Saturday 5 September
Patent foramen ovale in patients with haemodynamically significant pulmonary embolism
The prevalence of a patent foramen ovale is about 1 in 4. In cases with venous thromboembolism and raised right heart pressures, a patent foramen ovale may permit paradoxical emboli, which could complicate the course of patients with pulmonary embolism. Echocardiography enables detection of a patent foramen ovale in life. We have studied 85 patients who presented with
haemodynamically significant pulmonary embolism as judged by clinical, echocardiographic, or haemodynamic indices and who had an echocardiographic evaluation for patent foramen ovale. 33 patients (39%) had a patent foramen ovale. Clinical symptoms suggestive of paradoxical embolism were more likely in patients with than in those without a patent foramen ovale (39% vs 6%, p=0·00034), with new neurological deficits occurring in 11 patients (9 vs 2, p=0·005) and a vascular occlusion in 8 (7 vs 1, p=0·0096). Arterial oxygen tension was lower in patients with a patent foramen ovale (mean 55 [SD 14] vs 62  mm Hg, p=0·038). Mortality was not different between the two groups (27% vs 19%). Cardiopulmonary complications in terms of resuscitation, intubation, or the use of catecholamines were more frequently observed in patients with a patent foramen ovale (48% vs 23%, p=0·028). Patients with
haemodynamically significant pulmonary embolism are more likely to have arterial hypoxaemia and vascular occlusions, possibly due to paradoxical emboli.
Introduction foramen ovale can be detected by echocardiography.1 An atrial level shunt may have serious consequences (eg, paradoxical embolism2-4) in disorders such as venous thromboembolism. In addition, a patent foramen ovale may cause atrial hypoxaemia either in the absence of pulmonary arterial hypertensions,6 or, more importantly, when right-sided pressures are high.7,8Apart from a few reports of a patent foramen ovale in patients with venous thromboembolism and paradoxical emboli,3,9,10 no systematic attempt has been made to explore the clinical sequences of a right-to-left shunt at atrial level in patients with pulmonary embolism. We have prospectively investigated the hypothesis that patients with haemodynamically significant pulmonary embolism and a patent foramen ovale are more likely to have vascular occlusions as a possible sequence of paradoxical emboli. A
Patients and methods We investigated 85 patients who presented at our institution between May, 1988, and October, 1991, with haemodynamically significant pulmonary embolism and who had echocardiographic evaluation for a patent foramen ovale. This group consisted of both emergency patients and inpatients. There were 45 men and 40 women, ranging in age from 19 to 85 years (mean 55 [SD 18]). Diagnosis of pulmonary embolism was based on: pulmonary
angiography (33), necropsy (3), a high-probability lung scan with invasively confirmed pulmonary arterial hypertension (21), a high-probability lung scan and echocardiographic and/or doppler criteria for pulmonary arterial hypertension (17), echocardiographic and doppler criteria for pulmonary arterial hypertension that were
Medical Clinic III, Albert-Ludwigs University, Hugstettertstr 55, 78 Freiburg, FRG (Prof W. Kasper, MD, A. Geibel, MD, N. Tiede, MD, Prof H Just, MD) Correspondence to Prof W Kasper
confirmed invasively (10), or echocardiographic and doppler criteria for pulmonary arterial hypertension alone (1).To establish the diagnosis of haemodynamically significant pulmonary embolism either of the following criteria had to be fulfilled: (1) persistent hypotension ( < 90 mm Hg for over 15 min) not due to hypovolaemia and septic shock, or the need for cardiopulmonary resuscitation, or the continued support of catecholamines; (2) echocardiographic and/or doppler indices suggesting pulmonary arterial hypertension without significant left-heart disease and/or invasively confirmed pulmonary arterial hypertension, with either a mean pulmonary arterial pressure over 20 mm Hg or a total pulmonary vascular resistance above 200 dyn s cm-’ (20 kPa 1’’ s). Clinical information was obtained prospectively. Echocardiographic studies were done with 225, 2-5, 3-5, or 5 MHz phased-array transducers (Diasonic Echoscope CV 4300 or Toshiba SSH 160 A). Patients were evaluated for possible thrombi within the cardiac chambers and the great vessels.ll Pulmonary arterial hypertension was assumed if one of the following criteria were present: (1) dilatation of the right ventricular cavity (right ventricle -> left ventricle) on an apical or subcostal four-chamber view, or a right