Diagnosis
Cury RC; Shash K; Nagurney JT et alCardiac magnetic resonance with T2-weighted imaging improves detection of patients with acute coronary syndrome in the Emergency Department
Circulation
2008; 118: 837 - 844
- Submitted by:Richard Body - Clinical Research Fellow
- Institution:Manchester Royal Infirmary
- Date submitted:17th November 2009
1
Objectives and hypotheses
1.1
Are the objectives of the study clearly stated?
Yes. The hypothesis is stated at the end of the Background and the objectives are, unusually, stated under 'Outcomes' in the methods section.
The primary goal was to assess the diagnostic accuracy of cardiac MR (CMR) for ACS. Additional goals were to examine whether CMR could carried incremental value when added to the clincal judgement of ED physicians and clinical 'risk factor' burden. Finally, the study also aimed to compare the diagnostic accuracy of different MR sequences.
The primary goal was to assess the diagnostic accuracy of cardiac MR (CMR) for ACS. Additional goals were to examine whether CMR could carried incremental value when added to the clincal judgement of ED physicians and clinical 'risk factor' burden. Finally, the study also aimed to compare the diagnostic accuracy of different MR sequences.
2
Design
2.1
Is the study design suitable for the objectives
Yes, this is a prospective diagnostic cohort study, which is an appropriate design to fulfil the stated objectives.
2.2
Who / what was studied?
Patients (n=62) presenting to the ED with acute chest pain, negative initial biomarkers and no acute ischaemic ECG changes. In addition, the MR scanner had to be available for use, patients had to be non-pregnant, able to breath hold for 10-15 seconds, haemodynamically and clinically stable and have no contra-indications to MR scanning.
All patients underwent cardiac MR in the ED using a detailed protocol that included perfusion scanning. Patients were subsequently given routine clinical care and clinicians blinded to the CMR results.
All patients underwent cardiac MR in the ED using a detailed protocol that included perfusion scanning. Patients were subsequently given routine clinical care and clinicians blinded to the CMR results.
2.3
Was this the right sample to answer the objectives?
Yes, although the list of exclusions (while justifiable) is fairly lengthy and noteworthy.
2.4
Is the study large enough to achieve its objectives? Have sample size estimates been performed?
No sample size calculation has been reported. The study sample is very small (n=62) with only 13 patients being diagnosed with ACS.
2.5
Were all subjects accounted for?
We don't know how many patients with suspected ACS but normal initial biomarkers and ECG were ineligible for the study. This is important - the potential impact of CMR is very limited if only a small proportion of patients are actually eligible to undergo the scan.
Otherwise, all of the numbers do add up.
Otherwise, all of the numbers do add up.
2.6
Were all appropriate outcomes considered?
No. The study is potentially subject to verification bias, as the protocol did not mandate that patients underwent troponin testing a set time following symptom onset. As such, it is possible that some acute myocardial infarctions were missed (it is not currently possible to judge this given the data actually presented).
Further, patients were not followed up following hospital discharge for important outcomes such as death, AMI and revascularisation (although in-hospital data were recorded). CMR would be of greatest value if it could help to identify patients with ACS who are currently being discharged from hospital. This study did not assess that important outcome.
Further, patients were not followed up following hospital discharge for important outcomes such as death, AMI and revascularisation (although in-hospital data were recorded). CMR would be of greatest value if it could help to identify patients with ACS who are currently being discharged from hospital. This study did not assess that important outcome.
2.7
Has ethical approval been obtained if appropriate?
Yes, and all patients provided written consent.
2.8
Was an independent blinded gold standard test applied to all subjects?
No. The gold standard was the diagnosis assigned by two experienced ED physicians who reviewed the patient data forms and medical records. They were blinded to CMR findings and disagreements were solved by consensus, with an additional cardiologist adjudicating if necessary. While this is a reasonable gold standard, the study protocol ought to mandate troponin testing a set time from symptom onset. Further, as unstable angina in particular is a relatively subjective diagnosis to establish, it would perhaps be better to use more objective outcomes such as death, AMI and/or revascularisation at a set time following initial presentation.
