Novel Dosing and Monitoring of Aspirin in Infants With Systemic-to–Pulmonary Artery Shunt Physiology: the SOPRANO Study

Introduction

In the United States, congenital heart defects (CHDs) are the most common birth defects and account for a large number of infant deaths within the first year of life.^1,2 Shunts are most commonly placed for primary palliation or as a component of staged reconstruction in infants with single-ventricle (SV) physiology.³ Provision of pulmonary blood flow with a systemic-to–pulmonary artery shunt remains an essential component of initial treatment for some forms of complex cyanotic congenital heart disease.⁴ Infants with cyanotic congenital heart disease palliated with placement of a systemic-to–pulmonary artery shunt are at an increased risk for thromboembolic events with the incidence being reported as high as 25% to 40%.^5,6

Aspirin (ASA) is commonly used as thromboprophylaxis in patients requiring a systemic-to–pulmonary artery shunt and is associated with a greater than 7-fold reduction in the risk of shunt thrombosis and a decreased risk of death.⁷ However, there is a paucity of data regarding the optimal dose, monitoring, and effectiveness of ASA in congenital heart disease. Dosing is largely extrapolated from adult studies and small heterogeneous studies, and when to initiate enteral ASA in the postoperative period is also debated.⁸ Given these concerns, use of antiplatelet monitoring is crucial to ensure adequate ASA response. There are many reported ways to monitor ASA response, including ASA urine metabolite testing, light-transmission aggregometry, and platelet mapping. Two commonly used antiplatelet tests are thromboelastography platelet mapping (TPM) and VerifyNow aspirin reactivity units (ARUs; Accumetrics, San Diego, CA), but the evidence supporting these methods is quite varied with significant interpatient variability in response to the same weight-based ASA dosing.^3,9 A phenomenon called aspirin resistance has been seen in infants with CHDs, defined in previous studies using a combination of TPM and urine thromboxane concentrations.⁴ However, the incidence of ASA resistance or inadequate platelet inhibition reported has significantly varied between studies using TPM compared with ARU.^4,10

The objective of this study was to compare these 2 antiplatelet monitoring tests in SV patients palliated with a systemic-to–pulmonary artery shunt with regard to therapeutic effect. Additionally, this study assessed the safety and efficacy of our institutions’ ASA-dosing nomogram.

Materials and Methods

Results

Seventy patient charts were identified from the study period; 49 patients were included for final analysis with 25 in the TPM group and 24 in the ARU group. All patients excluded either never received ASA or did not have platelet inhibition measured by TPM or ARU. Baseline characteristics were similar amongst the 2 groups with the exception of shunt type (Table 1). In the TPM group, there were significantly more modified Blalock-Taussig-Thomas shunts (mBTTs) than Sano shunts ([88%] vs 8 [33%] and 3 [12%] vs 15 [62%], respectively; p = 0.004). The TPM group had significantly more patients with inadequate platelet inhibition (14 [56%] vs 2 [8%]; p = 0.0006) and escalation of care with additional thromboprophylaxis (15 [60%] vs 5 [21%]; p = 0.005]. Between TPM and ARU, there was no difference in a composite efficacy endpoint including shunt thrombosis (1 [2%] vs 0 [0%]; p = 0.32], cyanosis requiring re-intervention (9 [36%] vs 14 [58%]; p = 0.11], or sudden death [5/25 vs 1/24; p = 0.19) (Table 2; Figure). There was also no difference in minor bleeding (14 [56%] vs 14 [58%]; p = 0.87), major bleeding (1 [4%] vs 0 [0%]; p = 0.32], or composite bleeding (15 [52%] vs 14 [58%]; p = 0.66]. The most common bleeding event was bloody stools with 13 (52%) reported in the TPM group and 13 (54%) reported in the ARU group (Table 3.). A univariate prediction analysis of bloody stools was completed and did not identify any independent risk factors (Table 4).

