Presented in part at the Annual Congress of the Association of Anaesthetists of Great Britain and Ireland, Harrogate, September 2010.
The role of routine pre-operative bedside echocardiography in detecting aortic stenosis in patients with a hip fracture*
Version of Record online: 24 OCT 2011
Anaesthesia © 2011 The Association of Anaesthetists of Great Britain and Ireland
Volume 67, Issue 1, pages 51–54, January 2012
How to Cite
Loxdale, S. J., Sneyd, J. R., Donovan, A., Werrett, G. and Viira, D. J. (2012), The role of routine pre-operative bedside echocardiography in detecting aortic stenosis in patients with a hip fracture. Anaesthesia, 67: 51–54. doi: 10.1111/j.1365-2044.2011.06942.x
You can respond to this article at http://www.anaesthesiacorrespondence.com
- Issue online: 12 DEC 2011
- Version of Record online: 24 OCT 2011
- Accepted: 12 September 2011
The prevalence and severity of aortic stenosis in unselected patients admitted with a hip fracture is unknown. Derriford Hospital operates a routine weekday, pre-operative, targeted bedside echocardiography examination on all patients admitted with a hip fracture. We carried out a prospective service evaluation for 13 months from October 2007 on all 501 admissions, of which 374 (75%) underwent pre-operative echocardiography. Of those patients investigated, 8 (2%) had severe, 24 (6%) moderate and 113 (30%) had mild aortic stenosis or aortic sclerosis. Eighty-seven of 278 (31%) patients with no murmur detected clinically on admission had aortic stenosis on echocardiography and of the 96 patients in whom a murmur was heard pre-operatively, 30 (31%) had a normal echocardiogram. Detection of a murmur does not necessarily reflect the presence of underling aortic valve disease. However, if a murmur is heard then the likelihood of the lesion's being moderate or severe aortic stenosis is increased (OR 8.5; 95% CI 3.8–19.5). Forty-four (12%) of our unselected patients with fractured femur had either moderate or severe aortic stenosis (with or without moderate or severe left ventricular failure), or mild stenosis with moderately or severely impaired left ventricular function.
The management of patients admitted with a surgical diagnosis of hip fracture and an undiagnosed heart murmur is controversial. In a UK survey, when presented with the above scenario 20% of respondents would insist on a pre-operative echocardiogram, 54% would request an echocardiogram before surgery only if the patients had suspicious signs or symptoms, and 26% would go ahead without investigation . This variation in practice reflects the dilemma of waiting for echocardiography to provide definite diagnosis and potentially alter the anaesthetic technique , vs the risk of increased mortality from operative delay .
The main anaesthetic concern aroused by an undiagnosed heart murmur is the possibility of severe aortic stenosis. This has previously been shown to be a major risk factor for morbidity , and the diagnosis can change anaesthetic management. McBrien and colleagues reported recently the echocardiographic findings of patients admitted with fractured hips and new systolic murmurs, and found the incidence of previously undiagnosed aortic stenosis to be 6.9% . They suggested that the true incidence in the hip fracture population might be greater, as patients with known aortic stenosis were not included, and there may have been a failure of murmur identification on auscultation. The true prevalence of aortic stenosis in this population of patients is unknown.
We audited our practice at Derriford Hospital when only patients with a clinically detected systolic murmur had an echocardiogram. These patients waited a mean (SD) of 5.4 (3.4) days after admission for their echocardiogram, and the mean time to surgery from admission was 7.5 (5.5) days. In an attempt to reduce the time to surgery, a routine weekday, pre-operative, bedside, targeted echocardiography service was developed to examine all patients admitted with a hip fracture. A service evaluation was conducted to assess this new intervention. This evaluation aimed to assess the prevalence and severity of aortic stenosis, the degree of left ventricular dysfunction, and the frequency with which a murmur was heard in all patients admitted with a diagnosis of hip fracture during the study period.
Local Ethics Committee approval was requested for this service evaluation but was not required. The Caldicott Guardian approved the data handling procedures. All patients who were admitted with the surgical diagnosis of fractured neck of femur, for 13 months from October 2007, were included. Data collected from the case notes included patients' characteristics (age, sex, ASA physical status), clinical record of murmurs detected pre-operatively (by admitting doctor, orthogeriatrician or anaesthetist), previous diagnosis of valvular heart disease, previous echocardiogram reports, date and time of admission, and the new echocardiogram report.
