Osteoporos Int DOI 10.1007/s00198-017-4192-z

ORIGINAL ARTICLE

Risk and outcomes of fracture in peripheral arterial disease patients: two nationwide cohort studies F.-L. Liu 1,2 & C.-S. Lin 2,3,4 & C.-C. Yeh 5,6 & C.-C. Shih 7 & Y.-G. Cherng 1,2 & C.-H. Wu 8 & T.L. Chen 2,3,4 & C.-C. Liao 1,2,3,4,9

Received: 2 February 2017 / Accepted: 7 August 2017 # International Osteoporosis Foundation and National Osteoporosis Foundation 2017

Abstract Summary Using national insurance claims data of Taiwan, we found that patients with peripheral arterial disease (PAD) had increased risk of fracture during the follow-up period of 2000–2013. History of PAD was also associated with adverse outcomes in hospitalized fracture patients. Prevention strategies were needed in this susceptible population.

Ta-Liang Chen has equal contribution with the corresponding author. * C.-C. Liao [email protected]; [email protected] 1

Department of Anesthesiology, Shuang Ho Hospital, Taipei Medical University, New Taipei City, Taiwan

2

Department of Anesthesiology, School of Medicine, College of Medicine, Taipei Medical University, Taipei, Taiwan

3

Department of Anesthesiology, Taipei Medical University Hospital, 252 Wuxing St., Taipei 110, Taiwan

4

Health Policy Research Center, Taipei Medical University Hospital, Taipei, Taiwan

5

Department of Surgery, China Medical University Hospital, China Medical University, Taichung, Taiwan

6

Department of Surgery, University of Illinois, Chicago, IL, USA

7

School of Chinese Medicine for Post-Baccalaureate, I-Shou University, Kaohsiung, Taiwan

8

Department of Family Medicine, National Cheng Kung University Hospital, Tainan, Taiwan

9

School of Chinese Medicine, College of Chinese Medicine, China Medical University, Taichung, Taiwan

Introduction Limited information was available on the association between PAD and fracture. The purpose of this study is to evaluate fracture risk and post-fracture outcomes in patients with PAD. Methods We identified 6647 adults aged ≥ 20 years with newly diagnosed PAD using the Taiwan National Health Insurance Research Database in 2000–2004. Comparison cohort consisted of 26,588 adults without PAD randomly selected with frequency matching in age and sex. Events of fracture were identified during the follow-up period from January 1, 2000 until December 31, 2013, to evaluate adjusted hazard ratios (HR) and 95% confidence interval (CI) of fracture associated with PAD. Another nested cohort study of 799,463 hospitalized fracture patients analyzed adjusted odds ratios (ORs) and 95% CIs of adverse events after fracture among patients with and without PAD in 2004–2013. Results Incidences of fracture in people with and without PAD were 22.1 and 15.5 per 1000 person-years, respectively (P < .0001). Compared with control, the adjusted HR of fracture was 1.59 (95% CI, 1.48–1.69) for PAD patients. In the nested cohort study, patients with PAD had higher post-fracture mortality (OR = 1.16; 95% CI, 1.09–1.25) and various complications. PAD patients also had comparatively higher medical expenditure (2691 vs. 2232 USD, P < .0001) and longer hospital stay (10.6 vs. 9.0 days, P < 0.0001) during fracture admission. Conclusions Increased risk of fracture and post-fracture adverse outcomes were associated with PAD. This susceptible population needs care to prevent fracture and to minimize adverse outcomes after it occurs.

Osteoporos Int

Keywords Fracture . Outcome . Peripheral arterial disease . Risk

Abbreviations ICD-9International Classification of Diseases, 9th CM Revision, Clinical Modification CI Confidence interval HR Hazard ratio OR Odds ratio PAD Peripheral arterial disease

Introduction Peripheral arterial disease (PAD) of lower extremity affects over eight million Americans [1], with hospitalization and amputation for lower limb ischemia costing over USD 4 billion annually [2]. Due to preexisting comorbidities, patients with PAD also commonly present with minimal organ reserves leading to additional risk of complications, especially cardiopulmonary [3]. PAD is recognized as the leading cause of limited ambulation with intermittent claudication, reduced strength in the lower extremities, impaired balance, and higher prevalence of falls [4, 5]. Previous studies found significant association between PAD and reduced bone mass density primarily at femur [6, 7]. We considered it is reasonable that greater rate of falling with lower femoral neck bone mass might contribute to increased risk of fracture in PAD patients. However, the findings were inconclusive because some studies showed fracture increased in these patients while other studies did not [7–10]. For investigating whether PAD is associated with increased fracture risk, we conducted a nationwide retrospective cohort study using Taiwan National Health Insurance Research Database. We also were concerned about the comorbidities accompanying PAD, so we compared the risks of complications and in-hospital mortality during hospitalization for fracture in patients with and without PAD.

Methods Source of data Taiwan’s National Health Insurance Program has collected all medical reimbursement claims since 1996 in a Taiwan National Health Insurance Research Database. This is available to researchers with beneficiaries’ identification scrambled to protect privacy. Information sets available for this study

