In Reply:
Kampen et al. address the problem of differentiating cardiac and pulmonary right-to-left shunts by means of contrast-enhanced transcranial Doppler ultrasonography (c-TCD). It is important to distinguish transpulmonary and intrapulmonary passage of air. The microbubbles generated by means of the echo-contrast medium used are small enough to pass the capillaries, but they are also short-lived and thus dissolve before they can pass through the lungs. 1Bedell et al. 2present a case of paradoxical air embolism (PAE) developing after venous air embolism as a result of transpulmonary passage of air. The authors conclude that for transpulmonary air passage to occur, the amount of air and overload of the pulmonary filter mechanism are crucial. Such an overload is not to be expected from the small amount of air generated by injection of the contrast medium mentioned. 3
Some authors assume that late high-intensity transient signals due to contrast medium are indicative of pulmonary shunting. 4In contrast, Horner et al. 5report that the transit times in the presence of pulmonary shunts are comparable to those of cardiac shunts, concluding that TCD cannot distinguish between these two types of shunts. The time limit for the occurrence of microbubbles after the Valsalva maneuver is still a matter of debate: values range from 5 s (Job et al. 6) to 25 s (Jauss et al. 4). In determining the time limit, it is necessary to compromise between the sensitivity and specificity of c-TCD. The time interval of 15 cardiac cycles chosen in the study by Stendel et al. 7is within the range proposed in the recent literature 8and was associated with a specificity of c-TCD of 1 and a sensitivity of 0.92. This interval is also confirmed by the study of Karnik et al. 9This group investigated 36 patients by c-TCD, contrast-enhanced transesophageal echocardiography, and contrast-enhanced transthoracic echocardiography. In two patients with an echocardiographically proven patent foramen ovale (PFO), microbubbles in c-TCD occurred only after 10–20 cardiac cycles. The specificity of TCD was 1.
Another important aspect to consider is that the interval before microbubbles occur is also dependent on shunt size. Therefore, a larger shunt volume may be assumed not only when a large number of high-intensity transient signals due to contrast medium are observed, but also when they occur after a short interval. 8
False-positive results of c-TCD as reported in the letter by Kampen et al. were not observed with the protocol used in our study. 7Isolated late microbubbles seen in c-TCD may be due to particles initially adhering to the injection needle or the cardiac valve and entering circulation at a later time, as suggested by Droste et al. 10
The Valsalva maneuver increases the sensitivity of c-TCD 11and allows for differentiating pressure PFOs from permanent PFOs. 7However, a reliable differentiation between a PFO and a pulmonary shunt is not possible because high-intensity transient signals due to contrast medium have also been shown by c-TCD in patients with pulmonary right-to-left shunting with and without Valsalva maneuver, 12and the transit times associated with pulmonary shunts are comparable to those of cardiac shunts. 5
It is conceivable but remains to be shown that patients with a pressure PFO have a lower risk of PAE after venous air embolism than do patients with a permanent PFO. Therefore, preoperative risk estimation must identify these two patient groups. This can only be done by performing a Valsalva maneuver.
Assuming that a permanent PFO and a high shunt volume are associated with a higher risk for PAE, the authors recommend the following broad risk classification based on the results of c-TCD with and without Valsalva maneuver: low risk for the occurrence of PAE: pressure PFO, few microbubbles, longer latency; high risk for the occurrence of PAE: permanent PFO, many microbubbles, short latency. However, a quantitative estimate of the risk profile regarding the development of PAE cannot be made on the basis of the current data.
