Abstract
Keywords
In methadone maintenance treatment (MMT) clinics the prevalence of cannabis abuse has been observed to be widespread, and ranges from 16–78.5%% [1–6]. Little is known about the correlates of psychosocial, medical, and drug abuse history [6],[7] and the possible contribution of cannabis abuse to dropout rates from MMT programs [8]. It has also not been well established whether cannabis abuse in MMT is related to heroin abuse [5] or even methadone dosage [9]. Moreover, the question of whether there is a reduction or increase in cannabis abuse over time in MMT has not yet been resolved: Fairbank et al. [9] observed a reduction in marijuana use over time in MMT, while Bell et al. [1] found an increase in cannabis use over time.
A large number of other questions related to cannabis abuse in MMT patients have not been addressed. Are cannabis abusers in MMT single drug abusers or do they abuse multiple drugs? Is psychopathology and psychological distress related to its abuse, for example is emotional distress causal in the abuse of cannabis in MMT patients as the self-medication hypothesis would suggest? Are cannabis abusers (CAs) more vulnerable to risk behaviour (do they inject more and/or have more unprotected sex) leading to more infectiodiseases, such as HIV and HCV?
Answers to these questions could shed some light upon the theoretical question whether cannabis abuse is specifically related to the self-medication for the alleviation of emotional distress in MMT patients, whether cannabis abusers are part of a high risk group for infectious diseases and whether the abuse of cannabinoids should be considered a specific treatment focus in MMT, as cocaine and benzodiazepine abuse often are. Answers to these questions have therapeutic and policy implications which should be taken into consideration in MMT.
This study attempts to tentatively answer these questions by analyzing the pattern of cannabis abuse in an MMT clinic over time and considers whether CAs differ from non-cannabis abusers (NCAs) on a number of selected variables considered as being relevant to these questions.
Method
Subjects
The study was carried out at an MMT outpatient clinic located in Tel Aviv, Israel. The clinic, its policy, demographic composition and patterns of abuse of its patients have been described elsewhere [10–14]. Cannabis abuse, up to this study, has been considered as an offence that can bring about ‘not receiving’ privileges (take homes) but cannot lead to treatment termination.
Since the clinic opened, 283 patients have been registered at the clinic; of them, 260 were first admissions. There were 196 patients who remained in treatment for at least 1 year, of whom 176 were first admissions. Table 1 describes the number of patients who responded to the different measures.
Number of patients for the study measures
Measures
Within 1 month of entering treatment, patients are administered a Hebrew version of the Addiction Severity Index (ASI) regarding demographic and psycho-social information [15]. The questions relevant for the current study were gender, age, marital status, years of education, whether the patient had served time in prison, age of onset of abuse of drugs, frequency of abuse of drugs, suicide attempts and suicidal ideation.
All patients usually underwent two randomized observed urine tests per week for the entire length of their treatment. These samples are analyzed for methadone, morphine and codeine (e.g. also heroin), cannabis, benzodiazepine, cocaine and amphetamines using the ‘EMIT’ method [16] in an external laboratory. For the measurement and quantification of cannabis abuse we used two approaches: the first assessed the percentage of tests positive for a given month (the first month and the 12th month of treatment); the second considered that if a patient tested positive for cannabis for any consecutive 3 months during the first year of MMT he was considered a potential CA. All patients considered CAs were administered the Structural Clinical Interview for DSM-IV Axis I Disorders (SCID 1) for cannabis substance abuse (part E 22–23) so as to confirm or disconfirm CA status. This standard was adopted so as to avoid the over-inclusion of occasional users, testing errors and pretreatment use. For 8/71 (11.3%%) patients who had three consecutive months of positive urine for cannabis, the SCID I did not confirm CA status.
Maximum daily methadone dosage at month 12 of treatment was recorded.
Treatment tenure was calculated based upon the overall number of days patients remained in treatment.
The SCID I [17] provides information for making lifetime and current diagnoses; the SCID Axis II provides for diagnoses to assess personality disorders. Studies have shown satisfying reliability and validity for both SCIDs [18]. Patients were interviewed approximately 2 months after they entered treatment by a trained psychiatrist (TW).
The Symptom Check List 90-Revised (SCL-90-R) is a widely used 9-factor self-report inventory designed to reflect psychological symptom patterns. Patients are asked to assess symptoms on a 5-point Likert scale, ranging from 0 = ‘not at all’ to 4 = ‘extremely’. Alpha coefficient ranges from 0.77 to 0.90, and test-retest values are reported to range from 0.78 to 0.90 [19]. All patients responded to the SCL- 90-R questionnaire between 6 and 10 months into treatment.
