Increased Risk of Fracture and Postfracture Adverse Events in Patients With Diabetes: Two Nationwide Population-Based Retrospective Cohort Studies

Although the epidemiology, pathogenesis, prevention, and treatment of diabetes have been well established over the past two centuries (1), diabetes remains a pandemic chronic disease that is projected to reach an estimated global prevalence of 4.4% by 2030 (2). Diabetes also presents a high economic burden, with an estimated cost of 245 billion USD in 2012 in the U.S. (3). Recognized complications of diabetes include cardiovascular and cerebrovascular disease, peripheral vascular disease, renal dysfunction, and microvascular disease (1).

Fracture is one of the major causes of disability and mortality (4). For example, in the U.S. in 2005, more than two million fracture incidents cost an estimated 17 billion USD (5). Osteoporosis (defined as bone mineral density 2.5 SD or more below the average value for premenopausal women) is considered an important contributor to fracture (6) and has been investigated as a comorbidity in diabetic patients (7). Many epidemiological studies document that patients with diabetes had increased risk of fracture (8–27). However, these previous reports were limited by several factors, including a focus on specific populations (8–20), small samples of diabetic patients (11–14,20–22,27), case-control study design (13,21–23), inadequate adjustment for potential confounders (13,14,16–22,24,27), and reporting only a single type of fracture (10,13,16,21,25,26). The impact of diabetes on postfracture outcomes also remains unknown.

We used Taiwan’s National Health Insurance reimbursement claims to investigate the incidence and risk of fracture after adjustment for potential confounding factors for adults with diabetes in a nationwide retrospective cohort study. More importantly, we conducted a nested cohort study to investigate the impact of diabetes on postfracture outcomes.

Taiwan’s National Health Insurance system has registered all medical claims since 1996, and this database is available to researchers with identification numbers of those insured scrambled to protect patient privacy. Sets of information available for this study include sex, birth dates, diagnoses, health care received, medications prescribed, admissions, discharges, medical institutions, and physicians providing services. For research and administrative purposes, Taiwan’s National Health Research Institute has released a data subset of claims data for one million randomly selected insurance enrollees aged 0–113 years in 2005. This random subgroup represents ∼5% of Taiwan’s insured population. Information about medical care for these persons was collected between 1996 and 2008 (28–30).

Insurance reimbursement claims used in this study were from Taiwan’s National Health Insurance Research Database. For protection of personal privacy, the electronic database was decoded with patient identifications scrambled for further public access for research. This study was evaluated and approved by Taiwan’s National Health Research Institutes; this organization’s regulations do not require informed consent because patient identification has been decoded and scrambled (28–30). This study was conducted in accordance with the Declaration of Helsinki.

In this longitudinal cohort of one million insured individuals, we identified as the exposed cohort 32,471 patients age 20 years or older with new diagnosis of diabetes between 2000 and 2003 (without any previous record of diagnosis or treatment for diabetes from the database since 1996) and without a history of fracture before the index date. Previous studies (28,30) considered patients with diabetes as those who had made two visits for outpatient diabetes care. For identification of patients with diabetes more strictly, this study required at least three visits (attributable to diabetes or related complications) for outpatient or inpatient medical services with a principal diagnosis of diabetes within the preoperative 24-month period. During the same index period, we identified 64,942 people age 20 years or older matched by age and sex who had no diabetes diagnosis and no history of fracture before the index date as the nonexposure cohort. Patients with any diagnosis of fracture between 1 January 1996 and 31 December 2003 were excluded to ensure that all study participants were free of fracture at the start of both cohorts. Follow-up started 1 January 2000 and lasted until censoring due to death, loss to follow-up, or other causes by 31 December 2008. We sought to determine whether individuals with diabetes faced an increased risk of fracture.

To investigate the impact of diabetes on postfracture outcomes, a nested retrospective cohort study was conducted. We identified 17,002 hospitalized patients with fracture between 2004 and 2010, including 2,971 patients with prefracture diabetes and 14,031 without diabetes. We compared postfracture septicemia, deep wound infection, pneumonia, urinary tract infection, and mortality for a 30-day period postoperatively between fracture patients with and without prefracture diabetes.

We identified income status by defining low-income patients as those qualifying for waived medical copayment because this status is verified by the Bureau of National Health Insurance. Population density was calculated by dividing the population (persons) by the area (square kilometers) for each administrative unit of Taiwan and then sorting these areas into quartiles of low, moderate, high, and very high urbanization. These categories were used as surrogates for residential urbanization. Use of osteoporosis medications such as alendronate, pamidronate disodium, risedronate sodium, zoledronic acid, and hormone replacement therapy was also noted.

We used codes from the ICD-9-CM to define coexisting medical conditions and postfracture complications. Prefracture diabetes (ICD-9-CM 250) was defined as an exposure. Coexisting medical conditions were determined from medical claims for the follow-up period and included hypertension (ICD-9-CM 401–405), mental disorders (ICD-9-CM 290–319), chronic obstructive pulmonary disease (ICD-9-CM 490–496), ischemic heart disease (ICD-9-CM 410–414), hyperlipidemia (ICD-9-CM 272.0, 272.1, and 272.2), stroke (ICD-9-CM 430–438), liver cirrhosis (ICD-9-CM 571), and osteoporosis (ICD-9-CM 733.0). Renal dialysis was identified by administration code (D8, D9). Fracture is the main outcome of this retrospective cohort study, including skull bone fracture (ICD-9-CM 800–804) and fractures of neck and trunk (ICD-9-CM 805–809), upper limb (ICD-9-CM 810–819), and lower limb (ICD-9-CM 820–829). Complications after fracture were analyzed as secondary outcomes, including septicemia (ICD-9-CM 038 and 998.5), pneumonia (ICD-9-CM 480–486), urinary tract infection (ICD-9-CM 599.0), and deep wound infection (ICD-9-CM 958.3). Postfracture in-hospital mortality was also considered an outcome in the nested cohort study.

Our study used χ² tests to compare sociodemographic characteristics and coexisting medical conditions between people with and without diabetes. We calculated the hazard ratios (HRs) with 95% CIs for risk of fracture after diabetes, adjusting for age, sex, low income, urbanization, hypertension, mental disorders, chronic obstructive pulmonary disease, ischemic heart disease, hyperlipidemia, stroke, liver cirrhosis, osteoporosis, and renal dialysis in multivariate Cox proportional hazards regression models. The risk and incidence of fracture among categories of diabetes were determined, and sex- and age-stratified analyses for the adjusted HRs of fracture associated with diabetes were calculated.

In the nested cohort study, the sociodemographics and coexisting medical conditions between fracture patients with and without diabetes were compared by using χ² tests. The adjusted odds ratios (ORs) and 95% CIs of postfracture septicemia, deep wound infection, pneumonia, urinary tract infection, and mortality associated with prefracture diabetes were calculated in the multivariate logistic regressions with adjustment for age, sex, low income, urbanization, hypertension, mental disorders, chronic obstructive pulmonary disease, ischemic heart disease, hyperlipidemia, stroke, liver cirrhosis, osteoporosis, and renal dialysis. SAS, version 9.1 (SAS Institute, Inc., Cary, NC), statistical software was used for data analyses; two-sided P < 0.05 indicated significant differences between groups.

After the frequency matching by age and sex between cohorts with and without diabetes, there were no significant differences in age and sex (Table 1). Compared with the cohort without diabetes, patients with diabetes had higher proportions of living in the least urbanized areas (P < 0.0001) and of having low-income status (P < 0.0001), hypertension (P < 0.0001), mental disorders (P < 0.0001), osteoporosis (P < 0.0001), chronic obstructive pulmonary disease (P < 0.0001), ischemic heart disease (P < 0.0001), hyperlipidemia (P < 0.0001), stroke (P < 0.0001), liver cirrhosis (P < 0.0001), renal dialysis (P < 0.0001), and use of osteoporosis medication (P < 0.0001).