include sex, birth date, diagnoses at outpatient visits, admissions and discharges, health care provided, medications prescribed, medical institutions used, and physicians providing services. For research and administrative purposes, Taiwan’s National Health Research Institute also provided a data subset of one million beneficiaries randomly selected from insurance enrollees ages 0–113 years representing nearly 5% of Taiwan’s insured population. Information about medical care for these persons was collected from 1996 to 2008. The validity of studies based on this database has been well accepted by scientific journals worldwide [11, 12]. Our retrospective cohort study and nested cohort analysis were based on information from this source, and the Taipei Medical University institutional review board approved the study (TMU-JIRB201705084). Study design In this longitudinal cohort of one million insured individuals, we identified the exposed cohort as 6647 patients age 20 years or older who were newly diagnosed with PAD between 2000 and 2004 who had no history of fracture before the index date of diagnosis (without any previous record of PAD diagnosis or treatment in the database, which was established in 1996). To avoid coding error, we required at least two outpatient visits or inpatient medical services with physician’s PAD diagnosis. During the same index period, we identified 26,588 people age 20 years or older (fourfold non-PAD controls) matched by age and sex who had neither PAD diagnosis nor history of fracture as the non-exposure cohort. Patients with diagnosis of fracture between January 1, 1996 and December 31, 2004 were excluded to ensure that all study participants were free of fracture at the start of both cohorts. Participants with previous diagnosis of osteoporosis or receiving bisphosphonate treatment before the index date were excluded. Follow-up period started from January 1, 2000 and lasted until December 31, 2008 or censoring due to death, lost to follow-up, or other causes. During the follow-up period, participants who visited medical care (including outpatient, emergency, and inpatient care) with physician’s primary diagnosis of fracture were determined as fracture events. The study seeks to investigate whether individuals with PAD faced increased risk of fracture. To investigate the impact of PAD on post-fracture outcomes, a nested retrospective cohort study was conducted in the Surgical Cohort Database (one subgroup database in the Taiwan National Health Insurance Research Database). We identified 799,463 hospitalized patients with first fracture admission receiving inpatient care with physician’s primary diagnosis in 2004–2013, including 14,986 patients with prefracture PAD and 784,477 without PAD. Among hospitalized

Osteoporos Int

fracture patients, we compared post-fracture complications such as urinary tract infection, septicemia, pneumonia, stroke, acute renal failure, deep wound infection, acute myocardial infarction, pulmonary embolism, and mortality for a 30-day period after fracture among patients with or without prefracture PAD. These post-fracture complications were identified as participants who had physician’s primary or secondary diagnosis in the index admission or 30-day post-admission outpatient care and inpatient care. Patients who sought medical care with physician’s diagnosis of urinary tract infection, septicemia, pneumonia, stroke, acute renal failure, deep wound infection, acute myocardial infarction, and pulmonary embolism before fracture admission within 6 months were excluded in this study. Measures and definitions We identified income status by defining low income patients as those qualifying for waived medical copayment as verified by the National Health Insurance Administration. 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 anxiolytics, oral steroids, antiepileptics, antidepressants, antipsychotics, and osteoporosis medications (included estrogen, bisphosphonates, calcitonin, raloxifene, teriparatide, and denosumab) was also analyzed. We used the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) to define coexisting medical conditions and post-fracture complications. Pre-fracture PAD (ICD-9-CM 250.7, 440.2, 440.3, 443, 444, 785.4) was defined as primary exposure. Coexisting medical conditions determined from medical claims for the follow-up period included mental disorders (ICD-9-CM 290–319), hypertension (ICD-9-CM 401–405), chronic obstructive pulmonary disease (ICD-9-CM 490–496), ischemic heart disease (ICD-9-CM 410–414), diabetes (ICD-9-CM 250), cancer (ICD-9-CM 140–239), stroke (ICD-9-CM 430– 438), hyperlipidemia (ICD-9-CM 272.0, 272.1, 272.2), asthma (ICD-9-CM 493), congestive heart failure (ICD-9-CM 428), and liver cirrhosis (ICD-9-CM 571.2, 571.5). 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), lower limb (ICD-9-CM 820–829), and hip (ICD-9-CM 820). The association of PAD with various fractures (including skull bone, neck, trunk, upper limb, lower limb, hip fracture, and at least two types of fracture) and post-fracture mortality was also analyzed. Complications after fracture were analyzed as secondary outcomes, including

urinary tract infection (ICD-9-CM 599.0), septicemia (ICD9-CM 038 and 998.5), pneumonia (ICD-9-CM 480–486), stroke (ICD-9-CM 430–438), acute renal failure (ICD-9-CM 584), deep wound infection (ICD-9-CM 958.3), acute myocardial infarction (ICD-9-CM 410), and pulmonary embolism (ICD-9-CM 415). In-hospital post-fracture mortality was also considered as outcome in the nested cohort study. Statistical analysis In the retrospective cohort study including all participants, we used chi-square tests to compare sociodemographic characteristics and coexisting medical conditions between people with and without PAD. We calculated the hazard ratios (HRs) with 95% confidence intervals (CIs) for risk of fracture after PAD, using multivariate Cox proportional hazards regression models to adjust for age, sex, low income, urbanization, mental disorders, hypertension, chronic obstructive pulmonary disease, ischemic heart disease, diabetes, cancer, stroke, hyperlipidemia, asthma, congestive heart failure, liver cirrhosis, renal dialysis, and medications. We also calculated adjusted HRs of fracture associated with diabetes in the stratified analyses of age, sex, and number of medical conditions. In the nested cohort study, the sociodemographics and coexisting medical conditions between fracture patients with and without PAD were compared using chisquare tests. The adjusted odds ratios (ORs) and 95% CIs of post-fracture urinary tract infection, septicemia, pneumonia, stroke, acute renal failure, deep wound infection, acute myocardial infarction, and pulmonary embolism as well as mortality associated with pre-fracture PAD were calculated in the multivariate logistic regressions with adjustment for age, sex, low income, urbanization, size of hospital, mental disorders, hypertension, chronic obstructive pulmonary disease, ischemic heart disease, diabetes, cancer, stroke, hyperlipidemia, asthma, congestive heart failure, liver cirrhosis, renal dialysis, fracture-repair surgery, and medications. Risks and mortality of fracture subtypes for PAD patients were also calculated in the multivariate Cox models and logistic regression models, respectively. SAS version 9.1 (SAS Institute, Inc., Cary, NC) statistical software was used for data analyses; two-sided P < .05 indicated significant differences between groups.

Results The retrospective cohort study (Table 1) showed no significant differences in age and sex after frequency matching in age and sex between cohorts with and without PAD. Compared with the cohort without PAD, patients with PAD had higher proportions with hypertension, diabetes, mental disorders,

Osteoporos Int Table 1 Baseline characteristics of people with and without peripheral arterial disease

Non-PAD (N = 26,588)

PAD (N = 6647)

P value

Sex Female

n (%) 13,692 (51.5)

n (%) 3423 (51.5)

1.0000

Male

12,896 (48.5)

3224 (48.5)

Age, years 20–29

928 (3.5)

232 (3.5)

30–39

1868 (7.0)

467 (7.0)

40–49 50–59

4540 (17.1) 5040 (19.0)

1135 (17.1) 1260 (19.0)

60–69

6332 (23.8)

1583 (23.8)

70–79 ≥ 80

5968 (22.5) 1912 (7.2)

1492 (22.5) 478 (7.2)

Low income

844 (3.2)

367 (5.5)

< 0.0001

Medical conditions Hypertension

1.0000

10,129 (38.1)

3600 (54.2)

< 0.0001

Diabetes Mental disorders

4187 (15.8) 7835 (29.5)

3364 (50.6) 2883 (43.4)

< 0.0001 < 0.0001

COPD Ischemic heart disease Hyperlipidemia Stroke

6177 (23.2) 3878 (14.6) 4598 (17.3) 2304 (8.7)

2150 (32.4) 1898 (28.6) 1666 (25.1) 1209 (18.2)

< < < <

Cancer Asthma Osteoporosis

3418 (12.9) 2246 (8.5) 1817 (6.8)

952 (14.3) 748 (11.3) 672 (10.1)

< 0.0001 < 0.0001 < 0.0001

Congestive heart failure Liver cirrhosis Renal dialysis Parkinson’s disease Alcohol-related illness

870 (3.3) 1277 (4.8) 252 (1.0) 666 (2.5) 680 (2.6)

568 (8.6) 556 (8.4) 518 (7.8) 323 (4.9) 270 (4.1)

< < < < <

0.0001 0.0001 0.0001 0.0001 0.0001

13,201 (49.7) 7516 (28.3) 4099 (15.4) 4406 (16.6)

4921 (74.0) 2383 (35.9) 2043 (30.7) 2043 (30.7)

< < < <

0.0001 0.0001 0.0001 0.0001

4191 (15.8) 3623 (13.6) 656 (2.5) 401 (1.5) 164 (0.6) 31 (0.1) 22 (0.1)

1827 (27.5) 1223 (18.4) 221 (3.3) 176 (2.7) 58 (0.9) 8 (0.1) 5 (0.1)

< 0.0001 < 0.0001 < 0.0001 < 0.0001 0.0220 0.9361 0.8473

Medication use Anxiolytics Oral steroids Antiepileptics Antidepressants Antipsychotics Estrogen Bisphosphonates Calcitonin Raloxifene Teriparatide Denosumab

0.0001 0.0001 0.0001 0.0001

COPD chronic obstructive pulmonary disease, PAD peripheral arterial disease

chronic obstructive pulmonary disease, ischemic heart disease, hyperlipidemia, stroke, cancer, asthma, osteoporosis, congestive heart failure, liver cirrhosis, renal dialysis, Parkinson’s disease, and alcohol-related illness (all P < 0.0001). The proportions of selected medications were higher in patients with PAD than in those without PAD, including anxiolytics, oral steroids, antiepileptics, antidepressants,

antipsychotics, estrogen, bisphosphonates, and calcitonin (all P < 0.0001). During the follow-up period (Table 2), higher incidence of fracture was found in PAD patients than in those without PAD (22.1 vs. 15.5 per 1000 person-years, P < 0.0001); the corresponding HR of fracture associated with PD was 1.59 (95% CI, 1.48–1.69). Increased risk of fracture associated with PAD was

Osteoporos Int Table 2 Risk of fracture events for cohorts with and without PAD No PAD

n 26,588

PY 280,748

Events 4342

IRa 15.5

HR (95% CI)b 1.00 (reference)

PAD Female

No PAD

6647 13,692

66,124 143,833

1464 2648

22.1 18.4

1.59 (1.48–1.69) 1.00 (reference)

Male

PAD No PAD

3423 12,896

34,231 136,915

882 1694

25.8 12.4

1.57 (1.45–1.71) 1.00 (reference)

PAD

3224

31,893

582

18.2

1.50 (1.44–1.78)

Age, 20–39 years

No PAD PAD

2796 699

31,958 7719

251 103

7.85 13.3

1.00 (reference) 1.91 (1.48–2.46)

Age, 40–49 years

No PAD PAD

4540 1135

50,476 12,155

572 221

11.3 18.2

1.00 (reference) 1.95 (1.64–2.33)

Age, 50–59 years

No PAD

5040

53,923

856

15.9

1.00 (reference)

PAD

1260

13,092

274

20.9

1.63 (1.40–1.90)

Age, 60–69 years

No PAD PAD

6332 1583

66,430 15,832

1216 393

18.3 24.8

1.00 (reference) 1.58 (1.39–1.80)

Age, 70–79 years

No PAD PAD

5968 1492

58,850 13,305

1237 376

21.0 28.3

1.00 (reference) 1.39 (1.23–1.58)

Age, ≥ 80 years

No PAD

1912

19,111

210

11.0

1.00 (reference)

Hypertension

PAD No PAD

478 10,129

4023 106,350

97 1881

24.1 17.7

1.34 (1.01–1.79) 1.00 (reference)

PAD No PAD PAD

3600 4187 3364

36,386 44,305 33,161

772 700 764

21.2 15.8 23.0

1.44 (1.32–1.58) 1.00 (reference) 1.61 (1.45–1.80)

No PAD PAD No PAD PAD

7835 2883 6177 2150

83,935 29,879 64,658 21,736

1280 579 1109 448

15.2 19.4 17.2 20.6

1.00 (reference) 1.44 (1.29–1.60) 1.00 (reference) 1.37 (1.21–1.54)

No PAD PAD No PAD

3878 1898 4598

40,961 19,162 50,744

666 386 664

16.3 20.1 13.1

1.00 (reference) 1.39 (1.22–1.60) 1.00 (reference)

PAD No PAD PAD No PAD PAD No PAD PAD No PAD PAD No PAD PAD No PAD PAD No PAD PAD No PAD PAD No PAD

1666 2304 1209 3418 952 2246 748 1817 672 870 568 1277 556 252 518 666 323 680

18,226 23,733 12,177 34,624 9608 23,557 7572 18,961 7011 9029 5660 13,323 5761 2455 5061 6962 3318 7105

312 324 189 355 114 406 145 444 175 132 85 198 104 38 79 118 64 119

17.1 13.7 15.5 10.3 11.9 17.2 19.1 23.4 25.0 14.6 15.0 14.9 18.1 15.5 15.6 16.9 19.3 16.7

