The prognostic value of delta‐lactate in critically ill children

Abstract

Aim

This study aimed to test delta‐lactate (ΔL) as a short‐term risk stratification method in critically ill children.

Methods

An exploratory study of patients admitted to paediatric intensive care unit (PICU) was conducted. ΔL was calculated as the difference between the maximum lactate concentrations on Days 1 and 2. According to the ΔL cutoff, two groups were considered: low mortality risk (LMR) – ΔL ≥ 0.05 mmol/L – and high mortality risk (HMR) – ΔL < 0.05 mmol/L.

Results

Mortality, both during PICU stay and at 28 days, was statistically associated with elevated serum lactate on D1 and D2, per se. For the 93 cases with elevated lactate on Day 1, and a ΔL cutoff of 0.05 mmol/L, the area under the ROC curve was 0.698 (95% confidence interval, 0.47–0.93). HMR patients scored higher PIM3, were not discharged home until 28 days, counted fewer ventilation‐free days and needed renal replacement therapy more often.

Conclusion

Elevated lactate levels at admission, as well as applying the optimal cutoff for ΔL, allowed to predict short‐term mortality: if an increase or minimal decrease in lactate maximum levels occurred from D1 to D2, death was almost eight times more probable. In critically ill children, delta‐lactate predicts short‐term outcome.

What is already known on this topic

1. In critically ill adult patients lactate kinetics seems to be superior to its absolute value.

2. An increase in lactate concentration is accepted as a marker of hypoperfusion and it is used as a predictor of mortality and outcome at discharge.

What this paper adds

1. In the paediatric setting we recognise the importance of lactate concentration trend over time, rather than its absolute value by itself, in predicting mortality.

2. Our results suggest that elevated lactate levels at admission as well as an increase or a small decrease in lactate (delta‐lactate < 0.05mmol/l) in the first 48 hours after admission predicts a short‐term outcome.

In anaerobic conditions, cells under stress increase glycolysis, producing lactate as an end product. Several studies have proven that in contexts of macro and microcirculatory failures like sepsis and shock, hyperlactatemia occurs as a result of this pathway. ¹ Hyperlactatemia is probably a multifactorial result, dependent on each patient's disease pathophysiology. The clearance of lactate produced in human body tissues relies mainly on the liver (up to 70%) and kidney functions. ¹ In the case of hepatic or kidney failure, high lactate levels may arise independently of the presence of shock. Therefore, increased production, decreased clearance, or the combination of both can lead to hyperlactatemia.

Authors have suggested that lactate levels might help detect hypoperfusion while normal vital signs are still present (in trauma patients with important blood loss, for example) and thus predict worse outcomes in patients with occult hypoperfusion. ² , ³ Several processes can result in hyperlactatemia, making its interpretation challenging. ¹ , ⁴ In cardiac arrest, both the absence of blood flow and consequent ischemia–reperfusion inflammation led to an increase in lactate levels. Specifically in children, in cardiac arrest its effective clearance is associated with better outcomes. ⁵

The limit between ‘normal’ and elevated lactatemia is accepted as 2.0–2.5 mmol/L, although a universal cutoff does not exist; ‘high’ lactate levels are often defined as greater than 4 mmol/L. ³ , ⁶

In critically ill patients, an increase in lactate concentrations is accepted as a marker of hypoperfusion, regardless of the underlying cause of hyperlactatemia and infection status. ⁷ It is used as a predictor of mortality and outcome at discharge. ⁸ The Surviving Sepsis Campaign guidelines emphasise the importance of performing serum lactate measurement within the first 3 hours of diagnosis. ⁶ They propose that, in the absence of ScvO₂ (central venous oxygen saturation) values, a delay in lactatemia normalisation is a feasible option in detecting patients with severe sepsis‐induced tissue hypoperfusion and recommend the administration of crystalloids in patients presenting with hypotension or serum lactate ≥4 mmol/L. Similarly, current trauma guidelines recommend risk‐stratifying patients and guiding fluid administration according to lactatemia. ⁴ , ⁹ Identical approaches have been validated in patients admitted to paediatric intensive care units (PICUs). ¹⁰ , ¹¹

