G Model
JINJ-5949; No. of Pages 6 Injury, Int. J. Care Injured xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Injury journal homepage: www.elsevier.com/locate/injury
Hip fractures are risky business: An analysis of the NSQIP data Vasanth Sathiyakumar, Sarah E. Greenberg, Cesar S. Molina *, Rachel V. Thakore, William T. Obremskey, Manish K. Sethi The Vanderbilt Orthopaedic Institute Center for Health Policy, 1215 21st Avenue South, Suite 4200, Medical Center East, South Tower, Nashville, TN 3723, United States
A R T I C L E I N F O
A B S T R A C T
Article history: Accepted 14 October 2014
Introduction: Hip fractures are one of the most common types of orthopaedic injury with high rates of morbidity. Currently, no study has compared risk factors and adverse events following the different types of hip fracture surgeries. The purpose of this paper is to investigate the major and minor adverse events and risk factors for complication development associated with five common surgeries for the treatment of hip fractures using the NSQIP database. Methods: Using the ACS-NSQIP database, complications for five forms of hip surgeries were selected and categorized into major and minor adverse events. Demographics and clinical variables were collected and an unadjusted bivariate logistic regression analyses was performed to determine significant risk factors for adverse events. Five multivariate regressions were run for each surgery as well as a combined regression analysis. Results: A total of 9640 patients undergoing surgery for hip fracture were identified with an adverse events rate of 25.2% (n = 2433). Open reduction and internal fixation of a femoral neck fracture had the greatest percentage of all major events (16.6%) and total adverse events (27.4%), whereas partial hip hemiarthroplasty had the greatest percentage of all minor events (11.6%). Mortality was the most common major adverse event (44.9–50.6%). For minor complications, urinary tract infections were the most common minor adverse event (52.7–62.6%). Significant risk factors for development of any adverse event included age, BMI, gender, race, active smoking status, history of COPD, history of CHF, ASA score, dyspnoea, and functional status, with various combinations of these factors significantly affecting complication development for the individual surgeries. Conclusions: Hip fractures are associated with significantly high numbers of adverse events. The type of surgery affects the type of complications developed and also has an effect on what risk factors significantly predict the development of a complication. Concerted efforts from orthopaedists should be made to identify higher risk patients and prevent the most common adverse events that occur postoperatively. ß 2014 Elsevier Ltd. All rights reserved.
Keywords: Hip fracture Risk factor Adverse events
Introduction Hip fractures total nearly 1.6 million cases across the world each year, with an estimated 21.3 million cases by 2050 costing a projected $446.3 billion in treatment [1–3]. Age and sex standardized hip fracture rates range from 1.86 to 47.74 cases per 10,000 depending on country, with rates in the United States
* Corresponding author. Tel.: +1 615 936 0112; fax: +1 615 936 3630. E-mail addresses: [email protected] (V. Sathiyakumar), [email protected] (S.E. Greenberg), [email protected] (C.S. Molina), [email protected] (R.V. Thakore), [email protected] (W.T. Obremskey), [email protected] (M.K. Sethi).
from 1986 to 2005 averaging 957.3 cases per 100,000 women and 414.4 cases per 100,000 men over the age of 65 [4,5]. By 2030, the total number of hip fractures in the United States alone is estimated to affect 289,000 patients [6]. Despite this seemingly common orthopaedic problem, morbidity and mortality rates continue to remain high. Reported 1-year mortality rates in the literature have ranged from 11.9 to 26% of study populations [5,7–10]. Morbidities are also high, with studies reporting average reduced life expectancies of 1.8 years or 25% of remaining life expectancy per patient with an aggregate of 2.9 million disability life years lost as a result of hip fracture [11,12]. Complications secondary to hip fractures occur in up to 12–28% of patients and include pneumonia, urinary retention, and anaemia [13–15].
