In this nationwide cohort study, we found that TSCI is associated with a higher IRR of autoimmune diseases, particularly those affecting the nervous system. Additionally, we provide an overview of the associations for various autoimmune diseases, grouped primarily by the affected organ.

Although the biological pathway is not fully understood, a possible biological sequence from TSCI to chronic neuroinflammation, leading to secondary neurologic injury, the activation of neuro-autoreactive lymphoid cells, and ultimately autoimmunity might explain our findings [4, 5, 7, 10, 13]. The strong association of neurologic autoimmune diseases with TSCI aligns with the biological effects observed in previous experimental and observational studies.

Following TSCI, the initial immunological response involves polymorph nuclear myeloid infiltration of the lesion site, followed by M1 and M2 macrophages within the first hours to days [5]. Subsequently T-cells, especially CD8+ T-cells, are present, peaking early and persisting at about 10% of the peak concentration [5, 10]. B-cell expansion, maturation, and ectopic lymphoid follicle formation occur in the injured spinal cord, producing autoreactive antibodies such as anti-myelin basic protein (anti-MBP) [6, 7]. This process resembles the pathology seen in multiple sclerosis[16]. Recent human TSCI studies quantify the immunological response aligning with earlier experimental data, particularly the non-resolving aspects of the inflammatory cascade that may lead to central nervous system autoimmunity [13, 14]. Our study underscores the role of anti-MBP in secondary morbidity among TSCI patients, highlights its association with multiple sclerosis and Idiopathic polyneuropathy, both characterized by a pathological demyelination [5, 6, 8, 14, 16].

The development of Systemic and Dermatologic autoimmune diseases might be supported by the findings of systemic anti-nuclear antibodies (ANA) and anti-DNA (dsDNA) antibodies following TSCI, similar to those seen in rheumatism, systemic lupus erythematosus and autoimmune dermatologic diseases [7, 8]. These studies, along with our findings, suggest a true biological pathway where the combined innate and adaptive immune response to TSCI can lead to the maturation and clonal expansion of autoreactive T- and B-cells, ultimately leading to autoimmune disease.

That the initial increased risk of Gastroenterologic and DM-1 conditions, nullified in the Charlson comorbidity index and fully adjusted models, may indicate the influence of other factors on those associations. The impact of Charlson comorbidity index and hospitalization for infections, possibly arising from SCI-IDS and autonomic dysreflexia, emphasize their implications for multiorgan disease beyond autoimmune diseases.

SCI-IDS and inherited infectious susceptibility could mediate the pathway from TSCI to autoimmune disease [12]. Due to the small number of individuals with TSCI and an autoimmune disease, stratified analysis for formal mediation-analysis was not feasible [28]. Instead, we adjusted for hospitalization for infection to highlight its importance and the implicit risk, potentially modifiable pharmacologically. Interpretations should be cautious; we believe that only the direction, not the effect size, can be evaluated. The consistent direction of effect estimates suggests validity but does not rule out a potential collider bias.

We aimed to minimize misclassification of co-variates, suspecting that autoimmune diseases in the TSCI group could introduce detection bias or reversed causality, leading to differential misclassification. To mitigate this risk, we censored individuals diagnosed with TSCI and autoimmune disease during the same hospitalization, counting their risk time as non-TSCI, but excluding them as autoimmune disease cases. To avoid the confounding effects of multiple autoimmune diseases, we evaluated only an individual’s first autoimmune disease [19]. Despite these efforts, the multimorbidity and frequent hospitalizations of TSCI patients could lead to residual detection bias.

Table 3 suggests that thoracic TSCI predominantly increases the risk of autoimmune disease compared to a cervical TSCI. Initially, we hypothesized that cervical TSCI would be associated with a higher risk of autoimmune disease based on findings that higher injury levels amplify SCI-IDS and autonomic dysreflexia [3, 11]. Experimental studies have shown lesions above T9 reduce preganglionic inhibitory activity of postganglionic sympathetic nerves to the spleen, causing leukopenia and enhancing SCI-IDS [3, 11]. Similar findings apply to autonomic dysreflexia [3].

Our results indicate that non-penetrating thoracic TSCI might require more energy, leading to greater neurologic trauma and collateral damage, potentially explaining the higher risk of autoimmune disease despite less SCI-IDS and autonomic dysreflexia. Additionally, thoracic TSCI might result in more significant blood-spinal cord barrier disruption due to a higher probability of American Spinal Injury Association Impairment Scale (AIS) A injuries than cervical TSCI. Breakage of the blood-spinal cord barrier seems to be more prevalent and more pronounced in thoracic TSCI too [14]. Therefore, our findings could be more influenced by the magnitude of trauma than the level of TSCI.

Our definition of TSCI and the incidence (see Supplementary 5) aligns with Bjornshave Noe et al., report from the Spinal Cord Injury Centre of Western Denmark from 1990–2012 [29]. Our annual incidence data from 1981–2012 show realistic and relevant variation, but the increase from 2013-2018 might reflect misclassification. Table 1 shows the accumulated person years by calendar period. The TSCI population´s flattening curve and the steady state of the non-TSCI population in the study´s later period suggest possible bias from incidence issues in 2014–2018. Misclassified TSCIs may not have a higher autoimmune disease risk, potentially increasing type 2 error risk, but such bias cannot fully explain the association.

Our register-based study has limitations including small sample sizes in groups of TSCI patients with outcomes, and the inability to adjust for lifestyle factors like smoking, alcohol, body mass index, and familiar co-occurrence of autoimmune disorders. The positive predictive value of included variables is partially known and we cannot rule out misclassification from non-validated variables, though Danish registries are generally high quality [23]. The DM1 diagnosis has a high positive predictive value but unknown sensitivity, though a high sensitivity is expected given the clinical course [30].

The study population included all Danes born in Denmark, with complete public registration in the Danish Civil registration system, minimizing selection bias [22].

Several trauma-related factors including immediate treatment, pre-hospital vital parameters, pre-hospital transport time, initial AIS, decompressive surgery, and primary spinal cord tissue vitality could not be evaluated. These factors could illuminate the effect of secondary neurologic injury and inflammatory activity on autoimmune risk. The Danish TSCI demographics as described by Noe et al. [29] indicate that sharp and penetrating TSCI is rare, limiting our findings to blunt trauma TSCI populations.

Studying rare diseases risks large association deviations from small count variations. We advise the reader to consider raw counts, rates, and the full spectrum of estimated confidence intervals and associations. Based on our results, clinicians should monitor for potential late sequelae of TSCI. Prevention and medical optimization might reduce the Charlson comorbidity index´s impact and infection-related hospitalizations until further controlled studies are available.