Search

Truncus arteriosus with interrupted aortic arch (TA-IAA) is a rare congenital heart defect with historically poor outcomes. Contemporary multicenter data are limited.A retrospective cohort study of children who underwent repair of TA-IAA between 2009 and 2016 at 12 tertiary care referral centers within the United States was performed. Major adverse cardiac events (MACE) were defined as postoperative extracorporeal membrane oxygenation, cardiopulmonary resuscitation, or operative mortality. TA-IAA patients were compared with TA patients who underwent repair during the study period from the same institutions.We reviewed 35 patients with TA-IAA. MACE occurred in 12 patients (34%). Improvement over time was observed during the study period with 11 events (92%) occurring in the first half of the study period (P = .03). Factors associated with MACE included moderate or severe truncal valve insufficiency (P.01), concomitant truncal valve repair (P = .04), and longer cardiopulmonary bypass duration (P = .02). In comparison with 216 patients who underwent TA repair, patients with TA-IAA had a higher rate of MACE, but this finding was not statistically significant (34% vs 20%, respectively; P = .07). Additionally no differences between TA-IAA and TA groups were observed for unplanned reoperations (14% vs 22%, respectively; P = .3), hospital length of stay (24 vs 23 days, P = .65), or late deaths (7% vs 7%, P = 1.00).In this contemporary, multicenter cohort the rate of MACE after repair of TA-IAA was high but improved during the study period. Early childhood outcomes of patients with TA-IAA were similar to those with TA.

Background Multicenter studies on infants with anomalous left coronary artery from the pulmonary artery (ALCAPA) are lacking. We report the intermediate-term outcomes after ALCAPA repair in a multicenter cohort and identify risk factors for reintervention or death after discharge. Methods We retrospectively reviewed infants under 1 year of age who underwent ALCAPA repair from January 2009 to March 2018 at 21 US centers. The primary composite outcome was freedom from reintervention or death after discharge. We used the Kaplan-Meier survival analysis to examine freedom from reintervention or death and the Cox proportional hazard analysis to identify risk factors for this composite outcome. Results One hundred seventy-seven infants underwent ALCAPA repair; 170 (97%) survived to hospital discharge without transplantation. Twenty-three patients were lost to follow-up. The median duration of follow-up in the remaining 147 patients was 3.8 years (25%, 75%: 1.9 years, 6.0 years). Echocardiographic data were available at ∼3 years after discharge in 98 patients. Left ventricular function was normal in 96 patients (98%), whereas 26 patients (27%) had greater than mild mitral valve regurgitation. Sixteen patients (11%) underwent 20 reinterventions with 1 late death. Patients undergoing the Takeuchi procedure or atypical repairs (hazard ratio, 8.0; 95% confidence interval, 2.1-30.0) or with moderate or greater mitral regurgitation on discharge echocardiogram (hazard ratio, 3.4; 95% confidence interval, 1.2-9.1) were at increased risk for reintervention. Conclusions Intermediate-term outcomes after ALCAPA repair in infants are favorable. Persistent left ventricular dysfunction and reinterventions were uncommon, and mortality was rare. Patients who required atypical surgical repair or had moderate or greater mitral regurgitation at discharge warrant closer follow-up.

Truncal valve management in patients with truncus arteriosus is a clinical challenge, and indications for truncal valve intervention have not been defined. We sought to evaluate truncal valve dysfunction and primary valve intervention in patients with truncus arteriosus and determine risk factors for later truncal valve intervention. We conducted a retrospective cohort study of children who underwent truncus arteriosus repair at 15 centers between 2009 and 2016. Multivariable competing risk analysis was performed to determine risk factors for later truncal valve intervention. We reviewed 252 patients. Forty-two patients (17%) underwent truncal valve intervention during their initial surgery. Postoperative extracorporeal support, CPR, and operative mortality for patients who underwent truncal valve interventions were statistically similar to the rest of the cohort. Truncal valve interventions were performed in 5 of 64 patients with mild insufficiency; 5 of 16 patients with mild-to-moderate insufficiency; 17 of 35 patients with moderate insufficiency; 5 of 9 patients with moderate-to-severe insufficiency; and all 10 patients with severe insufficiency. Twenty patients (8%) underwent later truncal valve intervention, five of whom had no truncal valve intervention during initial surgical repair. Multivariable analysis revealed truncal valve intervention during initial repair (HR 11.5; 95% CI 2.5, 53.2) and moderate or greater truncal insufficiency prior to initial repair (HR 4.0; 95% CI 1.1, 14.5) to be independently associated with later truncal valve intervention. In conclusion, in a multicenter cohort of children with truncus arteriosus, 17% had truncal valve intervention during initial surgical repair. For patients in whom variable truncal valve insufficiency is present and primary intervention was not performed, late interventions were uncommon. Conservative surgical approach to truncal valve management may be justifiable.

