Clinical probability score and D-dimer estimation lack utility in the diagnosis of childhood pulmonary embolism

Transcription

1 Journal of Thrombosis and Haemostasis, 7: DOI: /j x ORIGINAL ARTICLE Clinical probability score and D-dimer estimation lack utility in the diagnosis of childhood pulmonary embolism T. T. BISS,* L. R. BRANDÃO,*W.H.A.KAHR,*A.K.C.CHAN and S. WILLIAMS* *Department of Hematology/Oncology, The Hospital for Sick Children, Toronto, ON; and Department of Hematology/Oncology, McMaster University, Hamilton, ON, Canada To cite this article: Biss TT, Brandão LR, Kahr WHA, Chan AKC, Williams S. Clinical probability score and D-dimer estimation lack utility in the diagnosis of childhood pulmonary embolism. J Thromb Haemost 2009; 7: Summary. Background: Childhood pulmonary embolism (PE) causes significant mortality and evidence suggests that it is under-diagnosed. Clinical probability scores and D-dimer estimation to assess pre-test probability have not been studied in children with suspected PE. Patients/Methods: This retrospective cohort study evaluated Wells simplified probability score for PE in 50 children with PE and 25 control patients, and D-dimer values in 27 and 12 PE negative children. Results: and groups had similar rates of risk factors for venous thromboembolism (VTE). Wells simplified probability score showed a small difference between and children (median score:, 4.5;, 4; P =0.009), children with PE are more likely to obtain a ÔPE likelyõ score (score > 4), P = The difference was of slightly greater significance when the Wells score was adjusted to account for pediatric normal ranges for heart rate, P = 0.007, and signs/ symptoms of upper limb DVT, P = Children with PE were as likely as patients to have a D-dimer value within the normal range (, 15%;, 25%; P = 0.654). A combination of a ÔPE unlikelyõ score and normal D-dimer value occurred in 1/12 (8%) of children. Conclusions: The Wells clinical probability score and D-dimer estimation may lack utility in the determination of pre-test probability of PE in children. Validation of a pediatric clinical probability score, incorporating D-dimer estimation, by prospective study, would be difficult as a result of the rarity of childhood PE. Keywords: D-dimer, diagnosis, pediatrics, pulmonary embolism. Introduction Childhood PE occurs infrequently, registry data reporting an incidence of 0.86 per hospital admissions [1]. Most events occur in children with serious underlying disorders. The incidence of deep venous thrombosis (DVT) in childhood is rising with the increasingly successful treatment of previously fatal disorders such as severe congenital cardiac defects, prematurity and malignancy [2], coupled with the increasing use of central venous catheters that are also known to be a provoking factor [3]. Mortality rate as a result of PE in childhood is reported to be around 10% [1,4,5] and the long-term impact on pulmonary and cardiac function in survivors remains unknown. Evidence of thromboembolism at autopsy in up to 4.2% of children [6] suggests under-diagnosis, likely resulting from the masking of symptoms of PE by those of the underlying disorder accompanied by a low index of suspicion in the treating physician. Individual symptoms and signs have low sensitivity and specificity for predicting a diagnosis of PE. However, the combination of a number of symptoms and signs into a clinical probability score allows for the determination of pre-test probability in adults presenting with PE. This has been shown to accurately predict the findings on subsequent imaging studies [7]. The combination of a clinical probability score and the results of D-dimer estimation improves the negative predictive value further and may safely exclude PE without the need for diagnostic imaging [8]. Neither of these approaches have been studied in children with clinically suspected PE. The aim of this retrospective cohort study was to evaluate the utility of the Wells simplified probability score for PE and D-dimer estimation in the diagnosis of PE occurring in childhood. Correspondence: Suzan Williams, Department of Hematology/ Oncology, The Hospital for Sick Children, 555 University Ave, Toronto, ON M5G 1X8, Canada. Tel.: ; fax: suzan.williams@sickkids.ca Received 26 February 2009, accepted 27 July 2009 Patients and methods Identification of cases of childhood PE () Ethical approval for the study was obtained from the Research Ethics Board at The Hospital for Sick Children, Toronto, ON, Canada. Children with radiologically proven PE were

