Background Patent foramen ovale (PFO) has been associated with otherwise cryptogenic stroke. The pathway of investigation of PFO varies among centres. We compared the sensitivity and specificity of transoesophageal echocardiography (TOE) and transcranial Doppler ultrasound (TCD) for the detection of PFO in patients with otherwise cryptogenic stroke who underwent subsequent right heart catheterisation.

Methods We retrospectively reviewed case notes of patients referred from a single institution between 2000–2008 for potential PFO closure with a diagnosis of stroke or transient ischaemic attack that was considered “cryptogenic” after investigation. All patients underwent TCD with IV bubble contrast, usually prior to TOE, and those with a significant right to left shunt (RLS) were referred on patient request for consideration of closure. We compared TOE and TCD against findings from right heart catheterisation, as the gold standard for PFO diagnosis.

Results From 79 patients identified, both TCD and TOE data were available in 39 (37 stroke, 2 TIA). Median age was 34 (range 16–38y). There were 15 (36%) males, 6 (15%) with hypertension, 14 (36%) with dyslipidaemia, 7 (18%) with a positive family history of stroke, 1 (3%) with coronary heart disease, 17 (44%) smokers, 11 (28%) with previous stroke and 16 (41%) with migraine. However, there were missing data for each risk factor parameter in 5 to 22 patients. Of the 39 cases, all had RLS on TCD, 25 had RLS on TOE, and 38 had PFO confirmed at cardiac catheterisation. One patient with RLS on TCD did not have a PFO on TOE or cardiac catheterisation. Therefore, in this population, TOE had a sensitivity of 66%, specificity of 100%, and a positive predictive value of 100%. TCD had a sensitivity of 100% and a positive predictive value of 97%. There was no difference in size of closure device between TOE RLS–positive patients compared to TOE RLS–negative (median 25mm in both groups, p=0.74 Mann–Whitney–U test, n=37).

Conclusions TCD with bubble test is sensitive and specific for the detection of right–to–left shunts due to PFO confirmed at right heart catheterisation. In this population, TOE had low sensitivity for RLS. There was no difference in physical PFO size between TOE positive and negative patients.