Abstract
Purpose:
The purposes of this study were to determine the rate of venous thromboembolism (VTE) after discharge from the hospital in patients treated operatively with a pelvic ring or acetabular fracture and to define the main time frame in which VTE occurs within the 90-day period after hospital discharge.
Methods:
California and Florida State Inpatient Databases from 2005 to 2009 were used to identify patients with clinically significant VTEs within 90 days of hospital discharge. ICD-9 diagnosis codes identified patients with a pelvic ring or acetabular fracture and a VTE. Procedure codes distinguished patients having surgical fracture treatment. Deep vein thrombosis (DVT) and pulmonary embolus (PE) were included.
Results:
Overall, 13,589 patients had a pelvic ring or acetabular fracture and operative treatment. One hundred thirteen patients (0.83%) had a VTE within 90 days after hospital discharge: 0.51% had a DVT, 0.21% had a PE, and 0.12% had both. Twenty eight percent of DVTs and 23% of PEs occurred >35 days after discharge, being evenly distributed out to 90 days. Therefore, overall, <0.2% of patients developed a DVT and <0.1% were diagnosed with a PE (<0.01% fatal) >35 days after the index hospitalization.
Conclusions:
A substantial proportion of VTE events occur over 35 days after discharge; however, the overall risk is low with fatal PE being extremely low (<0.01%). Given the diminished VTE risk after 35 days, the decision to further extend antithrombotic drug therapy may be guided by patient-specific factors, such as prolonged immobility.
Introduction
Patients undergoing operative treatment of pelvic ring and acetabular fractures are at high risk for venous thromboembolism (VTE). 1 –7 The overall prevalence of deep vein thrombosis (DVT) in pelvic fracture patients has been shown to be as high as 60% without prophylactic treatment. 8 The prevalence of proximal lower extremity DVT is estimated at more than 25% without prophylactic treatment. 1,8 Published reports indicate that routine postoperative thromboprophylaxis is required. 9 –12 With few notable exceptions, the use of a pharmacologic agent is one of the preferred methods. 1,5,8,12 –16 However, lack of consensus remains concerning the length of time prophylaxis should be continued after hospital discharge. 5,12 Letournel in his classic text recommended a 6-week course; however, no data were provided to support this recommendation. 4 Absent any hard data specific to acetabular and pelvic fracture patients addressing the prevention of VTE, results from large trials involving broader trauma populations currently remain the best secondary source of guidance. 5 However, recommendations for these patients undergoing major orthopedic surgery range widely, from a few days to as long as 3 months 12,17 ; guidelines published in 2004 and reiterated in 2008 and 2012 by the American College of Chest Physicians (ACCP) suggest 35 days from the day of surgery. 9,13,16 A reason for this uncertainty is the paucity of data regarding the occurrence of VTE after a patient’s hospital discharge. The purposes of this study were to determine the rate of VTE after discharge from the hospital in patients treated operatively with a pelvic ring or acetabular fracture and to identify the main time frame in which VTE occurs within the 90-day period after the patient’s hospital discharge.
Materials and methods
Data sources
The State Inpatient Database (SID) from California and Florida were reviewed to identify patients with clinically symptomatic VTEs occurring within 90 days after hospital discharge. These SIDs, from the years 2005 to 2009, are maintained by the Healthcare Cost and Utilization Project (HCUP). 18 HCUP is a family of health-care databases and related software tools and products developed through a Federal-State-Industry partnership and sponsored by the Agency for Healthcare Research and Quality. HCUP databases bring together the data collection efforts of State data organizations, hospital associations, private data organizations, and the Federal government to create a national information resource of encounter-level health care data (HCUP Partners). 18 HCUP includes the largest collection of longitudinal hospital care data in the United States, with all-payer, encounter-level information beginning in 1988. The databases contain 100% of patient records for each available state. 18,19 These particular databases were chosen because they contain a unique identifier to track each patient over time, ensuring that those with clinically symptomatic VTE can be found, whether they return to the same hospital or go to a different hospital within the state. 19 The SIDs contain complete records for each patient visit. Pertinent information for each visit included patient demographics, International Classification of Disease, Ninth Revision (ICD-9) diagnosis codes, procedure codes, mortality, and length of stay. Ethical Committee approval for use of these data was sought and obtained from our Institutional Review Board.
Patient selection
ICD-9 diagnosis and procedure codes were used to identify the patients having surgical treatment for their pelvic ring or acetabular fracture (Table 1). Patients who obtained initial treatment at a hospital located in a different state from their home state were excluded. The expectation would be that in the acute setting, as may be seen in a patient with a symptomatic DVT or pulmonary embolus (PE), patients would be more likely to seek immediate treatment at a hospital in their home state rather than returning to an out-of-state hospital. Therefore, their in-state return visit would be a confounding factor in the database. Patients were also excluded if they had less than 90 days of follow-up by way of their surgical fracture treatment occurring in October, November, or December of 2009.
ICD-9 codes used to identify patients with a pelvic or acetabulum fracture and surgical treatment.
ICD-9: International Classification of Disease, Ninth Revision.