2 1634 T. T. Biss et al identified by searching for the term Ôpulmonary embolismõ in electronic reports of relevant diagnostic investigations: ventilation perfusion (V/Q) scanning; spiral CT pulmonary angiography (CTPA); conventional pulmonary angiography; and echocardiography (ECHO). Institutional criteria for PE were: (i) V/Q scan High probability scan; (ii) CTPA Intraluminal filling defect in a lobar or main pulmonary artery; (iii) Pulmonary angiography Intraluminal filling defect in a lobar or main pulmonary artery; and (iv) ECHO Pulmonary artery obstruction [9]. Identification of patients Twenty-five patients, who were age and sex matched to the cases, were selected from a list of children who had undergone V/Q scanning for investigation of suspected PE. All had scans reported as Ôlow probabilityõ for PE. In order to reduce the possibility of including patients with false-negative V/Q scans, case notes referring to the period after V/Q scanning were carefully reviewed to ensure an uneventful follow-up period without anticoagulation. Sample size for the PE negative patients was based on feasibility and did not include all patients investigated for PE. Data collection Collection of data was by retrospective review of patient case notes and included demographic details, details of underlying disorder(s), risk factors for venous thromboembolism (VTE) and presenting symptoms and signs. A patient was considered to have had a previous DVT if the event was objectively diagnosed prior to ordering diagnostic imaging for PE. D- dimer values obtained on blood samples taken at the time of presentation were recorded. Wells simplified probability score for PE Wells simplified probability score for PE was applied retrospectively to and children who were designated ÔPE likelyõ if their score was > 4 and ÔPE unlikelyõ if their score was 4 (Table 1) [8]. The subjective element of Table 1 Wells simplified probability score for PE [8] Clinical signs and symptoms of DVT (minimum of leg swelling and pain with palpation of the deep veins) An alternative diagnosis is less likely than PE Heart rate > 100 Immobilisation* or surgery in the previous 4 weeks Previous DVT/PE Hemoptysis Malignancy (on treatment, treated in the last 6 months or palliative) [3.0 points] [3.0 points] [1.5 points] [1.5 points] [1.5 points] [1.0 points] [1.0 points] Score > 4, PE likely; Score 4, PE unlikely. DVT, deep venous thrombosis; PE, pulmonary embolism. *Immobilization refers to: complete bed rest of 3 or more days duration; or, lower extremity cast. Refers to an objectively diagnosed event. whether Ôan alternative diagnosis was less likely than PEÕ was determined from comments written in the case notes by the responsible clinician. The clinician was not blinded to the outcome, but the comments were made prior to the results of the diagnostic imaging. Symptoms and signs of upper venous system DVT were not included in the scoring system but were recorded separately for analysis. As a secondary analysis, for children 10 years of age the score was adjusted for heart rate. A score of 1.5 points was achieved if heart rate was: > 160 bpm in children < 1 year of age; > 120 bpm in children 1 10 years of age; and, > 100 bpm in children > 10 years of age. These values were based on published age-adjusted normal ranges for heart rate [10]. D-dimer measurement D-dimer was measured by quantitative assay at the time of presentation with symptoms and/or signs of PE. D-dimer values were classified as normal or above normal. The cut-off values for the various assays were: Turbiquant Ò D-dimer (Dade Behring, Deerfield, IL, USA), September 1999 to July 2003, normal value < 250 mcg L )1 ; D-dimer PLUS (Dade Behring), July 2003 to December 2005, normal value < 250 mcg L )1 ; HemosIL TM D-dimer HS (Instrumentation Laboratory, Richmond Hill, ON, Canada), December 2005 to January 2007, normal value < 300 ng ml )1. Statistical analysis Patient characteristics were compared between and groups using the Mann Whitney U-test for unpaired non-parametric data and the FisherÕs exact test for the analysis of binomial variables. Wells score was compared between and groups using the Mann Whitney U-test. Use of the Wells simplified probability score to classify patients as ÔPE likelyõ or ÔPE unlikelyõ, and D-dimer measurement, was compared using FisherÕs exact test. P-values of < 0.05 were considered significant. Statistical analyses were completed on GraphPad Instat software (La Jolla, CA, USA), copyrighted ( ). Results The group comprised 50 children with objectively diagnosed PE who had presented to The Hospital for Sick Children over an 8-year period from September 1999 to January Imaging modalities used to diagnose PE were as follows: V/Q scanning, 20 patients (40%); CTPA, 12 (24%); combination of V/Q scanning and CTPA, 7 (14%); cardiac angiography, 7 (14%); conventional pulmonary angiography, 2 (4%); and ECHO, 2 (4%). The control group comprised 25 children who had presented during the same time period and who were investigated by V/Q scanning for suspected PE but had negative results. A comparison of PE positive and groups according to demographic features, underlying diagnosis and risk factors for VTE is

