As with most mental health conditions, the sooner therapeutic interventions are delivered, the better the outcome for the individual. Eating disorders (EDs) affect around 9% of individuals worldwide (National Association of Anorexia Nervosa and Associated Disorders, 2021). This number could be higher, accounting for undiagnosed or ‘more mild’ disordered eating troubles.

Though undoubtedly a highly stigmatised diagnosis, EDs are far more common, leading to far higher mortality rates than you or I may think. In reality, evidence suggests that EDs have the highest mortality rate of all psychiatric disorders (Arcelus et al., 2011). Furthermore, it is estimated that around 25% of individuals battling anorexia nervosa are male, yet are often not taken seriously until much later, increasing the risk of harm (Mond et al., 2013). Research has highlighted the limitations of the current literature on the topic, citing a limited scope, a large focus on female participants, and aetiology (Kesheviah et al., 2014; Rikani et al., 2013; Smink, van Hoeken & Hoek, 2012), among others, to be the issue.

Iwajomo et al. (2020) recently conducted a retrospective population-based cohort study of people living in Ontario, Canada, which focused on the excess mortality associated with EDs. While the study design is certainly a mouthful, it is vital for such research to take place to limit methodological biases and increase the accuracy of information. The retrospective aspect of the research utilised healthcare data to form and analyse the cohort, which comprised around 19,000 individuals, starting from January 1st, 1990, to December 31st, 2013.

The study aimed to explore all-cause mortality rates in these individuals, accounting for other comorbidities such as congestive heart failure, obstructive pulmonary disease, diabetes, asthma, HIV, and cancer. The level of rurality or urbanicity was also considered and cross-referenced. Mortality rates were estimated relative to the general Ontario population and reported via standardised mortality ratios (SMRs) and potential years of life lost (PYLL).

Methods

The study’s inclusion criteria were individuals within the Ontario area who had received a hospital-based diagnosis and treatment and at least one diagnosis of anorexia nervosa (AN), bulimia nervosa (BN) and/or eating disorder not otherwise specified (EDNOS). These diagnoses had to take place between 1990 and 2013. Furthermore, the authors split the data by age of entry as 10 to 44 years of age and those 45 or older. Using coding from diagnostic manuals (ICD-10, ICD-9, and DSM-IV), the authors were able to identify eating disorders (EDs) within the population whose information was made available from the following administrative healthcare databases:

• International Credential Evaluation Services (ICES), an independent non-profit organisation funded by the Ontario Ministry of Health
• Registered Persons Database (RPDB), a population-based registry
• Ontario Office of the Registrar General (ORG) – Vital Statistics Agency, providing relevant information on mortality if applicable
• Canadian Institute for Health Information (CIHI), an organisation that provides maintenance of databases that include hospital-based information and diagnostic codes.

Excess mortality rates were shown through standardised mortality ratios (SMRs), potential years of life lost (PYLL), as well as amount relative to an ED. Mortality was also outlined by age, gender, overall, calendar year and deaths per 1000 person-years of observation time.

Additionally, sociodemographic factors for all-cause mortality were considered and statistically analysed using Cox proportional hazard regression models. Factors included were age, gender, calendar year (1990-2013), neighbourhood level household income, and the degree of urbanicity.

So, what does all this actually mean? Well, the authors collected data from healthcare databases and estimated rates of mortality, how many years of life were lost in general, and also due to battling an eating disorder. Then they ran statistical tests to determine what the correlation of mortality and EDs was, with other variables, such as housing, age, income, and gender.

Results

Descriptive statistics of the cohort

• Out of the 19,041 participants, 89.9% were female (n=17,108) and 10.1% were male (n=1,933)
• 7% of the cohort lived in cities and other urban areas
• Most of the cohort (88.9%) was between the ages of 14 and 44
• Around half of participants lived in middle- to low-income neighbourhoods.

Eating disorder-specific statistics

The authors found that some individuals had multiple diagnoses of eating disorders (EDs) over time, as well as comorbidities with other long term health conditions:

• 9% of the cohort had anorexia nervosa (AN) alone, or in combination with bulimia nervosa (BN) or eating disorder not otherwise specified (EDNOS)
• Yet, 37.1% had EDNOS alone.

Comorbidities with physical health conditions included:

• 5%: asthma
• 1%: diabetes
• 5%: cancer
• 5% congestive heart failure.

Mortality rates

• All-cause mortality rates for the entire cohort were much greater than they expected, when compared with the general population. In fact, people with EDs diagnosed in hospital settings had (on average) mortality rates that were five times higher that of the general Ontario population
• The standardised mortality ratios (SMRs) were higher for adults between the ages of 30 to 44 and were overall higher for males compared to females:
  • SMRs for males 7.24 [95% CI 6.58 to 7.96]
  • SMRs for females 4.59 [95% CI 4.34 to 4.85]
• Similarly, the potential years of life lost (PYLL) rates were similar for each gender separately, though overall were expected to be almost 30 years per 1,000 people (29.54 [95% CI 28.7 to 30.4])
• Increasingly worrisome, the estimated Years of Life Lost (YLL) were 375.6 and 174.1 for males and females, respectively
• Finally, the excess PYLL for EDs in particular was projected at 84%

To highlight the magnitude of these numbers, I pulled this direct quote from the paper:

Among all people with eating disorders throughout Ontario over the period of this study, it was estimated that 24,773 years of life were lost due to an eating disorder before the age of 75.

