Abstract
Objectives
Pulmonary embolism is believed to be a common cause of death of hospital inpatients. The aims of this study were to estimate the number of deaths caused by pulmonary embolism and the potential to reduce this by the use of caval filters according to accepted indications.
Design
Review of autopsy reports and death notification records from 2007 and 2008. When pulmonary embolism was given as cause of death (in the autopsy report or in section 1 a-c or part 2 of the Medical Certificate of the Cause of Death), hospital records were reviewed for evidence of pre-mortem diagnosis of pulmonary embolism or deep vein thrombosis (DVT) and for evidence of accepted indications for caval filter placement.
Setting
Large UK teaching hospital.
Participants
Hospital inpatients whose deaths were attributed to pulmonary embolism.
Main outcome measures
Proportion of deaths adjudged at autopsy to be due to pulmonary embolism; evidence of pre-mortem diagnosis of DVT or pulmonary embolism; total number of hospital admission and deaths.
Results
From a total of 186,517 adult inpatient admissions there were 2583 (1.4%) adult inpatient deaths of which 696 (27%) underwent autopsy. Of those undergoing autopsy, 14 (2.0%, 95% CI 1.2–3.3%) deaths were caused by pulmonary embolism. Pulmonary embolism was recorded as a cause of death in a further 12 (0.7%) of 1773 patients who did not undergo autopsy. Of these, five had a pre-mortem diagnosis of DVT or pulmonary embolism.
Conclusions
The proportion of deaths caused by pulmonary embolism appears to be considerably lower than the widely published rate, and of this small number, few have a pre-mortem diagnosis of DVT or pulmonary embolism. There is little scope for further reduction of pulmonary embolism mortality through use of caval filters according to guidelines. Current policy on pulmonary embolism risk prevention appears to be based on an over-estimate of the level of risk.
Introduction
Pulmonary embolism is believed to be a common cause of death of hospital inpatients in the UK 1,2 and there is some evidence that inferior vena cava filters are effective in preventing pulmonary embolism. 3 The aims of this study were to estimate the number of deaths caused by pulmonary embolism and the potential to reduce this by the use of inferior vena cava (IVC) filters according to accepted indications (Table 1).
Accepted indications for percutaneous IVC filter placement
The Department of Health in England has stated that pulmonary embolism causes more than 25,000 preventable deaths per year in hospitals in England and Wales 2 and has therefore established this as a priority for action by the National Health Service resulting in Clinical Guidelines and Quality Standards produced by the National Institute for Health and Clinical Excellence (NICE). 4,5 However this figure is based on data that may no longer be valid. A large number of changes in practice have occurred in recent years, including improved thromboprophylaxis measures and early mobilization which are likely to have had an impact on the number of pulmonary embolism deaths.
The clinical diagnosis of pulmonary embolism is notoriously unreliable. The best previous estimates of the proportion of deaths due to pulmonary embolism in hospital have been obtained from autopsy studies. A large autopsy study performed in Sheffield showed that, between 1979 and 1983, 10% of patients admitted to hospital died and that 10% of hospital deaths were due to pulmonary embolism. 6 The VITAE study reached a similar conclusion by extrapolating from the number of diagnosed DVTs to an estimate of the number of deaths from pulmonary embolism. The study estimated that there were 543,454 venous thromboembolism-related deaths (12%) across the European Union per annum. Seventy-one percent of these were hospital acquired. 7 However the assumptions used were based on outdated diagnostic methods and hospital practices. 8 More recent autopsy studies have shown a trend towards a lower proportion of pulmonary embolism deaths. 9–11 Rubinstein et al. 9 found that 3.4% of deaths were due to pulmonary embolism at autopsy. Cohen et al. 10 demonstrated a significant reduction of fatal pulmonary embolism over a 25-year period of 6.1% to 2.1%, attributed to the introduction of thromboprophylactic measures and changes in hospital practices. Stein et al. 11 found that 5.4% of hospital deaths were due to pulmonary embolism at autopsy.