3 Clinical probability score in childhood PE 1635 Table 2 Comparison of and groups according to demographic features, underlying diagnosis and risk factors for VTE n =50 n =25 P value Gender Male 27 (54) 14 (56) 1.0* Female 23 (46) 11 (44) Age Median (range), years 13 ( ) 12 (1 17) Underlying diagnosis Malignancy 9 (18) 10 (40) 0.051* Congenital cardiac disease 9 (18) 1 (4) 0.15* Risk factors for VTE Recent surgery à 12 (24) 5 (20) 0.777* Immobilisation 28 (56) 13 (52) 0.808* Previous DVT/PE 18 (36) 10 (40) 0.802* Central venous line 17 (34) 12 (48) 0.316* DVT, deep venous thrombosis; PE, pulmonary embolism; VTE, venous thromboembolism. *FisherÕs exact test. Mann Whitney U-test. à Recent surgery refers to surgery within the previous 4 weeks. Immobilization refers to: complete bed rest of 3 or more days duration; or, lower extremity cast. Refers to an objectively diagnosed event. shown in Table 2. Presenting features were compared between and groups as shown in Table 3. There was a small but significant difference in Wells simplified probability score between the and PE negative groups (median score:, 4.5;, 4; P = 0.009) (Fig. 1). children were more likely to obtain a PE likely score (score > 4) and children were more likely to obtain a PE unlikely score (score 4) (P = 0.012) (Table 4). The Wells simplified probability score was adjusted to account for pediatric normal ranges for heart rate (see ÔPatients and methodsõ). This modification to the Wells score resulted in slightly greater significant difference between the and groups (P =0.007). Table 3 Comparison of and groups according to presenting features Presenting feature [No. of patients (%)] n =50 n =25 P value Symptoms and signs of 11 (22) 2 (8) 0.198* lower limb DVT Symptoms and signs of 2 (4) 0 (0) 0.550* upper venous system DVT Hemoptysis 2 (4) 1 (4) 1.0* Pleuritic chest pain 16 (32) 6 (24) 0.594* Tachycardia: Heart 32 (64) 13 (52) 0.454* rate > 100 bpm Tachycardia: Age-adjusted 27 (54) 9 (36) 0.220* Hypoxia: O 2 saturations < 94% on air 18 (36) 12 (48) 0.454* DVT, deep venous thrombosis. *FisherÕs exact test. Age-adjusted heart rate was classified as follows: > 160 bpm in children < 1 year of age; > 120 bpm in children 1 10 years of age; > 100 bpm in children > 10 years of age [10]. Wells simplified probability score for PE Fig. 1. Comparison of pulmonary embolism (PE) positive and groups according to Wells simplified probability score for PE [8]. Horizontal bars represent median values (, 4.5;, 4.0; P = 0.009, Mann Whitney U-test). Table 4 Comparison of and groups according to Wells simplified probability score PE likely: score > 4 PE unlikely: score 4 n =50 n =25 36 (72) 10 (40) 14 (28) 15 (60) PE, pulmonary embolism. P = using FisherÕs exact test. Two children presented with symptoms and signs of upper venous system DVT (Table 3). One child had left arm pain and swelling, occurring in the presence of linerelated non-occlusive thrombus in the left axillary, subclavian and internal jugular veins confirmed by Doppler ultrasonography. A second child had right-sided chest and facial swelling, occurring in the presence of occlusive thrombus in the right internal jugular vein confirmed by computerized tomography (CT) of the chest. These children had Wells scores of 4 and 4.5, respectively. Inclusion of a score of 3.0 points for clinical signs and symptoms of upper venous system DVT in these patients made only a slight improvement in the performance of the Wells score in distinguishing children from PE negative children (P = 0.006). D-dimer values at the time of presentation with suspected PE were available for 27 and 12 children (Table 5). children were as likely as children to have a normal D-dimer value (P = 0.654). Laboratory results in the four children with a normal D- dimervaluewere:181 mcg L )1 (normalrange:< 250 mcg L )1 ); 238 mcg L )1 (< 250 mcg L )1 ); < 200 mcg L )1 (< 250 mcg L )1 ); and 295 mcg L )1 (< 300 mcg L )1 ). The combination of the Wells simplified probability score and D-dimer value is compared between and groups in Table 6.

