Predictive Value of Intraoperative Diagnosis of Residual Ventricular Septal Defects by Transesophageal Echocardiography

Transcription

1 ORIGINAL ARTICLES: SURGERY: The Annals of Thoracic Surgery CME Program is located online at To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal. Predictive Value of Intraoperative Diagnosis of Residual Ventricular Septal Defects by Transesophageal Echocardiography Baher M. Hanna,* Ahmed A. El-Hewala,* Peter J. Gruber, MD, PhD, J. William Gaynor, MD, Thomas L. Spray, MD, and Mohamed A. Seliem, MD, FACC Cairo University, Cairo, Zagazig University, Zagazig, Egypt; and The Cardiac Center, The Children s Hospital of Philadelphia, Philadelphia, Pennsylvania Background. Intraoperative transesophageal echocardiography (IOTEE) is well established as a monitoring tool during ventricular septal defect (VSD) repair to ensure complete closure of the defect. Residual shunts detected by IOTEE are common. The predictive value of IOTEE findings on the long-term course of residual shunts is not well documented, especially in regard to the need for reoperation or bacterial endocarditis prophylaxis. The objective of this study is to determine the predictive value of intraoperative IOTEE diagnosis of residual VSDs and therefore delineating the natural history of these findings. Methods. Retrospective review of IOTEE reports of 690 consecutive patients with VSD (isolated or part of a complex lesion) was undertaken. Those were compared with transthoracic echocardiographic reports of these patients before their discharge from the hospital, and the most recent transthoracic echocardiographic examination. Positive and negative predictive values, sensitivity, and specificity of such diagnoses were then calculated from predischarge and from follow-up transthoracic echocardiographic data. Results. There were 260 of 690 patients with a residual VSD on IOTEE; 24 required repeat cardiopulmonary bypass for complete closure. There were 573 patients with predischarge transthoracic echocardiographic examination; 296 had residual VSDs (125 not detected by IOTEE), and 13 defects required reoperation during the same hospitalization, 5 of which were detected by IOTEE. The positive and negative predictive values were 78% and 65%, respectively. Follow-up transthoracic echocardiographic examination of 383 local patients showed residual VSD in 57 (37 not detected by IOTEE), with positive and negative predictive values of 15% and 83%, respectively. Conclusions. Although IOTEE is sensitive enough to detect residual VSD shunts in many patients (37% of this cohort), the majority of these defects are trivial and resolve spontaneously, with a positive predictive value of only 15% on follow-up transthoracic echocardiographic examination and a rare need for reoperation. (Ann Thorac Surg 2010;89:1233 8) 2010 by The Society of Thoracic Surgeons Transesophageal echocardiographic imaging is currently a standard modality for intraoperative monitoring of infants and children during repair of congenital heart disease [1 5]. This is especially rewarding during repair of specific lesions including septal defects and atrioventricular and semilunar valves. Many residual defects or findings on the intraoperative transesophageal echocardiography (IOTEE) change significantly on the subsequent transthoracic echocardiographic imaging with changing hemodynamic status of Accepted for publication Oct 23, *Drs Hanna and El-Hewala are visiting physicians at the Cardiac Center of The Children s Hospital of Philadelphia. Address correspondence to Dr Seliem, Division of Cardiology, The Children s Hospital of Philadelphia, 34th St & Civic Center Blvd, Philadelphia, PA 19104; seliem@ .chop.edu. the patients from intraoperative to predischarge and then to a stable ambulatory status. The temporal history of these residual findings in septal defects is not well documented in a large cohort of patients with these defects whether isolated or as a part of more complex lesions, especially with regard to the need for reoperation or bacterial endocarditis prophylaxis. Delineation of such temporal history was the main purpose for that retrospective analysis of these patients. Material and Methods Patients The echocardiographic database of all patients who underwent ventricular septal defect (VSD) repair between January 2000 and December 2007 was reviewed. All patients who had an IOTEE after the repair were in by The Society of Thoracic Surgeons /10/$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 1234 HANNA ET AL Ann Thorac Surg INTRAOPERATIVE TEE AND RESIDUAL VSD 2010;89: Fig 1. A flow chart showing a longitudinal follow-up of 690 patients (Pts) with different types of ventricular septal defect (VSD) at three definite times; at the intraoperative transesophageal echocardiography (IOTEE), at a