cannabis and psychosis

The association between substance misuse (cannabis & alcohol) and psychosis

The association between cannabis and psychosis:

The prevalence of substance misuse (i.e. alcohol & street drugs) is widespread in mental disorders (McCrone et al 2000). Psychotic disorders (i.e. schizophrenia and schizoaffective disorders) have a lower incidence rates compared to other mental disorders; however they impose a considerable personal and public health burden (Degenhardt & Hall, 1997) since it is common for people with psychosis to have co-occurring drug or alcohol problems (McCreadie, 2002; Regier et al. 1990). This Comorbidity as a result largely enhances cost associated with treatment, predominantly in relation to hospital and other institutional services. The majority of the evidence regarding prevalence of substance use and mental illness comes from the United States, though European standpoint on prevalence is rapidly rising. One of the first and most famous US study is The Epidemiological Catchment Area study conducted by Regier et al. 1990 across 5 US sites, which involved over 20.000 participants. This study tried to measure how common dual diagnosis was and stated that 47% of the participants they surveyed with schizophrenia or schizophreniform disorder also met the criteria for substance use disorder/dependence at some point in their life. Furthermore, the likelihood of having substance use disorder amongst patients with psychotic illness was notably higher compared to those in the general population without a psychotic illness (Kessler et al. 1994; Regier et al. 1990). In sharp contrast, studies that have been conducted in the UK have reported more moderate rates of substance misuse amongst those with psychotic disorders. For instance, Cantwell in 2003 undertook a study with 316 patients with schizophrenia and only 7% had stated that they had a problem with substance misuse in the previous year, and 21% had reported problematic use some time before that (Cantwell, 2003). Similarly, Wright and Colleagues identified patients with psychotic illnesses who had been in contact with services in the London Borough of Croydon and reported that prevalence rates of dual diagnosis were 33% for the use of any substance and lifetime history of any illicit drug use was 35% of the sample (Wright et al. 2000).

Cannabis is the most frequently used substance in patients with psychotic disorder (Van Mastrigt et al. 2004; Barnett et al. 2007) and the literature indicates fairly consistently the association between regular uses of cannabis and developing first-onset psychosis[1] (Grech et al. 2005). However, the causes for the links are still hotly debated. Nevertheless, cannabis has damaging effects on the course of the illness on top of being an economic burden on health services (Atakan, 2009). More specifically, estimates of frequency of cannabis abuse amongst patients with first-episode psychosis is said to range from 20% to 70% (Miles et al. 2003; Isaac et al. 2005). A systematic review of 53 treatment samples and 5 epidemiological studies looking at the prevalence of cannabis use and misuse among those with psychotic illnesses stated lifetime use in 42% and lifetime misuse in 22.5% of the sample while current use was 23% and misuse 11.3% (Green et al. 2005). Furthermore, epidemiological studies also have shown that cannabis is the most frequently abused substance amongst younger people (Van Mastrigt, 2004; Addington & Addington, 2007; Wade et al. 2007) with lower educational background (Mueser et al. 1990) and it is more common in men than in women amongst first-episode psychosis patients (Barnett et al. 2007; Addington & Addington, 2007). However, a US study by (Sevy et al. 2001) reported no differences in sex, education, age of onset and baseline symptoms when they examined first episode psychosis patients with and without substance misuse problems. In addition, studies also have reported that adolescents who use cannabis on a weekly basis are at greater risk of developing acute psychosis 3 or more years later compared with non-users (Veen et al. 2004; Barnes et al. 2006). Inevitably, cannabis abuses among youth experiencing a first-episode psychosis results in negative consequences for prognosis and relapse and therefore complicate treatment and impair recovery (Archie et al. 2009). It is constantly reported in the literature that patients with first-episode psychosis who abuse cannabis had considerably more hospital admissions rates as well as more psychotic symptoms compared to first episode psychosis patients who do not abuse cannabis (Wade et al. 2006; Negrete et al. 1986; Turkington et al. 2009)

