Hip fracture and risk of acute myocardial infarction: A nationwide study

Authors

  • Chia-Hung Chiang,

    1. Division of Cardiology, Department of Medicine, Zhudong Veterans Hospital, HsinChu, Taiwan
    2. Division of Cardiology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan
    3. Cardiovascular Research Center, National Yang-Ming University, Taipei, Taiwan
    4. Institute of Clinical Medicine, National Yang-Ming University, Taipei, Taiwan
    Search for more papers by this author
    • The first two authors contributed equally to this work.

  • Chia-Jen Liu,

    1. Division of Hematology and Oncology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan
    2. Department of Internal Medicine, National Yang-Ming University Hospital, Yilan, Taiwan
    3. Institute of Public Health and School of Medicine, National Yang-Ming University, Taipei, Taiwan
    Search for more papers by this author
    • The first two authors contributed equally to this work.

  • Ping-Jen Chen,

    1. Department of Family Medicine and Department of Geriatric Medicine, Chi-Mei Medical Center, Tainan, Taiwan
    Search for more papers by this author
  • Chin-Chou Huang,

    1. Division of Cardiology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan
    2. Cardiovascular Research Center, National Yang-Ming University, Taipei, Taiwan
    3. Department of Medical Research and Education, Taipei Veterans General Hospital, Taipei, Taiwan
    4. Institute of Pharmacology, National Yang-Ming University, Taipei, Taiwan
    Search for more papers by this author
  • Chien-Yi Hsu,

    1. Division of Cardiology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan
    2. Cardiovascular Research Center, National Yang-Ming University, Taipei, Taiwan
    Search for more papers by this author
  • Zu-Yin Chen,

    1. Division of Cardiology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan
    2. Cardiovascular Research Center, National Yang-Ming University, Taipei, Taiwan
    3. Department of Internal Medicine, National Yang-Ming University Hospital, Yilan, Taiwan
    Search for more papers by this author
  • Wan-Leong Chan,

    1. Division of Cardiology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan
    2. Healthcare and Management Center, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Po-Hsun Huang,

    1. Division of Cardiology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan
    2. Cardiovascular Research Center, National Yang-Ming University, Taipei, Taiwan
    3. Institute of Clinical Medicine, National Yang-Ming University, Taipei, Taiwan
    Search for more papers by this author
  • Tzeng-Ji Chen,

    1. Department of Family Medicine, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Chia-Min Chung,

    1. Institute of Biomedical Sciences, Academia Sinica, Taipei, Taiwan
    Search for more papers by this author
  • Shing-Jong Lin,

    1. Division of Cardiology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan
    2. Cardiovascular Research Center, National Yang-Ming University, Taipei, Taiwan
    3. Institute of Clinical Medicine, National Yang-Ming University, Taipei, Taiwan
    4. Department of Medical Research and Education, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Jaw-Wen Chen,

    1. Division of Cardiology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan
    2. Cardiovascular Research Center, National Yang-Ming University, Taipei, Taiwan
    3. Department of Medical Research and Education, Taipei Veterans General Hospital, Taipei, Taiwan
    4. Institute of Pharmacology, National Yang-Ming University, Taipei, Taiwan
    Search for more papers by this author
  • Hsin-Bang Leu

    Corresponding author
    1. Division of Cardiology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan
    2. Cardiovascular Research Center, National Yang-Ming University, Taipei, Taiwan
    3. Institute of Clinical Medicine, National Yang-Ming University, Taipei, Taiwan
    4. Healthcare and Management Center, Taipei Veterans General Hospital, Taipei, Taiwan
    • Healthcare and Management Center, Division of Cardiology, Taipei Veterans General Hospital, No. 201, Sec. 2, Shi-Pai Road, Taipei, Taiwan, Republic of China.
    Search for more papers by this author

