Introducing intravascular microdialysis for continuous lactate monitoring in patients undergoing cardiac surgery: a prospective observational study

Blood lactate level is a clinically important biomarker in critically ill patients, and its measurement may aid in monitoring of hemodynamics[1]. Elevated blood lactate is strongly associated with adverse outcomes in critically ill patients and thus has prognostic value[2‐4]. Hyperlactatemia has traditionally been considered to be a sign of tissue hypoperfusion due to anaerobic metabolism[5], such as during cardiogenic shock. Investigators in recent studies have suggested that elevated blood lactate during critical disease is the result of not only tissue hypoperfusion but also adrenergic stress[6]. This phenomenon was demonstrated for skeletal muscle, where epinephrine activates the Na⁺/K⁺-ATPase pump independently of tissue hypoxia[7]. An elevated blood lactate concentration may result from both increased production and decreased clearance. Lactate clearance occurs mainly in the liver and is dependent on tissue perfusion. In addition to elevations in lactate, the rate of normalization of elevated lactate levels is also independently associated with increased mortality[8, 9].

Lactate monitoring may thus be used for risk stratification in critically ill patients[10, 11] and for early lactate-guided treatment. Researchers in a prospective randomized trial showed that lactate-guided treatment in patients with hyperlactatemia resulted in improved outcomes[12]. Cardiac surgery patients often develop hyperlactatemia postoperatively[13], which may be affected by several factors, such as the type of surgery, patient body temperature and extracorporeal circulation. Elevated lactate levels in this patient category are also associated with postoperative mortality and a prolonged ICU stay[14, 15]. A similar relation was observed for pediatric cardiac surgery patients[16].

We have previously evaluated intravascular microdialysis to monitor glucose, both offline and continuously[17‐19]. In our first study, we compared intravascular microdialysis lactate values offline with arterial blood lactate concentration[17], but we did not study the online continuous monitoring of lactate. The aim of our present observational study was to evaluate intravascular microdialysis as a method for continuous lactate monitoring in cardiac surgery patients by comparing the accuracy of microdialysis lactate values with reference values analyzed by an arterial blood gas test. In addition, we recorded major postoperative complications in relation to lactate levels.

Data were available for 78 patients. Data were missing for two patients monitored with the SLC due to catheter damage that occurred while performing mitral valve surgery in one patient and due to mechanical sensor failure in the other. Data were available for analysis in all patients monitored with the TLC. However, monitoring was ended prematurely in one patient because of catheter dislocation (accidental withdrawal), which was subsequently detected as the microdialysis values were accompanied by a warning signal due to rapid changes in blood lactate levels. In three other patients, the monitoring was paused because of sensor complications, which were also detected by the accompanying warning signal for uncertain values, but could be restarted after exchange of the sensor. Two patients monitored with the TLC had shorter follow-up times than expected because, by accident, monitoring was not conducted during nighttime. The patient characteristics from the included 78 patients are shown in Table 1. The mean follow-up time for all patients was 25.6 hours (range = 1 to 97, median = 22), and a total of 1,601 paired microdialysis–arterial blood gas lactate values were obtained and included for analysis. An average of 20.5 samples from each patient (range = 1 to 43, median = 19) were analyzed. The values for each patient are provided in Table 2. No consistent systematic drift between calibrations was observed (see Additional file1). Furthermore, we did not notice any increasing calibration drift with longer monitoring time over subsequent calibration periods (see Additional file2). After the removal of the microdialysis catheter, no blood clotting of the microdialysis membrane was observed and no complications caused by the microdialysis catheter were detected.

The mean (±SD) microdialysis lactate value was 1.40 ± 0.72 mmol/L, and the mean arterial blood gas lactate value was 1.38 ± 0.68 mmol/L. The regression coefficient (calculated using a random coefficient model) was 0.98 (P = 0.0001). The Pearson correlation coefficients (r) for each patient are presented in Table 2 (median = 0.92, range = 0.19 to 1). The mean relative difference was 1.5%, and the mean absolute relative difference was 10.3%. The Bland-Altman analysis showed a bias ± limits of agreement of 0.02 ± 0.42 mmol/L (Figure 2). An example of a patient’s microdialysis lactate values compared with arterial blood gas lactate values is shown in Figure 3.

