HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:
Guest Access | Sign In via User Name/Password

This Article Abstract Full Text (PDF) Free Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Article Archive Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (2) Citing Articles via Google Scholar Google Scholar Articles by Hogg, K. Articles by Mackway-Jones, K. Search for Related Content PubMed PubMed Citation Articles by Hogg, K. Articles by Mackway-Jones, K.

Respiratory Dead Space Measurement in the Investigation of Pulmonary Embolism in Outpatients With Pleuritic Chest Pain^*

^* From the Emergency Medicine Research Group (Dr. Hogg, Ms. Dawson, and Dr. Mackway-Jones), Emergency Department, Manchester Royal Infirmary, Manchester, UK; Respironics International (Mr. Tabor), Boulogne, France; and Former Department Chair (Ms. Tabor), Respiratory Care Program, Jacksonville, FL.

Correspondence to: Kerstin Hogg, MBChB, Emergency Medicine Research Group, Emergency Department, Manchester Royal Infirmary, Oxford Rd, Manchester, M13 9WL UK; e-mail: kerstinhogg{at}hotmail.com

	Abstract

TOP Abstract Introduction Materials and Methods Results Discussion References

 
Study objectives: Pleuritic chest pain is a common presenting condition in the emergency department. A noninvasive bedside rule out test for pulmonary embolism would aid investigating this patient group. Our study aimed to compare the clinical utility of three methods for calculating respiratory dead space in the diagnosis of pulmonary embolism in outpatients with pleuritic chest pain.

Design: Prospective diagnostic study.

Setting: Large city-center emergency department.

Patients: Between February 2002 and June 2003, 425 patients presenting to the emergency department with pleuritic chest pain were prospectively recruited.

Intervention: Data collection for respiratory dead space was performed in the emergency department by two researchers. The respiratory dead space fraction was calculated independently using three different methods. All patients underwent an independent reference standard diagnostic algorithm to establish the presence or absence of pulmonary embolism. Those with a low modified Wells clinical probability and a normal quantitative d-dimer finding were discharged home. All others followed a reference standard protocol using Prospective Investigation of Pulmonary Embolism Diagnosis-interpreted ventilation/perfusion scanning, CT pulmonary angiography, and digital subtraction pulmonary angiography. All patients were followed up clinically for 3 months.

Measurements and results: For the Bohr calculation, the area under the receiver operating characteristic curve was 0.62 (95% confidence interval [CI], 0.51 to 0.73), the Enghoff calculation was 0.66 (95% CI, 0.55 to 0.77), and the capillary sample Enghoff was 0.62 (95% CI, 0.49 to 0.65). The optimum Bohr cutoff value gave 100.0% sensitivity (95% CI, 84.5 to 100%) but a low specificity of 22.7% (95% CI, 18.8 to 27.2%). The optimum cutoff points for Enghoff and capillary Enghoff calculations gave sensitivities of 95.3% (95% CI, 77.3 to 99.2%) and 94.4% (95% CI, 74.2 to 99.0%), respectively, with poor specificity.

Conclusions: Respiratory dead space analysis does not perform well as a standalone diagnostic test for pulmonary embolism in outpatients presenting with pleuritic chest pain.

Key Words: dead space • diagnosis • pleuritic chest pain • pulmonary embolism

	Introduction

TOP Abstract Introduction Materials and Methods Results Discussion References

 
By the age of 50 years, 0.5% of the population have received a diagnosis of thromboembolic disease.¹ The symptoms of pulmonary embolism are diverse, ranging from sudden death to dyspnea. Pulmonary infarction causes pleural membrane inflammation resulting in pleuritic chest pain, which is commonly found in these patients.²³ Those presenting to the emergency department with pleuritic chest pain are often young and otherwise healthy. This group would benefit from a simple bedside rule-out test for pulmonary embolism, which could reduce the numbers requiring diagnostic imaging.