View Figure

• Download as Powerpoint
• Download Figure

Table 1. Baseline Patient Demographics

View Fullscreen

Table 2. Primary Results

View Fullscreen

Table 3. Secondary Results

View Fullscreen

Table 4. Univariate Predictors of Bloody Stool

View Fullscreen

The TPM group started ASA significantly sooner postoperatively (day 7 vs day 11; p = 0.02) on significantly fewer feeds (24 mL/kg/day vs 95 mL/kg/day; p = 0.0004) with no difference in incidence of bloody stool (13 [52]) vs 13 [54]; p = 0.88]. The median dose of ASA was 12.3 mg/kg (IQR, 10.9–14) in the TPM group, compared with 13.5 mg/kg (IQR, 12.2–14.5) in the ARU group.

Discussion

Our study used 2 separate antiplatelet monitoring tests. Traditionally, SV patients receiving ASA were evaluated for ASA efficacy by TPM, which examined platelet aggregation in response to exogenous AA.³ TPM provides an individualized percentage of inhibition to ASA when compared with a standardized reference range.³ Newer methods, such as ARU, use embedded AA beads and light transmission aggregometry to qualitatively evaluate platelet inhibition compared with a manufacturer-recommended threshold.¹⁴ Given there is no manufacturer-recommended minimum percentage of inhibition with regard to adequate platelet inhibition evaluated by TPM, efficacy was defined as greater than 50% inhibition of AA as previously reported by Mir et al.³ We found patients monitored with TPM were often reported as having subtherapeutic platelet inhibition (14/25, 56%). This correlates with the incidence of previous findings of a synonymous phenomenon called aspirin resistance.³ Given the concern for shunt thrombosis and sudden death, many of these patients were started on concomitant thromboprophylaxis with enoxaparin or UFH (15/25, 60%). After a practice change, our institution moved to ARU platelet inhibition testing for ASA.^9,10 Manufacturer recommendations for VerifyNow platelet aggregometry report a specificity of 100% and sensitivity of 91.4%, respectively, with an ARU of <550, at detecting ASA presence in adults receiving chronic ASA therapy of 81 mg daily.¹⁴ This was further studied by Koh et al¹⁰ who used an ARU cutoff of >550 to define ASA resistance in infants with shunt-dependent physiology.¹⁰ Therefore, we defined efficacy as an ARU <550. We found that a much smaller rate of patients in the ARU group (2/24, 8%) was defined as achieving inadequate platelet inhibition compared with TPM, and subsequently fewer patients were on concomitant thromboprophylaxis (5/24, 21%). Given this finding in a similar patient population while using the same weight-based ASA-dosing nomogram and finding no difference in adverse outcomes, ARU may be a more appropriate therapeutic platelet inhibition test than TPM. Furthermore, ARU monitoring may reduce the unnecessary ASA dose increases and the addition of concomitant thromboprophylaxis. Further validation and comparison should be explored.

In addition to platelet inhibition, our study aimed to evaluate efficacy as defined by clinical outcomes. Adverse clinical outcomes in an infant palliated with a systemic-to–pulmonary artery shunt with regard to inadequate platelet inhibition can be acute or chronic.^5,6 Acutely, patients can experience shunt thrombosis and sudden cardiac death.⁵ Chronically, intraluminal narrowing of the shunt via platelet aggregation and inflammation can result in cyanosis requiring early re-intervention, shunt thrombosis, or even sudden cardiac death.^5,6 Given the low incidence of these clinical outcomes, we used a composite score to evaluate the efficacy of our novel ASA-dosing nomogram, which included shunt thrombosis, cyanosis requiring early re-intervention, and sudden cardiac death. There was no statistical difference between the groups as displayed in a Kaplan-Meier event curve in the Figure.

Nationally accepted dosing of ASA for shunt prophylaxis in this patient population is 3 to 10 mg/kg, based on the type of shunt.⁸ Based on available dosage forms, institutions using enteral formulations are forced to choose between halved or quartered 81-mg tablets because most of these patients have surgical palliation in the first month of life. We developed a novel dosing nomogram to stratify patients to fractional tablets by weight. While ensuring an adequate antiplatelet dose with ASA, providers must weigh concern for gastrointestinal (GI) bleeding risk secondary to the COX inhibition of prostaglandins, which protect the mucosal lining.^15–18 The mean weight across our study population was 3.1 kg (IQR, 2.8–3.4), and the mean ASA dose was 12.7 mg/kg (IQR, 11.6–14.4). While there was no difference in GI bleeding between the TPM and ARU group, overall, our study population had a high GI bleeding rate (n = 26 [53%]). Single-ventricle patients despite shunt type often have decreased GI perfusion.^19,20 It is plausible that larger doses of ASA (>10 mg/kg/dose) may cause significant mucosal prostaglandin inhibition, putting patients at a higher risk of mucosal injury.²¹ However, a univariate predictive analysis of bloody stools showed no independent predictor including absolute or weight-based ASA dose (Table 4). No independent risk factor for GI bleeding was identified. Given our findings of low rate of shunt thrombosis with high rate of GI bleeding, the weight threshold for increasing from 20.25-mg ASA to 40.5-mg ASA should be studied further.

