People given a diagnosis of “personality disorder” (PD) are often at the bottom of the pile when it comes to mental health research. This is a group of people who are highly stigmatised by both professionals and the public, and whose health is under researched (Lam D. et al, 2015; The Lancet, 2016). They have often experienced significant trauma, as well as adverse social and environmental factors, but they also have much worse physical health (Mind, 2018). People given a PD diagnosis die an average of 18-19 years before the general population, not only because of suicide, but because of things like cardiovascular disease and obesity (Fok M. et al, 2012; Björkenstam E. et al, 2015).
Previous elves have written about the recent Royal College statement on PD (Harding K, 2020), but there hasn’t been an awful lot of writing about people’s physical health, apart from a couple of insightful blogs on prenatal stress and alcohol use disorder (Sri A, 2020; Connolly D, 2019).
Moving to the realm of physical health, another topic rarely discussed in elf blogs are immune-mediated inflammatory diseases (IMID), which includes things like inflammatory bowel disease (IBD), multiple sclerosis (MS) and rheumatoid arthritis (RA).
IMID gives us an opportunity to look at how inflammatory disease might play a role in mental illness, looking at that link between physical and mental health. Biological aetiologies in psychiatry seem to be an increasing focus of research, and so this is a good topic to explore if everything is just down to inflammation and that pesky C-reactive protein (CRP). Add in PD, IMID, RA, etc and we’re set up for an exciting game of elvish scrabble. But where shall we play?
Manitoba is a Canadian province that has produced some of the world’s best speed skaters, and would also score 12 scrabble points, so let’s slide over there and see what they have to say on the topic. Blaney et al (2020) used Manitoba’s population administrative database to look at the association between any personality disorder and three IMIDs (IBD, MS and RA).
The authors conducted a retrospective cohort study comparing 20,000 people with IMID to 100,000 matched controls. The exposure was having an IMID and the primary outcome was incidence of any PD (there are different types, but the authors looked at the umbrella category). The study period is just a bit shorter than the Star Wars franchise, running from 1984 to 2013, with a median follow up of around 13 years.
All of the data is taken from Manitoba’s population administrative database, which includes any individual eligible for health services (Dr Google says this includes anyone living there permanently more than 6 months a year).
Individuals with IBD/MS/RA were identified using validated measures based on their first contact with health services. The authors included these three IMIDs as they are some of the most common and most debilitating conditions. Readers will be familiar with some others that weren’t included, such as lupus, psoriasis and ankylosing spondylitis.
“Personality disorder” diagnoses were identified using hospital and physician records, documenting the relevant ICD codes. The authors recognise that remission is possible and so only counted people presenting to services within each year in their calculations for period prevalence.
The confounders in this study were sex, age, socioeconomic status, region (urban or rural), and number of visits to hospital for a non-psychiatric disorder.
Some of the analysis looked at incidence of PD in the 5 years before and after an IMID diagnosis, and so for those measures 10 years of continuous data were required, giving slightly smaller numbers: 12,000 cases and 65,000 controls.
The authors did a regression analysis (a statistical technique to control for the effect of other explanatory variables) and results were age and sex standardised to the Canadian population, as well as adjusted for covariates.
The crude rates of “personality disorder” PD were higher in the immune-mediated inflammatory diseases (IMID) group, with a lifetime prevalence of 4.72% PD, compared to 3.10% in the controls. This relationship remained after adjusting for covariates.
The results showed a 72% increase in the incidence rate of PDs in the group with IMID.
Adjusted incidence rate ratios (IRR) of “personality disorder” PD are presented below:
- All immune-mediated inflammatory diseases (IMID): IRR 1.72 (95%CI: 1.47 to 2.01)
- Inflammatory bowel disease (IBD): IRR 2.19 (95%CI: 1.69 to 2.84)
- Multiple sclerosis (MS) : IRR 1.79 (95%CI: 1.29 to 2.50)
- Rheumatoid arthritis (RA): IRR 1.61 (95%CI: 1.29 to 1.99)
In the smaller sample with 5 years of data before and after IMID diagnosis, incidence of PDs was increased in the 5 years before diagnosis of IMID.
- Lower socioeconomic status and urban residence were associated with an increased incidence of PDs
- Mid to older adulthood (age 45–64) was associated with overall decreased incidence of PDs
- Sex was not a predictor of PDs in this study
The authors made some suggestions to explain their results.
- The experience of having an immune-mediated inflammatory disease (IMID) could itself have an impact on your personality and regulation of emotions
- There are probably a number of shared risk factors. A history of trauma (for example past sexual abuse and violence) has been associated with “personality dysfunction” and a number of IMIDs
- Inflammation (potentially caused by chronic stress exposure) could contribute to both conditions
- There might be shared genetic markers that contribute to both conditions.
