Balloon Sizing of Atrial Septal Defects

To the Editor:

We read with interest the retrospective study on the sizing of atrial septal defects (ASD) by El-Said and colleagues ¹ in a recent issue of the Texas Heart Institute Journal. The authors concluded that transthoracic echocardiography (TTE) generally underestimated ASD size—especially if the ASD was >10 mm.

We would like to highlight certain differences between their findings and ours.^* We prospectively measured ASD size in 28 patients using 4 methods: 1) preprocedural TTE, 2) transesophageal echocardiography (TEE) one day before occlusion with an Amplatzer device (a self-centering double disk of nitinol wire with a central waist and Dacron mesh), 3) stretched diameter of the ASD by TTE (SD-TTE) during balloon sizing, and 4) stretched diameter of the ASD by fluoroscopy (SD-Fluoro). The mean TTE size (mean ± SD) was 15.36 ± 4.08 mm (range, 9–28 mm), and the TEE size was 17.64 ± 4.5 mm (range, 10–27 mm), P <0.001. The SD-Fluoro was defined as follows: the maximum diameter that can be pulled through the ASD in vivo and the minimum diameter that can be pulled through the measuring plate in vitro. Measurement by SD-Fluoro was performed before occlusion with the device.

The mean ratio of SD-Fluoro:TTE size was 1.51 ± 0.32, and that of the SD-Fluoro:TEE size was 1.55 ± 0.28 (P value not significant). The mean SD-Fluoro measurement was 22.57 ± 4.95 mm. The mean size of the device used was 23.04 ± 4.88 mm (12–32 mm). We found that TTE size slightly underestimated the TEE size of the ASD. However, there was a significant positive correlation between TTE size and SD-Fluoro (r=0.65, P <0.001) and between TEE size and SD-Fluoro (r=0.58, P <0.05). We found a similarity between TTE and TEE in 57% of patients, while El-Said and co-authors ¹ reported these methods to be similar in only 48% of their patients. Their group ¹ found that TTE underestimated SD-Fluoro by 22%, and TEE did so by 13.2%. The TTE underestimation was more in ASDs >10 mm in size. Both TTE and TEE were similar to SD-Fluoro measurement in ASDs ≤10 mm. However, in ASDs >10 mm, those authors recommended TEE for proper sizing, because the accuracy of TTE was lost in the larger ASDs. The ASD size in their patients was generally underestimated by TTE, ¹ but this was not so in our patients. We found uniformly good correlation between TTE and TEE results throughout the range of ASD sizes. El-Said ¹ found SD-Fluoro to be similar to SD-TEE; in contrast, we found a similarity between TEE (preprocedural) results and SD-Fluoro results.

These differences could be attributed to a number of factors. 1) We used SD-TTE (rather than SD-TEE) for evaluating shunt closure during SD-Fluoro measurement. 2) We used TEE (before the procedure) and not SD-TEE (during procedure) for comparison with SD-Fluoro: the 2 are temporally different. 3) We defined SD-Fluoro by the method described above and not by the diameter of the indentation when SD-TEE showed that there was no more shunting. Our method of measuring SD-Fluoro resulted in dynamic pulling of the balloon toward the right atrium in all patients. In 30% of their patients, ¹ balloon occlusion was static during SD-Fluoro measurement. Because of the complex anatomic structure of the atrial septum, the 2 methods could lead to different results. 4) Ours was a prospective study.

We required no SD-TEE to decide on the appropriate balloon size. We determined that SD-Fluoro was sufficient after cases were selected on the basis of TEE (preprocedural) or SD-TEE, because these latter 2 correlated well with SD-Fluoro. Hence, our procedure was simpler than that reported by El-Said and colleagues. ¹