4 1636 T. T. Biss et al Table 5 Comparison of and groups according to D-dimer value Above normal D-dimer Normal D-dimer Discussion n =27 n =12 23 (85) 9 (75) 4 (15) 3 (25) P = using FisherÕs exact test. D-dimer values at the time of presentation were available for 27/50 children and 12/25 children. Table 6 Comparison of and groups according to the combination of Wells simplified probability score and D-dimer value PE likely + above normal D-dimer PE likely + normal D-dimer PE unlikely + above normal D-dimer PE unlikely + normal D-dimer n =27 n =12 16 (59) 4 (33) 2 (7) 2 (17) 7 (26) 5 (42) 2 (7) 1 (8) PE likely, Score > 4; PE unlikely, Score 4. D-dimer values at the time of presentation were available for 27/50 children and 12/25 children. Accurate diagnosis of childhood PE is increasing in relevance as a result of its likely rising incidence, the significant risk of mortality and the potential for long-term morbidity. Evidence suggests that childhood PE is an under-diagnosed event so improvement in the determination of pre-test probability for PE in children would be welcomed. A commonly used clinical probability score in adults is the Wells simplified probability score [8] which is a sum of weighted scores from 1.0 to 3.0 for the presence of seven variables including a subjective assessment of whether PE is the most likely diagnosis (Table 1). Validation of the Wells simplified score in adults showed rates of PE for low, intermediate and high clinical probability groups to be 2%, 18.8% and 50%, respectively [8]. Subsequent prospective studies have confirmed the safety of combining a low or moderate pre-test probability, using a simple clinical model, with negative D-dimer assay to rule out PE without the need for diagnostic imaging [11,12]. The Wells simplified probability score is limited by its subjective element, a potentially greater problem when used by inexperienced physicians, and has not been validated for use in pediatric practice. This study involved the retrospective analysis of 50 PE positive children and 25 age- and sex-matched children. Applying the Wells simplified probability score for PE retrospectively showed a difference between the proportion of patients allocated to the PE unlikely group (28% of vs. 60% of children), and in Wells scores (median score of 4.5 and 4 for and children, respectively). The score did not perform as well in children as it does in adults and the possible reasons for this are outlined below. The and groups were well matched for age and sex and had similar frequencies of risk factors (recent surgery, immobilization and the presence of a central venous line) associated with PE, identifying common features in children who tend to be investigated for suspected PE. Scores obtained for these categories of the Wells score were therefore similar between the and groups. The remaining score components of signs/symptoms of lower limb DVT, subjective suspicion of PE as the most likely diagnosis, hemoptysis and tachycardia were not sufficiently discriminatory in identifying children with PE (Table 3). In general, clinicians have a low suspicion for PE in children. PE was felt to be the most likely diagnosis in only 58% (29/50) of children with PE and in 28% (7/25) of children. The subjective nature of this criterion has previously been criticized in adult practice [13] and appears to be similarly problematic in children. A number of patients had co-existent DVT. In addition to the 11 children with confirmed PE who had signs/symptoms of lower limb DVT, a further two had signs/symptoms of DVT in the upper venous system. Seventeen children had asymptomatic DVT (including eight with lower limb DVT, three with upper venous system DVT, five with renal vein/inferior vena cava thrombosis, and one with right atrial mural thrombus). Of the two children with signs/symptoms of upper venous system DVT, one had a ÔPE unlikelyõ Wells score (4) and the other a ÔPE likelyõ score (4.5). In view of the relatively high incidence of upper venous system DVT in children when compared with adults [1], adjustment of the Wells score to include signs/ symptoms of upper limb or neck DVT may improve the predictive ability of this score in a larger cohort of children. Tachycardia occurred in similar frequencies between the PE positive and groups (64% vs. 52%, respectively) and showed only a small difference when adjusted for agerelated normal ranges (54% vs. 36% for and PE negative children, respectively). The high frequency of tachycardia in both and children is likely to reflect the prevalence of underlying illness. In the present study, D-dimer estimation was non-discriminatory. Fifteen per cent of children with PE had normal values. In our previously reported cohort of children subsequently diagnosed with PE, D-dimer was elevated in 26/30 (86.7%) of patients, exceeding five times the upper limit of the normal range in 23/26 (76.7%) [5]. In a cohort of 12 pediatric patients and 2 adult patients subsequently diagnosed with PE, the Dade Behring D-dimer assay was elevated in 6/10 (60%) of patients [14]. D-dimers are known to be elevated in a number of systemic illnesses, such as vasculitis and pneumonia [15 17]. The similar rates of elevated D-dimers in the and groups reflects the high rate of underlying systemic illnesses in children presenting to this institution with suspected PE.

