|
How well can the chest radiograph diagnose left ventricular dysfunction? |
Badgett R G, Mulrow C D, Otto P M, Ramirez G |
|
|
Authors' objectives To review the diagnostic utility of the chest radiograph for left ventricular dysfunction.
Searching MEDLINE was searched from 1966 to February 1995 for English language articles using the following search terms: 'heart failure' as MeSH term or textword in abstract, and 'x-ray' or 'radiograph' as textword in title or abstract. Citations from two critical reviews (see Other Publications of Related Interest nos.1-2) were also checked, as were those from other primary and review articles, books and the authors' files. Experts who had published pertinent articles were contacted and data were requested where relevant.
Two secondary searches were also conducted: one for studies that addressed the inter-rater reliability of specific chest radiographic findings; and one for identifying the prevalence of either increased preload or decreased ejection fraction in the general population. The same sources as described above were used, although the former search utilised textbooks, seminal reviews and the references of identified articles; in the latter search, MEDLINE was searched using the exploded MeSH term 'radionuclide ventriculography' and the textwords 'systolic function', 'systolic dysfunction', 'ejection fraction' and 'echocardiogram'.
Study selection Study designs of evaluations included in the reviewNo inclusion criteria relating to the study design were specified.
Specific interventions included in the reviewStudies of chest radiography were eligible for inclusion.
Reference standard test against which the new test was comparedThe studies had to include an acceptable reference standard for the detection of heart failure to be included in the review. Acceptable reference standards were the measurement of ejection fraction by noninvasive testing (echocardiogram or radionuclide ventriculography), or by one of the following invasive pressure measurements of the left ventricular preload: left ventricular and end-diastolic pressure, left atrial pressure, or pulmonary capillary wedge pressure.
Participants included in the reviewPatients with suspected heart failure undergoing diagnostic chest radiography. Studies that consisted of more than 80% of patients with valvular heart disease were not included in the review.
Outcomes assessed in the reviewThe outcomes assessed were the estimated sensitivity and specificity for the diagnosis of increased preload and/or decreased ejection fraction, according to whether the radiographic findings were of cardiomegaly, redistribution and interstitial signs.
How were decisions on the relevance of primary studies made?The authors did not state how the papers were selected for the review, or how many reviewers performed the selection.
Assessment of study quality Studies included in the review were checked for the following methodological attributes.
Did the study population include a continuous spectrum of patients that included normal patients?
What cut-off value of the criterion standard was used to define abnormal?
Were the patients clinically stable between radiographic and criterion standard assessments?
Were the radiographs posteroanterior films?
Was the radiographic reading blinded to the criterion standard results?
Was the radiograph interpreted by an experienced radiologist or cardiologist?
Two readers independently abstracted each paper using a standardised data form. The authors did not state how any disagreements were resolved.
Data extraction Two readers independently abstracted each paper using a standardised data form. The categories of data extracted from the original studies were not stated clearly in the review but included: the clinical setting of the study population; the number of patients; reason for referral; criterion standard; whether the chest radiograph reading was blind to the criterion standard; the time between the radiograph and criterion standard; the presence of cardiomegaly, redistribution or interstitial signs; assessment of increased preload or reduced ejection fraction. For each study, 2x2 tables were constructed to calculate the sensitivity and specificity of the findings.
Methods of synthesis How were the studies combined?A quantitative synthesis of the studies was presented. The studies were grouped according to the clinical setting of the study population: acute myocardial infarction, elective cardiac evaluation (patients referred for dyspnoea or possible heart failure), and known severe systolic dysfunction (patients evaluated for possible cardiac transplantation). Within each clinical setting, the studies were grouped by reference standard and radiographic finding. Pooled sensitivities and specificities were calculated using the random-effects model of DerSimonian and Laird (see Other Publications of Related Interest no.3). The pre-test probability of heart failure was estimated from the prevalence of heart failure found in the secondary literature search. The estimated sensitivities and specificities were then used to calculate the post-test probability of heart failure based on the chest radiograph result.