In 1959, Robin et al⁴ suggested using respiratory dead space to diagnose pulmonary embolism. Pulmonary emboli obstruct blood flow to an area of alveoli within the lungs, producing an immediate pure alveolar dead space. In 1985, Burki⁵ calculated respiratory dead space using the Enghoff equation (the arterial to mixed expired carbon dioxide concentration difference over arterial concentration). He found that all patients with pulmonary emboli had a dead space fraction > 0.4. More recently, studies have demonstrated the utility of the area under the capnogram⁶ and FD late (the difference between arterial and exhaled carbon dioxide when curve extrapolated to 15% of total lung capacity)⁷ as diagnostic tests for pulmonary embolism. Kline et al⁸ used the Enghoff equation to calculate alveolar dead space in 380 patients referred for diagnostic imaging in six centers. The cutoff point of 0.2 yielded a sensitivity of 67% and specificity of 76%. When combined with a bedside d-dimer, the sensitivity rose to 98%. Both Kline et al⁹ and Rodger et al¹⁰ studied the use of the arterial to end-tidal carbon dioxide concentration difference over arterial concentration in the diagnosis of pulmonary embolism, giving sensitivities of 88% and 83%, respectively. Each dead space calculation necessitates analysis of exhaled air and in some cases simultaneous arterial blood sampling. The Manchester Investigation of Pulmonary Embolism Diagnosis (MIOPED) study aimed to compare three methods for the calculation of respiratory dead space and their clinical utility in the diagnosis of pulmonary embolism in outpatients with pleuritic chest pain.

	Materials and Methods

TOP Abstract Introduction Materials and Methods Results Discussion References

 
Patients
Ethics approval for the study was granted by the Central Manchester Local Research Ethics Committee. Patients presenting with pleuritic chest pain to the Manchester Royal Infirmary emergency department were approached for participation by one of two dedicated researchers (an emergency physician and a research nurse). The inclusion criteria was any patient with sharp pain between the neck and upper abdomen, which was worse on deep inspiration and coughing. Exclusion criteria were pneumothorax or ECG changes of cardiac ischemia, significant hypoxia (PaO₂ < 7.5 kPa or oxygen saturation by pulse oximetry < 88% on room air), a contraindication to contrast investigations, or weight > 140 kg. Pregnant women and children < 18 years of age were not recruited. Those who could not give informed consent or follow instructions were not recruited. Study patient demographics were recorded prospectively by the researchers. All emergency department-written notes were hand searched. Details were taken from the notes of any patient who was not recruited and described chest pain as sharp or pleuritic.

Respiratory Dead Space Analysis
Data Collection: Respiratory dead space data collection was performed blinded to the reference standard results in the emergency department. The patient lay in a semirecumbent position with a nose clip and breathed room air through a duck-billed mouth piece. Exhaled air was conducted through a mainstream circuit to a monitor (CO₂SMO Plus; Novametrics Medical Systems; Wallingford, CT), which analyzed exhaled carbon dioxide. The information was relayed by serial port to a laptop computer and downloaded into a software program (Analysis Plus for Windows, Version 5; Novametrics Medical Systems). Respiratory data were collected for a 10-min period. During the final 2 min of recording, a 2-mL arterial blood sample was collected in a heparinized syringe. Immediately afterwards, a pinprick was applied to the thumb and blood was collected in a heparinized capillary tube. Both blood samples were processed immediately in a blood gas analyzer, adjusted for body temperature. The blood gas analysis machine was calibrated daily.

At the end of the study, all patient CO₂SMO Plus monitor data were exported from Analysis Plus for Windows to Excel software (Microsoft; Redmond, WA). Mean values for end-tidal carbon dioxide and mixed expired carbon dioxide were calculated using all recorded measurements from each patient. The investigators compared three methods of calculating respiratory dead space fraction: end-tidal carbon dioxide concentration as a measure of alveolar carbon dioxide in the Bohr equation:

the Enghoff equation:

and the capillary Enghoff equation, which substitutes a capillary sample carbon dioxide concentration for the arterial concentration:

where VD = physiologic dead space volume, VT = tidal volume; PETCO₂ = end-tidal carbon dioxide partial pressure; PECO₂ = mixed expired carbon dioxide partial pressure; and PcCO₂ = capillary carbon dioxide partial pressure.

Reference Standard
The reference standard was conducted blind to the dead space results. Pulmonary embolism was excluded by a normal d-dimer result in a patient scoring low clinical probability of pulmonary embolism; a normal ventilation/perfusion scan finding; a low-probability ventilation-perfusion scan finding in a low-clinical-probability patient; a normal CT pulmonary angiographic finding in a low-clinical-probability patient; a CT scan finding demonstrating other pathology; or a normal digital subtraction pulmonary angiographic finding. Pulmonary embolism was diagnosed by a high-probability ventilation/perfusion scan finding in a moderate- or high-clinical-probability patient, a positive CT pulmonary angiographic finding, or a positive digital subtraction pulmonary angiographic finding. No other clinical data were used to determine the diagnosis of pulmonary embolism.

