Introduction
Cognitive deficits have been described as a core feature of schizophrenia (Elvevåg & Goldberg, Reference Elvevåg and Goldberg2000; Wobrock et al. Reference Wobrock, Ecker, Scherk, Schneider-Axmann, Falkai and Gruber2008). These deficits are observed across a wide range of neuropsychological tests that assess a range of cognitive functions, including attention, working memory and executive functioning (Heinrichs & Zakzanis, Reference Heinrichs and Zakzanis1998). It has been suggested that cannabis use in healthy individuals can produce cognitive impairment which resembles that which is evident in schizophrenia (Solowij & Michie, Reference Solowij and Michie2007). This is particularly interesting given the hypothesized association between cannabis use and schizophrenia (Degenhardt et al. Reference Degenhardt, Hall and Lynskey2003b; Degenhardt & Hall, Reference Degenhardt and Hall2006).
Despite similarities in cognitive dysfunction being observed between patients with schizophrenia and healthy cannabis users, particularly with respect to episodic memory, the effects of cannabis use on a number of key tasks and processes that are found to be deficient in schizophrenia are still inconclusive. For instance, poor performance on the Stroop task has been demonstrated numerous times in patients with schizophrenia (for a review, see Henik & Salo, Reference Henik and Salo2004) and deficits in attentional control have long been regarded as a fundamental aspect of the cognitive disturbances in schizophrenia (McGhie & Chapman, Reference McGhie and Chapman1961; Andreasen, Reference Andreasen1994). However, the effects of cannabis use on Stroop performance in healthy individuals are inconsistent. While some studies have found poor performance on the interference condition of the Stroop with acute cannabis intoxication (Hooker & Jones, Reference Hooker and Jones1987; Henquet et al. Reference Henquet, Rosa, Krabbendam, Papiol, Fananas, Drukker, Ramaekers and Van Os2006), or with heavy chronic use (Pope & Yurgelun-Todd, Reference Pope and Yurgelun-Todd1996), others have not (Miller et al. Reference Miller, Drew and Kiplinger1972; Pope et al. Reference Pope, Gruber, Hudson, Huestis and Yurgelun-Todd2001; Solowij et al. Reference Solowij, Stephens, Roffman, Babor, Kadden, Miller, Christiansen, McRee and Vendetti2002; Eldreth et al. Reference Eldreth, Matochik, Cadet and Bolla2004; Gruber & Yurgelun-Todd, Reference Gruber and Yurgelun-Todd2005). Similarly, a deficit in working memory, in both the visual and spatial domains, as measured across various tasks, has been well documented and is also thought to be a key feature of schizophrenia (Gold et al. Reference Gold, Carpenter, Randolph, Goldberg and Weingerger1997; Perry et al. Reference Perry, Heaton, Potterat, Roebuck, Minassian and Braff2001; Chey et al. Reference Chey, Lee, Kim, Kwon and Shin2002; Silver et al. Reference Silver, Feldman, Bilker and Gur2003; Lee & Park, Reference Lee and Park2005; Pirkola et al. Reference Pirkola, Tuulio-Henriksson, Glahn, Kieseppa, Haukka, Kaprio, Lonnqvist and Cannon2005; Tan et al. Reference Tan, Sust, Buckholtz, Mattay, Meyer-Lindenberg, Egan, Weinberger and Callicott2006; Twamley et al. Reference Twamley, Palmer, Jeste, Taylor and Heaton2006). However, the effects of cannabis use on working memory in healthy individuals are variable. Acute cannabis administration has been shown to impair spatial working memory in one study (D'Souza et al. Reference D'Souza, Perry, MacDougall, Ammerman, Cooper, Wu, Braley, Gueorguieva and Krystal2004), improve spatial working memory, but only in females, in another study (Makela et al. Reference Makela, Wakeley, Gijsman, Robson, Bhagwagar and Rogers2006), and have no effect in another still (Curran et al. Reference Curran, Brignell, Fletcher, Middleton and Henry2002) while chronic cannabis use has been found to have no effect on working memory in some studies (Solowij et al. Reference Solowij, Stephens, Roffman, Babor, Kadden, Miller, Christiansen, McRee and Vendetti2002; Kanayama et al. Reference Kanayama, Rogowska, Pope, Gruber and Yurgelun-Todd2004; Jager et al. Reference Jager, Kahn, Van Den Brink, Van Ree and Ramsey2006), and to impair it in others (Wadsworth et al. Reference Wadsworth, Moss, Simpson and Smith2006; Harvey et al. Reference Harvey, Sellan, Porter and Frampton2007). Lastly, a deficit in executive function, as assessed with the Wisconsin Card Sort Test (WCST), has been demonstrated numerous times in patients with schizophrenia, particularly in terms of reduced categories achieved and increased perseverative errors (e.g. Park, Reference Park1997; Heinrichs & Zakzanis, Reference Heinrichs and Zakzanis1998; Laws, Reference Laws1999; Everett et al. Reference Everett, Lavoie, Gagnon and Gosselin2001; Hartman et al. Reference Hartman, Steketee, Silva, Lanning and Andersson2003; Li, Reference Li2004; El Hamaoui et al. Reference El Hamaoui, Elyazaji, Yaalaoui, Rachidi, Saoud, d'Amato, Moussaoui, Dalery and Battas2006). However, the effects of cannabis use on WCST performance in healthy individuals have been inconsistent; some studies have found heavy cannabis use to be associated with fewer categories achieved (Bolla et al. Reference Bolla, Brown, Eldreth, Tate and Cadet2002) and more perseverative errors (Pope & Yurgelun-Todd, Reference Pope and Yurgelun-Todd1996), others have not (Pope et al. Reference Pope, Gruber, Hudson, Huestis and Yurgelun-Todd2001; Solowij et al. Reference Solowij, Stephens, Roffman, Babor, Kadden, Miller, Christiansen, McRee and Vendetti2002).