http://dx.doi.org/10.1016/j.injury.2014.10.051 0020–1383/ß 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Sathiyakumar V, et al. Hip fractures are risky business: An analysis of the NSQIP data. Injury (2014), http://dx.doi.org/10.1016/j.injury.2014.10.051
G Model
JINJ-5949; No. of Pages 6 2
V. Sathiyakumar et al. / Injury, Int. J. Care Injured xxx (2014) xxx–xxx
Many studies have investigated the morbidity and mortality rates of hip fractures using individual hospital data or larger databases such as Medicare cohorts. However, no single study has yet investigated these rates based on pre-operative risk factors on an international scale by using the expanded orthopaedic American College of Surgeons’ National Surgical Quality Improvement (NSQIP) database. This risk-adjusted database contains patient data from 462 hospitals in the US and 34 hospitals around the world and includes 135 patient variables ranging from preoperative risk factors to 30-day postoperative morbidities and mortalities. Recent NSQIP studies have investigated aggregate adverse event trends in all patients with hip fractures and have created risk calculators for morbidity and mortality for patients undergoing hip fracture surgery. However, these studies have not investigated adverse events and risk factors for morbidities and mortalities based on individual surgeries for the treatment of hip fractures [16,17]. The purpose of this paper is to investigate the major and minor adverse events associated with five common surgeries for the treatment of hip fractures using NSQIP data. Based on significantly associated pre-operative risk factors resulting in adverse outcomes, we utilized multivariate analysis to determine which significant patient factors significantly contribute towards a morbidity or mortality outcome. By noting adverse outcome rates and significant risk factors, the practicing orthopaedist may be better equipped to avoid complications through patient education and modifying controllable risk factors. Methods We obtained institutional IRB approval and gained access to the 2005–2011 NSQIP dataset from the American College of Surgeons. We isolated the five types of hip fracture surgeries: partial hip hemiarthroplasty (CPT 27125); percutaneous skeletal fixation of a femoral neck fracture (CPT 27235); open reduction and internal fixation of a femoral neck fracture (CPT 27236); open reduction and internal fixation of an intertrochanteric, peritrochanetric, or subtrochanteric femoral fracture (CPT 27244); and intramedullary fixation of an intertrochanteric, peritrochanteric, or subtrochanteric femoral fracture (CTP 27245) [29]. Complications for these five hip surgeries were then selected and categorized into major and minor adverse events based on previous literature using the NSQIP database [18–22]. The minor complications included superficial wound infections, wound dehiscence, urinary tract infections, and pneumonia. The major complications included mortality and any other complication that was not a minor complication. These included: deep surgical site infection, organ infection, myocardial infarction, deep venous thrombosis, pulmonary embolism, and cerebrovascular accident among other complications. Chi-square tests were used to determine any significant differences in the rates of these adverse events among the five surgeries. Demographics such as age, gender, and race, as well as selected clinical characteristics such as presence of diabetes, smoking status, functional status, and ASA class were collected for patients who underwent these surgeries. Chi-square and Fisher’s exact tests were used to compare categorical demographics among the five groups, and Wilcoxon ranked sum tests were used to compare continuous variables. To determine which demographic and clinical variables were significant risk factors for the development of any major or minor complication, we ran unadjusted bivariate logistic regression analyses on available independent variables in the NSQIP dataset. The independent variables used for bivariate analysis included demographics such as age, sex, race, and BMI; and clinical characteristics such as alcohol use, active smoking status, and a