Investigators have determined cutoff values for lactate kinetics capable of predicting short‐ and/or long‐term mortality, which serve as endpoints for resuscitation and as markers of therapeutic response. Although consensus has not been achieved, the results look promising. Recently, Pan et al. tested the usefulness of lactate clearance as a specific indicator of resuscitation outcome. Their results suggest lactate clearance per se is superior to ScvO₂ during a standard resuscitation, although an optimal rate remains debatable. ¹² Yusuke et al. compared 23 different lactate‐related indices for in‐hospital mortality prediction and concluded that the area under the ROC curve (ROC‐AUC) of maximum lactate at 24 h after admission was superior to other indices, as comparable with the Acute Physiology & Chronic Health Evaluation III (APACHE III) score. ¹³

In adults, several studies support the aforementioned hypothesis, ¹⁴ , ¹⁵ , ¹⁶ while the research regarding paediatric populations is scarce. In this study, we aimed to test the applicability of delta‐lactate as a short‐term risk stratification method (in‐hospital and at 28 days mortality) in critically ill children, pursuing a better understanding of lactate kinetics in children and standard of care optimisation.

Methods

Study design

We conducted an exploratory study, with retrospective data collection of patients admitted to the PICU for a period of 4 years, from 1st January 2016 to 31st December 2019. All patients included had at least one lactate measurement on D1 (Day 1) and another on D2 (Day 2). Exclusion criteria included the neonatal period, post‐operative admissions, or length of stay shorter than 48 h. A flow diagram of the sample selection process is presented in Figure 1.

Fig. 1.

Study flow diagram. From a restricted access database of a paediatric intensive care unit (PICU), 1564 patients admitted for 4 years were considered. After exclusion criteria were applied, a sample of 249 was obtained.

This study was approved by the Ethics Committee of Centro Hospitalar e Universitário de Coimbra (CHUC‐030‐20).

Demographic and clinical data were collected from a restricted access digital database: age; gender; length of PICU stay; diagnosis on admission; PIM3 score; comorbidities (such as chronic liver disease, chronic kidney disease, obesity or prematurity); maximum lactate concentration on D1 and D2; number of days until lactate normalisation; procalcitonin (PCT) level (at admission, maximum level registered and time until normalisation); need for invasive mechanical ventilation (IMV); ventilation‐free days; associated infection (timing, type of infection and use of antibiotics); acute kidney injury (AKI); the need for renal replacement therapy and diuretics; the need for vasoactive drugs; death during PICU stay; destination after discharge and death at 28 days.

Paediatric index of mortality‐3 (PIM3) score was obtained at admission, based on data collected within the first hour of admission (systolic blood pressure, pupillary light reflex, peripheral oxygen saturation, mechanical ventilation, base excess, elective admission to PICU, recovery from surgery, or a procedure being the main reason for PICU admission and type of diagnosis). The probability of death predicted by PIM3 score, expressed as a percentage, was calculated using a formula applied to the total score. ⁸

Lactate measurements through blood gas analysis did not follow any specific protocol, as they were performed according to the physician's assessment of the patient. All blood gas samples were collected using the same method and immediately processed in the same analyser (GEM Premier 3000 TM) in the PICU.

Infections were considered ‘at admission’ when known at admission or diagnosed in the first 48 h of PICU stay; they were expressed as ‘PICU acquired infection’ when the diagnosis was made after 48 h of PICU stay. We considered PCT to be normal when its concentration was <0.5 ng/dL.

Ventilation‐free days were calculated according to the following rules: zero days were attributed to patients who died and to surviving patients ventilated for ≥28 days; 28 days was considered in patients free of ventilation for ≥28 days; 28 minus the total days of ventilation was calculated for the remainder.

The difference between the maximum lactate concentration on D1 and D2 was used to calculate delta‐lactate (ΔL). We considered lactate to be elevated when ≥2 mmol/L and the day of lactate normalisation when three consecutive lactate measurements were normal (<2 mmol/L).

Statistical methods

Analysis was performed using SPSS (Statistical Package for the Social Sciences), version 25. All reported P values were two‐tailed, with a P value of 0.05 indicating statistical significance. Demographic and clinical variables were used to compare two groups. Correction for disease severity was considered given demographic data similarity, as shown in the next section. Categorical variables were presented as frequencies and percentages, and numerical variables as means and standard deviations or medians and interquartile ranges (IQRs), for variables with non‐normal distribution. Normal distribution was checked using the Shapiro–Wilk test or skewness and kurtosis. The chi‐square test or Fisher exact test was applied to compare qualitative variables; the Mann–Whitney test was used to compare quantitative variables. The area under the ROC curve was analysed to predict mortality in the group of patients with lactate levels of ≥2 mmol/L in D1. The Youden index was used to calculate an optimal cutoff (higher specificity and sensibility) to predict mortality in this sample. For comparative analysis between lower and higher mortality risk populations, two groups were considered based on the calculated cutoff value: ‘low mortality risk’ – LMR (a decrease of ≥0.05 mmol/L in lactate level) and ‘high mortality risk’ – HMR (an overall increase in lactate levels or a decrease <0.05 mmol/L).