There has been a lot of debate in the literature whether an association between cannabis consumption and psychosis exists, which has received substantial interest from scientists in recent years (Hall, 1998; Mueser et al. 1998; Blanchard et al. 2000; Degenhardt & Hall, 2002). The strongest evidence for causality is that cannabis consumption leads to a high risk of relapse of psychotic symptoms (Jablensky et al. 1991; Martinez- Arevalo et al. 1994). A study by Linszen et al. (1994, 2009) have found that patients with recent-onset schizophrenia who abused cannabis had significantly more and earlier psychotic relapses compared to non users. In addition, the researchers also found a dose-response effect with heavy cannabis users had a strong association with higher psychotic relapses (Linszen et al. 1994; Moore et al. 2007; Andreasson et al. 1987). Consistent with these findings, Van Os and colleagues (1992) also reported that cannabis consumption was closely linked with both a high risk of occurrence of psychosis and a poor prognosis in a sample which was followed over 3 years. More recently, in a longitudinal study in Canada by Pencer & Addington (2003) have found that cannabis use positively associated with more severe positive symptoms of schizophrenia at 1-year follow-up period. A paper by Verdoux et al. (2003) presents a unique piece of evidence providing specificity on the temporal association between cannabis consumption and psychotic symptoms over period of several hours and they concluded that there was a close association between the two and this association remained stable after controlling for other drug use, age and gender.

However, the greatest support that the associations between cannabis consumption and psychosis might be causal comes from studies that are longitudinal. For instance, one of the most famous and first prospective studies of cannabis use and schizophrenia was a 15 year follow-up of 50.465 young Swedish male conscripts. The authors concluded that those people who tried cannabis by age 18 were 2.4 times more probable to develop schizophrenia compared to those who have not consumed cannabis (Andreasson, Allebeck & Rydberg, 1987). The risk of being diagnosed with schizophrenia also significantly correlated with frequency of cannabis consumption. In another influential study, Zammit et al. (2002) undertook a 27-year follow-up of the Swedish cohort and concluded that there was a dose-response relationship between the regularity of cannabis use at age 18 and risk of getting diagnosed with schizophrenia during the follow-up. The findings that cannabis use in adolescence predicts psychotic symptoms or illnesses in adult life were also supported by other longitudinal studies conducted by Van Os et al. (2002) in Netherlands, Henquet et al. (2004) in Munich and Arsenault et al. (2002) in New Zealand. On the other hand, some researchers have critically evaluated the scientific evidence of a causal association between cannabis and psychosis and stated that the association at this moment in time was not that strong suggesting more observational and genetic studies to be conducted (Macleod, 2007). Nonetheless, these studies repeatedly stress out the damaging effects of cannabis and its negative implications on people.

Other researchers, in contrast, have argued that psychosis causes increased need for substance use (Kushner & Mueser, 1993). The most well studied model which claims that substance use disorder arises as a result of psychiatric problems is the “self-medication hypothesis” (Khantzian, 1985; 1997). In a nut shell, this hypothesis states that substance misuse is an effort to self-medicate psychiatric problems and assumes that people with mental health issues use substances to alleviate their unwanted and usually distressing symptoms. However, this hypothesis has been criticized. For instance, the best evidence comes from the Christchurch study in which cannabis use and psychotic symptoms data were gathered over the course of 25-year longitudinal study. The researchers have concluded that cannabis use at age 18 was correlated with more severe psychotic symptoms 3 and 7 years later. However, the psychotic symptoms at age 18 appeared to hold back the use of cannabis rather than promote cannabis use. Therefore, this study concluded that “the direction of causality is from cannabis use to psychotic symptoms (Fergusson et al. 2005).

What makes people who abuse cannabis more prone to develop psychotic symptoms? The main components of cannabis are ∆9-tetrahydrocannabinol (∆9-THC) and cannabidiol (Di Forti et al. 2009). The ∆9-THC is the main psychoactive component of cannabis plant which creates transient and usually mild, psychotic and affective experiences (Moore et al. 2007; D’Souza et al. 2005) and impaired memory in a dose dependent way (Murray et al. 2008). On the other hand, cannabidiol doesn’t produce hallucinations or delusions but it antagonises the cognitive impairment and psychotogenic effects resulted from ∆9-THC (Murray et al. 2008). Over the past few decades, it has been argued that the strength of cannabis has significantly risen due to containing three times more tetrahydrocannabinol compared to 1960s (Atakan, 2009). In addition, Potter et al. (2008) reported that even the THC ingredient of “skunk” has nearly two folded in the last 10 years. For instance, in a recent study by Di Forti et al. (2009) conducted a study to investigate whether first-episode psychosis patients were more prone to use high-potency cannabis and concluded that amongst those people who abused cannabis, 78% of the sample used high-potency cannabis (“skunk”) compared with 37% of the control group. Therefore, it is not surprising that the effects of cannabis on psychotic patients results in significantly worse clinical outcomes such as increase in suspiciousness and hallucinatory behaviour when a patient with severe mental illness uses cannabis (Sorbara et al. 2003).

