/cellulitis Pneumonia, PCP Tuberculosis Meningitis Other
36 29 7 7 20
62* 19 10 5 5
19* 38 6 6 31
22* 33 6 11 28
* P B 0.05 compared with other treatment groups.
declines (Fig. 1), possibly because the question referred to ‘‘in the last 24 hours’’ (thus, re-evaluation a few hours after the first medication dose could not elicit the full effect of the medication). Time-averaged mean VAS scores during treatment for these three measures were significantly lower (/14.1 to /23.7 mm) than at baseline (T0 versus Tmean) for the three groups combined. There were, however, no significant differences in the magnitude of the decrease in withdrawal (Tmean/T0) between the treatment groups (Table 2). At no time
during treatment did withdrawal scores exceed baseline scores for any individual (data not shown). 3.2. Physiologic measures There were no major changes in blood pressure or heart rate during the observation time interval of the study, although a relatively small decrease in heart rate reached statistical significance (Table 2). There was a trend (P /0.06) toward myosis (decrease in pupil size)
Fig. 1. Mean subject- and observer-rated opioid withdrawal scores and pupil sizes for the treatment groups combined and separately. The bars represent mean scores at baseline (T0, black), after one medication dose (T1, stripes), and mean treatment scores (Tmean, white). Brackets indicate standard errors.
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
268
Table 2 Mean (95% confidence intervals) of scores during treatment minus baseline scores (TmeanT0): overall and by treatment groupa Variables
Overall
Buprenorphine
Clonidine
Methadone
N OOWS (0 /30) SOWS (0 /30) Crave heroin VAS (mm) Need heroin VAS (mm) Heroin sickness VAS (mm) Heart rate (beats per minute) Systolic blood pressure (mmHg) Pupil size (mm) Pain faces cards (1 /8) Pain now (VAS) Pain in the last 24 h (VAS)
55 5.6 4.7 17.8 14.1 23.7 5.7 2.5 0.3 1.0 20.5 20.3
21 6.0 5.8 17.9 19.6 20.6 6.3 7.3 0.4 0.9 20.9 22.4
16 5.3 4.7 20.5 17.5 25.3 5.7 2.1 0.3 0.9 22.0 20.0
18 5.1 3.3 16.9 5.5 26.7 5.7 2.2 0.7 1.3 19.0 19.0
a
(6.8, 4.3) (6.2, 3.1) (28.6, 6.7) (23.7, 4.5) (31.7, 15.7) (10.0, 1.5) (6.7, 1.6) (0.7, 0.1) (1.5, 0.6) (29.4, 11.7) (27.5, 13.0)
(8.4, 3.6) (8.7, 2.9) (37.4, 1.6) (36.4, 2.7) (35.3, 6.0) (14.8, 2.3) (14.0, 0.7) (1.2, 0.4) (1.7, 0.005) (36.8, 5.1) (33.8, 10.8)
(7.6, 2.9) (8.0, 1.4) (43.5, 2.5) (37.1, 2.2) (44.3, 6.3) (13.5, 2.1) (9.7, 5.5) (0.3, 0.8) (1.8, 0.001) (39.4, 4.5) (36.9, 3.1)
(7.2, 3.1) (4.9, 1.6) (33.8, 0.014) (21.0, 10.0) (36.5, 17.0) (12.6, 1.1) (9.6, 5.3) (1.3, 0.004) (2.0, 0.6) (33.7, 4.2) (30.0, 8.0)
Values in brackets indicate 95% confidence intervals; changes were considered statistically significant where confidence intervals do not include
0.
in the buprenorphine and methadone groups, consistent with a pharmacologic effect of treatment (Table 2; Fig. 1). Systolic blood pressure decreased significantly in the buprenorphine group during treatment but not in other groups (Table 2). Two patients in the clonidine group had to be discontinued from the study because of decreased systolic blood pressure (B/90 mmHg) and bradycardia (HR B/50 BPM). 3.3. Subject-rated pain Scores on the Pain Faces choice and pain VAS are shown in Fig. 2; differences between baseline (T0) and treatment (Tmean) are shown in Table 2. Overall, pain scores decreased during treatment, although more gradually than withdrawal symptoms (Fig. 2). The pain scores were significantly lower during treatment than at baseline overall and for each of the three groups (Table 2). Decreases in Pain Faces scores were statistically significant although the changes were small (/1 unit); decreases in the ‘‘Pain Now’’ and ‘‘Pain in the last 24 hours’’ were somewhat larger (/20 units). There were no trends toward a difference in pain outcome between the three medication groups. At no time during treatment did any individual’s pain score exceed his or her score at intake. 3.4. Use of morphine during the study intervention About 80% of the patients were in pain at baseline, and physicians had the option to administer morphine for pain treatment in addition to opioid detoxification treatment. Because of the possibility that pain would be exacerbated during detoxification, outcome measures were reanalyzed as a function of patterns of morphine administration during the study. Data on administration of morphine as adjunct medication for pain during the research treatment
(detoxification) were available for 53 patients. About half of those patients (24/53) received morphine. Intake characteristics and treatment assignments of patients who received morphine and those who did not are shown in Table 3. At baseline (Fig. 3), patients who received morphine had significantly lower Pain Faces (P /0.003; two-sample t-tests), observed (OOWS) withdrawal scores (P /0.03), and scores on the ‘‘need heroin’’ VAS (P /0.04) (Fig. 3). A statistically significant difference was also found on the base line ‘‘heroin sickness’’ scale between those who received no morphine and those who received some morphine (77.49/6.6 versus 58.59/6.6; P /0.047), respectively. There were no statistically significant differences in the proportion of patients receiving morphine during detoxification or in the amount of morphine administered between the medication groups. There was, however, a tendency for fewer patients in the methadone group to receive smaller doses of morphine. Mean doses (9/S.D.) administered by group during the entire observation period were 5.39/5.6 mg for the buprenorphine group, 4.89/2.7 mg for the clonidine group, and 3.09/2.0 mg for the methadone group. There was evidence of sex differences in morphine administration during detoxification and in self-reports of pain. Females were significantly more likely to receive morphine: 13 out of 21 women (62%) versus 11 out of 32 men (34%) received morphine (P /0.05). This is in spite of the fact that women tended to score lower on pain measures initially (T0), a difference that reached significance for the Pain Face choice: 71% of the men chose a Pain Face card scoring 6 or higher at baseline compared to only 29% of the women; mean scores on Pain Faces were 5.89/0.2 for men versus 4.79/0.3 for women (P B/0.005). Perhaps because of this baseline difference, men had greater decreases in pain scores during treatment. However, sex was found not to be related to changes in pain or withdrawal in multivariate
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
269
Table 3 Intake characteristics and treatment assignment as a function of whether participants received morphine for pain during treatment Variables
No morphine
Morphine
N Years of age, mean9S.D. Race: African /American (%) Sex: male (%) CD4 5 200 cells/mm3 (%) Reporting pain on admission (%) Opioid withdrawal (OOW] 5) (%)
29 39.695.4 100 72* 72 72 69
24 40.095.4 96 46* 58 88 50
Reason for admission (%) Fever, cellulitis, skin ulcer Pneumonia, PCP Tuberculosis Meningitis Other
28 35 14 10 14
50 21 / 4 25
Treatment assignment (%) Buprenorphine Clonidine Methadone
48 50 67
52 50 33
Total N 53: use of morphine unknown for two patients. * P 0.05.
