Advertisement

Osteoporosis International

, Volume 30, Issue 3, pp 583–591 | Cite as

Preoperative antithrombotic therapy and risk of blood transfusion and mortality following hip fracture surgery: a Danish nationwide cohort study

  • C. DaugaardEmail author
  • A.B. Pedersen
  • N.R. Kristensen
  • S.P. Johnsen
Original Article

Abstract

Summary

Hip fracture surgery is associated with high risk of bleeding and mortality. The patients often have cardiovascular comorbidity, which requires antithrombotic treatment. This study found that preoperative use of oral anticoagulants was not associated with transfusion or mortality following hip fracture surgery, whereas increased risk may exist for antiplatelet drugs.

Introduction

Hip fracture surgery is associated with high bleeding risk and mortality; however, data on operative outcomes of hip fracture patients admitted while on antithrombotic therapy is sparse. We examined if preoperative antithrombotic treatment was associated with increased use of blood transfusion and 30-day mortality following hip fracture surgery.

Methods

Using data from the Danish Multidisciplinary Hip Fracture Registry, we identified 74,791 hip fracture surgery patients aged ≥ 65 years during 2005–2016. Exposure was treatment with non-vitamin K antagonist oral anticoagulant (NOAC), vitamin K antagonists (VKA), or antiplatelet drugs at admission for hip fracture. Outcome was blood transfusion within 7 days postsurgery and death within 30 days.

Results

A 45.3% of patients received blood transfusion and 10.6% died. Current NOAC use was associated with slightly increased risk of transfusion (adjusted relative risk (aRR) 1.07, 95% confidence interval (CI) 1.01–1.14), but similar mortality risk (adjusted hazard ratio (aHR) 0.88, 95% CI 0.75–1.03) compared with non-users. The pattern remained when restricting to patients with short surgical delay (< 24 h). VKA users did not have increased risk of transfusion or mortality. The risks of transfusion (aRR 1.15 95% CI 1.12–1.18) and 30-day mortality (aHR 1.18 95% CI 1.14–1.23) were increased among antiplatelet users compared with non-users.

Conclusions

In an observational setting, neither preoperative NOAC nor VKA treatments were associated with increased risk of 30-day postoperative mortality among hip fracture patients. NOAC was associated with slightly increased risk of transfusion. Preoperative use of antiplatelet drugs was associated with increased risk of transfusion and mortality.

Keywords

Anticoagulants Bleeding Cohort study Hip fractures Platelet inhibitor 

Introduction

Hip fracture is a common trauma among elderly people and a frequent cause of hospitalization and death [1]. Hip fracture patients often have cardiovascular comorbidity which requires treatment with antithrombotic drugs [2, 3]. The use of antithrombotic therapy has increased, and currently, more than 40% of patients admitted with hip fracture are on antithrombotic therapy [4]. This may have clinical implications for surgery and subsequent care. Ongoing antithrombotic treatment at the time of hip fracture represents a difficult clinical dilemma. The question is whether surgery should be delayed for drug effects to diminish or if it is preferable to proceed with immediate surgery to facilitate early mobilization and hereby reduce the risk of postoperative complications. On one hand, hip fracture surgery remains associated with a substantial risk of adverse outcomes, including perioperative bleeding and an overall 30-day mortality risk of approximately 10% [5]. Performing acute major surgery in a fragile group of patients with ongoing antithrombotic therapy may therefore not be an attractive approach. However, delaying surgery may is also be problematic. It is generally shown that delayed surgery is associated with an increase in mortality [6, 7, 8, 9], but this excess mortality may be rooted in medical conditions causing the delay rather than the actual delay itself [10].

Data on perioperative outcomes of hip fracture patients admitted while on antithrombotic therapy including the newer non-vitamin K antagonist oral anticoagulants (NOACs) is sparse. Accordingly, we aimed to examine the association between pre-admission prescription of NOACs, vitamin K antagonists (VKA), and antiplatelets in hip fracture patients with the rate of red blood cell (RBC) transfusion within 7 days and mortality within 30 days. We hypothesized that antithrombotic treatment at time of admission would be associated with a higher risk of blood transfusion and death among hip fracture patients.

Methods

Setting

We performed a nationwide population-based cohort study including all patients who underwent hip fracture surgery during 2005–2016. All Danish residents are assigned a unique ten-digit personal identification number at birth or immigration, enabling unambiguous individual-level linkage between Danish registers [11].

Study population

We used the Danish Multidisciplinary Hip Fracture Registry (DMHFR) to identify all first-time hip fracture patients aged 65 or older treated surgically between 1 January 2005 and 31 December 2016. The DMHFR is a nationwide clinical quality database containing information on all patients with either a primary or secondary inpatient diagnosis of femoral neck, pertrochanteric, or subtrochanteric fractures undergoing surgery including insertion of a primary hip replacement or open reduction and internal fixation in Denmark since 2003. Reporting to the database is mandatory for all Danish orthopedic departments [12, 13] and includes detailed pre-, per-, and postoperative data. No patients were excluded, and the total cohort included 71,791 patients.

Exposure

Medicine data was obtained from The Danish National Health Service Prescription Database which holds individual-level data on all prescription drugs sold by all community and hospital-based outpatient pharmacies in Denmark since 2004 [14]. Data on the following antithrombotic drugs were obtained: NOAC (Anatomical Therapeutic Chemical (ATC) code: B01AE07, B01AF01, B01AF02, B01AF03), VKA (ATC code: B01AA), and antiplatelets including acetylsalicylic acid (ATC code: B01AC, N02BA01, N02BA51). Patients were categorized into non-users (no redemption of prescribed antithrombotics in the year prior to surgery), former users (redemption of one prescription 91–365 days prior to hip fracture surgery) and current users (at least one prescription ≤ 90 days prior to hip fracture surgery) for each class of antithrombotic drugs.

