JAMDA 15 (2014) 725e731

JAMDA journal homepage: www.jamda.com

Original Study

Risk Analysis for Second Hip Fracture in Patients After Hip Fracture Surgery: A Nationwide Population-Based Study Shih-Hsun Shen MD a, Kuo-Chin Huang MD a, b, c, *, Yao-Hung Tsai MD a, b, Tien-Yu Yang MD a, Mel S. Lee MD, PhD a, b, Steve W.N. Ueng MD b, d, Robert W.W. Hsu MD a, b a

Department of Orthopaedics, Chiayi Chang Gung Memorial Hospital, Chiayi, Taiwan College of Medicine, Chang Gung University, Taoyuan, Taiwan Graduate Institute of Clinical Medical Science, Chang Gung University, Taoyuan, Taiwan d Department of Orthopaedics, Chang Gung Memorial Hospital, Linkou, Taiwan b c

a b s t r a c t Keywords: Second hip fracture fracture risk mortality osteoporosis geriatrics population-based study

Objectives: The current treatment program for fragility hip fractures (HFx) emphasizes a combination of early surgery, rehabilitation, and tertiary prevention strategy for osteoporosis; however, the effect is unclear and little information is available on the risk factors predicting the occurrence of a second hip fracture (SHFx). The aim of this study was to explore the incidence, risk factors, and subsequent mortality of SHFx in patients after their first hip fracture surgery (HFxS). Design, Setting, and Participants: We performed a nationwide population-based longitudinal observational study using the National Health Insurance Research Database (NHIRD) of Taiwan with a logistic regression model analysis. Of 87,415 patients undergoing HFxS during the period 2004 to 2007, we identified 8027 patients who had sustained an SHFx for analyses. Measurements: Data collected included patient characteristics (demographics, comorbidities, and concurrent medication use), incidence and hazard ratios of SHFx after HFxS, and subsequent age-specific mortality. Results: The overall incidence of SHFx was 9.18% and the age-specific mortality was increased 1.6- to 2.2-fold in patients with SHFx compared with those without after HFxS in this 7-year longitudinal study. The identified risk factors included age (AOR ¼ 1.84, 95% CI: 1.24e2.89), female gender (AOR ¼ 1.12, 95% CI: 1.03e2.30), obesity (AOR ¼ 2.89, 95% CI: 1.81e3.01), diabetes (AOR ¼ 3.85, 95% CI: 2.54e4.05), arterial hypertension (AOR ¼ 2.45, 95% CI: 1.83e2.62), hyperlipidemia (AOR ¼ 2.77, 95% CI: 1.27e3.19), stroke/TIA (AOR ¼ 2.85, 95% CI: 2.20e3.23), blindness/low vision (AOR ¼ 3.09, 95% CI: 2.54e3.73), and prolonged use of analgesics and anti-inflammatory medications (all AOR  3.05, all P values .012). Bisphosphonate therapy after HFxS had a significant negative risk association with the development of an SHFx (20.8% vs 32.3%, P ¼ .023; AOR ¼ 2.24, 95% CI: 1.38e2.90). Conclusion: We concluded that the occurrence of an SHFx and subsequent mortality in patients after HFxS is rather high. An understanding of the risk factors predicting the occurrence of an SHFx provides a valuable basis to improve health care for geriatric populations. Ó 2014 AMDA e The Society for Post-Acute and Long-Term Care Medicine.

Fragility hip fracture (HFx), one of the most severe consequences of osteoporosis, is a common and increasing cause of hospitalization, morbidity, lifelong disability, and even premature death.1e6 In a systemic epidemiological review Abrahamsen et al7 reported that

Shih-Hsun Shen, MD, and Kuo-Chin Huang, MD, contributed equally to this work. The authors declare no conflicts of interest. * Address correspondence to Kuo-Chin Huang, MD, No. 6, West Sec., Chia-Pu Rd., Pu-Tz City, Chiayi County 613, Taiwan. E-mail address: [email protected] (K.-C. Huang).

