• 
• 

Three Part Question

In [a patient with suspected heart failure] does [a chest x-ray] aid [accurate diagnosis].

Clinical Scenario

A 74y/o lady with a significant medical history of ischaemic heart disease and COPD presents to the Emergency Deparment with acute shortness of breath. You are unsure whether this represents an episode of acute heart failure or an exacerbation of her COPD. You request an urgent CXR in the hope that it makes the diagnosis clearer so that you are able to confidently prescribe the appropriate treatment.

Search Strategy

Ovid MEDLINE 1966 to November Week 3.
[exp Radiography, Thoracic/ OR chest roentgenogram.mp. OR chest radiograph.mp. OR chest x-ray.mp. OR cxr.mp.] OR [[radiog$.mp. OR radiol$.mp.] AND [exp LUNG/ OR lung.mp. OR chest.mp. or exp Thorax]] AND [exp Heart Failure, Congestive/ OR heart failure.mp. OR exp Cardiac Output, Low/ OR exp Heart Failure, Congestive/ OR cardiac insufficiency.mp. OR exp Ventricular Dysfunction, Left/ OR lvf.mp. OR left ventricular failure.mp.] AND [exp "Sensitivity and Specificity"/ OR exp Diagnostic Tests, Routine/ OR diagnos$.mp. OR sensitivit$.mp. OR specificit$.mp. OR utili$.mp. OR diagnos$.mp.] LIMIT to humans and English language