Outcome

The outcomes were RBC transfusion within 7 days after hip fracture surgery and all-cause 30-day mortality. Information on all allogenic RBC transfusion was obtained from the Danish Transfusion Database (DTDB). Patients were classified as having received either none or one or more RBC units. The 7-day window was chosen to capture transfusions related to surgery rather than other factors such as comorbidity. The DTDB is a nationwide clinical quality database collecting data on the use of blood products in Denmark since 2000. It contains individual level information on types and numbers of blood components administered, date of blood transfusion, and clinical biochemical data.

Information on all-cause mortality (30-days postsurgery) was obtained from the Danish Civil Registration System. This national registry has maintained records on vital status, migration, and residence for the entire Danish population since 1968 [11].

Covariates

From DMHFR, we obtained information on a range of pre-specified covariates: sex, age, body mass index (BMI), type of hip fracture (femoral neck or pertrochanteric and subtrochanteric fractures), type of surgery (osteosynthesis, or total- and hemi-hip arthroplasty), surgery delay (< 24 h, 25–36, and > 36 h), and surgery year. Following BMI (weight in kilograms (kg) divided by the square of height in meters (m)) categories were applied: underweight (BMI was < 18.5 kg/m2), normal weight (BMI was 18.5–24.9 kg/m2), overweight (BMI was 25–29.9 kg/m2), and obese (BMI was ≥ 30 kg/m2).

Surgery delay was defined as the time period from hip fracture admission to surgical procedure. For each patient, we extracted information on diagnosis in relation to hospitalization and outpatient visits through a period of 10 years prior to hip fracture and calculated a Charlson comorbidity index (CCI) score for each patient using all ICD-10 discharge diagnoses available from The Danish National Patient Registry [15, 16]. In addition, we extracted information on diagnosis of atrial fibrillation or flutter (ICD-10 diagnosis code I48). From the Danish National Health Service Prescription Database, we collected information on all redeemed prescriptions for non-steroidal anti-inflammatory drugs (NSAIDs), antidepressants (selective serotonin reuptake inhibitors (SSRIs)), corticosteroids, statins, and antibiotics in the year prior to hip fracture surgery. ATC-codes are provided in Online Resource 1.

Statistics

We computed the cumulative incidence with 95% confidence intervals (CIs) of RBC transfusion within 7 days of surgery according to antithrombotic use considering death as competing risk. In the complete case analyses, we excluded patients with missing data on BMI and surgery delay. Then, using a log-binomial model, we estimated relative risks (RRs) for RBC transfusion and corresponding 95% CIs, both crude and adjusted for age, sex, BMI, CCI, type of fracture, surgery delay, year of surgery, and other prescription medicine using non-user patients as reference. We did not adjust for type of surgery, as this variable was considered to be colinear with type of fracture. Since hospital-related factors including the surgeon behavior and the decision-making process may affect the need for transfusion, we adjusted for a cluster effect by departments.

The risk of death within 30 days was analyzed using a Cox proportional hazard regression model. We estimated hazard ratios (HRs) and corresponding 95% CI for 30-day mortality, both crude and adjusted for the same covariates as in log-binominal model, comparing current and former users of NOAC with non-users of NOAC. We did not adjust for transfusion as this variable was more likely to be an intermediate step rather than a confounder in the association between antithrombotic therapy and postoperative mortality. Similarly, HRs were calculated for users of VKA and antiplatelets. The assumption of proportional hazards was evaluated by visual inspection of plots.

In addition, all analyses were stratified by surgical delay to investigate whether the risk associated with antithrombotic therapy differed among patients according to time to surgery. Also, we made analyses restricted to patients diagnosed with atrial fibrillation or atrial flutter.

We assumed the data on BMI and surgery delay to be missing at random. Thus, we repeated all analyses using multiple imputation method. This method creates several plausible imputed datasets and thereafter combines the results based on information from available data [17]. In order to impute the data on BMI and surgery delay, we used all available information from the patients presented in Table 1 and Online Resource 2 (age, sex, BMI, CCI, type of fracture, type of surgery, surgery delay, year of surgery, other prescription medicine, and outcome data) and generated five imputed datasets. The presented adjusted RRs and HRs were calculated as the geometric mean of the RRs or HRs of the five datasets, with the corresponding CIs corrected for between- and within-imputation variation [18].
Table 1

Patient characteristics and surgery-related factors of all hip fracture patients according to antithrombotic treatment

 

All patients

Users of NOACsa

Users of VKAsb

Users of antiplatelets

 

Non-users

Former

Current

Non-users

Former

Current

Non-users

Former

Current

N (%)

N (%)

N (%)

N (%)

N (%)

N (%)

N (%)

N (%)

N (%)

N (%)

Total

74,791

(100.0)

73,507

(98.3)

221

(0.3)

1063

(1.4)

68,948

(92.2)

1681

(2.2)

4162

(5.6)

43,245

(57.8)

6979

(9.3)

24,567

(32.8)

Age

 65–74

14,411

(19.3)

14,221

(19.3)

38

(17.2)

152

(14.3)

13,533

(19.6)

231

(13.7)

647

(15.5)

9575

(22.1)

1158

(16.6)

3678

(15.0)

 75–84

28,750

(38.4)

28,266

(38.5)

94

(42.5)

390

(36.7)

26,157

(37.9)

730

(43.4)

1863

(44.8)

16,455

(38.1)

2752

(39.4)

9543

(38.8)

 85+

31,630

(42.3)

31,020

(42.2)

89

(40.3)

521

(49.0)

29,258

(42.4)

720

(42.8)

1652

(39.7)

17,215

(39.8)

3069

(44.0)

11,346

(46.2)