patients are at increased risk for premature death for many years after a fragility HFx, with an excess mortality rate during the first year after fracture ranging from 8.4% to 36.0%. Other studies have indicated that the cumulative mortality rates associated with HFx may be higher than that for other better known life-threatening conditions, such as pancreatic or gastric cancer and myocardial infarction.8e10 Even though patients survived after HFx, they exhibited decreased quality of life and increased dependence on family, caregivers, and social services. Also, the burden on the society is considerable because of the direct costs of hospital care, rehabilitation, and nursing home facilities.5 Almost 14% of fragility fractures and 72% of the total

1525-8610/$ - see front matter Ó 2014 AMDA e The Society for Post-Acute and Long-Term Care Medicine. http://dx.doi.org/10.1016/j.jamda.2014.05.010

726

S.-H. Shen et al. / JAMDA 15 (2014) 725e731

cost burden was borne by HFx.11 As the elderly population increases rapidly worldwide, HFx have already become an important public health and socioeconomic issue. The current treatment program for HFx, therefore, emphasizes a combination of early surgery, rehabilitation, and prophylactic strategies, with the goal of improving the treatment outcome and decreasing the cost burden on the family and society.12 During the past decades, many studies have been conducted to find the best prophylactic strategies focusing on primary, secondary, or tertiary prevention of osteoporosis to ensure better management of these disorders and to maximize the chances of substantially alleviating their burden.4,7e14 Primary prevention is aimed at a fracture risk reduction in the general population. Secondary prevention involves the screening of osteoporosis to identify those who may have a high risk of fragility fractures. Tertiary prevention is a strategy to prevent future fractures in patients who have already sustained a fracture.13 Given the health care resources constraint, preventing a second hip fracture (SHFx) in patients after hip fracture surgery (HFxS) is a major concern because it is the most devastating fracture for patients, an extremely hard rehabilitation process to go through, and a considerable cost burden on the family and society for that condition.11,12,14e16 The most common ways to prevent these catastrophic events include lifestyle modification, pharmacologic therapy, and fall-prevention strategies.16 However, the effect is unclear and little information is available on the demographic data, risk factors, and outcomes of persons who sustained an SHFx when compared with those without. Based on a nationwide population-based longitudinal observational study design, we aimed to determine the incidence rate, demographic characteristics, and risk associations of patients who developed an SHFx following HFxS. We also explored the age-specific mortality in these patients. The information from this study may be valuable in improving the medical care of elderly patients after HFxS, thereby preventing the occurrence of an SHFx and subsequent mortality. Materials and Methods Database The Taiwan National Health Insurance (NHI) program, officially commenced since March 1995, enrolls nearly all citizens of Taiwan. As from January 2004, 22.3 million Taiwanese residents, with a coverage rate exceeding 98% of the whole population, have been enrolled in this program. The National Health Insurance Research Database (NHIRD), derived from the payment system of the National Health Insurance Administration (NHIA) and maintained by the National Health Research Institute (NHRI) in Taiwan, provides vital information for research purposes. All the identifiers of individual patients are deleted by the NHIA before data are transferred to the NHIRD. Institutional review board approval is pre-approved by the NHRI for de-identified data. Details on the database generation, and monitoring and maintenance of the NHIRD are published online by the NHRI. The database includes patient demographics, disease diagnosis, medical care institutions, medical expenditure, and prescription claims data. Until now, many studies have been ongoing or published using the NHIRD. Study Design and Participants This study was a nationwide population-based longitudinal observational study. The study cases and controls were selected from the NHIRD covering the period from January 2004 to December 2007 and followed longitudinally until December 31, 2010. Subjects were

