Spinal cord injury increases the risk of Type 2 diabetes:
a population-based cohort study
Yun-Ju Lai, MDa, Cheng-Li Lin, MScb, Yen-Jung Chang, PhDb, Ming-Chia Lin, MScc,
Shih-Tan Lee, MDd, Fung-Chang Sung, PhDb,e, Wen-Yuan Lee, MDf,g, Chia-Hung Kao, MDf,h,*
Introduction
Spinal cord injury (SCI) is a common and devastating
event that can result in severe disability, increased mortality, and increased use of medical resources. In Taiwan, the
reported incidence of SCIs is approximately 2.46 per 10,000 person-years, 61.2% of which are traumatic cases [1]. Cervical SCIs are predominant. Elderly men, people with a low socioeconomic status, people in rural areas,
and those with preexisting comorbidities such as hypertension, diabetes, dyslipidemia, chronic obstructive
pulmonary disease, chronic renal failure, and Parkinson disease are more likely to suffer from SCIs than other patient groups [1].
An SCI induces various changes in systemic physiology that can lead to many complications that considerably affect the function and quality of life. The life expectancy of survivors of SCIs is approximately 90% of that of the general population [2,3]. The most common causes of death after a traumatic SCI are diseases of the respiratory system, followed by cardiovascular events. These two causes of death include more than 50% of all deaths of SCI patients [2]. Patients with SCIs typically live sedentary lifestyles and experience weight gain and metabolic changes that
may cause premature coronary heart disease [4,5]. In addition, patients with SCIs are reportedly at a higher risk of insulin resistance, atherogenic lipid profile, and metabolic
syndrome, which are precursors of diabetes and coronary artery disease, compared with age-matched people from the general population [4,6]. They tend to possess less lean body mass and greater adiposity than the non-SCI populations [7]. A recent study of veterans with SCIs identified
approximately 20% as obese and 33% as overweight [5].
Increased prevalence of cardiovascular risk factors in SCI patients leads to higher incidence of hypertension and ischemic heart disease than non-SCI populations [6,8].
Previous reports on the risk and prevalence of diabetes among SCI patients are limited and conflicting. In 2006, Lavela et al. [9] suggested that the prevalence of diabetes is higher in veterans with SCIs than in the civilian population.
Wahman et al. [10] further reported that patients with paraplegia exhibited higher prevalence of diabetes mellitus, hypertension, and dyslipidemia compared with the general population. However, a study by Banerjea et al. [11] identified that the prevalence of diabetes in SCI patients was
similar to that in the general older population. The results of a recent study by LaVela et al. [12] further indicated that the prevalence of diabetes was similar in men with and without SCIs.
Therefore, the primary aim of this study was to compare the risk and incidence rate (IR) of diabetes in SCI and non- SCI groups, based on data from Taiwan’s National Health Insurance Research Database (NHIRD).
Material and methods Data sources
A retrospective study was conducted using claims data obtained from Taiwan’s National Health Research Institutes (NHRI). Taiwan initiated its single-payer National Health
Insurance (NHI) program in March 1995 that covered approximately 99% of Taiwan’s 23.74 million residents in
2009 [13]. The NHRI maintains and updates the NHIRD.
We analyzed data from the NHIRD for the period 1997 to 2010; the data are released by the NHRI for public use. The scrambled identifications of insured individuals are available through link files that include the registry of
medical facilities, details of inpatient orders, ambulatory care data, and sociodemographic information for each patient.
Diagnoses are coded according to the International
Classification of Disease, Ninth Revision, Clinical Modification (ICD-9-CM). We confirm that all data were deidentified
and analyzed anonymously. In addition, this
study was approved by the Ethics Review Board of China Medical University (CMU-REC-101-012).
Study patients
In this study, patients with newly identified SCIs (ICD- 9-CM codes 806 and 952) from 1997 to 2010 were selected for data analysis. Patients aged 20 years or younger and those with a previous diagnosis of diabetes were excluded from the study. The remaining patients were included in the SCI cohort. The date on which the SCI patients received a diagnosis according to the inpatient claims data
was defined as the index date. We classified SCI patients into subgroups: C-spine SCI (ICD-9-CM codes 8060, 8061, 9520, 95200, 95201, 95202, 95203, 95204, 95205, 95206, 95207, 95208, and 95209), complete T-spine SCI (ICD-9-CM 806.21, 806.26, 952.11, and 952.16), incomplete T-spine SCI (ICD-9-CM codes 8062, 8063, and
9521), and lumbar, sacral, and coccygeal (L-S-Co-spine) SCI (ICD-9-CM codes 8064, 8065, 8066, 8067, 8068,
8069, 9522, 9523, 9524, 9528, and 9529). A non-SCI comparison cohort was randomly selected from the NHI beneficiaries
aged 20 years and older and matched with the SCI cohort in a 4:1 ratio based on age (every 5 years), sex, and
index date.
