Page 23 - ANZCP Gazette May 2023

𝑟𝑟(%)
1 + 𝑝𝑝(∑#)
𝑝𝑝𝑒𝑒𝑒𝑒𝑝𝑝𝑒𝑒 h𝑝𝑝 𝑝𝑝𝑝𝑝𝑒𝑒
𝑠𝑠𝑒𝑒𝑝𝑝𝑝𝑝𝑒𝑒 𝑏𝑏𝑝𝑝𝑝𝑝𝑒𝑒 𝑝𝑝𝑒𝑒𝑝𝑝𝑜𝑜𝑜𝑜𝑝𝑝𝑝𝑝𝑝𝑝𝑒𝑒𝑝𝑝𝑟𝑟 𝑜𝑜𝑒𝑒𝑝𝑝 𝑝𝑝h𝑝𝑝 𝑝𝑝𝑒𝑒𝑝𝑝𝑝𝑝𝑝𝑝𝑒𝑒𝑝𝑝 𝑝𝑝𝑒𝑒𝑝𝑝𝑟𝑟 𝑜𝑜𝑒𝑒𝑝𝑝 𝑝𝑝h𝑝𝑝 𝑝𝑝𝑒𝑒𝑝𝑝𝑝𝑝𝑝𝑝𝑒𝑒𝑝𝑝
×
of oxygen delivery (> 297 ml/min/m2) were significantly
If two units of red blood cells were given intraoperatively predicted risk would increase to 6.6%. Predicted risk if both occurred would be 11.5%, highlighting the impact of these intraoperative predictors in comparison to the model of preoperative factors only in which predicted risk remains at 2.8%. This may be of value to specialists caring for these patients after surgery, providing an additional understanding of the intraoperative course upon recovery, and providing the opportunity for future studies to evaluate optimisation of postoperative management.
The association of red blood cell transfusion and CPB oxygen delivery with poor outcomes has been well recognised. Transfusion of as little as 1 or 2 units of red blood cells has been associated with increases in mortality and morbidity in patients undergoing CABG (9). Our study found a 1.82 times increase in likelihood of 30-day mortality with transfusion of 1 unit of red blood cells intraoperatively, 2.5 times increase with 2 units and 5.2 times with 3 or more units. However, it remains unclear the degree to which transfusion is the causal factor related to outcome given its association with surgical complexity and complications (10). This study confirms the rationale for blood conservation strategies consistent with the 2017 EACTS/EACTA Guidelines on patient blood management for adult cardiac surgery (10), and the importance of reducing hemodilution to reduce both the requirement for red blood cell transfusion and to maintain oxygen delivery. The minimum oxygen delivery during CPB has been identified as an independent predictor of acute kidney injury (AKI) (11, 12), with a goal directed approach to maintain CPB oxygen delivery >270 ml/min/ m2 shown to be beneficial in reducing the incidence of stage
Discussion
In this study, we have shown an association between intraoperative CPB management and patient outcome (30-day mortality). This is the first predictive model to be reported which incorporates CPB parameters using 30- day mortality as the primary outcome. By reproducing the approach taken by Billah et al (1) to identify predictors of 30-day mortality we demonstrated generalisability of their model whilst also identifying relevant CPB predictors. This demonstrates the importance of CPB parameters in the prediction of 30-day mortality, significantly improving the ROC from 0.7833 to 0.8306 in comparison to the base model reported by Billah et al (1), from which 4 out of the 12 preoperative variables were retained. Furthermore, the estimates of risk compared with observed values were highly consistent as risk increases. Modifiable CPB parameters included CPB time, red blood cell transfusion, minimum oxygen delivery, duration of arterial