A cohort study on the incidence and outcome of pulmonary embolism in trauma and orthopedic patients

Pulmonary embolism (PE) and deep venous thrombosis (DVT) can be considered under the spectrum of venous thromboembolic (VTE) disease. No definitive scientific data exist regarding the overall incidence of VTE in the general population, but a recent study estimates the incidence to range between 1 and 5/1,000 in the general population [1]. In the surgical population, the prevalence can reach more than 50% in the absence of thromboprophylaxis [1].

Worldwide, more than 50% of all hospitalized patients are at risk for VTE and surgical patients are at higher risk than medical patients [2]. The incidence of PE represents 5 to 10% of deaths in the hospital setting, making this condition the most common preventable cause of in-hospital death [3‐6]. In addition, VTE and associated complications contribute substantially to patient morbidity and treatment costs [7, 8].

Within the discipline of trauma and orthopedics, the prevalence of DVT and PE has been estimated to be 1.16% and 0.93%, respectively [9]. Mortality rates have been reported to range between 0.38% and 13.8% [10, 11]. Principal risk factors include an injury severity score (ISS) greater than 50 and more than two surgical procedures [9]. PE appears to be the most common cause of mortality in patients that survive the first 24 hours following trauma and retrospective post-mortem data have demonstrated that out of an overall mortality of 13.8%, 1.6% was a consequence of fatal PE [10]. In the elective orthopedic clinical setting, PE is the second most frequent cause of death in patients that undergo lower limb total joint arthroplasty [11].

Despite the existing data reporting on the overall prevalence of PE in the trauma and orthopedic population, it remains a common belief that as the clinical signs and symptoms are non-specific and frequently silent, this complication may still be underdiagnosed [12]. The aim of this study is to determine the incidence of PE in trauma and orthopedic patients admitted to one of the largest tertiary referral centers in the UK and to investigate its association with the pattern of injury, type of treatment, co-morbidities, thromboprophylaxis and mortality.

Over the pre-specified study period, 18,151 orthopedic patients (10,648 for elective operations and 7,503 for trauma) were admitted to our institution. During the same study period, 5,656 CT-PA investigations were requested by all the medical disciplines out of which 151 (2.67%) requests were from the discipline of trauma and orthopedics. Six hundred and fifty (11.5%) patients had positive findings of PE. Out of the positive 650 cases, 86 (13.2%) patients were from our discipline (trauma and orthopedics) and formed the study cohort. Taking into consideration the number of clinically suspected PE (151) and the number of positive findings on the CT-PA (86 patients) it was estimated that 57% of patients presenting with clinical features consistent for PE had positive radiological CT-PA findings. One patient, however, was excluded from the study as the PE event occurred shortly prior to hospital admission for TKA and, thus, 85 patients formed the study cohort (see Figure 1).

Despite continuous improvement in medical knowledge and treatment modalities, the incidence of VTE and its related complications has remained fairly static during the last three decades [25]. Notwithstanding the implementation of prevention protocols, clinical manifestation of PE is not clear or specific. PE may be expressed in a silent way and can be missed by the clinical team. For this reason, PE might be under diagnosed.

In our study population, the incidence of PE was in line with data published in previous studies [1, 9, 12, 26‐28]. On further subgroup analysis it was noted that the incidence of PE was lower for the elective surgery cohort (0.23%) compared to the data recently published in the guidelines of the American College of Chest Physicians [26] (0.35%) and Markovic-Denic et al. [1] (1.6%), but is comparable to that reported by Jean-Marie Janue (0.14% in THA and 0.27% in TKA) [12]. The prevalence noted in trauma patients was in accordance with Maneker et al. [29] who showed a rate of 0.27% in his study population, which is lower than the rates reported by other studies [9, 30]. These differences, nevertheless, are difficult to explain in retrospective studies, performed in different geographic areas with different protocols of prophylaxis and treatment.

On the basis of current evidence, CT-PA is considered the gold standard for the diagnosis of PE [13, 31]. Chest contrast enhanced CT replaced catheter angiography due to its less invasive nature and accuracy, and has been proven to be superior or equal to angiography [31]. The reported sensitivity for the diagnosis of PE with CT-PA varies from 45 to 100% and the specificity from 78 to 100% [31]. CT-PA has some limitations in detecting isolated sub-segmental PE [31], but the introduction of the multi-detector CT technique currently allows evaluation of pulmonary vessels down to the sixth order branches, thus, significantly increasing the rate of detection of PE (sensitivity: 83%, specificity: 96%) [14, 32].

Many authors have attempted to correlate specific risk factors to the development of PE. Strong correlation was identified for the number and magnitude of surgical interventions, previous history of VTE and the length of the hospitalization period [25, 33, 34]. The next highly reported risk factors for VTE are cardiovascular disease [1, 23, 33, 34] and obesity [10, 12, 23, 35]. More than half of our study cohort belongs to the high-risk category with more than two risk factors being present as defined by the NICE guidelines [16].