The association between alcohol and psychosis:

It has been reported by some researchers in the literature that alcohol misuse in people with first-episode psychosis is widespread and leads to poor treatment outcome (Laugharne et al. 2002; Peterson et al. 2002; Addington & Addington, 1998). However, relative to cannabis misuse, studies that have examined the prevalence of alcohol misuse among patients with first-episode psychosis remain comparatively rare (Barnes et al. 2006). Most research investigating the function of substance misuse in psychosis mainly concentrated on cannabis and only few publications examined the role of alcohol misuse in psychosis (Crebbin et al. 2008). Consequently, the relationship between alcohol and psychosis remains relatively blurred.

In terms of prevalence rates, one study stated a baseline rate of alcohol misuse to be 14.6% in adolescent-onset first-episode psychosis patients (Ballageer et al. 2005). In another study, Menezes et al. (1996) have reported a one year prevalence rate of alcohol misuse in Camberwell to be 32%. More recently, a study found that 27% of first-episode psychosis patients met the criteria for problem drinking (Barnes et al. 2006). Data from the National Institute of mental health’s Epidemiologic Catchment Area survey reported a prevalence of recent alcohol misuse among schizophrenic patients which was ten times higher compared to the rest of the community sample (Boyd et al. 1984) and Regier et al. (1990) indicated that patients with schizophrenia were 3 times as likely as the general population to be alcoholics. In addition, studies also reported that among people with first-episode psychosis, men abused alcohol more often compared to women (Barnett et al. 2007) and those who abuse alcohol tend to be older than users of non-alcoholic substances (Salyers & Mueser, 2001). Johns and colleagues in 2004 have speculated that in adult population survey in UK, alcohol dependence was a predictor of psychotic symptoms and “the reported two-fold higher risk was independent of other risk factors for psychotic symptoms including drug dependence, suggesting that alcohol dependence per se doubles the risk of developing psychotic symptoms” (Johns et al 2004, cited from Crebbin et al. 2008). In contrary, Sorbara et al. (2000, 2003) undertook a study to investigate the impact of alcohol on the early course of psychosis and concluded that there was no association between alcohol misuse and having psychotic relapse. In support to these findings, in a longitudinal study conducted by Pencer & Addington in Canada (2003) have found that alcohol use had no impact on positive symptoms of schizophrenia at 1-year follow-up suggesting that the kind of a substance abused by patients plays an important role on the outcome of their symptoms.

What about the relationship between alcohol and psychosis? Studies looking at the relationship between alcohol misuse and psychosis have reported that alcohol dependence and not alcohol use as such predicted psychotic experiences in the general population (Johns et al. 2004; Tien & Anthony, 1990). Some researchers have found that patients with psychotic experiences tend to abuse more alcohol compared to those who do not have psychotic symptoms (Olfson et al. 2002). However, general consensus in the literature is that even though alcohol abuse results in worsening of psychotic symptoms experienced by patients and precipitates relapse, it does not in fact cause schizophrenia (Bernadt & Murray, 1986; Hambrecht & Hafner, 1996). In terms of psychotic symptoms, patients with schizophrenia who abused alcohol stated that they have experienced significantly more hallucinations and depressive symptoms compared to non-alcoholic schizophrenics (Pulver et al. 1989). Moreover, in a study conducted by Drake et al. (1989) have found that heavy alcohol use was notably associated with hostile threats, paranoia, depression and suicidal behaviour as well as disorganized incoherent speech. Additionally, alcohol use disorders in schizophrenia has also been closely linked with delusions (Barbee et al. 1989), assaultiveness (Yesavage & Zarcone, 1983), poor self-care (Alterman et al. 1980), housing instability and homelessness (Drake et al. 1991) as well as poor treatment compliance (Drake et al. 1989). These findings contrast with “self-medication hypothesis” proposed by Khantzian (1985). As stated earlier on, a central assumption of this hypothesis is that patients often self-medicate or ease their unwanted and disturbing symptoms through the use of substances. For instance, Schneier and Siris, (1987) have indicated that people with schizophrenia misuse alcohol because alcohol helps to reduce the anxiety and stress due to their extra pyramidal symptoms caused by their anti-psychotic medication. However, as seen from the above studies, alcohol abuse has worsened their outcome rather than alleviating their symptoms.