Given that there may be similarities in cognitive performance between patients with schizophrenia and healthy individuals who use cannabis, it may be hypothesized that patients with schizophrenia who use cannabis would show even further decrements in performance of these cognitive processes. However, the literature examining neuropsychological performance in patients with schizophrenia who use cannabis is mixed. For instance, decision-making performance has been shown to be impaired in cannabis-using patients in one study (Mata et al. Reference Mata, Rodriguez-Sanchez, Pelayo-Teran, Perez-Iglesias, Gonzalez-Blanch, Ramirez-Bonilla, Martinez-Garcia, Vazquez-Barquero and Crespo-Facorro2008), while another found no such difference (Sevy et al. Reference Sevy, Burdick, Visweswaraiah, Abdelmessih, Lukin, Yechiam and Bechara2007). A number of studies have found largely no significant differences in performance of a range of cognitive tasks between cannabis-using patients and non-users (Jockers-Scherubl et al. Reference Jockers-Scherubl, Wolf, Radzei, Schlattmann, Rentzsch, Gomez-Carrillo de Castro and Kuhl2007; Sevy et al. Reference Sevy, Burdick, Visweswaraiah, Abdelmessih, Lukin, Yechiam and Bechara2007; Mata et al. Reference Mata, Rodriguez-Sanchez, Pelayo-Teran, Perez-Iglesias, Gonzalez-Blanch, Ramirez-Bonilla, Martinez-Garcia, Vazquez-Barquero and Crespo-Facorro2008) or between substance-using patients (primarily cannabis) and non-users (Addington & Addington, Reference Addington and Addington1997; Pencer & Addington, Reference Pencer and Addington2003; Thoma et al. Reference Thoma, Wiebel and Daum2007; Wobrock et al. Reference Wobrock, Ecker, Scherk, Schneider-Axmann, Falkai and Gruber2008), while other studies have found improved cognitive performance in cannabis-using patients relative to non-users (Joyal et al. Reference Joyal, Halle, Lapierre and Hodgins2003; Stirling et al. Reference Stirling, Lewis, Hopkins and White2005; Coulston et al. Reference Coulston, Perdices and Tennant2007; Potvin et al. Reference Potvin, Joyal, Pelletier and Stip2008; Schnell et al. Reference Schnell, Dagmar, Daumann and Gouzoulis-Mayfrank2009). However, the improvements in cognition that have been reported are not consistent across cognitive domains, or between studies.
There are a number of possible reasons for this variability, including methodological differences in terms of type of cannabis effect examined (acute versus residual) and definition of cannabis use (lifetime cannabis misuse diagnosis, or definitions based on recency or frequency of use). Further, many of these studies fail to screen, or control, for confounds associated with age, education, antipsychotic medications, age of onset of illness and of cannabis use, and, importantly, use of other substances including alcohol, caffeine, nicotine and illicit substances. Additionally, many of these studies have a very small samples size on which comparisons are based, some do not include a healthy control group for comparison, and many fail to statistically correct for multiple comparisons. Finally, inclusion of both healthy individuals and patients with schizophrenia (cannabis users and non-users) in the one study is also of interest in order to examine any similarities or differences in performance between healthy people who use cannabis and patients with schizophrenia who do not, given the hypothesized association between cannabis use and schizophrenia, and also to examine any additive or interactive effects between the factors of cannabis use and schizophrenia.
Therefore, the present study aimed to address many of these issues by examining the effects of cannabis use on neuropsychological performance indices that have been well established to be deficient in schizophrenia (attentional control, as assessed by the Stroop task; working memory, as assessed by the letter–number sequencing (LNS) and spatial span (SS) subtests of the Wechsler Memory Scale (WMS); and executive functioning, as assessed by the WCST), in both healthy individuals, and patients with schizophrenia.
