Welcome to SCIRE Professional
 

Life Expectancy and Mortality

Survival rates for individuals with SCI have made steady improvements over the past five decades. Prior to World War II life expectancy for individuals with SCI was quite poor (Geisler et al. 1983). Leading causes of death were those resulting from renal failure and infection (Lammertse 2001). Since the introduction of antibiotics, improved emergency transportation, advances in long-term health interventions, and the availability of preventative care at specialized treatment centres, mortality rates have been steadily decreasing and the causes of death have begun to mirror those of the general population (Whiteneck et al. 1992; DeVivo et al. 1999). However, life expectancy is still diminished compared to the general population (Whiteneck et al. 1992; Hartkopp et al. 1997; McColl et al. 1997; Frankel et al. 1998; DeVivo et al. 1999; Yeo et al. 2000; Krause et al. 2004).

Causes of death among individuals with SCI and those in the general population appear to be similar. In 2011, the two leading causes of death in high- and middle-income countries for the general population were ischemic heart disease and stroke (WHO 2011). Other common causes were tracheal bronchus and lung cancers, chronic obstructive pulmonary disease, lower respiratory infections, and Alzheimer’s disease and other dementias (WHO 2011). Similarly, two leading causes of death in the SCI population are respiratory complications and heart disease (Hartkopp et al.1997; Frankel et al.1998; DeVivo et al. 1999; Soden et al. 2000; Zeilig et al. 2000; Garshick et al. 2005). Additionally, the latest report from the National Spinal Cord Injury Database (NSCIDB) indicates the main causes of death in persons with SCI in the United States are pneumonia and septicemia (NSCISC 2013). The high rates of cardiovascular disease in the SCI population may be partly due to physiological and functional changes following injury (Dearwater et al. 1986; Yekutiel et al. 1989; Bauman et al. 1992a, b; Gupta 2006). Interestingly, cancer is a growing cause of death in persons with SCI (DeVivo et al. 1999; Zeilig et al. 2000; Imai et al. 2004).

In general, it appears that as individuals with SCI age, cause of death becomes similar to age matched controls (Capoor & Stein 2005). Some deaths, however, may occur prematurely (e.g., from cardiovascular disease; Yekutiel et al. 1995), and there are some notable differences in mortality patterns between the SCI and the general populations.

In this section, 5 longitudinal studies and 1 cross-sectional study on mortality and life expectancy among individuals with spinal cord injury are reviewed.

Table 1a: Systematic Review on Mortality and Survival

Author Year; Country
Score
Research Design
Total Sample Size
MethodsOutcome
Chamberlain et al. 2015;

Switzerland

 

Reviewed published articles between Jan 1, 2000- June 17, 2013

 

Quality Assessment:

Center for Review and Dissemination;

STROBE guidelines

 

N=74 studies included in systematic review

 

 

 

 

Study characteristics: The study duration ranged from <1 year to more than 50 years and include data beginning in 1935 until 2009. Sample sizes varied between 24 and close to 100 000 individuals.  A majority of studies involved Traumatic SCI (TSCI) (n = 68), but a few studies were also included that considered Non-traumatic SCI (NTSCI) only (n = 7), or both NTSCI and TSCI (n = 3).

Methods: Literature search on English, French, German, Italian, or Spanish publications that reported at least one measure of survival or mortality on people with TSCI, NTSCI, or both. Studies focusing on specific populations were eligible as well as descriptive studies that included the number of in-hospital deaths, irrespective of whether mortality was the main outcome of interest. RCTs, reviews, case reports and studies not specific to SCI were not included.

Databases: PubMed, EMBASE, LILACS (The following MeSH terms were used in the search strategy: ‘spinal cord injury’, ‘paraplegia’, ‘quadriplegia’, ‘mortality’, ‘cause of death’, ‘survival’, ‘epidemiology’, ‘life expectancy’. Non-MeSH terms included: ‘standardized mortality ratio (SMR), which is ratio of observed deaths over expected deaths (where >1 represents more deaths and expected) the or, and ‘proportional hazards model’.)

