Pre-operative detection of valvular heart disease by anaesthetists

Transcription

1 Anaesthesia, 2006, 61, pages doi: /j x Pre-operative detection of valvular heart disease by anaesthetists W. A. van Klei, 1 C. J. Kalkman, 1 M. Tolsma, 1 C. L. G. Rutten 2 and K. G. M. Moons 1,3 1 University Medical Centre Utrecht, Department Peri-operative Care and Emergency Medicine, PO Box 85500, 3508 GA Utrecht, The Netherlands 2 Isala clinics, Weezenlanden Hospital, Department Anaesthesiology, PO Box 10500, 8000 G Zwolle, The Netherlands 3 University Medical Centre Utrecht, Julius Centre for General Practice and Patient Oriented Research, PO Box 85500, 3508 GA Utrecht, The Netherlands Summary We prospectively estimated the prevalence of heart murmurs in 2522 consecutive adult noncardiac surgery patients during pre-operative evaluation. Factors that contribute to the detection of a heart murmur were identified, and echocardiography was used to evaluate to what extent a murmur reflected presence of valvular heart disease. A cardiac murmur was detected in 106 patients (prevalence 4.2%, 95% CI: %). Multivariable logistic regression analyses showed that age and general physical impression were independently associated with detecting a murmur (p-values < 0.01). In 83 (79%) of the patients with a murmur, an echocardiographic diagnosis was available: 39% had aortic valve abnormalities, 24% had mitral valve regurgitation, 7% had other valvular heart disease and 30% did not have any abnormality. Thus, 58 of the 83 patients had valvular heart disease (positive predictive value using routine cardiac auscultation for diagnosing VHD: 70%, 95% CI: 59 79%). Murmurs in patients younger than 40 years never reflected valvular heart disease. Pre-operative cardiac auscultation seems only reasonable in patients aged 40 years or older. Subsequent echocardiography in these selected patients is necessary.... Correspondence to: Dr Wilton A. van Klei w.a.vanklei@umcutrecht.nl Accepted: 20 November 2005 Most causes of death and major morbidity during and after non-cardiac surgery result from cardiovascular events such as congestive heart failure, myocardial infarction and cerebrovascular accidents [1 3]. The aim of pre-operative evaluation is therefore to detect co-existing cardiovascular morbidity in surgical patients. Patient history and physical examination have been shown to be the most important tools to screen initially for such co-morbidity [2, 4, 5]. Cardiac auscultation is one of these screening instruments used to detect valvular heart disease (VHD) before surgery and has been recommended in all surgical patients [2, 6]. Various studies have shown that the presence of VHD is an independent risk factor for peri-operative morbidity and mortality in patients undergoing non-cardiac surgery [7 9]. Therefore, failure to detect a haemodynamically significant VHD before surgery is a major fear of anaesthetists. However, the probability of finding a rare disease or disorder during screening of large populations is influenced by the prevalence of that disease and by the accuracy of the screening instrument. Little is known about the prevalence of heart murmurs in surgical patients and how often the murmurs detected by the anaesthetist reflect VHD, i.e. the positive predictive value. Furthermore, patients suffering from haemodynamically significant VHD will typically have a decreased exercise tolerance that can be detected by a careful patient history. Therefore, it might well be that cardiac auscultation is not indicated in all surgical patients, but should be based on the anaesthetists impression of the patients health status (e.g. a reported decreased exercise tolerance). We first estimated the prevalence of heart murmurs detected by anaesthetists during pre-operative evaluation. To identify factors that contribute to the detection of a heart murmur, we compared the patient characteristics of patients with and without a murmur. To estimate the positive predictive value of the test, echocardiography was used to evaluate the extent to which a detected Journal compilation Ó The Association of Anaesthetists of Great Britain and Ireland 127

2 W. A. van Klei et al. Æ Pre-operative detection of VHD Anaesthesia, 2006, 61, pages murmur reflected the presence of VHD. Finally, the anaesthetists impression of the patients health status based on patient history was compared with the presence or absence of VHD. We expected that patients classified by the anaesthetist as healthy would not suffer from important VHD. Method Patients This prospective cohort study comprised 2522 consecutive adult surgical patients scheduled for surgery in a Dutch general hospital (Isala clinics, Zwolle). Patients from all surgical specialties were included, except for neuro- and cardiac surgery patients. All patients visited the outpatient pre-operative evaluation clinic on average 3 weeks before the scheduled surgery date. Data collection The hospital ethics committee approved the study and waived the