Method
Participants
The study included four groups: (1) 50 patients with a diagnosis of schizophrenia (n=48) or schizo-affective disorder (n=2) who were not current users of illicit substances (non-using patients with schizophrenia; SZN); (2) 22 patients with a diagnosis of schizophrenia (n=21) or schizo-affective disorder (n=1) who were current users of cannabis (cannabis-using patients with schizophrenia; SZC); (3) 38 healthy controls who were not current users of illicit substances (non-using controls; CN); and (4) 36 healthy controls who were current users of cannabis (cannabis-using controls; CC). Healthy controls (both CN and CC) were recruited from the general community through the use of advertisements in local media, and from a database of potential willing volunteers at the research centre. Patients (both SZN and SZC) were in-patients and out-patients of the major psychiatric hospital in Perth (Australia), and were recruited via direct contact from a researcher (K.E.S.). Prior to inclusion in the study, each patient's treating psychiatrist was contacted, with the patient's permission, to ascertain the patient's ability to provide informed consent. All participants were screened prior to inclusion in the study and exclusionary criteria included: self-reported presence of any hearing disorders; any neurological disorders or head injury; or loss of consciousness for over 15 min. In addition, both CN and CC participants were excluded if they had any past or present diagnosis of psychiatric illness or current use of psychiatric medications (such as antidepressants), or if they reported having a first-degree relative with a diagnosis of schizophrenia or schizo-affective disorder, as healthy relatives of patients with schizophrenia may show poorer cognitive performance than healthy non-relatives (Egan et al. Reference Egan, Goldberg, Kolachana, Callicott, Mazzanti, Straub, Goldman and Weinberger2001; Sitskoorn et al. Reference Sitskoorn, Aleman, Ebisch, Appels and Kahn2004). Further, SZN and CN participants were excluded if they had current or past treatment for a substance-use disorder, if they currently used any illicit substances, or if they had used any illicit substance more than once in the previous 12 months.
All healthy control participants (CN and CC) were administered the Mini International Neuropsychiatric Interview (MINI; Sheehan et al. Reference Sheehan, Lecrubier, Sheehan, Amorim, Janavs, Weiller, Hergueta, Baker and Dunbar1998), after recruitment, to assess for the presence of Axis I disorders. The MINI has established reliability and validity (Sheehan et al. Reference Sheehan, Lecrubier, Sheehan, Janavs, Weiller, Keskiner, Schinka, Knapp, Sheehan and Dunbar1997; Amorim et al. Reference Amorim, Lecrubier, Weiller, Hergueta and Sheehan1998). Subsequently, two CN participants were excluded from analysis. Fifteen CC participants (42%) screened positively for a range of psychiatric disturbances (mood disorder, 13; panic disorder, three; anxiety disorder, six; antisocial personality disorder, four). However, given that exclusion of these participants would have almost halved the sample size, and these participants had no official medical psychiatric diagnoses, the decision was made to retain them in the analyses. It should be noted that many studies have found increased rates of a range of psychiatric symptoms in chronic cannabis users (e.g. Troisi et al. Reference Troisi, Pasini, Saracco and Spalletta1998; Degenhardt et al. Reference Degenhardt, Hall and Lynskey2003a). In addition, we repeated the analyses with the exclusion of these 15 healthy cannabis users who screened positively for these psychiatric symptoms, and the findings were consistent with those observed in the sample as a whole. We excluded one CN and one SZN participant, as they were unable to provide urine samples, and as a result, absence of illicit drug use could not be confirmed. Thus, the final sample consisted of 142 participants: 49 SZN, 22 SZC, 35 CN and 36 CC.
All of the patients with schizophrenia were on antipsychotic medications. Of the SZN group, 43 were on atypical antipsychotics, one was on typical antipsychotics, and five were on both atypicals and typicals. Of the SZC group, 15 were on atypical antipsychotics, three were on typical antipsychotics, and four were on both atypical and typical. The average daily chlorpromazine equivalent dose (Atkins et al. Reference Atkins, Burgess, Bottomley and Riccio1997; Wood, Reference Wood2003) for the SZN group was 592.53 (s.d.=349.42) mg, and for the SZC group it was 648.24 (s.d.=263.10) mg. There was no significant group difference in the daily chlorpromazine equivalent dose [t(67)=−0.66, p=0.509]. For the SZN group, in addition to antipsychotics, four patients were taking anticholinergics, 16 were taking antidepressants, nine were taking benzodiazepines and 10 were taking mood stabilizers. For the SZC group, one patient was taking anticholinergics, two were taking antidepressants, seven were taking benzodiazepines and six were taking mood stabilizers. The average age of onset of illness for the SZN group was 22.65 (s.d.=6.69) years, and for the SZC group it was 20.45 (s.d.=2.76) years. There was no significant difference in age of onset of illness [t(68.70)=1.96, p=0.054]. International Classification of Diseases (ICD-10) psychiatric diagnoses in patients were confirmed with the Diagnostic Interview for Psychoses (Castle et al. Reference Castle, Jablensky, McGrath, Carr, Morgan, Waterreus, Valuri, Stain, McGuffin and Farmer2006). The demographic and substance-use characteristics of the sample can be found in Table 1. The illicit substance-use characteristics of the CC and SZC groups can be found in Table 2. This study was approved by the Western Australia North Metropolitan Area Mental Health Service Ethics Committee.
Table 1. Demographic and substance-use characteristics of the sample
Values are given as median (range) unless otherwise indicated.