Several professional organizations have developed guidelines for use of caval filters. 12–14 We sought to establish whether caval filters, used according to guidelines, still have the potential to reduce pulmonary embolism mortality.
Methods
Autopsy reports and death notification records from 2007 and 2008 were reviewed. When pulmonary embolism was given as cause of death (in the autopsy report or in section 1 a–c or part 2 of the Medical Certificate of the Cause of Death), hospital records were reviewed for evidence of pre-mortem diagnosis of pulmonary embolism or deep vein thrombosis (DVT) and for evidence of predisposing conditions, namely malignant disease, major surgery within 30 days, past history of venous thromboembolism (VTE) and pregnancy. The total numbers of inpatient admissions and deaths were obtained from the hospital information department.
In accordance with the advice of the National Research Ethics Service (
Results
In 2007 and 2008 there were a total of 186,517 adult inpatient admissions (excluding day cases) and 2583 (1.4%) inpatient deaths of adults at this hospital of whom 696 (27%) underwent autopsy. Of those undergoing autopsy, 14 (2.0%, 95% CI 1.2–3.3%) deaths were caused by pulmonary embolism. These patients had a median age of 76.5 years (range 36–86). Risk factors for pulmonary embolism are tabulated in Table 2.
Risk factors for autopsy proven and death certificated pulmonary embolism deaths
The death notification records of the 1887 patients who did not undergo autopsy were reviewed. One hundred and fourteen records could not be traced. Of the remaining 1773 deaths, pulmonary embolism was given as cause of death (in section 1 a–c or part 2) in 12 (0.7%). Risk factors are listed in Table 3. Overall 1% of hospital inpatient deaths were attributed to pulmonary embolism.
ONS causes of death 2007
It proved impossible to retrospectively evaluate factors such as immobility, active and passive physiotherapy, use of anti-thromboembolism stockings or prophylactic heparin from the available data.
Three of the patients had accepted indications for caval filter placement: one had complications of anticoagulation and two had a known DVT or pulmonary embolism and marginal cardiopulmonary reserve.
Discussion
Autopsy data from this study show that pulmonary embolism accounts for 2% of deaths in hospital inpatients, which is considerably lower than the widely quoted figure of 10%. 1,2
Most patients who die from pulmonary embolism do not have a pre-mortem diagnosis of DVT or pulmonary embolism and many have underlying life-threatening conditions. 11 Therefore, in the absence of an autopsy, death notification records are frequently inaccurate. One study suggests that, compared with the autopsy findings, the clinical diagnosis of pulmonary embolism is incorrect as often as it is correct. 15 A weakness of our study is the small number of patients who underwent an autopsy. Pulmonary embolism is very seldom given as a cause of death on the death notification records in this institution or, indeed nationally. According to the UK Office of National Statistics, pulmonary embolism (ICD-10 code I26) was the cause of death of 6016 people (including deaths both in and out of hospital) in England and Wales in 2007 and 2008 out of a total of 1,013,142 deaths (0.6%). 16,17
Autopsy is the gold standard method of establishing cause of death due to pulmonary embolism, in almost all cases demonstrating large coiled thromboemboli in the pulmonary trunk and its major branches. Occasionally, dissection of lobar and segmental pulmonary arteries demonstrates large numbers of smaller thromboemboli. Only a minority of patients who die in hospital undergo autopsy, therefore determining the death rate due to pulmonary embolism is at best an estimate. Autopsies are performed selectively and, in England, Wales and Northern Ireland almost all at the request of the Coroner. Reasons for referral to the Coroner include unexplained, unexpected and postoperative deaths. Given the nature of death from pulmonary embolism, it seems likely that this cause of death is over-represented in the cohort undergoing autopsy. Therefore, we believe mortality rates derived from autopsy data are more likely to be over-estimates than under-estimates. It seems highly improbable that there is a selective bias against autopsy when pulmonary embolism is the cause of death.