Outcome measures
For each patient, the duration of a follow-up evaluation
was measured from the index date until the first of the following occurred: diagnosis of Type 2 diabetes (ICD-9-CM
codes 250), withdrawal from the insurance system, the end of 2010, or death. Baseline comorbidity histories of hypertension (ICD-9-CM codes 401–405), hyperlipidemia (ICD-
9-CM code 272), chronic kidney disease (ICD-9-CM code 585), stroke (ICD-9-CM codes 430–438), coronary heart
disease (ICD-9-CM codes 411.1, 411.81, 411.89, 413, 414.0, 414.8, and 414.9), and congestive heart failure
(ICD-9-CM code 428) were identified according to their diagnoses in the inpatient claims data before the index date.
Statistical analysis
Demographic factors, including age, sex, and comorbidities, were compared between the SCI and non-SCI cohorts.
A chi-square test was used to evaluate the differences between the groups. The IRs of Type 2 diabetes from the
follow-up evaluation until the end of 2010 were calculated for the two groups. In addition, the IR ratio (IRR) of the SCI cohort compared with the non-SCI cohort was calculated.
Multiple Cox proportional hazard regression analyses were conducted to calculate the related hazard ratios (HRs) and 95% confidence intervals (CIs) for risk of Type 2 diabetes associated with SCI. The diabetes-free survival rates
were estimated using the Kaplan-Meier method, and the log-rank test was used to compare the survival curves of the SCI and non-SCI cohorts. All analyses were performed using the SAS System for Windows, Version 9.1 (SAS Institute Inc., Cary, NC), and results were considered statistically significant if two-tailed p values were less than .05.
Results
During the evaluation period, we identified 52,420 patients for inclusion in the SCI cohort and 209,680 patients
for inclusion in the non-SCI comparison cohort (Table 1).
Patients were predominantly men (63.6%), possessing a mean age of 51.6 years in the non-SCI and 51.7 years in the SCI cohorts. Comparison of the SCI and the non-SCI cohorts revealed that patients in the SCI cohort were more likely to have hypertension (16.20% vs. 7.19%), hyperlipidemia (3.68% vs. 1.72%), chronic kidney disease (1.01%
vs. 0.49%), stroke (6.23% vs. 3.20%), coronary heart disease (6.84% vs. 3.33%), and congestive heart failure
(2.56% vs. 1.10%) than patients in the non-SCI cohort.
The mean duration of follow-up evaluations was 5.98 years in the SCI and 6.50 years in the non-SCI comparison cohorts.
The incidence of Type 2 diabetes was slightly higher
in the SCI than in the non-SCI cohorts (22.1 per 10,000 person-years vs. 17.2 per 10,000 person-years), with an IRR of 1.28 (95% CI51.24–1.32) (Table 2). A Cox proportional hazard regression model analysis further revealed
that patients with SCIs were 1.33 times more likely to develop Type 2 diabetes than non-SCI patients (95%
CI51.22–1.45) after controlling for age, sex, and comorbidity.
The IRR of Type 2 diabetes was higher in men than in women (IRR51.40, 95% CI51.35–1.46 vs. IRR51.13, 95% CI51.07–1.19). Generally, the incidence of Type 2 diabetes increased with age in both cohorts. The IRR of Type 2 diabetes was highest in the youngest evaluated age group (20–34 years; IRR57.27, 95% CI56.82–7.75), and the adjusted HR for elderly patients compared with those aged 20 to 34 years was 17.4 (95% CI514.1–21.6).