The MIOPED study used the Manchester-modified Wells clinical probability score, which adds IV drug use as 3.0 points to the Wells score.¹¹ Our department has demonstrated (S. Jones, MBChB; unpublished observation; August 1999 to February 2001) that IV drug users with symptoms of deep vein thrombosis are at high risk for the condition. This modification ensured that the subpopulation of IV drug users underwent diagnostic imaging for pulmonary embolism. All patients underwent a quantitative latex agglutination d-dimer: 421 patients underwent the IL Test d-dimer (Instrumentation Laboratory; Aragón, Barcelona, Spain) [cutoff, 278 ng/mL], and four patients underwent the MDA d-dimer (Organon Teknika BV; Boxtel, the Netherlands) [cutoff, 0.5µg FEU/mL]. Ventilation/perfusion scanning was the diagnostic image of choice in patients without airways disease, pulmonary fibrosis, or left ventricular failure. The clinical probability scores, ventilation/perfusion scans, CT pulmonary angiograms, and digital subtraction pulmonary angiograms were independently reported by a second consultant in the appropriate specialty who was blinded to clinical details. Any discrepancies were resolved by the opinion of a third consultant.

All patients regardless of diagnosis were followed up clinically for 3 months. Each patient recruited into the study received written instructions to return or contact the lead physician on a 24 h/d page if he or she experienced any further symptoms of chest pain, shortness of breath, or pain or swelling of the legs. In this case, the patient was reviewed in the emergency department and investigated using the above diagnostic algorithm for pulmonary embolism. In addition, all patients were contacted by telephone 3 months after recruitment. Patients were followed up for evidence of deep vein thrombosis or pulmonary embolism.

An independent adjudication committee (two respiratory consultant physicians, one consultant in nuclear medicine, and one consultant radiologist) reviewed all cases in which a patient died during study follow-up. The committee also reviewed any case in which a patient had undergone further testing for pulmonary embolism during the clinical follow-up period without the knowledge of the study group. A consensus opinion was given on whether the patient’s death or continued symptoms were "unlikely," "possibly," or "probably" caused by thromboembolic disease.

Clinical probability scores and diagnostic images were rechecked in retrospect. Inevitably, this meant some patients had not completed their full investigation according to the final test results. All patients were included in the analysis, including patients who in retrospect did not complete full diagnostic imaging but had an uneventful follow-up period without anticoagulation. Those who failed to complete diagnostic imaging and received anticoagulation during follow-up were excluded from analysis.

Statistics
The study aimed to determine whether respiratory dead space was a useful test in excluding pulmonary embolism. A study cohort of 400 patients and a prevalence of 10% could prove a sensitivity of 95% with 95% confidence intervals (CIs) of 83 to 99%. A prevalence of 5% could prove a 95% sensitivity with 95% CI of 76 to 99%.

CIs for the area under the receiver operating characteristics curve were calculated using the Delong estimate.¹² CIs for sensitivity, specificity, and predictive values were calculated using the Wilson "score" method.¹³¹⁴ Those for likelihood ratios were constructed using the likelihood-based approach to binomial proportions by Gart and Nam.¹⁵ All prospectively accumulated information was stored anonymously in a database using statistical software (SPSS version 11.5; SPSS; Chicago, IL). CIs and receiver operating characteristics (ROC) curves were obtained using software (StatsDirect; StatsDirect Ltd; Cheshire, UK).

	Results

TOP Abstract Introduction Materials and Methods Results Discussion References

 
The Study Cohort
Between February 11, 2002, and May 29, 2003, 799 patients were assessed for inclusion into the MIOPED study. Four hundred twenty-five patients consented to participate in the study, and 97 patients declined to participate. Two hundred seventy-seven patients were excluded, 171 because their pain did not fit the inclusion criteria. Of the patients who were assessed, 64% were reviewed by a researcher between the hours of 9 AM and 3 PM. Twenty-eight percent were assessed between 3 PM and midnight, and 8.0% were assessed between midnight and 9 AM.

During the same time period, an additional 633 patients attended the emergency department with chest pain described in the notes as either pleuritic or sharp. The MIOPED cohort did not differ in age and sex distribution from the group who declined, were excluded, or were missed. A minority of patients who were not recruited had a clinical probability score or arterial blood gas analysis. The patients who underwent these tests did not differ clinically from the MIOPED recruits: 86.6% had low clinical probability vs 84.1% who were not recruited; mean PaO₂ of 11.8 kPa vs 11.0 kPa; mean PaCO₂ of 5.1 kPa vs 5.1 kPa; and mean WBC count of 8.9 x 10⁹/L vs 9.0 x 10⁹/L, respectively. Table 1 shows the study patient demographics. Table 2 shows the study patient results.