history of: >10% weight loss in the six months leading to the hip fracture surgery, any additional surgery in the 30 days prior to the hip fracture surgery, preoperative functional status (i.e. independent, dependent, etc.), chronic obstructive pulmonary disease (COPD), congestive heart failure (CHF), hypertension, peripheral vascular disease, esophageal varices, disseminated cancer, steroid use, bleeding disorders, dialysis, chemotherapy, radiotherapy, and blood transfusion in the 72 h prior to the hip fracture surgery. Additional independent variables included operative time and resident involvement in cases, and lab values such as preoperative sodium, blood urea nitrogen (BUN), hematocrit, white blood cell count (WBC), and platelet count. The main outcome was development of any adverse event or complication. A priori significance was set at p < 0.05. Independent variables shown to have significant effects on the development of complications through our bivariate analysis, as well as commonly used demographic measures such as age, gender, race, and BMI, were included as covariates in our multivariate analyses. Five separate multivariate regressions were run – one for each type of surgery. A combined regression was further run by combining all patients together from these five surgeries. Results A total of 9640 patients underwent hip fracture surgery. The overall rate of adverse events was events rate of 25.2% (n = 2433). Open reduction internal fixation of femoral neck fractures had the most patients with 2798 (29.0%), followed by intramedullary fixation of an intertrochanteric, peritrochanteric, or subtrochanteric femoral fracture (n = 2667; 27.7%), partial hip hemiarthroplasty (n = 1822; 18.9%), open reduction internal fixation of an intertrochanteric, peritrochanteric, or subtrochanteric femoral fracture (n = 1471; 15.2%), and finally percutaneous skeletal fixation of a femoral neck fracture (n = 882; 9.2%). Demographics for these five surgeries are provided in Table 1. There were no significant differences in patients with respect to BMI, distribution of gender, or diabetes status, but there were significant differences with respect to age, active smoking status, functional status, and ASA class. The distribution of major and minor adverse events for these five surgeries is detailed in Table 2. The top ten most common major complications were death, deep surgical site infection (SSI), organ infection, myocardial infarction (MI), deep venous thrombosis (DVT), pulmonary embolus (PE), cerebrovascular accident (CVA), coma, peripheral nerve injury, sepsis, and septic shock. Mortality was the most common major adverse event for all five procedures (44.9–50.6%), with either sepsis or MI comprising the second most common major adverse event. Peripheral nerve injury and organ infections were the least common major adverse events. There were only significant differences among the five surgeries for frequency of deep surgical site infection (p = 0.008) and organ infection (p = 0.020). Patients treated by open reduction and internal fixation of a femoral neck fracture had both the highest rates of deep SSI (6%) and organ infection (2.8%). Patients treated by intramedullary fixation of a trochanteric femur fracture had the lowest rate of deep SSI (1.6%), and patients treated by percutaneous skeletal fixation of the femoral neck had the lowest rate of organ infection (0.0%). For minor complications, urinary tract infections were the most common minor adverse event (52.7–62.6%) followed by pneumonia for all five surgeries. There were no significant differences in the frequencies of minor complications among these five groups. The rates of major, minor, and total adverse events were calculated based on the number of patients in each group and are also provided in Table 2. Open reduction and
Please cite this article in press as: Sathiyakumar V, et al. Hip fractures are risky business: An analysis of the NSQIP data. Injury (2014), http://dx.doi.org/10.1016/j.injury.2014.10.051
G Model
JINJ-5949; No. of Pages 6 V. Sathiyakumar et al. / Injury, Int. J. Care Injured xxx (2014) xxx–xxx
3
Table 1 Demographic and selected clinical characteristics of patients undergoing hip fracture repair. Characteristic
Partial hip hemiarthroplasty (CPT: 27125)
Percutaneous skeletal fixation of femoral neck fracture (CPT: 27235)
ORIF of femoral neck fracture (CPT: 27236)
ORIF of intertrochanteric, peritrochanteric, or subtrochanteric femoral fracture (CPT: 27244)
IM fixation of intertrochanteric, peritrochanteric, or subtrochanteric femoral fracture (CPT: 27245)
# Patients (% total) Mean age (SD) BMI (m/kg2) 30 Gender Female Male Race Black White Other Unknown Diabetic No Yes Active smoker No Yes Functional status Independent Partially dependent Totally dependent Unknown ASA class 1 or 2 3 4 Unknown
1822 (18.9) 78.8 (14.2)
882 (9.2) 77.1 (14.0)
2798 (29.0) 79.6 (13.1)
1471 (15.2) 80.3 (13.2)
2667 (27.7) 80.2 (13.1)
268 670 471 244
144 323 220 123
454 1067 679 357
247 488 345 202
438 973 656 381
1244 575
630 249
1971 824