CN, healthy non-using controls; CC, healthy cannabis-using controls; SZN, non-using patients with schizophrenia; SZC, cannabis-using patients with schizophrenia; THC, tetrahydrocannabinol; s.d., standard deviation; alcohol30, number of days of alcohol use in the previous 30 days; caffeine30, number of days of caffeine use in the previous 30 days.
Significant difference from CN: * p<0.05, ** p<0.01, *** p<0.0001.
Significant difference from CC: † p<0.05, †† p<0.01, ††† p<0.001.
Significant difference from SZN: ‡ p<0.05, ‡‡ p<0.01, ‡‡‡ p<0.001.
Significant difference from SZC: § p<0.05, §§ p<0.01, §§§ p<0.001.
Table 2. Recent illicit substance-use characteristics of cannabis-using controls and patients
Values are given as median (range).
CC, healthy cannabis-using controls; SZC, cannabis-using patients with schizophrenia; THC-COOH, 11-nor-Δ9-carboxy-tetrahydrocannabinol.
Substance-use assessment
Recent use of nicotine, alcohol, caffeine, cannabis and other illicit substances was assessed with a self-report questionnaire and with the alcohol and substance-misuse modules of the MINI, as previously described (Scholes & Martin-Iverson, Reference Scholes and Martin-Iverson2009a, Reference Scholes and Martin-Iversonb). Urine samples were also obtained and cloned-enzyme-donor-immunoassay was performed to screen for the presence of opiates, amphetamines, benzodiazepines, cannabis metabolites and cocaine metabolites according to the Australian/New Zealand standard AS/NZ 4308:2001 cut-off levels. Further, cotinine (nicotine metabolite) and 11-nor-Δ9-carboxy-tetrahydrocannabinol (THC-COOH) (cannabis metabolite) levels were quantified with gas chromatography–mass spectrometry.
Neuropsychological assessment
Stroop colour and word test
Attentional control was assessed with the standardized version of the Stroop task (Golden & Freshwater, Reference Golden and Freshwater2002), and raw scores were converted to T-scores in accordance with standard procedures (Golden & Freshwater, Reference Golden and Freshwater2002).
WMS-III: LNS and SS
The LNS (as a measure of auditory working memory) and the SS (both SS forward and SS backward subtasks, as a measure of spatial working memory) subtests of the WMS-III (Wechsler, Reference Wechsler1997) were administered. Raw scores for each subtask were converted to scaled scores (Wechsler, Reference Wechsler1997), and the scaled scores for LNS and SS were summed to give an overall measure of working memory.
WCST
The standard computerized version of the WCST was administered (Heaton & PAR Staff, 2005 b). Scoring was completed by the WCST computer scoring program (Heaton & PAR Staff, 2005 a) according to the standardized scoring procedures (Heaton et al. Reference Heaton, Chelune, Talley, Kay and Curtiss1993).
Procedure
All participants provided written informed consent upon arrival at the research centre (Centre for Clinical Research in Neuropsychiatry, Graylands Hospital, Perth, Australia). Demographic information was collected, and then participants completed the substance-use questionnaire, and were administered the alcohol and substance-use modules of the MINI. They then provided a urine sample. In order to reduce the likelihood of participants experiencing an abstinence syndrome during the testing session (e.g. Haney et al. Reference Haney, Ward, Comer, Foltin and Fischman1999a, Reference Haney, Ward, Comer, Foltin and Fischmanb, Reference Haney, Hart, Vosburg, Nasser, Bennett, Zubaran and Foltin2004; Hart et al. Reference Hart, Ward, Haney, Comer, Foltin and Fischman2002), cannabis users were instructed not to alter their cannabis use on the day of testing. Smoking of cigarettes was permitted ad libitum prior to the testing session, in order to reduce the likelihood of nicotine withdrawal affecting cognitive performance (e.g. George et al. Reference George, Vessicchio, Termine, Sahady, Head, Pepper, Kosten and Wexler2002; Evins et al. Reference Evins, Deckersbach, Cather, Freudenreich, Culhane, Henderson, Green, Schoenfeld, Rigotti and Goff2005). Potential acute effects of nicotine (e.g. Waters & Sutton, Reference Waters and Sutton2000; Smith et al. Reference Smith, Warner-Cohen, Matute, Butler, Kelly, Vaidhyanathaswamy and Khan2006) were minimized, as participants spent approximately 20 min with the researcher performing consent procedures and collecting demographic and substance-use information before the testing began. As this study was part of a larger study, participants were then prepared for psychophysiological recording, and recording of the startle reflex then took place (reported elsewhere; see Scholes & Martin-Iverson, Reference Scholes and Martin-Iverson2009a, b). Participants were then administered the Stroop task, the WMS-III subtasks and then the WCST.
Statistical analysis
Group differences in demographic and substance-use variables were investigated with one-way analysis of variance (ANOVA) (for normally distributed metric variables), non-parametric χ2 or Fisher's exact tests (for categorical variables) or non-parametric Kruskal–Wallis tests (for non-normally distributed metric variables).
