Relationship between positive end-expiratory pressure levels, central venous pressure, systemic inflammation and acute renal failure in critically ill ventilated COVID-19 patients: a monocenter retrospective study in France

Study Design

We conducted a single center, retrospective cohort study in a 43 beds ICU in France during the COVID-19 surge. This study was approved by Institutional Review Board of Paris 7 University (No. 00006477, HUPNVS). Informed consent was waived by the board.

Patient Population

All patients admitted in ICU from March 2020 to July 2020 for a SARS-CoV-2 pneumoniae were screened. SARS-CoV-2 pneumoniae was defined by the association of evocative lung lesions seen in computed tomography scan or thoracic radiography and a positive polymerase chain reaction result for SARS-CoV-2. Patients were included if they were 18 years old or older and if they required mechanical ventilation for respiratory support during ICU stay. Patients were excluded if they were mechanically ventilated for another reason than respiratory failure, and if another diagnosis than SARS-CoV-2 was possible to explain the lung injury. Patients were also excluded if they had an AKI due to another disease than SARS-CoV-2 pneumoniae during the 3 days before admission to ICU.

Data Collection

Collected data were extracted from electronic and paper medical records. Day 1 was defined as the first day of mechanical ventilation. Demographic data included age, sex, comorbidities, treatment by renin angiotensin inhibitor, immunosuppressive therapy, use of corticosteroids, antiplatelet drug or anticoagulant therapy. Charlson comorbidity index, Simplified Acute Physiology Score (SAPS) II and Sequential Organ Failure Assessment (SOFA) scores [22,23] were calculated 24 hours after admission in ICU. Biological parameters included serum creatinine, urea, leukocytes count, neutrophils count, lymphocytes count, oxygen partial pressure in arterial blood.

Ventilatory and hemodynamic parameters included PEEP, plateau pressure, tidal volume, fraction of inspired oxygen and CVP and correspond to the mean values of the day (4–6 measurements every day for each parameter). According to local nursing monitoring protocols, plateau pressure reported in the medical records were the values of pressure displayed by the ventilator after a 5 seconds inspiratory pause. Criteria for ARDS were evaluated according to Berlin consensus [24]. Diuresis and daily fluid balance measured every day at the same hour were collected.

Patient Management

All patients were managed according to local protocol based on international guidelines. Respiratory management included limitation of tidal volume to 6 ml/kg, limitation of plateau and driving pressure, use of neuromuscular blockade and prone positioning session of at least 16 hours if necessary. PEEP levels were individualized according to a local protocol to improve oxygenation without deteriorating compliance and cardiac output. Patent foramen ovale and pleural effusions were systematically looked for and considered for drainage.

Hemodynamics optimizations were performed using cardiac output monitors for all patients, principally esophageal Doppler. Central line in the superior venous territory and arterial lines were inserted in all patients. CVP was measured continuously.

Definitions

The diagnosis of AKI was based on the Kidney Disease Improving Global Outcome (KDIGO) classification [25]. Severe AKI at admission was defined as an increase in serum creatinine level ≥2 times from baseline or ≥353.6 µmol/L or initiation of RRT corresponding to KDIGO stage 2 and 3. Baseline serum creatinine levels were measured in blood samples taken before hospital admission when available. When baseline creatinine was not available, the lowest serum creatinine level measured during the patient’s hospital stay was used if the glomerular filtration rate calculated with Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula was ≥75 mL/min/1.73m². In other cases, the baseline creatinine level was estimated by using the modification of diet in renal disease equation with the assumption of an age-adjusted glomerular filtration rate because this approach performed best in terms of generating the smallest number of false negatives [26].

New, persistent and transient AKI

New severe AKI was defined as an increase in serum creatinine level ≥2 times from baseline or ≥353.6 µmol/L or initiation of RRT after 24 hours from inclusion in patients who had no severe AKI upon the initiation of mechanical ventilation (MV). Persistent severe AKI was defined as a steady or increase in KDIGO classification stage between the first 24 hours following initiation of MV and day 5 in patients with KDIGO stage ≥2 at inclusion or between the first 24 hours following initiation of MV and the last day of MV depending on which one comes first. Transient severe AKI was defined as a resolution or a downstaging of AKI between the first 24 hours following inclusion and day 5 (example from KDIGO stage 2 to KDIGO stage 1).

Worsening renal function

Worsening renal function (WRF) was defined as a new or persistent severe AKI (WRF+). Patients with no AKI or transient severe AKI were considered as patient without WRF (WRF–).

Primary endpoint

The primary endpoint of the study was to investigate the relationship between PEEP, CVP and WRF during the 5 days following the initiation of mechanical ventilation.

Secondary endpoints

The secondary endpoint was to identify other risk factors associated with WRF.

Statistical Analysis

Quantitative parameters are reported as median with interquartile range (IQR) and qualitative parameters are expressed as number and percentage. Categorical variables were compared using Fisher exact test as appropriate. Continuous variables were compared using Mann-Whitney U-test.