Introduction: Optimal management of neonates with tetralogy of Fallot and pulmonary atresia (TOF/PA) with confluent pulmonary arteries is unknown. We sought to compare outcomes for patients who underwent primary complete repair vs. initial surgical palliation followed by delayed repair. Methods: We conducted a retrospective study at 20 centers within CoRe-PCICS (Collaborative Research from the Pediatric Cardiac Intensive Care Society). Data were collected on infants undergoing initial surgical intervention at 0 - 60 days of age with TOF/PA from 2009 to 2018, excluding patients with MAPCAs or those undergoing ductal stenting (n=22). The primary outcome was days alive and out of hospital in first year of life (DAOH). Secondary outcomes were 1 year mortality and a composite major complication outcome (similar to that in prior STS-CHSD studies), defined as occurrence of ≥ 1 of the following: renal failure requiring dialysis, stroke/seizure, permanent pacemaker, ECMO, or diaphragm paralysis during a palliation and/or repair hospitalization, or unplanned reoperation in the first year. Multivariable modeling with generalized estimating equations were utilized to compare outcomes between groups. Results: Of 210 subjects, 79 underwent primary complete repair and 131 underwent surgical palliation. Patients who underwent palliation had greater use of preoperative mechanical ventilation at first procedure (26% vs. 8%, p = 0.002). Other baseline characteristics were similar between groups (p > 0.05 for all). There was no statistically significant difference in DAOH between the palliation and primary repair groups [median (25%,75% IQR): 319 (280,336) vs. 338 (314,348 days), adjusted p = 0.20]. Nine (7%) patients who underwent palliation died in the first year of life vs. 4 (6 %) who underwent primary repair (adjusted OR: 1.1, 95% CI: 0.3-4.5; p = 0.9). At least one major complication occurred in 35% of patients who underwent palliation vs. 18% of patients who underwent primary repair (adjusted OR: 2.5, 95% CI: 1.4-4.4, p = 0.001). Conclusions: For infants with TOF/PA with confluent pulmonary arteries, a strategy of surgical palliation or primary complete repair resulted in similar DAOH and early mortality, whereas the morbidity incidence favored primary repair.

Background Literature describing morbidity and mortality after truncus arteriosus repair is predominated by single-center reports. We created and analyzed a multicenter dataset to identify risk factors for late mortality and right ventricle-to-pulmonary artery (RV-PA) conduit reintervention for this patient population. Methods We retrospectively collected data on children who underwent repair of truncus arteriosus without concomitant arch obstruction at 15 centers between 2009 and 2016. Cox regression survival analysis was conducted to determine risk factors for late mortality, defined as death occurring after hospital discharge and greater than 30 days after operation. Probability of any RV-PA conduit reintervention was analyzed over time using Fine-Gray modeling. Results We reviewed 216 patients with median follow-up of 2.9 years (range, 0.1 to 8.8). Operative mortality occurred in 15 patients (7%). Of the 201 survivors there were 14 (7%) late deaths. DiGeorge syndrome (hazard ratio [HR], 5.4; 95% confidence interval [CI], 1.6 to 17.8) and need for postoperative tracheostomy (HR, 5.9; 95% CI, 1.8 to 19.4) were identified as independent risk factors for late mortality. At least one RV-PA conduit catheterization or surgical reintervention was performed in 109 patients (median time to reintervention, 23 months; range, 0.3 to 93). Risk factors for reintervention included use of pulmonary or aortic homografts versus Contegra (Medtronic, Inc, Minneapolis, MN) bovine jugular vein conduits (HR, 1.9; 95% CI, 1.2 to 3.1) and smaller conduit size (HR per mm/m2, 1.05; 95% CI, 1.03 to 1.08). Conclusions In a multicenter dataset DiGeorge syndrome and need for tracheostomy postoperatively were found to be independent risk factors for late mortality after repair of truncus arteriosus, whereas risk of conduit reintervention was independently influenced by both initial conduit type and size.

