Design

This was a prospective observational study that included patients admitted to intensive care for less than 24 h and mechanically ventilated. Patients were separated in two distinct groups: patients with septic shock in the early phase and patients without sepsis.

Ethical approval

The study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the Central Ethics Committee (Ethics Committee Ile de France II, Paris, France, December 2018, Clinicaltrials.gov n° NCT03818269). All the participants and/or their relatives provided informed consent.

Sample size Estimation

The interstitial pressure range was expected to be between − 5 and − 1 mmHg in patients with septic shock and between − 2 and + 2 mmHg in controls. A study of healthy subjects with different hydration states found a standard deviation of 1 mmHg¹⁸. Considering a standard deviation of ± 2 mmHg, a hypothesized pressure difference of ≥ 2 mmHg between groups, an alpha risk of 5%, and a power of 90%, we estimated that 22 analyzable patients were required in each group.

Patient selection

All patients meeting the following criteria were eligible for both groups: age > 18 years, ICU admission for less than 24 h, ongoing invasive mechanical ventilation with orotracheal intubation, absence of clinically detectable edema (according to a validated score²⁵), and affiliation to healthcare insurance.

In the septic shock arm, additional criteria included diagnosis of septic shock according to the definition of the Sepsis-3 consensus conference²⁶ and total crystalloid resuscitation < 50 ml/kg.

In the control arm, additional criteria were absence of sepsis and shock from any cause (i.e. systolic arterial pressure > 100 mmHg, absence of vasopressors, diuresis > 0.5 ml/kg/h, normal serum lactate), and crystalloid resuscitation < 50 ml/kg over the previous 12 h.

The main exclusion criteria for both groups were clinical disseminated intravascular coagulation (purpura or bleeding), admission after resuscitated cardiac arrest, cardiogenic shock, acute pancreatitis, severe dehydration (clinical signs of dehydration and/or sodium level > 150 mmol/L), metformin poisoning, sepsis or septic shock for > 24 h, moribund or death appearing imminent (within 24 h), and legal protection.

Outcomes

Primary endpoint and interstitial pressure measurement

The primary endpoint was the measurement of subcutaneous interstitial pressure (SIP) as early as possible after ICU admission (D1). All measurements were performed by a single operator (AD).

Subcutaneous interstitial pressure was measured on the leg at the same location in all patients, 2–3 cm externally from the tibial crest and 5 cm below the anterior tibial tuberosity. An anaesthetic patch (Emlapatch) was applied before each measurement. SIP was measured using a sterile, single-use, dedicated transducer-tipped pressure probe (Micro-Cath ^®, Millar Inc., USA) with a 3.5 French diameter. The pressure probe was inserted subcutaneously through a 30 mm, 16G venous catheter. The catheter was then removed from the skin (along the probe length), and the probe was finally removed by 1 cm to allow space around the probe and fixed using an adhesive transparent dressing. The probe was connected to a monitor (Powerlab^®, AD Instruments, New Zealand), and the data were recorded using dedicated software (Labchart^®, AD Instruments, New Zealand). The probe was zeroed in saline at room temperature prior to its introduction. SIP was measured after a stabilization phase of 20–30 min, averaged over 5 min. After probe removal, pressure readings were used to correct the measurement drift. In a subgroup of patients, the catheter was left for 24 h to measure temporal changes in the SIP.

Prospective reporting of bleeding and hematoma at the insertion site was planned (as a potential confounding factor for SIP leading to falsely increased measurements).

Ischemic interstitial pressure

Experimentally, the best evaluation of interstitial pressure is provided when transcapillary and lymphatic flows are stopped (e.g., after cardiac arrest in experimental studies). Thus, in a subset of patients, we evaluated the effect of interrupting blood flow on SIP using a tourniquet for a short period. A large blood pressure cuff was placed above the knee and rapidly inflated to 180–190 mmHg (or > 20 mmHg above the systolic blood pressure) for 5 min. The inflation and deflation times were reported in the pressure recording.

Secondary endpoints

1) CVP and perfusion pressure were measured simultaneously with SIP. When central venous access was not available, the CVP was approximated using inferior vena cava ultrasonography.

2) Daily and cumulated fluid balance (day 1 to 3).

Assessed daily by the difference in 24-hour fluid intake and diuresis (or losses by renal replacement therapy).

3) Edema, and stage of severity (day 1 to 3).

A validated edema score was used²⁵ ranging from 0 (no edema) to 4 (severe edema). The maximum severity score observed during the study period was used.

4) SOFA at admission and day 3.

5) Lactate level at admission.

6) ICU mortality.

Statistical analysis

Quantitative variables were described in each group (septic shock/control) as mean ± standard deviation (SD) or median [Q1-Q3] according to distribution, and compared using Student’s t-test or non-parametric test as appropriate. Qualitative variables were described in each group as n (%) and compared using Chi2 test or Fisher’s exact test, as appropriate.

Subcutaneous interstitial pressure was described in each group (septic shock/control) with mean ± standard deviation (SD) or median [Q1-Q3] according to distribution and compared between groups using Student’s t-test or non-parametric tests, as appropriate, on the whole population. A sensitivity analysis was conducted after excluding patients with hematoma at the insertion site, as it could affect the measurement of SIP.

The correlation between subcutaneous interstitial pressure and quantitative outcomes was assessed using the Spearman correlation coefficient.