pressure <50 mmHg, and cardiac index <1.6 l/min/m2. Application of these parameters as CPB quality indicators will facilitate development of improvement initiatives. Previously we have reported the use of CPB quality indicators to improve process outcomes through feedback to clinicians at an institution level (7), and through multicentre registry reporting using benchmarking (8). Furthermore, the inclusion of intraoperative variables in a cardiac surgery 30-day mortality risk prediction model provides additional information to guide postoperative management, which can be calculated for an individual patient as:
𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑝𝑝𝑝𝑝𝑟𝑟𝑟𝑟 (%) = 100 × 1 + 𝑝𝑝 = 100 × (∑#)
(∑#)
1 + 𝑝𝑝(∑#)
𝑝𝑝(∑#) 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑝𝑝𝑝𝑝𝑟𝑟𝑟𝑟 (%) = 100
𝑊𝑊h𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝=𝑝𝑝𝑒𝑒𝑝𝑝𝑒𝑒𝑒𝑒𝑝𝑝𝑒𝑒𝑝𝑝𝑝𝑝𝑒𝑒𝑒𝑒; ∑𝛽𝛽=𝑟𝑟𝑠𝑠𝑠𝑠𝑒𝑒𝑜𝑜𝑝𝑝h𝑝𝑝𝑝𝑝𝑝𝑝𝑒𝑒𝑝𝑝𝑟𝑟𝑒𝑒𝑒𝑒𝑝𝑝 𝑝𝑝𝑝𝑝𝑒𝑒𝑒𝑒; ∑𝛽𝛽=𝑟𝑟𝑠𝑠𝑠𝑠𝑒𝑒𝑜𝑜𝑝𝑝h𝑝𝑝𝑝𝑝𝑝𝑝𝑒𝑒𝑝𝑝𝑟𝑟𝑒𝑒𝑒𝑒𝑝𝑝𝑠𝑠𝑒𝑒𝑝𝑝𝑝𝑝𝑒𝑒𝑏𝑏𝑝𝑝𝑝𝑝𝑒𝑒𝑝𝑝𝑒𝑒𝑝𝑝𝑜𝑜𝑜𝑜𝑝𝑝𝑝𝑝
𝑝𝑝𝑟𝑟𝑒𝑒𝑒𝑒𝑝𝑝 𝑠𝑠𝑒𝑒𝑝𝑝𝑝𝑝𝑒𝑒 𝑏𝑏𝑝𝑝𝑝𝑝𝑒𝑒 𝑝𝑝𝑒𝑒𝑝𝑝𝑜𝑜𝑜𝑜𝑝𝑝𝑝𝑝𝑝𝑝𝑒𝑒𝑝𝑝𝑟𝑟 𝑜𝑜𝑒𝑒𝑝𝑝 𝑝𝑝h𝑝𝑝 𝑝𝑝𝑒𝑒𝑝𝑝𝑝𝑝𝑝𝑝𝑒𝑒𝑝𝑝
If we consider the following hypothetical patient, a 75-year- old male with ejection fraction estimate of 40% and hypercholesterolaemia undergoing urgent CABG surgery, with a CPB time of 103 minutes, minimum oxygen delivery of 278 ml/min/m2, 8 minutes of mean arterial pressure <50 mmHg and 5 minutes with cardiac index < 1.6 l/ min/m2 would have a predicted risk of 2.8% using our CPB parameter model, and 2.7% using the model reported by Billah et al (1). However, if the CPB time was >128 min, predicted risk would increase to 5%.
𝑝𝑝I AKI (13). Our model found that the 2 highest quintiles
a(s∑s#o)ciated with lower incidence of 30-day mortality,
2 compared with the lowest quintile (<235 ml/min/m ).
2 Thedurationofcardiacindex<1.6l/min/m independent
of oxygen delivery was shown to be an important
modifiable factor in our model, which we believe to be a novel finding. The identification of duration of mean arterial pressure <50 mmHg as an independent factor has been reported by Haase et al, who found that mean arterial pressure <50 mmHg recorded using electronic intraoperative data was an independent predictor of AKI in patients undergoing CPB, but only in conjunction with severe anemia (lowest haemoglobin concentration >75th percentile) (14). Studies evaluating the influence of CPB pressure management have largely reported manually recorded and potentially incomplete and biased M A P values or lacked additional covariates (15-18). Thus, they may have had limited accuracy, precision and statistical power for their interpretation within a complex pathophysiological context, making the role of these intraoperative factors or their combination as modifiable predictors of post-operative AKI uncertain (14). Our study found a 10% increase in likelihood of 30-day mortality for every 10 minutes duration of cardiac index <1.6 l/min/m2,
MAY 2023 | www.anzcp.org 20
𝑟𝑟
𝑒𝑒 𝑝𝑝