Search Outcome

340 papers from MEDLINE search of which 7 were felt to be relevant.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Madsen, EB et al 1984 USA	229 patients discharged following a definite AMI. All patients underwent a chest XR and radionuclide ventriculography.	Chest XRs reviewed by individuals unaware of ventriculography results and graded from 0 to IV for pulmonary congestion and the cardiac silhouette was used to measure cardiac size relative to the internal diameter of the chest on the film (CTR).	Comparison of CTR >50% with LVEF <50% on ventriculography.	Sensitivity 47%, Specificity 66%, PPV 66%, LR 1.37
			Presence of pulmonary congestion with LVEF <50% on ventriculography	Sensitivity 20%, specificity 92%, PPV 77%, LR 2.39
Rihal C.S. et al 1995 USA	Study looking at 14507 patients presenting to hospital with chest pain looking at how effective widely available variables were at predicting LV function. All patients underwent diagnostic cardiac catheterisation. Normal LV function defined as LVEF >50%.	Retrospective study comparing clinical variables with LVEF. Cardio-thoracic ratio calculated from the chest x-ray and defined as normal (<50%) or abnormal (>50%).	Prediction of abnormal LVEF by an increased CTR.	Sensitivity 14%, specificity 92%, PPV 34%, LR 1.72	Retrospective study, not clear who examined chest x-rays and whether they were blind to other results. Error in paper in calculating sensitivity and specificity.
Ogawa, K et al 2002 Japan	130 patients who had been referred to a cardiology department as out-patients with symptoms suggestive of heart failure but not on any treatment at the time of referral.	Comparison of echocardiography findings with electrocardiography and chest radiography in diagnosing heart failure. Chest x-rays considered abnormal if CTR >50%, prominence of pulmonary vessel shadows or lung congestion present or if hydrothorax was present unaccompanied by signs of inflammation in the lung fields.	Prediction of diastolic dysfunction by abnormal chest x-ray.	Sensitivity 75%, specificity 96%, PPV 71%, LR 17.1	Of 130 patients who were considered to have symptoms suggestive of heart failure only 16 were felt to have diastolic dysfunction and no patients were felt to have systolic dysfunction although 86 of the patients were judged to have heart disease.
Thomas, JT et al 2002 USA	Examined a group of 225 patients who had been admitted to hospital with a diagnosis of congestive heart failure. 43 patients excluded from the study as did not have an echo done during this admission and further 14 patients excluded due to primarily valvular problems. Defined LVEF <45% as abnormal.	Chest x-rays examined by an attending radiologist who decided whether cardiomegaly and pulmonary oedema were present or not.	Prediction of reduced LVEF by presence of cardiomegaly on CXR	Sensitivity 90%, specificity 15%, PPV 55%, LR 1.06	No clear definition of heart failure given, not clear who made final diagnosis. LVEF estimated from visual inspection albeit by experienced cardiologists.
			Prediction of reduced LVEF by presence of pulmonary oedema	Sensitivity 14%, specficity 86%, PPV 55%, LR 1.04
Fonseca, C; et al 2004 Portugal	6300 patients attending their GP for a variety of different complaints were selected at random. Patients who scored 2 or more in the Boston questionnaire and/or were receiving diuretics for heart failure went on to have further investigation for heart failure.	1058 patients were included in the initial study but only 1022 of these had an echocardiogram carried out and in 174 there were incomplete measurements made due to technical difficulties. These patients were therefore excluded.	Prediction of presence of heart failure by an abnormal chest x-ray	Sensitivity 68%, specificity 53%, PPV 61%, LR 1.46	Although this paper states that it excluded the patients who did not have a satisfactory echocardiogram from the analysis it has included these patients in the data given about chest x-rays. It is not clear how the presence or absence of heart failure was decided in these patients. Even using these figures there are slight differences between the given sensitivity and specificity and results that I obtained from my own calculations.
			Prediction of presence of heart failure by a CTR >0.5	Sensitivity 44%, specificity 73%, PPV 64%, LR 1.6
Knudsen, CW et al 2004 USA	Study looks at subset of data from the Breathing Not Properly study by McCulloch et al (2004). The study looked at 880 patients presenting to the Emergency department with acute shortness of breath and assessed the utility of BNP and CXR in the diagnosis of heart failure.	Gold standard of heart failure defined by two cardiologists who reviewed all clinical data. Chest XRs reported by radiologists in relevant hospitals, assessed for cardiomegaly, cephalisation of vessels, interstitial oedema, alveolar oedema, pleural effu- sion, hyperinflated lungs, and pneumonic infiltrates. The presence of cardiomegaly or cephalisation performed best. LVEF assessed by echo and <50% considered abnormal.	Prediction of reduced LVEF by presence of cardiomegaly on CXR	Sensitivity 79%, specificity 80%, PPV 80%, LR 3.99
			Prediction of reduced LVEF by presence of cephalisation	Sensitivity 40%, specificity 96%, PPV 91%, LR 9.43
Knudsen, CW et al 2004 USA	Study looks at subset of data from the Breathing Not Properly study by McCulloch et al (2004). The study looked at 880 patients presenting to the Emergency department with acute shortness of breath and assessed the utility of BNP and CXR in the diagnosis of heart failure.	Gold standard of heart failure defined by two cardiologists who reviewed all clinical data. Chest XRs reported by radiologists in relevant hospitals, assessed for cardiomegaly, cephalisation of vessels, interstitial oedema, alveolar oedema, pleural effu- sion, hyperinflated lungs, and pneumonic infiltrates. The presence of cardiomegaly or cephalisation performed best. LVEF assessed by echo and <50% considered abnormal.	Prediction of reduced LVEF by presence of cardiomegaly on CXR	Sensitivity 79%, specificity 80%, PPV 80%, LR 3.99
			Prediction of reduced LVEF by presence of cephalisation	Sensitivity 40%, specificity 96%, PPV 91%, LR 9.43
Hoiland-Carlsen, PF et al 2005 Denmark	111 patients selected at random from survivors of AMI over a period of 15 months. These patients all had a chest x-ray and a radionuclide ventriculogram performed on the same morning on the second week after their AMI. The chest x-rays were reported by a senior radiologist who was unaware of clinical findings and the ventriculography result.	The radiologist assessed all chest x-rays for enlargement of the LV chamber size and pulmonary vascular congestion. Radionuclide ventriculography was used to estimate LV End Diastolic Volume and LV Ejection Fraction	Prediction of increased LVEDV by enlarged LV on CXR	Sensitivity 68%, specificity 64%, PPV 66%, LR 1.86	Unable to calculate sensitivity, specificity etc from data presented for pulmonary vascular congestion.
			Prediction of decreased LVEF by enlarged LV on CXR	Sensitivity 63%, specificity 68%, PPV 79%, LR 2.00
			Prediction of increased LVEDV by pulm vascular congestion on CXR	Significant tendency p<0.0002
			Prediction of decreased LVEF by pulm vascular congestion	Signficant tendency P<.0003.
Costanzo, WE; Fein, SA 1988 US	23 patients with NYHA grade 3 or 4 heart failure and a pulmonary capillary wedge pressure of 20mmHg or greater.	Patients all had a chest XR within 24h of catheter placement and before starting any further treatment. Chest XR report looked at retrospectively and descriptions of pulmonary oedema considered positive. No consideration taken of the presence or absence of cardiomegaly on the report.	Presence of pulmonary vascular congestion or oedema mentioned in the chest XR report	11 out of 23 patients had an abnormal report. (sensitivity 48%)	Retrospective , small numbers. Does not assess presence or absence of cardiomegaly. Up to 24h may have passed before having CXR. Reporting radiologists unaware that reports to be involved in study, not known what the request for xr stated.