need for informed consent. All pre-operative data were collected prospectively. For each patient visiting the outpatient pre-operative evaluation clinic during the study period, the medical receptionist documented the patients demographics, weight, height, heart rate and blood pressure (using an automated non-invasive blood pressure device). Subsequently, the anaesthetist read the medical record and systematically performed a detailed medical history. The medical history was targeted to detect cardiovascular, pulmonary, liver, renal and musculoskeletal diseases. Furthermore, patients were asked about previous diseases, previous surgery, smoking habits, use of alcohol and use of medications. After completion of the history taking, the anaesthetist documented his general physical impression of the patient on a three-point scale (healthy vital, not entirely healthy and poor ill). Subsequently, each patient underwent auscultation of the heart and lungs, and head and neck evaluation. Auscultation was performed in a quiet room and any sound different from a normal first and second heart sound was classified as a murmur. The anaesthetists were not asked to classify further the detected murmur or to make a clinical diagnosis. Finally, each patient was assigned an ASA physical status. All patients in whom a heart murmur was detected during auscultation by the anaesthetist were referred for echocardiography (reference standard for VHD), except for those patients in whom echocardiography had been performed less than 1 year before their visit to the preoperative evaluation clinic and for those scheduled for procedures under local anaesthesia. The latter category mostly included cataract surgery patients. If an echocardiogram could not be obtained before the scheduled surgery date, the surgery was delayed until the results of the echocardiogram were known. This is according to normal clinical practice. Intra- and postoperative data from patients with a preoperatively detected heart murmur were collected retrospectively. Any special considerations or complications that had occurred were documented. Outcome VHD was defined as any valvular abnormality detected by echocardiography other than mild (trace) insufficiencies of the mitral valve or tricuspid valve. Possible diagnoses were aortic, mitral, pulmonary or tricuspid valve stenosis or regurgitation, insufficient prosthetic valves, combined valvular abnormalities and no valvular abnormality. Echocardiography was performed by staff cardiologists. Analyses The prevalence of heart murmurs in the study population was estimated to identify factors that contribute to the detection of a heart murmur. In addition, the characteristics of patients with and without a murmur were compared using odds ratio (OR) and mean difference (MD) with 95% confidence intervals (95% CI). Multivariable logistic regression analysis was used to quantify the independent association (defined by a p-value < 0.10) of these patient characteristics with the presence or absence of heart murmurs detected by pre-operative cardiac auscultation. All patient characteristics that were significantly associated with the detection of a murmur in the univariate analyses were included in the regression model. The ability of the model to discriminate between patients with and without a murmur was quantified by using the area under the receiver operating characteristic curve (ROC area) [10 12]. The echocardiographic diagnoses were used to determine to what extent cardiac murmurs reflected VHD and the positive predictive value of cardiac auscultation in the detection of VHD was estimated. The anaesthetists impression of the patients health status based on patient history and stratified for age was compared with the outcome (VHD) using cross-tabulation. Finally, to give an impression of the outcome in patients with a murmur, the follow-up data of the patients with a murmur detected by pre-operative auscultation were described. Results Of the 2522 patients, 106 had a cardiac murmur detected by auscultation during physical examination by the anaesthetist, a prevalence of 4.2% (95% CI: %). 128 Journal compilation Ó The Association of Anaesthetists of Great Britain and Ireland

3 Anaesthesia, 2006, 61, pages W. A. van Klei et al. Æ Pre-operative detection of VHD Table 1 Characteristics of the patients visiting the outpatient pre-operative evaluation clinic (n = 2522). Values are mean (SD) or absolute numbers (%). No cardiac murmur (n = 2416) Cardiac murmur (n = 106) OR or MD (95% CI) Gender; men (%) 965 (41) 36 (34) 0.7 (0.5; 1.1) Age; years (SD) 53 (19) 69 (15) 16 (11.8; 19.2) Body Mass Index; kg.m )2 (SD) 29 (5) 28 (6) 0.8 ()0.2; 1.8) Systolic blood pressure; mmhg (SD) 151 (26) 166 (28) 15 (9.9; 20.2) Diastolic blood pressure; mmhg (SD) 85 (11) 85 (12) )0.4 ()2.6; 1.9) Heart rate; beats.min )1 (SD) 80 (15) 82 (15) 1.3 ()1.6; 4.3) General physical impression* (%) Healthy vital 1717 (71) 38 (36) reference Not entirely healthy 572 (24) 36 (34) 2.8 (1.8; 4.5) Poor ill 124 (5) 31 (30) 11 (6.8; 18.8) ASA physical status* (%) ASA (39) 3 (3) n.a. ASA (46) 38 (36) n.a. ASA (14) 57 (54) n.a. ASA 4 6 (0.2) 8 (8) n.a. *In 68 patients (3%) the general physical impression and ASA status were unknown. SD, standard deviation; OR, Odds Ratio for discrete variables; MD, Mean Difference for continuous variables; n.a., not applicable, as ASA status was assigned only after cardiac auscultation was performed and the medical record was interpreted. Patients with a detected murmur were less frequently men (34% vs. 41%) and were considerably older than those without a murmur (difference 16 years; Table 1). Their mean systolic blood pressure was significantly higher (difference 15 mmhg) and the general physical impression of patients with a murmur was more frequently classified as not entirely healthy or poor ill. In 68 patients (3%), the general physical impression and ASA status were not known. Age, gender, general physical impression and systolic blood pressure were entered in the multivariate logistic regression model to estimate their independent association with the presence or absence of a heart murmur. Age was included as a categorical variable (age < 40, and > 60 years). Age and general physical impression were independently associated with detecting a murmur (p-values < 0.01); gender and systolic blood pressure were associated to a lesser extent (p-values of 0.13 and 0.11, respectively). The OR for the age categories and > 60 years compared with age category < 40 years was 2.3 (95% CI: ) and 5.5 (95% CI: ), respectively. The OR for patients classified as not entirely healthy and poor ill compared with the patients classified as healthy vital was 1.7 (95% CI: ) and 7.0 (95% CI: ), respectively. The ROC area of the model including age, gender, general physical impression and systolic blood pressure was 0.79 (95% CI: ). The ROC area of the model including age and general physical impression only was 0.78 (95% CI: ). In 83 (79%) of the 106 patients with a detected murmur, an echocardiographic diagnosis was available Table 2 Echocardiographic diagnoses of the patients with cardiac murmurs detected at the pre-operative evaluation clinic (n = 83). Values are numbers (%) of the total population (n = 2522). Aortic valve stenosis 23 (0.9) Aortic valve stenosis and mitral regurgitation 4 (0.2) Aortic valve regurgitation 1 (0.1) Prosthetic aortic valve 4 (0.2) Mitral valve regurgitation 20 (0.8) Pulmonary valve stenosis 2 (0.1) Combined valvular insufficiencies 4 (0.2) No valvular abnormality 25 (1.0) (Table 2). In the remaining 23 patients, no echocardiography was performed, in general because they were scheduled for cataract surgery under local anaesthesia (12 patients, 52%). The other 11 patients were scheduled for other procedures under local anaesthesia. Thirty-two (39%) of the 83 patients with an available echo had aortic valve abnormalities, 20 (24%) had mitral valve regurgitation, six (7%) had other VHD and 25 (30%) did not have any valvular abnormality (Table 2). Thus, in total 58 of the 83 patients had VHD, resulting in a positive predictive value using routine cardiac auscultation for diagnosing VHD of 70% (95% CI: 59 79%). Table 3 shows the frequencies (prevalences) of a murmur and of a VHD across the nine different strata of age and general physical impression (the two most important characteristics associated with heart murmurs). Murmurs were infrequent (6 696 = 0.9%) in patients aged less than 40 years. None of these patients had VHD. In the 765 patients aged between 40 and 60 years, 10 had Journal compilation Ó The Association of Anaesthetists of Great Britain and Ireland 129

4 W. A. van Klei et al. Æ Pre-operative detection of VHD Anaesthesia, 2006, 61, pages Table 3 General physical impression and age compared to the prevalence of murmurs and valvular heart disease. Values are absolute numbers (%) of row totals (n = 2518).* General physical impression Murmur (n = 106) VHD (n = 58) Age < 40 years Healthy vital (n = 642) 5 (0.8) 0 (0) Not entirely healthy (n = 47) 0 (0) 0 (0) Poor ill (n = 7) 1 (14.3) 0 (0) Age years Healthy vital (n = 591) 10 (1.7) 5 (0.8) Not entirely healthy (n = 150) 1 (0.7) 1 (0.7) Poor ill (n = 24) 5 (20.8) 4 (16.7) Age > 60 years Healthy vital (n = 522) 23 (4.4) 13 (2.5) Not entirely healthy (n = 411) 35 (8.5) 20 (5.0) Poor ill (n = 124) 26 (21.0) 15 (12.1) *Data were not available for four patients. VHD (VHD prevalence = 1.3%, 95% CI: %). Of these 10 patients, three had aortic valve stenosis (prevalence of aortic valve stenosis in this age group: 0.4%, 95% CI: %) and these three patients were all classified as having a poor functional status (general physical impression). In the 1057 patients aged over 60 years, 48 patients had VHD (VHD prevalence = 4.5, 95% CI: %). Of this age group, 24 patients (2.3%, 95% CI: %) had aortic valve stenosis, in four cases combined with mitral valve regurgitation. This study was not powered to show any relationship with postoperative