Translate this page into:
Comparison of venous lactate values on blood gas analyser versus laboratory autoanalyser: A preliminary study
For correspondence: Dr Smita Tripathi, Department of Biochemistry, Lady Hardinge Medical College, New Delhi 110 001, India. e-mail: shipibimbi@gmail.com
-
Received: ,
Accepted: ,
Abstract
Background & objectives
Lactate can be estimated by point-of-care testing (POCT) devices such as blood gas analysers (BGA) as well as central lab autoanalysers (AA); both have their advantages. Studies have compared lactate and electrolyte values on arterial blood; none used venous samples which are preferred in children. We planned to compare venous lactate by BGA and AA in children admitted to Paediatric Intensive Care Unit (PICU).
Methods
Fifty children aged one month to 18 yr admitted in PICU were included in the study after parental consent. Paired venous samples were collected in heparinised syringes for BGA, in fluoride containing vacutainers for AA and were analysed immediately.
Results
Lactate levels were assessed in 29 male and 21 female children [Median age: 4 yr interquartile range (IQR)=1-10)]. AA reported higher lactate levels but mean difference between two methods (0.34±1.56 mmol/L) was not statistically significant (P= 0.062). The methods showed statistically significant correlation [95% confidence interval (CI), Spearman correlation coefficient (rs)=0.816, P=<0.001]. Bland Altman plot showed 95 per cent of paired values had differences which was within limits of agreement (-2.71, +3.39); differences were more variable above 2 mmol/L. Both devices classified multiple samples into different clinically significant ranges.
Interpretation & conclusion
Venous lactate by BGA and AA was found comparable, but critical values requiring clinical intervention were not reported similarly by both analysers. Variability between them was higher at higher venous lactate levels (>2 mmol/L) which represent critically ill children. Use of the same analyser, either BGA or AA, is therefore recommended for monitoring.
Keywords
Auto-analyser
comparative study
critically ill
lactate
pediatric population
PICU
POCT
septic shock
Lactate is an important parameter for investigation conducted in the Paediatric Intensive Care Units (PICUs) which helps assess the prognosis of critically ill children. Lactate levels rise with decrease in tissue perfusion due to developing sepsis or septic shock and can hence help in predicting mortality in children1. Blood gas analysers (BGA) are usually preferred for estimating levels of lactate as point-of-care testing (POCT) devices have several practical advantages. They have significantly reduced turnaround time (TAT), allowing faster diagnosis and rapid intervention, resulting in improved management of acute conditions. They allow for reduced preanalytical errors, improved hospital efficiency as well as reduced hospital stay2. The use of POCT devices has been shown to be cost efficient3, and the low sample volume requirement is useful in paediatric settings where frequent blood sampling may contribute to anemia4. The BGA works on principles of detecting electrons produced from oxidation of lactate into pyruvate using electrodes5. BGAs use heparinised whole blood, and are operated by physicians or other supporting staff, who may underestimate the importance of quality control6. Lactate levels can also be estimated in plasma in Central Laboratory Autoanalysers (AA) that use lactate oxidase enzymatic methods for generation of coloured substances measured spectrophotometrically5. These analysers are typically operated by dedicated laboratory personnel, and quality control protocol is strictly followed. However, AAs have a much longer processing time and the Central Laboratory may not be in proximity to patient wards6.
The Central Laboratory Autoanalyser is the method of choice for lactate estimation in terms of accuracy5. However, BGAs are preferred by clinicians. Analyser specific differences in bias or precision can lead to statistically or clinically significant discrepancies in results obtained by the two methods. It is, therefore, imperative for clinicians to know the comparability of these two methods. If the methods are used interchangeably during treatment or follow up, the results must be interpreted with caution.
Worldwide, studies have been conducted to estimate the differences between the two methods. An American study7 in 2007 and a Finnish study8 in 2011 compared adult arterial blood samples on five different analysers (3 POCT and 2 core laboratory) and obtained good correlation between the methods. A study9 in Turkey in 2014 also found negligible variability between adult arterial blood samples assessed on one core laboratory and two blood gas analysers. A German study10 in 2021 had compared adult arterial blood on one POCT device with venous blood on two core laboratory analysers and obtained statistically significant variability, but no difference in clinical outcome. An Indian study11 in 2021 compared only electrolytes on one POCT and one core laboratory autoanalyser using paediatric venous blood samples and obtained acceptable differences between them.
