Increased Risk of Deep Vein Thrombosis and Pulmonary
Thromboembolism in Patients with Spinal Cord Injury:
A Nationwide Cohort Prospective Study
Wei-Sheng Chung
a,b, Cheng-Li Lin
c, Shih-Ni Chang
c,g, Hui-Ann Chung
d,
Fung-Chang Sung
c,e, Chia-Hung Kao
e,f,⁎
Introduction
A spinal cord injury (SCI) refers to any injury to the spinal cord caused by trauma instead of a disease. The location of the injury on the spinal cord determines the severity of the injury. After an SCI occurs, the functions of the body located below the point of injury become impaired. This impairment includes sensory and motor defcits, breathing
diffculties, and bowel or bladder dysfunction. Studies have also demonstrated
that cardiovascular morbidity developed in SCI patients [1,2].
The Virchow triad describes 3 major factors that contribute to
thromboembolism: hypercoagulability, hemodynamic change (stasis or
turbulence), and endothelial dysfunction [3], whereas stasis and
hypercoagulability are 2 major factors that can be attributed to the development of deep vein thrombosis (DVT) and pulmonary thromboembolism
(PE) in SCI patients. Several studies have demonstrated that an
impaired venous return from lower limb paralysis and abnormal coagulation
factors can predispose patients to thrombogenesis [4]. DVT and subsequent
PE remain signifcant causes of morbidity and mortality among
SCI patients [5]. The combined fndings of these studies indicate that
DVT and PE constitute venous thromboembolism (VTE).
The incidence of VTE among acute SCI patients has been reported to range between 5.4% and 90% in Western countries, depending on the
study population, the active surveillancemodality, and the use of pharmacological
prophylaxes [4,6,7]. However, studies on the incidence of DVT
and PE among SCI patients in Asian people are scant. In this study, we investigated whether the increased risk of DVT and PE among SCI patients
still existed after controlling for potential risk factors by implementing a nationwide prospective cohort study. Methods
Data Sources
The data for this study were retrieved from the Taiwan National
Health Insurance Research Database (NHIRD), which is a claims database maintained by the Department of Health and the National Health
(NHI) program, which was launched in 1995, provides medical insurance
for nearly 99% of Taiwan’s residents [8]. The database comprises
comprehensive information on clinical visits made by each insured person, such as the scrambled patient identifcation number, demographic
characteristics, inpatient and outpatient dates, diagnostic codes in accordance with the International Classifcation of Diseases, Revision 9,
Clinical Modifcation (ICD-9-CM), and patient prescriptions. All the reimbursement and claims data used for insurance payments are scrutinized
by administrative specialists and by peer review under the
universal health insurance system. The NHRI established the NHIRD, which provides all claims data in electronic format to the public for research purposes. The details of the claims fles have been described in
our previous studies [9,10]. Several studies conducted in Taiwan have
demonstrated the high accuracy and validity of the ICD-9 diagnosis included
in the NHIRD [11,12].
Study Participants
We selected cohort study patients based on frst admissions for SCI (ICD-9-CM 806 and 952) during the 1998–2008 period. Patients were excluded if they had a previous diagnosis of DVT (ICD-9-CM 453.8) or PE (ICD-9-CM 415.1) before the index date. The non-SCI controls were 1:4 frequency-matched for sex, age (span of 5 y), and the index year to the SCI group from the registry of benefciaries. The same exclusion criteria were also applied to non-SCI patients.
Outcome Measurement and Comorbidities
We identifed the frst diagnosis of DVT or PE by using the hospitalization records as the study endpoint. All of the study participants were
followed fromthe index date to the study endpoint, December 31, 2010, or when the patient withdrew from the database.
We also incorporated an inpatient diagnosis fle to ascertain the comorbidities, including atrial fbrillation (ICD-9-CM 427.31), hypertension
(ICD-9-CM 401–405), diabetes (ICD-9-CM 250), hyperlipidemia (ICD-9-CM 272), cerebrovascular accident (CVA; ICD-9-CM 430–438), congestive heart failure (ICD-9-CM 428), lower leg fracture or surgery (ICD-9-CM 820, 821, 823, 81.51, 81.52, 81.53, and 81.54), and cancer (ICD-9-CM 140–208).
