Epidemiology of Spinal Cord Injury in Adults in Sweden, 2016–2020: A Retrospective Registry-Based Study

Given the lack of population-based studies for spinal cord injuries (SCIs), global studies have emphasised the need to collect high-quality regional data on traumatic SCIs (TSCIs) and non-traumatic SCIs (NTSCIs) [1‒3]. Few epidemiological studies have been performed on this patient group in Sweden, and recent studies have been limited to one region [4, 5]. These studies have shown that the aetiology of TSCI is almost exclusively limited to falls [4], with an incidence of 19 per million, approximately 200 persons a year in Sweden, which remained stable between 2006 and 2015 [4]. In contrast, the incidence of TSCI in the Nordic countries is estimated to be 15.9 per million [6]. Previous studies in Sweden and the Nordic countries have focussed primarily on TSCI [6‒8]. Updated information on the Swedish SCI population in general, and NTSCI specifically, is lacking. Sweden has a public healthcare system that includes all citizens with a Swedish social security number, providing healthcare for everyone independent of insurance.

During the study period, patients with SCI were treated at the geographically closest rehabilitation unit from among more than 25 different units throughout the country. All of the units reported to the national medical quality registry Svenskt Register för Rehabiliteringsmedicin (SveReh). All patients are included in the register, and no consent is needed because it is a quality register; however, there is the possibility of withdrawing participation.

SveReh [9] was launched in 1997 [10]. The purpose of the register is to safeguard and develop the quality of the given healthcare, compare results between the different units in the country on a group level, and provide knowledge about the health of different patient groups over time [10].

The objective of this study was to use the best available up-to-date data to analyse the incidence and epidemiological characteristics of SCIs in Sweden. To the best of our knowledge, no study has presented data on NTSCI in Sweden. We compare, for the first time, the demographics and outcomes during primary rehabilitation of NTSCI and TSCI in Sweden.

The study utilised data retrieved from SveReh. All rehabilitation medicine units participated during the study period, and the inclusion rate at the hospitals was approximately 90%. Included in the study were all persons who were registered in SveReh during the study period (January 1, 2016, to December 31, 2020), aged 18 years or older at the time of SCI onset, and lived in Sweden (i.e., had a Swedish personal identification number). Only persons aged 18 and older were included as children are not always treated at rehabilitation medicine units and if so, are not entered into SveReh. Nor did we have ethical approval for the inclusion of children. New onset TSCI and NTSCI were included. A total of 1,224 patients were included in the study (Fig. 1). The Swedish population was 9.85 million on January 1, 2016, and 10.38 million on December 31, 2020 [11]. Incidence was calculated for each year, and the number of inhabitants on every 30th of June (2016–2020) was used [11].

This study was a registry-based, retrospective study of incidence, clinical findings, and outcomes during primary rehabilitation. The use of a respiratory aid was defined as the use of non-invasive ventilation, continuous positive airway pressure, or bi-level positive airway pressure; tracheostomy and the use of an invasive ventilator, either with consistent need 24 h a day or partially during the day; or the use of a cough assist machine.

Neurogenic bladder and bowel due to SCI were noted at admission and discharge, respectively, and defined as a disturbed functioning of the bladder or bowel. The International Standards for Neurological Classification of SCI include the clinical findings standardised by the American Spinal Cord Association Impairment Scale (AIS) [12]. The AIS is recorded in SveReh at admission to the rehabilitation unit, meaning that usually more than 48 h have passed since the injury. The Swedish Ethical Review Authority gave ethical approval (reference number 2022-01300-01) according to the guidelines of the Declaration of Helsinki.

SveReh is based on the International SCI Core Data Set [13, 14]. The registration is electronic and the data is stored at Registercentrum Västragötaland [15]. The data are not available to anyone other than the person in charge of data entry at the respective unit. Patient consent is not required for registration in national quality registers of healthcare, according to Swedish Law (Patient Data Act 2008:355, Chapter 7), but there is an opt-out possibility. To retrieve anonymised data from the registry for research, ethical approval is needed.

The data were entered by the respective unit during the patient’s primary rehabilitation stay during the study period. Incidence was calculated as the number of people with SCI each year divided by the population in Sweden on the 30th of June of the same year.

Statistical analyses were performed using IBM Statistical Package for Social Sciences (SPSS) version 24 for Windows (SPSS Inc., Chicago, IL, USA). The results are presented using quantitative descriptive statistics. The TSCI and NTSCI groups were compared using χ² and the Mann-Whitney U test.

