Long term follow up studies are as rare as hen’s teeth, but are an important method for observing population health. Despite their limitations they provide intelligence that go beyond the usual snapshot revealed in many studies. This study from Sweden led by Mats Fridell had a simple aim, what impact does substance use have on the life expectancy of people with co-morbid mental health and substance use problems?

Those of us working in either the mental health or the substance use field know that many of the people we work with die prematurely. So as the rest of the population has experienced extended life expectancy in recent decades, this achievement has not been shared by people who have comorbid problems.

Although it’s useful to have a general idea of how these combined problems influence premature death, we need something more specific to have any utility in practice. Assessing which substances people used is critical detail, we already know that opiates are far riskier than cannabis in terms of fatal overdose.

Methods

The study used data from a cohort of 1,405 patients who were admitted to an inpatient specialist detox unit in Sweden between the 1970’s through to the 1990’s. Forensic toxicology reports were used to determine whether a death was due to drugs. The criteria used to define a drug-related death was established prior to the study. Psychiatric records were used to check for a mental health diagnosis. They used 4 broad headings to organise these:

1. Severe mental illness
2. Substance-induced psychosis
3. Other mental disorders
4. No mental illness

Comparisons of mortality were made between the general population and the cohort. Sex and age co-variates were controlled for in the statistical analysis.

Reflecting typical gender ratios found in the United Kingdom, the Swedish cohort comprised 70.1 % men and 29.9% women.

Results

What we now describe as adverse childhood experiences (ACEs) were identified in the majority of the cohorts histories. First exposure to substances occurred around the age of fifteen. Cannabis, alcohol or amphetamines were the most commonly used drugs. Psychiatric comorbidity was estimated in 33.4% of the cohort.

The most shocking finding is that premature mortality was six times that found in the general population. Almost half of those in the cohort that used substances died prematurely.

Being male and older at the point of treatment admission increased the risk of death. As with the UK, opiates and alcohol were the most frequently identified substances at the point of death. Whereas amphetamine and cannabis use were associated with a lower risk of premature mortality.

Drilling into the specific way that individuals died, overdose was associated with opiate use and problems with mood/anxiety. A mix of illicit and prescribed drugs was associated with death caused by intoxication particularly for those with psychosis who used alcohol and sedatives. Surprisingly, those in the cluster of substance-induced psychosis had the lowest risk of premature death, lower even than those without a comorbidity.

An all too familiar finding was the poor provision of substitute treatment given to the opiate dependent patients in this cohort. This is really disappointing as the evidence suggests that substitute prescribing particularly at optimal doses protects individuals from many risks including mortality (Mattick, 2009).

Limitations

The authors are candid about the limitations urging caution in relation to interpreting causal relationships when using an observational cohort study design.

However there are broader problems which are beyond the authors control. For example there is some evidence suggesting that coroners are less likely to investigate unnatural female deaths compared to males (McLean 2017). If this cultural bias is at play in Sweden then female drug related deaths could have been under-counted.

Discussion

• This study like many others recruited patients from treatment, it is sobering to remember that the majority of people with comorbid problems don’t make it into treatment. So this study provides just a glimpse of the mortality associated with combined drug use and mental health.
• There are many times we don’t have reliable evidence to help inform workers and people with problems, but this is not the case for drug related deaths.
• We know how to reduce these deaths, the evidence is clear. For example:
  • providing access to naloxone for those at risk of opiate overdose,
  • providing optimal dosing of substitute drugs and so on.
• So if we have the know-how, something else is standing in the way of reducing these premature deaths. I wish I knew what it was, and so do the family and friends of those who die prematurely.

Conflicts of interest

None.

Links

Primary paper

Fridell, M., Bäckström, M., Hesse, M., Krantz, P., Perrin, S. and Nyhlén, A., (2019) Prediction of psychiatric comorbidity on premature death in a cohort of patients with substance use disorders: a 42-year follow-up. BMC psychiatry, 19(1), p.150. https://bmcpsychiatry.biomedcentral.com/articles/10.1186/s12888-019-2098-3

Other references