A quarter of TSCI patients died during the observation period and compared to the control group, the number of deceased was slightly more than double. Cardiovascular and pulmonary diseases were the major causes of death, pneumonia being the most prominent disease leading to death. Age at the time of injury and completeness of injury were independent predictors of death.

Previous studies of TSCI mortality, predictors of mortality, and causes of death are difficult to compare because of methodological differences regarding study populations, case selection, follow-up times, and medical certification of cause of death [1, 2, 5, 13,14,15,16,17,18,19,20,21]. Those studies have been conducted in different time periods in different countries and therefore with different healthcare systems, services, and policies. Many studies have focused on patients in the rehabilitation phase which excludes deaths occurring in the acute or subacute phase after injury and deaths of patients with minor TSCI-related disability or barriers to access rehabilitation [5, 15, 17, 19].

Notwithstanding the methodological differences in previous studies, improvements in survival over time and consistent predictors of mortality have been reported in several studies [1,2,3, 22]. Life expectancy after TSCI has improved over the decades but remains below that of the general population [1,2,3, 22]. A systematic review showed that overall mortality in TSCI was between 1.47 and 2.8 times higher than in the general population [3]. In line with earlier literature, TSCI was related to excess mortality in our study. Slightly more than double the amount of TSCI patients died during the observation period compared to the controls. In addition, our findings complement prior Finnish ones by Ahoniemi and colleagues, who described that the overall survival rate of TSCI was approximately half of the general population during the follow-up period 1976-2007 [5].

In earlier TSCI studies, advancements in critical care have been associated with a decline in mortality during the first two years after injury [3, 13, 19, 22]. Despite this, it still appears that a significant proportion of deaths of TSCI patients occur within 1-2 years after injury [13, 14]. Inglis et al. also recently reported that in-hospital mortality in the elderly after TSCI is high [23]. In our study, a significant number of TSCI deaths in the follow-up period occurred shortly after injury; 38% within one year and 51% within two years after injury. The 1-year mortality in the Finnish TSCI population was only slightly lower than in Estonian and Australian studies, where also deaths in the acute phase were included in the study [13, 14]. Sabre et al. showed that almost half of the deaths during the observation period 1997-2011 occurred during the first year after injury [14]. Similarly, O´Connor et al. reported that half of the deaths occurred in the first year after injury, with nearly two-thirds of these deaths occurring within the first two months [13]. On the other hand, in a study from another Nordic country, Norway, the 1-year mortality was lower than in our study, 21% [16].

Regarding the predictors of mortality, our study is in accordance with several other studies; the mortality was most strongly affected by age at the time of injury and completeness of injury, also regardless of the level of injury [1,2,3, 13, 14, 17, 18, 24, 25]. Also, a history of other medical conditions has previously been reported as a predictor of mortality [15, 17]. Recent studies have shown an increasing trend in the age at the time of TSCI, the proportion of cervical injuries and injuries caused by falls [6, 26, 27]. In Finland, especially low-level fall is the most dominant etiology of TSCI in patients over 60 years [6]. Moreover, in Finland, both the incidence and the average age sustaining a fatal cervical spine injury, associated with SCI in 83%, has increased markedly during the last decades [28]. Fall-induced accidents among elderly males were the most prominently increasing subpopulation [28]. In our current study, one-third of patients who had a fall as the etiology of TSCI died during the observation period. The information on the fall height (low-level fall vs. high-level fall) was not available. Also, over one-third of patients with a grade AIS A-C cervical injury died by the end of the follow-up period. It has been assumed, that nowadays an increasing number of less healthy and functionally less capable individuals are surviving to an older age in Finland and they are also more prone to falls and injuries [29]. Probably the greater occurrence of comorbidities among frail elderly contributes also to the increased risk of death in the group of patients with falls as the etiology of TSCI. Elderly males are the most prominent subgroup of individuals who sustain a TSCI, and the cause of injury is often a fall [6]. On the other hand, the incidence of TSCI among women is more evenly distributed in relation to age [6]. This baseline characteristic difference between male and female TSCI patients would allude to a higher mortality rate among males. A bit surprisingly, there was no difference in mortality between men and women in our study. In contrast, higher female mortality has been reported in four Scandinavian studies [5, 16, 17, 19], whereas large studies from the USA and UK have shown the opposite [1, 2]. Furthermore, ventilator-dependency has been stated to have a strong effect on mortality [1, 2, 23]. In line with this, most of the ventilator patients died in our patient group.

Until the 1970s, renal failure and other urinary tract complications were reported to be the leading causes of death among patients with SCI [21]. More recently, probably in connection with the advancements of SCI care, the causes of death among TSCI patients have become comparable to that of the general population [21]. Nowadays the leading causes of death among patients with TSCI are pulmonary diseases, especially pneumonia, and circulatory diseases followed by infections, and suicide [2, 3, 5, 14, 16, 30]. Consistent with other earlier results, the two most common causes of death among TSCI patients were circulatory and pulmonary diseases in our study. Circulatory disease mortality of TSCI patients is comparable to the general Finnish population in which circulatory diseases cause a third of the deaths [7]. Pulmonary disease-related mortality is ten times higher in the TSCI patients than in the general Finnish population wherein the mortality due to pulmonary diseases is only about 3% [7]. In our study, pneumonia was clearly the most noticeable single diagnosis leading to death. Depending on the degree of injury, a cervical or upper thoracic SCI markedly reduces respiratory function by causing paralysis of the respiratory muscles, and disturbed balance of autonomic nervous system with a preponderance of the parasympathetic nervous system. These changes lead to reduced ability to breathe and cough effectively as well as increased bronchial secretion and constriction favoring the development of pulmonary complications like pneumonia [31]. In addition, dysphagia with aspiration is relatively common among patients with cervical SCI potentially elevating the risk for pneumonia and other respiratory complications [32].

Our study gives a reliable population-based estimate of the TSCI-associated excess mortality. In our study, all patients with a new TSCI were enrolled and evaluated according to the International SCI Core Data Set already in the acute phase. This enables good case coverage with generalizable results. Additionally, the detailed review of all the available death certificate data improves the accuracy of the results, while accepting the underlying cause of death would have given more prominence to the external cause of injury-producing event than to the actual cause of death. As a limitation, the SCI center of Helsinki UH was missing from our study. It is possible that patients who are residents of Tampere UHs’ or Oulu UHs’ primary referral areas but were injured elsewhere are missing from the data. The case coverage is less certain in Turku and Kuopio UHs’ primary referral areas because the acute care and rehabilitation of those patients may partially have taken place in their own and Helsinki UH’s hospital districts. Prehospital deaths were not included in our study. Autopsies were not performed on all the deceased patients, and some of the causes of death might be incorrect.

In the future, further research with a longer follow-up period is necessary to evaluate the short-term and long-term mortality and causes of death. In future studies, concerning causes of death of individuals with TSCI, it is important to take into account that, the cause of death of TSCI patients is possibly described by the index injury [5, 13]. To avoid reporting these misleadingly high external causes of death, it is crucial to use all the detailed data including death certificates.