The incidence calculated each year, of both TSCI and NTSCI in SveReh, varied between 23.0 and 25.4 per million person-years in the period from 2016 to 2020. For the individuals with a noted TSCI, the incidence varied between 11.9 and 14.8 per million person-years. For those with a noted NTSCI, the incidence was between 8.9 and 11.8 per million person-years.

During the study period, 1,821 individuals aged 18 years or older were admitted to a rehabilitation unit in Sweden due to SCI and had their data entered into the SveReh. A total of 1,224 of these individuals were eligible for this study (Fig. 1). Thirty-three percent of the included persons were female, and the mean age at onset was 56 years (standard deviation [SD] 18), the median age at onset was 60 (SD 18). In the TSCI and NTSCI groups, the youngest individual was 19 years old in both and the oldest was 93 and 94 years, respectively. TSCI represented 55% of cases, with falls comprising 29%, transportation 14%, and sports 5%. In contrast, 42% of the cases were caused by NTSCI. The specific cause of NTSCI was lacking in 57% of cases, as the diagnosis entered in SveReh was unspecific, such as “tetraplegia.” Specific causes included vascular myopathy (11%), tumours (9%), and intraspinal abscesses (8%).

The median time spent at the unit was 40 days (range 0–456 days). The median was chosen instead of the mean because the length of stay (LOS) appeared to follow a skewed distribution. For TSCI, the median LOS was 48 days (range 1–456 days); for NTSCI, the median LOS was 34 days (range 0–278 days). The LOS is recorded in SveReh as the time from admittance to discharge from the rehabilitation unit and does not include the time spent in the acute setting.

The distribution of tetraplegia and paraplegia was 44% and 38%. Among individuals with TSCI, 61% had tetraplegia, compared to 28% of those with NTSCI (Table 1). Neurological level of injury (NLI) and AIS at discharge were missing for 9% and 16% of TSCI and NTSCI cases, respectively. Among individuals with tetraplegia, 66% had a high cervical neurological lesion (C1-C4). Seventy-five percent of individuals with paraplegia had a thoracic lesion. For the level of severity in which the NLI and AIS are combined, the status was not noted for 13% of cases at discharge. For 45 persons, it was not recorded whether it was TSCI or NTSCI. The distribution for the 1,179 persons where TSCI and NTSCI were recorded is presented in Table 1.

Sixty-three percent of patients, both TSCI and NTSCI, were classified at admission as incomplete with a motor incomplete lesion, 43% AIS D and 20% AIS C (Tables 2, 3). A total of 18% were classified with a motor and sensory complete lesion (AIS A) and 8% with a sensory incomplete lesion (AIS B) (Tables 2, 3). For 11% of those with either a TSCI or NTSCI, the AIS was not classified at admission.

For those initially classified as AIS A, 88% remained AIS A at discharge. For those with incomplete lesions, the changes during primary rehabilitation were more substantial (Tables 2, 3). For 45 individuals, whether the SCI was caused by trauma was missing, but the dispersion was the same, with AIS D and C being the most common.

A total of 94 individuals in the cohort (8%) needed some sort of breathing aid at discharge. Of these, the majority were discharged with a continuous positive airway pressure (30%) or cough assist (27%), and 32 (34%) needed invasive ventilatory assistance. For those who were dependent on invasive ventilation at discharge, 75% had a TSCI (Table 1). For those in need of a ventilator (partial or full need), the majority (n = 14, 44%) had AIS A, followed by AIS B (n = 6, 19%).

Neurogenic bladder dysfunction was noted in 709 cases (58%) at discharge, and the data were unknown, not performed, or missing for 246 cases (20%). For 804 cases (66%), whether the person needed help voiding urine was recorded, but for 420 cases (34%), the data were missing. Of the 701 individuals (57%) who needed help voiding their urine, 354 (51%) used intermittent catheterisation, 181 (26%) had an indwelling catheter, 152 (22%) had a suprapubic catheter, and 14 (2%) had some other assistance to empty the bladder.

Among TSCI cases, 419 (63%) had neurogenic bladder dysfunction at the time of discharge, compared to 270 (53%) NTSCI cases. The presence of neurogenic bladder dysfunction was recorded for 94% (n = 187) and 91% (n = 68) of individuals with AIS A and B at discharge, 80% (n = 137) of those with AIS C, 41% (n = 254) of those with AIS D, and 9% (n = 1) of those with AIS E. For those lacking an AIS classification at discharge, 42% (n = 62) had neurogenic bladder dysfunction.