Patients were queried after 1 year of treatment about HIV/HCV risktaking behaviour using our standard clinic questionnaire relating to whether the patient injected drugs, shared needles, performed safe sex, had sex for drugs and had a partner who abused drugs during the past year (see Bleich et al. [11] for a description of the questionnaire). Two weeks’ test-retest correlations for the items ranged from 0.90–1.00 on a random sample of 30 patients.
All patients are tested routinely for hepatitis C and HIV.
Results
Demographics
From the 176 patients who were administered the ASI and remained in treatment for at least a year, 63 were CAs and 113 NCAs. All following analyses are presented after the partialling out of intervening variables. Cannabis abusers were found to be younger (CAs = 35.7 yeras, NCAs 38.3 years; t = 2.34, p = 0.02) and more often single (CAs = 40/63; NCAs = 46/113; χ2 = 8.1, p = 0.004). Patients of Sepharadic (African/Asian) descent were more often NCAs (77/99) while patients of Ashkenaze (European/American) descent were more or less evenly divided between CAs and NCAs (26/56 NCAs; χ2 = 14.1, p = 0.001). More CAs than NCAs were Russian (16/63 vs. 12/113), or of European/American birth (5/63 vs. 5/113), as well as Palestinian (6/63 vs. 5/113) than native Israelis (33/63 vs. 83/113; χ2 = 11.2, df = 4, p = 0.024). No significant differences were found between CAs and NCAs in terms of demographic or history of abuse data.
Lifetime and current cannabis abuse
Information on lifetime and current cannabis abuse was gained from ASI interviews at treatment onset.
Sixty-five (25%%) of the 260 entering patients reported having abused cannabis regularly before entering the clinic, and 194 (74.6%%) stated that they abused cannabis at some point in their life.
Current prevalence
Information on current prevalence and course of cannabis abuse during MMT of patients 1 year in treatment was gained from urine tests.
Of the 196 patients who remained in treatment at least 1 year, 102 (52%%) had at least one urine test positive for cannabis during that year. A comparison of the number of patients using cannabis at 1 month with those using at 1 year revealed no significant increase (McNamar test for symmetry; p = 0.23). Specifically, 27/196 (13.8%%) used cannabis during their first month in treatment and 36/196 (18.4%%) during their 12th month of treatment. Of the 27 who used cannabis at treatment onset, 18 (66.6%%) no longer used it at 12 months, and of the 168 who did not use it at treatment onset, 27 (16.1%%) used it at 12 months into treatment. Of these 196 patients, 68 (34.7%%) used cannabis for at least three successive months during that year and 63 of these (32.1%%) were considered CAs following the SCID I criteria.
Treatment tenure
Non-cannabis abusers were not found to differ in the time they remained in treatment than CAs (Cox regression survival analysis: B = −0.17; SE = 0.13; Wald = 1.57; df = 1; p = 0.21; r = 0.00; Exp(B) = 0.84). Further analysis with heroin, cocaine and benzodiazepine abuse as covariants did not change the results significantly.
Drug use and methadone dosage
Information on drug use and methadone dosage of patients at least 1 year in MMT was gained from urine tests.
ANOVA with CA/NCA as the fixed factor and heroin, BZD, amphetamine and cocaine abuse at 12 months in treatment as dependent variables showed CAs to abuse more BZDs (F = 18.48, p = 0.000), amphetamines (F = 9.29, p = 0.003) and cocaine (F = 4.06, p = 0.045). Cannabis abusers were not found to abuse more heroin than NCAs.
A t-test of CA/NCA with the number of distinct different drugs abused at 12 months into treatment showed CAs to abuse a significantly greater range of drugs (1.6 vs. 0.79) than NCAs (t = 5.63; df = 194, p = 0.000)
ANCOVA of CA/NCA as fixed factor and methadone dosage as dependent variable and BZD abuse at 12 months in treatment as covariate (BZD abuse was previously found to be correlated to methadone dosage) did not find any relationship between CA/NCA and methadone dosage (F = 0.029, p = NS).
Infectious diseases
ANOVA of CA/NCA as fixed factor and HCV/HIV risk behaviour as dependent variable did not find any relation between CAs and risk behaviour.
A χ2 analysis showed CAs to have significantly more hepatitis C than NCAs.