Abstract

Osteoporotic fractures are associated with increased mortality risk. However, little data are available on the risk of acute myocardial infarction (AMI) after hip fracture. Therefore, we investigated whether hip fracture increased the risk of AMI in a large, nationwide cohort study. We obtained data from 8758 patients diagnosed with hip fracture from 2000 to 2009 and from 4 matched controls for each patient from the Longitudinal Health Insurance Database (LHID 2000), Taiwan. Controls were matched for age, sex, comorbid disorders, and enrollment date. All subjects were followed up from the date of enrollment until AMI, death, or the end of data collection (2009). Cox's regression model adjusted for age, sex, comorbid disorders, and medication was used to assess independent factors determining the risk of development of AMI. As expected, despite the matching, the hip fracture patients had more risk factors for AMI at baseline. A total of 8758 subjects with hip fractures and 35,032 controls were identified. Among these patients, 1183 (257 hip fracture patients and 926 controls) developed AMI during the median 3.2-year (interquartile range 1.4 to 5.8 years) follow-up period. Patients with hip fractures had a higher incidence of AMI occurrence when compared with controls (8.7/1000 person-years versus 6.82/1000 person-years). Multivariate analysis adjusted for baseline covariates indicated that hip fracture was associated with a greater risk for AMI development (hazard ratio [HR] = 1.29; 95% confidence interval [CI] 1.12–1.48; p < 0.001). We conclude that hip fracture is independently associated with a higher risk of subsequent AMI. © 2013 American Society for Bone and Mineral Research

Introduction

Hip fracture is a devastating event with subsequent functional disability, morbidity, and mortality, all of which contribute toward tremendous health problems and economic costs.1 The general categories of hip fractures include intracapsular (femoral neck and head) and extracapsular (intertrochanteric and subtrochanteric) fractures. The incidence of hip fracture has increased worldwide along with the increase in the elderly population.2–4 Previous studies have shown that the lifetime risk of hip fracture is 17.5% for women and 6% for men.5 In the vast majority of cases, a hip fracture is a fragility fracture arising from a fall or minor trauma in individuals with weakened osteoporotic bones, particularly in elderly individuals who are more likely to fall because of poorer balance, medication side effects, and difficulty in maneuvering around environment hazards.

Bone and vasculature are regulated by several shared factors, and calcification of the vascular walls resembles the bone formation process in many ways.6 A previous clinical study has shown that patients with self-reported myocardial infarction had significantly higher odds of having low hip bone mineral density (BMD).7 In addition, the presence of coronary calcifications has been associated with a nonsignificant 8% lower lumbar spine BMD,8 suggesting that there is an association between subclinical atherosclerosis and lower bone density in the elderly. Recently, it has been reported that cardiovascular disease (CVD), such as congestive heart failure and stroke, is associated with an increase in the risk of subsequent hip fracture,9 providing further support for such a connection, even if the exact mechanism remains unclear.

Hip fracture substantially increases the risk of death and impaired functional status among the elderly.10 In a study of 1327 patients with incident hip fracture, the risk of prior CVD was twice that of 3170 patients without hip fracture.11 One recent study reported that hip fracture is associated with an increased risk of subsequent stroke.12 However, to the best of our knowledge, no studies have evaluated the relationship between hip fracture and the risk of a subsequent acute myocardial infarction (AMI) in patients without history of myocardial infarction. It is therefore uncertain whether hip fractures may increase the risk of AMI. It is also unknown whether the risk for AMI differs depending on sex and other comorbid disorders. Therefore, we conducted a large-scale, population-based, nationwide, case-cohort study using the Taiwan National Health Insurance database to investigate whether patients with hip fracture have a higher risk of developing AMI compared with controls matched on the basis of age, sex, and comorbid disorders.

Materials and Methods

Data sources

Taiwan began the National Health Insurance (NHI) program in 1995 to provide comprehensive health care for all citizens. Enrollment in the program is mandatory, and coverage is approximately 98%. The program provides comprehensive medical care including outpatient, inpatient, emergency, and dental care, and provision of prescription drugs. Currently, the National Health Insurance Research Database (NHIRD) at the National Health Research Institutes (NHRI) in Taiwan takes charge of the complete NHI claims database and has published several dozens of extracted data sets for research purposes. The Longitudinal Health Insurance Database (LHID 2000), comprising 1,000,000 randomly sampled individuals who were alive in 2000, contains all the records on these individuals from 1995 to 2009. It is one of the largest nationwide population-based data sets in the world and allows researchers to trace all uses of medical services by these 1,000,000 enrollees. The random samples have been confirmed to be representative of the Taiwanese population.