Outcome variables of interest: In-hospital mortality, median survival, percent survival at 1, 5, 10, 20, and 30 years after SCI, standardized mortality ratio (SMR)

1.     Median survival ranged between 2.8-43 years after TSCI. One-year survival ranged between 79-100%, 5-year survival between 85-96%, and 10-year survival between 81-93%.

2.     Studies comparing survival between genders generally reported higher rates of survival among females

3.     Among studies that reported risk of mortality with age at injury, it was found that the mortality risk increased on average by 6% with increasing age and that higher mortality was found in males as compared to females.

4.     22 studies examining SMRs indicated that SMRs were higher in women compared to men, with increasing age, and with increasing severity of the lesion. SMRs related to cerebrovascular diseases, bacterial diseases, suicide/accidental poisoning, and pneumonia were often most elevated. SMRs for people with SCI ranged between 1.5-5x that of the general population

5.     Two studies found that the overall life expectancy among people with SCI was lower than in the general population and 3 studies find that the gap in life expectancy widened with increasing lesion severity and age.

Table 1b: Mortality and Life Expectancy

Author Year; Country
Score
Research Design
Total Sample Size
MethodsOutcome
Kopp et al. 2017;

USA

Level 2

Longitudinal

N=1203

 

 

 

Population: 639 people (77.3% M) with SCI

without Pn/Wi (pneumonia and wound infection) (53% of total): mean age at baseline (y)=35;

AIS A (71.3%), AIS B (24.6%), AIS C (34.6%); C1-C4 (41.6%), C5-C8 (58.4%)

 

564 people (83.3% M) with SCI who acquired Pn/Wi (47% of total): mean age at baseline (y)=34;

AIS A (65.6%), AIS B (18.8%), AIS C (15.6%); C1-C4 (58.3%), C5-C8 (41.7%)

 

Methodology: Eligible participants were assessed at 4 subsequent time points during a follow-up period of 5 years: at admission to inpatient rehabilitation, at discharge, and at 1 and 5 years after SCI to investigate whether prevalent hospital-acquired Pn/Wi affect the clinical long-term outcome after acute traumatic SCI. Pn/Wi-related mortality was assessed as a secondary endpoint.

 

Outcome Measures: Mortality and life expectancy

 

1.     Long-term survival curves indicated a higher mortality in the group of patients with Pn/Wi compared to patients without Pn/Wi in the total population and in the AIS subgroups over a 10 year period.

 

Shavelle et al. 2015;

USA

Level 3

Case-control

N=31,531

 

Population: 31,531 people with SCI who survived 2 years post-injury.

 

Methodology: A total of 31,531 persons, 8536 of whom died during the 39-year study period, contributed 484,979 person-years to the data set to compare the mortality between 4 time periods of injury: 1973-1979; 1980-1989; 1990-2004; 2005-2012

 

Outcome Measures: Survival; survival relative to the general population life expectancy.

1.     There was no trend toward improved survival since 1979. Mortality in persons with SCI did not decrease in the most recent period of 2005-2012 relative to the previous period of 1980-2004.
Cao et al. 2013;

USA

Level 3

Case-controlN=2685

Population: 2685 persons with traumatic SCI who had been discharged from 62 acute care non-federal hospitals in South Carolina between 1998 and 2009.

Methodology: Data were collected through a comprehensive data system spanning 11 years of observation; the mortality experiences of traumatic SCI participants were compared with the general population of South Carolina using SMRs for selected causes of death, adjusting for age, race, and sex.

Outcome Measures:  Mortality, and causes of death identified by ICD-10.

1.   The crude annual mortality rate during the period was 33 per 1000 person-years.

2.   Increasing number of comorbidities, admission into non-trauma centers, advancing age, type of insurance (Medicaid and commercial payers), higher injury level and completeness, and being female were positively and significantly associated with the risk of death after discharge from acute care facilities.