The IR of Type 2 diabetes was higher in patients with comorbidity than in patients without comorbidity. Table 3
shows the IRs, crude and adjusted HRs, and 95% CIs of
Type 2 diabetes according to different levels of lesion of SCI. Compared with the non-SCI cohort, patients with Cspine
injuries exhibited a 20% higher risk of Type 2 diabetes (adjusted HR51.20, 95% CI51.06–1.36), whereas those with complete T-spine injuries exhibited a 135%
higher risk and those with incomplete T-spine injuries exhibited a 60% higher risk (adjusted HR51.60, 95%
CI51.34–1.92). Patients with lumbar, sacral, and coccygeal spine injuries (L-S-Co-spine injury) exhibited a 38% higher risk of Type 2 diabetes than the non-SCI comparison cohort (adjusted HR51.38, 95% CI51.21–1.57). Table 4 shows the adjusted HRs for development of diabetes based on age, sex, comorbidity, and the level of lesion of SCI. Male SCI patients were more likely to suffer from diabetes than female SCI patients (adjusted HR51.23, 95% CI51.04–
1.44). Older SCI patients were more likely to suffer from diabetes than younger SCI patients (adjusted HR54.26 in patients older than 65 years, 95% CI53.16–5.74). Patients with complete T-spine injuries were more likely to suffer from diabetes than patients with C-spine injuries (adjusted
HR52.13, 95% CI50.95–4.79). The Kaplan-Meier method revealed that the cumulative incidence of Type 2 diabetes during the follow-up period was 0.42% higher in the SCI patients than in the non-SCI patients (log-rank test p!.0001) (Figure).
Discussion
Our results revealed that SCI patients were predominantly men, were older than 65 years, and possessed comorbidities such as hypertension, hyperlipidemia, chronic
kidney disease, stroke, coronary heart disease, and congestive heart failure (Table 1). In addition, most of these patients suffered from C-spine injuries (Table 3). This is
consistent with the epidemiologic report from a Taiwanese nationwide study conducted by Wu et al. in 2011 [1]. The IR of Type 2 diabetes was considerably higher in the SCI cohort than in the non-SCI cohort. The IR of Type 2 diabetes increased with age and comorbidities in both cohorts (Table 3). Among patients with SCIs, those with complete T-spine injuries exhibited the highest risk of developing diabetes (Tables 3and 4).
Some of our findings require further investigation. In our
study, the case number of patients with complete T-spine injuries was 274, which was considerably fewer than patients
with C-spine, incomplete T-spine, and L-S-Co-spine injuries (Table 3). The expected functional recovery after a complete T-spine injury is more satisfactory than that after a complete C-spine injury. However, our study revealed that patients with complete T-spine injury possessed
the highest risk of diabetes. Some of the patient information was not available in our data set, such as data on associated injuries, physical activity, body weight, lipid profile, blood pressure, use of medication, and rehabilitation resources. In addition, we did not possess data regarding mortality. Early death rates after admission for traumatic SCI range from 4% to 20% [14]. Other reasons that patients with C-spine SCIs are less likely to develop diabetes may exist. DeVivo et al. [15] discovered that patients with C1–C3 injuries possess a 6.6-fold increased risk of death, C4–C5 injuries
a 2.5-fold increased risk, and C6–C8 a 1.5-fold increased
risk when compared with those patients with SCIs that occurred in the thoracic or lower cord. A reduced life expectancy
might explain why patients with C-spine SCIs are less
likely to develop diabetes. In our study, patients with Cspine injuries might possess an increased risk of mortality
and reduced life expectancy that might explain why patients with T-spine injuries demonstrated higher risk of developing diabetes than those with C-spine injuries.
Previous studies have reported rates of diabetes in patients with SCIs as ranging from 13% to 22% [1,16,17].
In 2006, Lavela et al. [9] reported that the overall prevalence of diabetes in patients with SCIs was 20% that was
three times higher than in the general population. In our study, the identified diabetes IR in the SCI group was 22.1
per 10,000 person-years. The diabetes IR in the non-SCI group was 17.2 per 10,000 person-years that was higher than the rate in Taiwan’s general population, namely, 7.6
and 6.9 per 1,000 person-years in men and women, respectively [18]. According to our results, the IR of diabetes in
SCI patients was 1.28 times that seen in the non-SCI population and approximately three times the reported IR of
the general population. Because the non-SCI comparison cohort was matched with the SCI on age, both cohorts were predominantly elderly men and exhibited higher diabetes risks than the general population.