Table 2 shows that higher incidence of fracture was found in patients with diabetes than those without diabetes (24.2 vs. 17.1 per 1,000 person-years, P < 0.0001) during the follow-up period; the corresponding HR of fracture associated with diabetes was 1.66 (95% CI 1.60–1.72). The association between diabetes and fracture risk was more significant in females (HR 1.76 [95% CI 1.68–1.85]) than in males (HR 1.52 [95% CI 1.43–1.61]). Patients with diabetes were at greater risk of fracture in all age-groups.

Among 17,002 patients with fracture receiving related surgeries (Table 3), higher proportions of diabetic patients than patients without diabetes were noted among females, older people, those living in the least urbanized areas, and those with low income, hypertension, hyperlipidemia, liver cirrhosis, renal dialysis, mental disorders, ischemic heart disease, stroke, chronic obstructive pulmonary disease, and use of osteoporosis medication. In the nested retrospective cohort study (Table 4), patients with diabetes had higher risk of mortality (OR 1.27 [95% CI 1.02–1.60]), septicemia (OR 1.42 [95% CI 1.23–1.64]), deep wound infection (OR 1.34 [95% CI 1.06–1.71]), and urinary tract infection (OR 1.48 [95% CI 1.33–1.64]) after fracture with surgeries compared with those without diabetes. Postfracture mortality was associated with several characteristics of diabetes: diabetes-related hospitalization (OR 1.58 [95% CI 1.23–2.03]), ketoacidosis (OR 3.03 [95% CI 1.45–6.31]), renal manifestations (OR 1.87 [95% CI 1.32–2.64]), eye involvement (OR 1.79 [95% CI 1.27–2.52]), and peripheral circulatory disorder (OR 1.65 [95% CI 1.00–2.70]). Among patients with diabetes, inadequate control for diabetes (OR 1.61 [95% CI 1.11–2.34]) and type 1 diabetes (OR 1.93 [95% CI 1.14–3.26]) were independent risk factors for postfracture mortality.

This nationwide retrospective cohort study based on Taiwan National Health Insurance claims data analyzed long-term risk of all types of fracture in patients with diabetes during the follow-up period, with significant findings in women and elderly patients. Another nested cohort study in patients with diabetes showed increased rates of septicemia, deep wound infection, urinary tract infection, and postfracture mortality in patients receiving surgeries. Postfracture mortality was associated with ketoacidosis, renal manifestations, eye involvement, and peripheral circulatory disorder, which are all diabetes-related comorbidities. Large sample size, cohort study design, multivariate adjustment, including all types of fracture, and not being restricted by specific populations are strengths of this investigation.

Older age (11,19,25), female sex (19,25), low income (31), and urbanization (31) have been recognized as risk factors for fracture; these sociodemographic factors have also been associated with diabetes (29). These four factors were considered potential confounding factors in the association between diabetes and fracture. To investigate risk and outcomes of fracture in diabetic patients, we adjusted these sociodemographic characteristics in the multivariate regression models.

That osteoporosis is a risk factor for fracture has been noted in both general populations and diabetic patients (5–7,20). Hypertension (32), mental disorders (33), COPD (34,35), ischemic heart disease (34,35), hyperlipidemia (36), stroke (34), liver cirrhosis (35), and renal failure (35) have been shown to be independently associated with higher risk of fracture. These conditions commonly coexist with diabetes (29). However, previous studies were limited by inadequate control of coexisting medical conditions when investigating the association between diabetes and fracture risk (13,14,16–22,24). The current study found that fracture risk increased in patients with diabetes after adjustment for these potential confounding factors.