Fig. 2. Mean subject-rated pain scores for the treatment groups combined and separately. The bars represent mean scores at baseline (T0, black), after one medication dose (T1, stripes), and mean treatment scores (Tmean, white). Brackets indicate standard errors.
analyses that included both sex and morphine use as covariates.
4. Discussion The post-treatment improvements in this study suggest that opioid withdrawal was effectively accomplished by treatment with any one of the three regimens during hospitalization for medical reasons. Although relatively high at baseline, opioid withdrawal symptoms and pain ratings decreased during the 4-day study period. This interpretation, however, must be considered in light of the lack of a control group (for ethical reasons) and potential longitudinal reporting
bias. The initiation of pharmacological treatments was associated with a marked first-dose effect, followed by a more gradual decline in discomfort. Reports of adverse effects of treatment were few and observed only in the clonidine group. Two patients in the clonidine group were terminated due to development of bradycardia and hypotension, suggesting that this medication is less safe in some patients. At no time during treatment did withdrawal or pain scores exceed scores at intake for any individual. During treatment for an acute medical condition, buprenorphine, methadone, or clonidine in tapered doses were equally effective and were associated with decreases in opioid withdrawal intensity, with no significant differences among the treatment groups. The sample size in each group may not have been large enough to detect moderate differences between treatment regimens; however, there were no trends in the withdrawal data to suggest that a larger sample would have detected differences between the treatment groups. Combined with medical treatment for the underlying admission diagnosis, the study medications were associated with decreases in pain scores during the 4-day observation period, with no evidence of exacerbation of pain or hyperalgesia. Post-treatment decline in pain was more gradual with a first treatment effect that was less striking than that observed for withdrawal symptoms. As with effects on opioid withdrawal, there were no differences among the three treatment groups on changes in pain scores. However, fewer patients in the methadone group received adjunctive morphine, and these patients received smaller doses than patients in other groups. Although many factors, in addition to
270
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
Fig. 3. Pain and withdrawal scores as a function of whether participants received morphine for pain during treatment. The bars represent mean scores at baseline (T0, black), after one medication dose (T1, stripes), and mean treatment scores (Tmean, white). Brackets indicate standard errors.
pain severity, such as staff attitudes and the patient’s behavior, determine whether a patient receives morphine, it is possible that methadone (30 mg) may have been a more effective analgesic than the corresponding split doses of buprenorphine. Our results differ from those of a 10-day detoxification study conducted among non-AIDS patients residing on a drug abuse research ward, where, in the first 3 days, buprenorphine was found to be more effective than clonidine for decreasing opioid withdrawal symptoms (Cheskin et al., 1994). The doses of buprenorphine used in that study were 2 mg given sublingually every 6 h for 48 h, and the doses of clonidine were 1.7 mg given orally in the first 48 h. With a bioavailability estimated to be approximately 50% (Mendelson et al., 1997), 2 mg sublingual buprenorphine solution is equivalent to 1 mg parenteral dose. Our study used approximately half the doses of both medications used by Cheskin. The conservative dosage regimen used in this study was based on our concerns for potentially cachectic or unstable patients in this randomized, double-blind trial. In treating patients on an unblinded basis, the efficacy of buprenorphine might be increased by the use of higher doses, on the model of the study by Cheskin et al. (1994), or by repeating the first doses at short intervals (every 30 min) until resolution of symptoms prior to initiating taper (based on clinical experience of one of the investigators, AU). Given the effects of clonidine on cardiovascular function, increasing clonidine doses in medically ill patients might not be safe. The starting dose of 30 mg of
methadone used in the study compares well with recommendations by Stimmel (1997) of a starting dose of 15 /20 mg, to be adjusted by an additional 10 mg in the following 4 /12 h according to symptoms. From the design of the study, and given the relatively large first treatment effect, it is not possible to exclude a beneficial effect of systematic, non-judgmental, patient interviews about withdrawal experience, in the absence of a placebo group. Inclusion of a placebo group would not have been ethical given the established efficacy of these medications for decreasing withdrawal symptoms in non-HIV-infected drug users and denial of any withdrawal treatment might have added to the burden of HIV patients’ illnesses. Morphine administration may be an indicator of relative treatment ineffectiveness. In opioid-tolerant individuals, when parenteral opiates are indicated to treat moderate to severe pain, administration of larger morphine doses at more frequent intervals than in opioid naive patients (e.g., 10 /20 mg intramuscularly or subcutaneously) has been recommended (Stimmel, 1997). In spite of these instructions to physicians, morphine was administered in small single doses, as witnessed by the low average morphine doses received during the entire study (even after patients who did not receive opioid medications were excluded from the denominator). There was no evidence that the adjunctive administration of morphine decreased withdrawal (Fig. 3). Interesting sex differences in reported pain and in morphine use were observed, though their significance is
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
unclear. Men tended to rate their pain higher at baseline and to show larger decreases after treatment was started, after which the sex difference in pain was no longer present. Women, nevertheless, were more likely to receive some morphine during the observation period. Our study shows that the majority of heroin-dependent, HIV-infected patients had symptoms of opioid withdrawal during the first day of hospitalization, as evidenced by high scores on objective and self-reported withdrawal scales in 60% of patients, and that most (80%) were in pain. Instructions had been given to the hospital staff to stabilize patients with morphine to prevent interference with patient care and comfort by the study protocol. It is possible that opioid withdrawal and pain were not recognized and were, thus, not addressed. Conducting an opioid detoxification study on a busy medical ward was a challenge. The intention was initially to enroll a higher sample size (81 total), but unexpected difficulties in patient recruitment occurred. Only one-half of the patients who had acquired HIV through injection drug use were still using heroin and, were opioid-dependent, a much smaller proportions than expected. About 10% of patients with a history of heroin use were enrolled in methadone maintenance programs and thus ineligible for this study. Furthermore, after 2 years of conducting the study, it became apparent that the population from which we were recruiting was a relatively closed cohort with the same patients being readmitted several times making it difficult to acquire new patients. The exclusion of medically and mentally unstable patients further limited the pool of eligible patients. The resulting small sample size may have decreased our ability to detect differences in outcome among the medications. Our patient selection criteria may also have limited the generalizability of the study. Nevertheless, we believe that the study showed that opioid-dependent patients were initially experiencing opioid withdrawal and pain and that treatment with modest doses of medication following patient assessment effectively decreased these discomforts without adverse effects or increases in pain.
Acknowledgements The study was funded by the National Institute on Drug Abuse, Intramural Research Program. We thank Marcia Hoffman, Betsey Emilien, Michelle Dequina, and Neena Thomas for their expertise in data collection and data management, and Dr. David Epstein for his constructive review of the manuscript. We also thank the nursing staff of the Johns Hopkins Inpatient AIDS Service for their cooperation and unconditional support of the study, and the Osler Medical Service for physician support.