1.54 (1.33–1.78) 1.00 (reference) 1.16 (0.96–1.42) 1.00 (reference) 1.25 (0.99–1.57) 1.00 (reference) 1.30 (1.06–1.60) 1.00 (reference) 1.32 (1.10–1.60) 1.00 (reference) 1.19 (0.88–1.63) 1.00 (reference) 1.27 (0.98–1.65) 1.00 (reference) 1.17 (0.76–1.80) 1.00 (reference) 1.53 (1.09–2.16) 1.00 (reference)

PAD No PAD PAD

270 13,201 4921

2689 140,642 50,202

60 2292 1012

22.3 16.3 20.2

1.91 (1.36–2.70) 1.00 (reference) 1.40 (1.29–1.51)

Diabetes Mental disorders COPD Ischemic heart disease Hyperlipidemia Stroke Cancer Asthma Osteoporosis Congestive heart failure Liver cirrhosis Renal dialysis Parkinson’s disease Alcohol-related illness Anxiolytics

Osteoporos Int Table 2 (continued) Oral steroids Antiepileptics Antidepressants Antipsychotics 0 medical condition 1 medical condition 2 medical conditions 3 medical conditions ≥4 medical conditions

No PAD PAD No PAD PAD No PAD PAD No PAD PAD No PAD PAD No PAD PAD No PAD PAD No PAD PAD No PAD PAD

7516 2383 4099 2043 4406 2043 4191 1827 7734 449 5300 934 4829 1257 3751 1274 4974 2733

83,622 26,118 44,208 21,262 47,746 21,511 44,864 19,239 84,052 4301 54,214 8793 49,829 12,181 39,381 12,551 53,271 28,299

1057 398 645 408 746 420 720 356 1031 123 1000 244 908 308 642 288 761 501

12.6 15.2 14.6 19.2 15.6 19.5 16.0 18.5 12.3 28.6 18.4 27.7 18.2 25.3 16.3 22.9 14.3 17.7

1.00 (reference) 1.40 (1.23–1.58) 1.00 (reference) 1.52 (1.32–1.74) 1.00 (reference) 1.36 (1.19–1.55) 1.00 (reference) 1.37 (1.20–1.58) 1.00 (reference) 2.11 (1.73–2.57) 1.00 (reference) 1.74 (1.51–2.01) 1.00 (reference) 1.53 (1.34–1.75) 1.00 (reference) 1.55 (1.34–1.78) 1.00 (reference) 1.31 (1.17–1.47)

CI confidence interval, COPD chronic obstructive pulmonary disease, HR hazard ratio, IR incidence rate, PAD peripheral arterial disease, PY person-years a

Per 1000 person-years

b

Adjusted for all covariates listed in Table 1

found in both genders (female: HR = 1.57; 95% CI, 1.45–1.71; male: HR = 1.50; 95% CI, 1.44–1.78). Patients with PAD were at higher risk of fracture in all age groups, specifically 20–39 (HR = 1.91; 95% CI, 1.48–2.46), 40–49 (HR = 1.95; 95% CI, 1.64–2.33), 50–59 (HR = 1.63; 95% CI, 1.40–1.90), 60–69 (HR = 1.58; 95% CI, 1.39–1.80), 70–79 (HR = 1.39; 95% CI, 1.23–1.58), and ≥ 80 years (HR = 1.34; 95% CI, 1.01– 1.79). The HRs of fracture associated with PAD for people with 0, 1, 2, 3, and ≥ 4 medical conditions were 2.11 (95% CI, 1.73– 2.57), 1.74 (95% CI, 1.51–2.01), 1.53 (95% CI, 1.34–1.75), 1.55 (95% CI, 1.34–1.78), and 1.31 (95% CI, 1.17–1.47), respectively. Fracture risk was associated with PAD and various medical conditions, including hypertension (HR = 1.44; 95% CI, 1.32–1.58), diabetes (HR = 1.61; 95% CI, 1.45–1.80), mental disorders (HR = 1.44; 95% CI, 1.29–1.60), chronic obstructive pulmonary disease (HR = 1.37; 95% CI, 1.21–1.54), ischemic heart disease (HR = 1.39; 95% CI, 1.22–1.60), hyperlipidemia (HR = 1.54; 95% CI, 1.33–1.78), asthma (HR = 1.30; 95% CI, 1.06–1.60), osteoporosis (HR = 1.32; 95% CI, 1.10– 1.60), Parkinson’s disease (HR = 1.53; 95% CI, 1.09–2.16), and alcohol-related illness (HR = 1.91; 95% CI, 1.36–2.70). The association between PAD and fracture was also significant in medication subgroup analyses including anxiolytics (HR = 1.40; 95% CI, 1.29–1.51), oral steroids (HR = 1.40; 95% CI, 1.23–1.58), antiepileptics (HR = 1.52; 95% CI, 1.32–1.74), antidepressants (HR = 1.36; 95% CI, 1.19–1.55), and antipsychotics (HR = 1.37; 95% CI, 1.20–1.58). In the nested cohort study including 799,463 patients with fracture (Table 3), higher proportions of patients with PAD than control were noted among females, older people, and those with low income status, stay in small size of hospital,