In general, whilst there is short-lived psychotic symptoms associated with heavy and long-standing alcohol use reported by many patients, differing reports deeply propose that our understanding of the role of alcohol alone in the course of psychotic symptoms is sadly less clear (National drug strategy, 2006).

Association between cannabis use & alcohol and contacts with mental health services & home treatment teams

Contacts with mental health services:

Research that has been conducted to investigate the level of service use in first-episode psychosis patients have found that those patients with problematic substance use has been associated with higher relapse rates (Turkington et al. 2009; Jean-Paul Selten et al. 2007; Salyers & Mueser, 2001) and not surprisingly resulted in an increased number of inpatient mental health service use compared to first-episode psychosis patients who did not abuse substances (Haywood et al. 1995; Wade et al. 2006; Isaac et al. 2005). Similarly, in France, Sorbara et al. (2003) reported that having a comorbid substance use disorder with psychosis have resulted in increased number of days spent in hospital and that early discharge from hospital increased the likelihood of experiencing a relapse with three times more likely to be readmitted to hospital again. Patients in this study mainly abused cannabis (66%) and this study did not found any associations between alcohol consumption and relapse. Similar finding was also reported by Crebbin et al. (2008) that substance use without alcohol abuse was correlated with increased hospital days. Additional support for the assumption that having a substance use disorder increases the chances of contacts with mental health services comes from a UK study conducted by Sipos et al. (2001) in Nottingham with 166 patients with first-episode psychosis. The researcher concluded that drug abuse was significantly correlated with higher risk of rapid admission. In another study, Menezes et al. (1996) has shown that patients with dual diagnosis had a nearly double inpatient admission rates than those psychosis alone and the majority of the patients were alcohol abusers. However, it should be noted that the number of admission rates were not statistically significant. A retrospective study by Gupta et al. (1996) also confirms a previous finding that having a substance use disorder significantly correlates with psychiatric hospitalizations as well as higher rates of emergency-room visits (Ziedonis et al. 2005; Bartels et al. 1993).

Drawing on from these studies, the literature points out quite consistently that substance use worsens specifically positive symptoms experienced by patients which in turn has found to be strongly associated with longer in-patient stays (McCrone et al. 2000; Strakowski et al. 1994; Wright et al. 2000). Additionally, it has also been argued that once patients are in hospitals with limited access to substances they abuse, Ries et al. (2000) have found that those patients who abused substances before admission seem to report less symptom worsening compared to non-users (Lambert et al. 2005). A compelling study by Cantor-Graae et al. (2001) in Sweden with 87 patients seems to perfectly demonstrate the points mentioned above. Their study have found that those patients who have abused substances constantly in the community had increased number of admissions however, the duration in which they spent time in the hospital was less compared to those patients with only single diagnosis of schizophrenia. More surprisingly, alcohol was the main type of substance abuse which this result challenges previous findings that have suggested that there was no association between alcohol use and admission rates. Furthermore, Swofford et al. (1995) also found that substance use predicted subsequent relapse and hospitalization and majority of them had a history of abuse of alcohol.

