Can the presence of significant coagulopathy be useful to exclude symptomatic acute pulmonary embolism?
Thrombocytopaenia or an abnormal coagulation profile is not rare in hospitalised patients who have symptoms consistent with acute pulmonary embolism (PE). Theoretically, coagulopathy is more likely to occur in patients with pneumonia than acute PE. This study aimed to assess whether the presence of coagulopathy could be used to exclude acute PE in patients with symptoms and signs consistent with acute PE. In this study, a significant coagulopathy was defined as a platelet count < 100 x [10.sup.9]/l, an international normalised ratio > 1.5, or activated partial thromboplastin time >50 seconds. Patients treated with systemic anticoagulants prior to computed tomography pulmonary angiography were excluded. Of the 986 consecutive patients who required computed tomography puhnonary angiography to exclude acute PE over a four-month period in five hospitals in Western Australia, acute PE was confirmed in 149 patients (15.1%). The incidence of coagulopathy was not significantly different between those with and without acute PE (4 vs 7%, respectively; P=0.161) and between those with and without pneumonia (8 vs 7%, respectively; P=0.505). Positive and negative likelihood ratios of coagulopathy in differentiating acute PE or pneumonia were both unsatisfactory. As a continuous predictor, platelet counts, international normalised ratio, activated partial thromboplastin time and plasma fibrinogen concentrations were also not useful in differentiating between acute PE and other pulmonary pathologies (areas under the receiver operating characteristic curve were all close to 0.5). In conclusion, the presence of significant acquired coagulopathy cannot be used to suggest pneumonia or exclude symptomatic acute PE when the prevalence or pre-test probability of acute PE is not low.
Key Words: coagulation, diagnosis, pneumonia, sensitivity and specificity, venous thromboembolism
Acute pulmonary embolism (PE) is an extreme form of venous thromboembolism and is associated with significant morbidity and mortality (1,2). Various biomarkers, in particular d-dimers, have been recommended to be used to stratify the risk of patients who have symptoms of acute PE, so that unnecessary invasive diagnostic radiological studies can be mini-raised in low-risk patients (3).
Evidence suggests that thrombocytosis is associated with an increased risk of in vitro thrombotic tendency and clinical thromboembolic events including acute PE (2,4,5). On the other hand, thrombocytopaenia or an abnormal coagulation profile is associated with an increased risk of bleeding and possibly a lower risk of venous thromboembolism. Thrombocytopaenia or coagulopathy is not rare in hospitalised patients with sepsis or pneumonia who may also have symptoms such as chest pain, shortness of pain, hypotension and hypoxaemia suggestive of acute PE (6,7). Theoretically, the presence of coagulopathy should be more suggestive of a septic process, such as pneumonia, rather than acute PE. As such, the presence of coagulopathy may be useful to assist clinical decision-making about the risk of acute PE in patients who have symptoms suggestive of acute PE.
We hypothesised that significant acquired coagulopathy is more common in patients with pneumonia or other pulmonary pathologies than in patients with acute PE. In this retrospective cohort study, we assessed whether the presence of acquired coagulopathy can be useful to exclude acute PE in patients who have symptoms suggestive of acute PE.
After obtaining Royal Perth Hospital Ethics Committee approval (075/2012), clinical, laboratory and radiology data of consecutive patients who had symptoms suggestive of acute PE requiring computed tomography pulmonary angiography (CTPA) to exclude acute PE over a four-month period (January to April 2012) in five metropolitan hospitals (Royal Perth Hospital, Sir Charles Gairdner Hospital, Fremantle Hospital, Swan District Hospital and Rockingham Hospital) in Western Australia were obtained. Patients who had undergone CTPA for reasons other than to diagnose or exclude acute PE, who were treated with warfarin or unfractionated heparin for more than 24 hours prior to CTPA and those without any laboratory blood tests performed before CTPA were excluded from this study. Only blood tests performed within the 24 hours prior to CTPA were analysed for their association with risk of acute PE or pneumonia.
In this study, significant acquired coagulopathy was defined as a platelet count < 100 x [10.sup.9]/l, an international normalised ratio (INR) >1.5, or activated partial thromboplastin time (aPTT) >50 seconds. The primary outcome of this study was the incidence of coagulopathy in patients who had acute PE confirmed by CTPA, and the diagnostic ability of coagulopathy in differentiating between acute PE and other respiratory pathologies in patients who had clinical symptoms and signs suggestive of acute PE requiring CTPA to exclude acute PE. Because d-dimers have been recommended in the diagnostic work-up of patients who have symptoms suggestive of acute PE, we also assessed the diagnostic ability of d-dimers in diagnosing acute PE in this study. A d-dimers concentration >0.4 mg/l was defined as a positive d-dimers test by the laboratory. Radiologists' reports of the CTPA were considered as confirmation of the diagnosis, which might include PE, pneumonia, interstitial lung disease or other diagnosis, and all data were confirmed by the two investigators of this study before analysis.
Assuming the prevalence of acute PE for hospitalised patients who had symptoms suggestive of PE was about 10% and the incidence of significant acquired coagulopathy was 15% among those without acute PE, a sample size of 936 patients would have a power of 90% to exclude a relative risk of three in the association between coagulopathy and acute PE. A total sample size of 1000 was planned to allow for 'lost to follow-up'.
Differences in categorical and continuous variables between patients with acute PE and those without acute PE were analysed by chi-square and t-test, respectively. In this study, mortality of the patients was censored on 1 September 2012. Sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios were used to reflect the diagnostic ability of coagulopathy and d-dimers in diagnosing acute PE and pneumonia. Because the cut-points chosen in the categorical analyses may not have been optimal, the diagnostic ability of platelet counts, INR, aPTT and fibrinogen concentrations were also assessed by the area under the receiver operating characteristic curve by considering these parameters as a continuous variable. All analyses were conducted by SPSS for Windows (version 19.0, 2011, IBM, USA) and Confidence Interval Analysis (version 2.0.0, 2000, BMJ, UK) and a P value <0.05 was taken as significant.
Of the 986 patients included in the study, acute PE was confirmed in 149 patients (15%; Figure 1). The age, sex and presenting symptoms were not significantly different between those who had acute PE and those who did not. The d-dimers concentrations (3.6 vs 1.5 mg/l, P=0.001) and the proportion of patients with a positive d-dimers test (99 vs 90%, P=0.019) were significantly higher among patients who had acute PE than those without acute PE (Table 1 and Figure 2). A total of 67 patients (6.8%) had some form of coagulopathy. The median platelet counts of those with thrombocytopaenia were 62 x [10.sup.9]/l (interquartile range 38-83 x [10.sup.9]/l). The median INR for those with an abnormal [NR was 1.9 (interquartile range 1.8-2.5) and the median aPTT for those with an abnormal aPTT was 65 seconds (interquartile range 52-82). The incidence of significant coagulopathy was, however, not significantly different between patients with acute PE and those without acute PE (4 vs 7%, respectively; P=0.161).
Although the specificity and negative predictive value of coagulopathy in relation to acute PE were high, both the negative and positive likelihood ratios were not substantially different from one (Table 2), suggesting that coagulopathy could not be used reliably to change the post-test probability of acute PE in patients with signs and symptoms consistent with acute PE. A positive d-dimers test had a high sensitivity and low negative likelihood ratio for acute PE, but both a positive d-dimers test and coagulopathy were not useful in diagnosing pneumonia in patients with symptoms suggestive of acute PE.
As a continuous predictor, only d-dimers concentrations (area under receiver operating characteristic curve=0.76, 95% confidence interval: 0.70-0.83), but not platelet counts, INR or aPTT (areas under the receiver operating characteristic curve all close to 0.5), had a reasonably good discriminative ability to differentiate between acute PE and other pulmonary pathologies (Table 3).
Our results showed that significant acquired coagulopathy was not rare among hospitalised patients who presented with signs and symptoms of acute PE, but the presence of coagulopathy could not be reliably used to exclude acute PE or suggest pneumonia. We also showed that d-dimers concentrations had a reasonable ability to differentiate acute PE and other pulmonary pathologies, but a positive standard d-dimers test was extremely common among hospitalised patients, even without acute PE. These results have some clinical significance and require careful consideration.
First, thrombocytopaenia or an abnormal coagulation profile in INR or aPTT have traditionally been associated with an increased risk of bleeding, hence it would appear counter-intuitive to observe that patients with significant acquired coagulopathy could present with venous thromboembolism including acute PE. However, there is increasing evidence that patients with an abnormal coagulation profile can have increased risk of both bleeding and thrombotic complications, including patients with liver cirrhosis (8). Recent evidence suggested that a prolonged INR and aPTT do not exclude a deficiency in anticoagulants such as protein C or an increased thrombotic tendency (9). Furthermore, venous thromboembolism can also be contributed to by other factors such as venous stasis and endothelial injury (10). Both of these factors are prevalent in many hospitalised patients, including those with coagulopathy from infection or trauma.