Neuropsychological performance was analysed with repeated-measures multivariate analysis of covariance (RM MANCOVA) with two between-subjects factors [group: SZ (schizophrenia) and C (control); drug: C (cannabis user) and N (non-cannabis user)], the neuropsychological performance measures as the within-subjects factor (i.e. Stroop: word, colour, colour-word, interference) and six covariates [age, education, cotinine level, number of days of alcohol use in the previous 30 days (alcohol30), number of caffeinated drinks in the previous 24 h (caffeine number today) and average number of caffeinated drinks per day (average caffeine)]. Equality of error variances for each analysis was assessed with Levene's test, and no violations were observed. All covariates for all analyses met the assumptions of homogeneity of regression and multicollinearity. Planned pairwise comparisons with Sidak correction (p<0.05) were used to examine differences between groups for each neuropsychological performance measure.
Results
Demographics and substance use
As can be seen in Table 1, there were significant between-group differences in a number of the measures. Of these variables that differed significantly amongst the groups, there were significant correlations between the dependent measures (cognitive performance indices) and the variables age, education, cotinine level, alcohol30, caffeine number today and average caffeine; these variables were included as covariates in the subsequent analysis of the cognitive performance indices.
Table 2 shows the illicit substance-use characteristics of the cannabis-using participants. Of the CC participants, 21 were daily/nearly daily users, 12 were weekly users, one was a monthly user, and two used cannabis less than monthly. Additionally, 16 (47%) CC participants reported using other drugs in the last month (amphetamines, 12; hallucinogens, four). Toxicology analysis of the CC sample revealed that two screened positive for opiates (from reported pain medication taken the day before testing), eight screened positive for amphetamines, and one screened positive for benzodiazepines. The alcohol and substance-use modules of the MINI (for use of substances in the last 12 months) indicated that, of the CC group, eight screened positive for cannabis abuse, while 22 screened positive for cannabis dependence. Further, 12 (33%) screened positive for abuse/dependence of other substances.
Of the SZC participants, nine were daily/nearly daily users, seven were weekly users, four were monthly users and two used cannabis less than monthly. All SZC participants reported initiating cannabis use prior to their diagnosis of schizophrenia. Further, 12 (55%) SZC participants reported using other drugs in the last month (amphetamines, eight; narcotics, one; benzodiazepines, one; hallucinogens, two). Toxicology analyses of the SZC sample indicated that six patients screened positive for benzodiazepines, one screened positive for opiates, and two screened positive for amphetamines. The alcohol and substance-use modules of the MINI indicated that, of the SZC group, four screened positive for cannabis abuse and 15 screened positive for cannabis dependence. Further, 12 (55%) screened positive for abuse/dependence of other substances.
Neuropsychological performance
RM MANCOVA revealed a significant measure×group interaction for Stroop performance [F(3, 130)=13.30, p<0.0005, partial η2=0.24], but no significant drug×measure [F(3, 130)=0.07, p=0.976, observed power=0.06] or drug×group×measure [F(3, 130)=0.28, p=0.844, observed power=0.10] interactions. Similarly, there was a significant measure×group interaction for the WMS-III [F(4, 129)=4.49, p=0.002, partial η2=0.12], but no measure×drug interaction [F(4, 129)=1.29, p=0.276, observed power=0.40], and the group×measure×drug interaction was just off significance [F(4, 129)=2.43, p=0.051, observed power=0.68]. Finally, there was a significant measure×group interaction [F(9, 124)=2.71, p=0.007, partial η2=0.16] and measure×drug interaction for WCST performance [F(9, 124)=2.53, p=0.011, partial η2=0.16], but no significant group×measure×drug interaction [F(9, 124)=1.86, p=0.064, observed power=0.80]. Findings from the planned pairwise comparisons can be found in Table 3.
Table 3. Neuropsychological performance across the four groups: CN, CC, SZN and SZC
CN, healthy non-using controls; CC, healthy cannabis-using controls; SZN, non-using patients with schizophrenia; SZC, cannabis-using patients with schizophrenia.
Values are given as mean (standard error).
Significant difference from CN: * p<0.05, ** p<0.01, *** p<0.0001.
Significant difference from CC: † p<0.05, †† p<0.01, ††† p<0.001.
Significant difference from SZN: ‡ p<0.05, ‡‡ p<0.01, ‡‡‡ p<0.001.
Significant difference from SZC: § p<0.05, §§ p<0.01, §§§ p<0.001.
These analyses were repeated controlling for the cannabis-use variables that differed significantly between the two cannabis-using groups (Table 2). The pairwise comparisons between the CC and SZC groups from these analyses were consistent with those obtained without controlling for the cannabis-use variables, with the majority of significant comparisons exhibiting even greater differences.
Frequency and recency of cannabis use
Given that one study found recent and frequent use of cannabis in patients to be associated with better cognitive performance on some tasks (Coulston et al. Reference Coulston, Perdices and Tennant2007), the analyses were re-performed including only those from the cannabis-using groups who were daily or nearly daily users of cannabis. Despite the reduced sample sizes in the cannabis-using groups (CC=21, SZC=9), the significance of pairwise comparisons was entirely consistent with that observed in the sample as a whole. To examine recency of use, only those cannabis users who had used cannabis within the previous 24 h were included (CC=20, SZC=7). Again, the findings were consistent with those observed in the sample as a whole. Lastly, analyses were repeated including only those who had not used cannabis within the previous 24 h (CC=16, SZC=15). These analyses were consistent with those observed when utilising the whole sample.