Table 4 displays the risk factors for diabetes in the SCI and non-SCI patients. According to the American Diabetes Association’s position statement in 2012, risk factors for diabetes include more than 45 years of age, body mass index (BMI) greater than 25 kg/m2, physical inactivity, having a first-degree relative with diabetes, being a member of a high-risk ethnic population (eg, African American, Latino, Native American, Asian American, or Pacific Islander), being a woman who delivered a baby weighing more than 9 lb or has been diagnosed with gestational diabetes mellitus, hypertension ($140/90 mm Hg or undergoing therapy
for hypertension), a high-density lipoprotein cholesterol
(HDL-C) level of more than 35 mg/dL (0.90 mmol/L) or a triglyceride level of more than 250 mg/dL (2.82 mmol/L), polycystic ovary syndrome, hemoglobin A1C more than 5.7%, impaired glucose tolerance, impaired fasting glucose, other clinical conditions associated with insulin resistance (eg, severe obesity, acanthosis nigricans), and a history of cardiovascular disease [19]. Although follow-up studies on SCI patients did not establish risk estimates commonly used for the general population, nearly all risk factors tended to be more prevalent in SCI patients that in ambulatory patients [20]. In our study, in the SCI group, elderly men (mean
age, 51.7618.3 years) with comorbidity, including hypertension, hyperlipidemia, chronic kidney disease, stroke, coronary
heart disease, and congestive heart failure, were
predominant. They possessed a greater number of risk factors for diabetes than the non-SCI subgroups and the general population. In SCI patients, associated injuries, such as injuries to the head, chest, abdomen, pelvis, and extremities are common [21]. Daily energy expenditure is considerably lower in SCI patients than in non-SCI patients because of
lack of motor function, lack of accessibility, and fewer opportunities to engage in physical activity. These patients
tend to engage in limited activity, live sedentary lifestyles, and experience weight gain, nervous system dysfunction, and increased inflammatory response. SCI patients also suffer from abdominal obesity, increased body fat, and decreased lean body mass [22]. In a previous study, BMI of
25 kg/m2 or more was associated with a considerably high
prevalence of diabetes in male veterans with SCIs and disorders.
This risk was particularly pronounced if their BMI was 27.5 kg/m2 or more [23].
This study, which used data from the Taiwan’s NHIRD, reported the diabetic risk in SCI patients. Other studies that have used laboratory data similarly have revealed increased cardiovascular risk in SCI patients. Inflammatory markers that correlate with an increased risk of cardiovascular disease are reportedly more prevalent in patients with SCIs
than in non-SCI patients [24]. A previous study identified
that patients with C-spine injuries experience an enhanced lipid peroxidation process and decreased plasma antioxidant potential [25]. Other studies have reported a high prevalence of dyslipidemia and decreased levels of HDL-C in
SCI compared with non-SCI groups [24,26]. We presumed that inflammatory markers and autonomic dysfunction might be underlying mechanisms that cause the increased diabetes risk in SCI patients.
The spinal cord is essential for autonomic nervous system
regulation of the cardiovascular system because the preganglionic neurons controlling the heart and blood vessels originate
in the thoracolumbar spinal segments. The site and
extent of an SCI determine the degree of autonomic involvement in a cardiovascular dysfunction postinjury [27]. Autonomic
dysfunction caused by an SCI is associated with
several conditions that increase cardiovascular risk, including blood pressure abnormalities, heart-rate variability, arrhythmias, and a blunted cardiovascular response to exercise that
can limit physical activity [20]. After the induction of SCIs in experimental animals, Bravo et al. [28] observed immediate hypotension that they attributed to an autonomic imbalance involving the predominance of parasympathetic
activity. The authors proposed that episodic hypertension might then develop as part of a condition called autonomic dysreflexia. Hypertension reportedly increases in patients with paraplegia, but not in those with tetraplegia [8].
Our study was subject to limitations. First, the NHIRD
provides ICD-9-CM diagnosis codes but not detailed patient information, such as data on BMI, physical activity,
blood pressure, blood glucose levels, serum HDL-C
and triglyceride levels, and family history of Type 2 diabetes, all of which are risk factors for diabetes. Second,
evidence from a retrospective cohort study is generally of lower methodological quality than that from randomized trials because of bias related to adjustment for confounders.
Despite adequate control of confounding
factors, a major limitation of our study was potential bias because of possible unmeasured or unknown confounders.
Third, the diagnoses in the NHI claims primarily serve the
purpose of administrative billing and do not undergo verification for scientific purposes. We were unable to contact
the patients directly to obtain further information because of the anonymity assured by their identification numbers.
However, the data that we obtained on SCIs and diabetes diagnoses were highly reliable because of the considerable sample size. Finally, mortality data were not available in our data set; therefore, we were unable to
conduct an analysis to determine if greater mortality occurs in the C-spine cohort.
Some of our findings require further investigation. Further research should be conducted to investigate the reason that patients with complete T-spine injuries possess high risk and IR of diabetes. Some detailed patient information was not available, such as data on BMI, physical activity, blood pressure, blood glucose levels, serum HDL-C and triglyceride levels, family history of Type 2 diabetes, and mortality.
In conclusion, our findings suggest that patients with SCIs possess higher risk and IR of diabetes than non-SCI patients, particularly patients who are men, aged 65 years and older, exhibit comorbidity, and have experienced complete T-spine injuries.