ventricular end-diastolic diameter over 30 mm on a precordial echocardiogram, with dilatation of the right pulmonary artery (-> 12 mm/m2) on a suprastemal echocardiogram;(2) tricuspid regurgitation jet velocity over 2-5 m/s without inspiratory collapse of inferior vena cava; and (3) right ventricular free wall thickness over 5 mm. Tricuspid regurgitation jet velocity was quantified by continuous wave doppler.13 Two-dimensional echocontrast was used to detect a patent foramen ovale. Right-sided echocontrast was generated with 10 ml plasma expander (Oxypolygelatine 5-5%). The solution was agitated between 2 syringes mounted on a three-way stop-cock immediately before injection into either a peripheral vein or a central venous line. Macroscopic air was extruded to avoid air emboli. Echocontrast was considered adequate when the entire right atrium was opacified for at least three consecutive cardiac cycles. With a patent foramen ovale, microbubbles can appear within the last atrium.! A Valsalva manoeuvre was done in cooperative patients to further increase the likelihood of detecting a right-to-left shunt.! If an impending paradoxical embolus was observed, a patent foramen ovale was assumed and echocontrast was not done. The echocontrast studies had no side-effects. The technique has been routine here since 1983. The echocardiograms were stored on videotapes and reviewed by two experienced investigators, who were unaware of the final diagnosis. The echocardiographic diagnosis was made by consensus.
TABLE I-INITIAL CLINICAL PRESENTATION OF PATIENTS WITH HAEMODYNAMICALLY SIGNIFICANT PULMONARY EMBOLISM
TABLE II-HAEMODYNAMIC PROFILE OF PATIENTS WITH PULMONARY EMBOLISM
RAP=nght atnal pressure, PAP = pulmonary arterialpressure, CO= cardiac output, and PVR= total pulmonary vascular resistance.
The right heart was generally catheterised via an antecubital vein. Right-side pressures and cardiac output were measured by a flow-guided pulmonary artery thermodilution catheter. Bilateral pulmonary cineangiograms were done with a Berman catheter. A ventilation/perfusion lung scan was done with inhalation of 99mTc macroaggregated albumin followed by perfusion scans with intravenously administered 99mT c-diethylentriaminpentaacetic acid in the anterior, posterior, and lateral projections. A high-probability lung scan was defined as a ventilation/perfusion mismatch of at least one lobar defect or multiple segmental defects.!4 Data with a normal distribution are expressed as mean (SD); skewed data as median and interquartile range. Results were analysed with X2, Fisher’s exact, and unpairedt tests as appropriate. Adjusted comparisons were done by analysis of covariance. For a confirmatory test of variables, indicating peripheral, visceral, or cerebral vascular occlusions and arterial oxygen tension (PaOz), a p value below 0-05 was regarded as significant. For all other variables, p values were considered to be "exploratory".
Echocardiographic and clinical findings A patent foramen ovale was diagnosed in 33 patients (39%) because of either a positive echocontrast (n 30) or a visible thrombus straddling a patent foramen ovale. Presenting signs and symptoms of pulmonary embolism were not different between patients with and without a patent foramen ovale (table l). Most patients had symptoms suggestive of an acute pulmonary embolism within 5 days before study. Besides symptoms related to pulmonary embolism, 6 patients (7%) had a neurological deficit and/or arterial occlusion (5 patients with and 1 patient without a patent foramen ovale). Most patients in both groups had =
TABLE III-IN-HOSPITAL COURSE OF PATIENTS WITH HAEMODYNAMICALLY SIGNIFICANT PULMONARY EMBOLISM I
Mean (SD, range) *p=0 038, absent
p. U bb, TU UUU4. W UUb, vs
§U UU9ö, ’IU U28
predisposing conditions for pulmonary embolism, with no
apparent group differences (76% vs 69%). Mean arterial oxygen tension (in 61 patients, 72%) was significantly lower in patients with than in those without a patent foramen
patients (71%) had right-heart catheterisation with haemodynamic measurements. 53 (62%) had pulmonary arterial hypertension, of whom 20 had a patent foramen 60
ovale. There were no differences between the groups in any haemodynamic variable (table II). According to the echocardiographic and/or doppler criteria, 73 patients (86%) had evidence of pulmonary arterial hypertension, of whom 29 had a patent foramen ovale. 15 right-sided thrombi were detected in 13 patients. A thromboembolus in the pulmonary artery was identified in 9 cases, a right atrial thrombus in 4 cases, and 1 each in the right ventricle and superior vena cava.