271
References Aghajanian, G.K., 1978. Tolerance of locus coeruleus neurons to morphine and suppression of withdrawal response by clonidine. Nature 276, 186 /188. Agren, H.M., 1986. Clonidine treatment of the opiate withdrawal syndrome. Acta Psychiatr. Scand. 327, 91 /113. Bearn, J., Gossop, M., Strang, J., 1998. Accelerated lofexidine treatment regimen compared with conventional lofexidine and methadone treatment for inpatient opiate detoxification. Drug Alcohol Depend. 50, 227 /232. Bickel, W.K., Stitzer, M.L., Bigelow, G.E., Liebson, I.A., Jasinski, D.R., Johnson, R.E., 1988. A clinical trial of buprenorphine: comparison with methadone in the detoxification of heroin addicts. Clin. Pharmacol. Ther. 43, 72 /78. Bischoff, P., Kochs, E., 1993. Alpha2 agonists in anaesthesia and intensive care medicine. Anaesthesiol. Intensivmed. Notfallmed. Schmerzther. 28, 2 /12. Breitbart, W., McDonald, M.V., Rosenfeld, B., Passik, S.D., Hewitt, D., Thaler, H., Portenoy, R.K., 1996. Pain in ambulatory AIDS patients. I. Pain characteristics and medical correlates. Pain 68, 315 /321. Breitbart, W., Rosenfeld, B., Passik, S., Kaim, M., Funesti-Esch, J., Stein, K., 1997. A comparison of pain report and adequacy of analgesic therapy in ambulatory AIDS patients with and without a history of substance abuse. Pain 72, 235 /243. Buccafusco, J.J., 1992. Neuropharmacologic and behavioral actions of clonidine: interactions with central neurotransmitters. Int. Rev. Neurobiol. 33, 55 /107. Center for Disease Control, 1997, 1996. Update: trends in AIDS incidence, deaths, and prevalence */United States. MMWR 46, 165 /173. Cheskin, L.J., Fudala, P.J., Johnson, R.E., 1994. A controlled comparison of buprenorphine and clonidine for acute detoxification from opioids. Drug Alcohol Depend. 36, 115 /121. Frank, A., Moll, A., Hort, J., 1982. A comparison of three ways of measuring pain. Fheumatol. Rehabil. 21, 211 /217. Gold, M.S., Redmond, D.E., Kleber, H.D., 1978. Clonidine blocks acute opiate-withdrawal symptoms. Lancet, 2, 599 /601. Gold, M.S., Pottash, A.C., Sweeney, D., Kleber, H.D., 1980. Opiate withdrawal using clonidine: a safe, effective, and rapid nonopiate treatment. J. Am. Med. Assoc. 243, 343 /346. Gossop, M., 1990. The development of a short opiate withdrawal scale (SOWS). Addict. Behav. 15, 487 /490. Guthrie, S.K., 1990. Pharmacologic interventions for the treatment of opioid dependence and withdrawal. Ann. Pharmacother. 24, 721 / 734. Jaffe, J., Martin, W., 1990. Opioid analgesics and antagonists. In: Goodman, T.R.A., Gilman, Nies, A.S., Taylor, P. (Eds.), Goodman and Gilman’s The pharmacological Basis of Therapeutics. Pergamon Press, New York, pp. 485 /521 (Chapter 21). Jasinski, D.R., Pevnick, J.S., Griffith, J.D., 1978. Human pharmacology and abuse potential of the analgesic buprenorphine. Arch. Gen. Psychiatr. 35, 501 /516. Jasinski, D.R., Johnson, R.E., Kocher, T.R., 1985. Clonidine in morphine withdrawal. Arch. Gen. Psychiatr. 42, 1063 /1066. Karon, J.M., Fleming, P.L., Steketee, R.W., De Cock, K.M., 2001. HIV in the United States at the turn of the century: an epidemic in transition. Am. J. Public Hlth. 91, 1060 /1068. Mao, J., Price, D.D., Mayer, D.J., 1995. Mechanisms of hyperalgesia and morphine tolerance: a current view of their possible interactions. Pain 62, 259 /274. Mendelson, J., Upton, R.A., Everhart, E.T., Jacob, P., Jones, R.T., 1997. Bioavailability of sublingual buprenorphine. J. Clin. Pharmacol. 37, 31 /37.
272
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
Nigam, A.K., Ray, R., Tripathi, B.M., 1993. Buprenorphine in opiate withdrawal: a comparison with clonidine. J. Subst. Abuse Treat. 10, 391 /394. O’Connor, P.G., Carroll, K.M., Shi, J.M., Schottenfeld, R.S., Kosten, T.R., Rounsaville, B.J., 1997. Three methods of opioid detoxification in a primary care setting: a randomized trial. Ann. Intern. Med. 127, 526 /530. Peachey, J., Lei, H., 1988. Assessment of opioid dependence with naloxone. Br. J. Addict. 83, 193 /201. Perkins, I.H., 1970. Hospital treatment of narcotic addiction. Clin. Toxicol. 3, 571 /578. Power, C., Selnes, O.A., Grim, J.A., McArthur, J.C., 1995. HIV dementia scale: a rapid screening test. J. AIDS Hum. Retrovirol. 8, 273 /278. Preston, K.L., Umbricht, A., Huestis, M.A., Cone, E.J., 2001. Intravenous buprenorphine: effects of 2 to 16 mg doses in nondependent opioid abusers. Clin. Pharmacol. Ther. 69, 29.
Rosenfeld, B., Breitbart, W., McDonald, M.V., Passik, S.D., Thaler, H., Portenoy, R.K., 1996. Pain in ambulatory AIDS patients. II. Impact of pain on psychological functioning and quality of life. Pain 68, 323 /328. Stimmel, B., 1997. Pain and its Relief without Addiction: Clinical Issues in the Use of Opioids and other Analgesics. The Haworth Medical Press, New York. Striebel, H.W., Koenigs, D., Heil, T., 1993. The role of clonidine in clinical anaesthesiology. Anaestesist 42, 131 /141. Umbricht, A., Montoya, I.D., Hoover, D.R., Demuth, K.L., Chiang, C.-T., Preston, K.L., 1999. Naltrexone shortened opioid detoxification with buprenorphine. Drug Alcohol Depend. 57, 181 / 190. Walsh, S.L., Preston, K.L., Stitzer, M.L., Cone, E.J., Bigelow, G.E., 1994. Clinical pharmacology of buprenorphine: ceiling effects at high doses. Clin. Pharmacol. Ther. 55, 569 /580.
Opioid detoxification with buprenorphine, clonidine, or methadone in hospitalized heroin-dependent patients with HIV infection Annie Umbricht a,1, Donald R. Hoover b, Marvin J. Tucker a, Jo M. Leslie c, Richard E. Chaisson c, Kenzie L. Preston a,* b
a National Institute on Drug Abuse Intramural Research Program, 5500 Nathan Shock Drive, Baltimore, MD 21224, USA Department of Statistics, Institute for Health, Health Care Policy and Aging Research, Rutgers University, Piscataway, NJ 08854, USA c Johns Hopkins University School of Medicine, Division of Infectious Diseases, Baltimore, MD 21205, USA
Received 24 January 2002; received in revised form 14 September 2002; accepted 7 October 2002
Abstract With the growing role of intravenous drug use in the transmission of HIV infection, HIV-infected patients frequently present with comorbid opioid dependence. Yet, few empirical evaluations of the efficacy and consequences of opioid detoxification medications in medically ill HIV-infected patients have been reported. In a randomized, double-blind clinical trial, we evaluated the impact of three medications on the signs and symptoms of withdrawal and on the pain severity in heroin-dependent HIV-infected patients (N/55) hospitalized for medical reasons on an inpatient AIDS service. Patients received a 3-day pharmacologic taper with intramuscular buprenorphine (n /21), oral clonidine (n /16), or oral methadone (n /18), followed by a clonidine transdermal patch on the fourth day. Observed and self-reported measures of opioid withdrawal and pain were taken 1 /3 times daily for up to 4 days. Opiate administration used as medically indicated for pain was also recorded. Observer- and subject-rated opiate withdrawal scores decreased significantly following the first dose of medication and overall during treatment. Among all 55 subjects, selfreported and observer-reported pain decreased after treatment (on average observer-rated opioid withdrawal scale (OOWS) scores declined 5.6 units and short opioid withdrawal scale (SOWS) declined 4.8 units, P B/0.001, for both) with no indication of increased pain during medication taper. There were no significant differences of pain decline and other measures of withdrawal between the three treatment groups. During the intervention period, supplemental opiates were administered as medically indicated for pain to 45% of the patients; only 34% of men versus 62% of women received morphine (P B/0.05). These findings suggest buprenorphine, clonidine, and methadone regimens each decrease opioid withdrawal in medically ill HIV-infected patients. Published by Elsevier Science Ireland Ltd. Keywords: Opioid dependence; Pain management; HIV; Heroin; Withdrawal
1. Introduction Hospitalization of opioid-dependent individuals for acute (as well as chronic) illness often interrupts their heroin use and can result in withdrawal. Untreated opioid withdrawal, aside from being very unpleasant,
* Corresponding author. Tel.: /1-410-550-1639; fax: /1-410-5501528 E-mail address: [email protected] (K.L. Preston). 1 Present address: Sinai Hospital Addictions Recovery Program, Baltimore, MD 21215, USA.