3.   The overall mortality rate of the cohort was 3.6 times (95% CI, 3.3,3.9) higher than the general population.

4.   Likelihood of death was 1.6 times (95% CI: 1.2, 2.2) higher for persons with C1-4 level injuries than other levels.

Rabadi et al. 2013;

USALevel 3

Case-control

N=147

Population: 147 veterans with SCI (144M 3F); mean (SD) age: 59.6(13.5), range 26-90; mean (SD) age at injury (yrs): 37.7(14.4), range 18-82; Level of injury: 53% cervical, 33.3% thoracic, 13.6% lumbosacral; Severity: 53% complete, 47% incomplete.

Methodology: Patients enrolled in an SCI program from 1 Jan 2000 until 31 Dec 2011 were retrospectively studied (time points measured every 4 months and annually until end of study); the sample was divided into 2 groups based on the survival status by Dec 31 2011.

Outcome Measures:  Mortality, life expectancy

1.     There were 3 major causes of death: infection-related: 46% (pneumonia [21%], urinary infection [14%], infection of pressure ulcers [11%]); cardiovascular-related: 25% (myocardio infarction [14%], stroke, [4%] congestive heart failure [4%], other cardiovascular [4%]); cancer-related: 16%; other 15%.

2.     In individuals with complete SCI, deaths were mainly infection-related and occurred in the hospital (51%).

3.     In individuals with incomplete SCI, deaths were mainly cardiovascular and cancer-related and occurred in the community.

4.     Age at the time of injury was the main predictor of SCI-related mortality.

2.     The estimated survival since time of injury for SCI patients was 39.3 (2.3) yrs. The 10- and 20- yr survival rates after injury were 87% and 78% respectively.

Frisbie 2010;

USA

Level 3

Case-control

N=322

Population: 322 participants with SCI (318M 4F); had follow up service for SCI with scheduled annual checkups between 1998 and 2007.

Methodology: Cohort was divided into two groups: surviving and deceased through 2008.

Outcome Measures:  Mortality and life expectancy

1.     The 239 participants who survived had an average age of 60 and 30 yrs of paralysis and the 83 participants who died had an average age of 70 and 27 yrs of paralysis.

 

Pickelsimer et al. 2010;

USA

Level 3

Case-control

N=988

Population: 988 participants with traumatic SCI (731M 257F); differed significantly in age at TSCI, race and type of SCI.

Methodology: Using 2 statewide datasets, prevalence and incidence of medically attended health conditions during emergency department visits, acute care hospitalizations, and outpatient hospital visits, across all ages and SCI severity over a 10 year follow-up period were reported. A time-to-event life table method was used to calculate the prevalence and incidence of health conditions and Cox proportional hazard ratio of mortality by gender controlling for age and SCI severity.

Outcome Measures:  Morbidities identified by ICD-9-CM; ICDMAP-health care billing data.

1.     During 10 yr follow-up period, 49 (19.1%) women and 104 (14.2%) men died. No difference between genders for mortality after controlling for age and TSCI severity.

2.     The reported 15.5% mortality rate is higher than other reports (7 to 9%) examining outcomes in the SCI population over a 10-yr period.

Savic et al. 2010;

UK

Level 2

Longitudinal

N=282

Population: 282 individuals with traumatic SCI injured prior to 1971; age at time of injury 15-55 years; mean age at enrollment 52.7; 86.7% males; 29.2% tetra ABC, 49.8% Para ABC; 21.0% Incomplete injuries.

Methodology: Full physical assessment, diagnostic procedures, detailed medical and psychosocial interview, retrospective medical records review.

Outcome Measures:  Medical History and Current Status; battery of outcomes including quality of life, depression, pain, and environmental factors.

1.     Over a 16-year period, the mortality rate increased over time: 1993 (4.1%), 1996 (9.9%), 1999 (16.4%), 2002 (22.5%), 2006 (35.8%).
Samsa et al. 1993;

USA

Level 5

Cross-sectional with Able-boided (AB) comparison

N SCI=6147

Population: 5545 veterans with SCI; mean (SD) age at onset: 23(6) yrs; mean YPI: 24(13).

Methodology: Comparison of long-term survival of veterans with SCI to a population-based life table. Life table formed from age-matched American men from same time period.