predischarge transthoracic echocardiographic examination (TTE), and at a remote transthoracic echocardiographic examination. (Echo echocardiography; TEE transesophageal echocardiography.) cluded. There were 690 patients. The mean age at repair was 1.8 years (median, 0.32 years) and mean weight was 8.7 kg (median, 5 kg). The anatomic type of the VSD was conoventricular or perimembranous (n 190), muscular (n 32), atrioventricular canal (n 137), malalignment type in conotruncal lesions (n 284), and supracristal or conoseptal hypoplasia (n 23), in addition to 24 patients with reoperation for a significant residual VSD operated on before year All patients had two ventricular repairs. This study was approved by our institutional review board. The patients echocardiographic reports were anonymously reviewed. Echocardiographic Examinations All IOTEE and subsequent transthoracic echocardiographic examinations were performed using Phillips Sonos 5500, 7500, and IE 33 ultrasound systems (Phillips, Andover, MA). The biplane and multiplane pediatric transesophageal transducers were used with the former mostly used in patients who weigh less than 5 kg. All echocardiographic reports were part of each patient cardiology file. Predischarge transthoracic echocardiographic examinations were available in 573 patients and performed at a mean of 6.7 postoperative days (median, 5; standard deviation, 4.8; range, 1 to 27 days). Remote were available in 383 patients and performed at a mean of 3.2 years (median, 2.8; standard deviation, 1.7; range, 0.8 to 7 years). A residual VSD was considered small if it is described as tiny, trivial, or small, or has a color flow jet of 1 to 2 mm in width as measured on the left ventricular septal side. A moderate defect would be described as such or measured 3 to 5 mm in diameter of color flow jet width, whereas a large defect was any defect measuring equal to or more than 5 mm of color jet width or by the intraoperative estimate of pulmonary to systemic blood flow ratio by oximetry. Statistical Analysis Positive and negative predictive values were calculated from the data available on the IOTEE, predischarge, and late. Patients who had a residual VSD detected on the IOTEE that was still present on their predischarge transthoracic echocardiographic examinations are considered true positives, those with a VSD on the IOTEE and no VSD on the transthoracic echocardiographic examinations are the false positives, those with no VSD on the IOTEE and no VSD on the transthoracic echocardiographic examinations are the true negatives, those with no VSD on the IOTEE and a VSD on the transthoracic echocardiographic examinations are the false negatives. The same was applied between the transthoracic echocardiographic examinations studies done on predischarge and remote outpatient studies, ie, patients who had a VSD detected on the predischarge transthoracic echocardiographic examinations that was still present on their remote follow-up transthoracic echocardiographic examinations are considered true positives, and so on. The following variable were calculated: positive predictive value true positive/(true positive false positive); negative predictive value true negative/(true negative false negative); sensitivity true positive/(true positive false negative); and specificity true negative/(true negative false positive). All patients who did not have a predischarge or a late, ie, a study at all three periods, were not considered for such calculations. Comparison between the different anatomic subgroups was performed using the 2 analysis, with a probability value of less than 0.05 being significant. Results Among the whole group (Fig 1), a residual VSD shunt was detected by IOTEE in 260 of 690 patients (37%; Fig

3 Ann Thorac Surg HANNA ET AL 2010;89: INTRAOPERATIVE TEE AND RESIDUAL VSD 1235 Fig 2. (A) A pie chart showing the total number of patients who underwent repair of ventricular septal defect (VSD), number of those without residual ventricular septal defect, those with residual shunt on the intraoperative transesophageal echocardiography (IOTEE), and those requiring repeat cardiopulmonary bypass. (B) A pie chart showing the total number of patients who had a predischarge transthoracic echocardiographic examination (TTE), those who had a negative or positive intraoperative transesophageal echocardiography, those with a ventricular septal defect that was not detected by intraoperative transesophageal echocardiography but was seen on the predischarge transthoracic echocardiographic examination, and those requiring reoperation during the same hospitalization. (C) A pie chart showing the late transthoracic echocardiographic examination with a much smaller number of patients with residual ventricular septal defect. 