In addition, multiple NHI databases are maintained (eg, NHI enrollment files, claims data, and a registry for prescription drugs) to provide comprehensive utilization and enrollment information for all patients with catastrophic illnesses who are exempted from copayments under the NHI program. Because the NHIRD consists of de-identified secondary data released to the public for research purposes, this study was exempt from full review by the Institutional Review Board. The encrypting procedure is consistent so that linkage of claims belonging to the same patient is feasible within the NHIRD.

Study population

We conducted a retrospective cohort study from January 1, 2000, to December 31, 2009. Using the discharge codes (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes 820 to 820.9) in the LHID 2000, we identified hip fracture patients who had undergone examinations of X-ray and were newly diagnosed from January 1, 2000, to December 31, 2009.13 Their first hospitalization for hip fracture served as the index use of a medical service for hip fracture. Patients with an initial hip fracture diagnosis before 18 years of age or a diagnosis of AMI before or at hip fracture diagnosis were excluded from the study. Patients with a diagnosis of pathological fractures (ICD-9-CM codes 733.1 to 733.19) and fractures from transport-related accidents (ICD-9-CM codes E800 to E848) were also excluded.13 The final hip fracture cohort consisted of 8758 patients. Information on comorbid disorders and medications including antihypertensive agents, aspirin, clopidogrel, warfarin, proton pump inhibitors (PPIs), statins, oral estrogen, oral steroids, and oral bisphosphonates was collected for analysis.

Control cohort

Subjects without any record of hip fracture were used as the matched control cohort and were randomly selected from the remaining NHI beneficiaries registered in the LHID 2000. Each hip fracture patient was matched with 4 controls by age, sex, date of enrollment, and presence of comorbid disorders including hypertension, diabetes, previous history of coronary artery disease (CAD), and osteoporosis, within the same follow-up period. The same exclusion criteria were applied. A total of 35,032 subjects served as the matched control cohort.

Measurement of acute myocardial infarction

The endpoint of the study was the occurrence of AMI identified by an insurance claim (ICD-9-CM codes 410 to 410.92) as the primary diagnosis during ambulatory care visits, emergency care visits, or hospitalization during the follow-up period. All study subjects were followed up until AMI, death, or until December 31, 2009, whichever was earlier. The diagnosis of AMI was based on the World Health Organization guideline in which at least two of the three following criteria should be met: prolonged (≥30 minutes) anterior chest discomfort; abnormal electrocardiography readings (ST segment elevation or depression, evolving Q waves, symmetric inversion of T waves); and elevated levels of cardiac enzymes (CK-MB, troponin T, or troponin I).14

Statistical analyses

Extraction and computation of data were performed using the Perl programming language (version 5.12.2). Microsoft SQL Server 2005 (Microsoft Corp., Redmond, WA, USA) was used for data management and computing. Statistical analysis was performed utilizing SPSS software (version 17.0, SPSS Inc., Chicago, IL, USA). All data were expressed as the frequency (percentage) for ordinal and categorical data or as mean ± standard deviation (SD) for continuous data. Continuous data for the study group and comparison group were compared by Student's t test. Categorical data were compared with a chi-square test and Yates' correction or Fisher's exact test, as appropriate. To assess the risk of developing AMI during the period of up to 10 years of follow-up, the Kaplan-Meier method was employed. Multivariate regression analysis was performed using Cox proportional hazards regression analysis with adjustment for confounding factors including age, sex, and comorbid disorders. Comorbidity variables were treated as time-varying covariates in the Cox model. A p value < 0.05 was considered statistically significant.