Our autopsy data show that 2% of hospital deaths are from pulmonary embolism. This is in keeping with several other autopsy studies that have shown a progressive reduction in pulmonary embolism death rate in the last 30 years. 9–11 It suggests a substantial reduction compared with a comparable series of 30 years ago in a similar patient population, which suggested that 10% of hospital deaths were caused by pulmonary embolism. 6 The factors accounting for this are speculative but are likely to be the result of increased awareness and increased adoption of thromboprophylaxis, as well as early mobilization.
The Department of Health in England regards pulmonary embolism as a major cause of preventable death in hospital patients. 2 This has led to the adoption of VTE as a target for action by the Department of Health and NICE. A strategy of guidelines, risk assessment checklists and quality standards has been introduced in an effort to reduce it. 4,5,18
The All-Party Parliamentary Thrombosis Group, Department of Health and NICE all quote the same figure of 25,000–32,000 deaths in hospital patients in England and Wales per year. In some publications these are also described as ‘preventable’ deaths. It is likely that this figure has been derived by extrapolation from older autopsy data 6 and, more recently from the VITAE study. 7 The VITAE study used a decision tree model to estimate the burden of pulmonary embolism mortality in Europe. The probabilities used to populate the decision tree are however largely derived from outdated sources 8 which are not applicable to modern hospital practice.
There were 284,000 hospital deaths in England and Wales during 2007. 16 Assuming that the proportion of deaths due to pulmonary embolism in our autopsy cohort is representative of hospital deaths as a whole, this suggests an annual total of 5680 (95% CI 3400–9400) deaths. This is a greater than four-fold reduction compared with the accepted figure.
Many hospital patients who die from pulmonary embolism have other severe life-threatening conditions. 11,19 It is therefore unrealistic to equate deaths from pulmonary embolism with preventable deaths.
Basing NHS policy on inflated figures is unhelpful in that it distorts priorities. Indeed the All-Party Parliamentary Thrombosis Group and DOH documents explicitly state that the number of people dying from pulmonary embolism exceeds the total deaths from breast cancer, AIDS and road accidents and 25 times that of MRSA. This is not supported by Office of National Statistics figures (Table 2). 16
There is no doubt that hospitalized patients are at risk of pulmonary embolism and that thromboprohylaxis and early mobilization are important and life-saving. It seems likely that much has already been achieved in this respect and that the risk of dying form pulmonary embolism in UK hospitals has diminished substantially compared with 30 years ago.
There is limited potential for prevention of pulmonary embolism mortality through greater use of caval filters, if the accepted criteria for filter placement are applied. 12–14 Few patients in this series whose death was attributed to pulmonary embolism met accepted criteria for filter placement. More widespread application of caval filters to all patients with risk factors for VTE, such as immobility, malignant disease or major surgery would have little potential benefit, but substantial cost.
Pharmacological methods of thromboprophylaxis are not without risk, particularly of haemorrhage. The risk–benefit equation of thromboprophylaxis should not be distorted by use of exaggerated figures for pulmonary embolism deaths. Furthermore, any future evaluation of the effectiveness of the NICE guidelines will be inaccurate if the current death rate is overestimated.
Conclusion
There has been a substantial reduction in the proportion of deaths of hospital inpatients due to pulmonary embolism in the last 30 years from around 10% to 1% of deaths and from around 1% to 0.01% of admissions. Current NHS prioritization of VTE appears to be based on outdated estimates of the magnitude of the problem.
DECLARATIONS
Competing interests
None declared
Funding
None
Ethical approval
Not applicable
Guarantor
NC
Contributorship
NC, EWB and DK designed the study; DK collected the data; NC and OH wrote the manuscript; EWB, CH and OH revised the manuscript
Acknowledgements
None
Reviewer
Raman Uberoi