A major unanswered clinical question that often arises in the pediatric cardiac intensive care unit is in relation to enteral medication administration and enteral feeding volume. Practice varies throughout the country in regard to volume of enteral feeds a patient must be receiving prior to administration of enteral medications. Given the previously discussed concern for mucosal injury with ASA, administration on minimal to no enteral feeds is typically avoided.²² Delay in ASA administration during the immediate postoperative period, when a patient with a systemic-to–pulmonary artery shunt is at the highest risk of thrombosis, can increase the risk for shunt thrombosis and in-stent stenosis.^5,6 Currently these patients undergo bridging therapy with low-dose UFH; however, it is not known if this offers the same antiplatelet and anti-inflammatory benefits as ASA. Our study found that the TPM group received ASA significantly earlier postoperatively while receiving significantly fewer enteral feeds than the ARU group. This finding is likely due to a change in provider practice during the study period, but cannot be fully quantified. Despite this finding, the 2 groups had no difference in GI bleeding and thus withholding ASA, based on a specific enteral feeding volume goal, may not be necessary.

There are notable limitations to this study. A major difference in the baseline demographics to include shunt type between our 2 patient groups, TPM and ARU, was noted. The TPM group had significantly more mBTTs than the ARU group, while the ARU group had significantly more Sano shunts than the TPM group. This difference is due to a practice shift in surgical leadership and technique preference during the study period. In this patient population, Sano shunts were associated with increased transplant-free survival at 12 months when compared with the mBTTs; however, the Sano shunts had more unintended interventions and complications.²³ Given our TPM group had significantly more patients with mBTTs, but also with concomitant thromboprophylaxis owing to inadequate TPM platelet inhibition, this may have affected the composite outcome of our study.

Our study evaluated groups separately by the type of antiplatelet monitoring they used to quantify antiplatelet response: TPM or ARU. The type of monitoring changed during the study period, therefore no patients received both TPM and ARU monitoring. Additionally, there are few studies that compare the sensitivity and specificity between TPM and ARU, with varying results.^24–26 The lack of adequate comparative studies makes validation of an individual platelet inhibition assay difficult. Given we did not have both types of monitoring for each patient, we were unable to correlate TPM and ARU for individual patients. However, the demographics were similar and the dosing nomogram was the same throughout the study period; therefore, we can infer that the response to ASA should be similar between the 2 groups.

Additionally, our study used a novel ASA-dosing nomogram that ultimately resulted in aggressive weight-based dosing. Our average dose of ASA of 12.7 mg/kg across the study population was higher than the recommendations found in most literature of 3 to 10 mg/kg.⁸ Our aggressive dosing may have skewed our efficacy and safety results; however, both the TPM and ARU groups received dosing according to the same nomogram. Lastly, sample size and power calculations were not performed. This is largely due to the predetermined intervention groups based on practice change; however, without a power calculation we cannot rule out the possibility of type I or type II error.

Conclusion

Use of ARU monitoring compared with TPM resulted in more consistent results and a reduction in the need for concomitant thromboprophylaxis. Using our ASA-dosing nomogram, the rates of shunt thrombosis and cyanosis requiring early re-intervention were low; however, rates of bloody stool were high in both groups. Given this finding, further studies should be done to determine the optimal ASA-dosing threshold. Timing of administration postoperatively did not correlate with bleeding despite significant differences in quantity of enteral feeds. Our study suggests ARU may be more reliable than TPM for determining ASA sensitivity in children with congenital heart disease requiring systemic-to–pulmonary artery shunt. Larger, prospective studies should be conducted to confirm these findings.