Actualized mean PEEP was defined by the cumulative mean PEEP dose, i.e., the mean PEEP between the day studied and the previous day. We also calculated mean values during 5 days for the daily PEEP, CVP, leukocytes count, and neutrophils count to reflect the overall exposition of the kidney, to elevated pressures and inflammation during the first days of mechanical ventilation.

We developed two multivariable models with adjustment for cofounding factors (SAPS II and mean leukocytes count for 5 days) to assess the association between mean CVP level for 5 days, mean PEEP level for 5 days and the risk of WRF. Predicted probabilities of WRF by the models were plotted together with significantly associated variables and a local polynomial regression fitting was performed to visually assess the relationship between the variable and the risk of WRF.

Finally, the relationship between PEEP level and CVP was explored using linear regression. A P-value lower than 0.05 has been considered significant. Odds ratio are reported with 95% confidence intervals (CIs). All statistical analyses were performed using R statistical software version 4.0.2 (R Core Team, 2020, R Foundation for Statistical Computing; https://www.R-project.org).

Baseline Characteristics

Seventy-three patients with SARS-CoV-2 pneumoniae were screened. Among them, 57 were mechanically ventilated and included in the study. All patients fulfilled Berlin criteria for ARDS. Flowchart is shown in Supplementary Figure 1.

Among the 57 included patients severe AKI was diagnosed in 16 patients (28%). Severe AKI was diagnosed at the time of intubation in 11 patients (19.3%). Among them, four patients presented a transient AKI, and seven patients had a persistent severe AKI. Forty-six patients (80.7%) had no severe AKI at the time of intubation, five developed a new severe AKI. Therefore, 12 patients were included in the WRF+ and 45 patients were included in the WRF– group.

Baseline characteristics are shown in Table 1. Patients in the WRF+ group were mainly females and had higher SOFA and SAPS II scores. They were more susceptible to have hypertension and a pre-existing chronic kidney failure. Leukocytes at admission were higher in the WRF+ group than in the WRF– group (median [IQR], 12 [11–14] vs. 9 [7–11]; P=0.01). Regarding the ventilatory parameters at day 1, there was no difference of ratio between the partial pressure in O₂ and the inspiratory fraction of O₂ (P/F ratio), PEEP, compliance, Pplat and inspiratory fraction of oxygen between groups. The CVP was also not different between groups at the time of intubation.

Table 1.

Demographic characteristics

Concerning the need for vasopressors, there was no difference in terms of the number of patients requiring vasopressors between the WRF– and WRF+ group (35 [81.4%] vs. 12 [100.0%], P=0.249). There was also no difference in terms of the vasopressors free days during the 5 days post intubation between the WRF– and WRF+ groups (median [IQR], 2 [0–4] vs. 1 [0–2]; P=0.13).

Association between AKI and Ventilatory Parameters

During the 5 days following the initiation of mechanical ventilation, there was no difference in daily compliance, daily PEEP and daily actualized mean PEEP between groups except for PEEP and actualized mean PEEP at day 2 (Figure 1, Supplementary Figures 2 and 3). The PEEP mean value during the 5 days was not different between the WRF+ and WRF–groups (median [IQR], 10 [9–11] vs. 8 [8–11]; P=0.09) (Table 2). The absence of association between PEEP and WRF persisted after multivariable analysis (Supplementary Table 1, Supplementary Figure 4).

Figure 1.

Comparison of daily positive end expiratory pressure (PEEP) during the 5 days following intubation.

Figure 3.

Relationship between central venous pressure and positive end expiratory pressure (PEEP).

Table 2.

Mean values of PEEP, P/F ratio, CVP, leukocytes, and neutrophils during the 5 days following intubation

Association between Ventilatory Parameters, CVP and AKI

The daily values of CVP were not different between the two groups (Supplementary Figure 5). However, the mean value of CVP within 5 days after intubation was more elevated in the WRF+ group than in the WRF–group (median [IQR], 12 [11–13] vs. 10 [9–12]; P=0.03) (Table 2). Using multivariable analysis, there was a positive relationship between mean values of CVP during the 5 days and the risk of WRF, with an odd ratio for WRF of 1.97 (95% CI, 1.12–4.33) per increase of 1 mm Hg of CVP (Supplementary Table 2), and a significant increase in risk of WRF with CVP values above 10 mm Hg (Figure 2). Investigating the relationship between PEEP and venous pressure, there was a linear relationship between PEEP level and CVP (Figure 3).

Figure 2.

Risk of worsening renal function according to the mean value of central venous pressure during the 5 days.

Association between AKI and Biological Parameters

Daily leukocytes and neutrophils were more elevated in the WRF+ group (Figure 4, Supplementary Figure 6), except for neutrophils on the day 4. The mean value of leukocytes on the 5 days after intubation were 14 G/L (IQR, 11–18) in the WRF+ group and 9 G/L (IQR, 8–11) in the WRF– group (P=0.002), and mean value of neutrophils were also higher in the WRF+ group (11 G/L [IQR, 10–15]) than in the WRF– group (8 G/L [IQR, 6–9]) (P=0.001). We did not observe any difference in lymphocytes count between the two groups (Supplementary Figure 7).

Figure 4.

Comparison of daily leukocytes during the 5 days following intubation.