Truncal valve management in patients with truncus arteriosus is a clinical challenge, and indications for truncal valve intervention have not been defined. We sought to evaluate truncal valve dysfunction and primary valve intervention in patients with truncus arteriosus and determine risk factors for later truncal valve intervention. We conducted a retrospective cohort study of children who underwent truncus arteriosus repair at 15 centers between 2009 and 2016. Multivariable competing risk analysis was performed to determine risk factors for later truncal valve intervention. We reviewed 252 patients. Forty-two patients (17%) underwent truncal valve intervention during their initial surgery. Postoperative extracorporeal support, CPR, and operative mortality for patients who underwent truncal valve interventions were statistically similar to the rest of the cohort. Truncal valve interventions were performed in 5 of 64 patients with mild insufficiency; 5 of 16 patients with mild-to-moderate insufficiency; 17 of 35 patients with moderate insufficiency; 5 of 9 patients with moderate-to-severe insufficiency; and all 10 patients with severe insufficiency. Twenty patients (8%) underwent later truncal valve intervention, five of whom had no truncal valve intervention during initial surgical repair. Multivariable analysis revealed truncal valve intervention during initial repair (HR 11.5; 95% CI 2.5, 53.2) and moderate or greater truncal insufficiency prior to initial repair (HR 4.0; 95% CI 1.1, 14.5) to be independently associated with later truncal valve intervention. In conclusion, in a multicenter cohort of children with truncus arteriosus, 17% had truncal valve intervention during initial surgical repair. For patients in whom variable truncal valve insufficiency is present and primary intervention was not performed, late interventions were uncommon. Conservative surgical approach to truncal valve management may be justifiable.

Background: Infants after cardiopulmonary bypass are exposed to increasing inflammatory mediator release and are at risk of developing fluid overload. The aim of this pilot study was to evaluate the impact of passive peritoneal drainage on achieving negative fluid balance and its ability to dispose of inflammatory cytokines. Methods: From September 2014 to November 2016, infants undergoing STAT category 3, 4, and 5 operations were randomized to receive or not receive intraoperative prophylactic peritoneal drain. We analyzed time to negative fluid balance and perioperative variables for each group. Pro- and anti-inflammatory cytokines were measured from serum and peritoneal fluid in the passive peritoneal drainage group and serum in the control group postoperatively. Results: Infants were randomized to prophylactic passive peritoneal drain group (n = 13) and control (n = 12). The groups were not significantly different in pre- and postoperative peak lactate levels, postoperative length of stay, and mortality. Peritoneal drain patients reached time to negative fluid balance at a median of 1.42 days (interquartile range [IQR]: 1.00-2.91), whereas the control at 3.08 (IQR: 1.67-3.88; P = .043). Peritoneal drain patients had lower diuretic index at 72 hours, median of 2.86 (IQR: 1.21-4.94) versus 6.27 (IQR: 4.75-11.11; P = .006). Consistently, tumor necrosis factor-α, interleukin (IL)-4, IL-6, IL-8, IL-10, and interferon-γ were present at higher levels in peritoneal fluid than serum at 24 and 72 hours. However, serum cytokine levels in peritoneal drain and control group, at 24 and 72 hours postoperatively, did not differ significantly. Conclusions: The prophylactic passive peritoneal drain patients reached negative fluid balance earlier and used less diuretic in early postoperative period. The serum cytokine levels did not differ significantly between groups at 24 and 72 hours postoperatively. However, there was no significant difference in mortality and postoperative length of stay.

Background Different techniques have been used for exposure of ventricular septal defect (VSD) margins when there is crowding of the VSD anatomy by tricuspid valve subvalvar apparatus. This study compared surgical outcomes for the two techniques of tricuspid valve leaflet detachment and the rarely described tricuspid valve chordal detachment for hard-to-expose VSDs. Methods Patients undergoing transatrial VSD repair were identified from our institutional database. Follow-up echocardiography and patient data were obtained from medical records. Between January 2005 and August 2016, 130 isolated conoventricular VSDs were repaired. Among these, 26 patients underwent leaflet detachment, 15 underwent chordal detachment, and 89 underwent regular VSD repair (reference group). Results The groups did not differ significantly in age, weight, postoperative length of stay, genetic/syndromic abnormalities, time to extubation, and left and right ventricular systolic function. The cardiopulmonary bypass and cross-clamp time were significantly longer in the leaflet detachment group than in the reference group (118 ± 28 vs 102 ± 32 minutes [p = 0.02] and 73 ± 20 vs 61 ± 23 minutes [p = 0.01], respectively). Echocardiographic follow-up was available for 87 patients at a mean of 2.6 years (range, 1 month to 11 years). Tricuspid regurgitation was rated as none or trivial in 66 (76%), mild in 20 (23%), and moderate in 1 reference group patient. There was no difference in presence of residual VSD or degree of tricuspid regurgitation among the three groups. There was no reoperation for tricuspid regurgitation. Conclusions Tricuspid valve leaflet and chordal detachment techniques provide an equally viable and safe alternative to closure of hard-to-expose VSDs while maintaining appropriate tricuspid valve function. Their use in our series did not lead to increased tricuspid valve dysfunction at early-to-midterm echocardiographic assessment.

The aim of this study was to evaluate outcome measures after the use of del Nido (dN) cardioplegia compared with conventional multidose high-potassium (non-dN) cardioplegia in neonates and infants.We retrospectively analyzed data in patients, aged younger than 1 year, undergoing cardiopulmonary bypass (CPB) from January 2012 to August 2015. We changed our cardioplegia protocol from non-dN to dN administered in a single or infrequently dosed strategy in September 2013. The outcomes of the dN group (n = 107) are compared with the non-dN group (n = 118). We analyzed variables for demographic, intraoperative, early postoperative, and discharge variables.The two groups were similar in age, weight, height, CPB, and cross-clamp time; preoperative and postoperative echocardiographic systolic functions; first 24-hour postoperative urine output and inotropic score; length of stay; and mortality rate. The Society of Thoracic Surgeons/European Association for Cardio-Thoracic Surgery Congenital Heart Surgery (STAT) mortality category was significantly higher in the dN group (p = 0.03). The cardioplegia dosing interval was lower for the non-dN group (p0.001). The volume and doses of cardioplegia per patient were significantly higher in the non-dN group (p0.001). In a subanalysis, when the Norwood patients were excluded from both groups, the overall STAT mortality category difference was no longer significant. The demographic, early postoperative, and discharge variables still showed no significant difference when the two groups were compared.Similar outcomes can be achieved with less frequent interruption of the operation and lower volume of cardioplegia when using dN cardioplegia solution compared with conventional cardioplegia. The dN cardioplegia with extended ischemic interval can be used as an alternative strategy in the neonatal and infant population during cardiac operations.

Supplemental Material, IRB_Protocol_Supplemental_File_CLEAN_6.29.2019 for Passive Peritoneal Drainage Impact on Fluid Balance and Inflammatory Mediators: A Randomized Pilot Study by Kamal K. Pourmoghadam, Stacey Kubovec, William M. DeCampli, Bertha Ben Khallouq, Kurt Piggott, Carlos Blanco, Harun Fakioglu, Alicia Kube and Sukumar Suguna Narasimhulu in World Journal for Pediatric and Congenital Heart Surgery

Background Elevated pulmonary vascular resistance (PVR) is common following repair of truncus arteriosus. Inhaled nitric oxide (iNO) is an effective yet costly therapy that is frequently implemented postoperatively to manage elevated PVR. Objectives We aimed to describe practice patterns of iNO use in a multicenter cohort of patients who underwent repair of truncus arteriosus, a lesion in which recovery is often complicated by elevated PVR. We also sought to identify patient and center factors that were more commonly associated with the use of iNO in the postoperative period. Design Retrospective cohort study. Setting 15 tertiary care pediatric referral centers. Patients All infants who underwent definitive repair of truncus arteriosus without aortic arch obstruction between 2009 and 2016. Interventions Descriptive statistics were used to demonstrate practice patterns of iNO use. Bivariate comparisons of characteristics of patients who did and did not receive iNO were performed, followed by multivariable mixed logistic regression analysis using backward elimination to identify independent predictors of iNO use. Main results We reviewed 216 patients who met inclusion criteria, of which 102 (46%) received iNO in the postoperative period: 69 (68%) had iNO started in the operating room and 33 (32%) had iNO initiated in the ICU. Median duration of iNO use was 4 days (range: 1-21 days). In multivariable mixed logistic regression analysis, use of deep hypothermic circulatory arrest (odds ratio: 3.2; 95% confidence interval: 1.2, 8.4) and center (analyzed as a random effect, p = .02) were independently associated with iNO use. Conclusions In this contemporary multicenter study, nearly half of patients who underwent repair of truncus arteriosus received iNO postoperatively. Use of iNO was more dependent on individual center practice rather than patient characteristics. The study suggests a need for collaborative quality initiatives to determine optimal criteria for utilization of this important but expensive therapy.

Background Transcatheter patent ductus arteriosus (PDA) occlusion is feasible in small infants and may improve lung function in symptomatic patients. We aimed to describe transcatheter PDA closure in small infants including predictors of technical success and rate of complication and to identify factors associated with improved respiratory status after closure. Methods All patients in the NICU at our center who were referred for transcatheter PDA occlusion between 1/2010 and 11/2014 were retrospectively identified. Relevant details were extracted. Additionally, a modification of the respiratory severity score (RSS) (FiO2 × mean airway pressure) was used to characterize degree of pulmonary support before and at intervals after catheterization. Results Twenty patients were identified with median age of 96 days (13–247) and weight of 3.1 kg (1.7–4.7). The PDA was type F morphology in 14 (70%) patients. The PDA was successfully occluded in 16 (80%) patients. Ratio of minimum PDA diameter/length was >0.5 in all unsuccessful attempts and

Objective We sought to describe characteristics and operative outcomes of children who underwent repair of truncus arteriosus and identify risk factors for the occurrence of major adverse cardiac events (MACE) in the immediate postoperative period in a contemporary multicenter cohort. Methods We conducted a retrospective review of children who underwent repair of truncus arteriosus between 2009 and 2016 at 15 centers within the United States. Patients with associated interrupted or obstructed aortic arch were excluded. MACE was defined as the need for postoperative extracorporeal membrane oxygenation, cardiopulmonary resuscitation, or operative mortality. Risk factors for MACE were identified using multivariable logistic regression analysis and reported as odds ratios (ORs) with 95% confidence intervals (CIs). Results We reviewed 216 patients. MACE occurred in 44 patients (20%) and did not vary significantly over time. Twenty-two patients (10%) received postoperative extracorporeal membrane oxygenation, 26 (12%) received cardiopulmonary resuscitation, and 15 (7%) suffered operative mortality. With multivariable logistic regression analysis (which included adjustment for center effect), factors independently associated with MACE were failure to diagnose truncus arteriosus before discharge from the nursery (OR, 3.1; 95% CI, 1.3-7.4), cardiopulmonary bypass duration >150 minutes (OR, 3.5; 95% CI, 1.5-8.5), and right ventricle-to-pulmonary artery conduit diameter >50 mm/m2 (OR, 4.7; 95% CI, 2.0-11.1). Conclusions In a contemporary multicenter analysis, 20% of children who underwent repair of truncus arteriosus experienced MACE. Early diagnosis, shorter duration of cardiopulmonary bypass, and use of smaller diameter right ventricle-to-pulmonary artery conduits represent potentially modifiable factors that could decrease morbidity and mortality in this fragile patient population.

Objective Nutrition is vital for maintaining optimal cellular and organ function, particularly in neonates who undergo cardiac surgery. Achieving nutritional goals preoperatively can be challenging because of fluid restrictions, suboptimal oral intake, and concerns for inadequate gastrointestinal circulation. We examined preoperative caloric intake and its effects on postoperative course in neonates who underwent cardiac surgery. Methods We retrospectively reviewed records of neonates (younger than 30 days) who underwent congenital heart surgery requiring cardiopulmonary bypass from 2008 to 2014 at Arnold Palmer Hospital for Children. Data on multiple nutritional and postoperative variables were collected. Study outcomes included hospital length of stay, duration of mechanical ventilation, and acute kidney injury (AKI). Results Records of 95 neonates were reviewed. Sixty-six patients (69.5%) with a median age of 5 days did not achieve preoperative caloric goal, whereas 29 patients (30.5%) with a median age of 11 days did. Of those who achieved caloric goal, 6 (20.6%) achieved it via total parental nutrition, 9 (31.1%) with a combination of total parental nutrition and enteral feeds, and 14 (48.3%) via enteral route. There was a significant difference in peak lactate ( P = .002), inotropic score ( P = .02), and duration of mechanical ventilation ( P = .013) between those who did and did not achieve caloric goal. In multivariable analysis we found that failure to achieve caloric goal preoperatively was independently associated with stage 2 or 3 AKI ( P = .04; odds ratio, 4.48; 95% confidence interval, 1.02-19.63) and younger age at the time of surgery ( P Conclusions Failure to achieve preoperative caloric goal might contribute to development of AKI and might be associated with greater severity of illness postoperatively.