Sex

 Female

53,359

(71.3)

52,522

(71.5)

142

(64.3)

695

(65.4)

49,784

(72.2)

1062

(63.2)

2513

(60.4)

31,815

(73.6)

4735

(67.8)

16,809

(68.4)

 Male

21,432

(28.7)

20,985

(28.5)

79

(35.7)

368

(34.6)

19,164

(27.8)

619

(36.8)

1649

(39.6)

11,430

(26.4)

2244

(32.2)

7758

(31.6)

CCI scorec

 0

30,111

(40.3)

29,846

(40.6)

44

(19.9)

221

(20.8)

28,775

(41.7)

339

(20.2)

997

(24.0)

22,189

(51.3)

1958

(28.1)

5964

(24.3)

 1–2

30,251

(40.4)

29,638

(40.3)

105

(47.5)

508

(47.8)

27,619

(40.1)

746

(44.4)

1886

(45.3)

15,218

(35.2)

3045

(43.6)

11,988

(48.8)

 3 ≤

1442

(19.3)

14,023

(19.1)

72

(32.6)

334

(31.4)

12,554

(18.2)

596

(35.5)

1279

(30.7)

5838

(13.5)

1976

(28.3)

6615

(26.9)

BMId

 No data

15,552

(20.8)

15,315

(20.8)

45

(20.4)

192

(18.1)

14,440

(20.9)

347

(20.6)

765

(18.4)

8863

(20.5)

1417

(20.3)

5272

(21.5)

 Underweight

6455

(8.6)

6360

(8.7)

12

(5.4)

83

(7.8)

6113

(8.9)

123

(7.3)

219

(5.3)

3984

(9.2)

589

(8.4)

1882

(7.7)

 Normal

33,987

(45.4)

33,408

(45.4)

95

(43.0)

484

(45.5)

31,484

(45.7)

739

(44.0)

1764

(42.4)

19,965

(46.2)

3156

(45.2)

10,866

(44.2)

 Overweight

14,691

(19.6)

14,409

(19.6)

50

(22.6)

232

(21.8)

13,300

(19.3)

358

(21.3)

1033

(24.8)

8178

(18.9)

1403

(20.1)

5110

(20.8)

 Obese

4106

(5.5)

4015

(5.5)

19

(8.6)

72

(6.8)

3611

(5.2)

114

(6.8)

381

(9.2)

2255

(5.2)

414

(5.9)

1437

(5.8)

Type of fracture

 Femoral neck

39,605

(53.0)

38,935

(53.0)

111

(50.2)

559

(52.6)

36,485

(52.9)

916

(54.5)

2204

(53.0)

23,062

(53.3)

3628

(52.0)

12,915

(52.6)

 Per- or sub-trochanteric

35,186

(47.0)

34,572

(47.0)

110

(49.8)

504

(47.4)

32,463

47.1)

765

(45.5)

1958

(47.0)

20,183

(46.7)

3351

(48.0)

11,652

(47.4)

Surgery delay (hours)

 No data

171

(0.2)

171

(0.2)

0

(0.0)

0

(0.0)

154

(0.2)

5

(0.3)

12

(0.3)

106

(0.2)

19

(0.3)

46

(0.2)

 < 24

45,027

(60.2)

44,438

(60.5)

120

(54.3)

469

(44.1)

42,694

(61.9)

763

(45.4)

1570

(37.7)

26,132

(60.4)

4065

(58.2)

14,830

(60.4)

 24–36

13,654

(18.3)

13,430

(18.3)

38

(17.2)

186

(17.5)

12,453

(18.1)

315

(18.7)

886

(21.3)

7788

(18.0)

1282

(18.4)

4584

(18.7)

 > 36

15,939

(21.3)

15,468

(21.0)

63

(28.5)

408

(38.4)

13,647

(19.8)

598

(35.6)

1694

(40.7)

9219

(21.3)

1613

(23.1)

5107

(20.8)

Underweight if BMI < 18.5 kg/m2, normal weight if BMI 18.5–24.9 kg/m2, overweight if BMI 25–29.9 kg/m2, and obese if BMI ≥ 30 kg/m2

aNOACs non-vitamin K antagonist oral anticoagulants

bVKAs vitamin K antagonists

cCCI score Charlson comorbidity index score

dBMI body mass index

All statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA). The study was approved by the Danish Data Protection Agency (Record no. 1-16-02-467-15). According to Danish law, registry-based studies do not require ethical approval.

Results

Patient characteristics

Among the 74,791 first time hip fracture patients, 29,792 patients (39.8%) were on antithrombotic therapy at admission; 1.4% were current users of NOAC, 5.6% were current users of VKA, and 32.9% were current users of antiplatelet drugs. Users of any antithrombotics were more likely to be male, to receive statin treatment, and had more comorbid conditions (Table 1 and Online Resource 2). Compared to non-users, current users of anticoagulant drugs (NOAC and VKA) experienced more often surgery delay, while current users of antiplatelets did not experience excess time to surgery. Apart from BMI (20.8% missing) and surgery delay (0.2% missing), data were complete for all variables.