Results

We assessed data for 8758 patients diagnosed with hip fracture and 35,032 controls without hip fracture. The mean age of the study subjects was 70 years (SD 17.4), and the median follow-up duration was 3.2 years (interquartile range 1.4 to 5.8 years), with a maximal follow-up period of 10 years. In total, 1183 patients (2.7%) suffered from newly developed AMI during the follow-up period: 257 of the hip fracture patients (2.9%) and 926 of the controls (2.6%). Table 1 shows the baseline characteristics and medications of the hip fracture patients and controls. Compared with controls, patients with hip fracture had a shorter follow-up duration (2.8 versus 3.4 years, p < 0.001). Hip fracture patients also had more comorbid disorders, including chronic kidney disease, dyslipidemia, and chronic obstructive pulmonary disease (COPD), and they also took more medications, including aspirin, clopidogrel, warfarin, proton pump inhibitors, statins, oral estrogen, oral steroids, and oral bisphosphonates. Table 2 and Supplemental Table S1{TBL S1} show the baseline characteristics of patients with and without development of AMI during follow-up. The patients who developed AMI were older and had more comorbid disorders including hypertension, diabetes, CAD, chronic kidney disease, dyslipidemia, and COPD. Those who developed AMI also took more aspirin, clopidogrel, warfarin, and statins.

Table 1. Baseline Characteristics and Medication Use of the Study Population
 Hip fracturep Value
 No (n = 35,032)Yes (n = 8758)
  1. NS = nonsignificant; CAD = coronary artery disease; COPD = chronic obstructive pulmonary disease; PPIs = proton pump inhibitors; ACEI = angiotensin converting enzyme inhibitor; ARB = angiotensin II receptor blocker.

  2. Data are mean ± SD or median (interquartile range); t test and chi-square test were used for evaluating continuous and categorical variables, respectively.

Age (years)70.0 ± 17.470.0 ± 17.4NS
Follow-up duration (years)3.4 (1.5–6.0)2.8 (1.0–5.2)<0.001
Male, n (%)15,720 (44.9)3930 (44.9)NS
Hypertension, n (%)22,449 (64.1)5612 (64.1)NS
Diabetes, n (%)12,473 (35.6)3118 (35.6)NS
CAD, n (%)13,273 (37.9)3318 (37.9)NS
Osteoporosis, n (%)10,945 (31.2)2735 (31.2)NS
Chronic kidney disease, n (%)5650 (16.1)1759 (20.1)<0.001
Dyslipidemia, n (%)10,995 (31.4)2374 (27.1)<0.001
COPD, n (%)12,282 (35.1)3350 (38.3)<0.001
Medications, n (%)
 Aspirin8977 (25.6)2390 (27.3)0.002
 Clopidogrel1154 (3.3)352 (4.0)0.001
 Warfarin530 (1.5)173 (1.6)0.002
 PPIs4204 (12.0)1409 (16.1)<0.001
 Statins4898 (14.0)911 (10.4)<0.001
 Oral estrogen950 (2.7)381 (4.4)<0.001
 Oral steroids5902 (16.8)1836 (21.0)<0.001
 Oral bisphosphonates1157 (3.3)720 (8.2)<0.001
 Calcium channel blocker10,221 (29.2)2432 (27.8)0.010
 ACEI2255 (6.4)581 (6.7)NS
 ARB5445 (15.5)1127 (12.9)<0.001
 Alpha blocker3040 (8.7)756 (8.6)NS
 Beta blocker5374 (15.3)1211 (13.8)<0.001
 Diuretics3834 (10.9)834 (9.5)<0.001
Table 2. Demographic Data of the Study Population With or Without Acute Myocardial Infarction
 Acute myocardial infarctionp Value
 No (n = 42,607)Yes (n = 1183)
  1. NS = nonsignificant; CAD = coronary artery disease; COPD = chronic obstructive pulmonary disease; PPIs = proton pump inhibitors; ACEI = angiotensin converting enzyme inhibitor; ARB = angiotensin II receptor blocker.

  2. Data are mean ± SD; t test and chi-square test were used for evaluating continuous and categorical variables, respectively.

Age (years)69.8 ± 17.576.6 ± 10.8<0.001
Male, n (%)19,148 (44.9)502 (42.4)NS
Hypertension, n (%)27,122 (63.9)939 (79.4)<0.001
Diabetes, n (%)15,033 (35.3)558 (47.2)<0.001
CAD, n (%)16,022 (37.6)568 (48.0)<0.001
Osteoporosis, n (%)13,287 (31.2)393 (33.2)NS
Chronic kidney disease, n (%)7168 (16.8)241 (20.4)0.001
Dyslipidemia, n (%)12,958 (30.4)411 (34.7)0.002
COPD, n (%)15,172 (35.6)460 (35.7)0.021
Medications, n (%)
 Aspirin10,872 (25.5)495 (41.8)<0.001
 Clopidogrel1401 (3.3)105 (8.9)<0.001
 Warfarin674 (1.6)29 (2.5)0.021
 PPIs5481 (12.9)132 (11.2)NS
 Statins5623 (13.2)186 (15.7)0.012
 Oral estrogen1286(3.0)45 (3.8)NS
 Oral steroids7527 (17.7)211 (17.8)NS
 Oral bisphosphonates1829 (4.3)48 (4.1)NS
 Calcium channel blocker12,158 (28.5)495 (41.8)<0.001
 ACEI2714 (6.4)122 (10.3)<0.001
 ARB6376 (15.0)196 (16.6)NS
 Alpha blocker3672 (8.6)124 (10.5)0.026
 Beta blocker6339 (14.9)246 (20.8)<0.001
 Diuretics4518 (10.6)150 (12.7)0.023

Fig. 1 exhibits the results of the log-rank test and Kaplan-Meier survival analysis. During the maximal 10-year follow-up period, the cumulative incidence of AMI was significantly higher in patients with hip fracture than controls (p = 0.001 by log-rank test). Furthermore, compatible with other studies, hip fracture patients in the current study had higher mortality than controls (crude hazard ratio [HR] = 1.84; 95% confidence interval [CI] 1.74–1.95; p < 0.001) during maximal 10 years of follow-up period.

Figure 1.

Kaplan-Meier estimates of the cumulative incidence of acute myocardial infarction in subjects categorized by hip fracture. The cumulative incidence of acute myocardial infarction was significantly higher in patients with hip fracture (p < 0.001 by log-rank test).

Table 3 shows the association between hip fracture and risk of subsequent AMI. The crude HR for AMI in hip fracture patients was 1.27 (95% CI 1.10–1.45; p = 0.001). After adjustment for other confounders including age, sex, comorbid disorders, and medication, hip fracture remained associated with a 29% increase in AMI risk (HR = 1.29; 95% CI 1.12–1.48; p < 0.001). Fig. 2 shows a stratified analysis of hip fracture and AMI risk among variable subgroups. Patients with hip fracture had a higher risk of AMI development, especially women, older patients, and patients with comorbid disorders including hypertension, diabetes, CAD, osteoporosis, and COPD.

Table 3. Association Between Hip Fracture and Acute Myocardial Infarction
    Hazard ratio (95% CI)
 AMI, n (%)Incidence (per 1000 person-years)Unadjusted HRAdjusted HR
Patient groupsYesNo  Model 1aModel 2bModel 3c
  • AMI = acute myocardial infarction.

  • a

    Model 1: Adjusted for age and sex.

  • b

    Model 2: Adjusted for age, sex, and comorbid disorders (hypertension, diabetes, coronary artery disease, osteoporosis, chronic kidney disease, dyslipidemia, chronic obstructive pulmonary disease).

  • c

    Model 3: Adjusted for age, sex, comorbid disorders and medications (aspirin, clopidogrel, warfarin, proton pump inhibitors, statins, oral estrogen, oral steroids, oral bisphosphonates).

Controls (n = 35,032)926 (2.6)34,106 (97.4)6.821.001.001.001.00
Hip fracture (n = 8758)257(2.9)8501 (97.1)8.701.27 (1.10–1.45)1.30 (1.14–1.50)1.28 (1.1–1.47)1.29 (1.12–1.48)
Figure 2.

The association between hip fracture and acute myocardial infarction in specific subgroups identified by Cox regression analysis.