3
Measurement and observation
3.1
Is it clear what was measured, how it was measured and what the outcomes were?
The CMR protocol is described in detail and the reporting procedure is also well defined. The primary outcome is clearly stated.
3.2
Are the measurements valid?
Yes, it would appear so. The CMR protocol is well described and has been previously reported.
3.3
Are the measurements reliable?
The CMRs were evaluated by two 'experienced observers' who were blinded to the patient's name and all clinical data. The consensus opinion was taken, with disagreements resolved by a third 'expert reader'. There are no details provided regarding the number of discrepancies and there is no attempt to assess interobserver reliability.
3.4
Are the measurements reproducible?
There was no assessment of intraobserver reproducibility so we cannot assess whether the results are reproducible from this study alone.
4
Presentation of results
4.1
Are the basic data adequately described?
Table 1 provides a good breakdown of the baseline characteristics of the included patients. Time from symptom onset, time to CMR completion, length of hospital stay and ED physician gestalt were not reported, but should have been.
4.2
Are the results presented clearly, objectively and in sufficient detail to enable readers to make their own judgement?
Sufficient information is presented for the reader to be able to check calculations of sensitivity, specificity and predictive values. However, the methods section states that confidence intervals were calculated. These are important, given the small sample size, but are not reported.
4.3
Are the results internally consistent, i.e. do the numbers add up properly?
Yes.
5
Analysis
5.1
Are the data suitable for analysis?
Yes.
5.2
Are the methods appropriate to the data?
Yes, the methods seem generally appropriate. Calculation of sensitivity, specificity and predictive values is certainly appropriate, although confidence intervals should have been stated.
With regard to the multivariate regression analysis, it is not clear whether the study was adequately powered for this and the results are of uncertain significance. The reader really needs to know how many additional early diagnoses could be established using CMR when added to the clinical information already available. I'm not certain the analysis presented establishes this robustly.
With regard to the multivariate regression analysis, it is not clear whether the study was adequately powered for this and the results are of uncertain significance. The reader really needs to know how many additional early diagnoses could be established using CMR when added to the clinical information already available. I'm not certain the analysis presented establishes this robustly.
5.3
Are any statistics correctly performed and interpreted?
Yes, they appear to be.
6
Discussion
6.1
Are the results discussed in relation to existing knowledge on the subject and study objectives?
Yes.
6.2
Is the discussion biased?
The authors don't really clearly defined how CMR may be used to assist in diagnosis. How might it be used to influence disposition decisions? How many admissions (if any) might be prevented? What is the cost of the CMR? This is a significant issue - the CMR protocol may actually add to the cost of patient care and may not have a beneficial effect on patient care or save other investigations. What is the potential for CMR to identify patients with ACS who would otherwise have been missed? What proportion of patients are actually eligible for CMR? These important issues are not really addressed.
7
Interpretation
7.1
Are the authors' conclusions justified by the data?
No. I disagree that the study demonstrated 'the feasibility of CMR in the ED'. The authors have not demonstrated how it may influence patient care, have not commented on the proportion actually eligible, have not commented on cost or resource implications, have not suggested how it may have a beneficial effect on patient care, have used a very small sample size and used suboptimal outcome measures.
7.2
What level of evidence has this paper presented? (using CEBM levels )
3b
7.3
Does this paper help me answer my problem?
No. I am none the wiser.
8
Implementation
8.1
Can the test be implemented in practice?
No, it is expensive, requires specialist equipment and specialist expertise to report. The evidence presented does not currently suggest that it is ready for clinical implementation.
8.2
What aids to implementation exist?
None.
8.3
What barriers to implementation exist?
Cost; expertise; equipment; lack of evidence.
8.4
Are my patients the same as the patients tested?
It is hard to say. The baseline characteristics are roughly similar to similar UK studies.
8.5
Will the test improve diagnosis in my patients?
This study does not present convincing evidence that it will.