In the present issue of Intensive Care Medicine, Ospina-Tascón et al. present an intriguing observational study in which they assessed the association between venous-to-arterial carbon dioxide difference (Pv-aCO2) and microvascular perfusion in patients with early septic shock [1]. A total of 75 adult patients with septic shock from a 60-bed mixed ICU in Columbia were included in the study. Potentially eligible patients with septic shock in the ICU were screened, and those eligible had a pulmonary artery catheter (PAC) inserted and were included in the study. Arterial and mixed venous blood samples were collected at the insertion of the PAC (T0) and at 6 h (T6), and Pv-aCO2 was defined as the difference between mixed-venous and arterial CO2 partial pressures. A sidestream dark-field imaging device was used to evaluate the microcirculation of the tongue at T0 and T6, and the association between Pv-aCO2 and the microcirculation of the tongue was assessed. Furthermore, the association between Pv-aCO2 and global haemodynamic variables was evaluated. The authors conclude that Pv-aCO2 was closely related to microcirculatory blood flow parameters during the early phase of septic shock, whereas Pv-aCO2 was poorly related to systemic haemodynamic variables [1].

Septic shock is a serious and frequent condition in the ICU, and early recognition and adequate treatment of tissue hypoperfusion is crucial [2, 3]. Oxygen-derived parameters such as central venous oxygen saturation (ScVO2) have failed to demonstrate clinical benefit as resuscitation targets in recent large randomised controlled trials and systematic reviews [4–6]. Mean baseline ScVO2 was above 70 % in all the trials, which may be explained by early recognition and aggressive treatment of shock. Nevertheless, while a low ScVO2 can prompt further resuscitation [3], in the context of normal or high ScVO2 its role is more questionable. Indeed, one of the limitations of using ScVO2 as a marker of hypoperfusion is that normal to high values cannot discriminate whether oxygen delivery is adequate or in excess of demand. High ScVO2 in the context of high lactate has for instance been shown to be associated with poor survival rates [7]. In this situation other tissue perfusion variables such as Pv-aCO2 have been proposed [8, 9]. In normal circumstances the difference between the arterial and the venous carbon dioxide is less than 6 mmHg. However, in states of low perfusion this difference can increase. In areas of the microcirculation that are poorly perfused there is an increased local production of CO2. Despite poor perfusion, since CO2 is about 20 times more soluble than O2, the likelihood of CO2 diffusing out of ischemic tissues and into the venous effluent is high, making it a very sensitive marker of hypoperfusion. Vallée et al. demonstrated that in septic patients with normalized ScVO2, high central venous-to-arterial CO2 difference (Pcv-aCO2) was associated with worse outcomes in terms of lower lactate clearance, lower cardiac index, and higher sepsis-related organ failure assessment (SOFA) score [8]. Ospina-Tascón et al.’s study adds to the evidence that P(c)v-aCO2 may be a useful tool to identify patients who remain inadequately resuscitated when an ScVO2 of 70 % has been reached [10]. Consequently, Pv-aCO2 may prove useful in the assessment of patients with shock (Fig. 1).

Fig. 1
figure 1

Algorithm for the assessment of patients with shock

Full size image

While there seems to be convincing evidence that P(c)v-aCO2 is a marker of (microcirculatory) hypoperfusion, a number of unanswered questions and limitations regarding routine clinical bedside use of Pv-aCO2 exist. First, use of Pv-aCO2 has not been tested in (randomised) clinical trials. Consequently, there is a risk of falsely inflated estimates [11], and the balance between benefits and harms are unknown. Second, there is a lack of Pv-aCO2 studies assessing patient-important outcome measures with adequate follow-up [12]. Trials using non-patient-important outcome measures (surrogate outcomes) overestimate the intervention effect by 40–50 % [13]. Third, existing studies have mainly been conducted in small single-centre institutions, which increases the risk of reporting falsely inflated estimates [14, 15]. Finally, the unblinded assessment of the association between Pv-aCO2 and the microcirculation increases the risk of selection bias [15].

In conclusion, Pv-aCO2 is an interesting and potentially important new tool for evaluation of the microcirculation in critically ill patients with shock; however, additional clinical evaluation, including adequate assessment of benefits and harms in high-quality randomised clinical trials, is needed prior to routine clinical use at the bedside.