can confound medical assessment and increase patients’ anxiety towards necessary medical procedures, resulting in conflicts with the healthcare team or in premature discharges against medical advice (with potentially harmful consequences). Although pharmacological treatments have been shown to minimize opioid withdrawal intensity in medically stable patients, their effects on withdrawal have not been evaluated in drug users hospitalized for an acute illness. Of particular interest are hospitalized drug users infected with HIV, as injection drug use is a major cause of HIV transmission in the United States (Center
03765-8716/02/$ - see front matter. Published by Elsevier Science Ireland Ltd. PII: S 0 3 7 6 - 8 7 1 6 ( 0 2 ) 0 0 3 2 5 - 3
264
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
for Disease Control, 1997; Karon et al., 2001). When HIV-infected drug users are hospitalized, they are brought in contact with healthcare professionals who could alleviate drug use withdrawal discomfort and possibly intervene in the cycle of drug use. Besides withdrawal-related discomfort during hospitalization, HIV-infected opioid users may experience additional pain from their underlying HIV-related illnesses (Breitbart et al., 1996, 1997; Rosenfeld et al., 1996). Neurologic changes associated with HIV progression may affect the response to opioid detoxification and perhaps even exacerbate pain from withdrawal-related conditions. For example, animal research has demonstrated that neuromediator changes underlying neuropathic pain are similar to those observed in opioid tolerance (Mao et al., 1995). This finding raises the question of whether intensity of HIV-related pain could be increased by opioid withdrawal and, thus, promote relapse to heroin use upon discharge. It is, therefore, important to assess the potential impact of opioid detoxification on pain syndromes and to develop standards of care that minimize suffering. This study was designed to evaluate responses to three different opioid detoxification medications (clonidine, methadone, and buprenorphine) on withdrawal and pain measures while HIV-infected patients received treatment for their acute medical condition. Clonidine, an a2-noradrenergic agent, is a presynaptic inhibitor of norepinephrine release in the locus coeruleus with weak analgesic activity at the spinal level (Aghajanian, 1978; Gold et al., 1978, 1980; Guthrie, 1990; Striebel et al., 1993; Bischoff and Kochs, 1993), though the importance of this activity in opioid-dependent human is not known. Methadone, a m-opioid agonist with pharmacological properties qualitatively similar to those of morphine, has been recommended for managing opioid dependence in the hospital setting (Perkins, 1970; Stimmel, 1997). Buprenorphine is a partial m-opioid agonist analgesic that is 25/50 times more potent than morphine (Jaffe and Martin, 1990; Jasinski et al., 1978). All three medications have been tested in clinical trials for opioid detoxification and shown to be effective in suppressing opioid withdrawal in healthy opioid-dependent volunteers seeking treatment (Umbricht et al., 1999; Jasinski et al., 1985; Agren, 1986; Buccafusco, 1992; Cheskin et al., 1994; Nigam et al., 1993; O’Connor et al., 1997; Bearn et al., 1998; Bickel et al., 1988). The outcome of opioid detoxification in HIV-infected individuals hospitalized for an acute medical condition has not been evaluated. We hypothesized that, while all three regimens would be effective among hospitalized HIV-infected persons, buprenorphine would be superior to clonidine and methadone because it has high receptor affinity, a long duration of action, and low intensity of withdrawal symptoms following abrupt discontinuation (Jasinski et al., 1978). Because of its partial agonist
activity, buprenorphine causes little or no cardiorespiratory depression in opioid-tolerant individuals, even at high doses (Walsh et al., 1994; Preston et al., 2001) and, thus, may be the safest in medically compromised patients. The objectives of this study were to compare the effectiveness of these three opioid detoxification methods on withdrawal and pain scores in HIV/AIDS patients during hospitalization for an acute medical condition. A subsidiary objective was to evaluate patterns of medical indication for opiate (morphine) administration during the detoxification and associated treatment outcome.
2. Methods 2.1. Patients and setting Patients were admitted to the inpatient AIDS service of the Johns Hopkins Hospital, either through the emergency department or via the outpatient AIDS clinic. Within 18 h of admission, subjects were screened by questionnaire and urine toxicology for current heroin use and for evidence of physical dependence. Inclusion criteria were HIV seropositivity, age of at least 18, hospitalization for an acute medical illness, and current opioid dependence by self-report and physical examination. Exclusion criteria were concurrent alcohol dependence, inability to give informed consent, acute psychosis or AIDS dementia (measured by the AIDS dementia scale; Power et al., 1995), hypotension, bradycardia, coagulopathy or severe thrombocytopenia precluding intramuscular injections, and enrollment in methadone maintenance treatment. Eligibility was determined in a 45 min screening procedure. All volunteers gave informed written consent prior to participation and were paid $10 for the screening procedure, and $10 a day for the 4-day study. Each patient’s physician was informed about his or her participation in the study. The study was approved by the National Institute on Drug Abuse and the Johns Hopkins University School of Medicine Institutional Review Boards. 2.2. Medications Study medications were administerd double-blind, triple-dummy in tapering doses over 3 days. Buprenorphine (Buprenex 0.3 mg injection, Reckitt & Colman Pharmaceuticals, Inc., Richmond, VA) was administered intramuscularly: 0.6 mg every 4 h on day 1, 0.6 mg every 6 h on day 2, and 0.6 mg every 8 h on day 3. Clonidine (Clonidine HCl 0.1 mg tablet, Schein Pharmaceutical, Inc., Florham Park, NJ) was administered orally as a loading dose of 0.2 mg, followed by 0.1 mg
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
every 4 h on day 1, 0.1 mg every 6 h on day 2, and 0.1 mg every 8 h on day 3. Methadone (Methadone HCl USP, 10 mg tablet, Roxane Laboratories, Inc., Columbus, OH) was administered orally once a day, 30 mg on day 1, 20 mg on day 2, and 10 mg on day 3. Methadone and its placebo were dispensed in blue capsules (size 0), and clonidine and its placebo were dispensed in yellow capsules (size 0). Placebo capsules were filled with dextrose anhydrous granular USP (Amend Drug & Chemical Co., Inc., Irvington, NJ). Equivalent volumes of normal saline given intramuscularly served as the buprenorphine placebo. In addition to these study medications, to prevent the development of opioid withdrawal between the time of admission to the hospital and initiation of the study, attending physicians were instructed to prescribe morphine (recommended dose: 10 mg intramuscularly every 4 h as needed) up to 6 h prior to enrollment. The use of morphine and other opiates, as adjunctive medications for pain during the study intervention, was also recorded as an outcome measure. At the end of the study, starting on day 4, a transdermal clonidine patch (Catapres-TTS-2 patch, Boehringer Mannheim Corp./Roche Diagnostic Systems, Inc., Somerville, NJ) was applied to all patients to minimize protracted withdrawal symptoms. 2.3. Randomization and blinding procedures After patients consented to participate, they were randomly assigned to one of the three treatment groups: clonidine, buprenorphine, or methadone. To obtain comparable treatment groups, patients were stratified on four characteristics: opioid withdrawal severity, presence of pain, cocaine use, and most recent CD4 cell count (above or below 200 mm 3). To maintain the blind, one active medication and two inactive medications were administered to all participants. At each medication time, all patients received a subcutaneous injection (either buprenorphine or a placebo) and capsules (active treatment or placebo) to swallow. As all three treatments have demonstrated effectiveness for decreasing symptoms of opioid withdrawal, the inclusion of a placebo control group (no active medications) was judged to be unethical, especially among a medically ill population. 2.4. Data collection Data were collected by trained interviewers with paper and pencil questionnaires. Baseline measures (T0) were collected prior to administration of the test medications, representing patient status on the first hospital day. Thereafter, as described below, measures were taken 1 /3 times daily depending on the measure.