Comment(s)

Chest XRs were generally judged in 2 ways. The first sign that was sought was the presence of cardiomegaly as decided by the cardiothoracic ratio. There was some variation on how this was measured, the internal diameter was either measured at the widest point or level with the highest point of the left hemidiaphragm. This was compared with the cardiac silhouette at its widest point. A ratio of >0.5 was considered abnormal. Chest XRs were also examined for pulmonary venous congestion. Most papers presented this in a graded system ranging from redistribution of blood flow to frank alveolar infiltrates. The gold standard was the ejection fraction as estimated by echocardiography, radionuclide ventriculography or right heart catheterisation. Although there was a definite tendency towards cardiomegaly and pulmonary venous congestion in patients with heart failure the overall accuracy of this test was disappointing. Specificity fared better than sensitivity but was only higher than 95% in one small study that only found a small number of patients with diastolic dysfunction.

Clinical Bottom Line

Chest XRs are a fast, relatively cheap, relatively safe and non-invasive form of investigation. They are useful for investigating the breathless patient but do not have the specificity or sensitivity to rule in nor rule out the diagnosis of heart failure by themselves. They should be interpreted in the context of the history and clinical findings from the patient.

Level of Evidence

Level 2 - Studies considered were neither 1 or 3.

References

1. Madsen, E. B. et al Usefulness of the chest x-ray for predicting abnormal left ventricular function after acute myocardial infarction American Heart Journal 1984. 108(6): p1431-6
2. Rihal, CS; Davis, KB; Kennedy, JW; Gersh, BJ The utility of the clinical, electrocardiographic and roentgenographic variables in the prediction of left ventricular function. American Journal of Cardiology 1995; 75(4): p220-3
3. Ogawa, K; Oida, A; Sugimura, H; Kaneko, N; Nogi, N; Hasumi, M; et al. Clinical significance of blood brain natriuretic peptide level measurement in the detection of heart disease in the untreated outpatients: comparison of electrocardiography, chest radiography and echo Circulation Journal 2002; 66(2):p122-6
4. Thomas, JT; Kelly, RF; Thomas, SJ; Stomas, TD;et al Utility of history, physical examination, electrocardiogram, and chest radiograph for differentiating normal from decreased systolic function in patients with heart failure. American Journal of Medicine 2002; 112(6):p437-45
5. Fonseca, C; Mota, T; Morais, H; Matias, F; Costa, C; Oliveira, AG; Ceia, F The value of the electrocardiogram and chest X-ray for confirming or refuting a suspected diagnosis of heart failure in the community The European Journal of Heart Failure 2004; 6(6) pp807-812
6. Knudsen, CW; Omland, T; Clopton, P; Westheim, A; Nowak, RM; Aumont, MC; Duc, P; Hollander, JE; Wu, AH; McCullough, P; Maisel, AS Diagnostic Value of B-Type Natriuretic Peptide and Chest Radiographic Findings in Patients with Acute Dyspnoea American Journal of Medicine 2004; 116(6):p 363-8
7. Knudsen, CW; Omland, T; Clopton, P; Westheim, A; Nowak, RM; Aumont, MC; Duc, P; Hollander, JE; Wu, AH; McCullough, P; Maisel, AS Diagnostic Value of B-Type Natriuretic Peptide and Chest Radiographic Findings in Patients with Acute Dyspnoea American Journal of Medicine 2004; 116(6):p 363-8
8. Hoilund-Carlsen, PF; Gadsboll, NS; Hein, E; Stage, P; Badsberg, JH; Jensen, BH Assessment of Left Ventricular Systolic Function by the Chest X-Ray: Comparison with Radionuclide Ventriculography Journal of Cardiac Failure 2005; 11(4):p299-305
9. Costanzo, WE; Fein, SA The Role of the Chest X-Ray in the Evaluation of Chronic Severe Heart Failure: Things Are Not Always As They Appear Clin Cardiol 1988; (11):pp486-88