Impact of Associated Conditions Resulting From Spinal Cord Injury on Health Status and Quality of Life in People With Traumatic Central Cord Syndrome

Study Design 

We conducted a retrospective review of the spine database at the Vancouver General Hospital to identify people who were admitted and treated for a traumatic central cord syndrome between 1994 and 2002. Traumatic central cord syndrome was defined as a cervical SCI that produced disproportionately greater weakness in the upper limbs than in the lower limbs and is associated with sacral pin-prick or voluntary motor sparing. Inclusion criteria included: being over the age of 18, admitted to the Spine Unit of Vancouver General Hospital within 72 hours of injury, have an initial neurologic examination performed by a physiotherapist, have the opportunity to benefit from an inpatient or outpatient rehabilitation program, and being a minimum of 2 years postinjury at the time of follow-up. Exclusion criteria included: having a major comorbidity (eg, cancer with or without metastases, severe liver or kidney failure, multiple sclerosis, diabetes, or inflammatory arthritis) or other traumatic injuries that would affect physical function; having organic brain disease or dementia; and not being able to read or comprehend the instruments. Hospital and university ethics approval was obtained to conduct the study. All subjects provided informed consent.

Data Collection 

A chart review was conducted to obtain the following data: sex, date of birth, date of injury, and admission neurologic status. We contacted potential subjects by mail and/or phone to schedule a follow-up assessment. In subjects who provided informed consent, an experienced physiotherapist who worked on the Spine Unit at Vancouver General Hospital visited the subjects in their homes and conducted the follow-up assessment. The physiotherapist interviewed the subjects to determine if the associated SCI conditions were present or absent and conducted a neurologic assessment based on the International Standards for Neurological Classification of Spinal Cord Injury24 to determine the motor score. All subjects completed the following standardized questions and/or instruments: sociodemographic factors, Self-Administered Comorbidity Questionnaire (SCQ),25 SF-36,26 numeric QOL rating scale, and symptom satisfaction.

Instruments 

Statistical Analysis 

We conducted an exploratory analysis to examine the distribution and the correlation for the associated SCI conditions neuropathic pain, spasticity, problems with bowel, bladder, and sexual function, and motor score. To evaluate the effect on health status (SF-36 PCS and MCS, symptom satisfaction) and QOL (numeric QOL rating scale), a 3-stage analysis was conducted. Models for the associated SCI conditions were developed for each stage, except when no correlation was detected at stage 1 based on a P greater than .90. In stage 1, a univariate analysis was conducted to evaluate the effect prior to any adjustments and this provided an estimate of the unadjusted effect of the associated SCI condition on the patient reported outcomes. In stage 2, the adjusted effect was estimated by controlling for the personal factors listed in the conceptual model (fig 1) that are potential confounders. Compensation status, marital status, and employment were assessed but they were not found to be confounders and so we did not include them in further analyses. Confounding was addressed using automatic variable selection32 and by comparing the adjusted and unadjusted effects.33 A backward stepwise variable selection procedure was used and covariates that produced a minimum of a 10% change in the effect or had a statistical value less than P less than .20 were retained.33 In stage 3, further evaluation of the confounding effects of the other associated SCI conditions was performed. In stages 2 and 3, the variable selection procedure was first assessed for the main effect and then an interaction between the main effect and the associated SCI conditions was evaluated. If the interaction was statistically significant with P less than .05 then graphs were produced to determine if the interaction was clinically relevant.

We used linear regression to analyze the SF-36 PCS and MCS; the regression coefficient was used for estimating the effect. Ordinal logistic regression was used to analyze the numerical QOL rating scale and symptom satisfaction data where the odds ratio (OR) was used to estimate the effect. All statistical analyses were conducted using SAS.a

Subject Characteristics 

A total of 114 people with the diagnosis of traumatic central cord syndrome were identified in the spine database and 70 subjects completed the follow-up. People were excluded for the following reasons: 13 people died; 3 were hospitalized and were too sick to complete the follow-up; 16 could not be located; and 12 people declined participation. Those excluded in the analysis did not differ from those who were included, except for the average age (62y for people excluded vs 51y for subjects included).

The characteristics of the sample are included in table 1. Missing data occurred in less than 10% of the associated SCI conditions (see table 1) and consequently, the analysis was conducted on the observed data only without imputation.34 Information about whether subjects were seeking compensation for their injury was incomplete; 47% (n=19) did not answer the question. This variable was treated as a 3 level categorical variable (yes, no, not applicable) in the analysis.34 In addition, the bladder, bowel, and sexual function problems were grouped into 1 variable (bladder, bowel, and sexual function problems) because 18 (78%) of 23 subjects reported 2 or more problems with bowel, bladder, or sexual function and so it was not possible to isolate their independent effect (table 2).