Measurement of lactate in an arterial blood sample is a better indicator of lactate production than in a venous blood sample12. However, venous samples are preferred for monitoring children in our institution because venous sampling is convenient and less invasive as compared to the already sick paediatric patients. Limited studies are available comparing BGAs and AAs using venous blood. Studies from India have compared electrolytes11, however, no study so far has been conducted comparing lactate levels using venous blood samples. In this preliminary study, we aim to compare venous lactate in individuals admitted to the PICU by the above-mentioned methods to find out if there is any statistically significant difference in their values.
Materials & Methods
This prospective observational study was conducted in a 20-bedded PICU of the department of Paediatrics, Lady Hardinge Medical College, New Delhi, a tertiary care teaching hospital over a period of two months (August-September 2022). Ethical clearance for the study was obtained from the Institutional Ethics Committee.
Sample size
Considering the mean lactate levels in patients admitted to the PICU as 6.67 mmol/L and standard deviation as 5.4413, and assuming that the difference in measurements is not more than 8 per cent i.e., as 0.53 (8% of 6.67)10, the minimum sample size was found to be 32; at 95 per cent level of significance and 80 per cent power. However, this sample size was calculated according to the difference reported between arterial and venous lactate. Therefore, a convenience sample of 50 children admitted to the PICU (aged 1 month to 18 yr) during the study period was considered.
Sample collection
Written and oral consent from parents of the study participants was obtained and patient confidentiality was maintained throughout the study. Paired venous samples were taken from cases within 2 h of admission. Venous blood was drawn, and not arterial, as per the PICU policy to monitor venous blood gases in sick children.
Sample collected in grey cap vacutainers (containing sodium fluoride and potassium oxalate anticoagulant) was sent to central laboratory for lactate estimation on autoanalyser-AU680 Beckman Coulter Clinical Chemistry Autoanalyser (Brea, CA, USA). It was ensured that these samples were transported on ice packs to the central laboratory by PICU staff. Plasma was separated from cells within 15 min of collection5 and samples were analysed within a time window of 2 h. The Central Biochemistry Laboratory of our institute is affiliated to the External quality assurance programme by Randox (RIQAS, Dublin, Ireland) and Internal Quality Controls are provided by Beckman Coulter Inc. (Brea, CA, USA). The Internal Quality Checks (IQC) are performed twice a week according to the manufacturer’s instructions. Lab turnaround time for emergency parameters is 2-3 h. Lactate is analysed on auto-analysers with a test linearity range of 0.22-13.32 mmol/L (2-120 mg/dl). It is sensitive (lowest detectable value 0.01 mg/dL) and has a coefficient of variation (CV) of <2 per cent14.
Venous blood was also taken in heparinised syringes for estimation of lactate by the blood gas analyser-Radiometer ABL 800 Flex (Radiometer Medical ApS, Bronshoj, Denmark) located inside the PICU and analysed immediately by doctors or nursing staff. The POCT device was not part of any external quality assurance programme. Internal Quality Checks (IQC) were performed twice a week by PICU staff.
The results from the BGA were obtained in mmol/L. The results of the AA were obtained in mg/dL and were converted to mmol/L by multiplying with a conversion factor of 0.1115.
Statistical analysis
Data were coded and recorded on Microsoft Excel Version 2410. SPSS v30 (IBM Corp.) was used for data analysis. Normality of data was assessed using the Shapiro-Wilk test which showed that differences between the paired samples were not normally distributed (P<0.001). Wilcoxon Signed Ranks test was used for comparing means of the two methods. The two methods were compared by Spearman’s rank correlation and Passing Bablok regression. Agreement between clinically relevant ranges of lactate was assessed by creating a contingency table and calculating the Weighted kappa. The bias between the two methods was displayed using the Bland Altman plot. Statistical significance was kept at P<0.05.
Results
The clinical and biochemical profile of 50 children (29 males and 21 females) at the time of admission to the PICU is shown in table I.
| Parameters | Value |
|---|---|
| Number of femalesa | 21 (42%) |
| Median age (yr)b | 4 (1-10) |
| Duration of stay in PICU (days)c | 8.72±5.79 |
| Number of cases with acidosis (pH<7.35) at admissiona | 22 (44%) |
| Number of cases with elevated lactate (≥2 mmol/L) at admission (by either method)a | 23 (46%) |
| Number of patients requiring ventilator supporta | 30 (60%) |
| Number of cases diagnosed with sepsis or septic shocka | 12 (24%) |
| System involved at the time of admission* | |
| Renal | 14 |
| Cardiovascular | 14 |
| Haematological | 12 |
| Respiratory | 17 |
| Neurological | 11 |
| Others | 16 |
The BGA and AA measured lactate levels across ranges of 0.4 - 8.7 mmol/L and 0.67 - 10.75 mmol/L respectively. Mean lactate level was 2.37±1.67 mmol/L by BGA and slightly higher, 2.7±2.17 mmol/L by the central lab AA, whereas the median was 2 (1.4-3) mmol/L [Median (IQR)] by BGA and 1.85 (1.2-3.3) mmol/L by the AA. The difference of mean value between the two analysers was 0.34±1.56 mmol/L and it was not statistically significant (Wilcoxon Signed Ranks test; Z=-1.868, P= 0.062).
In our results, tests for correlation showed a strong positive correlation (Spearman’s Correlation Coefficient (rs)=0.816) which was statistically significant (P<0.001). The Passing Bablok regression gave a slope of 0.8190 (with 95% CI of -0.2552 to 0.4237) and intercept of 0.1045 (with 95% CI of 0.6781 to 1.051), implying good correlation between the results.
The Bland-Altman plot of the difference between methods (X-axis) and the mean of the two methods (Y-axis) is shown in figure. The mean difference of the two methods was 0.34 with a standard deviation of 1.56 and 95 per cent of paired values had a difference which was within the limits of agreement (LoA) of (+3.39, -2.71). From this plot, it can be seen that the lactate values are comparable between the two analysers at lower values (<2 mmol/L). However, it was seen that above 2 mmol/L, the difference obtained between the results of the two analyzers becomes more variable/erratic.
- Bland-Altman plot for lactate values measured by Blood Gas Analyser (BGA) and Autoanalyser (AA). Difference between BGA and AA values on X-axis and mean of BGA and AA values on Y-axis. The solid line shows the mean difference between the two methods and the dashed lines represent 95 per cent confidence intervals. Difference was calculated between lactate values from AA and BGA of the same study participant. Mean Lactate was calculated by taking the mean of the values from both BGA and AA of the same study participant.
We further divided all our results into three groups according to clinically relevant lactate levels, ≤2, 2-4, and ≥4 mmol/L, as shown in table II. The lactate values obtained by the two methods showed substantial agreement across these categories. They agreed in 74 per cent cases (P<0.001). However, many samples were distributed into different ranges by both methods. The disagreement between BGA and AA was for 2 per cent samples in the ≤2 mmol/L range, for 4 per cent samples in the 2-4 mmol/L range and for 6 per cent samples in the >4 mmol/L range.
| Lactate | Lactate (Autoanalyser), n (%) | Weighted Kappa | |||||
|---|---|---|---|---|---|---|---|
| ≤2 mmol/L | 2-4 mmol/L | >4 mmol/L | Total | k | P value | ||
| Lactate (Blood gas analyser) | ≤2 mmol/L | 23 (46) | 4 (8) | 1 (2) | 28 (56) | 0.827 | <0.001 |
| 2-4 mmol/L | 4 (8) | 8 (16) | 3 (6) | 15 (30) | |||
| >4 mmol/L | 0 (0) | 1 (2) | 6 (12) | 7 (14) | |||
| Total | 27 (54) | 13 (26) | 10 (20) | 50 (100) | |||
Samples measured by each method were categorised into three clinically relevant ranges, less than or equal to 2, 2-4 and greater than 4 mmol/L and the number of samples falling into each category by both methods were recorded. The dark blue cells on the diagonal represent the number of paired samples where lactate values fell into the same category by both BGA and AA. The light blue cells represent the number of paired samples where lactate values fall into different categories by BGA and AA. Adding frequencies of the dark blue cells shows that BGA and AA showed agreement in 76% cases, and this agreement was statistically significant
Discussion
All previous studies showed good correlation between BGAs and AAs7-11, which was also seen in our study. Prior studies7,8,10 had also observed that blood gas analysers on an average resulted in lower lactate values, which was also seen in our study.
Above 2 mmol/L, we observed that differences between results obtained by the BGA and AA were more variable. This is similar to what was observed in an American study7, however, in their results, discrepancies were more significant beyond 6 mmol/L. Multiple paired samples were categorised into different clinically significant ranges by both analysers, especially above 2 mmol/L, above which clinical intervention is usually required. There is therefore a risk that the same patient may have a high value reported by one method, and if the methods are interchanged, the follow up value may be low. This could lead to clinical misinterpretation and errors in patient evaluation.