Although 79.3% of our study population was on TP treatment, patients still developed PE. We could not identify any additional specific related factors to VTE development, but we have observed that in our cohort, 62.4% of patients were older than 60 years and 22.4% were more than 80 years of age. Several studies reported age as an independent risk factor for VTE [1, 26, 36]. In our cohort, many of these elderly patients also had lower limb pathology (83.5%). One may speculate a synergistic effect of these two parameters in reducing mobility and leading to a higher risk of developing PE. In a recent case-crossover study reduced mobility was reported as a significant trigger of hospitalization for VTE. The risk of VTE hospitalization was 4.2-fold greater in the time period when reduced mobility occurred [37].

In 13 cases, TP treatment was not prescribed even though the patients had risk factors for VTE. Eight of these 13 cases sustained foot and ankle injuries, which were managed non-operatively and followed up in the outpatient fracture clinics. The reason for this can be attributed to the lack of clarity of national and local guidelines about TP treatment in the out-patient setting, particularly with injury patterns that are considered as less debilitating. Shibuya et al. [38] stated that routine use of TP treatment in foot and ankle injuries is not warranted in contrast to our findings, which support the view that even minor foot injuries cannot be neglected and risk assessment should be performed on an individual basis. This has led to the expansion of the routine risk assessment of patients treated in our out-patient setting and regular audit cycles have been implemented to ensure consistent compliance.

Concomitant DVT was identified in one-third of our study cohort. Knudson et al. [39] analyzed the American College of Surgeons National Trauma Data Bank and found 522 cases of PE out of 450,375 trauma patients (0.11%). In only 16% of these cases a concomitant DVT was diagnosed. In a prospective cohort study [40] of 397 patients with the clinical suspicion of PE, 149 were positive for PE and less than one-third had a concomitant DVT. Cipolle et al. [41] performed a trauma registry analysis of 10,141 trauma admissions, and found 30 cases of PE, of which only 5 (16.7%) had coexisting DVT. Moreover, in a retrospective review of medical records of 247 trauma patients who underwent TPA/CTV over a three-year period, Velmahos et al. [42] recognized positive findings of PE in 46 patients (19%) and among these, only 7 (15%) also had a DVT. Hypothesizing that CTPA/CTV was considered the most accurate method for diagnosing VTE, the same authors [42] investigated this lack of association between PE and DVT. They stated that it was unlikely that a diagnosis of DVT could have been significantly missed for an insufficient sensitivity of the diagnostic tool. Therefore, they hypothesized that the clots might be formed de novo within the pulmonary circulation as a consequence of the changes within the lung vascular endothelium and in the rheological blood properties induced by a post-traumatic hyper-adrenergic and hyper-inflammatory state. However, there is still no definitive evidence about the etiological relationship between DVT and PE, and further studies are desirable to comprehend this phenomenon.

Our mortality was low and consistent with other reports in the literature (Table 9). The low mortality rate noted in the trauma patient subgroup could be attributed to the high percentage of less severe traumatic injuries (frequency of polytrauma patients: 11.9%) and to the good compliance rate of implementing our TP treatment protocols. Consequently, the average number of co-morbidities in our sample was 2.6. A higher proportion of non-survivors had three or more co-morbidities in contrast to the survivors, and this difference was statically significant (P = 0.034).

The present study has several limitations, including the retrospective nature of data collection, from the case notes and electronic databases, the small sample size, the short study period (two years) and the absence of a control group. Moreover, our data pool documented events occurring only during hospital stay (primary or readmission). We are aware that some of our study population could have been admitted or treated elsewhere as we treat a number of tertiary referred patients and, as such, we might have missed some patients who developed PE. Strengths of the study include the identification of consecutive patients with specific injury patterns and risk factors who sustained PE in a large teaching hospital over a two-year period.

In this study, following the NICE guidelines for thromboprophylaxis, the incidence of VTE was found to be similar to the rates reported in the international literature, whereas the mortality rate was considerably lower. It appears that local protocols, in compliance with the NICE guidelines, are effective in the prevention of VTE and in reducing mortality in trauma and orthopedic patients. However, despite the wide administration of both mechanical and chemical TP treatment, patients can still develop PE. It appears that the possibility of PE development is not only related to certain patient related risk factors but also to the consequence of all the aspects of the post-traumatic or post-surgical disease process. Overall, the type of treatment, the type and length of drug administration, the duration of immobilization and the individual response of each patient appear to contribute to the development of this rare yet fearful complication.

Further studies are desirable to monitor the incidence and outcome of PE in trauma and orthopedic patients so that on-going vigilance and on-going evaluation of the efficacy and effectiveness of treatment protocols will ensure that the morbidity will remain low and the mortality will continue to improve.