identified from the database by using the following criteria: (1) a discharge diagnosis code of HFx (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] diagnosis codes 820.0e820.9), (2) a procedure code of internal fixation or partial hip replacement (ICD-9-CM procedure codes 79.15, 79.35, and 81.52), and (3) patients aged 45 years or older. The first admission date of HFx was defined as the index date. The exclusion criteria were patients with multiple trauma (Diagnosis Related Group [DRG] codes 484e487); primary or secondary diagnoses of endocrine, nutritional, metabolic and immune diseases (ICD-9-CM codes 240.0e279.9); and those in whom surgery was not performed or who died within the first 48 hours (ie, the difference between the date of death and the date of admission ¼ 0e1 days). Patients who were nonresidents of Taiwan and those with a diagnosis of HFx within the previous 2 years of the index date were excluded to avoid confounding effects. At this stage, 90,314 of a total of 141,314 patients with an HFx were included and followed longitudinally until the end of the study. During the follow-up period, patients with an SHFx were selected as the cases (n ¼ 8,764) and those without as the controls (n ¼ 81,550). We excluded patients with concomitant multiple trauma (DRG codes 484e487) and those with primary and secondary diagnoses of cancer (ICD-9-CM codes 140.0e209.9 or V10) in both groups. Finally, 8027 patients with a SHFx were included in the case group and 79,388 without in the control group. The flowchart of the patient-selection process is presented in Figure 1. Primary and Secondary End Points Patients were classified as cases and controls based on the occurrence of an SHFx after HFxS. The primary end point of this study was to determine the incidence of an SHFx after HFxS and the risk association among age, gender, selected morbidities, concurrent medication use, and the occurrence of an SHFx. Age was categorized into 8 groups (45e49, 50e54, 55e59, 60e64, 65e69, 70e74, 75e79, and 80 years). The selected comorbidities included metabolic syndrome ([MetS]; obesity, diabetes mellitus, arterial hypertension, and hyperlipidemia), coronary heart disease, myocardial infarction, cardiac dysrhythmia, peripheral arterial occlusive disease (PAOD), kidney dysfunction, stroke/transient ischemic attack (TIA), dementia, Parkinson disease, blindness/low vision, chronic obstructive pulmonary disease (COPD), osteoporosis, and arthritis. The concurrent medication use included calcium/vitamin D, bisphosphonates, hormone replacement therapy (HRT), selective estrogen receptor modulators (SERMs), calcitonin, steroids, acetaminophen, nonsteroidal anti-inflammatory drugs (NSAIDs), cyclooxygenase-2 (COX-2) inhibitors, antidepressants, stimulants, antipsychotics, mood stabilizers, anxiolytics, and depressants. For our secondary end point, information on age-specific mortality in both groups was collected for further analysis. Statistical Analysis A c2 analysis was used for analyzing categorical data. Univariate and multivariate analyses by using a logistic regression model were performed to detect the predicting factors having a significant relationship with the occurrence of an SHFx after HFxS. Patients without an SHFx were designed as the reference group and the crude and adjusted hazard ratios (HR) were obtained over the 1-year, 3-year, 5-year, and 7-year follow-up periods. Adjusted factors included age, gender, selected morbidities, and medication use history. The survival rate free of occurrence of an SHFx after HFxS was estimated by using the Kaplan-Meier survival method and the log rank test. Statistical significance was defined as P < .05. All statistics were 2-sided and performed using the SAS software (version 9.2; SAS Inc., Cary, NC).

S.-H. Shen et al. / JAMDA 15 (2014) 725e731

727

Total Admission Patients from 2004 to 2007 (n = 1,187,100) Excluded: Died in the enrolled admission

Total eligible participants (n = 141,314) Excluded: admitted for endocrine reasons (n = 696,790) admitted for head trauma (n = 69,597) admitted for chest trauma (n = 103,987) admitted for abdomen trauma (n = 106,320) admitted for upper limb trauma (n = 108,664) admitted for other lower limb trauma (n = 95,739)

With new-onset hip fracture from 2004 to 2010 (n = 8,764) Excluded: had head trauma (n = 670) had chest trauma (n = 1,230) had abdomen trauma (n = 946) had upper limb trauma (n = 871) had lower limb trauma (n = 657)

Without new-onset hip fracture from 2004 to 2010 (n = 81,550) Excluded: had head trauma (n = 6,580) had chest trauma (n = 12,990) had abdomen trauma (n = 12,466) had upper limb trauma (n = 10,079) had lower limb trauma (n = 9,739)

Excluded:

Excluded:

had malignancy (n = 2,162)

had malignancy (n = 737)

Included cases (n = 8,027)

Included controls (n = 79,388)

Fig. 1. Flowchart of the patient-selection process.

Results In the period 2004 to 2007, 87,415 patients aged 45 years or older who sustained an HFx and underwent HFxS were enrolled in this study using the aforementioned inclusion and exclusion criteria (Figure 1). As shown in Table 1, the overall incidence of SHFx in patients after HFxS during the follow-up period was 9.18%. The number of SHFx had increased approximately 2.7-fold in these years, from 1154 in 2004 to 3121 in 2007. The 1-year cumulative SHFx risk was 4.1% and 2.7% for patients 75 years and older and those younger than 75 years, respectively (Figure 2A). The high overall incidence and increasing new-patient numbers highlight the importance of understanding and prevention of SHFx in patients after HFxS. Baseline data on the patients who experienced their first HFxS in 2004 to 2007 are shown in Table 2. Age distribution younger than 55 years, coronary heart disease, myocardial infarction, cardiac dysrhythmia, PAOD, kidney dysfunction, dementia, Parkinson’s Table 1 Second Hip Fracture Numbers and Incidence Rate in Patients After Hip Fracture Surgery From 2004 to 2007 in Taiwan Year

Hip Fracture Surgery

Second Hip Fracture

Incidence Rate, %

2004 2005 2006 2007 Total Average

14,750 17,327 25,545 29,793 87,415 21,854

1154 1509 2243 3121 8027 2007

7.82 8.71 8.78 10.48 9.18 8.95

disease, COPD, osteoporosis, arthritis, concurrent use of central nervous system (CNS) medications, or anti-osteoporosis agents except for bisphosphonates did not differ among the 79,388 patients with only one HFx and the 8027 patients who sustained an SHFx during the follow-up period until the end of 2010. Table 3 shows the results of univariate and multivariate analyses of the potential risk factors for an SHFx in patients after HFxS. Using a logistic regression model, age (adjusted odds ratio [AOR] 1.84, 95% confidence interval [CI] 1.24e2.89), female gender (AOR 1.12, 95% CI 1.03e2.30), obesity (AOR 2.89, 95% CI 1.81e3.01), diabetes mellitus (AOR 3.85, 95% CI 2.54e4.05), arterial hypertension (AOR 2.45, 95% CI 1.83e2.62), hyperlipidemia (AOR 2.77, 95% CI 1.27e3.19), stroke/TIA (AOR 2.85, 95% CI 2.20e3.23), blindness/low vision (AOR 3.09, 95% CI 2.54e3.73), and use of steroids (AOR 4.17, 95% CI 3.90e5.32), acetaminophen (AOR 3.45, 95% CI 2.43e4.68), NSAIDs (AOR 4.12, 95% CI 3.80e4.55), and COX-2 inhibitors (AOR 3.05, 95% CI 2.28e3.50) were all identified as significant predictors for a SHFx. Bisphosphonate therapy in patients after HFxS had a significant negative risk association with the development of an SHFx (20.8% vs 32.3%, P ¼ .023; AOR 2.24, 95% CI 1.38e2.90) (Tables 2 and 3). Some illegibility was seen in SHFx risks in different age categories (Table 2). Comparison of cases and controls revealed no significant differences in patients younger than 55 years, but unclear risk associations in patients 55 and older. To clarify the effects of age categories on the occurrence of an SHFx in patients after HFxS, the 1-year, 3-year, 5-year, and 7-year SHFx risks and hazard ratios (HRs) for SHFx for 3 age categories were studied and are presented in Table 4. The adjusted HRs for SHFx were significantly higher for the age categories

728

S.-H. Shen et al. / JAMDA 15 (2014) 725e731 Table 2 Baseline Characteristics of 87,415 Patients (Age 45 Years) Who Experienced Their First Hip Fracture in 2004e2007 Variables

Fig. 2. (A) Kaplan-Meier survival estimates with SHFx as an end point. (B) Adjusted hazard ratios of SHFx of patients aged 55 to 74 years and patients aged 75 years and older compared with the risk of SHFx in patients younger than 55 years in selected time intervals after the first HFxS.

55 to 74 years and 75 years and older compared with the reference categories. Figures 2A and 2B revealed that there was a timedependent trend toward an increased SHFx risk in elderly patients, with an adjusted HR of 8.89 and a P value of .001 at 7-year follow-up after HFxS for those patients 75 years and older. The 7-year mortality rates of total included patients were 0.66%, 0.99%, and 4.49% for the age categories younger than 55 years, 55 to 74 years, and 75 years and older, respectively (Table 5). Between-group comparison revealed that there was a significant age-dependent trend toward an increased mortality rate in elderly patients after HFxS (Figure 3A). Elderly patients who sustained an SHFx during the followup period had a significantly higher mortality rate compared with those who did not (age category 55e74 years: 1.57% vs 0.97%, P < .05; 75 years: 6.80% vs 4.08%, P < .001) (Table 5 and Figure 3B). The 7-year mortality rate had increased up to more than 1.6-fold in patients who had an SHFx compared with those who did not for all patients 55 years and older. Discussion The present population-based longitudinal observational study showed a high rate of SHFx in patients after HFxS, with an overall incidence of 9.18%. The number of SHFx had increased approximately

Age 45e49 50e54 55e59 60e64 65e69 70e74 75e79 80 Sex Male Female Comorbidities Obesity Diabetes mellitus Arterial hypertension Hyperlipidemia Coronary heart disease Myocardial infarction Cardiac dysrhythmia PAOD Kidney dysfunction Stroke/TIA Dementia Parkinson disease Blindness/low vision COPD Osteoporosis Arthritis Drugs Calcium/vitamin D Bisphosphonates HRT SERMs Calcitonin Steroids Acetaminophen NSAIDs COX-2 inhibitors Antidepressants Stimulants Antipsychotics Mood stabilizers Anxiolytics Depressants

Cases, n ¼ 8027

Controls, n ¼ 79,288

P Value

16 56 144 193 361 771 3074 3412

1270 2620 4763 8574 10,082 15,984 17,783 19,212

(1.6) (3.3) (6.0) (10.8) (12.7) (19.0) (22.4) (24.2)

.299 .346 .014* .015* .002* .006* .009* .001*

1710 (21.3) 6317 (78.7)

26,992 (34.0) 42,396 (66.0)

.017* .013*

1638 2834 2561 1589 1485 297 273 281 369 859 345 297 337 201 5081 209

(20.4) (35.3) (31.9) (19.8) (18.5) (3.7) (3.4) (3.5) (4.6) (10.7) (4.3) (3.7) (4.2) (2.5) (63.3) (2.6)

6510 14,846 16,989 6669 13,337 2699 1905 3017 3414 3572 1985 1667 1747 2143 49,618 1588

(8.2) (18.7) (21.4) (8.4) (16.8) (3.4) (2.4) (3.8) (4.3) (4.5) (2.5) (2.1) (2.2) (2.7) (62.5) (2.0)

.002* .002* .016* .019* .224 .523 .457 .408 .480 .029* .059 .672 .003* .372 .3010 .482

2432 1670 955 931 1220 2769 2256 3163 1766 610 385 257 153 250 217

(30.3) (20.8) (11.9) (11.6) (15.2) (34.5) (28.1) (39.4) (22.0) (7.6) (4.8) (3.2) (1.9) (3.1) (2.7)

25,007 25,642 11,194 10,638 13,258 10,956 9130 16,195 12,861 4843 3652 2461 1350 2381 2779

(31.5) (32.3) (14.1) (13.4) (16.7) (13.8) (11.5) (20.4) (16.2) (6.1) (4.6) (3.1) (1.7) (3.0) (3.5)

.205 .023* .149 .142 .186 .019* .013* .003* .024* .272 .679 .511 .295 .689 .836

(0.2) (0.7) (1.8) (2.4) (4.5) (9.6) (38.3) (42.5)

COPD, chronic obstructive pulmonary disease; COX-2, cyclooxygenase-2; HRT, hormone replacement therapy; NSAIDs, nonsteroidal anti-inflammatory drugs; PAOD, peripheral arterial occlusive disease; SERMs, selective estrogen receptor modulators; TIA, transient ischemic attack. Values are n (%). *P < .05 is significant and all analysis was done by logistic regression model in SAS 9.2 (SAS Institute, Inc, Cary, NC).

2.7-fold during the follow-up period, from 1154 in 2004 to 3121 in 2007. We also observed a significantly increased mortality rate in patients after an SHFx compared with those who did not have an SHFx, with a 7-year cumulative risk ranging from 1.39% to 6.80% in the case group and 0.64% to 4.08% in the control group depending on age. Thus, the age-specific mortality was increased 1.6- to 2.2-fold in patients with SHFx, compared with those without, after HFxS in this 7-year follow-up study. The high overall incidence and increased mortality rate of SHFx in our study highlight the importance of refinement of the treatment program and tertiary prevention strategy to improve the medical care of elderly patients after their first fragility HFx. To the best of our knowledge, this study is the first report on the risk of SHFx in Asian patients after HFxS using a nationwide population-based register. Although Ryg et al17 had explored this major health concern in Western populations using the National

S.-H. Shen et al. / JAMDA 15 (2014) 725e731

729

Table 3 Logistic Regression Analysis of Second Hip Fracture Occurrence of Cases Versus Controls Variables

Univariate Analysis

Age Sex Comorbidities Obesity Diabetes mellitus Arterial hypertension Hyperlipidemia Coronary heart disease Myocardial infarction Cardiac dysrhythmia PAOD Kidney dysfunction Stroke/TIA Dementia Parkinson’s disease Blindness/low vision COPD Osteoporosis Arthritis Drugs Calcium/vitamin D Bisphosphonates HRT SERMs Calcitonin Steroids Acetaminophen NSAIDs COX-2 inhibitors Antidepressants Stimulants Antipsychotics Mood stabilizers Anxiolytics Depressants

Multivariate Analysis

UOR

95% CI

P Value

AOR

95% CI

P Value

1.68 1.08

1.12e2.03 0.84e1.49

.007* .010*

1.84 1.12

1.24e2.89 1.03e2.30

.009* .030*

2.15 3.49 2.38 2.23 0.07 0.22 0.60 0.40 0.65 2.55 0.77 0.59 2.07 0.52 0.43 0.67

1.37e3.47 2.01e4.26 1.06e2.85 1.43e3.26 0.27e1.20 0.03e1.84 0.10e0.74 0.25e0.79 0.44e0.79 1.36e3.24 0.23e1.08 0.43e0.80 1.27e2.84 0.23e0.80 0.18e0.77 0.37e1.40

.029* .004* .002* .001* .221 .611 .638 .115 .329 .004* .480 .433 .011* .535 .653 .421

2.89 3.85 2.45 2.77 d d d d d 2.85 d d 3.09 d d d

1.81e3.01 2.54e4.05 1.83e2.62 1.27e3.19 d d d d d 2.20e3.23 d d 2.54e3.73 d d d

.033* .003* .006* .007* d d d d d .011* d d .023* d d d

0.95 1.94 0.64 0.77 0.96 3.88 2.65 3.70 2.53 0.28 0.08 0.77 0.22 0.21 0.42

0.56e1.62 1.09e2.29 0.43e0.94 0.27e1.20 0.86e1.07 2.66e4.84 1.21e3.40 2.41e4.33 1.36e3.85 0.11e0.72 0.01e0.91 0.27e1.20 0.03e1.84 0.08e0.55 0.17e1.56

.864 .002* .776 .634 .508 .003* .012* .007* .020* .255 .723 .221 .611 .236 .892

d 2.24 d d d 4.17 3.45 4.12 3.05 d d d d d d

d 1.38e2.90 d d d 3.90e5.32 2.43e4.68 3.80e4.55 2.28e3.50 d d d d d d

d .017* d d d .011* .009* .002* .012* d d d d d d

AOR, adjusted odds ratio; CI, confidence interval; COPD, chronic obstructive pulmonary disease; COX-2, cyclooxygenase-2; HRT, hormone replacement therapy; NSAIDs, nonsteroidal anti-inflammatory drugs; PAOD, peripheral arterial occlusive disease; SERMs, selective estrogen receptor modulators; TIA, transient ischemic attack; UOR, unadjusted odds ratio. *P < .05 is significant and all analysis was done by logistic regression model in SAS 9.2 (SAS Institute, Inc, Cary, NC).

Hospital Discharge Register in Denmark from 1977 to 2001, they did not exclude patients younger than 45 years and those who sustained HFx secondary to a high-energy trauma. In our nationwide Table 4 Crude and Adjusted Hazard Ratios of Second Hip Fracture Occurrence Among Cases in Different Age Groups Second Hip Fracture, n ¼ 8027

Patients

Risk analysis for second hip fracture in patients after hip fracture surgery: a nationwide population-based study.

The current treatment program for fragility hip fractures (HFx) emphasizes a combination of early surgery, rehabilitation, and tertiary prevention str...
577KB Sizes 0 Downloads 3 Views