They also gave their own conclusions fairly succinctly:
In summary, persons with IMIDs are at an increased risk of a PD, regardless of the specific IMID. Elevated comorbidity rates may relate to shared risk factors between IMID and PDs but this requires further investigation.
Strengths and limitations
The major strength of a study like this is its sample size. With 20,000 cases and 100,000 controls, the numbers are big and mean we can be confident they are reflective of the Manitoban population. The effect size appears reasonable and confidence intervals for these are well above 1, which means the numbers are statistically significantly.
However, in any study using population level data there are some limitations; often that the data is a mile wide, but an inch deep.
The number of cases of “personality disorder” (PD) may have been under-diagnosed in both cohorts. The authors themselves say they may have missed cases diagnosed by non-medics (for example a psychologist). PD diagnoses aren’t always well recorded in patient’s notes, and so more cases can be picked up by using computer software which searches free text as well (Fok M. et al, 2014).
To be included in period prevalence data the authors required ‘at least one contact with services in a given year’. Given that limited services are available for people given a personality disorder diagnosis (at least here in the UK) this may underestimate the prevalence, as people may not engage with services on an annual basis.
The authors did not look at many different sociodemographic characteristics; their measure of socio-economic status was based on home address, and is fairly crude (similar to using Index of Multiple Deprivation in the UK). A major omission appears to be the ethnicity of participants, which a) neglects to look at the varying impact of illnesses on different groups, and b) makes it hard to judge how well the results can be generalised to other populations. It’s not clear if this was a limitation of the dataset, or a choice by the authors, but I would expect most papers in 2020 to consider this.
Finally, the authors themselves say that “caution should be applied to interpretation regarding the individual IMID groups, given some small [group] sizes”. And it’s worth noting they only looked at three specific IMIDs.
The authors have some links to pharmaceutical companies (e.g. serving on advisory boards, but not funding this research). I don’t think this has added much bias to the work, but think it’s always worth mentioning.
Implications for practice
This study adds to the body of evidence linking physical and mental health, and adds to the idea that there may be shared risk factors (or even shared inflammatory pathways) causing both IMID and PD.
In finding an association between the two, the authors suggest taking a more psychological approach to helping patients with IMID (rather than suggesting a more biological approach to PD). They suggest broader support for those with chronic physical health conditions, such as interpersonal effectiveness training, distress tolerance and DBT (dialectical behaviour therapy). They also suggest giving greater attention to PDs in IMID research and in the care of IMID patients.
It’s tempting (but probably a dangerous simplification) to think of potential benefits of anti-inflammatory drugs in supporting people with a diagnosis of personality disorder to manage their symptoms. However, this is not suggested by this study, and those medications can have significant side effects. More research in the area could be helpful, but might be barking up the wrong tree, and it’s perhaps wiser to focus on the things we know can help: psychological support and addressing socio-economic risk factors.
Statement of interests
I am working in a similar field, using routinely recorded patient data to look at physical health outcomes in people given a diagnosis of personality disorder.
Blaney C et al (2020). Incidence and temporal trends of co-occurring personality disorder diagnoses in immune- mediated inflammatory diseases. Epidemiology and Psychiatric Sciences 29, e84, 1–11.
Björkenstam E et al (2015). Excess cause-specific mortality in in-patient-treated individuals with personality disorder: 25-year nationwide population-based study. The British Journal of Psychiatry 207(4), 339-45.
Connolly D. Prevalence of comorbid personality disorder and alcohol use disorder. The Mental Elf, 19 Jul 2019.
Fok M et al (2012). Life expectancy at birth and all-cause mortality among people with personality disorder. Journal of Psychosomatic Research 73(2), 104-7. [PubMed abstract]
Fok M et al (2014). The impact of co-morbid personality disorder on use of psychiatric services and involuntary hospitalization in people with severe mental illness (PDF). Social psychiatry and psychiatric epidemiology 49.
Harding K. Personality disorder: new position statement from the Royal College of Psychiatrists. The Mental Elf, 31 Jan 2020.
Lam D et al (2016). An experimental Investigation of the Impact of Personality Disorder Diagnosis on Clinicians: Can We See Past the Borderline? Behavioural and Cognitive Psychotherapy 44(3), 361-373. [PubMed abstract]
The Lancet Editorial (2016). The health crisis of mental health stigma. The Lancet 387(10023), 1027.
Mind and others (2018) “Shining lights in dark corners of people’s lives” The Consensus Statement for People with Complex Mental Health Difficulties who are diagnosed with a Personality Disorder (PDF).
Sri A. Prenatal stress and personality disorder: is there a link? The Mental Elf, 27 Jan 2020.