Title : Context-sensitive indexation of hemodynamic parameters: A new concept of precision medicine in the ICU
Background: Hemodynamic parameters such as cardiac output (CO) and stroke volume (SV) are measured as “raw” values. To facilitate their interpretation and to provide consistent normal ranges, some hemodynamic parameters are adjusted to biometric data such as body weight, height and body surface area (BSA). However, several studies demonstrated inconsistent “normal ranges” for old and young patients as well as for female and male patients.
These inconsistencies can be attributed to several reasons: 1.) The current indexations are unspecific for the individual parameter. 2.) They are restricted to few biometric data such as weight and height. 3.) Some parameters such as CO and heart size (global end-diastolic volume GEDV) seem to be more closely associated with age and gender compared to weight and height. 4.) Several studies demonstrated a marked impact of “contexts” such as mechanical ventilation (MV), atrial fibrillation (AF) and technique of the measurement on the raw values. 5.) Adjustment of several parameters (CO, SV and GEDV) for the same indexation (BSA) might impede assessment of their association by “mathematic coupling”. 6.) Numerous hemodynamic parameters are not indexed at all, but interpretation of these parameters might be facilitated by adjustment for systematic confounders.
To overcome these concerns, we introduced the concept of “context sensitive indexation” CSI. CSI aims at individual indexation of each parameter and also adjusts for relevant contexts. CSI has been suggested for currently indexed parameters such as CO, SV, GEDV and extravascular lung water EVLW, but also for unindexed parameters such as stroke volume variation (SVV). The concept of CSI has been developed based on a large database including 10,936 transpulmonary thermodilutions (TPTDs) in 608 patients.
Objective, methods and results
1.) To introduce the concept of “context sensitive indexation” CSI
2.) To show the derivation and validation of CSI using independent databases
3.) To show attempts for future refinements of CSI
4.) To show the potential practical use of CSI:
a.) GEDV indexed to CSI (GEDV_CSI): GEDV is a marker of preload which might help to predict fluid responsiveness in critically ill patients. However, the data regarding prediction of fluid responsiveness are conflicting. We compared the prediction of fluid responsiveness by unindexed GEDV, GEDVI indexed to BSA and GEDVI_CSI in a group of 33 ICU-patients with shock undergoing a volume challenge (VC) with 7mL/kg saline 0.9% over 30 minutes. Among the different indexations of GEDV, only GEDVI_CSI significantly predicted fluid responsiveness defined as an increase in Cardiac Index ≥10% (AUC=0.745; p=0.028).
b.) Similarly, SVV is considered as a predictor of fluid responsiveness. Our database analysis demonstrated independent association of SVV with older age (p=0.002), weight (p=0.010), heart rate (p<0.001), atrial fibrillation (p<0.001) and spontaneous breathing (p<0.001). SVV_CSI was calculated based on a regression formula adjusting SVV to these variables. To validate SVV_CSI, we compared its predictive capabilities regarding fluid responsiveness in 106 volume challenges in an independent group of patients with controlled (n=28) or assisted ventilation (n=39) or spontaneous breathing (n=39). While unindexed SVV (ROC-AUC 0.573; p=0.069) and central venous pressure CVP (AUC 0.388; p=0.139) were not predictive SVV_CSI AUC 0.670; p=0.025) significantly predicted fluid responsiveness.
Conclusion: CSI improves interpretation of hemodynamic parameters.