265
Measures taken once a day were collected in the late afternoon, measures taken twice a day were collected in the morning and late afternoon, and measures taken three times a day were collected at the beginning of each 8 h nursing shift. Patients rated opioid withdrawal symptoms on the short opioid withdrawal scale (SOWS; total score: 0 /30; Gossop, 1990) and three visual analog scales (VASs; scores 0/100 mm). Patients completed SOWS twice a day, rating the severity (not at all, a little, quite a bit, or extremely) of 10 symptoms (feeling sick, stomach cramps, muscle spasms/twitching, feeling of coldness, heart pounding, muscular tension, aches and pains, yawning, runny eyes, and insomnia). Patients completed VAS once a day in the afternoon, rating ‘‘crave heroin’’, ‘‘need heroin’’, and ‘‘heroin sickness’’ on 100 mm lines with the left extremity anchored ‘‘not at all’’ and the other ‘‘most ever’’. Pain was assessed twice a day with Pain Faces cards and with two items from VAS. The Pain Faces cards, a set of eight cards representing faces with a range of emotions from happiness to increasing degree of suffering (score 0 /8; Frank et al., 1982), were presented in random order. Patients selected the face that best represented how they felt. The two VAS items were ‘‘How much pain do you have now?’’ and ‘‘In the last 24 hours, how much pain did you have?’’ Patients completed these as described above. An observer-rated opioid withdrawal scale (OOWS, total score: 0 /30; modified from the Clinical Institute Narcotic Assessment; Peachey and Lei, 1988) was completed by nurses three times daily. The 11 items on OOWS were nausea/vomiting, abdominal complaints/ diarrhea, subjective changes in temperature, myalgia, gooseflesh, sweating, restlessness, tremor, lacrimation, rhinorrhea, and yawning. Pupil diameter was evaluated by monocular infrared television pupillometry (Eye Dynamics, Inc., Torrance, CA). Blood pressure and heart rate were measured by nursing staff at change of shift. Demographic characteristics, most recent CD4 cell count, presumptive admission diagnoses, and morphine and other opiate administration data were collected by chart review. Use of adjunctive opioid medications during the study intervention was ascertained through review of medication administration records. Doses of non-morphine opioid drugs were converted to morphine equivalents (Stimmel, 1997). The time of discharge was decided by the medical team, based on the patient’s medical status. A number of patients were discharged prior to the scheduled end of the study. Because the study was not a primary determinant of discharge, time of discharge was not included as an outcome measure.
266
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
2.5. Statistical analyses Treatment outcome was evaluated as time-averaged data, i.e., by averaging measures during treatment after initiation of the study medications (mean treatment score; Tmean /mean(T1, T2,. . ., TN ), where N is the number of measures taken for that patient, maximum N was 4/12 depending on how frequently study measures were taken. This summary value was used for the following reasons: (1) the study was completed during a patient’s hospital course, which was driven by the medical treatment; (2) the time of discharge was determined by the medical condition and not influenced by withdrawal symptoms and thus the treatment assignment; and (3) due to the competing needs of clinical care, the same measures could not be taken at the same time points for all patients. A cross-sectional data set was created: baseline scores (T0), scores after the first medication dose (T1), mean treatment scores (Tmean), standard deviation of all treatment scores (T1, T2,. . ., TN ), difference between T1 and baseline (T1/T0), and difference between mean treatment score and baseline (Tmean/T0). Increased or higher scores (or changes) indicate more severe withdrawal and lack of therapeutic efficacy while lower scores indicate treatment benefit. The main outcome measures were compared between the treatment groups and analyzed with respect to the following covariates of interest: sex, presence of pain, OOWS ]/5, CD4 cell count 5/200 mm 3, morphine administration (patients who received morphine during the study intervention were compared with those who never received morphine). As needed, other comparisons were made with t-tests and contingency tables (exact tests). Multivariate analyses were performed using linear models. One-sample t-tests were used to test the null hypothesis for the differences T1/T0 and Tmean/T0; 95% confidence intervals (from the t-distribution) were calculated for these differences on OOWS, SOWS, scores on visual analogue scales, and on Pain Faces cards, heart rate, systolic and diastolic blood pressures, and pupil sizes. Adjustment for baseline characteristics (sex, presence of pain, OOWS /5, CD4 cell countB/200 mm 3, or the admission diagnosis ‘‘fever/cellulitis’’) did not statistically alter the results. Therefore, data are presented as unadjusted means. As appropriate, statistical comparisons were made with paired t-tests, two-sample t-tests, one-way ANOVA, exact tests and two-sample t-tests. Only P -values 5/0.05 are reported.