During investigation, acute pulmonary embolism was diagnosed in 20 patients. Five patients could not be contacted during follow-up, and 98.8% of patients completed follow-up. Four patients died. Two patients were determined by the adjudication committee "possibly" to have pulmonary embolic disease after re-presenting in follow-up, one of whom died. The other three deaths and one re-presentation were deemed "unlikely" to be caused by thromboembolic disease. Table 3 details how pulmonary embolism was diagnosed and excluded.

Respiratory Dead Space Analysis Outcomes
Diagnostic Characteristics: All patients underwent respiratory dead space analysis. Fourteen patient sets of respiratory parameters were unavailable for analysis. A technical fault with the Analysis Plus for Windows program resulted in loss of data for three patients. The data for 11 patients were lost when a laptop computer was stolen from the emergency department. Two patients were unable to follow the instructions to breathe through the mouthpiece and produced no analyzable data. Two patients did not undergo arterial blood gas analysis. The technique for capillary blood sampling proved difficult, and we were unable to obtain a sufficient sample for analysis in 59 patients.

Figures 123 show the ROC plots for all three methods of respiratory dead space calculation. The optimum cutoff points to achieve maximum use as an exclusion test for pulmonary embolism were calculated. Table 4 shows the sensitivity, specificity, predictive values, and likelihood ratios when these cutoff values are applied to this patient cohort.

View larger version (10K): [in this window] [in a new window] [Download PPT slide]   Figure 1.. ROC curve for Bohr-calculated dead space fraction.

View larger version (10K): [in this window] [in a new window] [Download PPT slide]   Figure 2.. ROC curve for Enghoff-calculated dead space fraction.

View larger version (10K): [in this window] [in a new window] [Download PPT slide]   Figure 3.. ROC curve for capillary sample Enghoff-calculated dead space fraction.

Respiratory Dead Space and Latex Agglutination D-dimer
In previous studies,⁸⁹¹⁰ respiratory dead space analysis has been combined with d-dimer to form a combined diagnostic test. Table 5 shows the diagnostic characteristics when the reference standard latex agglutination d-dimer test is combined with the Bohr-calculated respiratory dead space test. If both d-dimer and dead space are elevated, the combined test result is positive. If either or both tests are normal, the combined test result is negative.

	Discussion

TOP Abstract Introduction Materials and Methods Results Discussion References

The MIOPED study recruited patients with pleuritic chest pain. Pleuritic chest pain is a feature in 85% of patients with submassive pulmonary emboli and no previous cardiac or respiratory disease.³ In contrast to other pulmonary embolism studies,⁸¹⁰¹¹ in which patients are enrolled on the clinician’s suspicion of pulmonary embolism, our study had an objective, well-defined starting point. Our pretest probability was 5.3%, and the majority of patients scored a low clinical probability of pulmonary embolism. These are precisely the patients who would benefit most from a noninvasive bedside rule out test for pulmonary embolism. Any emergency department test that can safely reduce the proportion of patients requiring imaging would reduce radiation exposure, cost, and hospital admission. The MIOPED study is the first to assess the utility of respiratory dead space analysis in emergency department patients presenting with pleuritic chest pain.

We were unable to assess every patient who described their chest pain as sharp or pleuritic. This could have introduced bias to our cohort. However, the MIOPED study assessed the same percentage of all potential patients throughout the evenings, nights, and weekends. The MIOPED cohort did not differ clinically from those not recruited.

The study was designed to minimize all imperfect reference standard bias. Diagnostic images were independently rechecked, all patients were followed up clinically for 3 months, and all deaths were reviewed by an independent panel. The respiratory dead space data collection and calculation was conducted blind to the reference standard results. To our knowledge, the MIOPED study is the first to have conducted a standardized algorithm independent of planned clinician testing.

Our results demonstrate poor areas under the ROC curves for all dead space calculations. Theoretically dead space analysis might have performed well in our patient cohort with a mean age of 38 years and little comorbidity. Most patients had normal gas exchange and chest radiographs. The majority found it easy to follow our instructions and breathe through the mouthpiece for 10 min. However, all patients with pleuritic chest pain were included regardless of clinical evidence of other lung disease such as pneumonia. Furthermore, both Comroe¹⁶ and more recently Giuntini¹⁷ suggested that there exists a substantial redistribution of ventilatory air away from unperfused lung following pulmonary embolism. As a result, the concentration of end-tidal carbon dioxide might remain unchanged and the calculated respiratory dead space could be normal.

The MIOPED cohort had a low prevalence of pulmonary embolism. Although this does not directly influence the sensitivity and specificity results, negative predictive values are high even when the test performs poorly. The spectrum of disease may have influenced the results. Pleuritic chest pain is associated with smaller pulmonary emboli. In some cases, the infarction volume may have been too small to produce an abnormal respiratory dead space fraction in otherwise healthy patients.

The three receiver operating characteristics curves vary in shape. The Bohr equation curve allows a cutoff point with high sensitivity and the fewest false-positive results (potentially more use as a rule out test). Bohr dead space calculation uses only expired air. The Enghoff dead space requires additional capillary or arterial blood analysis. Capillary sampling proved technically difficult, and not all arterial samples were obtained on first attempt. It is possible that some patients hyperventilated prior to arterial puncture, lowering the blood carbon dioxide concentration, as two Enghoff results were negative, implying the arterial carbon dioxide concentration was lower than the mean mixed expired concentration.

It appears that combining Bohr-calculated dead space and d-dimer might be a more useful way to excluded pulmonary embolism; however, care must be taken when interpreting this result. Firstly, the authors are applying the optimal cutoff point for respiratory dead space, which was derived from the same patient cohort. This will overestimate the performance of the dead space test. Secondly, the latex agglutination d-dimer formed part of the reference standard investigation (not a research test), and the combined test results are open to incorporation bias.

Our results contrast with research of others. The multicenter study by Kline et al⁸ used a similar method of data collection but calculated sensitivity as 67.2% (95% CI, 55.0 to 77.5%) and specificity as 76.3% (95% CI, 71.2 to 85.6%). The prevalence of pulmonary embolism was 16.8%, compared to 5% in our study. The mean age was 52 years, only 44.5% had pleuritic chest pain, and 15% had malignancy. There may have been fewer "pulmonary infarction"¹⁸-type presentations and more patients with large pulmonary emboli. Rodger et al¹⁰ collected data with capnography and the CO₂SMO Plus monitor. They used a modified version of the Enghoff equation, calculating a sensitivity of 83.0% (95% CI, 69.2 to 92.4%) and a specificity of 70.3% (95% CI, 61.2 to 78.0). Again, the population differed, with a mean age of 60.6 years. Both have successfully combined dead space with d-dimer to give a combined test with high sensitivity. In each case, however, the combined test was positive if either dead space or d-dimer were elevated. In contrast, our study found the combined test performed best if it was considered positive only when both dead space and d-dimer were elevated. Verschuren et al¹⁹ analyzed volumetric capnogram curves from 45 emergency department patients with a pretest probability of 40% and concluded that the area under the ROC curve for Fd late (the arterial to exhaled carbon dioxide when curve extrapolated to 15% of total lung capacity) was 87.6 ± 4.9% (mean ± SD). This was a preliminary study with selected patients and experimental technology. As yet, there is no recommendation to standardize the calculation of respiratory dead space.

Many studies⁸⁹¹⁰ have combined respiratory dead space with d-dimer to exclude pulmonary embolism. The Bohr equation was the easiest to calculate, using only mean data from the CO₂SMO Plus monitor. The ROC curve lends to a cutoff point yielding a high sensitivity with the fewest number of false-positive results. Future work is required to show whether using this cutoff value in combination with a d-dimer test is an effective rule-out tool in patients with pleuritic chest pain.

	Acknowledgements

 
The authors thank Magnus Harrison, who was involved in the planning of the study, and Stewart Teece, Mary Prescott, and Paul Taylor, who helped during data collection. Caroline Shiach helped obtain funding.

	Footnotes

 
Abbreviations: CI = confidence interval; MIOPED = Manchester Investigation of Pulmonary Embolism Diagnosis; ROC = receiver operating characteristic

Mr. Tabor is employed by Respironics International, who market the CO₂SMO Plus monitor. Mr. Tabor provided work on this study independent of Respironics International and conducted a blind analysis of the data files.

Dr. Hogg designed the study, collected the data, and drafted the article; Ms. Dawson collected data, and redrafted and approved the final version of the article; Mr. Tabor and Ms. Tabor analyzed the data; and Dr. Mackway-Jones conceived and supervised the study, redrafted, and approved the final version of the article.

Funding was provided by Pharmacia Limited, who funded a part-time research nurse for 13 months of the study recruitment period.

Received for publication January 10, 2005. Accepted for publication March 22, 2005.

	References

TOP Abstract Introduction Materials and Methods Results Discussion References

 

1. Hansson, PO, Welin, L, Tibblin, G, et al (1997) Deep vein thrombosis and pulmonary embolism in the general population: the study of men born in 1913. Arch Intern Med 157,1665-1670[Abstract]
2. Miniati, M, Prediletto, R, Formichi, B, et al Accuracy of clinical assessment in the diagnosis of pulmonary embolism. Am J Respir Crit Care Med 1999;159,864-871[Abstract/Free Full Text]
3. Stein, PD, Willis, PW, DeMets, DL History and physical examination in acute pulmonary embolism in patients without preexisting cardiac or pulmonary disease. Am J Cardiol 1981;47,218-223[CrossRef][ISI][Medline]
4. Robin, E, Julian, D, Travis, D, et al A physiologic approach to the diagnosis of pulmonary embolism. N Engl J Med 1959;260,586-591[ISI][Medline]
5. Burki, NK The dead space to tidal volume ratio in the diagnosis pulmonary embolism. Am Rev Respir Dis 1985;133,679-685
6. Kline, JA, Arunachlam, M Preliminary study of the capnogram waveform area to screen for pulmonary embolism. Ann Emerg Med 1998;32,289-296[CrossRef][ISI][Medline]
7. Eriksson, L, Wollmer, P, Olsson, CG, et al Diagnosis of pulmonary embolism based on alveolar dead space analysis. Chest 1989;96,357-362[Abstract/Free Full Text]
8. Kline, JA, Israel, EG, Michelson, EA, et al Diagnostic accuracy of a bedside d-dimer assay and alveolar dead-space measurement for rapid exclusion of pulmonary embolism: a multicenter study. JAMA 2001;285,761-768[Abstract/Free Full Text]
9. Kline, JA, Meek, S, Boudrow, D, et al Use of the alveolar dead space fraction (Vd/Vt) and plasma d-dimers to exclude acute pulmonary embolism in ambulatory patients. Acad Emerg Med 1997;4,856-863[ISI][Medline]
10. Rodger, MA, Jones, G, Rasuli, P, et al Steady-state end-tidal alveolar dead space fraction and d-dimer: bedside tests to exclude pulmonary embolism. Chest 2001;120,115-119[Abstract/Free Full Text]
11. Wells, PS, Anderson, DR, Rodger, M, et al 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;83,416-420[ISI][Medline]
12. DeLong, ER, DeLong, DM, Clarke-Pearson, DL Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44,837-845[CrossRef][ISI][Medline]
13. Agresti, A, Coull, BA Approximate is better than "exact" for interval estimation of binomial proportions. Am Stat 1998;52,119-126[CrossRef]
14. Newcombe, RG Two-sided confidence intervals for the single proportion: comparison of seven methods. Stat Med 1998;17,857-872[CrossRef][ISI][Medline]
15. Gart, J, Nam, J Approximate interval estimation of the ratio of binomial parameters: a review and corrections for skewness. Biometrics 1988;44,323-338[CrossRef][ISI][Medline]
16. Comroe, JH The main functions of the pulmonary circulation. Circulation 1966;33,146-158[Free Full Text]
17. Giuntini, C Ventilation/perfusion scan and dead space in pulmonary embolism: are they useful for the diagnosis? Q J Nucl Med 2001;45,281-286[Medline]
18. Stein, PD, Henry, JW Clinical characteristics of patients with acute pulmonary embolism stratified according to their presenting syndromes. Chest 1997;112,974-979[Abstract/Free Full Text]
19. Verschuren, F, Liistro, G, Coffeng, R, et al Volumetric capnography as a screening test for pulmonary embolism in the emergency department. Chest 2004;125,841-850[Abstract/Free Full Text]

This article has been cited by other articles:

				K Hogg, D Dawson, and K Mackway-Jones Outpatient diagnosis of pulmonary embolism: the MIOPED (Manchester Investigation Of Pulmonary Embolism Diagnosis) study Emerg. Med. J., February 1, 2006; 23(2): 123 - 127. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Free Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Article Archive Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (2) Citing Articles via Google Scholar Google Scholar Articles by Hogg, K. Articles by Mackway-Jones, K. Search for Related Content PubMed PubMed Citation Articles by Hogg, K. Articles by Mackway-Jones, K.