A cohort study based on a small sample reported that diabetic patients had higher perioperative adverse events after fracture than people without diabetes; these events were cardiac diseases, urinary tract infection, gastrointestinal symptoms, and prolonged length of hospital stay (27). However, this previous study was limited because it did not control confounding factors. Although previous studies have reported that diabetic patients had increased fracture risk (8–27), the current study is the first investigation suggesting that patients with diabetes have increased risks of postfracture septicemia, deep wound infection, and urinary tract infection after multivariate adjustment for sociodemographics and coexisting medical conditions. Adverse outcomes after noncardiac surgeries in diabetic patients were noted in our previous study (29).

In the present nested cohort study, type 1 diabetes, prefracture inpatient care for diabetes, inadequate control of glucose, diabetes-related ketoacidosis, coma, renal manifestations, eye involvement, and peripheral circulatory disorder were severity-related clinical predictors of postfracture mortality in diabetic patients. These interesting findings have not been reported previously, though previous investigations have studied the crude association between diabetes and fracture (8–27).

Possible explanations for the association between diabetes and fracture risk include altered bone mineral density and subsequent osteoporosis in diabetic patients. The strong correlation between poor bone mineral density and fracture suggested that losing bone mineral density may be one cause of diabetic patients suffering fracture (6,7). A second possible explanation is that the rapid rise in incidence of obesity and type 2 diabetes in all age-groups might result in substantial increases in prevalence of diabetes-related cognitive dysfunction (37) or mental illness that were identified as risk factors for fracture. Third, people with diabetes-related eye involvement such as cataract, glaucoma, and retinopathy have been found to be more likely to have fracture (38). Inadequate control for glucose resulting in hypoglycemic coma is another possible explanation for increased fracture risk in patients with diabetes (39).

This study has some limitations: First, we used insurance claims data that lacked information on sociodemographics, lifestyle, hormonal status, fasting glucose, glycated hemoglobin, and other biomedical measures. A second limitation is possible underreporting from use of records from reimbursement claims for coexisting medical conditions. Such cases could include patients with prediabetes (patients with impaired fasting glucose [100–125 mg/dL] or HbA_1c [5.7–6.4%]) (40), those who received medical services fewer than three times within the preoperative 24-month period, and those who did not receive hypoglycemic medication. Such underreporting might also be a factor due to patients with minor fractures, such as those who neglected to seek care for or did not know that they experienced fractures. However, these data should be distributed equally between both groups without causing bias in the results. Third, though the accuracy of major diagnosis codes from the research database in studies based on these has been accepted by peer reviewers for prominent scientific journals worldwide (28–30), validity of diabetes, fracture, and other comorbidities and complications codes might still be a limitation of this study. Future research is needed that includes prospective design and clinical measures, lifestyle information, and detailed severity of diabetes as covariates in multivariate analyses.

In conclusion, this detailed analysis showed that diabetes is an important independent risk factor for fracture and postfracture adverse events. Poorly controlled diabetes, type 1 diabetes, and diabetes-related complications contributed to mortality after fracture with repair surgeries. This study provides comprehensive assessment of fracture risk and postfracture outcomes in patients with diabetes. Further studies are needed to develop specific strategies to decrease fracture risks and postfracture adverse outcomes for this challenging patient population.

Acknowledgments. This study is based on data obtained from the National Health Insurance Research Database provided by the Bureau of National Health Insurance of Taiwan’s Ministry of Health and Welfare and managed by the National Health Research Institutes.

The authors’ interpretation and conclusions do not represent view points of these three agencies.

Funding. This study was supported by a grant (NSC102-2314-B-038-021-MY3) from Taiwan’s National Science Council.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. C.-C.L. created the concept and design, performed the statistical analysis, wrote the draft of the manuscript, interpreted the results, reviewed and edited the manuscript, and read and approved the final manuscript. C.-S.L., C.-C.S., C.-C.Y., Y.-C.C., Y.-W.L., and T.-L.C. participated in study design, interpreted the results, revised the manuscript, and read and approved the final manuscript. Y.-W.L. and T.-L.C. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.