Clinical course In total, 19 patients (22%) died in hospital. There was no difference in mortality in the two groups (table III). In-hospital course seemed to be more complicated for patients with a patent foramen ovale. Cardiopulmonary complications in terms of resuscitation, intubation, or use of catecholamines occurred significantly more often in patients TABLE IV-CLINICAL PRESENTATION PRESUMPTIVE FOR PARADOXICAL EMBOLISM
patent foramen ovale than in those without. There differences between the groups for therapeutic interventions. Disabling bleeds were equally present, but 2 cerebral bleeds occurred in patients with a patent foramen ovale. In 1 of these patients, the cerebral event developed after thrombolysis for pulmonary embolism; in the other patient, a secondary haemorrhage of a cerebral infarction occurred while she was on heparin. a
Paradoxical embolism Clinical symptoms presumptive for a paradoxical embolism occurred in 13 patients with a patent foramen ovale and in 3 patients without (39% vs 6%, p < 0-00034). 9 patients with and 2 without a patent foramen ovale had neurological deficits (p=0005) (table IV). 7 patients with a patent foramen ovale had a confirmed peripheral or visceral arterial occlusion. Surgical embolectomy was done in 4 patients. 1 patient (no 7) had arterial occlusion confirmed on digital subtraction angiography. Either multiple arterial occlusion (no 12) or peripheral arterial occlusion (no 13) was observed at necropsy in 2 patients. In another 2 patients, renal (no 9) or cerebral infarction (no 10) related to arterial emboli were confirmed at necropsy. Cerebral infarction was demonstrated in 4 patients (no 5, 6, 7, and 11) with a patent foramen ovale and in 2 (no 14 and 16) without a patent foramen ovale. Only 2 patients (no 3 and 4) had no confirmatory findings for arterial thromboemboli but clinically they had sensory aphasia for about an hour in one case and a motor aphasia with seizure and a short respiratory arrest in the other. A patent foramen ovale was found in 4 patients (no 9, 10, 12, and 13) at necropsy, in accordance with the echocardiographic evaluation during life.
recovered. D died CT =
Paradoxical embolism is commonly first diagnosed at necropsy.1S The diagnosis can usually be established during life when an impending paradoxical embolus is identified,3 in which case the situation can sometimes be retrieved surgically.16,17 The most likely cause of passage of venous clots into the systemic circulation is a patent foramen ovale. The frequency of a patent foramen ovale in necropsy series ranges from 25 to 35%,’8,’9 whereas that in healthy volunteers studied by echocardiography ranges between 5 and 31%.° Furthermore, young adults with unexplained stroke, and divers with decompression sickness have a patent foramen ovale frequency ranging from 50 to 61 %,20,-22 making paradoxical thromboembolism and/or air embolism the most likely mechanism. The 39% occurrence of a patent foramen ovale in our study is slightly higher than that in necropsy series in healthy volunteers. Pulmonary arterial hypertension, which was observed in most of our patients, might have increased the likelihood of our detection of a patent foramen ovale. Furthermore, the frequency of a patent foramen ovale that we found might also have been increased because affected patients with haemodynamically significant pulmonary embolism might have more clinical complications to attract medical attention. Similarly, pulmonary arteriovenous malformation can result either in left-sided echocontrast24 or in paradoxical emboli.3,9However, this is an unlikely cause in our patients, since the echocontrast appeared in the left atrium within three cardiac cycles after the right atrium had been opacified, which is atypical for pulmonary arteriovenous malformation. 24
needed to evaluate whether a patent foramen ovale affects the prognosis of such patients. are
N. T. receives
fellowship from the
Forschung. REFERENCES 1.
Arterial oxygen tension (Pa02) on breathing room air in relation to total pulmonary vascular resistance (PVR).
PaO, was influenced by presence of patent foramen ovale (40 present, A =absent) (ana!ys)sofcovariance,p