Discussion

Cardiovascular disease and complications are the main causes of morbidity and mortality in older patients with hip fracture and those undergoing hip fracture surgery.15 However, the role of comorbid disorders and complications in mortality remains uncertain. Few studies have investigated the relationship between hip fracture and future cardiovascular events. In addition, all cardiovascular diseases are usually grouped together.16 AMI is one of the leading causes of mortality worldwide.17, 18 The results of the current study indicate that, after adjustment for potential confounding factors, hip fracture is independently associated with a 29% increase in risk of developing AMI compared with controls. In addition, the risk of AMI after hip fracture is more prominent in elderly patients, women, and in patients with other underlying diseases. To the best of our knowledge, the current study is the first large-scale, nationwide cohort study to explore the relationship between hip fracture and risk of new-onset AMI. We advocate that individuals with hip fracture should modify their cardiovascular risk factors to avoid future cardiovascular events. Furthermore, it should be emphasized that our estimates were independent of known clinical risk factors of myocardial infarction including hypertension, diabetes, and CAD, and hip fracture would probably add to the risk of such factors.

The pathological links between hip fracture and AMI remain unclear. Several mechanisms have been proposed and could be partially related to physical inactivity and reduced exercise capacity after hip fracture.19 Deterioration of preexisting cardiovascular risk could be triggered in such circumstances.20 In addition, common risk factors for CVD and osteoporotic fractures such as hypertension, smoking, diabetes, and dyslipidemia have each been suspected as a pathogenic factor linking these two disease entities.16 However, none of these could fully explain the relationship between hip fracture and future development of AMI.

The current study provides strong evidence linking hip fracture and AMI risk. Although the pathogenesis underlying this relationship could not be elucidated, some mechanisms and genetic factors may explain this association. First, oxidative stress may be a potential determinant between bone loss, fractures, and cardiovascular events. Walter and colleagues demonstrated that serum oxidative stress markers such as thiobarbituric acid-reactive substances are predictive of cardiovascular events including fatal and nonfatal myocardial infarction.21 Oxidative stress is also associated with a faster bone turnover, higher bone resorption, and osteoporosis because oxidative stress inhibits osteoblastogenesis.22 Recently, a study demonstrated that oxidative stress was increased in older patients with hip fracture compared with healthy elderly controls.23 Taken together, these findings suggest that increased oxidative stress might be one possible mechanism linking hip fracture and subsequent AMI.

Another link between bone and cardiovascular diseases may be chronic systemic inflammation. Chronic inflammation processes play an important role in the development and progression of atherosclerotic plaque. Several inflammatory mediators, including C-reactive protein, interleukin-6, and tumor necrosis factor-α, have been proven to be correlated with the severity or prognosis of CAD.24, 25 Increased circulating inflammatory cytokines also stimulate bone resorption and result in osteoporotic change, leading to hip fracture.26, 27 In addition, inflammatory cells may secrete matrix metalloproteinases (MMPs), which degrade bone collagen and participate in bone resorption. Previous studies have shown that some specific MMPs are activated and contribute to adult fracture healing, especially during endochondral development.28 Members of the MMP family have also been identified in human atherosclerotic plaque shoulders and in regions of foam cell accumulation, contributing to lesion progression, plaque vulnerability, and de novo atherosclerotic remodeling.29 In addition, expression of serum MMPs has been found in patients with acute coronary syndrome and osteoporosis.30, 31 These data suggest that chronic inflammation, inflammatory markers, and MMPs might play pivotal roles in the relationship between osteoporotic hip fracture and subsequent AMI.

In the current study, the risk of AMI after hip fracture was more prominent among elderly patients, women, and those with comorbid disorders. The higher risk observed for the simultaneous presence of hip fracture and comorbid medical disorders might reflect a synergistic effect between these risk factors. Gender differences in the relationship between hip fractures and AMI risk remain controversial. Greater bone loss has been linked to greater progression of aortic calcification in women but not in men.32 Furthermore, aortic calcification is associated with increased risk of myocardial infarction.33, 34 Taken together, these findings suggest that osteoporotic fractures may have a higher impact on subsequent AMI among women than among men. In contrast, a study by Varosy and colleagues reported that the risk of coronary events was lower in those who had postmenopausal fractures.35 However, in that study, no association between hip fracture and AMI risk was established, and patients with hip fracture did not have a significantly lower risk of coronary events. The study subjects were all postmenopausal women with CAD, and their results may not be applicable to the general population. Our results from the subgroup analyses were partially in line with a previous study showing that the presence of more comorbid disorders on admission was associated with a greater risk of postoperative complications and mortality among elderly hip fracture patients.15 In the same study, age was also an independent predictor of mortality in hip fracture patients.