Second, our recent observational study showed that omission of early thromboprophylaxis in patients who were considered to have contraindications to pharmacological thromboprophylaxis was associated with an increased risk of mortality in critically ill patients (11). Although this result could be explained by residual confounding, it is also possible that an increased risk of venous thromboembolism might have, in part, contributed to the higher risk of mortality for those with coagulopathy and omission of early thromboprophylaxis (12). Further analysis of our previous study on in-vitro thrombotic tendency of critically ill patients showed that 37% of the patients who had a raised INR or aPTT were, in fact, associated with an increased risk of in-vitro thrombotic tendency as measured by thromboelastography (5). Whether pharmacological thromboprophylaxis should be routinely used for patients who have acquired coagulopathy remains, however, uncertain. Recent evidence suggests that thromboelastography may be useful to guide thromboprophylaxis when the balance between thrombotic and bleeding risk in a patient is unpredictable (13,14).
Third, the utility of a biomarker can vary substantially, depending on the pre-test probability of the outcome of interest (15). Although d-dimers concentration has been recommended to exclude venous thromboembolism including acute PE in low-risk patients (3), our results showed that a standard positive d-dimers test is unlikely to be useful in hospitalised patients with symptoms suggestive of acute PE. This is due to the fact that most patients with pulmonary pathologies other than acute PE also had a raised d-dimers concentration (false positive rate 90%) and hence the utility of a negative d-dimers test to reduce utilisation of CTPA is low. Nevertheless, the usefulness of a higher d-dimers concentration threshold (e.g. >2 mg/l) to differentiate between acute PE and other pulmonary pathologies in hospitalised patients with symptoms suggestive of acute PE remains uncertain and this merits further investigation by a large prospective cohort study.
The last consideration is the limitations of the study. First, we have used CTPA as a gold standard in defining acute PE in this study and some patients with small sub-segmental PE might have been misclassified as not having acute PE. Second, we did not have data on causes of coagulopathy in our patients. It is possible that different causes of coagulopathy may have different effects on risk of thromboembolism. Finally, the number of patients with coagulopathy (6.8%), especially those with severe coagulopathy (e.g. platelet count <50 x [10.sup.9]/l, INR >2.0 and aPTT >60 seconds), was relatively small and, as such, we could not exclude a type II error in concluding that coagulopathy was not associated with a reduced risk of acute PE. The association between risk of acute PE and severe coagulopathy, stratified by either thrombocytopaenia or raised INR/aPTT, merits further investigation.
In summary, the presence of significant acquired coagulopathy cannot be used to suggest pneumonia or exclude symptomatic acute PE when the prevalence or pre-test probability of acute PE is not low. Although acute PE was associated with a higher d-dimers concentration than other pulmonary pathologies, a traditional positive d-dimers concentration (>0.4 mg/l) was not useful in excluding acute PE in hospitalised patients with symptoms suggestive of acute PE. Whether a higher threshold of d-dimers concentration can be useful for this type of patient remains uncertain and this merits further investigation.
This study was solely funded by the Department of Intensive Care Medicine, Royal Perth Hospital.
Caption: Figure 1: Flow chart showing inclusion and exclusion of patients in the study.
Caption: Figure 2: Difference in d-dimers concentrations between acute pulmonary embolism (PE) and other pulmonary pathologies. An abnormally high d-dimers concentration was prevalent (90%) among those without PE. Error bars signify 95% confidence interval.
(1.) Ho KM, Burrell M, Rao S, Baker R. Incidence and risk factors for fatal pulmonary embolism after major trauma: a nested cohort study. Br J Anaesth 2010; 105:596-602.
(2.) Ho KM, Chavan S. Prevalence of thrombocytosis in critically ill patients and its association with symptomatic acute pulmonary embolism. Thromb Haemost 2013; 109:272-279.
(3.) Torbicki A, Perrier A, Konstantinides S, Agnelli G, Galie N, Pruszczyk Pet al. Guidelines on the diagnosis and management of acute pulmonary embolism: the Task Force for the Diagnosis and Management of Acute Pulmonary Embolism of the European Society of Cardiology (ESC). Eur Heart J 2008; 29:2276-2315.