Objective. Our hypothesis is that specific proinflammatory and anti-inflammatory urinary cytokines are useful in the diagnostic evaluation of risk for sepsis in term neonates. We conducted a pilot, prospective hospital-based longitudinal observational study to test the urine of term neonates with a 13 biomarker panel of cytokines. Methods. Infants were divided into 2 groups: The control group (n = 15) consisted of infants admitted to newborn nursery, and the test group (n = 15) consisted of infants admitted to the neonatal intensive care unit for presumed sepsis. Bagged urine samples were collected from 30 term neonates for testing our hypothesis. Results. Urinary interleukin (IL)-8 ( P = .004*), inducible protein (IP)-10 ( P = .007*), and monocyte chemoattractant protein (MCP)-1 ( P = .02) were significantly increased in the test group compared with the control group. Conclusions. Urinary IL-8, IP-10, and MCP-1 are proinflammatory cytokines that are increased in the neonate during an infectious inflammatory process. These may be useful predictors as an adjunct to the current protocols to recognize neonatal sepsis.

Transcatheter patent ductus arteriosus (PDA) occlusion is feasible in small infants and may improve lung function in symptomatic patients. We aimed to describe transcatheter PDA closure in small infants including predictors of technical success and rate of complication and to identify factors associated with improved respiratory status after closure.All patients in the NICU at our center who were referred for transcatheter PDA occlusion between 1/2010 and 11/2014 were retrospectively identified. Relevant details were extracted. Additionally, a modification of the respiratory severity score (RSS) (FiOTwenty patients were identified with median age of 96 days (13-247) and weight of 3.1 kg (1.7-4.7). The PDA was type F morphology in 14 (70%) patients. The PDA was successfully occluded in 16 (80%) patients. Ratio of minimum PDA diameter/length was0.5 in all unsuccessful attempts and0.4 in all successful cases (P = .01). Of the 16 cases of occlusion, Amplatzer Vascular Plug II was used in 15 (94%). No deaths or pulse loss occurred. Five (25%) patients required blood transfusion and transfusion was associated with lower hemoglobin (P = .049), lower weight (P = .008), and lower aortic pressure (P = .04). Excluding 1 patient with significant congenital heart disease, the RSS improved at 3 days in 9 (60%) patients and at 7 days in 12 (80%) compared with preintervention value. Patient factors were not associated with improved RSS at 3 or 7 days.In our cohort of symptomatic infants, transcatheter PDA occlusion was successful in most and a ratio of minimum PDA diameter/length of0.4 was predictive of technical success. Using a surrogate for pulmonary support, the majority of patients were on less support 7 days after closure.

Oral erythromycin prolongs cardiac repolarization and is associated with case reports of torsades de pointes. Because erythromycin is extensively metabolized by cytochrome P-450 3A (CYP3A) isozymes, commonly used medications that inhibit the effects of CYP3A may increase plasma erythromycin concentrations, thereby increasing the risk of ventricular arrhythmias and sudden death. We studied the association between the use of erythromycin and the risk of sudden death from cardiac causes and whether this risk was increased with the concurrent use of strong inhibitors of CYP3A.We studied a previously identified Tennessee Medicaid cohort that included 1,249,943 person-years of follow-up and 1476 cases of confirmed sudden death from cardiac causes. The CYP3A inhibitors used in the study were nitroimidazole antifungal agents, diltiazem, verapamil, and troleandomycin; each doubles, at least, the area under the time-concentration curve for a CYP3A substrate. Amoxicillin, an antimicrobial agent with similar indications but which does not prolong cardiac repolarization, and former use of erythromycin also were studied, to assess possible confounding by indication.The multivariate adjusted rate of sudden death from cardiac causes among patients currently using erythromycin was twice as high (incidence-rate ratio, 2.01; 95 percent confidence interval, 1.08 to 3.75; P=0.03) as that among those who had not used any of the study antibiotic medications. There was no significant increase in the risk of sudden death among former users of erythromycin (incidence-rate ratio, 0.89; 95 percent confidence interval, 0.72 to 1.09; P=0.26) or among those who were currently using amoxicillin (incidence-rate ratio, 1.18; 95 percent confidence interval, 0.59 to 2.36; P=0.65). The adjusted rate of sudden death from cardiac causes was five times as high (incidence-rate ratio, 5.35; 95 percent confidence interval, 1.72 to 16.64; P=0.004) among those who concurrently used CYP3A inhibitors and erythromycin as that among those who had used neither CYP3A inhibitors nor any of the study antibiotic medications. In contrast, there was no increase in the risk of sudden death among those who concurrently used amoxicillin and CYP3A inhibitors or those currently using any of the study antibiotic medications who had formerly used CYP3A inhibitors.The concurrent use of erythromycin and strong inhibitors of CYP3A should be avoided.