Senior echocardiography technicians performed echocardiography at the patient’s bedside on the orthopaedic ward. This was carried out between 08:30 and 09:00 each morning of the working week (Monday to Friday inclusive). The written report was placed in the patient’s notes. A Vivid I Ultrasound machine (GE Healthcare Ltd, Hatfield, UK) was used. This is a compact, portable, high performance digital ultrasound device. It provides image acquisition in 2D, M-mode and Doppler within a range of operating frequencies. During the targeted echocardiogram, left ventricular function and aortic valve structure and function were evaluated. Aortic stenosis was defined as an abnormality of the aortic valve causing obstruction to the left ventricular outflow. Aortic stenosis was measured by peak gradient across the valve and expressed as no stenosis, mild (< 36 mmHg), moderate (36–64 mmHg) or severe (> 64 mmHg). Patients with aortic sclerosis were included in the mild aortic stenosis group. Left ventricular ejection fraction was measured and expressed as normal (> 50%), mild (40–50%), moderate (30–40%) and severe (< 30%) impairment.
Following the introduction of the routine echocardiography service, the times from admission to echocardiography and to surgery were recorded.
The effectiveness of auscultation for detection of aortic stenosis was explored using logistic regression (using ‘R: A language and environment for statistical computing’. R Foundation for Statistical Computing, Vienna, Austria).
In the 13-month study period (from October 2007 to November 2008), 501 consecutive hip fracture patients were admitted to the hospital. Data were collected on 495 cases; six case notes were missing. Three hundred and seventy-four patients underwent pre-operative echocardiography; it was not obtained in 121. The principal reason for failure to obtain an echocardiogram was admission to hospital on a Friday or Saturday and subsequent surgery during the weekend; in three patients the images obtained were of a poor quality and they were not included in the further analysis.
The median (IQR [range]) age of patients was 84 (77–88 [20–102]) years. The majority (94%) of our patients were > 60 years old. One hundred and fourteen (23%) patients were male and 381 (77%) female.
The overall prevalence of aortic stenosis was 145/374 (39%). Severe or moderate aortic stenosis was detected in 8% and severe or moderate left ventricular impairment was detected in 7% of patients (Table 1). Of the 24 patients with moderate aortic stenosis, five had moderately or severely impaired left ventricular function (Table 2). None of the patients with severe stenosis had moderate or severely impaired function. A murmur was detected in 113 (22%) of the 495 admitted patients and in 96 (26%) of the 374 patients who underwent echocardiography. Of the eight patients with severe aortic stenosis, a murmur was noted in seven (Table 3), and of the 24 patients with moderate stenosis, a murmur was noted in 16. Overall, 87 (31%) patients with no murmur heard during clinical examination had aortic stenosis or sclerosis on echocardiography, of which nine were moderate or severe. Auscultation of a murmur was associated with a greater risk of moderate or severe aortic stenosis (OR 8.5; 95% CI 3.8–19.5).
|n = 374|
|No abnormality detected||205 (55%)|
|Mild stenosis or sclerosis||113 (30%)|
|Moderate stenosis||24 (6%)|
|Severe stenosis||8 (2%)|
|Other valve lesions||24 (6%)|
|Left ventricular function|
|Mild impairment||147 (39%)|
|Moderate impairment||23 (6%)|
|Severe impairment||3 (1%)|
|Degree of aortic stenosis||LV function||Number|
|None (n = 205)||Good||123 (60%)|
|Mild impairment||74 (36%)|
|Moderate impairment||8 (4%)|
|Mild stenosis or sclerosis (n = 113)||Good||43 (38%)|
|Mild impairment||58 (51%)|
|Moderate impairment||11 (10%)|
|Severe impairment||1 (1%)|
|Moderate stenosis (n = 24)||Good||8 (33%)|
|Mild impairment||11 (46%)|
|Moderate impairment||3 (12%)|
|Severe impairment||2 (8%)|
|Severe stenosis (n = 8)||Good||6 (75%)|
|Mild impairment||2 (25%)|
|Other valve lesion (n = 24)||Good||21 (87%)|
|Mild impairment||2 (8%)|
|Moderate impairment||1 (4%)|
|No abnormality||Mild stenosis or sclerosis||Moderate stenosis||Severe stenosis||Other lesion|
|Murmur heard (n = 96)||30 (31%)||35 (36%)||16 (17%)||7 (7%)||8 (8%)|
|No murmur heard (n = 278%)||175 (63%)||78 (28%)||8 (3%)||1 (0.5%)||16 (6%)|
Following the introduction of the routine echocardiography service the mean (SD) time to echocardiography was 1.0 (0.7) days and the time to surgery was 2.9 (1.9) days.