(4.) Ho KM, Yip CB, Duff O. Reactive thrombocytosis and risk of subsequent venous thromboembolism: a cohort study. J Thromb Haemost 2012; 10:1768-1774.
(5.) Duff OC, Ho KM, Maybury SM. In vitro thrombotic tendency of reactive thrombocytosis in critically ill patients: a prospective case-control study. Anaesth Intensive Care 2012; 40:472-478.
(6.) Walsh TS, Stanworth SJ, Prescott RJ, Lee RJ, Watson DM, Wyncoll D. Prevalence, management, and outcomes of critically ill patients with prothrombin time prolongation in United Kingdom intensive care units. Crit Care Med 2010; 38:1939-1946.
(7.) Tufano A, Guida A, Di Minno MN, Prisco D, Cerbone AM, Di Minno G. Prevention of venous thromboembolism in medical patients with thrombocytopenia or with platelet dysfunction: a review of the literature. Semin Thromb Hemost 2011; 37:267-274.
(8.) Senzolo M, Sartori MT, Lisman T. Should we give thromboprophylaxis to patients with liver cirrhosis and coagulopathy? HPB (Oxford) 2009; 11:459-464.
(9.) Tripodi A, Chantarangkul V, Mannucci PM. Acquired coagulation disorders: revisited using global coagulation/anticoagulation testing. Br J Haematol 2009; 147:77-82.
(10.) Bagot CN, Arya R. Virchow and his triad: a question of attribution. Br J Haematol 2008; 143:180-190.
(11.) Ho KM, Chavan S, Pilcher D. Omission of early thromboprophylaxis and mortality in critically ill patients: a multicenter registry study. Chest 2011; 140:1436-1446.
(12.) Park MS, Martini WZ, Dubick MA, Salinas J, Butenas S, Kheirabadi BS et al. Thromboelastography as a better indicator of hypercoagulable state after injury than prothrombin time or activated partial thromboplastin time. J Trauma 2009; 67:266-275.
(13.) Boyce H, Hume-Smith H, Ng J, Columb MO, Stocks GM. Use of thromboelastography to guide thromboprophylaxis after caesarean section. Int J Obstet Anesth 2011; 20:213-218.
(14.) Attaran S, Somov R Awad WI. Randomised high- and low-dose heparin prophylaxis in patients undergoing thoracotomy for benign and malignant disease: effect on thrombo-elastography. Eur J Cardiothorac Surg 2010; 37:1384-1390.
(15.) Ho KM. Ten commandments of interpreting and applying results of biomarker research. Anaesth Intensive Care 2011; 39:799-801.
K. M. HO *, J. A. TAN ([dagger])
Intensive Care Unit, Royal Perth Hospital, Perth, Western Australia, Australia
* MB, BS, Postgrad Dip (Echo), MPH, PhD, FRCP(Glasg), FCICM, FANZCA, Consultant Intensivist.
([dagger]) MB, BS, Anaesthetic Registrar.
Address for correspondence: Clinical Associate Professor K. M. Ho, Intensive Care Unit, Royal Perth Hospital, Wellington Street. Perth, Western Australia, Australia 6000. Email: firstname.lastname@example.org
Accepted for publication on February 20, 2013.