3. Results Sixty-three eligible heroin-dependent patients were enrolled in the study, and 21 were assigned to each treatment group. Data were available for 55 patients:
three patients were dropped out after baseline before receiving any study medication, and five were excluded due to medication errors (e.g., staff administered openlabel buprenorphine instead of ‘‘study (blinded placebo or active) buprenorphine’’). Table 1 shows patients’ characteristics at enrollment. There were no other significant differences in intake characteristics among the groups except that patients randomized to the buprenorphine group were more likely to have been admitted for fever/cellulitis (i.e., 63% versus approximately 20% for the other two groups; P B/0.05). A post hoc analysis controlling for this diagnosis showed that it did not significantly change the results of the outcome measures. Self-reported heroin use histories were available for 43 participants. On average, participants had used heroin for 17.9 (S.D. 8.6; range 4/37) years and had used 13.0 (S.D. 10.7; range 0 /30) days out of the last 30. Twenty four of 55 (43%) participants completed the study; within the treatment groups, 7 of 21 (33.3%) in the buprenorphine group, 8 of 16 (50%) in the clonidine group, and 9 of 18 (50%) in the methadone group completed. Among the 31 who did not complete, 25 (81%) were discharged from hospital, transferred to another unit or dropped out for non-study-related medical reasons prior to completion; two (6%, both in the buprenorphine treatment group) left the hospital against medical advice; one (3%, in the buprenorphine treatment group) chose to discontinue their participation; one (3%, in the buprenorphine treatment group) was dropped from the study due to non-compliance with study procedures; and two (6%, in the clonidine group) dropped out due to side effects. Drop out from the study was not correlated with reason for admission, CD4 count on admission, sex, OOWS, nor treatment group. 3.1. Subject- and observer-rated opioid withdrawal Fig. 1 shows the mean withdrawal scores at baseline (T0), after the first dose of study medications (T1), and during treatment (Tmean) for all patients and by treatment group. Table 2 shows the mean changes in scores between baseline and during treatment (Tmean/T0) for all patients and by group with the 95% CI for the difference. Overall, both subject- and observer-rated opioid withdrawal decreased significantly after initiation of the study medications (P B/0.001 for each group) as indicated by SOWS and OOWS (Fig. 1; Table 2). After a marked first-dose effect, further improvements were minimal, and overall mean scores during treatment (Tmean) were not significantly lower than scores after one treatment dose (T1) (Fig. 1). The overall mean timeaveraged decreases ranged from /5.1 to /6.0 for OOWS and /3.3 to /4.7 for SOWS. Scores on subjective VAS for ‘‘crave heroin’’, ‘‘need heroin’’, and ‘‘heroin sickness’’ indicated more gradual
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
267
Table 1 Patient characteristics at intake Variables
All patients
Buprenorphine
Clonidine
Methadone
N evaluable Years of age, mean9S.D. Race: African /American (%) Sex: male (%) CD4 cell 5 200/mm3 (%) Presence of pain (%)
55 39.795.6 98 62 67 78
21 39.695.4 95 71 67 81
16 40.095.4 100 69 69 81
18 40.095.9 100 44 68 72
Opioid withdrawal OOWS, mean9SE OOWS] 5 (%)
7.590.6 60
7.791.1 62
7.791.1 63
7.390.9 56
Admission diagnosis (%) Fever/cellulitis Pneumonia, PCP Tuberculosis Meningitis Other
36 29 7 7 20
62* 19 10 5 5
19* 38 6 6 31
22* 33 6 11 28
* P B 0.05 compared with other treatment groups.
declines (Fig. 1), possibly because the question referred to ‘‘in the last 24 hours’’ (thus, re-evaluation a few hours after the first medication dose could not elicit the full effect of the medication). Time-averaged mean VAS scores during treatment for these three measures were significantly lower (/14.1 to /23.7 mm) than at baseline (T0 versus Tmean) for the three groups combined. There were, however, no significant differences in the magnitude of the decrease in withdrawal (Tmean/T0) between the treatment groups (Table 2). At no time
during treatment did withdrawal scores exceed baseline scores for any individual (data not shown). 3.2. Physiologic measures There were no major changes in blood pressure or heart rate during the observation time interval of the study, although a relatively small decrease in heart rate reached statistical significance (Table 2). There was a trend (P /0.06) toward myosis (decrease in pupil size)
Fig. 1. Mean subject- and observer-rated opioid withdrawal scores and pupil sizes for the treatment groups combined and separately. The bars represent mean scores at baseline (T0, black), after one medication dose (T1, stripes), and mean treatment scores (Tmean, white). Brackets indicate standard errors.
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
268
Table 2 Mean (95% confidence intervals) of scores during treatment minus baseline scores (TmeanT0): overall and by treatment groupa Variables
Overall
Buprenorphine
Clonidine
Methadone
N OOWS (0 /30) SOWS (0 /30) Crave heroin VAS (mm) Need heroin VAS (mm) Heroin sickness VAS (mm) Heart rate (beats per minute) Systolic blood pressure (mmHg) Pupil size (mm) Pain faces cards (1 /8) Pain now (VAS) Pain in the last 24 h (VAS)
55 5.6 4.7 17.8 14.1 23.7 5.7 2.5 0.3 1.0 20.5 20.3
21 6.0 5.8 17.9 19.6 20.6 6.3 7.3 0.4 0.9 20.9 22.4
16 5.3 4.7 20.5 17.5 25.3 5.7 2.1 0.3 0.9 22.0 20.0
18 5.1 3.3 16.9 5.5 26.7 5.7 2.2 0.7 1.3 19.0 19.0
a
(6.8, 4.3) (6.2, 3.1) (28.6, 6.7) (23.7, 4.5) (31.7, 15.7) (10.0, 1.5) (6.7, 1.6) (0.7, 0.1) (1.5, 0.6) (29.4, 11.7) (27.5, 13.0)
(8.4, 3.6) (8.7, 2.9) (37.4, 1.6) (36.4, 2.7) (35.3, 6.0) (14.8, 2.3) (14.0, 0.7) (1.2, 0.4) (1.7, 0.005) (36.8, 5.1) (33.8, 10.8)
(7.6, 2.9) (8.0, 1.4) (43.5, 2.5) (37.1, 2.2) (44.3, 6.3) (13.5, 2.1) (9.7, 5.5) (0.3, 0.8) (1.8, 0.001) (39.4, 4.5) (36.9, 3.1)
(7.2, 3.1) (4.9, 1.6) (33.8, 0.014) (21.0, 10.0) (36.5, 17.0) (12.6, 1.1) (9.6, 5.3) (1.3, 0.004) (2.0, 0.6) (33.7, 4.2) (30.0, 8.0)
Values in brackets indicate 95% confidence intervals; changes were considered statistically significant where confidence intervals do not include
0.
in the buprenorphine and methadone groups, consistent with a pharmacologic effect of treatment (Table 2; Fig. 1). Systolic blood pressure decreased significantly in the buprenorphine group during treatment but not in other groups (Table 2). Two patients in the clonidine group had to be discontinued from the study because of decreased systolic blood pressure (B/90 mmHg) and bradycardia (HR B/50 BPM). 3.3. Subject-rated pain Scores on the Pain Faces choice and pain VAS are shown in Fig. 2; differences between baseline (T0) and treatment (Tmean) are shown in Table 2. Overall, pain scores decreased during treatment, although more gradually than withdrawal symptoms (Fig. 2). The pain scores were significantly lower during treatment than at baseline overall and for each of the three groups (Table 2). Decreases in Pain Faces scores were statistically significant although the changes were small (/1 unit); decreases in the ‘‘Pain Now’’ and ‘‘Pain in the last 24 hours’’ were somewhat larger (/20 units). There were no trends toward a difference in pain outcome between the three medication groups. At no time during treatment did any individual’s pain score exceed his or her score at intake. 3.4. Use of morphine during the study intervention About 80% of the patients were in pain at baseline, and physicians had the option to administer morphine for pain treatment in addition to opioid detoxification treatment. Because of the possibility that pain would be exacerbated during detoxification, outcome measures were reanalyzed as a function of patterns of morphine administration during the study. Data on administration of morphine as adjunct medication for pain during the research treatment
(detoxification) were available for 53 patients. About half of those patients (24/53) received morphine. Intake characteristics and treatment assignments of patients who received morphine and those who did not are shown in Table 3. At baseline (Fig. 3), patients who received morphine had significantly lower Pain Faces (P /0.003; two-sample t-tests), observed (OOWS) withdrawal scores (P /0.03), and scores on the ‘‘need heroin’’ VAS (P /0.04) (Fig. 3). A statistically significant difference was also found on the base line ‘‘heroin sickness’’ scale between those who received no morphine and those who received some morphine (77.49/6.6 versus 58.59/6.6; P /0.047), respectively. There were no statistically significant differences in the proportion of patients receiving morphine during detoxification or in the amount of morphine administered between the medication groups. There was, however, a tendency for fewer patients in the methadone group to receive smaller doses of morphine. Mean doses (9/S.D.) administered by group during the entire observation period were 5.39/5.6 mg for the buprenorphine group, 4.89/2.7 mg for the clonidine group, and 3.09/2.0 mg for the methadone group. There was evidence of sex differences in morphine administration during detoxification and in self-reports of pain. Females were significantly more likely to receive morphine: 13 out of 21 women (62%) versus 11 out of 32 men (34%) received morphine (P /0.05). This is in spite of the fact that women tended to score lower on pain measures initially (T0), a difference that reached significance for the Pain Face choice: 71% of the men chose a Pain Face card scoring 6 or higher at baseline compared to only 29% of the women; mean scores on Pain Faces were 5.89/0.2 for men versus 4.79/0.3 for women (P B/0.005). Perhaps because of this baseline difference, men had greater decreases in pain scores during treatment. However, sex was found not to be related to changes in pain or withdrawal in multivariate
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
269
Table 3 Intake characteristics and treatment assignment as a function of whether participants received morphine for pain during treatment Variables
No morphine
Morphine
N Years of age, mean9S.D. Race: African /American (%) Sex: male (%) CD4 5 200 cells/mm3 (%) Reporting pain on admission (%) Opioid withdrawal (OOW] 5) (%)
29 39.695.4 100 72* 72 72 69
24 40.095.4 96 46* 58 88 50
Reason for admission (%) Fever, cellulitis, skin ulcer Pneumonia, PCP Tuberculosis Meningitis Other
28 35 14 10 14
50 21 / 4 25
Treatment assignment (%) Buprenorphine Clonidine Methadone
48 50 67
52 50 33
Total N 53: use of morphine unknown for two patients. * P 0.05.