Disturbed bowel functioning was noted in 704 individuals (58%) at discharge, with missing data for 4%. Faecal incontinence at the time of discharge was noted for 225 cases (18%), and the data were unknown, not performed, or not applicable for 557 cases (46%). Twenty-four individuals (2%) had a bowel ostomy at discharge, and the data were unknown, not performed, or not applicable for 519 individuals (42%).

Of those with TSCI, 62% (n = 413) had neurogenic bowel dysfunction at the time of discharge, compared to 53% (n = 270) of NTSCI cases. The presence of neurogenic bowel dysfunction was recorded for 91% (n = 181) and 84% (n = 63) of individuals with AIS A and B at discharge, 76% (n = 131) of those with AIS C, 43% (n = 269) of those with AIS D, and 27% (n = 3) of those with AID E. For those lacking an AIS classification at discharge (n = 136), 42% (n = 57) had neurogenic bowel dysfunction.

Few complications were noted during inpatient stays: deep vein thrombosis affected 27 cases (2%), pulmonary embolism 48 cases (4%), heterotopic ossification 8 cases (0.7%), fractures due to falls 1 case (0.1%), contractures 12 cases (1%), pressure ulcers 75 cases (6%), gastric ulcers 7 cases (0.6%), pneumonia 55 cases (4.5%), pyelonephritis 106 cases (9%), methicillin-resistant Staphylococcus aureus (MRSA) infection 33 cases (3%), infections with unknown origin 145 cases (12%), and unknown complications 106 cases (9%). The complications of TSCI and NTSCI are given in Table 4.

Patients with a TSCI were slightly younger and more severely injured than those with an NTSCI. The median age differed, with 7 years between the groups (57 years for TSCI vs. 63 years for NTSCI), which is in line with previous studies [16, 17]. Ninety-six (14.4%) of the TSCI cases and 87 (17%) of the NTSCI cases were aged ≥75 years. Thus, patients are admitted for rehabilitation regardless of age. The mean age of 56.3 years is slightly higher than in previous studies [4] in Sweden, with a mean of 54.9 years reported in 2017. However, it does correlate with the SCI population getting older at the time of injury [4]. We chose to use the median instead of the mean because the distribution was skewed. As this study only included adults (age ≥18 years), the results differ some from those in other Nordic countries [6, 18], as they included a paediatric population with a younger mean age at the time of injury. In general, the studies show an increasingly older population at SCI onset [6].

We found a difference between the TSCI and NTSCI groups regarding gender, though both were overrepresented by males, but the difference was greater in the TSCI group. Men are reported as being overrepresented in both TSCI [6, 19] and NTSCI [20, 21]. Similar findings have been reported by Joseph et al. [4] and in Norway [20].

Compared to the TSCI group, fewer cervical lesions with AIS grades A, B, or C were found in the NTSCI group. This can be explained by the underlying aetiology of NTSCI, with injury predominantly at the thoracic level or in the lumbosacral region [2, 21], which was also true in our study.

The incidence in this study ranged between 8.9 and 11.8 per million for NTSCI and 11.9–14.8 per million for TSCI, but some of our patients were excluded and we do not know if all persons with an acute SCI were properly registered in SveReh. In addition, not all SCIs are treated at a rehab unit; for example, progressive SCI cases are rarely admitted to an SCI unit for rehabilitation. Compared to the incidence of TSCI in Stockholm presented by Divanoglou et al. [5] and Joseph et al. [4], the incidence was notably lower in our population, likely due to our study only including the persons who were admitted for rehabilitation and not those who may have suffered from an SCI and passed away shortly thereafter.

Our results for TSCI were in line with the findings from Norway (11.4–15.9 per million [7]) but notably lower than findings from Finland (33.6 per million [8]) or Iceland (33.5 per million [18]). However, the results for TSCI were lower than the mean 16 cases per million reported in Western Europe [3].

Comparing our findings with data from adjacent European countries with comparable standards of living, demographics, and healthcare systems, indicate that falls are the leading cause of TSCI in Western Europe [3]. In more southern parts of Europe, transportation was overrepresented in Greece and Italy [3]. In the Nordic countries, transportation used to be a leading cause but has since declined, with falls now being the primary aetiology of TSCI [6].

Consistent with these findings, results from Norway [7] show that falls, sports, and transportation are the leading causes of TSCI, which aligns with our findings where falls, transportation, and sports are also identified as the primary causes. For NTSCI, not as many studies have been conducted; however, a Finnish study from 2020 [21] found a much higher incidence of NTSCI, but it included individuals with less severe lesions who did not require inpatient rehabilitation. Figures from Norway [20] are comparable when it comes to NTSCI (7.7–10.4 per million). One recent study from Germany [22] reported an incidence of 15.73 per million for TSCI and NTSCI combined, and an Italian study [1] reported a decreasing trend in incidence to 10.8/million, lower than in the present study.