Results

Our study included 249 patients, of which 60.6% were males. The median age was 5.6 years (IQR 0.87–13.74). The median length of PICU stay was 7.0 days (IQR 3.0–11.0). The median PIM3 score was 3.1% (IQR 1.1–6.9). Mortality rate during PICU stay was 7.2% (n = 18) and two more children died in the 28‐day period after onset. In total, 20 patients (8.0%) died (Table 1). The highest percentage of deaths occurred in children aged between 12 months and 10 years old (Fig. 2).

Table 1.

Demographic and clinical characteristics of the sample (n = 249)

	Median or n	IQR or %		Median or n	IQR or %
Age at admission (years)	5.6	[0.9–13.7]	PIM3 score (%)	3.1	[1.1–6.9]
Male gender	151	60.6%	Max. lactate in D1 (mmol/L)	1.6	[1.0–2.5]
Admission diagnosis group			Max. lactate in D2 (mmol/L)	1.2	[0.8–1.8]
Respiratory disease	84	33.7%	Lactate normalisation day	2.0	[2.0–3.0]
Trauma	39	15.7%	PCT at admission (mmol/L)	1.35	[0.2–7.9]
Shock	38	15.3%	Max. PCT registered (mmol/L)	2.59	[0.4–15.1]
Neurologic disease	30	12.0%	Infection
Renal disease	17	6.9%	No infection	103	41.4%
Cardiac disease	12	4.8%	At admission	120	48.2%
GI disease	12	4.8%	Hospital‐acquired infection	26	10.4%
Other	17	6.9%	AKI
Comorbidities			No AKI	205	82.3%
No comorbidities	110	44.2%	At admission	10	4.0%
Active cancer disease	36	14.5%	During PICU stay	34	13.7%
Syndrome/malformation	27	10.8%	Diuretic drugs	141	56.6%
Chronic heart disease	22	8.8%	Renal replacement therapy	19	7.6%
Chronic neurological disease	14	5.6%	IMV	140	56.2%
Chronic GI disease	10	4.0%	Ventilation‐free (days)	23.0	[17.0–26.0]
Other	30	12.0%	Vasoactive drugs	91	36.5%
Discharge			Length of vasoactive drugs (days)	4.0	[2.0–6.8]
To another ward/hospital	215	86.3%	PICU length of stay (days)	7.0	[3.0–11.0]
Home	16	6.4%	Mortality during PICU stay	18	7.2%
			Mortality at 28 days	20	8.0%

Demographic and clinical variables considered allowed the characterisation of the 249 patients included and the counting of deaths both during PICU stay and at 28 days after discharge. These data served as the basis for this investigation, as they allowed the comparison of mortality rates and patients' profile after a proven valuable outcome predictor – hyperlactatemia – was applied.

GI, gastro‐intestinal; IMV, invasive mechanical ventilation; IQR, interquartile range; Max, maximum; PCT, procalcitonin; PICU, paediatric intensive care unit; PIM3, paediatric index of mortality‐3.

Fig. 2.

Frequency and death rate of different age groups. Ninety‐three of the 249 patients included presented with elevated lactatemia at admission. Regardless of the maximum lactate level in Day 1, death was more likely to occur in children aged from 1 to 10 years old. MaxLacD1, maximum lactate in Day 1; PICU, paediatric intensive care unit. (), Deaths during PICU stay and (), Frequency.

The median of maximum lactate on D1 of patients who died during PICU stay was higher than in those who survived (3.0 mmol/L (IQR 1.3–9.7), vs. 1.5 mmol/L (IQR 1.0–2.4), P = 0.006). Similarly, patients who died during PICU stay had a higher maximum median lactate level on D2 than those who survived (2.4 mmol/L (IQR 1.3–10.0), vs. 1.1 mmol/L (IQR 0.8–1.7), P < 0.001). Considering mortality at 28 days, the median of maximum lactate was higher in the group of patients who died, both on D1 (2.9 mmol/L (IQR 1.3–9.3) vs. 1.5 mmol/L (1.0–2.4), P = 0.010) and on D2 (2.4 mmol/L (IQR 1.4–8.6) vs. 1.1 mmol/L (IQR 0.8–1.7), P < 0.001), respectively.