diabetes, hypertension, mental disorders, chronic obstructive pulmonary disease, ischemic heart disease, hyperlipidemia, osteoporosis, stroke, cancer, renal dialysis, asthma, congestive heart failure, Parkinson’s disease, liver cirrhosis, and alcoholrelated illness (all P < 0.0001). Fracture patients with PAD used more medications, including anxiolytics, antiepileptics, antipsychotics, antidepressants, oral steroids, bisphosphonates, calcitonin, raloxifene, teriparatide, and denosumab (all P < 0.0001). Table 4 shows PAD was associated with post-fracture urinary tract infection (OR = 1.14; 95% CI, 1.09–1.18), septicemia (OR = 1.24; 95% CI, 1.19–1.30), pneumonia (OR = 1.13; 95% CI, 1.07–1.18), stroke (OR = 1.12; 95% CI, 1.06–1.18), acute renal failure (OR = 1.28; 95% CI, 1.19–1.38), deep wound infection (OR = 1.20; 95% CI, 1.08–1.32), acute myocardial infarction (OR = 1.30; 95% CI, 1.16–1.47), pulmonary embolism (OR = 1.29; 95% CI, 1.03–1.62), and subsequent mortality (OR = 1.16; 95% CI, 1.09–1.25). Compared with fracture patients without PAD, those with PAD had higher medical expenditure (2691 ± 3103 vs. 2232 ± 2942 USD, P < 0.0001) and longer stay (10.6 ± 13.7 vs. 9.0 ± 17.2 days, P < 0.0001). The risks of neck or trunk fracture (HR = 1.52; 95% CI, 1.34–1.72), upper limb fracture (HR = 1.56; 95% CI, 1.39– 1.75), lower limb fracture (HR = 1.44; 95% CI, 1.29–1.61), hip fracture (HR = 1.48; 95% CI, 1.28–1.72), and ≥ 2 types of fracture (HR = 1.59; 95% CI, 1.48–1.69) increased in patients with PAD. Post-fracture mortality increased in patients with PAD after neck or trunk fracture (OR = 1.21; 95% CI, 1.01–1.46), upper limb fracture (OR = 1.23; 95% CI, 1.03– 1.48), and lower limb fracture (OR = 1.15; 95% CI, 1.06– 1.24) (Table 5).

Osteoporos Int Table 3 Characteristics of hospitalized fracture patients with and without peripheral arterial disease

Non-PAD (N = 784,477) Sex

PAD (N = 14,986)

P value < 0.0001

n (%)

n (%)

Female

378,027 (48.2)

9035 (60.3)

Male Age, years

406,450 (51.8)

5951 (39.7)

20–29

109,424 (14.0)

137 (0.9)

30–39 40–49

95,511 (12.2) 113,497 (14.5)

277 (1.9) 742 (5.0)

50–59

133,006 (17.0)

1895 (12.7)

60–69 70–79

107,324 (13.7) 120,783 (15.4)

3084 (20.6) 4805 (32.1)

104,932 (13.4)

4046 (27.0)

≥ 80 Size of hospital

< 0.0001

< 0.0001

Small

301,123 (38.4)

6202 (41.4)

Moderate Large

264,414 (33.7) 218,940 (27.9)

5154 (34.4) 3630 (24.2)

23,052 (2.9)

680 (4.5)

< 0.0001

Diabetes Hypertension Mental disorders COPD

94,800 (12.1) 172,320 (22.0) 132,252 (16.9) 72,553 (9.3)

8286 (55.3) 6383 (42.6) 4163 (27.8) 2674 (17.8)

< < < <

IHD Hyperlipidemia

44,089 (5.6) 41,487 (5.3)

2619 (17.5) 1596 (10.7)

< 0.0001 < 0.0001

Osteoporosis

26,808 (3.4)

1130 (7.5)

< 0.0001

Stroke Cancer Renal dialysis Asthma CHF

15,636 (2.0) 31,345 (4.0) 8069 (1.0) 26,834 (3.4) 10,449 (1.3)

1005 (6.7) 970 (6.5) 980 (6.5) 930 (6.2) 805 (5.4)

< < < < <

16,117 (2.1) 17,361 (2.2) 18,235 (2.3)

739 (4.9) 567 (3.8) 360 (2.4)

< 0.0001 < 0.0001 0.5316

208,795 (26.6) 56,458 (7.2) 50,117 (6.4) 40,267 (5.1) 35,234 (4.5) 19,126 (2.4) 15,824 (2.0) 24,198 (3.1)

7098 (47.4) 2523 (16.8) 1979 (13.2) 1887 (12.6) 1291 (8.6) 812 (5.4) 784 (5.2) 477 (3.2)

< 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 < 0.0001 0.4904

5811 (0.7) 1810 (0.2) 395 (0.1) 684,138 (87.2)

270 (1.8) 74 (0.5) 23 (0.2) 12,365 (82.5)

< < < <

62,050 (7.9) 82,717 (10.5) 326,156 (41.6)

488 (3.3) 1950 (13.0) 3957 (26.4)

< 0.0001 < 0.0001 < 0.0001

Low income Medical conditions

Parkinson’s disease Liver cirrhosis Alcohol-related illness Medication use Anxiolytics Antiepileptics Antipsychotics Antidepressants Oral steroids Bisphosphonates Calcitonin Estrogen Raloxifene Teriparatide Denosumab Fracture-repair surgery Types of fracture Skull bone fracture Neck or trunk fracture Upper limb

0.0001 0.0001 0.0001 0.0001

0.0001 0.0001 0.0001 0.0001 0.0001

0.0001 0.0001 0.0001 0.0001

Osteoporos Int Table 3 (continued) Hip fracture Lower limb

Non-PAD (N = 784,477) 177,149 (22.6) 384,368 (49.0)

PAD (N = 14,986) 7115 (47.5) 9530 (63.6)

P value < 0.0001 < 0.0001

≥ 2 types of fracture

143,310 (18.3)

2003 (13.4)

< 0.0001

CHF congestive heart failure, COPD chronic obstructive pulmonary disease, IHD ischemic heart disease, PAD peripheral arterial disease

Discussion Using claims data from Taiwan’s National Health Insurance in this nationwide retrospective cohort study, we found an increased risk of fracture in patients with PAD in both sexes, all age groups, people with and without medical conditions, and those with every type of fracture. Our nested cohort study showed PAD was associated with increased mortality and complications, including urinary tract infection, pneumonia, septicemia, stroke, pulmonary embolism, acute renal failure, acute myocardial infarction, and deep wound infection in hospitalized fracture patients. This is the first report providing comprehensive information on associations between PAD and fracture. Previous studies suggested that older age, female gender, and low income were risk factors for fracture and PAD [8, 13–16]. In the present two studies, we considered these sociodemographic characteristics as potential confounding factors and used multivariate regression models to reduce confounding bias when investigating the associations between PAD and fractures. Osteoporosis is considered a risk factor for fracture and is well studied in both general populations and PAD patients [6, 7, 13, 17]. Postmenopausal women have higher incidence of osteoporosis because they lose bone mass faster than men [18]. In our subgroup analysis, female gender Table 4 Risks of complications and mortality during fracture admission for patients with PAD