1009 460
1874 789
78 1408 48 288
34 747 21 80
126 2266 73 333
43 1117 34 277
115 2198 63 291
1508 314
722 160
2330 468
1205 266
7943 1697
1642 180
754 128
2434 364
1296 175
223 344
1134 529 146 13
481 320 75 6
1538 998 230 32
884 462 113 12
1477 924 240 26
363 1139 316 4
211 516 151 4
548 1710 535 5
307 923 241 0
523 1637 499 8
p value
JINJ-5949; No. of Pages 6 Injury, Int. J. Care Injured xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Injury journal homepage: www.elsevier.com/locate/injury
Hip fractures are risky business: An analysis of the NSQIP data Vasanth Sathiyakumar, Sarah E. Greenberg, Cesar S. Molina *, Rachel V. Thakore, William T. Obremskey, Manish K. Sethi The Vanderbilt Orthopaedic Institute Center for Health Policy, 1215 21st Avenue South, Suite 4200, Medical Center East, South Tower, Nashville, TN 3723, United States
A R T I C L E I N F O
A B S T R A C T
Article history: Accepted 14 October 2014
Introduction: Hip fractures are one of the most common types of orthopaedic injury with high rates of morbidity. Currently, no study has compared risk factors and adverse events following the different types of hip fracture surgeries. The purpose of this paper is to investigate the major and minor adverse events and risk factors for complication development associated with five common surgeries for the treatment of hip fractures using the NSQIP database. Methods: Using the ACS-NSQIP database, complications for five forms of hip surgeries were selected and categorized into major and minor adverse events. Demographics and clinical variables were collected and an unadjusted bivariate logistic regression analyses was performed to determine significant risk factors for adverse events. Five multivariate regressions were run for each surgery as well as a combined regression analysis. Results: A total of 9640 patients undergoing surgery for hip fracture were identified with an adverse events rate of 25.2% (n = 2433). Open reduction and internal fixation of a femoral neck fracture had the greatest percentage of all major events (16.6%) and total adverse events (27.4%), whereas partial hip hemiarthroplasty had the greatest percentage of all minor events (11.6%). Mortality was the most common major adverse event (44.9–50.6%). For minor complications, urinary tract infections were the most common minor adverse event (52.7–62.6%). Significant risk factors for development of any adverse event included age, BMI, gender, race, active smoking status, history of COPD, history of CHF, ASA score, dyspnoea, and functional status, with various combinations of these factors significantly affecting complication development for the individual surgeries. Conclusions: Hip fractures are associated with significantly high numbers of adverse events. The type of surgery affects the type of complications developed and also has an effect on what risk factors significantly predict the development of a complication. Concerted efforts from orthopaedists should be made to identify higher risk patients and prevent the most common adverse events that occur postoperatively. ß 2014 Elsevier Ltd. All rights reserved.
Keywords: Hip fracture Risk factor Adverse events
Introduction Hip fractures total nearly 1.6 million cases across the world each year, with an estimated 21.3 million cases by 2050 costing a projected $446.3 billion in treatment [1–3]. Age and sex standardized hip fracture rates range from 1.86 to 47.74 cases per 10,000 depending on country, with rates in the United States
* Corresponding author. Tel.: +1 615 936 0112; fax: +1 615 936 3630. E-mail addresses: [email protected] (V. Sathiyakumar), [email protected] (S.E. Greenberg), [email protected] (C.S. Molina), [email protected] (R.V. Thakore), [email protected] (W.T. Obremskey), [email protected] (M.K. Sethi).
from 1986 to 2005 averaging 957.3 cases per 100,000 women and 414.4 cases per 100,000 men over the age of 65 [4,5]. By 2030, the total number of hip fractures in the United States alone is estimated to affect 289,000 patients [6]. Despite this seemingly common orthopaedic problem, morbidity and mortality rates continue to remain high. Reported 1-year mortality rates in the literature have ranged from 11.9 to 26% of study populations [5,7–10]. Morbidities are also high, with studies reporting average reduced life expectancies of 1.8 years or 25% of remaining life expectancy per patient with an aggregate of 2.9 million disability life years lost as a result of hip fracture [11,12]. Complications secondary to hip fractures occur in up to 12–28% of patients and include pneumonia, urinary retention, and anaemia [13–15].