Table 1. Summary of Subject Characteristics

	Characteristic	Mean ± SD (Range; IQR)
	Age (y) at injury	45±18 (range, 13–91; IQR, 18–59)
	Age (y) at follow-up	51±18 (range, 19–95; IQR, 38–65)
	Time to follow-up (mo)	70±34 (range, 24–144; IQR, 41–95)
	Comorbidity score (0–100)	5.4±5.5 (range, 0–26.0; IRQ, 0–9.5)
	Motor score at follow-up	92.1±11.6 (range, 39–100; IQR, 89–100)
		n(%)
	Sex
	Men	57(81)
	Women	13(19)
	Education
	Less than high school	16(24)
	High school	20(30)
	Some college	10(15)
	College	13(20)
	Postgraduate or degree	7(11)
	Marital status
	In a relationship	42(62)
	Single	26(38)
	Employment
	Employed	37(54)
	Retired	15(22)
	Disabled	12(18)
	Unemployed	4(6)
	Seeking/obtained compensation
	Yes	23(33)
	No	28(40)
	Missing	19(27)

Abbreviations: IQR, interquartile range; SD, standard deviation.

Table 2. Frequency of Associated SCI Conditions

	Associated SCI Condition	n (%)
	Neuropathic pain (n=69)	27(39)
	Spasticity (n=69)	41(59)
	Bowel function problems (n=69)	12(17)
	Bladder function problems (n=69)	17(25)
	Sexual function problems (n=66)	18(27)
	Only 1 problem with bowel, bladder, or sexual function	5/23(22)
	Two of 3 problems with bowel, bladder, or sexual function	12/23(52)
	Problems with bowel, bladder, and sexual function	6/23(26)

Health Status and QOL 

Health status was assessed using the SF-36 and a question related to symptom satisfaction. The PCS was 42.4±12.3 and the MCS was 52.4±10.6. Overall satisfaction with symptoms was rated as very dissatisfied or dissatisfied by 24 subjects (35%); neutral by 6 subjects (8%); and satisfied or very satisfied by 40 subjects (57%). QOL was assessed using the numeric QOL rating scale (range, 1–7). Eleven subjects who had a score between 1 and 4 were regrouped into a new category (labeled <5). This produced 4 categories for numeric QOL rating scale, with 11 (16%), 15 (22%), 24 (36%), and 17 (25%) for the categories less than 5, 5, 6, and 7, respectively.

Effect of Associated SCI Conditions on Health Status and QOL 

The results for the effect of the associated SCI conditions on SF-36 PCS, MCS, symptom satisfaction, and numeric QOL rating scale are presented in table 3.

Table 3. Effect of the Associated SCI Conditions on Health Status and QOL Outcomes

	Associated SCI Condition	SF-36 PCS			SF-36 MCS			Symptom Satisfaction			QOL
		β	95% CI	P	β	95% CI	P	OR	95% CI	P	OR	95% CI	P
	Neuropathic pain
	Stage 1	−10.9	−16.2to−5.7	<.001	−5.4	−10.4to−0.5	.03	5.8	2.2to14.8	<.001	1.4	0.6to3.3	.48
	Stage 2	−7.2	−11.8to−2.6	.002	−5.4	−10.4to−0.5	.03	5.0	1.9to12.9	<.001	1.1	0.4to2.8	.84
	Stage 3	−4.0	−8.3to0.3	.07	Interaction§			3.7	1.4to10.2	.01	0.9	0.3to2.4	.82
	Spasticity
	Stage 1	−8.3	−13.8to−2.8	.003	−0.2	−5.2to4.9	.95	3.5	1.4to8.6	.01	1.3	0.5to3.0	.60
	Stage 2	−6.3	−11.4to−1.1	.02	NA			2.8	1.1to7.2	.03	0.9	0.4to2.2	.82
	Stage 3	−1.4	−5.9to3.1	.54	NA			1.5	0.5to4.2	.50	0.5	0.2to1.5	.23
	Bowel, bladder, and sexual dysfunction
	Stage 1	−15.0	−19.7to−10.4	<.001	−2.9	−8.4to2.7	.31	3.5	1.3to9.2	.01	3.6	1.3to9.5	.01
	Stage 2	−11.2	−16.1to−6.4	.003	−2.9	−8.4to2.7	.31	3.8	1.2to12.0	.03	1.5	0.5to4.6	.45
	Stage 3	−6.9	−11.6to−2.2	.004	1.4	−4.5to7.3	.65	0.7	0.2to2.3	.54	1.5	0.5to4.7	.79
	Follow-up motor score⁎
	Stage 1	7.1	5.3to9.0	<.001	2.9	0.8to4.9	.01	2.7	1.6to4.6	<.001	1.7	1.1to2.5	.01
	Stage 2	5.5	3.6to7.4	<.001	2.9	0.8to4.9	.01	2.9	1.6to5.2	<.001	1.5	0.9to2.2	.07
	Stage 3	Interaction†			Interaction‡			2.2	1.2to3.8	.009	1.7	1.1to2.7	.03
		Interaction∥