Statistical Analysis
We compared the baseline characteristics between the SCI cohorts and non-SCI cohorts by using the chi-square test. The incidence densities
were determined under the Poisson assumption. To estimate the cumulative incidence of DVT or PE risks in the SCI patients and non-SCI cohort, we performed survival analysis by using the Kaplan-Meier
method, and signifcance was determined using the log-rank test. Multivariate Cox proportional-hazards regression was used to examine the
effect of an SCI on the risk of DVT or PE development, which was determined by the adjusted hazard ratio (HR) with a 95% confdence interval
(CI). The multivariable model was used to control for age, sex, comorbidities of atrial fbrillation, hypertension, diabetes, hyperlipidemia,
CVA, congestive heart failure, lower leg fracture or surgery, and cancer. A two-tailed P value of b.05 was considered statistically signifcant. All statistical analyses were performed using SAS statistical software (version 9.2 forWindows; SAS Institute, Inc., Cary, NC, USA).
Results
Table 1 shows the distributions of age, sex, and comorbidities of the 2 cohorts. The mean age of the 2 cohorts was approximately 50 years, and 62.7% of the patients were men. The SCI cohort was more likely to have atrial fbrillation (1.18% vs 0.68%, P b .0001), hypertension (17.1% vs 7.45%, P b .0001), diabetes (10.9% vs 4.24%, P b .0001), hyperlipidemia (4.13% vs 1.88%, P b .0001), CVA (6.78% vs 3.41%, P b .0001),
congestive heart failure (2.86% vs 1.13%, P b .0001), lower leg fracture or surgery (6.16% vs 1.59%, P b .0001), or cancer (2.13% vs 1.63%, P b .0001) compared with the non-SCI cohort.
In estimating DVT risk, the overall incidence rate of the SCI cohort was 8.99 per 10,000 person-years, and the sex-specifc incidence rate forwomen andmenwas 8.54 and 9.26 per 10,000 person-years, respectively (Table 2). Comparedwith the non-SCI cohort, the SCI patients exhibited a signifcantly increased risk of DVT,with an incidence rate ratio (IRR) of 2.55 (95% CI = 2.48, 2.63) and an adjusted HR of 2.46 (95% CI = 2.11, 2.87). The IRR of DVT was signifcantly higher for men (IRR = 3.05, 95% CI = 2.94, 3.17), patients younger than 50 years (IRR = 8.68, 95% CI = 8.28, 9.10), and patients without comorbidities (IRR = 2.29, 95% CI = 2.21, 2.37). However, the risk of DVT rosemarkedly with increasing age after adjusting for sex and comorbidities. The
results of the log-rank test and the cumulative incidence curve of DVT,
as shown in Fig. 1, indicate that the SCI patients had a signifcantly
higher DVT incidence than the non-SCI group did (log-rank P b .001).
incremental risk of DVT development, regardless of the follow-up period. The HRs of DVT development seemed to decrease as the follow-up period increased: the SCI patientswith b 0.25 years of follow up had the highest HR (HR = 16.9, 95% CI = 8.89, 31.9), followed by patients with 0.25–1 years (HR = 4.00, 95% CI = 2.71, 5.91), 1–4 years (HR = 1.94, 95% CI = 1.48, 2.56), 4–8 years (HR = 1.70, 95%
CI = 1.26, 2.30), and N 8 years of follow up (HR = 1.68, 95% CI = 1.07, 2.66).
To estimate the PE risk, a comparison between the patients with SCI and without SCI revealed that the SCI patients were more likely to develop PE (IRR = 1.70, 95% CI = 1.65, 1.76; adjusted HR = 1.57, 95%
CI = 1.23, 1.99) compared with the non-SCI cohort.