In addition, partial correlations (with the same covariates as included in the ANOVAs) between the cannabis-use variables and cognitive performance were performed for both the CC and SZC groups. There were no significant correlations between cannabis use and neuropsychological performance, for either the CC or SZC group, after Bonferroni–Holm correction for multiple comparisons.
Use of other substances
In order to investigate whether the use of other substances (both alcohol and illicit substances) could have contributed to the observed findings, the analyses were again completed using various subgroups of the cannabis-using groups. The results of these analyses were all consistent with the analyses completed utilising the whole sample. These subgroup analyses included: only those who did not screen positive to any other substance with the urine toxicology screen (CC=27, SZC=16); only those who reported no use of any other illicit substance in the previous 30 days (CC=20, SZC=10); only those with no other substance abuse/dependence in the previous year, according to the MINI (CC=22, SZC=8).
Discussion
The current study examined the effects of chronic cannabis use on neuropsychological performance, in both healthy people and patients with schizophrenia. The neuropsychological domains assessed represented those that have been well established to be disturbed in patients with schizophrenia: attentional control, as assessed with the Stroop task; working memory, as assessed here with the WMS III; and executive functioning, as assessed by the WCST. Consistent with the literature (for a review, see Heinrichs & Zakzanis, Reference Heinrichs and Zakzanis1998), patients with schizophrenia showed poorer performance on all tasks, as compared with healthy controls. Interestingly, there were few differences in cognitive performance between cannabis users and non-users, for both patients with schizophrenia and healthy controls. These findings suggest that cannabis use has very little effect on the cognitive functions examined.
Consistent with a multitude of studies, patients with schizophrenia showed poorer performance than controls on all measures of the Stroop task (for a review, see Henik & Salo, Reference Henik and Salo2004). However, there were no significant differences between the CC and CN participants on any Stroop measures, nor were there any significant differences between the SZC and SZN groups. This suggests that chronic cannabis use had no effect on Stroop performance. These findings are in line with a number of other studies that have found no effects of chronic cannabis use on performance of the Stroop task in healthy people (Miller et al. Reference Miller, Drew and Kiplinger1972; Pope et al. Reference Pope, Gruber, Hudson, Huestis and Yurgelun-Todd2001; Solowij et al. Reference Solowij, Stephens, Roffman, Babor, Kadden, Miller, Christiansen, McRee and Vendetti2002; Eldreth et al. Reference Eldreth, Matochik, Cadet and Bolla2004; Gruber & Yurgelun-Todd, Reference Gruber and Yurgelun-Todd2005). While acute cannabis intoxication may disturb Stroop performance in healthy people (Hooker & Jones, Reference Hooker and Jones1987; Henquet et al. Reference Henquet, Rosa, Krabbendam, Papiol, Fananas, Drukker, Ramaekers and Van Os2006), the subjects in the current study were not acutely intoxicated at the time of testing. Further, those previous studies that have shown impaired Stroop performance in healthy chronic cannabis users have only shown this to be the case in select groups, such as very heavy users who are males (Pope & Yurgelun-Todd, Reference Pope and Yurgelun-Todd1996). Our findings are also consistent with those reported in patients with schizophrenia, with no difference found between cannabis users and non-users (Coulston et al. Reference Coulston, Perdices and Tennant2007; Thoma et al. Reference Thoma, Wiebel and Daum2007). Although one early study did find that lifetime cannabis-use disorder in patients was associated with poor interference performance on the Stroop task (Liraud & Verdoux, Reference Liraud and Verdoux2002), the sample of this study consisted of patients diagnosed with a range of both psychotic and mood disorders, and thus the specificity of this relationship in schizophrenia could be questioned.