Risk of RBC transfusion

Among all hip fracture patients, 45.3% received RBC transfusion within 7 days postsurgery (Table 2).
Table 2

Cumulative incidence and relative risk (RR) with 95% confidence interval (95% CI) for red blood cell transfusion within 7 days after surgery according to antithrombotic therapy

 

Transfusion (%)

Complete case method

Multiple imputation methoda

Crude RR (95% CI)

Adjustedb RR (95% CI)

Adjustedb RR (95% CI)

NOACc

 Non-users

45.4 (45.05–45.7)

Reference

Reference

Reference

 Former users

40.7 (34.2–47.1)

0.90 (0.79–1.02)

1.00 (0.85–1.18)

0.99 (0.86–1.15)

 Current users

43.0 (40.0–45.9)

0.95 (0.87–1.03)

1.07 (1.00–1.14)

1.07 (1.01–1.14)

Vitamin K antagonists

 Non-users

45.4 (45.0–45.7)

Reference

Reference

Reference

 Former users

45.7 (43.3–48.1)

1.01 (0.95–1.07)

0.95 (0.90–1.01)

0.97 (0.91–1.03)

 Current users

44.5 (43.0–46.0)

0.98 (0.94–1.03)

0.99 (0.95–1.04)

1.00 (0.95–1.05)

Antiplatelets

 Non-users

41.0 (40.6–41.5)

Reference

Reference

Reference

 Former users

49.1 (48.0–50.3)

1.20 (1.16–1.23)

1.13 (1.10–1.17)

1.13 (1.10–1.16)

 Current users

51.8 (51.2–52.4)

1.26 (1.23–1.30)

1.16 (1.12–1.20)

1.15 (1.12–1.18)

aTo impute the data, we used all available information from the patients presented in Table 1 and Online Resource 2

bAdjustments were made for age, sex, body mass index, Charlson comorbidity index score, type of fracture, surgery delay, year of surgery, and other prescription medication (non-steroidal anti-inflammatory drugs (NSAIDs), selective serotonin reuptake inhibitors (SSRIs), corticosteroids, statins and antibiotics)

cNOAC non-vitamin K antagonist oral anticoagulant

The adjusted RR of RBC transfusion was 1.07 (95% CI 1.01–1.14) for current users compared to non-users of NOAC. Among patients undergoing surgery within 24 h after hip fracture, the adjusted RR of RBC transfusion was 1.14 (95% CI 1.02–1.27) for current NOAC users compared with non-users (Table 3).
Table 3

Adjusted relative risk (RR) with 95% confidence interval (95% CI) for red blood cell transfusion within 7 days after surgery according to antithrombotic therapy. Results stratified by surgery delay

 

Surgery delay

< 24 h

24–36 h

> 36 h

NOACa

 Non-users

Reference

Reference

Reference

 Former users

0.98 (0.78–1.24)

1.36 (0.93–1.97)

0.72 (0.38–1.37)

 Current users

1.14 (1.02–1.27)

1.03 (0.92–1.16)

0.96 (0.87–1.06)

Vitamin K antagonists

 Non-users

Reference

Reference

Reference

 Former users

0.96 (0.87–1.05)

1.00 (0.88–1.13)

0.92 (0.85–1.01)

 Current users

1.00 (0.94–1.07)

0.96 (0.89–1.04)

0.97 (0.92–1.04)

Antiplatelets

 Non-users

Reference

Reference

Reference

 Former users

1.12 (1.08–1.16)

1.10 (1.03–1.18)

1.13 (1.09–1.18)

 Current users

1.16 (1.12–1.20)

1.15 (1.10–1.20)

1.08 (1.04–1.11)

Adjustments were made for age, sex, body mass index, Charlson comorbidity index score, type of fracture, surgery delay, year of surgery and other prescription medication (non-steroidal anti-inflammatory drugs (NSAIDs), selective serotonin reuptake inhibitors (SSRIs), corticosteroids, statins and antibiotics)

Results shown for multiple imputation model. To impute the data, we used all available information from the patients presented in Table 1 and Online Resource 2

aNOAC non-vitamin K antagonist oral anticoagulant

Treatment with VKA was not associated with increased risk of RBC transfusion, since the adjusted RR among current compared to non-VKA users was 1.00 (95% CI 0.95–1.05). For patients who underwent surgery within 24 h after hip fracture, the adjusted RR of RBC transfusion was also 1.00 (95% CI 0.94–1.07).

The analysis restricted to hip fracture patients with atrial fibrillation or flutter confirmed the findings of no increased risk of transfusion among current users of NOAC and VKA (Table 6).

The adjusted RR for RBC transfusion was 1.15 (95% CI 1.12–1.18) for antiplatelet users and 1.13 (95% CI 1.10–1.16) for former antiplatelet users compared to non-users. Among patients who experienced > 36 h surgery-delay, the adjusted RR among users of antiplatelets compared to non-users was 1.08 (95% CI 1.04–1.11).

Mortality risk

At 30 days after hip fracture surgery, 10.6% of all patients had died. Mortality within 30 days of surgery was 11.3% for patients dispensed NOAC, 10.8% for those dispensed VKA, and 12.7% for those dispensed antiplatelets.

The adjusted HR of death within 30 days was 0.88 (95% CI 0.75–1.03) for current users of NOAC compared to non-users (Table 4). For patients who experienced > 36 h surgery delay, the adjusted HR for 30-day mortality was 0.70 (95% CI 0.54–0.91) for NOAC users compared to non-users (Table 5). Among patients with atrial fibrillation or atrial flutter, the adjusted HR for 30-day mortality was 0.59 (95% CI 0.48–0.73) for NOAC users compared with non-users (Table 6).
Table 4

Hazard ratio (HR) with 95% confidence interval (95% CI) for 30-day mortality according to antithrombotic therapy

 

Death (%)

Complete case method

Multiple imputation methoda

Crude HR (95% CI)

Adjustedb HR (95% CI)

Adjustedb HR (95% CI)

NOACc

 Non-users

10.5

Reference

Reference

Reference

 Former users

10.9

1.03 (0.80–1.33)

0.84 (0.58–1.21)

0.88 (0.69–1.13)

 Current users

11.3

1.07 (0.90–1.29)

0.90 (0.73–1.11)

0.88 (0.75–1.03)

Vitamin K antagonists

 Non-users

10.5

Reference

Reference

Reference

 Former users

14.1

1.38 (1.22–1.57)

