Discussion

These analyses demonstrate that serum lactate is not significantly associated with oxygen delivery metrics, including venous saturation, cerebral oxygen extraction ratio, and renal oxygen extraction ratio. Serum lactate was also found not to be significantly associated with other conventional haemodynamic metrics including heart rate, blood pressure, or arterial saturation after regression analyses were conducted.

These data demonstrate that serum lactate may not be a robust marker of systemic oxygen delivery in paediatric patients cared for in a cardiac ICU. It must be noted that a significant but weak correlation was noted between venous saturation and serum lactate when the venous saturation was below 70%. There are limited data regarding this in any patient population, particularly in paediatrics. ^{Reference Klee, Rimensberger and Karam3–Reference Sneha, Mhaske, Saha, Gupta and Singh6} Klee and colleagues studied 93 paediatric patients after cardiac surgery with cardiopulmonary bypass, with each patient getting a gas immediately upon ICU admission, 4 h after admission, and 12 h after admission. This study found that lactate did not correlate well with oxygen extraction ratio and that 55% of patients with a lactate of over 2 mmol/L had an oxygen extraction ratio over 30%. This study did note a correlation between lactate and glucose levels. ^{Reference Klee, Rimensberger and Karam3} Similarly, Schranz and colleagues did not find a correlation between mixed venous saturation and serum lactate after cardiac surgery in infants. ^{Reference Schranz, Schmitt and Oelert4}

A few studies have characterised the correlation between lactate and venous saturation in adult populations. While some of these have demonstrated correlation between lactate and adequacy of oxygen delivery, most have demonstrated no correlation. Even those studies demonstrating a correlation only demonstrate a weak correlation. ^{Reference Bisarya, Shaath, Pirzad, Satterwhite, He and Simpson5,Reference Sneha, Mhaske, Saha, Gupta and Singh6}

As the body and the component organs require oxygen to function, it seems logical that monitoring systemic oxygen delivery would be important. ^{Reference Loomba and Flores7,Reference Loomba8} Systemic oxygen delivery is the product of cardiac output and oxygen content. Cardiac output is quantified by the Fick principle as the quotient of oxygen consumption and the arteriovenous oxygen content difference. Oxygen content is a function of haemoglobin, saturation, and the partial pressure of oxygen. Other than arterial saturation, other continuously, noninvasively monitored haemodynamic variables (heart rate and blood pressure, for example) are not directly related to systemic oxygen delivery. In fact, in critically ill paediatric patients, the inadequacy of routine monitoring to assess systemic oxygen delivery has been demonstrated. ^{Reference Dhillon, Yu, Zhang, Cai and Li9}

More invasive metrics used to assess the adequacy of systemic oxygen delivery have been less frequently studied. Lactate is one such metric. Lactate is produced by the body mainly as a product of glycolysis. Under normal conditions, as glucose feeds into the glycolytic pathway, pyruvate is produced to then feed into the Krebs cycle and undergo oxidative phosphorylation. However, under conditions such as low oxygen content or low NAD+ (Nicotinamide Adenine Dinucleotide), lactate is produced to compensate. Lactate can also be produced by the body under normoxic conditions. The now well-studied Warburg effect describes proliferative tissues’ production of lactate in favour of pyruvate by means of glycolysis despite the presence of oxygen. While this process is most often thought of in terms of tumour cells, studies have found the Warburg effect to occur in non-malignant processes as well.

Both, venous and arterial lactate levels serve as critical markers in assessing tissue perfusion and cellular oxygenation. In instances of inadequate tissue oxygen delivery or impaired cellular metabolism, such as shock or sepsis, the body’s compensatory mechanisms lead to an increase in lactate production. ^{Reference Duke, Butt and South10–Reference Samaraweera, Gibbons, Gour and Sedgwick12}

Arterial lactate levels reflect the amount of lactate present in the bloodstream, indicating the balance between lactate production and clearance by the liver and other organs. Elevated arterial lactate levels suggest tissue hypoperfusion and anaerobic metabolism. In contrast, venous lactate levels, typically obtained from central venous or mixed venous blood samples, represent the amount of lactate returning to the heart from the tissues. Venous lactate levels can thus provide insights into the adequacy of tissue oxygenation and the effectiveness of tissue perfusion. Discrepancies between arterial and venous lactate levels may indicate the presence of tissue-specific metabolic abnormalities or regional variations in perfusion. Despite these differences, both arterial and venous lactate levels are valuable and interchangeable in clinical practice as indicators of poor perfusion. ^{Reference Mahmoodpoor, Shadvar and Sanaie11,Reference Theerawit and Na Petvicham13,Reference Oi, Mori and Yamagata14} They can guide resuscitative efforts and therapeutic interventions aimed at improving tissue oxygenation and restoring perfusion.

Understanding that lactate can be produced in even normoxic conditions is important as lactate’s current role in medicine is often centred on its use as a marker for anaerobic metabolism. Studies have discerned many more functions of lactate throughout the body beyond being a marker, albeit not sensitive or specific, of anerobic metabolism. Lactate plays a role in acute inflammation, chronic inflammation, wound healing, and intracellular signalling among others. ^{Reference Brooks, Arevalo, Osmond, Leija, Curl and Tovar15,Reference Gupta16}

While lactate may not correlate with adequacy of oxygen delivery, it should not be denied that it demonstrated prognostic ability. The current study did not investigate the prognostic ability of lactate, but previous studies have done so in paediatric cardiac ICU patients and have demonstrated prognostic ability. ^{Reference Ranucci, Isgro and Carlucci17–Reference Loomba, Dyamenahalli and Savorgnan24} Adequacy of oxygen delivery, systemic or regional, is best done by either directly measuring an underlying venous saturation or by using a surrogate such as near-infra-red spectroscopy. ^{Reference Loomba8,Reference Loomba, Rausa and Sheikholeslami25,Reference Sheikholeslami, Dyson and Villarreal26}

This study is one of few paediatric studies that investigate the correlation between lactate and venous saturation and utilise high-fidelity haemodynamic monitoring data. The use of high-fidelity haemodynamic monitoring data allows for maximisation of data available for the analyses, strengthening power.

Despite the strengths of this study, it is not without its limitation. This is a single centre study so centre specific considerations may impact the findings of this study and thus its generalizability. Some will criticise the low patient number. It is important to note that sample size in such time series analyses is not the patient number but the number of time points. A previous study has, nonetheless, investigated model quality based on patient number and time point number and found that with approximately 35 patients, 8 time points per patient, or a sample size of 280 total data points, results in high model quality. ^{Reference Hecht and Zitzmann27} With approximately 6 time points per patient in the current study, there is good model quality. The best correlate of this statistical phenomenon is found in meteorological modelling in which the temperature or precipitation for a specific city is being modelled. The sample size in such an analysis isn’t one, the number of cities from which the data is drawn, but rather is the number of time points being used from that city. Further discussion of this phenomenon is beyond the scope of this manuscript although a reference is provided here. Another limitation of our study is the potential variability associated with femoral venous line saturations, which may impact the reliability of the data obtained. Finally, it is important to note the potential confounding effect of hyperglycaemia on lactate levels specially in postsurgical patients. Distinguishing patients with hyperglycaemia from those with hypoperfusion solely based on lactate levels can be challenging, as both conditions may contribute to elevated lactate levels.