Service use and cost for patients with schizophrenia were investigated in another study carried out by Bartels et al. (1993) in an out-patient service. The findings were not different from previous studies reported above and the researchers concluded that during 1-year those patients with substance use problems made use of institutional care and emergency services more often compared to those who did not had substance use problems. In another highly influential work carried out by McCrone et al. (2000) compared service use of individuals who were diagnosed with psychosis and a substance use disorder with psychosis alone. The researchers measured six-month of service use and concluded that “a greater proportion of the patients with dual diagnosis used community psychiatric nurses, in-patient care and the emergency clinic”. Despite such evidence, the other researchers have found diverse outcomes in terms of service use in patients with psychotic disorders who abuse substances. For instance, a study conducted in the US reported that patients with schizophrenia who also had substance use disorders who were admitted to hospital had a quicker recovery rate compared to those patients with a diagnosis of schizophrenia only. They also found that, at the time of the admissions, both patient groups showed the same severity of their psychotic symptoms but more importantly, it was observed that those patients with dual diagnosis tended to be discharged earlier and with less psychotic symptoms than those with schizophrenia only diagnosis (Ries et al. 2000). Consistent with these findings, a more recent study examining 316 patients with schizophrenia in Scotland with and without substance abuse problems have found that there was no statistical variation in terms of hospital admission rates and the number of hospitalizations between users and non-users. More surprisingly, the study failed to find any significant effects of substance use on their psychotic symptoms (Cantwell, 2003).

Consequently, as seen from the above, the evidence regarding service use produced mixed results in the literature. There is an ongoing debate whether alcohol misuse results in longer hospitalizations or not but evidence regarding cannabis use and contacts with mental health services seems to be pretty strong.

Contacts with Home treatment Teams (HTT):

According to Johnstone & Zolese, (1999) the most costly type of psychiatric care is hospital admission. However, in recent year’s alterations that has been made to mental health services allowed various acute psychiatric disorders to be managed without requiring hospitalization (Dean & Gad, 1990; Muijen et al. 1992). This different approach of care to mental health services resulted in significant reduction of patients with first-episode psychosis needing admission at initial contact (Harrison et al. 1991). The introduction of such care into the acute mental health care in most areas of United Kingdom is called crisis resolution teams[2] (CRT, sometimes called home treatment team) (Johnson et al. 2008) and have been in existence since the 1980s (Carroll et al. 2001). Consequently, there is a great amount of evidence in the literature stating the benefits of treatment at patient’s own home (Minghella et al. 1998; Harrison et al. 1999; Bracken and Cohen. 1999). According to a recent article, “home-based treatment can help individuals and their careers by providing care within a free and familiar environment and it has the potential to facilitate a rapid return to normal lifestyle and function” (Tomar et al. 2003).

However, the most important question is whether home treatment teams deal with the crisis effectively and reduce inpatient admission with a less cost to mental health services. According to a recent study conducted by Dibben et al. (2008) home treatment teams have successfully reduced adult inpatient admissions on average by 18%. Similarly, one naturalistic and one randomized control trial have also examined outcomes of home treatment team care in London (Johnson et al. 2005) and concluded that the teams have effectively reduced hospital admissions at 8 weeks and 6 weeks of home-based care.

Another important question that needs to be addressed is whether substance users with psychotic disorders contact home treatment teams more often compared to non-users and therefore cost more to the mental health services? Unfortunately, not many studies have conducted in this area examined the link between dual diagnosis and contacts with home treatment teams versus psychosis alone and contacts with home treatment teams (Cotton et al. 2007). This finding therefore highlights the need in UK to examine the frequency and the impact of problematic substance use among those with severe mental health problems, predominantly in community settings (Graham et al. 2001). For instance, in a study conducted by Priebe et al. (2003) investigated patient characteristics in routine assertive outreach services in the UK. The study found that almost 20% of the sample they examined abused or was dependent on drugs and 16% misused or were dependent on alcohol. The drug that most of the patients abused was cannabis. In addition, it has been reported that problematic drug and alcohol misuse in assertive outreach teams rose to 45% (Graham et al. 2001).

Cost of dual diagnosis to mental health services:

Although majority of patients who come into contact with psychiatric services have problematic substance use disorders (Wright et al. 2002), as yet there have been few attempts to understand the human and financial cost of this problem to mental health services. Nevertheless, there has been a lot of attention from clinicians, politicians and academics in recent years towards people with psychotic disorders who abuse substances because of the major economical cost to society (Bartels et al. 1993)and personal cost to themselves and not to forget the emotional burden upon the sufferer’s family (Clark, 1994). Therefore, it is constantly been pointed out by many researchers that the cost of care for people with dual diagnosis whether treated or not, serves as a huge challenge to the modern healthcare system (Dixon, 1999). To make the matters more concrete, Hall et al. (1985) have reported that “the economic cost of psychotic disorders such as schizophrenia has been estimated as similar to that of all cancers combined”. According to a very recent report published by NHS confederation in 2009 stated that “people who have developed problems with alcohol or drugs because of a pre-existing mental health condition, or have had a mental health condition caused by substance misuse, use services more and cost the NHS more”.