5 Clinical probability score in childhood PE 1637 There are some limitations to the findings of this study in relation to the use of D-dimer estimation. First, D-dimer values were only available for approximately 50% of the and groups (27/50 and 12/25, respectively). Selection bias may have accounted for the high rate of elevated D-dimer values in the group, D-dimers being testedonlyinthosefelttobeatgreatestriskofpebecauseof the presence of co-morbidities. Second, three different D-dimer assays were used during the study period. These are known to have differing sensitivities for PE in adults and specific cut-off values have not been validated in children, limiting any comparisons that could be made between the and groups. In adults, the combination of a clinical probability score and D-dimer measurement may offer the safest management strategy for patients [18]. Pulmonary embolism can be safely ruled out in adults with a PE unlikely score and normal D- dimer assay without the need for diagnostic imaging [11,12]. Our study showed that the combination of a PE unlikely score and normal D-dimer value occurred in only 1/12 (8%) of the group suggesting that, even if the combination was able to identify a cohort of patients not requiring further diagnostic imaging, the low frequency of occurrence would limit itsõ usefulness in clinical practice. The potential advantages of determining pre-test probability of PE in children using a clinical probability score and/or D- dimer estimation include the reduction in the use of V/Q and CT scanning with a consequent reduction in cost and exposure to ionizing radiation and radiographic contrast. A further reduction in cost is possible by avoiding admission of children who present as outpatients with symptoms of PE. Both of these depend on the existence of a validated approach with a high negative predictive value in order to avoid false negatives and therefore discharge of children from hospital who will later present with PE. It would appear that the Wells simplified probability score for PE lacks utility in children and itsõ discriminatory value is only slightly improved by minor modifications to account for physiological differences in normal heart rate and increased frequency of upper venous system DVT. Quantitative D-dimer estimation lacks specificity for childhood PE, likely as a result of to the presence of other conditions associated with elevated D-dimers in the investigated population. Given the limitations of the Wells score and D-dimer estimation, combining the two does not appear to offer an enhancement to clinical decision making for investigation of childhood PE in the hospital setting. The predictive value of a combination of the Wells clinical probability score and D-dimer estimation may, however, have greater relevance to children presenting to primary care physicians as these children are likely to have a lower prevalence of underlying illness. There are some further limitations to this study. Data were collected retrospectively, making the criterion that refers to the subjective opinion of whether PE was the most likely diagnosis difficult to define. In addition, the presence of multiple comorbidities in the ÔÕ group raises the possibility that some of these patients had small PEs that were not detected by V/Q scanning. Although some of the patients in this group had other imaging studies to rule out PE, including CT scanning, many of them had only a negative V/Q scan. Not all children who had presented to this institution with suspected PE were included in this study; a larger group, including patients with PE ruled out by a negative CT, would have decreased the potential impact of missed PE. Conclusions The results of this study suggest that neither the Wells simplified probability score or D-dimer estimation are likely to be useful tools in determining pre-test probability of PE in children, particularly in the tertiary care setting, despite modification of the score to apply to pediatric practice. They appear to lack the ability to identify a cohort of children with suspected PE who can safely forego further investigation and are therefore not likely to be cost-effective. The retrospective nature of this study introduces a number of limitations. Prospective study of a large number of children presenting