Fig. 2. Mean subject-rated pain scores for the treatment groups combined and separately. The bars represent mean scores at baseline (T0, black), after one medication dose (T1, stripes), and mean treatment scores (Tmean, white). Brackets indicate standard errors.
analyses that included both sex and morphine use as covariates.
4. Discussion The post-treatment improvements in this study suggest that opioid withdrawal was effectively accomplished by treatment with any one of the three regimens during hospitalization for medical reasons. Although relatively high at baseline, opioid withdrawal symptoms and pain ratings decreased during the 4-day study period. This interpretation, however, must be considered in light of the lack of a control group (for ethical reasons) and potential longitudinal reporting
bias. The initiation of pharmacological treatments was associated with a marked first-dose effect, followed by a more gradual decline in discomfort. Reports of adverse effects of treatment were few and observed only in the clonidine group. Two patients in the clonidine group were terminated due to development of bradycardia and hypotension, suggesting that this medication is less safe in some patients. At no time during treatment did withdrawal or pain scores exceed scores at intake for any individual. During treatment for an acute medical condition, buprenorphine, methadone, or clonidine in tapered doses were equally effective and were associated with decreases in opioid withdrawal intensity, with no significant differences among the treatment groups. The sample size in each group may not have been large enough to detect moderate differences between treatment regimens; however, there were no trends in the withdrawal data to suggest that a larger sample would have detected differences between the treatment groups. Combined with medical treatment for the underlying admission diagnosis, the study medications were associated with decreases in pain scores during the 4-day observation period, with no evidence of exacerbation of pain or hyperalgesia. Post-treatment decline in pain was more gradual with a first treatment effect that was less striking than that observed for withdrawal symptoms. As with effects on opioid withdrawal, there were no differences among the three treatment groups on changes in pain scores. However, fewer patients in the methadone group received adjunctive morphine, and these patients received smaller doses than patients in other groups. Although many factors, in addition to
270
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
Fig. 3. Pain and withdrawal scores as a function of whether participants received morphine for pain during treatment. The bars represent mean scores at baseline (T0, black), after one medication dose (T1, stripes), and mean treatment scores (Tmean, white). Brackets indicate standard errors.
pain severity, such as staff attitudes and the patient’s behavior, determine whether a patient receives morphine, it is possible that methadone (30 mg) may have been a more effective analgesic than the corresponding split doses of buprenorphine. Our results differ from those of a 10-day detoxification study conducted among non-AIDS patients residing on a drug abuse research ward, where, in the first 3 days, buprenorphine was found to be more effective than clonidine for decreasing opioid withdrawal symptoms (Cheskin et al., 1994). The doses of buprenorphine used in that study were 2 mg given sublingually every 6 h for 48 h, and the doses of clonidine were 1.7 mg given orally in the first 48 h. With a bioavailability estimated to be approximately 50% (Mendelson et al., 1997), 2 mg sublingual buprenorphine solution is equivalent to 1 mg parenteral dose. Our study used approximately half the doses of both medications used by Cheskin. The conservative dosage regimen used in this study was based on our concerns for potentially cachectic or unstable patients in this randomized, double-blind trial. In treating patients on an unblinded basis, the efficacy of buprenorphine might be increased by the use of higher doses, on the model of the study by Cheskin et al. (1994), or by repeating the first doses at short intervals (every 30 min) until resolution of symptoms prior to initiating taper (based on clinical experience of one of the investigators, AU). Given the effects of clonidine on cardiovascular function, increasing clonidine doses in medically ill patients might not be safe. The starting dose of 30 mg of
methadone used in the study compares well with recommendations by Stimmel (1997) of a starting dose of 15 /20 mg, to be adjusted by an additional 10 mg in the following 4 /12 h according to symptoms. From the design of the study, and given the relatively large first treatment effect, it is not possible to exclude a beneficial effect of systematic, non-judgmental, patient interviews about withdrawal experience, in the absence of a placebo group. Inclusion of a placebo group would not have been ethical given the established efficacy of these medications for decreasing withdrawal symptoms in non-HIV-infected drug users and denial of any withdrawal treatment might have added to the burden of HIV patients’ illnesses. Morphine administration may be an indicator of relative treatment ineffectiveness. In opioid-tolerant individuals, when parenteral opiates are indicated to treat moderate to severe pain, administration of larger morphine doses at more frequent intervals than in opioid naive patients (e.g., 10 /20 mg intramuscularly or subcutaneously) has been recommended (Stimmel, 1997). In spite of these instructions to physicians, morphine was administered in small single doses, as witnessed by the low average morphine doses received during the entire study (even after patients who did not receive opioid medications were excluded from the denominator). There was no evidence that the adjunctive administration of morphine decreased withdrawal (Fig. 3). Interesting sex differences in reported pain and in morphine use were observed, though their significance is
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
unclear. Men tended to rate their pain higher at baseline and to show larger decreases after treatment was started, after which the sex difference in pain was no longer present. Women, nevertheless, were more likely to receive some morphine during the observation period. Our study shows that the majority of heroin-dependent, HIV-infected patients had symptoms of opioid withdrawal during the first day of hospitalization, as evidenced by high scores on objective and self-reported withdrawal scales in 60% of patients, and that most (80%) were in pain. Instructions had been given to the hospital staff to stabilize patients with morphine to prevent interference with patient care and comfort by the study protocol. It is possible that opioid withdrawal and pain were not recognized and were, thus, not addressed. Conducting an opioid detoxification study on a busy medical ward was a challenge. The intention was initially to enroll a higher sample size (81 total), but unexpected difficulties in patient recruitment occurred. Only one-half of the patients who had acquired HIV through injection drug use were still using heroin and, were opioid-dependent, a much smaller proportions than expected. About 10% of patients with a history of heroin use were enrolled in methadone maintenance programs and thus ineligible for this study. Furthermore, after 2 years of conducting the study, it became apparent that the population from which we were recruiting was a relatively closed cohort with the same patients being readmitted several times making it difficult to acquire new patients. The exclusion of medically and mentally unstable patients further limited the pool of eligible patients. The resulting small sample size may have decreased our ability to detect differences in outcome among the medications. Our patient selection criteria may also have limited the generalizability of the study. Nevertheless, we believe that the study showed that opioid-dependent patients were initially experiencing opioid withdrawal and pain and that treatment with modest doses of medication following patient assessment effectively decreased these discomforts without adverse effects or increases in pain.