So as to further analyze these results an ANOVA procedure was done with heroin, BZD, amphetamine and cocaine abuse at 12 months in treatment as dependents that showed hepatitis C (as a fixed factor) to be related to BZD abuse. In order to rule out the influence of BZD abuse on the relationship between hepatitis C and CA/NCA, we analyzed patients with no BZD abuse at month 12 after treatment onset. The significant relationship between HCV and cannabis abuse disappeared for these patients (n = 99; χ2 = 1.42, df = 1; p = 0.33). Cannabis abusers were thus not specifically found to engage in more risk behaviour than NCAs nor were they found to have more hepatitis C.
Psychological distress
A significant Pearson's correlation was found between BZD abuse and methadone dosage with the SCL-90-R global score. Therefore an ANCOVA with NCA/CA as fixed factor and the SCL-90-R global score as dependent variable and BZD abuse at 12 months into treatment and methadone dosage at 12 months into treatment as covariants was performed. No difference was found between CAs and NCAs on the SCL-90-R (F = 0.14, p = NS). Similarly, there was no difference between them in the number of patients who tried to commit suicide (χ2 = 0.83, p = NS) or had suicidal ideations (χ2 = 0.067, p = NS).
Psychopathology (Axis I and II)
No differences were found comparing CA/NCA with the presence of an Axis 1 lifetime diagnosis (χ2 = 0.23, p = NS), a current diagnosis (χ2 = 0.08, p = NS), the presence of a mood disorder (χ2 = 0.08, p = NS), an anxiety disorder (χ2 = 0.37, p = NS), a psychotic disorder (χ2 = 0.001, p = NS), any personality disorder (χ2 = 0.05, p = NS) or the presence of an antisocial personality disorder (χ2 = 0.17, p = NS). Results are presented in Table 2.
χ2 comparison of cannabis abusers and non-cannabis abusers on diagnoses of psychopathology and personality disorders (n = 115)
Furthermore CA/NCA was not found to be related to any specific current diagnosis of abuse or dependence of either sedatives, hallucinogens, cocaine or stimulants, but was found to be significantly related to the sum of dependency and abuse diagnoses other than cannabis or opiates (F = 7.5, p = 0.007).
Discussion
Lifetime prevalence was found in nearly 75%% of our patients and current abuse at treatment inception in 25%%. Our results showed that cannabis in our MMT clinic is more often abused by the single and younger patient of American or European descent. They did not have a differential history of drug abuse, did not differ in Axis I or Axis II psychopathology, level of psychological distress, risk-taking behaviour or frequency of infectious disease compared to NCAs. Cannabis abusers were found to abuse more benzodiazepines, cocaine and amphetamines, and generally abused a greater range of drugs of abuse. Furthermore they were more often diagnosed as abusing, or being dependent upon, different drugs than NCAs. These findings are not entirely unexpected, and are similar to other studies on CAs in MMT, finding little difference between the population of CAs and NCAs [2],[5],[6],[20].
More than one-half of the patients who remained in treatment at least 1 year used cannabis at least once during that year. We found no significant increase in use over a 1-year period. These results are unlike the Otago methadone program outcomes which reported an almost 50%% reduction in cannabis daily use among MMT patients [21] and other studies which observed decreases in other secondary drugs of abuse such as cocaine and benzodiazepines (BZD) [9],[11],[12] over time in MMT. The pattern of cannabis abuse in MMT is different than that of cocaine or BZDs. From discussions with patients and staff it seems that contrary to BZDs and cocaine [9], cannabis is perceived not so much as a self-medication drug or something that patients feel dependent upon, but more as a social and recreational drug. Furthermore, some patients report that the combination of methadone and cannabis seems to achieve a better level of ‘relaxation’ than methadone alone.
The fact that CAs use a greater range of drugs and abuse them more often than NCAs suggests that despite the success of methadone in reducing heroin abuse, these patients continue a drug abusing life-style. It is important to note that the abuse of cannabis in these patients is probably not due to psychopathology or psychological distress insofar as they did not differ from the NCAs on the SCL-90-R but is more a social phenomenon than a psychological or medical one. The fact that we found cultural differences in the frequency of use supports this.
We conclude that cannabis abuse is not a specific health-related risk-factor for MMT patients, nor does it affect abuse of heroin and retention rate. We also conclude that it is not related to psychological selfmedication or psychopathology or even personality disorder, but that it is probably related to social– recreational aspects of a drug abuse life-style.
Questions such as whether the abuse of cannabis should be ignored or focused upon during MMT, what should be done to patients abusing cannabis and how this should be handled within the therapeutic context, should be reviewed in the light of these results.
Footnotes
Acknowledgements
This study was supported by the Souraski Medical Center Research Fund. We thank Yaron Yagil for statistical advice and Esther Eshkol for editorial assistance.
The first and second authors contributed equally to this study.