1.08 (0.91–1.27)

1.10 (0.96–1.24)

 Current users

10.8

1.04 (0.93–1.15)

0.96 (0.83–1.10)

0.91 (0.82–1.01)

Antiplatelets

 Non-users

9.2

Reference

Reference

Reference

 Former users

11.8

1.31 (1.20–1.43)

1.10 (0.99–1.21)

1.09 (1.00–1.17)

 Current users

12.7

1.42 (1.36–1.47)

1.21 (1.15–1.28)

1.18 (1.14–1.23)

aTo impute the data, we used all available information from the patients presented in Table 1 and Online Resource 2

bAdjustments were made for age, sex, body mass index, Charlson comorbidity index score, type of fracture, surgery delay, year of surgery, and other prescription medication (non-steroidal anti-inflammatory drugs (NSAIDs), selective serotonin reuptake inhibitors (SSRIs), corticosteroids, statins and antibiotics)

cNOAC non-vitamin K antagonist oral anticoagulant

Table 5

Adjusted hazard ratio (HR) with 95% confidence interval (95% CI) for 30-day mortality according to antithrombotic therapy. Results stratified by surgery delay

 

Surgery delay

< 24 h

24–36 h

> 36 h

NOACa

 Non-users

Reference

Reference

Reference

 Former users

1.23 (0.89–1.69)

0.21 (0.03–1.61)

0.79 (0.36–1.73)

 Current users

1.08 (0.83–1.39)

0.95 (0.69–1.30)

0.70 (0.54–0.91)

Vitamin K antagonists

 Non-users

Reference

Reference

Reference

 Former users

1.12 (0.93–1.34)

1.29 (0.92–1.80)

1.14 (0.92–1.42)

 Current users

1.04 (0.91–1.18)

1.10 (0.90–1.34)

0.87 (0.74–1.03)

Antiplatelets

 Non-users

Reference

Reference

Reference

 Former users

1.11 (0.99–1.23)

1.18 (1.05–1.34)

1.07 (0.91–1.27)

 Current users

1.20 (1.13–1.27)

1.36 (1.25–1.49)

1.11 (0.98–1.25)

Adjustments were made for age, sex, body mass index, Charlson comorbidity index score, type of fracture, surgery delay, year of surgery, and other prescription medication (non-steroidal anti-inflammatory drugs (NSAIDs), selective serotonin reuptake inhibitors (SSRIs), corticosteroids, statins and antibiotics)

Results shown for multiple imputation model. To impute the data, we used all available information from the patients presented in Table 1 and Online Resource 2

aNOAC non-vitamin K antagonist oral anticoagulant

Table 6

Results restricted to hip fracture patients with atrial fibrillation or flutter. Adjusted relative risk for red blood cell transfusion within 7 days after surgery and adjusted hazard ratio for 30-day mortality according to antithrombotic therapy

 

Adjusted relative risk (RR) with 95% confidence interval (95% CI) for red blood cell transfusion within7 days after surgery

Adjusted hazard ratio (HR) with 95% confidence interval (95% CI) for 30-day mortality

NOACa

 Non-users

Reference

Reference

 Former users

0.88 (0.75–1.03)

0.59 (0.43–0.82)

 Current users

0.97 (0.90–1.04)

0.59 (0.48–0.73)

Vitamin K antagonists

 Non-users

Reference

Reference

 Former users

0.93 (0.87–1.00)

0.82 (0.69–0.98)

 Current users

0.92 (0.87–0.97)

0.65 (0.56–0.76)

Antiplatelets

 Non-users

Reference

Reference

 Former users

1.08 (1.03–1.15)

0.95 (0.81–1.13)

 Current users

1.16 (1.12–1.20)

1.02 (0.94–1.10)

Adjustments were made for age, sex, body mass index, Charlson comorbidity index score, type of fracture, surgery delay, year of surgery, and other prescription medication (non-steroidal anti-inflammatory drugs (NSAIDs), selective serotonin reuptake inhibitors (SSRIs), corticosteroids, statins and antibiotics). Results shown for multiple imputation model. To impute the data, we used all available information from the patients presented in Table 1 and Online Resource 2

aNOAC non-vitamin K antagonist oral anticoagulant

Users of VKA had an adjusted HR of 30-day mortality of 0.91 (95% CI 0.82–1.01) compared to non-users. When patients underwent surgery within 24 h after hip fracture, current users of VKA did not have an excess 30-day mortality, as the adjusted HR was 1.04 (95% CI 0.91–1.18) for VKA users compared to non-users. For patients with atrial fibrillation or atrial flutter, the adjusted HR for 30-days mortality was 0.65 (95% CI 0.56–0.76) for VKA users compared with non-users.

Antiplatelet users had increased risk of 30-day mortality, with an adjusted HR of 1.18 (95% CI 1.14–1.23) compared to non-users. The increased mortality was consistent when stratifying for time to surgery. For antiplatelet users compared to non-users, the adjusted HR was 1.02 (95% CI 0.94–1.10) among patients with atrial fibrillation or atrial flutter.

Discussion

This study found that current NOAC users had a slightly increased risk of RBC transfusion, whereas current VKA users had similar risk of RBC transfusion as non-users. We found no evidence of increased 30-day mortality among hip fracture patients who are current NOAC or VKA users compared to non-users. However, among hip fracture patients admitted on antiplatelets, we found a 15% relative increase in RBC transfusion risk and 18% relative increase in mortality compared with non-users. An early surgery within 24 h was associated with an increased risk of transfusion but had no association with mortality among NOAC users.