In Table 2, we present the characteristics of the 93 patients (37.3%) who had a maximum lactate level of ≥2.0 mmol/L on D1. In this group of patients (n = 93), there were 11 deaths (11.8%) during PICU stay, and one more patient died until 28 days, leading to a total mortality rate of 12.9%. These mortality rates are, respectively, 1.7 times (P = 0.030) and 1.6 times (P = 0.029) higher than the corresponding rate in the initial group of 249 patients. For optimal sensitivity (0.636) and specificity (0.890), death was predicted when lactate values either increased or marginally decreased (less than 0.05 mmol/L) from D1 to D2. The area under the ROC curve was 0.698 (95% confidence interval 0.47–0.93). Based on the calculated cutoff, 17.2% (n = 16/93) of patients were included in the HMR group, while 82.8% (n = 77/93) were included in the LMR group.

Table 2.

Characteristics of patients with elevated lactate at admission, LMR and HMR groups

Data category	Patients with maximum lactate in D1 ≥ 2 mmol/L (n = 93)		LMR group† (n = 77)		HMR group‡ (n = 16)
	Median or n	IQR or %	Median or n	IQR or %	Median or n	IQR or %	P value
Age at admission (years)	9.3	[0.9–14.1]	9.3	[1.0–14.1]	8.8	[0.5–14.1]	0.895§
Male gender	60	64.5%	53	68.8%	7	43.8%	0.056¶
Diagnosis at admission
Trauma	25	26.9%	20	26.0%	5	31.3%
Shock	23	24.7%	18	23.4%	5	31.3%
Respiratory disease	18	19.4%	17	22.1%	1	6.3%
Neurologic disease	9	9.7%	8	10.4%	1	6.3%
Cardiac disease	7	7.5%	6	7.8%	1	6.3%
Renal disease	6	6.5%	4	5.2%	2	12.5%
GI disease	1	1.1%	1	1.3%	0	0.0%
Other	4	4.3%	3	3.9%	1	6.3%
Comorbidities
No comorbidities	49	52.7%	39	50.6%	10	62.5%
Active cancer disease	10	10.8%	7	9.1%	3	18.8%
Syndrome/malformation	10	10.8%	9	11.7%	1	6.3%
Chronic heart disease	8	8.6%	8	10.4%	0	0.0%
Chronic endocrine disease	4	4.3%	3	3.9%	1	6.3%
Chronic neurological disease	4	4.3%	4	5.2%	0	0.0%
Other	8	8.6%	7	9.1%	1	6.3%
PICU length of stay (days)	8.0	[4.0–13. 0]	8.0	[4.0–13.0]	7.5	[3.3–12.5]	0.834§
Discharge
To another ward/hospital	78	83.9%	69	89.6%	9	56.3%
Home	4	4.3%	4	5.2%	0	0.0%
PIM3 score (%)	3.5	[1.4–7.1]	3.3	[1.2–5.8]	15.3	[3.6–63.6]	0.002§
Max. lactate in D1 (mmol/L)	3.2	[2.4–6.15]	3.2	[2.4–6.0]	3.5	[2.6–8.8]	0.359§
Max. lactate in D2 (mmol/L)	1.9	[1.2–3.2]	1.6	[1.1–2.4]	5.6	[3.1–10.7]	<0.001§
Lactate normalisation day	2.0	[2.0–3.0]	2.0	[2.0–3.0]	2.0	[2.0–5.0]	0.168§
PCT at admission (mmol/L)	4.3	[0.8–20.1]	4.4	[0.5–18.9]	2.0	[1.1–32.0]	0.520§
Max. PCT registered (mmol/L)	6.4	[1.2–33.5]	6.4	[1.1–27.9]	7.0	[1.9–43.0]	0.740§
Infection
No infection	43	46.2%	35	45.5%	8	50.0%
At admission	37	39.8%	31	40.3%	6	37.5%	1.000††
Hospital‐acquired infection	13	14,0%	11	14.3%	2	12,5%
AKI
No AKI	70	75.3%	59	76.6%	11	68.8%
At admission	5	5.4%	4	5.2%	1	6.3%	1.000††
During PICU stay	18	19.4%	14	18.2%	4	25.0%
Diuretic drugs	53	57.0%	45	58.4%	8	50.0%	0.535¶
Renal replacement therapy	11	11.8%	6	7.8%	5	31.3%	0.020††
IMV	65	69.9%	53	68.8%	12	75.0%	0.769††
Ventilation‐free‐days (days)	21.0	[9.5–24.0]	21.0	[15.0–24.5]	0.0	[0.0–22.0]	0.011§
Vasoactive drugs	54	58.1%	42	54.5%	12	75.0%	0.131¶
Length of vasoactive drugs (days)	4.0	[3.0–8.0]	4.0	[3.0–8.0]	4.5	[3.0–6.8]	0.983§
Mortality during PICU stay	11	11.8%	4	5.2%	7	43.8%	<0.001††
Mortality at 28 days	12	12.9%	5	6.5%	7	43.8%	0.001††

