STUDY DESIGNS AND PROTOCOLS

Study I was a retrospective study in which we assessed the effect of use of inotrope treatment on 90-day mortality and examined factors associated with the initiation of inotrope treatment in patients with septic shock. Patient data were retrospectively collected from the clinical data management system (PICIS, Wakefield, MA, USA, and Dräger Medical, Lübeck, Germany) and from the Finnish Intensive Care Consortium database (Tieto Oy, Helsinki, Finland). Mortality data were obtained from the Finnish Intensive Care Consortium database and from the Population Register Center in Helsinki, Finland. Demographic and continuously measured hemodynamic data as well as data of vasopressor and inotrope use and other medications for hemodynamic support were collected and statistically analyzed. We assessed the association of demographic data, disease severity scores (APACHE II and SOFA for day I), hemodynamic variables, vasopressor treatment, blood lactate levels, and indicators of renal function (serum creatinine and urine output) with the initiation of inotrope treatment. Most importantly, the association between inotrope treatment during the first 24 hours in the ICU and outcome was assessed. The primary endpoint was 90-day mortality, but hospital mortality, ICU and hospital lengths of stay, and 1-year mortality were also assessed. A subgroup of patients monitored by using a pulmonary artery catheter was evaluated for PAC-derived hemodynamic parameters associated with use of inotropes and outcome. In this group of patients, the ability to respond to inotrope treatment by an increase in SV and CI was assessed. The changes in hemodynamic parameters induced by inotrope treatment and their association with outcome were also evaluated. To reduce the selection bias of more severely ill patients receiving inotropes with a higher propensity, adjustment with propensity scoring based on observed variables was performed in both the general study population and the subgroup monitored by PAC. Due to the retrospective design of the study, patients were treated according to current international and local guidelines at the discretion of the attending physician.

Study II was a retrospective cohort study in which the association of early hemodynamic variables (first 24 hours in the ICU) with outcome was assessed in critically ill patients with severe acute pancreatitis. Demographic data, disease severity scores, continuously measured hemodynamic variables, data of renal function and data of vasopressor and inotrope therapy were collected as described for Study I. The association of hemodynamic variables and hemodynamic support of the general study population and those of the subgroup with circulatory shock with 90-day mortality, which was the primary endpoint, was assessed. The associations of hemodynamic variables with early mortality (within 14 days of ICU admission) and ICU and hospital lengths of stay were also assessed.

In Study III, an interventional prospective study, the predictive values of changes in hemodynamic parameters during elevation of PEEP from 10 to 20 cm H2O were evaluated. The study algorithm is shown in Figure 7. Upon inclusion of the patient into the study, PEEP was set at 10 cm H2O. At PEEP 10 cm H2O, baseline catheter-derived

H2O, during which measurements and TEE were performed again during an end-expiratory pause. PEEP was then returned to baseline. A volume expansion with plasma expander was then conducted, after which the measurements and TEE were re-performed for assessment of response to fluid. All patients were mechanically ventilated, sedated, and paralyzed during the study. The changes in catheter-derived hemodynamic values and in left ventricular end-diastolic area (LVEDA) and aortic velocity time integral (VTIAo) assessed by TEE were statistically evaluated for their ability to predict fluid responsiveness. The changes in respiratory parameters, blood gas analysis, and DO2

induced by elevation of PEEP and volume expansion were also assessed.

  Figure  7.    Study  algorithm  for  Study  III  

Study IV was a predefined prospective observational substudy of the multicenter FINNAKI study. Its aim was to evaluate the association of early hemodynamics with development and progression of AKI during the first five days in the ICU. Special emphasis was put on MAP and vasoactive treatment during the first 24 hours in the ICU.

Routine data (demographics, diagnoses, disease severity scores, outcome measures) were collected from the Finnish Intensive Care Consortium database. A standardized case report form (CRF) for registration of data on chronic and present health information, risk factors for AKI, infections and antimicrobial treatment, organ dysfunction, fluid balance, and information on renal replacement therapy (RRT) was filled at admission and daily during days one to five in the ICU, and at ICU and hospital discharge. KDIGO criteria³²⁶ for AKI were assessed continuously based on urine output and measured creatinine values. The last serum creatinine (SCr) value from the preceding year, excluding the week before admission, was used as the baseline. If no baseline values were available, the

Modification in Diet in Renal Disease (MDRD) equation was used. The time-adjusted MAP was calculated based on each median value during the first 24 hours or until endpoint was reached and assessed for an association with development of AKI, as shown in Figure 8. The areas and aggregate times below MAP threshold values from 55 to 85 mmHg were calculated and were also assessed for an association with development of AKI. Moreover, the associations between vasoactive treatment and factors previously known to be associated with AKI, such as comorbidities, fluid balance, fluid therapy, and radio-contrast dye, were assessed.

Figure 8. MAP during the first 24 hours for a patient with severe sepsis. The reference line is set at 65 mmHg and the arrows indicate (significant) areas below this line