and fracture were closely related, a finding compatible with previous studies [6, 9, 14, 16]. Hypertension, diabetes, mental disorders, asthma or chronic obstructive pulmonary disease, ischemic heart disease, hyperlipidemia, stroke, liver cirrhosis, renal failure, and Parkinson’s disease have been shown to be independently associated with risk of fracture [11, 12, 19–22]. These medical conditions are also considered as comorbidities for patients with PAD [23, 24]. Psychiatric medications such as anxiolytics and antidepressants have been suggested as contributors to fracture risk [39, 40]. However, previous studies were limited by inadequate control for coexisting medical conditions and medications when investigating associations between PAD and fracture risk [7–10]. Our study suggested that patients with PAD had increased risk of fracture after adjustment for these potential medical confounding factors. In addition, our stratified analysis showed the significant association between PAD and fracture risk also existed among people taking anxiolytics, oral steroids, antiepileptics, antidepressants, and antipsychotics. The relationship between PAD and reduced bone mass has been found in postmenopausal women, particularly at the femur [25]. Some studies also suggested that aortic calcification was an independent risk factor for hip fracture in men and postmenopausal women [26, 27]. It is assumed that fractures

Post-fracture outcomes

Non-PAD (N = 784,477) Events (%)

PAD (N = 14,986) Events (%)

Risk of outcomes OR (95% CI)a

Complications Urinary tract infection Septicemia Pneumonia Stroke Acute renal failure Deep wound infection

118,020 (15.0) 60,753 (7.7) 62,666 (8.0) 36,352 (4.6) 14,587 (1.9) 18,732 (2.4)

4585 (30.6) 2774 (18.5) 2593 (17.3) 1650 (11.0) 900 (6.0) 446 (3.0)

1.14 (1.09–1.18) 1.24 (1.19–1.30) 1.13 (1.07–1.18) 1.12 (1.06–1.18) 1.28 (1.19–1.38) 1.20 (1.08–1.32)

Acute myocardial infarction Pulmonary embolism Mortality Medical expenditure, USDb Length of hospital stay, daysb

4772 (0.6) 1666 (0.2) 21,621 (2.8) 2232 ± 2942 9.0 ± 17.2

334 (2.2) 82 (0.6) 1059 (7.1) 2691 ± 3103 10.6 ± 13.7

1.30 (1.16–1.47) 1.29 (1.03–1.62) 1.16 (1.09–1.25) P < 0.0001 P < 0.0001

CI confidence interval, OR odds ratio, PAD peripheral arterial disease a

Adjusted for all covariates listed in Table 3

b

Mean ± SD

Osteoporos Int Table 5

Risks of fracture and post-fracture mortality in patients with PAD stratified by fracture subtypes

Types of fracture

Risk of fracture

Post-fracture mortality

PY

Non-PAD

26,588

305,986

159

0.52

1.00 (reference)

PAD

6647

75,089

48

0.64

1.36 (0.95–1.93)

Neck or trunk fracture

Non-PAD PAD

26,588 6647

300,442 72,856

1110 407

3.69 5.59

Upper limb

Non-PAD PAD

26,588 6647

297,858 72,394

1468 473

Lower limb

Non-PAD

26,588

297,229

Hip fracture

PAD Non-PAD

6647 26,588

71,992 302,911

PAD Non-PAD

6647 26,588

PAD

6647

Skull bone fracture

≥ 2 types of fracture

b

Events

IRc

n

HR (95% CI)a

n

OR (95% CI)b

Deaths

Mortality, %

62,050

2669

4.30

488

56

11.5

1.26 (0.94–1.71)

1.00 (reference) 1.52 (1.34–1.72)

82,717 1950

2812 138

3.40 7.08

1.00 (reference) 1.21 (1.01–1.46)

4.93 6.53

1.00 (reference) 1.56 (1.39–1.75)

326,156 3957

3537 150

1.08 3.79

1.00 (reference) 1.23 (1.03–1.48)

1605

5.40

1.00 (reference)

384,368

14,350

3.73

1.00 (reference)

536 767

7.45 2.53

1.44 (1.29–1.61) 1.00 (reference)

9530 177,149

761 12,594

7.99 7.11

1.15 (1.06–1.24) 1.00 (reference)

73,674 280,748

322 4342

4.37 15.5

1.48 (1.28–1.72) 1.00 (reference)

7115 143,310

674 3800

9.47 2.65

1.07 (0.99–1.17) 1.00 (reference)

66,124

1464

22.1

1.59 (1.48–1.69)

2003

103

5.14

0.96 (0.78–1.19)

1.00 (reference)

CI confidence interval, IR incidence rate, OR odds ratio, PAD peripheral arterial disease, PY person-years a

Adjusted for all covariates listed in Table 1

b

Adjusted for all covariates listed in Table 3

c

Per 1000 person-years during the follow-up period

in the thoracolumbar spine, proximal femur, proximal humerus, and distal radius were considered to be osteoporotic [13], and this previous finding is consistent with our investigations showing increased risk of neck or trunk, upper limb, lower limb, and hip fractures in patients with PAD. We investigated increased post-fracture mortality in patients with PAD in comparison with other fracture patients. Our findings were consistent with previous studies suggesting PAD is a risk factor for increased mortality and complications in hip or lower limb orthopedic surgery [28, 29]. As PAD is a marker for advanced atherosclerotic pathology [30], complex cardiovascular medication regimens might be interrupted during fracture admission, resulting in reduced treatment efficacy. Skull bone fracture, hip fracture, and at least two types of fractures were considered as severe fracture with poor clinical condition in admissions. The impact of PAD may decrease when other more severe medical conditions occur during fracture admissions, and that is why we could not investigate the association between PAD and post-fracture mortality involving these severe fractures in the subgroup analysis. The association between PAD and fracture was observed in every subgroup analysis, specifically in both sexes, all age groups, and people with and without medical conditions or medications. In epidemiological theory, these phenomena strengthen the causal inferences of PAD and risk of fracture. A more significant association between PAD and fracture was found in people having no other medical conditions which revealed that more medical conditions may dilute PAD’s contribution to fracture risk. This also may explain why the