Acknowledgements The study was funded by the National Institute on Drug Abuse, Intramural Research Program. We thank Marcia Hoffman, Betsey Emilien, Michelle Dequina, and Neena Thomas for their expertise in data collection and data management, and Dr. David Epstein for his constructive review of the manuscript. We also thank the nursing staff of the Johns Hopkins Inpatient AIDS Service for their cooperation and unconditional support of the study, and the Osler Medical Service for physician support.
271
References Aghajanian, G.K., 1978. Tolerance of locus coeruleus neurons to morphine and suppression of withdrawal response by clonidine. Nature 276, 186 /188. Agren, H.M., 1986. Clonidine treatment of the opiate withdrawal syndrome. Acta Psychiatr. Scand. 327, 91 /113. Bearn, J., Gossop, M., Strang, J., 1998. Accelerated lofexidine treatment regimen compared with conventional lofexidine and methadone treatment for inpatient opiate detoxification. Drug Alcohol Depend. 50, 227 /232. Bickel, W.K., Stitzer, M.L., Bigelow, G.E., Liebson, I.A., Jasinski, D.R., Johnson, R.E., 1988. A clinical trial of buprenorphine: comparison with methadone in the detoxification of heroin addicts. Clin. Pharmacol. Ther. 43, 72 /78. Bischoff, P., Kochs, E., 1993. Alpha2 agonists in anaesthesia and intensive care medicine. Anaesthesiol. Intensivmed. Notfallmed. Schmerzther. 28, 2 /12. Breitbart, W., McDonald, M.V., Rosenfeld, B., Passik, S.D., Hewitt, D., Thaler, H., Portenoy, R.K., 1996. Pain in ambulatory AIDS patients. I. Pain characteristics and medical correlates. Pain 68, 315 /321. Breitbart, W., Rosenfeld, B., Passik, S., Kaim, M., Funesti-Esch, J., Stein, K., 1997. A comparison of pain report and adequacy of analgesic therapy in ambulatory AIDS patients with and without a history of substance abuse. Pain 72, 235 /243. Buccafusco, J.J., 1992. Neuropharmacologic and behavioral actions of clonidine: interactions with central neurotransmitters. Int. Rev. Neurobiol. 33, 55 /107. Center for Disease Control, 1997, 1996. Update: trends in AIDS incidence, deaths, and prevalence */United States. MMWR 46, 165 /173. Cheskin, L.J., Fudala, P.J., Johnson, R.E., 1994. A controlled comparison of buprenorphine and clonidine for acute detoxification from opioids. Drug Alcohol Depend. 36, 115 /121. Frank, A., Moll, A., Hort, J., 1982. A comparison of three ways of measuring pain. Fheumatol. Rehabil. 21, 211 /217. Gold, M.S., Redmond, D.E., Kleber, H.D., 1978. Clonidine blocks acute opiate-withdrawal symptoms. Lancet, 2, 599 /601. Gold, M.S., Pottash, A.C., Sweeney, D., Kleber, H.D., 1980. Opiate withdrawal using clonidine: a safe, effective, and rapid nonopiate treatment. J. Am. Med. Assoc. 243, 343 /346. Gossop, M., 1990. The development of a short opiate withdrawal scale (SOWS). Addict. Behav. 15, 487 /490. Guthrie, S.K., 1990. Pharmacologic interventions for the treatment of opioid dependence and withdrawal. Ann. Pharmacother. 24, 721 / 734. Jaffe, J., Martin, W., 1990. Opioid analgesics and antagonists. In: Goodman, T.R.A., Gilman, Nies, A.S., Taylor, P. (Eds.), Goodman and Gilman’s The pharmacological Basis of Therapeutics. Pergamon Press, New York, pp. 485 /521 (Chapter 21). Jasinski, D.R., Pevnick, J.S., Griffith, J.D., 1978. Human pharmacology and abuse potential of the analgesic buprenorphine. Arch. Gen. Psychiatr. 35, 501 /516. Jasinski, D.R., Johnson, R.E., Kocher, T.R., 1985. Clonidine in morphine withdrawal. Arch. Gen. Psychiatr. 42, 1063 /1066. Karon, J.M., Fleming, P.L., Steketee, R.W., De Cock, K.M., 2001. HIV in the United States at the turn of the century: an epidemic in transition. Am. J. Public Hlth. 91, 1060 /1068. Mao, J., Price, D.D., Mayer, D.J., 1995. Mechanisms of hyperalgesia and morphine tolerance: a current view of their possible interactions. Pain 62, 259 /274. Mendelson, J., Upton, R.A., Everhart, E.T., Jacob, P., Jones, R.T., 1997. Bioavailability of sublingual buprenorphine. J. Clin. Pharmacol. 37, 31 /37.
272
A. Umbricht et al. / Drug and Alcohol Dependence 69 (2003) 263 /272
Nigam, A.K., Ray, R., Tripathi, B.M., 1993. Buprenorphine in opiate withdrawal: a comparison with clonidine. J. Subst. Abuse Treat. 10, 391 /394. O’Connor, P.G., Carroll, K.M., Shi, J.M., Schottenfeld, R.S., Kosten, T.R., Rounsaville, B.J., 1997. Three methods of opioid detoxification in a primary care setting: a randomized trial. Ann. Intern. Med. 127, 526 /530. Peachey, J., Lei, H., 1988. Assessment of opioid dependence with naloxone. Br. J. Addict. 83, 193 /201. Perkins, I.H., 1970. Hospital treatment of narcotic addiction. Clin. Toxicol. 3, 571 /578. Power, C., Selnes, O.A., Grim, J.A., McArthur, J.C., 1995. HIV dementia scale: a rapid screening test. J. AIDS Hum. Retrovirol. 8, 273 /278. Preston, K.L., Umbricht, A., Huestis, M.A., Cone, E.J., 2001. Intravenous buprenorphine: effects of 2 to 16 mg doses in nondependent opioid abusers. Clin. Pharmacol. Ther. 69, 29.
Rosenfeld, B., Breitbart, W., McDonald, M.V., Passik, S.D., Thaler, H., Portenoy, R.K., 1996. Pain in ambulatory AIDS patients. II. Impact of pain on psychological functioning and quality of life. Pain 68, 323 /328. Stimmel, B., 1997. Pain and its Relief without Addiction: Clinical Issues in the Use of Opioids and other Analgesics. The Haworth Medical Press, New York. Striebel, H.W., Koenigs, D., Heil, T., 1993. The role of clonidine in clinical anaesthesiology. Anaestesist 42, 131 /141. Umbricht, A., Montoya, I.D., Hoover, D.R., Demuth, K.L., Chiang, C.-T., Preston, K.L., 1999. Naltrexone shortened opioid detoxification with buprenorphine. Drug Alcohol Depend. 57, 181 / 190. Walsh, S.L., Preston, K.L., Stitzer, M.L., Cone, E.J., Bigelow, G.E., 1994. Clinical pharmacology of buprenorphine: ceiling effects at high doses. Clin. Pharmacol. Ther. 55, 569 /580.