2A); of these, 24 patients (3.5%) required repeat cardiopulmonary bypass for closure of a residual hemodynamically significant VSD defect, as determined by IOTEE or the intraoperative estimate of pulmonary to systemic blood flow ratio by oximetry. Before discharge from the hospital, predischarge transthoracic echocardiographic examinations were available in 573 patients (Fig 2B); of these, 296 patients (51%) still showed a residual shunt at the VSD site and the IOTEE did not detect a residual shunt in 125 patients (18%). Thirteen patients required reoperation during the same hospitalization based on the transthoracic echocardiographic examination findings (13 of 573; 2.3%); IOTEE detected the VSD in 5 of these 13 patients. Using the data on the 573 patients who had both IOTEE and transthoracic echocardiographic examinations, the positive predictive value was 78% (sensitivity, 58%) and the negative predictive value was 65% (specificity, 57%). Remote as an outpatient procedure was available in 383 patients (Fig 2C). This detected a residual VSD in 57 patients (15%), of whom 20 patients were detected on the IOTEE, and 5 of them required reoperation for the same defect (5 of 383; 1.3%). With these follow-up data on the 383 patients, the positive predictive value decreased to 15% and the negative predictive value increased to 83, with sensitivity and specificity of 35% and 65%, respectively. For further evaluation of the effect of the size of the residual VSD, as determined by the IOTEE, on the long-term temporal course, the records of the 260 patients with residual VSD were examined in more detail. There were 210 patients with a small VSD, of which 134 were still seen on the predischarge transthoracic echocardiographic examinations (positive predictive value, 63%), and only 16 were still seen on long-term follow-up transthoracic echocardiographic examinations (positive predictive value, 12%). Forty-two patients had a moderate size VSD on their IOTEE, of which 33 were still seen on the predischarge transthoracic echocardiographic examinations (positive predictive value, 88%), and only 3 were still seen on long-term follow-up transthoracic echocardiographic examinations (positive predictive value, 10%). Eight patients had a large VSD on their IOTEE, of which 4 were still seen on the predischarge transthoracic echocardiographic examinations (positive predictive value, 50%), and only 1 was still seen on long-term follow-up transthoracic echocardiographic examination (positive predictive value, 25%). The incidence of a residual shunt at the VSD repair site was not different on IOTEE among these anatomic types except in the muscular group in which the incidence was higher (12 of 20; 60%). Comment Intraoperative transesophageal echocardiography has been established as a safe, valuable, and accurate monitoring tool during surgical repair of congenital cardiac defects since the early 1990s [1 5]. In addition to its value to confirm the anatomic and physiologic findings, it is very valuable in assessing the adequacy of the repair and to rule out hemodynamically significant residual lesions [1, 3, 5]. It is not unusual for a patient with VSD to go back on cardiopulmonary bypass for repairing a significant residual defect detected by intraoperative IOTEE; such a scenario was observed in 3.5% of this study cohort (Fig 2A). The main objective of this study is to longitudinally follow up on the fate of residual shunts detected at the site of a repaired VSD in a large cohort of patients, and to determine the predictive values of the IOTEE on long-term

4 1236 HANNA ET AL Ann Thorac Surg INTRAOPERATIVE TEE AND RESIDUAL VSD 2010;89: Fig 3. Transthoracic echocardiographic image with color comparison showing an example of a residual intramural ventricular septal defect (long arrow) after repair of double-outlet right ventricle. The ventricular septal defect patch (small arrow) seems intact, and the residual defect is through the trabeculations around the distal edge of the patch, with left to right shunt into the right ventricular outflow tract (star). The number of ventricular septal defects was 37. follow-up of these patients. These data will be important in counseling the patients families about these residual shunts, in addition to determining the need for bacterial endocarditic prophylaxis in light of the new American Heart/American College recommendations. The detection of a residual shunt after the cardiac repair is fairly common, occurring in 37% of all patients on the intraoperative IOTEE (Fig 2A). However, on longitudinal follow-up of these patients, predischarge transthoracic echocardiographic examinations were available in 573 patients and a residual defect was still detected in 296 patients or 52%, of which there were 125 patients whose residual defects were not detected by the intraoperative IOTEE (Fig 2B). These data