There have been numerous compelling studies investigating the cost of care amongst individuals with dual diagnosis. For instance, studies conducted in USA by Bartels et al. (1995) & Kivlahan et al. (1991) and another study conducted in UK by McCrone et al. (2000) have all reported that those patients with problematic substance use disorder and a psychosis had made use of mental health services significantly more than non-users. However, an important point which they all stated that the total costs between users and non-users did not produce statistically significant results. On the other hand, one must be very cautious when interpreting these results. The number of patients used in these studies was relatively small and therefore one must be vigilant when generalizing the findings. In contrast, in another study carried out by Laugharne et al. (2002) investigated the link between alcohol use and cost of treatment in patients with a diagnosis of psychosis. 708 patients were examined over 2 year period. The researchers concluded that “there was no difference in mean cost of care between non-drinkers, moderate and heavy drinkers. Increased alcohol consumption was associated with lower cost of care. Heavier alcohol consumers did not incur more treatment costs and may, indeed, cost less than other patients. This may reflect higher levels of drinking being associated with better overall functioning, poor engagement with services or exclusion from services”. These results are somewhat inconsistent with those of McCrone et al. (2000) study where they examined costs of individuals who have a dual diagnosis of psychosis and substance use disorder compared with those with a diagnosis of psychosis only. They found that dual diagnosis patients had significantly higher ‘core’ psychiatric service cost, a difference of £1362 compared to those patients without a substance use disorder. However, when all services were examined the total cost was not statistically different. Consistent with these results, a compelling study by Bartels et al. (1993) has found that those patients with existing substance use problems had considerably greater service use and cost of institutional services. However, there were no statistically different findings in terms of cost of other services i.e. psychosocial rehabilitation, housing supports and outpatient treatment.

Another important issue in terms of cost that needs to be distinguished is between the cost of inpatient admissions when in crisis and cost of home treatment teams when in crisis. It is crucial to know the different costs between these two care approaches. One of the reasons being that if home treatment teams can offer the same kind of treatment as inpatient admission to individuals at their homes, this can significantly reduce the total cost of care and therefore cost less to mental health services. Evidently, “economic modelling estimated that an acute mental health service making full use of crisis resolution home treatment team services in appropriate cases costs approximately £600 less per crisis episode than one in which crisis resolution home treatment team is not available- chiefly because some admissions will be avoided altogether and others will be shorter, reducing the costs incurred with overnight stays” (cited from National Audit Office, 2007). Furthermore, if we break down the individual costs of inpatient admissions and home treatment team contacts we can see a great amount of differentiation between the two. According to Curtis and Netten, (2006) the cost of crisis resolution home treatment team (CRHT) effort per patient is approximate at £244 per week which corresponds to £976 per month. The cost per day for a home treatment team also differs in terms of input. It is estimated that high input (at least daily contact with CRHT teams) is £43 and low input (less than daily contact with CRHT team) is £21 (National Audit Office, 2007). In sharp contrast, the average cost of inpatient admission per day is estimated to be £201 (National Audit Office, 2007). As a result, these figures reflect how important it is to consider alternative treatment option to inpatient admission and therefore this can hugely reduce the cost to mental health services and society as a whole.

[1] “A first episode psychosis refers to the development of psychotic symptoms during the early phase of schizophrenia or affective disorders before the individual has received adequate trials of treatment” (Archie et al. 2005 cited from Archie et al. 2009).

[2] “A crisis resolution team (CRT, sometimes called home treatment) provides intensive support for people in mental health crisis in their own home, and stays involved until the crisis is resolved. It is designed to provide prompt and effective home treatment, including medication, in order to prevent hospital admission and provide support to informal carers” (cited from Crisis Resolution and Home Treatment in Mental Health, ed. Sonia Johnson, Justin Needle, Jonathan P. Bindman & Graham Thornicroft, 2008).