association between PAD and fracture risk was not significant in patients with stroke, renal dialysis, and congestive heart failure, as these were considered significant contributing factors for fracture [19, 31, 32]. We suggest that health education should focus on patients with multiple comorbidities to improve disease understanding and motivate adoption of better lifestyles [19]. Although previous studies have shown some relationship between PAD and fracture [7–10, 16], the importance of PAD history on post-fracture adverse outcomes has not been addressed clearly. Our results suggest that patients with PAD have increased risks of fracture and of post-fracture urinary tract infection, pneumonia, septicemia, stroke, pulmonary embolism, acute renal failure, acute myocardial infarction, deep wound infection, post-fracture mortality, and more medical resource use. The association between PAD and fracture risk may be explained by the following reasons. First, PAD patients have reduced bone mass and subsequent osteoporosis. Ischemia or reduction of blood flow to lower extremities could result in greater demineralization or bone loss and negative bone equilibrium in the affected region [26, 33]. The strong correlation between poor bone mineral density and fracture suggests that losing bone mineral density may be one cause of PAD patients suffering from fracture. The second possible explanation is vitamin D deficiency. Vitamin D is essential for calcium absorption and is an important factor affecting bone mineral density. About 70% of patients with PAD have low levels of 25-hydroxyvitamin D, and 25-hydroxyvitamin D levels are

Osteoporos Int

inversely associated with the stages of PAD [34]. Symptoms of fatigue, muscle and bone pain, weakness, and difficulty walking are common symptoms in patients with PAD, suggesting immobilization in advanced PAD stages may restrict patients from exposure to sun and thus lower their levels of vitamin D [35]. The third possible explanation is falls. PAD patients with intermittent claudication are at higher risk of falling [5]. Sensory and motor nerve dysfunction of the lower extremities secondary to impaired peripheral circulation in PAD patients may be responsible for their greater falling prevalence [36]. Because falls are a frequent complication for PAD patients [37], it is important to prevent falls and fractures during their daily activities and specific health care protocols are needed to address this issue. The strengths of the present study include large sample size, cohort study design, multivariate adjustment, and analyzing all types of fracture. However, some limitations should be addressed. First, we used insurance claims data that lacks information on sociodemographics, lifestyle factors, hormonal status, vitamin D, bone mineral density, and other biomedical measures correlating with risks of fracture or PAD. The second limitation is possible underreporting from use of records from reimbursement claims for coexisting medical conditions. Patients with minor fractures may not seek treatment, so without these diagnoses we may have underestimated fracture risk. However, these data should be distributed equally between both groups and the bias should be minor. Third, though the accuracy of major diagnosis codes from the research database has been accepted by peer reviewers for prominent scientific journals worldwide [1, 2, 11, 12, 38], validity of PAD, fracture, and other comorbidity and complication codes might still be a limitation of this study. In addition, defining PAD patients by physician diagnoses may be not totally correct, because undiagnosed patients with mild or no symptoms may exist. Future research is needed that includes prospective design and lifestyle information as covariates in multivariate analyses. In conclusion, PAD is an important independent risk factor for fracture and post-fracture adverse events. This study provided comprehensive assessment of fracture risk and postfracture outcomes in patients with PAD. Further studies are needed to develop specific strategies to decrease fracture risks and post-fracture adverse outcomes for this fragile patient population.

Funding Information This study was supported in part by grants from Taiwan’s Ministry of Science and Technology (MOST105-2629-B-038001, MOST105-2221-E-038-014, MOST105-2314-B-038-025, MOST104-2314-B-038-027-MY2, NSC102-2314-B-038-021-MY3), Shuang Ho Hospital, Taipei Medical University (104TMU-SHH-23), and Taiwan’s Ministry of Health and Welfare Clinical Trial and Research Center of Excellence (MOHW106-TDUB-212-113004) Compliance with ethical standards Conflicts of interest Feng-Lin Liu, Chao-Shun Lin, Chun-Chieh Yeh, Chun-Chuan Shih, Yih-Giun Cherng, Chih-Hsing Wu, Ta-Liang Chen, and Chien-Chang Liao declare that they have no conflict of interest.

References 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13. Acknowledgements This study is based in part on data obtained from the National Health Insurance Research Database. This database is provided by the National Health Insurance Administration of Taiwan’s Ministry of Health and Welfare and is managed by the National Health Research Institutes. The authors’ interpretations and conclusions do not represent those of the aforementioned agencies.

14.

15.

Nowygrod R, Egorova N, Greco G et al (2006) Trends, complications, and mortality in peripheral vascular surgery. J Vasc Surg 43: 205–216 Hirsch AT, Hartman L, Town RJ, Virnig BA (2008) National health care costs of peripheral arterial disease in the Medicare population. Vasc Med 13:209–215 Krupski WC, Layug EL, Reilly LM, Rapp JH, Mangano DT (1992) Comparison of cardiac morbidity between aortic and infrainguinal operations. Study of Perioperative Ischemia (SPI) Research Group. J Vasc Surg 15:354–363 McDermott MM, Greenland P, Liu K et al (2001) Leg symptoms in peripheral arterial disease: associated clinical characteristics and functional impairment. JAMA 286:1599–1606 Gardner AW, Montgomery PS (2001) Impaired balance and higher prevalence of falls in subjects with intermittent claudication. J Gerontol A Biol Sci Med Sci 56:M454–M458 Mangiafico RA, Russo E, Riccobene S et al (2006) Increased prevalence of peripheral arterial disease in osteoporotic postmenopausal women. J Bone Miner Metab 24:125–131 Collins TC, Ewing SK, Diem SJ et al (2009) Peripheral arterial disease is associated with higher rates of hip bone loss and increased fracture risk in older men. Circulation 119:2305–2312 Hyde Z, Mylankal KJ, Hankey GJ, Flicker L, Norman PE (2013) Peripheral arterial disease increases the risk of subsequent hip fracture in older men: the Health in Men Study. Osteoporos Int 24: 1683–1688 Von Mühlen D, Allison M, Jassal SK, Barrett-Connor E (2009) Peripheral arterial disease and osteoporosis in older adults: the Rancho Bernardo Study. Osteoporos Int 20:2071–2078 Peláez VC, Ausín L, Ruiz Mambrilla M, Gonzalez-Sagrado M, Pérez Castrillón JL (2015) Ankle-brachial index, risk of clinical fractures, mortality and low bone mass in nursing home residents. Eur Rev Med Pharmacol Sci 19:1577–1582 Liao CC, Lin CS, Shih CC et al (2014) Increased risk of fracture and postfracture adverse events in patients with diabetes: two nationwide population-based retrospective cohort studies. Diabetes Care 37:2246–2252 Huang YF, Cherng YG, Hsu SP et al (2015) Risk and adverse outcomes of fractures in patients with Parkinson's disease: two nationwide studies. Osteoporos Int 26:1723–1732 Court-Brown CM, Caesar B (2006) Epidemiology of adult fractures: a review. Injury 37:691–697 Stehman-Breen CO, Sherrard DJ, Alem AM et al (2000) Risk factors for hip fracture among patients with end-stage renal disease. Kidney Int 58:2200–2205 Guilley E, Herrmann F, Rapin CH, Hoffmeyer P, Rizzoli R, Chevalley T (2011) Socioeconomic and living conditions are