are more or less similar to those reported by Yang and colleagues [6], with an incidence of a residual VSD on IOTEE in 33% of patients, two thirds of whose defects disappeared on predischarge transthoracic echocardiographic examinations. Of almost 200 patients who had no VSD on intraoperative IOTEE, 15 were found to have a small, hemodynamically insignificant VSD on remote outpatient follow-up transthoracic echocardiographic examination (approximately 8%). Several factors might have contributed to missing these defects in 125 patients, including hemodynamic factors (eg, tachycardia and hypovolemia after coming off cardiopulmonary bypass, masking a small defect by turbulence caused by many indwelling cannulas, sutures, or patch material), the duration given to the examiner to complete the echocardiographic examination, the experience of the examiner, and so forth. The discrepancy between the findings of the intraoperative IOTEE and the has been reported by other groups. Khatami and associatesl [7] reported residual VSD findings in 21% in the IOTEE, which increased to 36% as detected by transthoracic echocardiographic examinations before discharge from the hospital, and dropped to approximately 10% on long-term outpatient follow-up. All small residual VSDs did not require any intervention, and spontaneous closure was not different among different diagnoses. Other groups reported an incidence of 6% to 8% on long-term [8, 9]. On the other hand, our findings were not affected by the number of patients who required reoperation during the same hospitalization as we used the predischarge and remote transthoracic echocardiographic examinations for calculation of the predictive values. Among different anatomic subtypes, the incidence of a residual shunt was higher only with muscular VSD owing to the difficulty of finding these defects intraoperatively and eliminating the shunt totally. A special type of these residual VSDs detected either by IOTEE or transthoracic echocardiographic examination is the so-called intramural residual VSD (Fig 3), which usually occurs after repair of conotruncal defects, and may be missed during the IOTEE [10]. Using the data available on those 573 patients before discharge from the hospital, the predictive value was 78%. The other subgroup of patients who underwent long-term outpatient follow-up studies was 383 patients. Among those, the positive predictive value dropped significantly to 15%. These data are very close to the data available from the natural history of small VSD where up to 80% of these defects close spontaneously. The majority of the residual VSDs detected on the IOTEE were considered small (210 of 260; 81%), although moderate (42 patients; 16%) and large defects (8 patients; 3%) were uncommon. The predictive valve for the small and moderate residual defects (97% of all cases) was not different from that of the entire group (12% and 10%, respectively), whereas that for the large defects was higher (25%). Furthermore, among the patients who required reoperation for their VSD after discharge (n 5), 4 were considered small on the IOTEE and 1 was large, whereas on the latest transthoracic echocardiographic examination, 4 were found to be moderate and remained large. These numbers, however, were too small to draw meaningful statistical significance within the entire group. In conclusion, the detection of a residual shunt at the repaired VSD site carries a positive predictive value of

5 Ann Thorac Surg HANNA ET AL 2010;89: INTRAOPERATIVE TEE AND RESIDUAL VSD 1237 only 15%, regardless of the type of the VSD, in the long run, assuming that such a shunt was considered hemodynamically insignificant either intraoperatively or before discharge from the hospital. Such residual defects need surgical intervention only rarely, and probably do not require bacterial endocarditis prophylaxis as long as the patient observes good dental hygiene [11]. In terms of limitations, this study was a retrospective review with all inherent deficiencies of such studies. Echocardiographic examinations were not available in all 690 patients, with approximately 17% missing a study at discharge, and only 44% had a study at all three periods (IOTEE, predischarge, and remote transthoracic echocardiographic examinations). This was mostly related to the referral nature of these patients who did not have their remote follow-up at our institution. Although statistical calculations used in this cohort excluded all these patients, the numbers of the remaining patients at all three periods were large enough for the required statistical power. The examiner who performed the IOTEE was not necessary the same one who performed the transthoracic echocardiographic examination later on, which may raise a concern regarding interobserver variability. However, the majority