Patients with schizophrenia also showed poorer spatial working memory, relative to controls. Spatial working memory deficits have been suggested to be the most robust working memory disturbance observed in schizophrenia (Lee & Park, Reference Lee and Park2005). Consistent with the Stroop findings, there were no significant differences in any of the working memory measures, between CC and CN participants, nor were there significant differences between SZC and SZN patients. However, SZC patients did show more widespread deficits in the working memory measures, when compared with CN participants, than did SZN patients. While one previous study found contrasting findings, whereby superior working memory performance was displayed by cannabis-using patients, as compared with non-users (Schnell et al. Reference Schnell, Dagmar, Daumann and Gouzoulis-Mayfrank2009), other studies have found no such differences between users and non-users (Cleghorn et al. Reference Cleghorn, Kaplan, Szechtman, Szechtman, Brown and Franco1991; Jockers-Scherubl et al. Reference Jockers-Scherubl, Wolf, Radzei, Schlattmann, Rentzsch, Gomez-Carrillo de Castro and Kuhl2007; Sevy et al. Reference Sevy, Burdick, Visweswaraiah, Abdelmessih, Lukin, Yechiam and Bechara2007; Mata et al. Reference Mata, Rodriguez-Sanchez, Pelayo-Teran, Perez-Iglesias, Gonzalez-Blanch, Ramirez-Bonilla, Martinez-Garcia, Vazquez-Barquero and Crespo-Facorro2008), as in the current study. However, these latter studies did not include a healthy non-using control group (Cleghorn et al. Reference Cleghorn, Kaplan, Szechtman, Szechtman, Brown and Franco1991; Mata et al. Reference Mata, Rodriguez-Sanchez, Pelayo-Teran, Perez-Iglesias, Gonzalez-Blanch, Ramirez-Bonilla, Martinez-Garcia, Vazquez-Barquero and Crespo-Facorro2008) or did not conduct the relevant pairwise comparisons between the healthy non-using control group and the cannabis-using patients (Jockers-Scherubl et al. Reference Jockers-Scherubl, Wolf, Radzei, Schlattmann, Rentzsch, Gomez-Carrillo de Castro and Kuhl2007). Hence, it cannot be ascertained whether, as in the current study, cannabis-using patients showed more widespread deficits in memory function than did non-using patients. Although the published literature on the effects of chronic cannabis use on working memory in healthy individuals is somewhat mixed, the lack of difference between cannabis users and non-users, as observed here, is supported by a number of studies (Solowij et al. Reference Solowij, Stephens, Roffman, Babor, Kadden, Miller, Christiansen, McRee and Vendetti2002; Kanayama et al. Reference Kanayama, Rogowska, Pope, Gruber and Yurgelun-Todd2004; Jager et al. Reference Jager, Kahn, Van Den Brink, Van Ree and Ramsey2006).
Similarly, both SZC and SZN patients showed poorer WCST performance across most WCST measures, as compared with the CN group. Poor WCST performance in patients has been well documented (e.g. Park, Reference Park1997; Heinrichs & Zakzanis, Reference Heinrichs and Zakzanis1998; Bustini et al. Reference Bustini, Stratta, Daneluzzo, Pollice, Prosperini and Rossi1999; Everett et al. Reference Everett, Lavoie, Gagnon and Gosselin2001; Hartman et al. Reference Hartman, Steketee, Silva, Lanning and Andersson2003; El Hamaoui et al. Reference El Hamaoui, Elyazaji, Yaalaoui, Rachidi, Saoud, d'Amato, Moussaoui, Dalery and Battas2006), and the current study demonstrates that patients with schizophrenia who use cannabis exhibit similar deficits. This finding is consistent with Jockers-Scherubl et al. (Reference Jockers-Scherubl, Wolf, Radzei, Schlattmann, Rentzsch, Gomez-Carrillo de Castro and Kuhl2007) and Coulston et al. (Reference Coulston, Perdices and Tennant2007) who both found no additive effect of cannabis use on WCST disturbances in schizophrenia.
In addition, both patient groups showed deficits in WCST measures relative to cannabis-using controls, with one exception. That is, there were no significant differences in perseverative responses and perseverative errors between the CC and SZN groups. In line with this, CC participants had significantly lower scores for perseverative errors, relative to CN participants. Thus, it appears that CC participants show more perseverative errors during the WCST, which is consistent with the deficit observed in schizophrenia. This finding is in accordance with a previous study which found a specific increase in perseveration in healthy cannabis users (Pope & Yurgelun-Todd, Reference Pope and Yurgelun-Todd1996). While a latter study by this research group found no such difference (Pope et al. Reference Pope, Gruber, Hudson, Huestis and Yurgelun-Todd2001), this may be attributed to methodological differences, as a 28 day abstinence period was employed in the 2001 study. This suggests that the schizophrenic-like increase in perseveration observed here may be associated with the residues of cannabis in the body, and, thus, may resolve after these residues are cleared following sufficient abstinence. However, the current study did not detect any significant correlations between recency of cannabis use and WCST perseveration.
The findings from the ANOVAs and pairwise comparisons suggest that there is very little effect of chronic cannabis use on the cognitive performance measures administered. This is supported by the lack of correlations observed between the cannabis-use measures and the neuropsychological indices examined. The neuropsychological measures employed in the current study are measures which have been consistently demonstrated to be disturbed in schizophrenia, and hence the findings of the current study suggest that cannabis use in healthy individuals does not produce deficits, in these measures, that resemble those in schizophrenia (with the exception of increased perseveration in the WCST), and cannabis use in patients with schizophrenia does not produce further decrements in these fundamental processes. It should be noted that the lack of significant differences between SZN and SZC patients cannot be explained by a floor effect in cognitive performance, as the T-scores for many of the measures were above 40. Further, given that our cannabis-using groups consisted of participants who also used other illicit drugs, we completed a second round of analyses excluding individuals based on use of other substances (both alcohol and illicit drugs). These analyses were all consistent with the findings utilizing the whole cannabis-using sample, suggesting that our findings are not confounded by our participants' infrequent use of other substances.