To our knowledge, this is the first nationwide study on the impact of long-term preoperative antithrombotic treatment including NOAC on bleeding measured with transfusion and mortality among hip fracture patients. However, in accordance with our findings among patients with atrial-fibrillation or flutter, a recent population-based study by Ge et al. reported no association between use of oral anticoagulants and bleeding or 30-day mortality among patients with atrial fibrillation undergoing urgent major non-cardiac surgery including hip fracture surgery [19]. Small-scale studies support our findings of no difference in transfusion or 30-day mortality among hip fracture patients in general who were users of warfarin compared to non-users [20, 21]. In contrast, a recent register-based cohort study conducted in the UK including 2036 hip fracture patients found an increase in 1-year mortality among patients admitted on warfarin compared to non-users [22]. In our study, antiplatelet users consisted of patients treated with either platelet aggregation inhibitors or acetylsalicylic acid. A recent meta-analysis also showed preoperative clopidogrel treatment to be associated with a small increase in transfusion risk among hip fracture patients [23]. In additional agreement with our findings, small-scale studies indicate an increase in blood transfusion risk and 1-year mortality among hip fracture patients treated with acetylsalicylic acid [24, 25, 26].

When undergoing early surgical intervention to enable early mobilization and reduce mortality, hip fracture patients on antithrombotic therapy may experience prolonged bleeding time secondary to the drug pharmacokinetics. However, the increased risk of bleeding must be balanced by the potential benefits of early surgery. Guidelines recommend that antithrombotics are discontinued prior to invasive procedures; however, recommendation becomes difficult when patients undergo non-planned urgent hip fracture surgery. In this study, we found that twice as many current users of NOAC or VKA experienced surgery delay > 36 h compared to non-users. However, despite early surgery, the RBC transfusion and 30-day mortality risks were similar between anticoagulated (VKA and NOAC) and non-anticoagulated patients. This suggests that even early hip fracture surgery may be safe among anticoagulated patients. Surprisingly, we also found that patients on NOAC had a lower mortality risk compared to non-users of NOAC when surgery was delayed > 36 h. The lower mortality may potentially be due to enhanced attention and more intensive perioperative observation of NOAC patients awaiting surgery.

Our results also suggest that even when hip fracture surgery was delayed more than 36 h, and drug effects may have diminished, both the risk of RBC transfusion and 30-day mortality remained higher for patients on antiplatelet drugs. Thus, delaying surgery among hip fracture patients on antiplatelet therapy did not appear to be associated with better perioperative outcomes.

Antithrombotic therapy is commonly used to prevent cardiovascular events in patients with atrial fibrillation and flutter. Thus, the finding of excess 30-day mortality among patients on antiplatelet therapy may be caused by confounding by indication. Accordingly, Ge et al. suggested atrial fibrillation to be an independent predictor of perioperative mortality in urgent non-cardiac surgery which was not mediated by oral anticoagulant therapy [19]. In our analysis restricted to patients with atrial fibrillation or atrial flutter, we found no association between antiplatelet drug use and 30-day mortality risk. Hence, underlying atrial fibrillation or atrial flutter appeared to be a potential effect modifier in the association between use of antiplatelet drugs and postoperative outcome, either directly or as a proxy marker of other cardiovascular illnesses. In addition, the excess mortality among patients on antiplatelets drug may be mediated by transfusion, as antiplatelet users had increased risk of RBC transfusion. Studies indicate that transfusion may have fatal outcome [27, 28, 29]; however, a recent Cochrane meta-analysis showed the same mortality risk when hip fracture patients were treated with restrictive or liberal transfusion policy [30]. As our mortality outcome was 30-day mortality by all causes, this study cannot verify or exclude the possibility of complications to transfusion as contributors to mortality. In conclusion, further research should focus on the mechanisms linking antiplatelets with an adverse postoperative outcome in hip fracture patients.

Strengths and limitations

This study was based on unselected prospectively collected data from nationwide DMHFR with high validity [12]. Data on blood transfusions was directly drawn from the blood bank systems in which registration of all blood products is mandatory according to Danish law [31]. The Danish National Patient Registry provides full coverage of hospitals and ensures complete follow-up [32].

However, limitations of this study should be noted. As antithrombotic drug exposure was defined by drug redemption at general pharmacies, we were unable to ascertain drug indication, compliance, and treatment adherence. Additionally, we did not capture medication administered during hospitalization including bridging therapy, perioperative pharmacological venous thromboembolism prophylaxis, and reversal agents. We could not ascertain the date of antithrombotic discontinuation prior to surgery. In addition, although we adjusted for several covariates, there is always the risk of unmeasured confounding. Furthermore, we lacked information on the severity of comorbid diseases included in the CCI score, which may have introduced residual confounding. We used multiple imputation to handle missing data on BMI and surgery delay. Due to prospective registration of patient data, any lack of reporting to the DMHFR is probably independent of both BMI level and time to surgery.

In this article, we use RBC transfusion within the first week of a hip fracture as a surrogate for perioperative blood loss. The time frame of 7 days supports this premise, but information about potential active bleeding or anemia is missing.

Although this study contributes to the debate regarding safety among hip fracture patients treated with antithrombotic drugs, further work is required to form a future clinical practice particularly for hip fracture patients admitted on newer NOAC treatment. Also, we found preoperative antiplatelet drugs to be associated with increased risk of blood transfusion and mortality, why clinicians should ensure this group of patients receives an appropriate level of care. However, further studies are needed to understand the nature of these associations.

Conclusions

Hip fracture patients preoperatively treated with NOAC can safely be operated as there is only slightly increased risk of RBC transfusion but not of 30-day mortality. Users of VKA preoperatively had similar risk of RBC transfusion and mortality as VKA non-users. Preoperative use of antiplatelet drugs among hip fracture patients was associated with significantly increased risk of RBC transfusion and 30-day all-cause mortality. However, delaying hip fracture surgery among patients on antiplatelet therapy did not significantly improve the survival, while immediate surgery may still be considered the best treatment strategy.