of these patients was followed up at our institution and underwent several studies during that follow-up period, which confirmed these findings. The same argument, ie, redundancy of the follow-up echocardiographic examinations, also applies to the fact that the study involved review of the echocardiographic reports and not actual review of the images on each patient. Although we used the term false positive on those cases in which a VSD was detected on the IOTEE and not on the follow-up transthoracic echocardiographic examination, it should be realized that it is possible that these cases did have a trivial prepatch shunting that was seen on the IOTEE and disappeared before discharge and thus was not detected by the transthoracic echocardiographic examination. This classification, however, seems appropriate for the sake of discussing the significance of such findings during the operative procedure. References 1. Stevenson JG, Sorensen GK, Gartman DM, Hall DG, Rittenhouse EA. Transesophageal echocardiography during repair of congenital cardiac defects: identification of residual problems necessitating reoperation. J Am Soc Echocardiogr 1993; 6: O Leary PW, Hagler DJ, Seward JB, et al. Biplane intraoperative transesophageal echocardiography in congenital heart disease. Mayo Clin Proc 1995;70: Rosenfeld HM, Gentles TL, Wernovsky G, et al. Utility of intraoperative transesophageal echocardiography in the assessment of residual cardiac defects. Pediatr Cardiol 1998; 19: Randolph GR, Hagler DJ, Connolly HM, et al. Intraoperative transesophageal echocardiography during surgery for congenital heart defects. J Thorac Cardiovasc Surg 2002;124: Ma XJ, Huang GY, Liang XC, et al. Transesophageal echocardiography in monitoring, guiding, and evaluating surgical repair of congenital cardiac malformations in children. Cardiol Young 2007;17: Yang SG, Novello R, Nicolson S, et al. Evaluation of ventricular septal defect repair using intraoperative transesophageal echocardiography: frequency and significance of residual defects in infants and children. Echocardiography 2000; 17: Dodge-Khatami A, Knirsch W, Tamaske M, et al. Spontaneous closure of small residual ventricular septal defects after surgical repair. Ann Thorac Surg 2007;83: Meijboom F, Szatmari A, Utens E, et al. Long-term follow-up after surgical closure of ventricular septal defect in infancy and childhood. J Am Coll Cardiol 1994;24: Roos-Hesseling JW, Meijboom FJ, Spitaels SEC, et al. Outcome of patients after surgical closure of ventricular septal defect at young age: longitudinal follow up of years. Eur Heart J 2004;25: Preminger TJ, Sanders SP, van der Velde, ME, Castaneda AR, Lock JE. Intramural residual interventricular defects after repair of conotruncal malformations. Circulations 1994; 89: Wilson W, Taubert K, Gewitz M, et al. Guidelines from the American Heart Association: a guideline from the American Heart Association Rheumatic Fever, Endocarditis, and Kawasaki Disease Committee, Council on Cardiovascular Disease in the Young, and the Council on Clinical Cardiology, Council on Cardiovascular Surgery and Anesthesia, and the Quality of Care and Outcomes Research Interdisciplinary Working Group. Circulation 2007;116: INVITED COMMENTARY Intraoperative echocardiography delivers abundant realtime data, but with it, the burden of reacting or not to findings that are institution-dependent, operatordependent, probe-dependent, and subjective. Reporting a large series of intraoperative transesophageal echocardiography (IOTEE) studies and subsequent surveillance transthoracic echocardiography (TTE), Hanna and colleagues [1] reinforce our understanding of the incidence and fate of residual ventricular septal defects (VSDs) after operative repair. The decision to return to cardiopulmonary bypass to address an echocardiographic finding demands that such decision be based on the best evidence possible. The authors are to be congratulated for expanding the evidence bank that informs such decisions. Recognizing the variability in echo acuity between institutions or operators, it is reassuring to see the similarities generated by several independent studies, the current study included (Table 1) [1 4]. At some point in the perioperative period, all published studies show an incidence of small residual VSDs of between 33% and 38% (columns A and B), and most spontaneously close over time (column C). It is notable that although the IOTEE incidence of small residual VSDs is remarkably consistent, there is disparity in comparing the intraoperative findings and the predischarge TEE findings. Studies are at variance in detecting what may be early spontaneous closure or expansion of the number of defects detected between the operation 2010 by The Society of Thoracic Surgeons /10/$36.00 Published by Elsevier Inc doi: /j.athoracsur