The sex-specifc IRR of PE for the SCI cohort compared with the non-SCI cohort was signifcant for both females (IRR = 1.86, 95% CI = 1.76–1.96) and males (HR = 1.56, 95% CI = 1.49–1.63). The
age-specifc analysis revealed that the incidence increased with age in both cohorts. However, the IRR of PE was the highest in the youngest
group (IRR = 5.20, 95% CI = 4.96, 5.44) and decreased with age. SCI patients without any comorbidities had a higher incidence of PE than non-SCI patients without any comorbidities did (2.07 vs 1.33 per 10,000 person-years; IRR = 1.56, 95% CI = 1.50–1.62). The risk of PE rose markedly with increasing age after adjusting for
sex and comorbidities (Table 2). The Kaplan-Meier curves suggested
that the SCI patients had a higher risk of PE development compared with that of the non-SCI cohort (log-rank P b .001). By conducting an analysis stratifed by follow-up periods,we observed that the SCI cohort had an increased risk of developing PE, whichwas signifcantly high for the period of b 0.25 years (HR = 3.64, 95% CI = 1.36, 9.75), followed by 0.25–1 years (HR = 2.61, 95% CI = 1.33, 5.11), 4–8 years
(HR = 1.19, 95% CI = 0.75, 1.88), and N 8 years (HR = 2.17, 95%
CI = 1.25, 3.77) (Fig. 2).
When considering the effects of the SCI location on the risk of DVT and PE development, we determined that the C-spine SCI patients were at the greatest risk of developing DVT (HR = 2.96, 95% CI = 2.43, 3.61), followed by the T-spine SCI patients (HR = 2.68, 95% CI = 2.05, 3.50), and the L-S-Co-spine SCI patients (HR = 1.60, 95% CI = 1.25, 2.05) after adjusting for age, sex, and comorbidities. Concerning the risk of PE, the T-spine SCI patients were determined to
have the greatest risk of developing PE (HR = 1.69, 95% CI = 1.10, 2.58), followed by the L-S-Co-spine SCI patients (HR = 1.52, 95% CI = 1.08, 2.15), and the C-spine patients (HR = 1.48, 95% CI = 1.04,
2.12) after adjusting for age, sex, and comorbidities (Table 3).
Overall, the risk of VTEwas 2.17-fold greater in the SCI cohort than in the non-SCI cohort (95% CI = 1.90–2.48) after adjusting for age, sex, and comorbidities.
Discussion
This nationwide cohort study demonstrated that SCI patients have a
2.46-fold adjusted HR of DVT and a 1.57-fold adjusted HR of PE comparedwith that of the general population. Moreover, this study demonstrated
that the incidence of DVT was 89.9 per 100,000 person-years,
and the incidence of PE was 32.4 per 100,000 person-years among SCI patients in Taiwan. The incidence of DVT and PE among SCI patients
was lower in Taiwan compared with that of Western countries [4,6,7].
According to ethnic-comparison studies, the incidence of VTE among Asian-Pacifc Islanders and Asian Americans appeared to be 2.5-fold to
5-fold lower than that among Caucasians [13,14]. The cause of lower incidence
in the Taiwan population remains unclear; however, itmight be
associated with ethnic differences and distinct dietary habits [15,16].
Most of the SCI patientsweremen and younger adults. This fnding is
consistentwith that of previous studies [17,18]. The SCI patients of both
sexes had a higher incidence rate of DVT development compared with that of the non-SCI cohort. Men with an SCI had a higher incidence rate of DVT than women with an SCI did, despite no statistically signifcant differences occurring after controlling for age and comorbidities.