Aortic stenosis is increasingly common in our ageing population and its severity increases with age. Before our report, the prevalence and severity of aortic stenosis in an unselected patient population admitted with a surgical diagnosis of hip fracture was unknown. We found that 8% of patients who underwent echocardiography had either moderate or severe stenosis. Since we investigated all patients present when the echocardiography technician was available, our data should reflect the entire population of fractured femur patients.
A population-based sample from Finland of 501 people reported echocardiographic findings of moderate to severe aortic valve stenosis in 8.8% of females and 3.6% of males in patients aged 75–86 years . Data from the USA showed an increasing incidence of aortic stenosis after the age of 65 years in the general population, with an incidence in the over 75s of 2.8% . McBrien and his team studied a large population of patients who were admitted with a fractured neck of femur. Echocardiograms were performed in 1167 of 3997 cases, the majority for assessment of an undiagnosed murmur. The incidence in this selected group was 6.9% . Our findings are therefore comparable with previous studies. Importantly, our patients were unselected.
Patients with severe aortic stenosis and a low cardiac output often present with a low transvalvular pressure gradient . This is a function of underestimation of aortic jet velocity on echocardiogram (the impaired ventricle fails to generate enough pressure to create the ‘real’ gradient between the left ventricle and the aorta). Such patients can be difficult to distinguish from those with low cardiac output and only mild to moderate aortic stenosis. Low-dose dobutamine stress echocardiography can be used to determine whether the actual stenosis is severe or moderate , but this may lead to greater delays and its associated risks. By estimating both aortic stenosis and left ventricular function we hoped to avoid underestimating the severity of disease in patients with apparently mild stenosis. While our study identified five patients with moderate or severe left ventricular impairment amongst those with moderate or severe aortic stenosis, we found a further 12 patients with such a degree of impaired left ventricular function in patients whose stenosis was mild. Therefore, 44/374 (12%) of unselected patients with fractured femur had either moderate or severe aortic stenosis (with or without moderate or severe left ventricular failure), or mild stenosis with moderately or severely impaired left ventricular function.
Classically, a loud (grade 4/6), late-peaking systolic murmur radiating to the carotid arteries, a single or paradoxically split second heart sound, and a delayed and diminished carotid upstroke, confirm the presence of severe aortic stenosis. However, physical examination findings are specific but not sensitive for making the diagnosis and determining its severity . In previous work, it is usually only those patients who had a previously undiagnosed systolic murmur that would have undergone pre-operative echocardiography . Our evaluation revealed that 87/278 (31%) of patients without clinically detected murmurs had aortic stenosis. Conversely, in the 96 patients where a murmur was heard, 30 (31%) had a normal echocardiogram. Thus, pre-operative auscultation of a murmur does not necessarily reflect either the presence or the severity of the underlying aortic valve disease. However, if a murmur is heard, then moderate or severe aortic stenosis is 8.5 times more likely.
A busy ward is not the best place for a full cardiological examination, and the grade and experience of the examining doctor varies considerably. However, our data represent the ‘real world’ and reflect the realities of a busy surgical service.
After establishing our routine echocardiology service, time to surgery dramatically reduced. Although part of this improvement was a function of reduced anaesthetic delay as a result of the echocardiography, other factors also contributed. We did not evaluate whether the echocardiography result led to a change in anaesthetic practice, but the presence of aortic stenosis or left ventricular dysfunction did alter pre-operative consultations and the process for obtaining informed consent, as well as targeting invasive monitoring and increasing clinicians' vigilance.
In conclusion, we have demonstrated a point prevalence of aortic stenosis of 39% in our hip fracture population, of whom 8% had moderate or severe stenosis.
The authors wish to thank Linda Zacharkiw, Senior Echocardiology technician, and her team for performing these echocardiograms. No external funding and no competing interests declared.
- 4National Confidential Enquiry into Perioperative Deaths. Changing the Way we Operate. The 2001 Report of the National Confidential Enquiry into Perioperative Deaths. London: NCEPOD, 2001.