From the original sample of 249 patients, 93 presented with elevated lactatemia at admission (maximum lactate level in Day 1 ≥ 2 mmol/L): applying this cutoff, different clinical and demographic characteristics and higher death percentage were assessed. Posteriorly, among these 93 patients, an ideal cutoff of 0.05 mmol/L was obtained for Delta‐Lactate (calculated as the difference between maximum lactate levels in Days 1 and 2), which allowed the consideration of two groups: Low Mortality Rate (LMR) ^† – decrease in lactate levels of ≥0.05 mmol/L – and High Mortality Rate ^‡ (HMR) – increase or small decrease (<0.05 mmol/L). The two groups were compared: HMR patients scored higher PIM3, had longer in‐hospital stays, needed replacement therapies more often, counted lesser ventilation free‐days and were more likely to die.

AKI, acute kidney injury; GI gastro‐intestinal; IMV, invasive mechanical ventilation; IQR, interquartile range; Max, maximum; PCT, procalcitonin; PICU, paediatric intensive care unit; PIM3, pediatric index of mortality‐3.

^†Decrease in lactate levels of ≥0.05 mmol/L.

^‡Increase or small decrease (<0.05 mmol/L) in lactate levels.

^§Mann–Whitney U test,

^¶Chi‐square Pearson test,

^††Fisher's exact test.

In both groups, shock and trauma were the most frequent diagnoses at admission. The HMR group scored a significantly higher median PIM3: 15.3% (IQR 3.6–63.6) versus 3.3% (IQR 1.2–5.8) (P = 0.002) in the LMR group. In the HMR group, no patients were discharged home (vs. 5.3% in the LMR group) until 28 days.

During PICU stay, seven patients of the HMR group died versus four patients in the LMR group: mortality rate was more than eight times higher in the HMR group (43.8% vs. 5.2%, P < 0.001). At 28 days, one more patient from the LMR group died. Considering the different age groups (Fig. 3), in LMR patients, death during PICU stay was only reported in the (12 months to 10 years) age group. Trauma (n = 2), respiratory disease (n = 1) and shock (n = 1) were the causes of admission to PICU of those patients. The highest number of deaths in HMR patients occurred in the (10 years to 18 years) age group, of which 66.7% (n = 2) had trauma as the admission diagnosis. The median of maximum lactate levels on D2 (P < 0.001) was higher in the HMR group. AKI was diagnosed in 31.3% (n = 5) of patients in the HMR group – in one of them (6.3%) it was present at admission – versus 23.4% in the LMR group – it was present at admission in four patients (5.2%). HMR patients needed renal replacement therapies more often (P = 0.020). Regarding ventilation support, 75.0% of patients from the HMR group underwent IMV (vs. 68.8% of the LMR patients, P = 0.769). The median of IMV initiation day was Day 1.0 (IQR 1.0–1.0) in both groups (P = 0.638). The median of ventilation‐free days was lower in the HMR patients (0.0 days (IQR 0.0–22.0) vs. 21.0 days (IQR 15.0–24.5), P = 0.011).

Fig. 3.

Frequency and death rate in different age groups in LMR and HMR groups of patients. The 93 patients with hyperlactatemia at admission were divided into LMR group – decrease in lactate levels of ≥0.05 mmol/L – and HMR – increase or small decrease (<0.05 mmol/L). In the HMR group, patients aged from 10 to 18 years old are more likely to die, whereas in the LMR group, death occurred only among patients aged 1 to 18 years old. HMR, high mortality rate; LMR, low mortality rate; PICU, paediatric intensive care unit. (), Deaths during PICU stay and (), Frequency.