http://dx.doi.org/10.1016/j.injury.2014.10.051 0020–1383/ß 2014 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Sathiyakumar V, et al. Hip fractures are risky business: An analysis of the NSQIP data. Injury (2014), http://dx.doi.org/10.1016/j.injury.2014.10.051
G Model
JINJ-5949; No. of Pages 6 2
V. Sathiyakumar et al. / Injury, Int. J. Care Injured xxx (2014) xxx–xxx
Many studies have investigated the morbidity and mortality rates of hip fractures using individual hospital data or larger databases such as Medicare cohorts. However, no single study has yet investigated these rates based on pre-operative risk factors on an international scale by using the expanded orthopaedic American College of Surgeons’ National Surgical Quality Improvement (NSQIP) database. This risk-adjusted database contains patient data from 462 hospitals in the US and 34 hospitals around the world and includes 135 patient variables ranging from preoperative risk factors to 30-day postoperative morbidities and mortalities. Recent NSQIP studies have investigated aggregate adverse event trends in all patients with hip fractures and have created risk calculators for morbidity and mortality for patients undergoing hip fracture surgery. However, these studies have not investigated adverse events and risk factors for morbidities and mortalities based on individual surgeries for the treatment of hip fractures [16,17]. The purpose of this paper is to investigate the major and minor adverse events associated with five common surgeries for the treatment of hip fractures using NSQIP data. Based on significantly associated pre-operative risk factors resulting in adverse outcomes, we utilized multivariate analysis to determine which significant patient factors significantly contribute towards a morbidity or mortality outcome. By noting adverse outcome rates and significant risk factors, the practicing orthopaedist may be better equipped to avoid complications through patient education and modifying controllable risk factors. Methods We obtained institutional IRB approval and gained access to the 2005–2011 NSQIP dataset from the American College of Surgeons. We isolated the five types of hip fracture surgeries: partial hip hemiarthroplasty (CPT 27125); percutaneous skeletal fixation of a femoral neck fracture (CPT 27235); open reduction and internal fixation of a femoral neck fracture (CPT 27236); open reduction and internal fixation of an intertrochanteric, peritrochanetric, or subtrochanteric femoral fracture (CPT 27244); and intramedullary fixation of an intertrochanteric, peritrochanteric, or subtrochanteric femoral fracture (CTP 27245) [29]. Complications for these five hip surgeries were then selected and categorized into major and minor adverse events based on previous literature using the NSQIP database [18–22]. The minor complications included superficial wound infections, wound dehiscence, urinary tract infections, and pneumonia. The major complications included mortality and any other complication that was not a minor complication. These included: deep surgical site infection, organ infection, myocardial infarction, deep venous thrombosis, pulmonary embolism, and cerebrovascular accident among other complications. Chi-square tests were used to determine any significant differences in the rates of these adverse events among the five surgeries. Demographics such as age, gender, and race, as well as selected clinical characteristics such as presence of diabetes, smoking status, functional status, and ASA class were collected for patients who underwent these surgeries. Chi-square and Fisher’s exact tests were used to compare categorical demographics among the five groups, and Wilcoxon ranked sum tests were used to compare continuous variables. To determine which demographic and clinical variables were significant risk factors for the development of any major or minor complication, we ran unadjusted bivariate logistic regression analyses on available independent variables in the NSQIP dataset. The independent variables used for bivariate analysis included demographics such as age, sex, race, and BMI; and clinical characteristics such as alcohol use, active smoking status, and a
