This is the first study, to our knowledge, that directly considers an association between autism and CSS. In our large sample of autistic adults, 21% reported an included CSS diagnosis of FMS, ME/CFS, IBS or TMJD and 60% scored at or above the clinical cut-off for a CSS on the CSI, suggesting that CSS symptoms are very common in autistic people.
A factor analysis of the CSI [40] was undertaken to test the measure’s construct validity in an autistic sample, as previous studies have focussed on chronic pain and control groups [39, 41, 53]. The results supported the bi-factor model [53], and a highly internally consistent scale. In our sample, the mean CSI score for those with a diagnosed CSS was 55.3, slightly higher than the mean score of 52.4 Neblett et al. [39] found in their study of CSS patients establishing the clinical cut-off of 40 on the CSI. However, the mean CSI score for autistic participants without a diagnosed CSS was 40.6; a score far higher than that of Neblett’s control group (30.9) and closer instead to the mean score of 40.9 in the non-CSS chronic pain patients. This suggests that CSS symptoms such as pain and fatigue are very common in autistic individuals and possibly more prevalent than in the general population.
Our results also showed that, as predicted, higher scores on the CSI were associated with greater sensory sensitivity, greater anxiety and lower subjective well-being, with females reporting greater sensory sensitivity and scoring higher on the CSI than males. Whilst higher scores on the CSI also appeared to be associated with higher autistic traits, we found that sensory sensitivity and anxiety fully mediated this relationship. Further analysis showed that, just as previously reported in the general population [33, 39], there were clear sex differences, with women over-represented for both CSS diagnoses and severity of CSS symptoms. Women also showed greater sensory sensitivity and reported greater anxiety, depression and lower subjective wellbeing. Previous research into sensory sensitivity in autism has been mixed when considering sex differences [9, 57]. Recent studies on the SPQ, both on data within the Netherlands Autism Register (of which this dataset is a subsample) and outside, found that autistic females had higher sensory sensitivity than both autistic males [42] and non-autistic females [43]. Research within the general population also suggests that females may be more sensitive than males across a range of modalities[34, 58], with hormones thought to play a key role [59]. The path analysis we conducted suggests that sensory sensitivity and anxiety, as well as biological sex by itself, all contribute to CSS symptoms such that autistic females might be more vulnerable to CSS than autistic males and non-autistic females. Given that the non-binary participants in this study seemed to show the same vulnerability, and autistic individuals are less likely to identify with their assigned gender at birth [60–62], gender-focussed research is needed to explore this further.
There are many theoretical reasons why autism and CSS might be linked with each other. Sensory processing differences are a core feature of autism [7] with autistic people reporting greater sensory sensitivity [9] than the general population. People with CSS also experience pain and sensory sensitivity [8] but in this case it is associated with central sensitisation [5]. The multisensory integration mechanisms responsible for the development of sensory sensitivity in chronic pain patients could also explain a relationship in the opposite direction i.e. chronic pain in autistic people with sensory processing differences. Indeed, migraine is included under the CSS umbrella, and is thought by some to be a form of “sensory dysmodulation” [63]. Our results demonstrated that autistic people with greater sensory sensitivity also had more CSS symptoms. Since sensory processing differences tend to be present from a young age [7], we can hypothesise that sensory sensitivity is a risk factor for developing CSS, however further research is needed to explore the direction of this relationship.
Autism and CSS may also be related through neuroimmune and genetic differences, particularly the recently recognised “trifecta” of conditions, including Mast cell Activation Syndrome (MCAS) [64], dysautonomia, including Postural Orthostatic Tachycardia Syndrome (POTS) [65], and joint hypermobility disorders including Hypermobile Ehlers-Danlos Syndromes (hEDS) [66]. These conditions often appear together [67] and are increasingly found to be co-morbid with or underly CSS diagnoses [25]. They have also been recognised to be associated with autism and other neurodevelopmental conditions [20, 22], particularly through the work of Eccles et al [21, 25, 68, 69], and Casanova et al (2019) recently theorised that some forms of autism could even be hereditary connective tissue disorders [70]. Autism and fibromyalgia are directly associated through the FRM1 gene mutation [71] associated with Fragile X syndrome [72], the most common single gene cause of autism.
Our findings suggest that mental health conditions in the autistic population could contribute to CSS symptoms, with anxiety being a significant predictor of CSS symptoms in this sample. Autistic traits alone were not a significant predictor of CSS symptoms, with our path analysis indicating that higher anxiety and sensory sensitivity in the autistic population might explain the higher incidence of CSS symptoms. Existing research supports this theory; high anxiety [73], chronic stress [74] and PTSD [73, 75] have all been associated with CSS, and are also more common in the autistic community [76–79] than the general population. Other psychological factors could provide a basis for a possible link; some research suggests chronic illness severity might be affected by illness beliefs and coping mechanisms [80]. If this is the case, then cognitive differences in autistic people, such as a lower tolerance for uncertainty [81], might affect their experience of CSS symptoms, and this is something that could be explored in future research.
Diagnostic issues might also explain why an association between autism and CSS has thus far been largely overlooked. Historically, autism has been classified as a social and communication disorder [82], with sensory issues only included in the most recent DSM criteria [83]. An autism diagnosis is still predominantly based on behaviour in childhood, with a considerable sex bias such that females tend to be underdiagnosed [36] or diagnosed later [37]. This is in contrast to CSS, which is more often diagnosed in adulthood and where females are more likely to be diagnosed than males [33]. Increased understanding of the lived experience of autism has improved awareness of the many co-occurring health issues autistic people experience [84] but this is not reflected in the current diagnostic criteria [83, 85]. It could be the case that the CSI has captured physical symptoms that have always been common in the autistic population, but not recognised because they were not obvious to the external observer.
Clinically, this study has important implications. We found that the relationship between autistic traits and CSS symptoms was fully mediated by anxiety and sensory sensitivity. Autistic people often struggle to access mental health support or occupational therapy, particularly in adulthood [86, 87]. Our research suggests that increased anxiety and sensory sensitivity could have wider physical health implications, and longitudinal research could explore further whether interventions focussed on these aspects might mitigate the risk of autistic people developing a CSS later in life.
Limitations
A strength of this study is that these data were reported as part of an ongoing data collection in the NAR volunteer register, with participants not explicitly primed as to the aims of the CSI data collection. Therefore, it is unlikely that these findings are inflated due to selection or attrition bias.
In terms of limitations, firstly, although we were able to include a large sample of autistic participants, we did not have a control or CSS only group. Future research including these groups would add additional power and allow for a greater exploration of the relationships between the main variables in the wider population. Secondly, it could be the observed association between autism and CSS relates more to the wording of the questions in the CSI [40] and reflects an overlap of symptoms rather than a true co-occurring condition (e.g. sensory sensitivities, anxiety attacks, child trauma). Further studies could utilise different illness-specific instruments and alternative instruments to the CSI to establish whether a relationship between autism and CSS remains on a condition-by-condition basis. In this study, CSS diagnoses were self-reported and not independently verified. It is possible that participants may have indicated conditions that were not formally clinically diagnosed. Future studies recruiting via health services or incorporating additional questions are needed to confirm these findings. Finally, as a cross-sectional study, this research is limited when exploring cause and effect, such as through the path analyses, and any inferences drawn need to be treated with caution. Longitudinal studies may be able to shed more light on how and why autism and CSS might be related.