How were differences between studies investigated?For each group of studies, Fast*Pro (see Other Publications of Related Interest no.4) was used to calculate the homogeneity (alpha 0.10).
Results of the review Twenty-nine studies (n=4,692) were included. In 16 of these the chest radiograph reading was blinded to the standard criterion result (n=1,417).
Redistribution best diagnosed increased preload with a sensitivity of 65% (95% confidence interval, CI: 55, 75) and a specificity of 67% (95% CI: 53, 79). Cardiomegaly best diagnosed decreased ejection fraction with a sensitivity of 51% (95% CI: 43, 60) and a specificity of 79% (95% CI: 71, 85). Inter-rater reliability was fair to moderate for redistribution and moderate for cardiomegaly. The clinical setting affected the results by decreasing the specificity of cardiomegaly to 8% in detecting increased preload in patients with severe systolic dysfunction. The absence of redistribution could only exclude increased preload in situations in which the suspicion (pre-test probability) of disease was less than 9%, whereas redistribution could confirm increased preload when the pre-test probability was greater than 91%. The absence of cardiomegaly could only exclude a reduced ejection fraction if the pre-test probability was less than 8%, whereas cardiomegaly could confirm a reduced ejection fraction if the pre-test probability was greater than 87%.
The heterogeneity test identified that many of the point estimates were not homogeneous. The point estimates were therefore calculated using the random-effects model of DerSimonian and Laird (see Other Publications of Related Interest no.3).
Authors' conclusions Redistribution and cardiomegaly are the best chest radiographic findings for diagnosing increased preload and reduced ejection fraction, respectively. Unfortunately, neither finding alone can adequately exclude or confirm left ventricular dysfunction in the usual clinical settings. Redistribution is not always reliably interpreted.
CRD commentary The review addressed an appropriate question with suitable inclusion and exclusion criteria, although these were not clearly stated in the article. The search strategy was limited to a single electronic database, but the authors made good use of non-electronic data sources. However, they will have failed to identify significant non-English language papers. Two reviewers extracted the data independently and the authors formally checked the validity of the studies included in the review. However, they did not present their findings for all the criteria listed in the review, nor did they check the significance of differences found between the studies, e.g. three of the largest studies were not blinded.
The individual studies were presented adequately in the review, except that the outcomes of the individual studies were omitted. The meta-analysis utilised in the review was appropriate, with checks for heterogeneity. It was unclear whether the results presented by the authors were derived, in part, from studies that were not presented in the review. The authors' conclusions appear justified by the findings of the review, although the estimation of the prevalence of heart failure in the population based on data from unrelated studies that were not described or detailed in the review weaken these findings, as it is not possible from the information provided to judge how reliable an estimate was used. In addition, the large number of unblinded studies may well have influenced the overall results.
Implications of the review for practice and research Practice: The authors concluded that the generalisability of the review's findings is limited by the differing degrees of expertise of clinicians and their differing definitions of abnormal. They specifically state 'Clinicians should continue to use the chest radiograph: however, it should be used in conjunction with other clinical findings'.
Research: The authors did not state any implications for further research.
Funding San Antonio Cochrane Center; Audie L. Murphy Memorial Veterans Affairs Hospital.
Bibliographic details Badgett R G, Mulrow C D, Otto P M, Ramirez G. How well can the chest radiograph diagnose left ventricular dysfunction? Journal of General Internal Medicine 1996; 11(10): 625-634 Other publications of related interest 1. Badgett RG, Lucey CR, Mulrow CD. Unpublished data. 2. Mulrow CD, Lucey CR, Farnett LE. Discriminating causes of dyspnea through the clinical examination. J Gen Intern Med 1993;6:383-92. 3. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177-88. 4. Eddy DM, Hasselblad V. Fast*Pro. New York (NY): Academic Press; 1992.
Indexing Status Subject indexing assigned by NLM MeSH Confidence Intervals; Diagnosis, Differential; Humans; Sensitivity and Specificity; Ventricular Dysfunction, Left /diagnosis /radiography AccessionNumber 11997000168 Date bibliographic record published 30/09/2004 Date abstract record published 30/09/2004 Record Status This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn. |
|
|
|