history of: >10% weight loss in the six months leading to the hip fracture surgery, any additional surgery in the 30 days prior to the hip fracture surgery, preoperative functional status (i.e. independent, dependent, etc.), chronic obstructive pulmonary disease (COPD), congestive heart failure (CHF), hypertension, peripheral vascular disease, esophageal varices, disseminated cancer, steroid use, bleeding disorders, dialysis, chemotherapy, radiotherapy, and blood transfusion in the 72 h prior to the hip fracture surgery. Additional independent variables included operative time and resident involvement in cases, and lab values such as preoperative sodium, blood urea nitrogen (BUN), hematocrit, white blood cell count (WBC), and platelet count. The main outcome was development of any adverse event or complication. A priori significance was set at p < 0.05. Independent variables shown to have significant effects on the development of complications through our bivariate analysis, as well as commonly used demographic measures such as age, gender, race, and BMI, were included as covariates in our multivariate analyses. Five separate multivariate regressions were run – one for each type of surgery. A combined regression was further run by combining all patients together from these five surgeries. Results A total of 9640 patients underwent hip fracture surgery. The overall rate of adverse events was events rate of 25.2% (n = 2433). Open reduction internal fixation of femoral neck fractures had the most patients with 2798 (29.0%), followed by intramedullary fixation of an intertrochanteric, peritrochanteric, or subtrochanteric femoral fracture (n = 2667; 27.7%), partial hip hemiarthroplasty (n = 1822; 18.9%), open reduction internal fixation of an intertrochanteric, peritrochanteric, or subtrochanteric femoral fracture (n = 1471; 15.2%), and finally percutaneous skeletal fixation of a femoral neck fracture (n = 882; 9.2%). Demographics for these five surgeries are provided in Table 1. There were no significant differences in patients with respect to BMI, distribution of gender, or diabetes status, but there were significant differences with respect to age, active smoking status, functional status, and ASA class. The distribution of major and minor adverse events for these five surgeries is detailed in Table 2. The top ten most common major complications were death, deep surgical site infection (SSI), organ infection, myocardial infarction (MI), deep venous thrombosis (DVT), pulmonary embolus (PE), cerebrovascular accident (CVA), coma, peripheral nerve injury, sepsis, and septic shock. Mortality was the most common major adverse event for all five procedures (44.9–50.6%), with either sepsis or MI comprising the second most common major adverse event. Peripheral nerve injury and organ infections were the least common major adverse events. There were only significant differences among the five surgeries for frequency of deep surgical site infection (p = 0.008) and organ infection (p = 0.020). Patients treated by open reduction and internal fixation of a femoral neck fracture had both the highest rates of deep SSI (6%) and organ infection (2.8%). Patients treated by intramedullary fixation of a trochanteric femur fracture had the lowest rate of deep SSI (1.6%), and patients treated by percutaneous skeletal fixation of the femoral neck had the lowest rate of organ infection (0.0%). For minor complications, urinary tract infections were the most common minor adverse event (52.7–62.6%) followed by pneumonia for all five surgeries. There were no significant differences in the frequencies of minor complications among these five groups. The rates of major, minor, and total adverse events were calculated based on the number of patients in each group and are also provided in Table 2. Open reduction and
Please cite this article in press as: Sathiyakumar V, et al. Hip fractures are risky business: An analysis of the NSQIP data. Injury (2014), http://dx.doi.org/10.1016/j.injury.2014.10.051
G Model
JINJ-5949; No. of Pages 6 V. Sathiyakumar et al. / Injury, Int. J. Care Injured xxx (2014) xxx–xxx
3
Table 1 Demographic and selected clinical characteristics of patients undergoing hip fracture repair. Characteristic
Partial hip hemiarthroplasty (CPT: 27125)
Percutaneous skeletal fixation of femoral neck fracture (CPT: 27235)
ORIF of femoral neck fracture (CPT: 27236)
ORIF of intertrochanteric, peritrochanteric, or subtrochanteric femoral fracture (CPT: 27244)
IM fixation of intertrochanteric, peritrochanteric, or subtrochanteric femoral fracture (CPT: 27245)
# Patients (% total) Mean age (SD) BMI (m/kg2) 30 Gender Female Male Race Black White Other Unknown Diabetic No Yes Active smoker No Yes Functional status Independent Partially dependent Totally dependent Unknown ASA class 1 or 2 3 4 Unknown
1822 (18.9) 78.8 (14.2)
882 (9.2) 77.1 (14.0)
2798 (29.0) 79.6 (13.1)
1471 (15.2) 80.3 (13.2)
2667 (27.7) 80.2 (13.1)
268 670 471 244
144 323 220 123
454 1067 679 357
247 488 345 202
438 973 656 381
1244 575
630 249
1971 824
1009 460
1874 789
78 1408 48 288
34 747 21 80
126 2266 73 333
43 1117 34 277
115 2198 63 291
1508 314
722 160
2330 468
1205 266
7943 1697
1642 180
754 128
2434 364
1296 175
223 344
1134 529 146 13
481 320 75 6
1538 998 230 32
884 462 113 12
1477 924 240 26
363 1139 316 4
211 516 151 4
548 1710 535 5
307 923 241 0
523 1637 499 8
p value