Identification of venous thromboembolic events
ICD-9 diagnosis codes were then used to identify the patients with a DVT or PE (Table 2). The main focus of this study was to determine the rate of post-hospital discharge VTE. Therefore, any VTE event occurring during the index (initial fracture treatment) hospital visit was excluded. The time of VTE onset was recorded as the number of days after the patient’s discharge from the hospital. To isolate VTE risk related to the index surgical treatment, patients who were readmitted to the hospital were censored from that point forward. In addition, patients were censored after one VTE event to avoid counting a patient more than once.
ICD-9 codes used to identify the pelvic and acetabular fracture patients with a venous thromboembolism (DVT or PE).
DVT: deep vein thrombosis; PE: pulmonary embolism; ICD-9: International Classification of Disease, Ninth Revision.
Statistical analysis
VTE event rates were determined by recording the absolute number of events. The rate of DVTs and PEs were also determined individually.
Results
Overall, 13,589 patients were identified as having undergone surgical treatment for a pelvic ring or acetabular fracture. The mean age of the population was 48 years (SD: 21; range 16 years and older). In these 13,589 patients, a total of 268 VTE events were recorded; however, 155 were excluded for the following reasons: (1) 22 were recurrent with the patient already recorded as having a VTE, (2) 97 were recorded as occurring on the same day as the index hospital discharge, and (3) 36 occurred during their index hospital stay. Therefore, of these 13,589 patients, 113 patients (0.83%) were identified as having had a VTE within 90 days after their index hospitalization: 0.51% had a DVT, 0.21% had a PE, and 0.12% had both. In total, 0.63% developed a post-index hospitalization DVT, presenting at a mean of 25 days after discharge (SD: 23) and ranging from 1 to 88 days. In total, 44 patients (0.32%) were diagnosed with a post-index hospitalization PE, presenting at a mean of 19 days after discharge (SD: 23) and ranging from 1 to 82 days. Twenty eight percent of the DVTs and 23% of the PEs occurred more than 35 days after the index hospitalization, being evenly distributed out to 90 days. Eleven percent of PEs resulted in a fatality with a fatal PE occurring in less than 0.01% of patients more than 35 days after the index hospitalization. Therefore, overall, less than 0.2% of patients developed a DVT and less than 0.1% were diagnosed with a PE (<0.01% fatal) more than 35 days after the index hospitalization.
The mean age of patients presenting with a post-index hospitalization VTE was 55 years (SD: 21; range: 17–95 years). Sixty-eight (60%) were male and forty-five (40%) were female. Patients that returned with a VTE had stayed in the hospital during their index (surgical) visit on average 14 days (SD: 20; range: 1–143 days).
Discussion
Patients undergoing operative treatment for pelvic ring and acetabular fractures present with an increased risk for developing VTE. 1,4,5,7,8,17,20 Despite the recognition of this issue and the institution of various prophylaxis regimens, VTE remains an important cause of morbidity and mortality for patients with pelvic and acetabular fracture. 5,17 In this patient population, risk factors for VTE development include advanced age, operative repair, and venous stasis during surgery and early postoperative care. 12,13,20 Postoperative thromboprophylaxis remains the optimal strategy for preventing VTE development with the use of a pharmacologic agent as one of the preferred methods. 1,8 –17,21 However, there are no current prophylaxis guidelines specific for this high-risk population of patients with a pelvic or acetabular fracture. 5 It has been recommended that until such time that data specific to acetabular and pelvic fracture patients become available, physicians involved in the care of these patients should use published guidelines from other trauma populations when making their prophylaxis decisions. 5
Predictably, it follows that a lack of consensus exists concerning the length of time prophylaxis should be continued after hospital discharge. Recommendations now range widely, from a week to as long as 3 months. 5,9,13,15 –17,21,22 Guidelines published by the ACCP recommends 35 days from the day of surgery for hip fracture patients and those undergoing major orthopedic surgery. 9,13,16 A reason for this uncertainty is the paucity of data regarding the occurrence of a VTE after a patient’s hospital discharge. The purposes of this study were to determine the rate of VTE after discharge from the hospital in patients treated operatively with a pelvic ring or acetabular fracture and to define the main time frame in which VTE occurs within the 90-day period after the patient’s hospital discharge.
Our results showed that the rate of post-discharge VTE occurrence is 0.83%: DVT 0.51%, PE 0.21%, both DVT and PE 0.12%. In addition, similar results were noted in a study presented at the 7th ACCP Conference on antithrombotic and thrombolytic therapy in 2004. 13 Of a cohort of 897 patients who had hip fracture surgery and were treated with prophylaxis for roughly 5 weeks, 7 (0.8%) patients developed a DVT. No PEs were reported in their population. 13 Although, the ACCP recommends prophylaxis for up to 35 days from the day of surgery, our results showed that a large portion of the postoperative VTE events occur after 35 days post-discharge. In our patient population, 28% of DVT events presented after 35 days post-discharge, presenting as far out as 88 days post-discharge. In addition, 23% of PE events occurred after 35 days post-discharge, ranging up to 82 days. However, less than 0.01% of patients developed a fatal PE over 35 days post-discharge.