outcomes. However, to obtain an impression of the outcome in patients with a detected murmur, we retrospectively collected surgical data. In 59% (62 106) of the patients with a detected murmur, antibiotic prophylaxis against endocarditis was administered or additional intra-operative monitoring (such as central venous pressure or invasive arterial pressure measurement) was used, compared to 9% ( ) of the patients without a murmur. Of the 106 patients with a heart murmur, 103 (97%) received permission for surgery, whereas surgery was cancelled for three patients. One of these three patients was admitted to the hospital before the surgery date because of cardiac disease. For the other two patients, the reason for cancellation was unknown. For 100 of the 103 patients (97%) surgery was without complications. Two patients were scheduled before surgery for postoperative ICU admission. Both patients had severe cardiac comorbidity. One patient was referred unexpectedly to the ICU due to acute heart failure some days after a lower leg amputation. He was diagnosed before surgery with a murmur based on mitral valve insufficiency (established by echocardiography). This patient died in the ICU 26 days after surgery. One other patient who was operated for a dialysis shunt revision died after a hospital stay of 3 months due to cardiac failure and renal insufficiency. This patient was known before surgery to have aortic valve stenosis and mitral valve insufficiency. Surgery cancellations and postoperative complications did not occur in patients classified as healthy vital. Discussion We estimated the prevalence of heart murmurs detected by anaesthetists in non-cardiac surgery patients visiting the outpatient pre-operative evaluation clinic. A heart murmur was detected in 4% of these patients. In 70% of the patients, the detected murmur reflected VHD (positive predictive value). Older age and the anaesthetists general physical impression of the patient were independently related to the presence or absence of a murmur and VHD. As far as we know, this is the first study that attempts to estimate the prevalence of heart murmurs and, more specifically, the prevalence of murmurs that reflect VHD in a general surgical population. This study has, of course, some limitations. First, it should be noted that the estimated prevalence of murmurs (4%) is likely an underestimation of the true prevalence. Auscultation was performed by anaesthetists and it has been reported that even cardiologists detect only 80 90% of all heart murmurs [13 15]. This may imply that the true prevalence of heart murmurs in our study population should be at least 5% (4 0.8). Furthermore, in at least 5% of asymptomatic adults, heart murmurs are detectable [13]. On the other hand, 5% was within the confidence interval around the estimated prevalence of VHD of 4.2%. Second, one of our aims was to quantify the extent to which echocardiography confirms the clinical diagnosis of VHD in patients with a murmur detected by the anaesthetist. As a result, we only performed echocardiography in patients who had a murmur detected during auscultation. To estimate the true prevalence of VHD, it would be necessary to know the frequency of VHD missed by auscultation (false negatives). As a result of our study design, no inferences can be drawn about these false negative auscultations. This would require a large scale study in which all surgical patients, with and without a murmur, undergo echocardiography. Hence, we also could not estimate the sensitivity, specificity or negative predictive value of auscultation for the detection of VHD. Third, the medical records of the patients were available to the anaesthestists prior to auscultation. This might have influenced the accuracy of detecting a murmur. One could argue that the number of detected murmurs would have been lower had the anaesthetists been blinded to medical history. However, blinding the 130 Journal compilation Ó The Association of Anaesthetists of Great Britain and Ireland

5 Anaesthesia, 2006, 61, pages W. A. van Klei et al. Æ Pre-operative detection of VHD anaesthetist to medical history would be in sharp contrast to daily practice and certainly limit the applicability of the study results to clinical practice. It has widely been recommended that in diagnostic research settings one should not blind observers of diagnostic tests to prior test results if these prior test results are routinely available in practice [16, 17]. Moreover, the primary aim was not to estimate the accuracy of anaesthetists in detecting murmurs but to obtain an estimate of the prevalence of murmurs and VHD in a surgical population found during normal daily practice. Cardiac auscultation by the anaesthetist seems a reasonable screening tool to detect clinically relevant VHD in elderly patients (60 years or older) and possibly in patients aged between 40 and 60 years with a poor general physical impression. In our population, the 696 patients aged less than 40 