A number of recent studies have suggested that cognitive performance is actually superior in patients who use cannabis relative to non-users (Stirling et al. Reference Stirling, Lewis, Hopkins and White2005; Coulston et al. Reference Coulston, Perdices and Tennant2007; Potvin et al. Reference Potvin, Joyal, Pelletier and Stip2008; Henderson et al. Reference Henderson, Coulston, Lagopoulos, Degabriele, Das and Malhi2009; Schnell et al. Reference Schnell, Dagmar, Daumann and Gouzoulis-Mayfrank2009; Yucel et al. Reference Yucel, Bora, Lubman, Brewer, Cotton, Conus, Condello, Wood, McGorry and Pantelis2009). There are a number of possible reasons for this. First, many of these studies suffer from small sample sizes. For example, the study by Coulston et al. (Reference Coulston, Perdices and Tennant2007) examined a number of indices of cannabis use and their relationship to cognitive performance. Although the overall sample of cannabis-using patients was sufficient, the samples on which the conclusions were based, when separating the sample according to frequency and recency of cannabis use, were only very small. Further, cannabis-using patients have been suggested to reflect a relatively distinct group who differ from non-using patients in terms of pre-morbid social adjustment and/or intelligence quotient (IQ), abilities that are needed in order to initiate and maintain drug-seeking behaviour (e.g. Joyal et al. Reference Joyal, Halle, Lapierre and Hodgins2003; Wobrock et al. Reference Wobrock, Sittinger, Behrendt, D'Amelio, Falkai and Caspari2007; Potvin et al. Reference Potvin, Joyal, Pelletier and Stip2008; Schnell et al. Reference Schnell, Dagmar, Daumann and Gouzoulis-Mayfrank2009). Although a recent study suggested that drug-using patients do not differ from non-using patients in terms of pre-morbid social functioning or competency (Stirling et al. Reference Stirling, Lewis, Hopkins and White2005), there is some evidence that cannabis-using patients may show a higher IQ (Kumra et al. Reference Kumra, Thaden and DeThomas2005), which may account for the improved performance observed in some studies. Such a contention is supported by the study by Schnell et al. (Reference Schnell, Dagmar, Daumann and Gouzoulis-Mayfrank2009) who found improved performance in cannabis-using patients, relative to non-using patients, after an abstinence period of 78 days on average. Given that a number of studies in healthy cannabis users have suggested the deficits in cognitive performance associated with cannabis use can resolve after sufficient periods of abstinence (over 30 days) (e.g. Pope et al. Reference Pope, Gruber, Hudson, Huestis and Yurgelun-Todd2001), it might be suggested that the improved cognition in studies such as the one by Schnell et al. (Reference Schnell, Dagmar, Daumann and Gouzoulis-Mayfrank2009) may be indicative of superior pre-morbid IQ/cognition which becomes evident after resolution of the deficits associated with cannabis use, upon sufficient periods of abstinence. Such an interpretation would be supported by our study, and others who examine cannabis-using patients (Jockers-Scherubl et al. Reference Jockers-Scherubl, Wolf, Radzei, Schlattmann, Rentzsch, Gomez-Carrillo de Castro and Kuhl2007; Sevy et al. Reference Sevy, Burdick, Visweswaraiah, Abdelmessih, Lukin, Yechiam and Bechara2007; Mata et al. Reference Mata, Rodriguez-Sanchez, Pelayo-Teran, Perez-Iglesias, Gonzalez-Blanch, Ramirez-Bonilla, Martinez-Garcia, Vazquez-Barquero and Crespo-Facorro2008) and also substance-using patients (primarily cannabis users) (Addington & Addington, Reference Addington and Addington1997; Pencer & Addington, Reference Pencer and Addington2003; Thoma et al. Reference Thoma, Wiebel and Daum2007; Wobrock et al. Reference Wobrock, Ecker, Scherk, Schneider-Axmann, Falkai and Gruber2008), without any imposed abstinence period, and find largely no differences in cognitive performance between users and non-users. Whether this interpretation is correct remains to be investigated in a longitudinal design where chronic cannabis-using patients are tested during normal use of cannabis, and then again after a prolonged period of abstinence.
In conclusion, it appears that cannabis use in both healthy individuals and patients with schizophrenia has only very subtle effects on performance of the neuropsychological tasks administered here, which have long been established to index characteristic disturbances in schizophrenia. As such, current chronic cannabis use in healthy individuals produces little similarity in cognitive dysfunction to that evident in schizophrenia, with the exception of a schizophrenic-like increase in perseveration. Similarly, patients with schizophrenia who are current chronic users of cannabis appear to show little difference in cognitive performance to those who do not use cannabis, and, as such, it may be that chronic cannabis use has no additive effect on cognitive dysfunction in schizophrenia.
Acknowledgements
We gratefully acknowledge the assistance of Paul Connelly and Sarah Howell in the recruitment of the patient sample, and also Peter Hackett and Rose Kwa for toxicology analyses. This research was supported by a National Health and Medical Research Council (NHMRC) grant (no. 403994).
Declaration of Interest
None.