Outcome Measures:  Survival curves, extending from 3 mos to 40 yrs after injury.

1.     Mean life expectancy of veterans with traumatic SCI and surviving at least 3 months is an additional 39 yrs after injury – 85% of age-matched American males.

2.     Older age at injury is a stronger predictor of poorer long-term survival than is complete tetraplegia.

Discussion

Rabadi et al. (2013) reported a 10-year survival rate of 87% from time of injury and a mean age of injury of 39. Pickelsimer and colleagues (2010) reported a 10-year survival rate of 84.5% from time of injury with a mean age of injury of 41 and excuded those that died in the first 89 days.

Although Samsa et al. (1993) found that injury level was not a significant predictor for mortality, they noted a near significant effect (p=0.06) for complete cervical injuries compared to all other injuries. Similarly, in an 11-year longitudinal study, Cao et al. (2013) reported the likelihood of death in individuals with injuries at the C1-4 levels to be 1.6 times greater than for individuals with injuries at other levels. Several studies have found no association between impairment and mortality (e.g. Liang et al. 2001; Imai et al. 2004; Garshick et al. 2005), whereas other studies have highlighted the importance of impairment as a prognostic factor (Whiteneck et al. 1992; McColl et al. 1997; Coll et al. 1998; DeVivo et al. 1999; Soden et al. 2000; Yeo et al. 2000; Strauss et al. 2006). Samsa and colleagues (1993) found age of onset to be a significant predictor of long-term survival, which is consistent with other longitudinal studies that did not meet our inclusion criteria (Whiteneck et al. 1992; Frankel et al. 1998). Rabadi et al. (2013) reported age of onset to be the only predictor of mortality.

With regards to causes of mortality, Samsa et al. (1993) found that diseases of the genitourinary system (i.e. renal failure, septicemia) disproportionately accounted for death in their SCI sample, but the patterns of death began to approach that of the general population by 20 years post-injury. For instance, the rates of circulatory disease and neoplasms steadily increased across time points. Interestingly, the causes of death due to injury and poisoning, and external conditions were the highest at 3-months to 5 years post-SCI, and steadily decreased across time points. In a 12-year longitudinal study of 147 veterans with SCI, Rabadi et al. (2013) reported infection, cardiovascular, and cancer to be the three primary causes of death. Although not discussed, causes of death may have also included suicides. Regardless, the findings of lower levels of evidence highlighting the high rates of suicide as a cause of death (i.e. Imai et al. 2004) reinforces the need to provide psychosocial services to help minimize the occurrence of suicide in persons with SCI.

An acknowledged limitation of the study by Samsa et al. (1993) is the reliance on secondary data sources for case identification, control selection, and mortality assessment. The study was also only on male veterans and did not include women or persons who did not survive acute SCI (i.e. less than 3 months post-SCI).Causes of death were not reported by Pickelsimer et al. (2010) or by Savic et al. (2010).

Conclusion

There is level 4 evidence that the 10 year survival rate post injury is 84-87% (Rabadi et al. 2013; Pickelsimer et al. 2010).

There is Level 4 evidence (Frisbie 2010) that the mortality rate post-SCI over a 10-year period may be 15.5% to 25.8%, and level 4 evidence (Cao et al. 2013) that the mortality rate is higher for individuals with SCI than the general population.

There is Level 4 evidence (Cao et al. 2013) that mortality may be higher for persons with SCIs at the C1-4 level than other spinal cord levels.

There is Level 4 (Frisbie 2010) to Level 5 evidence (Samsa et al. 1993) that the causes of death post-SCI are beginning to approximate those of the general population.

There is Level 4 and 5 evidence (Samsa et al. 1993; Cao et al. 2013) that life expectancy for males with SCI is lower than the general male population.

There is level 4 evidence (Rabadi et al. 2013) that older age at time of injury is a predictor of SCI-related mortality.

Life expectancy for males with SCI is likely lower than the general male population.

Persons injured at a younger age will likely have a longer life expectancy than persons injured at an older age.

Causes of death post-SCI may be similar to those of the general population.