Cigarettes, alcohol and cannabis are each used by hundreds of millions of people for their psychoactive properties. Someone might wish to drink away their sorrows after a bad break-up, try to relax after a stressful exam week by getting high, or light a cigarette because they think it will help them concentrate. While this sort of substance use is commonplace, we also know that smokers, heavy alcohol users and cannabis users are at higher risk of mental illnesses, such as a depressive or anxiety disorder. Determining why there is an association between substance use and mental illness is more difficult than it seems. But recent research is helping us to do just that, and it could also help improve the prevention and treatment of such conditions.
Science has already shown that life events that can increase the risk of mental illness – such as childhood trauma or the unexpected death of a loved one – can also lead to increased substance use. While this insight is useful, another potential explanation for the relationship between substance use and mental illness is that one makes the other more likely – and if we know which causes which, we can intervene. Many people intuitively think that symptoms of mental illness cause someone to use (more of) a substance. This is reflected by the anecdotal examples above, as feelings of depression, anxiousness and fatigue can be reasons for using. But the reverse might also be possible: that using psychoactive substances leads to mental illness.
Normally, to identify whether one behaviour or condition causes another, scientists would conduct a randomised controlled trial. Say that we want to test whether smoking causes depression. People could be randomly assigned to groups, with one instructed to smoke a pack of cigarettes a day while another is told not to smoke at all. The groups would then have to be followed for a long period of time to see whether those in the smoking group develop depression more often than those in the non-smoking group. Of course, you can see why instructing people to smoke would be ethically problematic. Alternatively, we could follow a cohort of people for years to see how substance use and mental illness naturally tend to relate to each over time. But there is a catch here as well: the factors that make people more susceptible to both mental illness and substance use (such as early trauma) could make it appear as if there’s a cause-and-effect relationship between mental illness and substance use, when in fact those other factors explain their co-occurrence.
Do we need to accept that the question we want to answer is just too difficult? Well, maybe not. In recent years, a new scientific method has taken over the field of causal inference. A person’s propensity for substance use and risk for mental illness are partly determined by their genes, and scientists have identified many of the responsible genetic variants. While overall genetic risk is shaped by a very large number of genetic variants, some have a particularly strong effect. For substance use, we have a relatively clear idea of what the strongest genetic variants do. For instance, the most important genetic variant for smoking is in a gene that determines the number of nicotine receptors in your brain. If you have an A variant instead of a G variant, you will have fewer nicotine receptors, which means you are less sensitive to the initial neurological effects of smoking. It will allow you to smoke more cigarettes more quickly, making it easier to become addicted. Many of the genetic variants associated with problematic alcohol use are situated in genes that code for enzymes that help you break down alcohol in your blood.
Amazingly, these genetic variants can aid our quest to unravel the relationship between substance use and mental illness. The key to this is the fact that the genetic variants that you inherit from your mother and father are randomly assigned at conception. So, whether you receive a genetic variant that gives you a higher risk for a certain trait or one that gives you a lower risk is determined by chance. One group of people in the population will have the A variant of the smoking gene, putting them at higher risk of smoking than another group with the G variant. Because these genetic differences have occurred randomly before birth, they are independent from other influences. In a group of people with a higher genetic risk of smoking, we will find that lung cancer occurs more than average. Similarly, if we want to know if smoking causes depression, we could test whether a higher genetic risk of smoking comes with a higher risk of depression.
The method described here is called Mendelian randomisation. In a systematic review, my colleagues and I evaluated 63 Mendelian randomisation studies on the relationship between substance use and mental illness to determine what we have learned so far.
One important conclusion of this work is that symptoms of mental illness can lead to excessive alcohol use, while the reverse – alcohol use causing forms of mental illness such as anxiety, depression and bipolar disorder – was not evident. It seems that alcohol is indeed used as a type of ‘self-medication’ – to numb feelings of anxiety or depression, for instance. While clinical observations indicate that mental health improves in individuals who successfully cut down on their problematic alcohol use – which, among other reasons, makes doing so a worthy goal – the Mendelian randomisation findings suggest that alcohol use is typically not an underlying cause of these kinds of mental illness.
Interestingly, the findings for smoking were different. There was some evidence that mental illness can cause people to smoke more. However, there was much stronger evidence that smoking increases the risk of developing mental illness, particularly the risk of depression, bipolar disorder and schizophrenia. A plausible explanation for this is that toxic compounds from inhaled cigarette smoke cause inflammation and oxidative stress in cells throughout the body, processes that are known to predict mental illness.
Another relationship of particular interest is that between cannabis use and schizophrenia. While cannabis is often mentioned as a risk factor for developing psychosis and schizophrenia (a psychotic disorder), convincing proof of the causal nature of this relationship was previously lacking. Mendelian randomisation studies now indicate that, similar to smoking, there is a two-way relationship: cannabis appears to increase the risk of schizophrenia, but the reverse is also true, in that those liable to schizophrenia are more likely to start using cannabis.
While it is a powerful method, Mendelian randomisation works well only if certain strict assumptions are met. Whether this is always the case is questionable. The most important pitfall is that a lot of genetic variants are pleiotropic, meaning they influence not one or a few traits, but many. This can be a challenge for Mendelian randomisation. Say that the genetic variant that increases people’s smoking risk also happens to play a direct role in the development of depression. This means that, if we find an effect of that smoking-related variant on depression, it might not be due to smoking. While the studies that led to the robust findings described above extensively checked and corrected for such pleiotropic effects, some other studies that we reviewed did not. More studies that account for these and other issues – and comparisons with findings from different types of research methods – are needed to make conclusions with greater certainty.
Overall, however, these studies provide important new insights, showing that the relationship between substance use and mental illness is even more complex than previously thought. Particularly impactful is the finding that smoking could lead to mental illness. Healthcare professionals might hesitate to raise the topic of quitting smoking with patients who have a mental illness; smoking can seem like the least of their problems. Our review indicates that this needs to change, as quitting smoking might actually help patients recover. The evidence of Mendelian randomisation studies dovetails with carefully conducted longitudinal studies, which suggest that quitting smoking could improve mental health and cognitive functioning. Going forward, compassionately helping patients with a mental illness to quit smoking should be a priority.
Moreover, we need to try even harder to reduce smoking in the population at large, with smoking rates around the world remaining high. It will be particularly important to prevent smoking among youth with a high risk of poor mental health, as smoking might act as a trigger for developing mental illnesses. Maybe we can strengthen public health messages by warning people that smoking not only causes diseases such as lung cancer, but could also put them at increased risk of mental illness. This might help us kill two birds with one stone: reduce a deadly habit and promote mental health.