Sociodemographic factors

Male gender and being of older age were associated with an increased risk of mortality, as well as higher hazard ratios for those living in the highest- and lowest-income neighbourhoods. This could illustrate that discernible differences within mortality risks alter, dependent on income level, highlighting yet another instance of income-related inequality within society. Interestingly, results indicated that those living in rural environments under the age of 45 had a lower mortality risk, yet those over 45 had a higher risk.

Additionally, control variables of long-term ailments, such as cancer, HIV, and congestive heart failure also showed to increase risk of mortality. Conversely, conditions such as asthma and diabetes showed the reverse. Finally, a crucial result of the study was that, throughout calendar years, mortality appeared to have a downward trend, which the authors attributed to potential improvement in living standards.

Conclusions

The study’s main objective was to estimate rates of burden and excess mortality in individuals experiencing eating disorders (EDs). This was one of the few studies on an international scale to use population-based data, as well as to report SMR estimates for both male and female genders. As a recap, the study concluded with the following findings:

1. Those with eating disorders diagnosed in hospital settings showed to have a mortality rate five times that of the general Ontario population
2. For both males and females, more than 80% of life was lost before the age of 75
3. Mortality rate for females was 5-fold in comparison to the general population
4. Additionally, the mortality ratio for males (SMR) was significantly higher than females

Strengths and limitations

Strengths

With a focus on anorexia nervosa (AN), bulimia nervosa (BN), and eating disorder not otherwise specified (EDNOS), the authors effectively filled a gap within the literature. This was further achieved through the inclusion of males and females, setting up a population-based study, and the use of a comparison group, something that is uncommon. Furthermore, their methodology appears robust and reproducible, and didn’t fall in the trap of cohort selection bias. Finally, this study presents findings in a way that highlights the gravity of the problem of excess mortality within these disorders, urging those in the field to improve health outcomes.

Limitations

While this research certainly made important contributions to the literature, it would have been interesting to see similar research on more types of EDs, for instance binge eating disorder (BED), or orthorexia, among others. Understandably so, the authors cite their reasoning for the focus on AN, BN and EDNOS to data limitations, where the databases used had information on these three disorders specifically. Therefore, as mentioned themselves, a considerable limitation could be the greater application of the findings to other EDs. A further limitation to consider is lack of outpatient data included, as they drew their information from administrative hospital-based databases, again, pointing to wider applications.

Implications for practice

It would be interesting to see other researchers follow the footsteps of Iwajomo et al. yet take this work a step further. In other words, to reproduce the findings in other parts of the world, potentially with different economies or sociopolitical backgrounds to that of Canada. Additionally, by bringing to light the level of mortality surrounding EDs, the authors have (hopefully) opened the door to a new narrative. Further interventions can be formulated and applied to target ED-specific mortality, to minimize years of life lost, level of burden, and increase quality of life. Finally, policy changes could be made to implement psychoeducation and training programs at school and professional-based levels.

Statement of interests

None.

Links

Primary paper

Iwajomo, T., Bondy, S. J., de Oliveira, C., Colton, P., Trottier, K., & Kurdyak, P. (2020). Excess mortality associated with eating disorders: population-based cohort study. The British Journal of Psychiatry, 1–7. https://doi.org/10.1192/bjp.2020.197

Other resources

Arcelus, J., Mitchell, A. J., Wales, J., & Nielsen, S. (2011). Mortality rates in patients with anorexia nervosa and other eating disorders. Archives of General Psychiatry, 68(7), 724. https://doi.org/10.1001/archgenpsychiatry.2011.74

Iwajomo, T., Bondy, S. J., de Oliveira, C., Colton, P., Trottier, K., & Kurdyak, P. (2020). Excess mortality associated with eating disorders: population-based cohort study. The British Journal of Psychiatry, 1–7. https://doi.org/10.1192/bjp.2020.197

Keshaviah, A., Edkins, K., Hastings, E. R., Krishna, M., Franko, D. L., Herzog, D. B., Thomas, J. J., Murray, H. B., & Eddy, K. T. (2014). Re-examining premature mortality in anorexia nervosa: A meta-analysis redux. Comprehensive Psychiatry, 55(8), 1773–1784. https://doi.org/10.1016/j.comppsych.2014.07.017

Mond, J., Mitchison, D., Hay, P., Cohn, L., & Lemberg, R. (Eds.). (2013). Prevalence and implications of eating disordered behavior in men. In Current Findings on Males with Eating Disorders (p. 195). Taylor & Francis.

National Association of Anorexia Nervosa and Associated Disorders. (2021, March 3). Eating Disorder Statistics | General & Diversity Stats | ANAD. Retrieved August 15, 2021, from https://anad.org/get-informed/about-eating-disorders/eating-disorders-statistics/

Rikani, A. A., Choudhry, Z., Maqsood Choudhry, A., Ikram, H., Waheed Asghar, M., Kajal, D., Waheed, A., & Jahan Mobassarah, N. (2013). A critique of the literature on etiology of eating disorders. Annals of Neurosciences, 20(4). https://doi.org/10.5214/ans.0972.7531.200409

Smink, F. R. E., van Hoeken, D., & Hoek, H. W. (2012). Epidemiology of Eating Disorders: Incidence, Prevalence and Mortality Rates. Current Psychiatry Reports, 14(4), 406–414. https://doi.org/10.1007/s11920-012-0282-y

Photo credits

• Photo by Scott Walsh on Unsplash
• Photo by Annie Spratt on Unsplash
• Photo by Ben Allan on Unsplash
• Photo by youssef naddam on Unsplash