In 2006, Bjørnarå et al. reported on a series of hip fracture patients followed for 6 months following surgery for the purpose of determining the occurrence and timing of non-fatal postoperative VTE with similar findings. 22 Patients received VTE prophylaxis with low molecular weight heparin for approximately 10 days or until discharge. 22 Over the 13-year time period of this study, 2420 patients underwent surgery for hip fracture and overall, 2.8% had a documented postoperative VTE (1.5% being a DVT and 1.3% being a PE). 22 The majority of patients were documented to have sustained their VTE after discharge from the hospital: DVT occurred in 1% (26/2420) patients after hospital discharge; PE occurred in 0.7% (17/2420) after hospital discharge. Interestingly, the time to VTE after hospital discharge, recorded as days after surgery, ranged widely. For DVT, the median was 29 days (range 8–150 days); for PE, the median was 29 days (range 10–173 days). 22 The authors determined that a cumulative risk of VTE lasted for up to 3 months in these fracture patients.
Our study has a number of limitations. Analysis of the SID, a restricted-access, publicly available dataset that is maintained by the AHRQ has as one of the provisions of the State Data Use Agreement that no data observation involving less than or equal to 10 observations is to be published. This measure was put in place by AHRQ to protect individual patients' privacy and to prevent the potential disclosure of personal information. However, publication of percentages is allowed, which we have provided. In addition, the database did not include information on thromboprophylaxis. However, the Surgical Care Improvement Project (SCIP) suggested in 2005 that surgery patients receive appropriate VTE prophylaxis within 24 h prior to surgery to 24 h after surgery. 23 Furthermore, postoperative thromboprophylaxis has been a recommended practice since 1994 and a standard practice since 2004. 8,9,13,16 Therefore, it is a reasonable assumption that the patients included in our study would have received some type of postoperative thromboprophylaxis. Nonetheless, the exact type of thromboprophylaxis is unknown. In addition, it is unknown how long the thromboprophylaxis was provided to each study patient. However, since 2004, the ACCP has recommended 35 days from the day of surgery. 9,13,16 Therefore, the authors also believe it is a reasonable assumption that the patients included in our study received some type of post-hospital discharge extended thromboprophylaxis. Furthermore, our time frame of major interest is after 35 days, at which time it would be expected that any extended thromboprophylaxis would have been discontinued.
An additional study limitation concerns the cut-off date for our analysis. With the data sets used, the exact surgery date could not be verified for each patient. Therefore, the period for our evaluation beyond the ACCP recommendation for thromboprophylaxis was considered as being after 35 days of discharge from the hospital. Since the patients’ total hospitalization ranged from on average 14 days (SD: 20; range: 1–143 days), neither the results or conclusions of this study would be appreciably affected by this limitation. Another point to consider is that VTE events may have been missed if patients acutely presented to an outpatient clinic for the event. However, outpatient clinics regularly refer patients with a clinically symptomatic VTE to a local emergency department for evaluation and treatment. During the time of the collected data used for this study (2005–2009), all patients diagnosed with a VTE would have been hospitalized. It is only more recently that home treatment has been recommended for patients with an acute DVT and only under special circumstances, such as well-maintained living conditions, strong support from family or friends, phone access, and ability to quickly return to the hospital if there is deterioration. 9 In addition, the patient must feel well enough to be treated at home. 9 Therefore, the expectation is that the majority of these VTE events would have been captured after referral. Furthermore, the exact surgical procedure is not specified for each patient. The recommendation for prophylaxis use is provided in broad terms by the ACCP for patients with pelvic ring and acetabular fractures. Therefore, our data reflects the general overview that is consistent with the ACCP guidelines. 9,13,16
Our study is also limited by the fact that it is essentially a retrospective analysis of data derived from hospital-based coding. Human errors can occur during the input and export of individual data sets that we are unable to validate. Furthermore, our study does not provide definitive causality between the operative treatment for pelvic ring and acetabular fractures and the occurrence of VTE. Therefore, it is not known how many patients could have sustained a VTE unrelated to their pelvic injury. However, as noted above, many previous studies have determined causality. 1 –8 In addition, the mistaken inclusion of any VTE unrelated to the pelvic trauma would have erroneously inflated the total number recorded, indicating that the occurrence of VTE specifically related to the operative treatment for pelvic ring and acetabular fractures, if anything, may be even lower than our reported low post-hospital discharge numbers. Only a prospective study evaluating the continuation of prophylaxis for late VTE in at risk patients in this specific population will eliminate our current study’s limitations.
Conclusion
Post-hospital discharge thromboprophylaxis as recommended does not eliminate VTE risk; however, the overall occurrence of postoperative VTE after hospital discharge in pelvic ring and acetabular fracture patients is low (0.8%). A substantial proportion of these VTE events, including fatal PE, occur over 35 days after discharge; however, the risk of fatal PE is extremely low (<0.01%). Given the diminished VTE risk after 35 days, the decision to further extend antithrombotic drug therapy may be guided by patient-specific factors, such as prolonged immobility.
Footnotes
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