Regarding PE, SCI patients also had a higher incidence rate for both
sexes compared with that of the non-SCI cohort. The incidence rate of PE amongwomenwith an SCIwas higher than that amongmen, despite
no statistically signifcant differences emerging after controlling for age and comorbidities. Several population-based studies have reported inconsistent differences in the incidence of DVT and PE between men
and women [19–21]. These epiphenomena might be associated with
the use of a different study design, age groups, and follow-up periods. The incidence rate of DVT and PE rosemarkedly with increasing age in patients of both sexes. The result is consistent with that of previous
studies [22–24]. The age-specifc crude relative risk indicates that the effect
among middle-aged and older adults. However, the adjusted HR of DVT and PE increased considerably with age after controlling for sex and comorbidities. This result is consistentwith that ofWestern studies
[25]. As people age, activity levels decrease and cardiopulmonary systems
deteriorate. The risk factor for thrombosis associatedwith advancing age may thus accentuate the effects of SCI on the risk of DVT and PE
development [26]. The proportion of comorbidities interacting with SCI
that develop into DVT and PE might be higher among older adults. Patients with any comorbidities have an increased risk of developing
DVT and PE, as Table 2 shows.
The risk of developing DVT and PE following an SCI is highest within
the frst 3 months, as shown in Figs. 1 and 2, a fnding that is consistent
with that of previous studies [6,27]. Ploumis et al. suggested that
thromboprophylaxis in SCI patients should begin as early as possible
once it is deemed safe against potential bleeding complications [27].
Atio et al. demonstrated that the early application of pharmacological therapy such as low-molecular-weight heparin therapy, and mechanical treatment such as permanently dressed gradient elastic stockings and the external sequential pneumatic compression of the lower limbs
for DVT prevention, produce a marked reduction in VTE [6]. Because of
the mobility limitation accompanying venous stasis after an SCI, early active or passive mobilization is also recommended.
Maung et al. demonstrated that the rate of VTE was highest in the T1-6 SCI group used in their study, and that L-spine SCI patients had
the lowest rate of VTE [25]. However, they did not control for age, sex,
and comorbidities. In the present study, we determined that the incidence rate of DVT and PE was highest in T-spine SCI patients. However,
after adjusting for age, sex, and comorbidities, C-spine SCI patientswere revealed to have the highest risk of DVT development compared with the patients with T-spine and L-S-Co-spine SCI. Conversely, T-spine
SCI patients have the highest risk of developing PE compared with patients with C-spine and L-S-Co-spine SCI, after adjusting for age, sex,
and comorbidities. A plausible explanation might be that the disability (quadriplegia or paraplegia) related to SCI depends on the level of injury.
In the SCI patients, decreased physical activity levels resulted in obesity, metabolic syndrome, and cardiopulmonary system deterioration. In addition, an increased prevalence of comorbidities related to the
cardiovascular risk factors for the SCI patientswas observedwhen comparing
them with the age- and sex-matched controls [28–30]. Therefore,
we considered the higher prevalence of cardiovascular risk
factors to be a controlling factor for the SCI patients during analysis
(the adjusted HRs are shown in Tables 2 and 3).
The strength of this study is that it was a nationwide prospective longitudinal cohort study on the risk of DVT and PE development in Asian peoplewith an SCI. These fndings can be generalized to the general population. However, several limitations must be consideredwhen interpreting these fndings. First, the lack of drug information, such as hormone replacement therapy, contraceptive drugs, and anticoagulants, required to adjust for the outcomes of interest could be a major
limitation. Second, the NHIRD does not accurately distinguish between complete and incomplete SCIs, which might be another limitation. Third,we used ICD-9 codes to defne the outcomes of interest, especially for DVT, which might have led to misclassifcation. Regardless, allmedical reimbursements and physician diagnoses are scrutinized by administrative specialists and by peer review under the universal health
insurance system in Taiwan to minimize misclassifcation.
In conclusion, this nationwide study of 47,916 SCI patients with approximately 309,000 follow-up person-years demonstrated that SCI patients
have a 2.46-fold and 1.57-fold increased risk of developing DVT and PE, respectively, compared with that of the general population. In addition, the greatest risk of DVT, up to a 16.9-fold increase, and a 3.64-fold increased risk of PE developed within 3 months after an SCI occurred. These results emphasize the importance of amultidisciplinary team adopting an integrated approach to reduce potential risk factors for SCI patients.