Ventilatory assistance at discharge was more common after TSCI than after NTSCI. However, whether everyone who would have benefitted from ventilatory assistance, such as non-invasive ventilation, had been discovered was not clear. Between the different hospitals, there may be a difference in awareness and routines for clinical assessment of ventilatory problems, as not all units are responsible for more severe injuries and respiratory complications. This is one of the reasons why SCI rehabilitation has since been centralised in Sweden.

The LOS was calculated from admission to the rehab unit until discharge from the same unit and did not include the time spent in an ICU or other type of acute setting. This differs from the international guidelines [14] and makes it difficult to compare the results to international studies. TSCI had a 14-day longer LOS than NTSCI, which is in line with previous studies [16, 20]. Few patients with complete NTSCI and/or need for ventilatory assistance may explain why the LOS was shorter in this group. We also did not consider additional complications due to a TSCI, such as other fractures that may lengthen the LOS. The completeness and NLI of SCI should be considered in this discussion. Compared to LOS in Nordic countries with a similar healthcare system, the LOS in our study was shorter than in a Norwegian study that also included time spent at a spinal cord unit or rehabilitation medicine unit [17], 15.2 and 11.8 weeks for TSCI and NTSCI, respectively. A Finnish study [8] found a mean LOS that included the time spent in the acute setting of 65.9 days for TSCI, which is more in line with the findings in our study.

Considering that only 2.7% of the cohort was missing information on neurogenic bowel dysfunction, it is odd that 20% of the cohort lacked information regarding bladder dysfunction. We assume that this is due to difficulty finding the status within the electronical medical charts, resulting in the person in charge of filling in the SveReh having trouble finding the correct information and leaving the section of the registry empty.

Pyelonephritis was noted in 14.5% of TSCI cases and 4.5% of NTSCI cases. Urinary tract infections are known to be more common among the TSCI population [19, 23]. Nonetheless, it is important to be vigilante regarding voiding routines and hygiene. We need to improve the tracking in medical charts regarding voiding as 34% of the cohort had missing data and, of the patients for whom data were available, more than half used intermittent catheterisation, which is the best practice [23‒25], although not always suitable due to loss of hand function, high age, or psychiatric reasons.

Pressure ulcers were noted during primary rehabilitation in 10.2% of TSCI cases and 3.3% of NTSCI cases, lower than reported in previous studies [17, 26], possibly due to increased knowledge of the importance of limiting this complication. On April 1, 2023, the rehabilitation of SCI in Sweden was centralised into four different rehabilitation medicine units. Future studies comparing data following the centralisation and whether the results vary from the present study would be of interest.

As a register-based study, the present study has a limitation of data being forgotten or not correctly registered. For example, the cause of NTSCI was in a majority of the cases not documented because only the primary diagnosis was recorded in the registry, often listed simply as “tetraplegia” or “paraplegia,” making it impossible to ascertain the specific cause. By including all adults available in the nationwide register, our study provides a good opportunity to investigate outcomes and complications of SCI regardless of cause. However, we know from experience that not all persons are admitted to a rehabilitation unit and thereby could be missed. As no children were included, the study only shows the incidence for the adult population in Sweden, not the entire country’s population. Additionally, since the rehabilitation medicine units are often separate from acute care, immediate AIS evaluations after the injury, which are recorded in SveReh, are not always possible.

Sweden has a lower incidence of SCI than Western Europe and globally. Males are overrepresented among both NTSCI and TSCI cases. The most common cause of TSCI is falls. Few complications are reported during the time spent in primary rehabilitation, but pyelonephritis and pressure ulcers are still too common and need to be better evaluated. Individuals with NTSCI are older, less severely injured, and have a shorter LOS than individuals with TSCI.

Written informed consent was not required as per decision by the Swedish Ethical Review Authority which gave ethical approval (reference number 2022-01300-01) according to the guidelines of the Declaration of Helsinki.

The authors have no conflicts of interest to declare.

This research has been made available by contributions from the Promobilia Foundation, Spinal Cord Injury Centre Gothenburg, and grants from the Swedish state under the agreement between the Swedish government and the county councils, the ALF-agreement (SU 2023-03122, -965653). None have had any influence on the content of the study.

C.J. analysed the data. C.J., K.S.S., and T.R. interpreted the data and prepared and revised manuscript content. Å.L.N. critically revised the manuscript content. All authors have given their final approval of the version to be submitted.

The datasets generated and analysed during this study are available from the corresponding author upon reasonable request.