years never had VHD (Table 3) and nearly all patients with aortic valve stenosis (24 of the 27 patients; 89%) were found in patients aged 60 years or older. The remaining three patients with aortic valve stenosis were found in the category between 40 and 60 years with a poor general physical impression. Hence, cardiac auscultation may need to be performed only in 42% of the patients visiting the pre-operative evaluation clinic, i.e. the patients aged over 60 years, and echocardiography can be withheld in 20% of the patients with a murmur, as 22 of the 106 patients with a murmur were aged less than 60 years. We suggest that patients older than 60 years should always be referred for further evaluation by echocardiography after detection of a murmur by pre-operative auscultation. We found an overall prevalence of aortic valve stenosis of 1.1% and a prevalence of 2.4% in patients aged 60 years or older. This latter percentage is comparable to that found previously in patients from the same age category (2%) [18 20]. Pre-operative knowledge of the presence of VHD is important in the maintenance of adequate coronary perfusion, to prevent tachycardia and bradycardia and to maintain blood pressure and normovolaemia in haemodynamically important aortic valve stenosis [6, 21]. In conclusion, pre-operative cardiac auscultation by the anaesthetist in patients aged younger than 40 years seems unnecessary, but it should be done in patients aged 60 years or older to select patients who might have a VHD. Subsequent echocardiography in these selected patients is necessary to establish or exclude a definite diagnosis of VHD to plan peri-operative care. References 1 Goldman L. Cardiac risk in noncardiac surgery: an update. Anesthesia and Analgesia 1995; 80: Roizen MF. Preoperative evaluation. In: Miller RD, ed. Anaesthesia. New York: Churchill Livingstone, 2000: Mangano DT. Perioperative cardiac morbidity. Anesthesiology 1990; 72: Klei van WA, Grobbee DE, Rutten CLG, et al. The role of history and physical examination in preoperative evaluation: much opinion and little evidence. European Journal of Anaesthesia 2003; 20: Klei van WA, Moons KGM, Rutten CLG, Kalkman CJ, Knape JTA, Grobbee DE. The effect of outpatient preoperative evaluation of hospital inpatients on cancellation of surgery and length of hospital stay. Anesthesia and Analgesia 2002; 94: Stoelting R, Dierdorf S. Valvular heart disease. In: Anesthesia and Co-Existing Disease. New York: Churchill Livingstone, 1993: Goldman L, Caldera DL, Nussbaum SR, et al. Multifactorial index of cardiac risk in noncardiac surgical procedures. New England Journal of Medicine 1977; 297: Detsky AS, Abrams HB, McLaughlin JR, et al. Predicting cardiac complications in patients undergoing non-cardiac surgery. Journal of General International Medicine 1986; 1: Detsky AS, Abrams HB, Forbath N, Scott JG, Hilliard JR. Cardiac assessment for patients undergoing noncardiac surgery. A multifactorial clinical risk index. Archives of International Medicine 1986; 146: Hosmer DW, Lemeshow S. Applied Logistic Regression. New York: J. Wiley, 1989: Harrel FE, Lee KL, Mark DB. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Statistics in Medicine 1996; 15: Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic curve. Radiology 1982; 143: Etchells E, Bell C, Robb K. Does this patient have an abnormal systolic murmur? Journal of the American Medical Association 1997; 277: Verheugt FW. The systolic heart murmur [in Dutch]. Nederlands Tijdschrift Voor Geneeskunde 1998; 142: Choudhry NK, Etchells EE. Does this patient have aortic regurgitation? Journal of the American Medical Association 1999; 281: Moons KGM, Grobbee DE. When should we remain blind and when should our eyes remain open in diagnostic research? Journal of Clinical Epidemiology 2002; 55: Loy CT, Irwig L. Accuracy of diagnostic tests read with and without clinical information: a systematic review. Journal of the American Medical Association 2004; 292: Lindroos M, Kupari M, Heikkila J, Tilvis R. Prevalence of aortic valve abnormalities in the elderly: an Journal compilation Ó The Association of Anaesthetists of Great Britain and Ireland 131

6 W. A. van Klei et al. Æ Pre-operative detection of VHD Anaesthesia, 2006, 61, pages echocardiographic study of a random population sample. Journal of the American College of Cardiologists 1993; 21: Stewart BF, Siscovick D, Lind BK, et al. Clinical factors associated with calcific aortic valve disease. Cardiovascular Health Study. Journal of the American College of Cardiologists 1997; 29: Otto CM, Bonnie KL, Kitzman DW, Gersh BJ, Phil D, Siscovick DS. Association of aortic-valve sclerosis with cardiovascular mortality and morbidity in the elderly. New England Journal of Medicine 1999; 341: Torscher LC, Shub C, Rettke SR, Brown DL. Risk of patients with severe aortic stenosis undergoing noncardiac surgery. American Journal of Cardiology 1998; 81: Journal compilation Ó The Association of Anaesthetists of Great Britain and Ireland