NOTE. Covariates considered in the model: age at time of follow-up; sex; education; comorbidities; time since injury; neuropathic pain, spasticity; bowel, bladder, and sexual dysfunction, follow-up motor score. Abbreviation: NA, not applicable.

⁎ FMS is the effect for an increment of 10 points in the motor score. † Follow-up motor score and spasticity. ‡ Follow-up motor score and neuropathic pain. § Neuropathic pain and follow-up motor score. ∥ Follow-up motor score and education.

SF-36 PCS 

In stages 1 and 2, there was a significant association between neuropathic pain and PCS. In stage 3, there was a nonsignificant trend suggesting there may be an effect of neuropathic pain on the PCS. Spasticity also had a significant association with the PCS in the first 2 stages of the analysis but was not significant in stage 3. The associated SCI condition bowel, bladder, and sexual dysfunction on the PCS was statistically significant in all stages of the analysis and decreased PCS. In terms of the motor function, as expected the PCS increased as the follow-up motor score increased, as shown in the models from stages 1 and 2 of the analysis. However, this effect was modified by spasticity and education in stage 3. As the follow-up motor score increased, the presence of spasticity (fig 2) and having a lower level of education reduced the positive effect of an improvement in the follow-up motor score on the PCS.

View Large Image Download to PowerPoint

Fig 2. The modifying effect of spasticity on the relationship between the follow-up motor score and SF-36 PCS.

SF-36 MCS 

In stage 1 and 2, neuropathic pain was significantly associated with MCS. In stage 3, a significant interaction with the follow-up motor score was detected. As the follow-up motor score increased, neuropathic pain did not have as much of a negative impact on the MCS (fig 3). In other words, when neuropathic pain was present in addition to a lower follow-up motor score, there was a disproportionate reduction in the MCS, greater than what would be expected from the low follow-up motor score alone. Neither spasticity nor bowel, bladder, and sexual dysfunction had a statistically significant effect on the MCS in any stage of the analysis. The effect of the follow-up motor score on the MCS was modified by the presence of neuropathic pain in stage 3 of the analysis as described previously.

View Large Image Download to PowerPoint

Fig 3. The modifying effect of the follow-up motor score on the relationship between neuropathic pain and SF-36 MCS.

Associated SCI Conditions and Health Status 

The presence of neuropathic pain was present in 39% of the subjects with traumatic central cord syndrome, making it the second most prevalent associated SCI condition among those included in this study. Neuropathic pain decreased the SF-36 PCS by 4 points (95% confidence interval [CI], −8.3 to 0.3) and although it was not statistically significant, it was very close based on the width of the CI. These results are similar to what has been reported in other studies. Westgren and Levi12 reported that neuropathic pain had a large effect (Cohen effect size, 1.27) on the SF-36 bodily pain domain and a medium effect on the SF-36 role−physical domain (Cohen effect size, .45). The SF-36 domains bodily pain and role−physical are both used to calculate the PCS and therefore can be compared with our results.27 Westgren and Levi12 did not adjust for the confounding effect of personal factors, time since injury, or other associated SCI conditions, so results from our study may provide a more accurate estimate of the effect. It has been suggested that instruments that specifically measure pain interference with physical functioning may be more sensitive in clinical research and this should be assessed in future studies.35

The presence of neuropathic pain was significantly associated with subjects being dissatisfied with their current symptoms. This finding is not surprising because neuropathic pain is very difficult to treat and approximately one third of people with neuropathic pain rate it as severe.36 The impact of neuropathic pain on mental health in our study was influenced by the degree of paralysis. An interaction was detected between neuropathic pain and follow-up motor score whereby people with neuropathic pain and a lower follow-up motor score had disproportionately worse mental functioning than those with a lower follow-up motor score but without neuropathic pain. Our research therefore suggests that motor function is an important factor to consider when measuring the effect of neuropathic pain on mental functioning. Because motor score was not included in studies12, 19 that assessed the effect of neuropathic pain on mental functioning it is not possible to compare our results. In studies that have investigated the relationship between neuropathic pain and motor function or completeness of the SCI, the findings are mixed because some studies show a relationship and others do not.37