In 18 patients, a microdialysis lactate value >3.0 mmol/L was obtained, and the following adverse events occurred in seven of these patients: reoperation for bleeding (n = 5), perioperative myocardial infarction (n = 1) and circulatory instability with transient visceral ischemia (n = 1). If the cutoff point had been set to 2.5 mmol/L, another patient who underwent reoperation for bleeding would have been noted, as would another nine patients for whom no adverse events were recorded. A cutoff point <2.5 mmol/L would not have added more patients with recorded adverse events.

In this study, we demonstrate that intravascular microdialysis may be used for the continuous monitoring of lactate in patients undergoing cardiac surgery. The microdialysis lactate values were well correlated and had low bias compared with reference arterial blood gas lactate values.

The system recognized its malfunction in six patients (as described in Results, first paragraph) and could be corrected for, and the system restarted in three patients. Overall, the correlation coefficient for each patient showed good association between microdialysis lactate and arterial blood gas lactate concentrations. However, a few patients with low correlation coefficients need to be discussed. The reason for correlation coefficients <0.5 was one or two outliers. For example, patient 6, who was monitored with the TLC, the correlation coefficient would be 0.83 (P = 0.0001) instead of 0.19 if one outlier were removed. Similarly, removal of one or two outliers in the remaining patients with correlation coefficients <0.5 would yield new correlation coefficients >0.75 (all with P < 0.01).

The monitoring of lactate has been made easier with point-of-care testing. Studies have been conducted to analyze the differences in arterial, venous and capillary lactate monitoring methods with findings of high accuracy[20]. There are several reasons why we used arterial samples both as a reference measurement method and for calibration of the central venous microdialysis measurements. First, arterial samples of lactate (and glucose) may be considered the standard method, and they are regularly taken in critically ill patients as opposed to central venous samples. Differences between central venous and arterial lactate concentrations are usually minimal[21, 22]. Weil et al. studied the correlation between central venous and arterial lactate values and observed a mean difference of 0.03 mmol/L (r = 0.996)[23]. In our previous study, we also found good agreement between arterial and venous blood gas lactate values (mean difference = 0.09 mmol/L, limit of agreement ±0.6 mmol/L and Pearson correlation coefficient = 0.92), further supporting arterial blood gas as a reference method[17].

Lactate monitoring in critically ill patients is preferably performed by assessing the lactate trend by repeated lactate sampling, which may be used for predicting in-hospital mortality[11]. Jansen and colleagues used a time interval of 2 hours in their early lactate-guided therapy study[12]. Continuous monitoring may be advantageous for the critically ill patient, as the clinician is provided information every minute that could potentially be used to support and evaluate the given treatment.

Interestingly, we observed a higher complication rate in patients with a postoperative lactate level >3 mmol/L. It should be stressed that the goal of our study was not to evaluate lactate-guided treatment. The sample size was not adequate to evaluate the association between lactate concentration and outcome. However, this observation has some value because continuous lactate monitoring may add further knowledge in this field, which is why we mention this observation in our results. The hypothetical cutoff point of >3.0 mmol/L was based on a recent study in which the researchers reported the association between high lactate values and postoperative complications in adult cardiac surgery patients[15]. Investigators in other studies in which a higher cutoff point (3.0 to 3.5 mmol/L) was used have reported a better predictive value relating hyperlactatemia with mortality[2], even though significantly increased mortality has previously been observed with a cutoff point of 1.5 mmol/L[24].

We hypothesize that the continuous monitoring of lactate may aid in lactate-guided treatment in patients who present with elevated lactate levels, but this remains to be further studied. Studies on lactate-guided treatment have been performed early in critical illness (set to the first 8 hours after ICU admission in the study by Jansen et al.[12]), and thus it seems to be important to start lactate monitoring as early as possible. Hence, eligible patients should receive the TLC with integrated microdialysis function as soon as possible after arrival to the ICU. This protocol should not constitute a clinical problem, because these patients usually require rapid central venous access and thus could just as easily receive the microdialysis TLC in place of a regular CVC. If a clinical benefit could be demonstrated with the continuous monitoring of lactate, the cost-effectiveness of utilizing TLCs instead of CVCs would need to be calculated as well.

One should be mindful that this is a pilot study that needs to be further substantiated. In our present study, intravascular microdialysis was evaluated for continuous lactate monitoring in patients undergoing cardiac surgery and not in other patient categories. Additional studies of intravascular microdialysis in surgical and medical ICU patients would be of value.

We demonstrate that central venous microdialysis is a robust method for continuous blood lactate monitoring in patients undergoing cardiac surgery. This system constitutes a novel and interesting new method of monitoring blood lactate concentration and may be useful in monitoring critically ill patients.