Table 1 Characteristics of the patients (n=986) with symptoms requiring computed tomography pulmonary angtography to exclude acute pulmonary embolism Patients with Variable PE (n=149) Age, y (SD) 62 (16) Male, number (%) 82 (55) Symptoms, number- (%) Chest pain 78 (52) Shortness of breath 69 (46) Hypotension 12 (8) Hypoxaemia 40 (27) Other reasons (calf swelling/pain, 65 (44) haemoptysis, + d-dimers) INR (SD) 1.1 (0.2) aPTT sec (SD) 33 (6) Fibrinogen, g/l (SD) 5.0 (2) Platelets, x [10.sup.9]/l 262 (117) Coagulopathy, number (%) * 6 (4) Severe coagulopathy, number (%) 2 (1.3) ([dagger]) Positive d-dimers ([double dagger]), 64 (99) number (%) Mortality, number (%) 19 (13) Patients without Variable PE (n=837) P value Age, y (SD) 61 (17) 0.610 Male, number (%) 395 (47) 0.091 Symptoms, number- (%) Chest pain 507 (61) 0.070 Shortness of breath 390 (47) 0.999 Hypotension 64 (8) 0.868 Hypoxaemia 174 (21) 0.106 Other reasons (calf swelling/pain, 312 (37) 0.144 haemoptysis, + d-dimers) INR (SD) 1.2 (0.5) 0.001 aPTT sec (SD) 34 (9) 0.056 Fibrinogen, g/l (SD) 4.9 (2) 0.510 Platelets, x [10.sup.9]/l 259 (112) 0.744 Coagulopathy, number (%) * 61 (7) 0.161 Severe coagulopathy, number (%) 5 (0.6) 0.235 ([dagger]) Positive d-dimers ([double dagger]), 366 (90) 0.019 number (%) Mortality, number (%) 83 (10) 0.307 * Defined by either a low platelet count (<100 x [10.sup.9]/l), high INR (>1.5) or prolonged aPTT (>50 seconds). ([dagger]) Defined by having at least two of the following coagulation abnormalities: a low platelet count (<100 x [10.sup.9]/l), high INR (>1.5), or prolonged aPTT (>50 seconds). ([double dagger]) D/dimers results were available before CTPA only in 472 patients and a positive d/dimers result was defined by d/dimers concentrations greater than 0.4 mg/l. PE = pulmonary embolism, SD = standard deviation, INR = international normalised ratio, aPTT = activated partial thromboplastin time. Table 2 Sensitivity, specificity, likelihood ratios and predictive values of significant coagulopathy * in excluding acute pulmonary embolism or predicting pneumonia in patients with symptoms of acute pulmonary embolism requiring computed tomography pulmonary angiography compared to a positive d/dimers result (>0.4 mg/l). Diagnosis Sensitivity Specificity Significant All PE (n=149) 4% 93% coagulopathy Bilateral PE (n=91) 6% 91% Saddle PE (n=10) 0% 93% DVT (n=32) 9% 93% Pneumonia (n=173) 8% 94% Positive All PE (n=65) 99% 10% d-dimers Bilateral PE (n=36) 97% 9% Saddle PE (n=4) 99% 9% DVT (n=7) 99% 9% Pneumonia (n=61) 95% 10% Diagnosis +PV -PV +LR -LR Significant All PE (n=149) 9% 84% 0.6 1.04 coagulopathy Bilateral PE (n=91) 4% 93% 0.6 1.03 Saddle PE (n=10) 0% 99% 0 1.08 DVT (n=32) 5% 97% 1.4 0.97 Pneumonia (n=173) 21% 83% 1.2 0.98 Positive All PE (n=65) 15% 98% 1.1 2 d-dimers Bilateral PE (n=36) 8% 98% 1.1 0.3 Saddle PE (n=4) 1% 99% 1.1 0.1 DVT (n=7) 2% 99% 1.1 0.1 Pneumonia (n=61) 13% 93% 1.1 0.5 * Defined by a low platelet count (<100 x [10.sup.9]/l), high international normalised ratio (>1.5) or prolonged activated partial thromboplastin time (aPTT >50 seconds). PV = predictive values, LR = likelihood ratios, PE = pulmonary embolism, DVT = deep vein thrombosis confirmed by Doppler ultrasound. Table 3 Area under receiver operating characteristic curve for platelet counts, coagulation profiles and d-dimers in predicting symptomatic acute pulmonary embolism Variable Area under the receiver operating characteristic curve All PE Bilateral PE Saddle PE (n=149) (n=91) (n=10) Platelet 0.54 (0.47-0.61) 0.57 (0.48-0.65) 0.47 (0.30-0.63) counts INR 0.50 (0.43-0.57) 0.52 (0.44-0.60) 0.45 (0.32-0.58) aPTT 0.48 (0.41-0.54) 0.43 (0.35-0.51) 0.42 (0.30-0.53) Fibrinogen 0.52 (0.45-0.60) 0.50 (0.42-0.58) 0.44 (0.27-0.60) D-dimers 0.76 (0.70-0.83) 0.50 (0.42-0.58) 0.89 (0.74-0.99) PE = pulmonary embolism, INR = international normalised ratio, aPTT = activated partial thromboplastin time.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Original Papers|
|Author:||Ho, K.M.; Tan, J.A.|
|Publication:||Anaesthesia and Intensive Care|
|Date:||May 1, 2013|
|Previous Article:||The prevalence of anaemia, hypochromia and microcytosis in preoperative cardiac surgical patients.|
|Next Article:||Effects of dexmedetomidine infusion on laryngeal mask airway removal and postoperative recovery in children anaesthetised with sevoflurane.|