Sources of funding

This study was funded by the Program for Clinical Research Infrastructure (PROCRIN) established by the Lundbeck Foundation and the Novo Nordisk Foundation.

Notes

Compliance with ethical standards

This article does not contain any studies with human participants or animals performed by any of the authors. The study was approved by the Danish Data Protection Agency (record no. 1-16-02-467-15), and no patient consent was needed as detail has been removed from the case description to ensure anonymity. According to the Danish law, registry-based studies do not require ethical approval.

Conflicts of interest

Cecilie Daugaard, Alma B. Pedersen, and Nickolaj R. Kristensen declare that they have no conflict of interest. Søren P. Johnsen has, outside the submitted work, received speaker honorarium from BMS and Pfizer; participated in board meetings for BMS, Pfizer, and Bayer; and received previous research funding from BMS and Pfizer.

Supplementary material

198_2018_4786_MOESM1_ESM.pdf (102 kb)
ESM 1 (PDF 101 kb)

References

  1. 1.
    Kanis JA, Odén A, McCloskey EV et al (2012) A systematic review of hip fracture incidence and probability of fracture worldwide. Osteoporos Int 23:2239–2256.  https://doi.org/10.1007/s00198-012-1964-3 CrossRefGoogle Scholar
  2. 2.
    Brauer C, Coca-Perraillon M, Cutler D, Rosen A (2009) Incidence and mortality of hip fractures in the United States. JAMA 302:1573–1579.  https://doi.org/10.1001/jama.2009.1462.Incidence CrossRefGoogle Scholar
  3. 3.
    Vestergaard P, Rejnmark L, Mosekilde L (2007) Increased mortality in patients with a hip fracture-effect of pre-morbid conditions and post-fracture complications. Osteoporos Int 18:1583–1593.  https://doi.org/10.1007/s00198-007-0403-3 CrossRefGoogle Scholar
  4. 4.
    Madsen CM, Jantzen C, Lauritzen JB, Abrahamsen B, Jorgensen HL (2016) Temporal trends in the use of antithrombotics at admission. Acta Orthop 87:368–373.  https://doi.org/10.1080/17453674.2016.1195662 CrossRefGoogle Scholar
  5. 5.
    Moran CG, Wenn RT, Sikand M, Taylor AM (2005) Early mortality after hip fracture: is delay before surgery important? J Bone Joint Surg Am 87:483–489.  https://doi.org/10.2106/JBJS.D.01796 Google Scholar
  6. 6.
    Weller I, Wai EK, Jaglal S, Kreder HJ (2005) The effect of hospital type and surgical delay on mortality after surgery for hip fracture. J Bone Joint Surg Br 87:361–366CrossRefGoogle Scholar
  7. 7.
    Ryan DJ, Yoshihara H, Yoneoka D, Egol KA, Zuckerman JD (2015) Delay in hip fracture surgery: an analysis of patient-specific and hospital-specific risk factors. J Orthop Trauma 29:343–348.  https://doi.org/10.1097/BOT.0000000000000313 CrossRefGoogle Scholar
  8. 8.
    Moja L, Piatti A, Pecoraro V, Ricci C, Virgili G, Salanti G, Germagnoli L, Liberati A, Banfi G (2012) Timing matters in hip fracture surgery: patients operated within 48 hours have better outcomes. A meta-analysis and meta-regression of over 190,000 patients. PLoS One 7:e46175.  https://doi.org/10.1371/journal.pone.0046175 CrossRefGoogle Scholar
  9. 9.
    Simunovic N, Devereaux PJ, Sprague S, Guyatt GH, Schemitsch E, DeBeer J, Bhandari M (2010) Effect of early surgery after hip fracture on mortality and complications: systematic review and meta-analysis. CMAJ 182:1609–1616.  https://doi.org/10.1503/cmaj.092220 CrossRefGoogle Scholar
  10. 10.
    Vidán MT, Sánchez E, Gracia Y et al (2011) Causes and effects of surgical delay in patients with hip fracture. Ann Intern Med 155:226.  https://doi.org/10.7326/0003-4819-155-4-201108160-00006 CrossRefGoogle Scholar
  11. 11.
    Schmidt M, Pedersen L, Sørensen HT (2014) The Danish Civil Registration System as a tool in epidemiology. Eur J Epidemiol 29:541–549CrossRefGoogle Scholar
  12. 12.
    Danish Hip Fracture Database. Annual report 2017. https://www.sundhed.dk/sundhedsfaglig/kvalitet/kliniske-kvalitetsdatabaser/akutte-sygdomme/hoftebrud-nip/. Accessed 13 Jun 2018
  13. 13.
    Mainz J, Krog BR, Bjørnshave B, Bartels P (2004) Nationwide continuous quality improvement using clinical indicators: the Danish National Indicator Project. Int J Qual Heal Care 16:i45–i50.  https://doi.org/10.1093/intqhc/mzh031 CrossRefGoogle Scholar
  14. 14.
    Johannesdottir SA, Horváth-Puhó E, Ehrenstein V, Schmidt, Pedersen L, Sørensen H (2012) Existing data sources for clinical epidemiology: the Danish National Database of Reimbursed Prescriptions. Clin Epidemiol 4:303–313.  https://doi.org/10.2147/CLEP.S37587 CrossRefGoogle Scholar
  15. 15.
    Charlson ME, Pompei P, Ales KL, MacKenzie CR (1987) A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 40:373–383.  https://doi.org/10.1016/0021-9681(87)90171-8 CrossRefGoogle Scholar
  16. 16.