The strength of our study is the use of a population-based data set with enrollment of a large number of subjects, which enabled us to trace prospectively the differences between hip fracture patients and controls. The current cohort study confirms that hip fractures are associated with an increased risk of subsequent AMI. As with other studies, there are some limitations to consider. First, because a claims database was used, our study was unable to investigate the effects of risk factors such as body mass index, dietary habits, exercise capacity, and cigarette smoking in the regression model, potentially compromising our findings. Because nutritional supplements were not covered by the national insurance project in Taiwan, information on vitamin D and calcium supplements is also lacking. We could not adjust our estimates for severity of CAD and myocardial infarction either. In one epidemiological study in Taiwan, smoking habit is much lower among elderly females than males.36 In the current study, more study subjects are females with a mean age greater than 70 years, suggesting the prevalence of smokers should not be high. Another epidemiological study investigating risk factors for hip fracture in Taiwan showed patients with hip fracture tended to be female, elderly, and diabetic, instead of smokers.37 After adjusting for these cofounders, hip fracture is still independently associated with increased risk of MI in our study. Second, the diagnosis of hip fracture and AMI was based on the administrative claims data reported by physicians or hospitals. A widespread skepticism prevails regarding the validity of diagnostic coding in medical claims, but previous epidemiological studies conducted with the LHID have revealed results consistent with population-based surveys for several conditions.38 Furthermore, to maximize diagnostic coding accuracy in medical claims, the NHI has extensive and systemic quality-assurance processes including routine crosschecking of chart reviews by clinical specialists. Because cardiovascular events such as myocardial infarction and stroke are critical illnesses causing severe disability, almost all patients need medical treatments. Furthermore, the coverage of national insurance is more than 99% in Taiwan. A previous validation study comparing ICD coding and hospital record review in Taiwan has showed that sensitivity and specificity for major cardiovascular events such as stroke were 100% and 95%, respectively, suggesting the high positive predictive value for a major cardiovascular event in which hospitalization is needed.39 Another study also demonstrated a high positive predictive value up to 95% for acute myocardial infarction diagnosis after comparing ICD coding from Medicare claims and a diagnosis based on structured hospital record review,40 further supporting the accuracy of AMI diagnosis in our study. Third, although hip fracture patients had higher mortality in the current study, we did not know the exact cause of death for each enrollment, especially those deaths that occurred outside of hospitals. Therefore, it is difficult to investigate and explain the causal relationship between hip fractures and all-cause deaths. Finally, we did not investigate the impact of type of hip fracture, anesthesia, or surgery in the current study. Further studies may be needed to assess the risk of myocardial infarction among hip fracture subtypes.

We conclude that hip fracture is independently associated with a higher risk of subsequent AMI. Our findings suggest that hip fracture patients may benefit from screening for CVD and from modifying preexisting cardiovascular risk factors to avoid future cardiovascular events. Clinicians should be aware of the considerably increased risk of cardiovascular events after hospitalization for hip fracture.

Disclosures

All authors state that they have no conflicts of interest.

Acknowledgements

This study was partly supported by research grants V99B1-011, V99C1-125, and V100B-013 from Taipei Veterans General Hospital, Taipei, Taiwan; CI-97-13 and CI-98-16 from the Yen Tjing Ling Medical Foundation, Taipei, Taiwan; NSC 100-2314-B-075-055 from the National Science Council, and NSC-99-2911-I-009-101 from the UST-UCSD International Center of Excellence in Advanced Bio-engineering sponsored by the Taiwan National Science Council I-RiCE Program. The corresponding authors have full access to all the data used in the study, and take responsibility for the decision to submit the manuscript for publication.

Authors' roles: All authors participated in the conception, design, interpretation of data, drifting, and revision of the manuscript.

Ancillary