Discussion

Resembling the existing literature, this study proved that higher serum lactate levels in the first 24 h of PICU admission, regardless of any normal range or limit, are associated with higher short‐term mortality rates: in our sample, the median of maximum lactate levels on D1 and D2 were 2.0 and 2.2 times higher in patients who died during PICU stay, compared to those who survived. ¹⁰ , ¹⁷ When an upper limit of normality was considered for maximum lactate value on D1 (>2 mmol/L), this group had a higher mortality rate compared to the initial 249 patients.

We recognise the importance of lactate concentration trend over time, rather than its absolute value by itself, in predicting mortality, pinpointing the usefulness of a cutoff value for lactate clearance rate in the therapeutic approach of critically ill children (our study provided a ΔL cutoff of 0.05 mmol/L). Higher mortality during PICU stay was associated with an increase or decrease of 0.05 mmol/L in the maximum serum lactate values (ΔL < 0.05 mmol/L), from D1 (the first 24 h after admission) to D2 (the period from 24 to 48 h after admission).

The fact that, in both LMR and HMR groups, shock and trauma were the most frequent diagnoses reported at admission to PICU reinforces the importance of the incorporation of serum lactate measurements into guidelines for the management of patients in shock, as in The Surviving Sepsis Campaign ⁶ or recent trauma guidelines for risk stratifying patients and guiding fluid administration. ⁹ , ¹⁸

The similarity between both groups' median of maximum lactate levels on D1, in contrast to the difference seen in the median of D2 maximum lactate levels, reinforces the idea that variations in lactate levels are clinically more significant than isolated lactate concentrations. Also, the period from 24 to 48 h after admission seems to be the best period to evaluate patients and reliably estimate their clinical course. Baysan et al. recently advocated adding the predictive value of lactate at 24 h after admission to the Acute Physiology and Chronic Health Evaluation IV model for in‐hospital mortality in critically ill patients with sepsis. ¹⁹

As expected, HMR patients' higher PIM3 scores reflected the higher severity of their diagnosis at admission. ⁸ The fact that no patients in the HMR group were discharged home until 28 days suggests that either the cause of admission to PICU or the subsequent sequelae led to a longer in‐hospital stay. ²⁰

Although AKI diagnosis was similar in both groups, patients in the HMR needed renal replacement therapy more often, proving the severity of renal function impairment and renal replacement therapy techniques can influence lactate elimination in those patients. ²¹

Lesser ventilation‐free days suggest the unfavourable evolution of HMR patients. Probably, this group of patients reached clinical and laboratory criteria for ventilator weaning later than those with a more favourable outcome ²² and, additionally, are susceptible to the potential consequences of prolonged immobilisation and sedation. ²³

We identified several potential limitations throughout the course of our investigation. The heterogeneity of the sample, despite the exclusion criteria, suggests the need for homogenised samples in future studies. The data were collected without some confounding variables being gathered such as the selection of a more severe diagnosis or comorbidity to the detriment of concomitant (but less determinant) diagnoses or comorbidities, based on the authors' perspective. Given its retrospective nature, an insufficient number of lactate measurements, or registrations of blood gas analysis results, led to a decrease in the number of cases included. Also, in a large number of patients, it was seen that a normal first blood gas analysis did not prompt a second one, precluding the interpretation of delta‐lactate and reducing our sample. On the other hand, by limiting the sample to critically ill children, we can guarantee the validity of delta‐lactate as a predictor of outcome in this particular group of patients; the analysis of the area under the ROC curves allowed, even so, to obtain a robust statistical model. The short follow‐up period did not allow long‐term mortality or sequelae prediction.

Conclusion

In conclusion, an increase or a small decrease (<0.05 mmol/L) in ΔL reflects in higher PIM3 scores, the need for renal replacement therapies, lesser ventilation‐free days, longer in‐hospital stay and death.

Our results motivate further prospective and controlled studies on the subject, comparing ΔL with both short‐ and long‐term prognoses. Specific protocols should guide the frequency per day or interval between arterial gas analysis. ⁹ Finally, a constellation of sequelae, recently referred as Intensive Care Syndrome ²⁴ should be incorporated into prognosis analysis, particularly in children. ²⁵ The complexity of lactate production, metabolism and elimination should be kept in mind, exhaustively pursuing the cause of hyperlactatemia.