    Andersen TF, Madsen M, Jørgensen J et al (1999) The Danish National Hospital Register. A valuable source of data for modern health sciences. Dan Med Bull 46:263–268.  https://doi.org/10.1177/1759720X13491025 Google Scholar
  17. 17.
    Pedersen AB, Mikkelsen EM, Cronin-Fenton D, Kristensen N, Pham TM, Pedersen L, Petersen I (2017) Missing data and multiple imputation in clinical epidemiological research. Clin Epidemiol 9:157–166.  https://doi.org/10.2147/CLEP.S129785 CrossRefGoogle Scholar
  18. 18.
    Van Buuren S (2012) Flexible imputation of missing data. Interdisciplinary statistics series, 1st edn. Chapman and Hall/CRC, New YorkCrossRefGoogle Scholar
  19. 19.
    Ge Y, Ha ACT, Atzema CL, Abdel-Qadir HM, Fang J, Austin PC, Wijeysundera DN, Lee DS (2017) Association of Atrial Fibrillation and Oral Anticoagulant Use with Perioperative Outcomes after Major Noncardiac Surgery. J Am Heart Assoc 6:e006022.  https://doi.org/10.1161/JAHA.117.006022 CrossRefGoogle Scholar
  20. 20.
    Gleason LJ, Mendelson DA, Kates SL, Friedman SM (2014) Anticoagulation Management in Individuals with hip fracture. J Am Geriatr Soc 62:159–164.  https://doi.org/10.1111/jgs.12591 CrossRefGoogle Scholar
  21. 21.
    Collinge CA, Kelly KC, Little B et al (2012) The effects of clopidogrel (Plavix) and other oral anticoagulants on early hip fracture surgery. J Orthop Trauma 26:568–573.  https://doi.org/10.1097/BOT.0b013e318240d70f CrossRefGoogle Scholar
  22. 22.
    Lawrence JE, Fountain DM, Cundall-Curry DJ, Carrothers AD (2016) Do patients taking warfarin experience delays to theatre, longer hospital stay, and poorer survival after hip fracture? Clin Orthop Relat Res 475:273–279.  https://doi.org/10.1007/s11999-016-5056-0 CrossRefGoogle Scholar
  23. 23.
    Doleman B, Moppett IK (2015) Is early hip fracture surgery safe for patients on clopidogrel? Systematic review, meta-analysis and meta-regression. Injury 46:954–962.  https://doi.org/10.1016/j.injury.2015.03.024 CrossRefGoogle Scholar
  24. 24.
    Kragh AM, Waldén M, Apelqvist A, Wagner P, Atroshi I (2011) Bleeding and first-year mortality following hip fracture surgery and preoperative use of low-dose acetylsalicylic acid: an observational cohort study. BMC Musculoskelet Disord 12:254.  https://doi.org/10.1186/1471-2474-12-254 CrossRefGoogle Scholar
  25. 25.
    Anekstein Y, Tamir E, Halperin N, Mirovsky Y (2004) Aspirin therapy and bleeding during proximal femoral fracture surgery. Clin Orthop Relat Res 418:205–208CrossRefGoogle Scholar
  26. 26.
    Madsen CM, Jørgensen HL, Norgaard A, Riis T, Jantzen C, Pedersen OB, Duus BR, Lauritzen JB (2014) Preoperative factors associated with red blood cell transfusion in hip fracture patients. Arch Orthop Trauma Surg 134:375–382.  https://doi.org/10.1007/s00402-013-1906-3 CrossRefGoogle Scholar
  27. 27.
    Robinson WP, Ahn J, Stiffler A et al (2005) Blood transfusion is an independent predictor of increased mortality in nonoperatively managed blunt hepatic and splenic injuries. J Trauma 58:437–444.  https://doi.org/10.1097/01.TA.0000153935.18997.14 CrossRefGoogle Scholar
  28. 28.
    Koch CG, Li L, Duncan AI, Mihaljevic T, Cosgrove DM, Loop FD, Starr NJ, Blackstone EH (2006) Morbidity and mortality risk associated with red blood cell and blood-component transfusion in isolated coronary artery bypass grafting*. Crit Care Med 34:1608–1616.  https://doi.org/10.1097/01.CCM.0000217920.48559.D8 CrossRefGoogle Scholar
  29. 29.
    Pedersen AB, Mehnert F, Overgaard S, Johnsen SP (2009) Allogeneic blood transfusion and prognosis following total hip replacement: a population-based follow up study. BMC Musculoskelet Disord 10:167.  https://doi.org/10.1186/1471-2474-10-167 CrossRefGoogle Scholar
  30. 30.
    Brunskill S, Millette S, Shokoohi A, Pulford E, Doree C MM (2015) Red blood cell transfusion for people undergoing hip fracture surgery. Cochrane Database Syst Rev CD009699. doi:  https://doi.org/10.1002/14651858.CD009699.pub2
  31. 31.
    Danish Patient Safety Authority (2018) Redegørelse for blodproduktområdet 2017. Available at: https://stps.dk/da/udgivelser/2018/redegoerelse-for-blodproduktomraadet-2017/. Accessed 26 November 2018
  32. 32.
    Schmidt M, Schmidt SAJ, Sandegaard JL, Ehrenstein V, Pedersen L, Sørensen HT (2015) The Danish National Patient Registry: a review of content, data quality, and research potential. Clin Epidemiol 7:449–490.  https://doi.org/10.2147/CLEP.S91125 CrossRefGoogle Scholar

Copyright information

© International Osteoporosis Foundation and National Osteoporosis Foundation 2018

Authors and Affiliations

  1. 1.Department of Clinical Medicine, Department of Clinical EpidemiologyAarhus University HospitalAarhus NDenmark
  2. 2.Danish Center for Clinical Health Services Research, Department of Clinical Medicine, Aalborg University HospitalAalborg UniversityAalborgDenmark

Personalised recommendations