It has been suggested that the process of living with disability must be considered when evaluating health status and QOL because people may undergo a response shift.38 Response shift theory states that people with chronic conditions such as SCI undergo a change in how they view their health due to a change in their internal standards or in their values (reprioritization), or in how they conceptualize health.38 It may be the effect of regaining motor function that enables people to cope more effectively with neuropathic pain, whereas in people who do not regain motor function, that may be an additional negative influence in addition to having neuropathic pain that impacts their mental functioning. Future research must address the management and treatment for people with neuropathic pain and this research must be comprehensive, considering including biologic, psychosocial, and environmental factors.35

The presence of spasticity was the most prevalent associated SCI condition, affecting 59% of the study sample, which is similar to a recent study of people with traumatic central cord syndrome by Aito et al,39 who reported 54%. The effects of spasticity are variable. Spasticity has been reported to be problematic, affecting activities of daily living in some studies40, 41 and in other studies it has been reported to be beneficial or have no effect.40, 41, 42 In this study, spasticity was not associated with the SF-36 PCS after adjusting for confounders. Some of the discrepancies in the literature may be related to differences in the approach to measuring spasticity. In the study by Westgren and Levi,12 spasticity was defined as being problematic and as impacting on well-being or daily life. When the Cohen effect size was calculated for the association between spasticity and SF-36 domains, subjects who rated their spasticity as beneficial were not included in the analysis.12 However, in our study, subjects with traumatic central cord syndrome were just asked if they had spasticity or not; we did not inquire about interference, so all subjects were included whether or not their spasticity was problematic or beneficial. The introduction of instruments that specifically measure spasticity and include items regarding the problematic and beneficial aspects will assist in further clarifying the effect of spasticity on physical functioning.42

In addition, the presence of spasticity had no effect on the SF-36 MCS or the rating of satisfaction with symptoms in this study after adjusting for confounders. A recent qualitative study by Mahoney et al42 described how people with SCI experience spasticity. Mahoney reported that each person's experience with spasticity was unique and each had different ways to manage spasticity.42 Subjects described how they preferred to control their spasticity rather than to suppress it, allowing them to make spasticity have a more positive effect on their everyday life.42 Spasticity can also be controlled with medication and surgical treatments, which are considered to be highly successful and effective.36 Therefore, it seems that people can manage their spasticity and adapt to having spasticity, which would explain the lack of impact on mental health and satisfaction with symptoms reported in our study. Future studies similar to the work being done by Mahoney42 will assist rehabilitation professionals in understanding the beneficial and negative impact of spasticity on physical and mental function following SCI, which will ensure that spasticity is managed to optimize participation in life activities.

The effect of bowel, bladder, and sexual dysfunction had a strong effect on physical functioning on subjects with traumatic central cord syndrome resulting in a reduction in the SF-36 PCS by 6.9 points. No effect was found between bowel, bladder, and sexual dysfunction and mental function or symptom satisfaction after adjusting for confounders. The association between problems related to bowel,12, 20, 43 bladder,10, 12, 19, 20, 21 and sexual function12 on physical and mental function has been reported in the literature. Results from this study indicating the negative effect of problems with bowel and bladder on physical function are consistent with the literature10, 12, 19, 20, 21 and adds to the literature by quantifying the estimate of this effect. Reasons that bowel and bladder dysfunction has such a negative influence on physical function include fear of embarrassment about incontinence or reliance on devices such as catheters or diapers as well as discomfort and the time required to manage these problems.20, 21 The effect of bowel and bladder dysfunction on mental function is not as clear in the literature. Hicken et al20 controlled for factors such as age, education, sex, race, occupation, and impairment level and reported no difference in mental function in people with SCI who require assistance for bowel and bladder management compared with those who are independent, which is similar to this study. It is evident that future research should include more qualitative studies to further explore why bowel and bladder problems affect physical function but not mental function similar to the effects seen for spasticity. Based on our findings and previous studies it is evident that interventions must be directed at treating or reducing the effect of bowel and bladder problems to maximize a person's physical functioning.