consecutively with suspected PE would be required in order to evaluate accurately the utility of the current or pediatric modified Wells score and D-dimer estimation in children. As a result of the rarity of childhood PE, a study of this nature would be difficult. Acknowledgements The authors thank P.S. Babyn for assistance with identifying cases and control patients. T.T. Biss acknowledges Baxter Bioscience Canada for sponsoring her fellowship post at The Hospital for Sick Children. A.K.C. Chan is a Career Investigator funded by the Heart and Stroke Foundation of Canada. Disclosure of Conflict of Interests The authors state that they have no conflict of interest. References 1 Andrew M, David M, Adams M, Ali K, Anderson R, Barnard D, Bernstein M, Brisson L, Cairney B, DeSai D, Grant R, Israels S, Jardine L, Luke B, Massicotte P, Silva M. Venous thromboembolic complications (VTE) in children: first analyses of the Canadian registry of VTE. Blood 1994; 83: Sandoval JA, Sheehan MP, Stonerock CE, Shafique S, Rescorla FJ, Dalsing MC. Incidence, risk factors, and treatment patterns for deep venous thrombosis in hospitalized children: an increasing population at risk. J Vasc Surg 2008; 47: Massicotte MP, Dix D, Monagle P, Adams M, Andrew M. Central venous catheter related thrombosis in children: analysis of the Canadian registry of venous thromboembolic complications. JPediatr1998; 133: van Ommen CH, Heijboer H, Buller HR, Hirasing RA, Heijmans HS, Peters M. Venous thromboembolism in childhood: a prospective twoyear registry in the Netherlands. JPediatr2001; 139:

6 1638 T. T. Biss et al 5 Biss TT, Brandao LR, Kahr WH, Chan AK, Williams S. Clinical features and outcome of pulmonary embolism in children. Br J Haematol 2008; 142: Buck JR, Connors RH, Coon WW, Weintraub WH, Wesley JR, Coran AG. Pulmonary embolism in children. JPediatrSurg1981; 16: WellsPS,GinsbergJS,AndersonDR,KearonC,GentM,TurpieAG, Bormanis J, Weitz J, Chamberlain M, Bowie D, Barnes D, Hirsh J. Use of a clinical model for safe management of patients with suspected pulmonary embolism. Ann Intern Med 1998; 129: Wells PS, Anderson DR, Rodger M, Ginsberg JS, Kearon C, Gent M, Turpie AGG, Bormanis J, Weitz J, Chamberlain M, Bowie D, Barnes D, Hirsh J. Derivation of a simple clinical model to categorize patients probability of pulmonary embolism: increasing the models utility with the SimpliRED D-dimer. Thromb Haemost 2000; 83: Babyn PS, Gahunia HK, Massicotte P. Pulmonary thromboembolism in children. Pediatr Radiol 2005; 35: Siedel HM,Ball JW,Dains JE,Benedict GW.MosbyÕs Guide to Physical Examination, 6th edn. Philadelphia, PA: Mosby, Wells PS, Anderson DR, Rodger M, Stiell I, Dreyer JF, Barnes D, Forgie M, Kovacs G, Ward J, Kovacs MJ. Excluding pulmonary embolism at the bedside without diagnosis imaging: management of patients with suspected pulmonary embolism presenting to the emergencydepartmentbyusingasimpleclinicalmodelandd-dimer.ann Intern Med 2001; 135: Stein PD, Woodard PK, Weg JG, Wakefield TW, Tapson VF, Sostman HD, Sos TA, Quinn DA, Leeper KV, Hull RD, Hales CA, Gottschalk A, Goodman LR, Fowler SE, Buckley JD. Diagnostic pathways in acute pulmonary embolism: recommendations of the PI- OPED II investigators. Am J Med 2006; 119: Righini M, Bounameaux H. External validation and comparison of recently described prediction rules for suspected pulmonary embolism. Curr Opin Pulm Med 2004; 10: Rajpurkar M, Warrier I, Chitlur M, Sabo C, Frey MJ, Hollon W, Lusher J. Pulmonary embolism-experience at a single childrenõs hospital. Thromb Res 2007; 119: Yilmaz D, Kavakli K, Ozkayin N. The elevated markers of hypercoagulability in children with Henoch-Schonlein Purpura. J Pediatr Hematol Oncol 2005; 22: Imamura T, Yoshihara T, Yokoi K, Nakai N. Ishida H, Kasubuchi Y. Impact of increased D-dimer concentrations in Kawasaki disease. Eur JPediatr2005; 164: Michelin E, Snijders D, Conte S, Dalla Via P, Tagliaferro T, Da Dalt L, Monciotti CM, Simioni P, Stefanutti G, Ghirardo V, Gamba P, Barbato A. Procoagulant activity in children with community acquired pneumonia, pleural effusion and empyema. Pediatr Pulmonol 2008; 43: ten Cate-Hoek AJ, Prins MH. Management studies using a combination of D-dimer test result and clinical probability to rule out venous thromboembolism: a systematic review. J Thromb Haemost 2005; 3: