| | Effect of teaching and type of stethoscope on cardiac auscultatory performanceReceived 5 December 2005; accepted 13 April 2006. BackgroundAuscultation of the heart is a routine procedure. It is not known whether auscultatory skills can be improved by teaching or with the use of an advanced stethoscope. MethodsThis study was a randomized trial with a 2 × 2 factorial design. Seventy-two house officers were randomized to a simple or an advanced stethoscope and to a 4-hour course in auscultation or no course. The doctors auscultated 20 patients' hearts and categorized findings as normal or as one or more of 5 categories of heart diseases. Patients were selected such that 16 had a known heart disease as well as a corresponding murmur and 4 had no heart disease or murmur. Auscultatory performance was assessed as concordance with echocardiographic findings and interobserver variation. ResultsDoctors using the advanced stethoscope diagnosed 35% of the patients correctly, as compared with doctors using the simple stethoscope who did 33% of the patients (P = .27). Similarly, 34% of the patients were diagnosed correctly by doctors who had received teaching as compared with 33% of those who were by doctors who had received no teaching (P = .41). The κ values were higher for doctors who had received teaching for aortic stenosis (0.43 vs 0.28, P = .004) and ventricular septum defect (0.07 vs 0.01, P = .003). There was no difference between groups for any other single murmur or for the detection of murmurs as such. ConclusionHeart auscultation findings were in poor accordance with echocardiographic findings and had high interobserver variation. Neither outcome improved to any important extent with the subjects' use of an advanced stethoscope or attending of a course in heart auscultation. Stethoscopic examination of the heart is a routine procedure. Auscultatory findings hold important prognostic information1, 2, 3 and often guide clinicians in recommending further treatment and examinations. The reliability of heart auscultation is reflected partly in the ability to detect or exclude underlying heart disease (accuracy) and partly in the level of agreement between observers (precision). Studies on how to improve auscultatory skills are sparse and have focused on the effect of type of stethoscope and teaching. The studies were small and with important methodological problems, however. No randomized clinical trial that investigates whether auscultatory performance is improved by training has been published. The only randomized trial published investigating the effect of type of stethoscope compared an electronic stethoscope with a conventional stethoscope and could not show a clear benefit of using an advanced stethoscope.4 Our primary aim was to investigate the effects of training and type of stethoscope used on both the accuracy and precision of heart auscultation. Secondarily, we wanted to examine auscultatory performance in general. Methods  We conducted a randomized trial with a 2 × 2 factorial design. Seventy-two observers were randomized to use 1 of 2 types of stethoscopes and to either teaching or no teaching. All observers performed heart auscultation on the same 20 patients. Observers All house officers from 10 hospitals in Copenhagen County were sent a letter with an invitation to participate. The first 72 doctors who responded were included in the study. All the participating doctors declared that they had no knowledge of any of the patients and any known hearing disability. All observers consented in using the study stethoscope in their clinical work for 4 weeks before the day the study patients were to be examined. The observers were not informed whether they should receive teaching or not until the day the examinations were performed. House officers in Denmark have between 0 and 18 months of postgraduate clinical experience and do not receive any formalized training or teaching in auscultation during this period. Randomization Observers were irreversibly included in the study before randomization. Randomization was conducted in one public session with 3 of the authors present. It was based on a list of computer-generated random numbers. No reallocation or exclusion of observers took place after randomization. Seventy-two observers were allocated to 4 equal-sized groups. Group 1 received a 3M Littmann (Cerritos, CA) Classic II SE stethoscope and did not receive teaching. Group 2 received a 3M Littmann Master Cardiology stethoscope and did not receive teaching. Group 3 received a 3M Littmann Classic II SE stethoscope and received teaching. Group 4 received a 3M Littmann Master Cardiology stethoscope and received teaching. Patients Potentially eligible outpatients were identified from conveniently sampled patient files by 2 investigators (MD and KI). Sixteen patients with charts describing murmurs and significant valve disease or septal defect and 4 with charts describing no murmur and no valve disease or septum defect were recruited. The patients were examined with auscultation and echocardiography within 48 hours before the study by one investigator (KI). The classification of the echocardiographic findings was based on the discretion of the examiner. We included patients with murmurs only if they had a moderate or severe valve defect or a significant septum defect and a murmur concordant with the echocardiographic examination. Similarly, we included patients without murmurs only if they had normal heart valves without color jets showing regurgitation or turbulent flow (minor tricuspidal regurgitation was accepted) and had no audible murmur. Oral informed consent was given by all included patients. All 20 screened patients were included in the study. The local ethical committee approved the study. Teaching The observers randomized to training were taught for 4 hours by a specialist in cardiology (Frank Steensgaard-Hansen) with extensive experience in teaching auscultation. The course consisted of a theoretical introduction to auscultation as well as auscultatory techniques and a description of auscultatory findings, followed by 3 hours of providing examples and training in recognition of digitally recorded murmurs. Stethoscopes We provided the observers with a simple or an advanced membrane-based stethoscope. We chose the 3M Littmann Classic II SE stethoscope, which is one of the cheapest Littmann stethoscopes, as the simple stethoscope. We chose the 3M Littmann Master Cardiology stethoscope, which is an expensive acoustic stethoscope, as the advanced stethoscope. Study examinations All examinations were conducted on the same day. Auscultation was performed while patients were sitting in their beds. Doctors were instructed to restrict each examination to not longer than 2 minutes per patient. The patients were asked not to reveal any clinical information to the examiners. At the end of each examination, the doctors filled up a multiple-choice questionnaire with 5 categories of heart diseases (aortic regurgitation, aortic stenosis, mitral regurgitation, ventricular septal defect, and atrial septal defect) or no heart disease. Observers could place more than one cross in the multiple-choice questionnaire and were told that the patients had one or more of the cited heart diseases or no heart disease. Data analyses Data were manually transferred from the questionnaires to a data sheet and thereafter checked for typing errors. For each single observer, we calculated sensitivity, specificity, and agreement with the echocardiographic findings. Subsequently, these values were compared between the groups of observers, according to either teaching or type of stethoscope (using unpaired t tests), or both (using general linear models). The κ values were calculated for all pairs of doctors within each group (defined either by teaching or by type of stethoscope, or a combination). Because of the interdependence between these values (several κ values for each single observer), P values from traditional tests (test with an assumption of independence between observations) could not be trusted and we therefore used a randomization test to compare groups.5 Primarily planned analysis was the difference between the randomized groups with respect to (1) sensitivity and specificity of the auscultation as well as the number of findings that concurred with the echocardiographic findings and (2) κ values. For secondary analysis, we studied the difference in the total number of positive auscultatory findings and the interaction between type of stethoscope and teaching. Results All 20 patients who were included finished the study. There were 13 men (65%), and the patients' median age was 69 years (range 30-88 years). Seven patients had aortic stenosis, 5 patients had mitral regurgitation, 1 patient had atrial septal defect, 1 patient had aortic regurgitation, 2 patients had aortic stenosis, and 4 patients had no valve disease or septal defect. All the participating doctors auscultated all the patients' hearts (1440 examinations). Effect of teaching Accuracy Although sensitivity was higher among doctors who received teaching for all murmurs, there was no statistically significant difference in any single sensitivity or specificity of auscultation between doctors who received teaching and those who did not (Table I). Similarly, there was no statistically significant difference between the average number of patients who were diagnosed correctly by physicians who had not received teaching (33% of patients) and that of patients who were by physicians who received teaching (34% of patients) (P = .41). Agreement For 2 of the 5 categories of heart diseases, the observers who received teaching agreed more than did those who received no teaching. The mean κ value for aortic stenosis was 0.43 versus 0.28; for ventricular septum defect, it was 0.07 versus 0.01 (Table II). There was no statistically significant difference between the mean κ values for the observers who had been taught and those who had not been taught with respect to the other 3 single murmurs (Table II)—neither was there any statistically significant difference between those with respect to the detection of any murmur—with the mean κ value being 0.14 versus 0.18. Regardless of the statistically significant effect of training on 2 examinations, the size of the effect was small and did not result in observers reaching mean κ values >0.5. Effect of type of stethoscope Accuracy There was no statistically significant difference between the doctors using either type of stethoscope with respect to either the sensitivity or the specificity of heart auscultation (Table III). There was no statistically significant difference between the average number of patients who were diagnosed correctly by doctors using the simple stethoscope (33% of patients) and that of patients who were by doctors using the advanced stethoscope (35% of patients) (P = .27). Agreement There was no statistically significant difference in the level of agreement between doctors using the simple stethoscope and those using the advanced stethoscope. This was the case both for the 5 single heart diseases and for the detection of any heart disease (Table IV). General auscultatory performance The sensitivity of heart auscultation ranged from 0.11 to 0.47, lowest for atrial septal defect and highest for aortic stenosis. The specificity ranged from 0.82 to 0.93, lowest for mitral regurgitation and highest for aortic regurgitation. The sensitivity ranged from 0.71 to 0.74 and the specificity ranged from 0.56 to 0.67 for detection of any murmur. Mean κ values ranged from 0.01 to 0.43, lowest for recognition of ventricular septal defect and highest for recognition of aortic stenosis. Secondary analysis We found no interaction between type of stethoscope and teaching (data not shown). There was no statistically significant difference between the compared groups with respect to the number of murmurs detected (Table V). Discussion Main results We did not find that the performance of heart auscultation was importantly improved by the type of stethoscope or by teaching of doctors. A small effect of teaching was however seen on the mean κ values in 2 of 5 examinations and on a nonsignificant trend for the specificity of auscultation. On average, only one third of the patients were diagnosed correctly; in addition, in general, the sensitivity of diagnosing the different heart diseases was low. The sensitivity in detecting any murmur was between 0.71 and 0.74, implying that more than 1 of 4 patients with significant valve disease were considered to have a normal stethoscopic examination—a finding we find rather troubling. κ Values The κ value summarizes agreement beyond chance for binary variables, with 0 corresponding to no agreement beyond chance and 1 corresponding to perfect agreement. κ Values are dependent on the prevalence of the disease; therefore, comparison of κ values between studies with different objectives and different study populations is difficult. Here, we mainly used κ values for internal comparisons to compare agreement for different stethoscopes and levels of clinical experience. According to Landis and Koch,6 κ values <0 indicate poor agreement; between 0.00 and 0.20, slight agreement; between 0.21 and 0.40, fair agreement; between 0.41 and 0.60, moderate agreement; between 0.61 and 0.80, substantial agreement; and between 0.81 and 1.00, excellent or almost perfect agreement. Agreement in our study was between slight and moderate. Aortic stenosis is traditionally the easiest murmur to detect7 and was in our study also the single murmur with the highest agreement rates. Surprisingly, aortic regurgitation, which is a weak diastolic murmur, had higher agreement rates than did mitral regurgitation, which is a murmur that is detected easily.7 The agreement rates for the detection of whether there is a murmur or not were fair for all groups of doctors and independent of type of stethoscope used and of teaching. Teaching The auscultation course lasted only 4 hours and was based on theory and training with recorded heart sound. We chose a relatively short course with only one teaching session partly owing to the practical difficulties in reassembling the same group of doctors twice and partly to ensure that none of the observers knew whether they were allocated to teaching or no teaching before the day of the examinations, which potentially could influence any preparation for the study. It is however possible that a longer auscultation course with more sessions and practical training with real patients would make the weak effect of teaching stronger. Studies on other clinical skills have shown an effect of training on interobserver variation.8 Stethoscopes We compared 2 stethoscopes from the same firm. 3M Littmann Classic II SE is the cheapest Littmann stethoscope recommended for doctors, whereas 3M Littmann Master Cardiology is the most expensive acoustic Littmann stethoscope. The manufacturer claims that 3M Littmann Master Cardiology offers the ultimate performance (http://www.3m.com/us/healthcare/professionals/littmann/jhtml/products.jhtml). We found no difference between doctors using the simple stethoscope and those using the advanced stethoscope in concordance with echocardiographic findings, interobserver variation, or number of murmurs heard. Although we only compared 2 types of stethoscopes, we think that our results suggest that an expensive and advanced stethoscope does not necessarily improve auscultatory performance, at least not in the hands and ears of house officers. Previous studies The effect of type of stethoscope used has been evaluated using comparison of acoustic performance9, 10, 11 or clinicians' self-reported preferences.12, 13 The effect of type of stethoscope on precision has been examined in one study with only 12 observers.4 The study compared an acoustic stethoscope with an electronic stethoscope and could not show a significant difference in κ values between the 2 stethoscopes. The effect of teaching and training has been evaluated in only one clinical study, which suggested a beneficial effect of training and teaching on auscultatory accuracy.14 The study was however nonrandomized and included only 10 observers examining the same patients before and after teaching. Other studies used recordings of cardiac murmurs, which do not necessarily correlate well with bedside skills, and showed only minimal or no effect of training on auscultatory performance.15, 16, 17, 18 Observational studies on the accuracy of heart auscultation are sparse. They indicate that, in general, the sensitivity of cardiologists in detecting relevant valve disease is reasonable; however, several studies found low to moderate sensitivities of heart auscultation.19, 20, 21, 22, 23, 24 The few published studies on the interobserver variation of the auscultation of the heart reported surprisingly low agreement rates, expressed as chance-corrected agreement rates, or κ values, <0.5.25, 26, 27, 28 Disregarding the randomization, our study by far included the largest number of observers in an observational study describing the accuracy and precision of auscultation in a clinical setting.14, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37 Limitations The setting in our study differed somewhat from a normal clinical situation. The patients were selected to ensure a reasonably high proportion of positive findings. As in most agreement studies, the observers had no knowledge of the patients' history or of other clinical signs apart from their auscultatory findings. The doctors could not discuss the findings with colleagues. However, knowledge of the patients' history has previously been shown not to influence observer variation of auscultation of the lungs.38 We have included a relatively high number of observers, compared with the number of patients, in our study. This was done because the study objective was related to the performance of observers and not to effect on patients. A high number of observers therefore gave the study the largest power. The relatively low number of patients participating in the study could of course be a problem regarding the representation of the murmurs. We did however choose patients with common murmurs and typical auscultatory presentation to minimize this problem. The echocardiographic examinations were performed and judged by a single observer. The risk of misclassification was small, however, because all the patients already had a known heart disease and our echocardiographic examination confirmed previous findings. The comparisons between type of stethoscope and teaching are based on groups with only house officers. This was done to ensure a homogenous group of observers. This could make it difficult to extrapolate the results to doctors with other levels of experience. It seems however most logical to measure the effect of teaching among the most inexperienced doctors for whom it seems reasonable to assume that the reliability of their cardiac auscultation would give room for most improvement and not among doctors who have already received years of training and education. It would however be of interest to see if a study with more experienced observers would find similar results regarding the effect of teaching and type of stethoscope used on auscultatory skills. Minor echocardiographic valve defects might not be possible to hear during auscultation. This could reduce the number of patients who obtained the correct auscultatory diagnosis. However, all the valve diseases were judged as moderate or severe and physicians who examined patients before the study heard a murmur that was concordant with the echocardiographic finding. It is possible that a study with a higher number of observers would be able to find a significant difference regarding auscultatory skills between groups of observers. However, even if the results of such a study would be consistent with our results, we think that such a difference would be of limited clinical interest. Clinical implications Our study shows that stethoscopic examinations performed by house officers have low sensitivity and specificity as well as high interobserver variation. We found no beneficial effect of using an advanced and expensive stethoscope, suggesting that it may be rational to use a simple and cheaper one. Neither did we find a clear tendency for the beneficial effect of teaching, although there might have been a tendency for a minor effect on agreement and sensitivity. We find that it would be interesting to examine the effect of more extensive teaching and training. In the clinical situation, stethoscopic findings are integrated with information from clinical history, physical examination, x-ray pictures, and blood tests. Furthermore, house officers will typically consult more experienced physicians when in doubt. A future study on stethoscopic performance could take these factors into account. In the meantime, stethoscopic findings should be interpreted cautiously. Conclusion We found that the sensitivity of stethoscopic examinations in detecting and excluding echocardiographically defined heart valve diseases was only moderate and that observers disagreed considerably. This is troubling because auscultating of the heart is a routine procedure that often guides clinicians in further planning for patients. Neither training nor type of stethoscope improved sensitivity, specificity, or agreement importantly. KI and AH initiated and coordinated the formulation of the primary study hypothesis and designed the protocol. KI, RG, HT, MD, LT, and AT participated in data collection. KI, AH, and LS analyzed the data and wrote the first draft of the paper. All authors interpreted the data and revised the first draft for important intellectual content. All authors approved the final version. KI is the guarantor of the article. KI had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. References 1. 1Barzilai B, Davis VG, Stone PH, et al. Prognostic significance of mitral regurgitation in acute myocardial infarction. The MILIS Study Group. Am J Cardiol. 1990;65:1169–1175. MEDLINE |
CrossRef
2. 2Dey DK, Sundh V, Steen B. Do systolic murmurs predict mortality in the elderly? A 15-year longitudinal population study of 70-year-olds. Arch Gerontol Geriatr. 2004;38:191–200. Abstract | Full Text |
Full-Text PDF (74 KB)
|
CrossRef
3. 3Drazner MH, Rame JE, Stevenson LW, et al. Prognostic importance of elevated jugular venous pressure and a third heart sound in patients with heart failure. N Engl J Med. 2001;345:574–581. MEDLINE |
CrossRef
4. 4Iversen K, Greibe R, Timm HB, et al. A randomized trial comparing electronic and conventional stethoscopes. Am J Med. 2005;118:1289. Full Text |
Full-Text PDF (207 KB)
5. 5Edgington ES. Randomization tests. 1995. 6. 6Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33:159–174.
CrossRef
7. 7Chizner MA. The diagnosis of heart disease by clinical assessment alone. Dis Mon. 2002;48:7–98. Full Text |
Full-Text PDF (560 KB)
|
CrossRef
8. 8Brorson S, Bagger J, Sylvest A, et al. Improved interobserver variation after training of doctors in the Neer system. A randomised trial. J Bone Joint Surg Br. 2002;84:950–954.
CrossRef
9. 9Abella M, Formolo J, Penney DG. Comparison of the acoustic properties of six popular stethoscopes. J Acoust Soc Am. 1992;91:2224–2228. MEDLINE |
CrossRef
10. 10Ertel PY, Lawrence M, Song W. How to test stethoscopes. Med Res Eng. 1969;8:7–17. MEDLINE 11. 11Kindig JR, Beeson TP, Campbell RW, et al. Acoustical performance of the stethoscope: a comparative analysis. Am Heart J. 1982;104:269–275. MEDLINE |
CrossRef
12. 12Grenier MC, Gagnon K, Genest J, et al. Clinical comparison of acoustic and electronic stethoscopes and design of a new electronic stethoscope. Am J Cardiol. 1998;81:653–656. Full Text |
Full-Text PDF (281 KB)
|
CrossRef
13. 13Philip JH, Raemer DB. An electronic stethoscope is judged better than conventional stethoscopes for anesthesia monitoring. J Clin Monit. 1986;2:151–154. MEDLINE |
CrossRef
14. 14Favrat B, Pecoud A, Jaussi A. Teaching cardiac auscultation to trainees in internal medicine and family practice: does it work?. BMC Med Educ. 2004;4:5. MEDLINE |
CrossRef
15. 15Barrett MJ, Lacey CS, Sekara AE, et al. Mastering cardiac murmurs: the power of repetition. Chest. 2004;126:470–475. MEDLINE |
CrossRef
16. 16Mahnke CB, Nowalk A, Hofkosh D, et al. Comparison of two educational interventions on pediatric resident auscultation skills. Pediatrics. 2004;113:1331–1335. 17. 17Mangione S, Nieman LZ. Cardiac auscultatory skills of internal medicine and family practice trainees. A comparison of diagnostic proficiency. JAMA. 1997;278:717–722. MEDLINE 18. 18St Clair EW, Oddone EZ, Waugh RA, et al. Assessing housestaff diagnostic skills using a cardiology patient simulator. Ann Intern Med. 1992;117:751–756. MEDLINE 19. 19Attenhofer Jost CH, Turina J, Mayer K, et al. Echocardiography in the evaluation of systolic murmurs of unknown cause. Am J Med. 2000;108:614–620. Abstract | Full Text |
Full-Text PDF (94 KB)
|
CrossRef
20. 20Barron JT, Manrose DL, Liebson PR, et al. Comparison of auscultation with two-dimensional and Doppler echocardiography in patients with suspected mitral valve prolapse. Clin Cardiol. 1988;11:A30. MEDLINE |
CrossRef
21. 21Bloch A, Crittin J, Jaussi A. Should functional cardiac murmurs be diagnosed by auscultation or by Doppler echocardiography?. Clin Cardiol. 2001;24:767–769. MEDLINE |
CrossRef
22. 22Reichlin S, Dieterle T, Camli C, et al. Initial clinical evaluation of cardiac systolic murmurs in the ED by noncardiologists. Am J Emerg Med. 2004;22:71–75. Abstract | Full Text |
Full-Text PDF (123 KB)
|
CrossRef
23. 23Roldan CA, Shively BK, Crawford MH. Value of the cardiovascular physical examination for detecting valvular heart disease in asymptomatic subjects. Am J Cardiol. 1996;77:1327–1331. Abstract |
Full-Text PDF (568 KB)
|
CrossRef
24. 24Shry EA, Smithers MA, Mascette AM. Auscultation versus echocardiography in a healthy population with precordial murmur. Am J Cardiol. 2001;87:1428–1430. Full Text |
Full-Text PDF (77 KB)
|
CrossRef
25. 25Gadsbøll N, Høilund-Carlsen PF, Nielsen GG, et al. Symptoms and signs of heart failure in patients with myocardial infarction: reproducibility and relationship to chest X-ray, radionuclide ventriculography and right heart catheterization. Eur Heart J. 1989;10:1017–1028. 26. 26Ishmail AA, Wing S, Ferguson J, et al. Interobserver agreement by auscultation in the presence of a third heart sound in patients with congestive heart failure. Chest. 1987;91:870–873. MEDLINE |
CrossRef
27. 27Lok CE, Morgan CD, Ranganathan N. The accuracy and interobserver agreement in detecting the ‘gallop sounds’ by cardiac auscultation. Chest. 1998;114:1283–1288. MEDLINE |
CrossRef
28. 28Raftery EB, Holland WW. Examination of the heart: an investigation into variation. Am J Epidemiol. 1967;85:438–444. MEDLINE 29. 29Cohen MV, Spindola-Franco H. Correlation between left ventriculography, auscultation, and M-mode and two-dimensional echocardiography in mitral valve prolapse. Herz. 1988;13:293–308. MEDLINE 30. 30Dobrow RJ, Calatayud JB, Abraham S, et al. A study of physician variation in heart-sound interpretation. Med Ann Dist Columbia. 1964;33:305–308. MEDLINE 31. 31Forssell G, Jonasson R, Orinius E. Identifying severe aortic valvular stenosis by bedside examination. Acta Med Scand. 1985;218:397–400. MEDLINE 32. 32Griffiths RA, Sheldon MG. The clinical significance of systolic murmurs in the elderly. Age Ageing. 1975;4:99–104. MEDLINE 33. 33Maisel AS, Gilpin EA, Klein L, et al. The murmur of papillary muscle dysfunction in acute myocardial infarction: clinical features and prognostic implications. Am Heart J. 1986;112:705–711. MEDLINE |
CrossRef
34. 34Heidenreich PA, Schnittger I, Hancock SL, et al. A systolic murmur is a common presentation of aortic regurgitation detected by echocardiography. Clin Cardiol. 2004;27:502–506. MEDLINE |
CrossRef
35. 35Lembo NJ, Dell'Italia LJ, Crawford MH, et al. Bedside diagnosis of systolic murmurs. N Engl J Med. 1988;318:1572–1578. MEDLINE 36. 36Shub C. Echocardiography or auscultation? How to evaluate systolic murmurs. Can Fam Physician. 2003;49:163–167. 37. 37Stewart IM. Systolic murmurs in 525 healthy young adults. Br Heart J. 1951;13:561–565. MEDLINE |
CrossRef
38. 38Gjørup T, Bugge PM, Jensen AM. Interobserver variation in assessment of respiratory signs. Acta Med Scand. 1984;216:61–66. MEDLINE a Clinic of Cardiology, Rigshospitalet, Copenhagen Ø, Denmark b Clinic of Cardiology, Amager Hospital, Copenhagen S, Denmark c Department of Biostatistics, University of Copenhagen, Copenhagen K, Denmark d The Nordic Cochrane Center, Rigshospitalet, Copenhagen Ø, Denmark  Reprint requests: Kasper Iversen, MD, Clinic of Cardiology, Rigshospitalet, Blegdamsvej 9, DK-2100 Copenhagen Ø, Denmark.
The study was supported by unrestricted grants given by Pfizer Denmark, Medicon Denmark, and the Diamond Foundation. None of the grantors had any influence on the design and conduct of the study; collection, management, analysis, and interpretation of the data; as well as preparation, review, and approval of the manuscript. PII: S0002-8703(06)00347-4 doi:10.1016/j.ahj.2006.04.013 © 2006 Mosby, Inc. All rights reserved. | |
|
FULL TEXT