Osteoporos Int determinants of hip fracture incidence and age occurrence among community-dwelling elderly. Osteoporos Int 22:647–653 16. Sennerby U, Melhus H, Gedeborg R et al (2009) Cardiovascular diseases and risk of hip fracture. JAMA 302:1666–1673 17. Ensrud KE (2013) Epidemiology of fracture risk with advancing age. J Gerontol A Biol Sci Med Sci 68:1236–1242 18. Black DM, Rosen CJ (2016) Clinical practice: postmenopausal osteoporosis. N Engl J Med 374:254–262 19. Gregson CL, Dennison EM, Compston JE et al (2014) Diseasespecific perception of fracture risk and incident fracture rates: GLOW cohort study. Osteoporos Int 25:85–95 20. Hsu CC, Hsu YC, Chang KH, Lee CY, Chong LW, Wang YC, Hsu CY, Kao CH (2016) Increased risk of fracture in patients with bipolar disorder: a nationwide cohort study. Soc Psychiatry Psychiatr Epidemiol 51:1331–1338 21. Bang UC, Benfield T, Bendtsen F, Hyldstrup L, Beck Jensen JE (2014) The risk of fractures among patients with cirrhosis or chronic pancreatitis. Clin Gastroenterol Hepatol 12:320–326 22. Nickolas TL, Leonard MB, Shane E (2008) Chronic kidney disease and bone fracture: a growing concern. Kidney Int 74:721–731 23. Ouriel K (2001) Peripheral arterial disease. Lancet 358:1257–1264 24. Fowkes FG, Rudan D, Rudan I et al (2013) Comparison of global estimates of prevalence and risk factors for peripheral artery disease in 2000 and 2010: a systematic review and analysis. Lancet 382: 1329–1340 25. Vogt MT, Cauley JA, Kuller LH, Nevitt MC (1997) Bone mineral density and blood flow to the lower extremities: the study of osteoporotic fractures. J Bone Miner Res 12:283–289 26. Szulc P, Blackwell T, Schousboe JT et al (2014) High hip fracture risk in men with severe aortic calcification: MrOS study. J Bone Miner Res 29:968–975 27. Bagger YZ, Tankó LB, Alexandersen P, Qin G, Christiansen C (2006) Radiographic measure of aorta calcification is a sitespecific predictor of bone loss and fracture risk at the hip. J Intern Med 259:598–605 28. Sathiyakumar V, Avilucea FR, Whiting PS et al (2016) Risk factors for adverse cardiac events in hip fracture patients: an analysis of NSQIP data. Int Orthop 40:439–445 29. Belmont PJ Jr, Davey S, Rensing N, Bader JO, Waterman BR, Orr JD (2015) Patient-based and surgical risk factors for 30-day

postoperative complications and mortality after ankle fracture fixation. J Orthop Trauma 29:e476–e482 30. Rossi E, Biasucci LM, Citterio F et al (2002) Risk of myocardial infarction and angina in patients with severe peripheral vascular disease: predictive role of C-reactive protein. Circulation 105: 800–803 31. Mittalhenkle A, Gillen DL, Stehman-Breen CO (2004) Increased risk of mortality associated with hip fracture in the dialysis population. Am J Kidney Dis 44:672–679 32. Pouwels S, Lalmohamed A, Leufkens B et al (2009) Risk of hip/ femur fracture after stroke: a population-based case-control study. Stroke 40:3281–3285 33. Laroche M, Pouilles JM, Ribot C et al (1994) Comparison of the bone mineral content of the lower limbs in men with ischaemic atherosclerotic disease. Clin Rheumatol 13:611–614 34. Fahrleitner-Pammer A, Obernosterer A, Pilger E et al (2005) Hypovitaminosis D, impaired bone turnover and low bone mass are common in patients with peripheral arterial disease. Osteoporos Int 16:319–324 35. Fahrleitner A, Dobnig H, Obernosterer A et al (2002) Vitamin D deficiency and secondary hyperparathyroidism are common complications in patients with peripheral arterial disease. J Gen Intern Med 17:663–669 36. Laghi Pasini F, Pastorelli M, Beermann U et al (1996) Peripheral neuropathy associated with ischemic vascular disease of the lower limbs. Angiology 47:569–577 37. Gardner AW, Montgomery PS (2001) The relationship between history of falling and physical function in subjects with peripheral arterial disease. Vasc Med 6:223–227 38. Ghanami RJ, Hurie J, Andrews JS et al (2013) Anesthesia-based evaluation of outcomes of lower-extremity vascular bypass procedures. Ann Vasc Surg 27:199–207 39. Gracious BL, Fontanella CA, Phillips GS, Bridge JA, Marcus SC, Campo JV (2016) Antidepressant exposure and risk of fracture among Medicaid-covered youth. J Clin Psychiatry 77:e950–e956 40. Vestergaard P, Rejnmark L, Mosekilde L (2006) Anxiolytics, sedatives, antidepressants, neuroleptics and the risk of fracture. Osteoporos Int 17:807–816