Because 78% of subjects reported at least 2 or more problems with bowel, bladder, and sexual function we did not attempt to assess their independent effect. Based on our results it appears that problems with sexual function affect physical functioning. However, it is not clear how problems related to sexual function contribute to this negative effect on physical function versus problems related to bowel and bladder function. Very little research has been done assessing the impact of sexual function problems on health status and so it is difficult to determine how our results compare with those of other studies. In a study by Westgren and Levi,12 problems with sexual function did not affect physical and mental function. Future research should further investigate the effect of sexual function on physical and mental functioning, from both a qualitative and quantitative perspective. A recent study by Anderson et al44 reported that sexual function is a priority for people living with SCI and has a tremendous impact on their life, which further supports the need for more research in this area.

The effect of the follow-up motor score was significantly associated with all the measures of health status. In terms of the effect of the follow-up motor score on the PCS, we identified an interaction between follow-up motor score with spasticity and education. As expected, an increase in the follow-up motor score produced an improvement in the PCS. This improvement however, was reduced in people who reported spasticity and had a lower level of education, preventing them from benefiting from their motor recovery. As previously discussed, people may adapt to their injury and factors such as education may be a mechanism influencing this process.38 This research indicates the importance of following people as they return to the community and developing a treatment plan to manage problems with spasticity so that physical function can be maximized. Rehabilitation providers must ensure people with SCI are aware that spasticity can be successfully controlled with medication and surgical treatments.36 It is also important that the people are intimately involved in developing the treatment plan so spasticity can be effectively managed and have a positive effect on their everyday life. Finally, it is not too surprising that people who had diminished motor recovery were dissatisfied with their current symptoms. Those with tetraplegia rank regaining arm and hand function as their highest priority, in terms of enhancing their QOL.5 Motor recovery leads to greater independence and the role of assistive devices with the advances in technology and other clinical interventions such as the timing of surgery hold tremendous promise for maximizing recovery following SCI.1

Associated SCI Conditions and QOL 

QOL was assessed in this study using a single question rating overall QOL (numeric QOL rating scale). Results from this study indicate that ratings of QOL were not affected by the associated SCI conditions neuropathic pain, spasticity, or bowel, bladder, and sexual dysfunction after adjusting for confounders. Follow-up motor score, however, did have a positive effect on QOL (OR=1.7; 95% CI, 1.1–2.7) in the third stage of the analysis. In the literature, there are conflicting results regarding the impact of the associated SCI conditions assessed in this study and QOL. Pain19, 45, 46, 47 and spasticity19 have been reported to be negatively associated with QOL in people with SCI. There are, however, other studies that have found no relationship between pain48 or bladder problems19, 48 and QOL. In terms of the follow-up motor score, the research published to date has not shown a relationship between measures assessing impairment and QOL.49, 50, 51 A meta-analysis by Dijkers49 reported a very minor association that was not significant (mean correlation, −.05; 95% CI, −.12 to .02) but only 1 of the 16 studies used the follow-up motor score and the majority used broad categories such as type of SCI or level of the lesion. As stated by Dijkers,49 there is a need for more longitudinal studies using the actual follow-up motor scores as a measure of impairment, rather than just classifying people as tetraplegic versus paraplegic, to determine the effect of impairments such as motor function on QOL.

In addition, QOL is now frequently measured but there is no consensus on how to define it.52, 53 The terms life satisfaction and QOL are often used interchangeably; however, there is no consensus in the literature if life satisfaction is equivalent to QOL.52, 53 This study did not find a relationship between overall QOL and most of the associated SCI conditions. Because many of the previous studies compared associated SCI conditions with life satisfaction19, 47, 48 and we assessed overall QOL our results may not be directly comparable due to differences in how QOL is operationalized.

Study Limitations 

There are several limitations to our study that need to be considered when interpreting the results. Our sample of subjects included only traumatic central cord syndrome, which is an incomplete injury in the cervical region, and therefore our results cannot be generalized to people with either a complete injury or an incomplete injury in the thoracic or lumbar regions of the spine. The sample size was relatively small (N=70) and future studies that include other types of SCI will make it feasible to have larger samples. Because it was a secondary analysis of an existing data set, people who had comorbid conditions that affected their physical function were excluded and this reduces the generalizability of our research findings. In addition, we conducted a cross-sectional assessment of subjects who were treated at 1 major trauma center. Future studies should include prospective study designs with a priori hypotheses that collect data longitudinally in order to establish causal relationships between the effects of associated SCI conditions on health status and QOL. The conceptual model presented in this study should be further enhanced to include secondary conditions resulting from SCI such as pressure ulcers and overuse injuries. In addition, not only should new instruments be developed to obtain information on the presence or absence of associated and secondary conditions, but they should also should include questions about the severity of these conditions. By developing a comprehensive model and collecting the data longitudinally the relationship between the secondary SCI conditions and the associated SCI conditions can be determined and the effect on patient outcomes can be measured.