This study was not without limitations. It had a small sample size and hence it lacks the power to satisfactorily conclude interchangeability of the analysers. Our results have limited generalisability as analysers and protocols vary across institutions; we only compared two specific analysers at a single tertiary care centre. The central laboratory was located at a distance from the PICU, hence, preanalytical errors such as temperature control and delay in processing, may be expected15. There may be an increase in lactate content due to cellular metabolism16, which might explain the lower lactate values seen on the blood gas analysers.
It is advisable to use results from the same analyser for follow up, especially in critically ill patients who may have high lactate levels. Every institution that follows two different methods for the same parameter must define comparability between them for better patient management. A study with larger sample size along with correlation of patient outcomes can be useful.
Overall, venous lactate levels by blood gas analyser and central lab auto-analyser in children admitted to the ICU were found to be comparable, with strong correlation (Interclass Correlation Coefficient=0.68, P<0.001) and substantial agreement (P<0.001). However, differences in lactate values become more variable at higher lactate levels (>2 mmol/L) and 2, 4 and 6 per cent paired samples respectively in clinically relevant categories of ≤2, 2-4 and >4 mmol/L respectively were classified into different categories by both analysers. Clinicians are therefore advised to use the same analyser consistently for patient monitoring.
Financial support & sponsorship
The first author (ASK) received funding support under the Indian Council of Medical Research (ICMR), New Delhi for the Short-Term Studentship (STS) Programme (Reference ID: 2022-09504).
Conflicts of Interest
None.
Use of Artificial Intelligence (AI)-Assisted Technology for manuscript preparation
The authors confirm that there was no use of AI-assisted technology for assisting in the writing of the manuscript and no images were manipulated using AI.
References
- Pediatric sepsis: a summary of current definitions and management recommendations. Curr Pediatr Rep. 2023;11:29-3.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- POCT: An inherently ideal tool in pediatric laboratory medicine. EJIFCC. 2021;32:145-57.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- The cost-effectiveness of point of care testing in a general practice setting: results from a randomised controlled trial. BMC Health Serv Res. 2010;10:165.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Anemia, blood loss, and blood transfusions in North American children in the intensive care unit. Am J Respir Crit Care Med. 2008;178:26-33.
- [CrossRef] [PubMed] [Google Scholar]
- Burtis CA, Ashwood ER, Bruns DE, eds. Tietz Textbook of Clinical Chemistry and Molecular Diagnostics (4th ed). WB Saunders Company; 1999.
- Practical challenges related to point of care testing. Pract Lab Med. 2015;4:22-9.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Comparison of lactate values between point-of-care and central laboratory analyzers. Am J Clin Pathol. 2007;128:168-71.
- [CrossRef] [PubMed] [Google Scholar]
- Interchangeability of blood gas, electrolyte and metabolite results measured with point-of-care, blood gas and core laboratory analyzers. Clin Chem Lab Med. 2011;49:1187-91.
- [CrossRef] [PubMed] [Google Scholar]
- Comparison of blood gas, electrolyte and metabolite results measured with two different blood gas analyzers and a core laboratory analyzer. Scand J Clin Lab Invest. 2015;75:97-105.
- [CrossRef] [PubMed] [Google Scholar]
- Blood-gas vs Central-Laboratory analyzers: Interchangeability and reference intervals for sodium, potassium, glucose, lactate and hemoglobin. Heliyon. 2021;7:e08302.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Point of care testing of serum electrolytes and lactate in sick children. EJIFCC. 2021;32:158-66.
- [PubMed] [PubMed Central] [Google Scholar]
- Correlation and agreement between arterial and venous blood gas analysis in patients with hypotension-an emergency department-based cross-sectional study. Int J Emerg Med. 2023;16:18.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Serum lactate levels as the predictor of outcome in pediatric septic shock. Indian J Crit Care Med. 2011;15:102-7.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]
- Beckman Coulter. AU/DxC AU Instructions for useBLOSR6X93EU.02 LACT Metabolite. Available from: https://www.beckmancoulter.com/download/file/phxBLOSR6X93EU02-EN_US/BLOSR6X93EU02?type=pdf, accessed on June 3, 2025.
- Types and frequency of preanalytical mistakes in the first Thai ISO 9002:1994 certified clinical laboratory, a 6 - month monitoring. BMC Clin Pathol. 2001;1:5.
- [CrossRef] [PubMed] [Google Scholar]
- Lactate metabolism in human health and disease. Signal Transduct Target Ther. 2022;7:372.
- [CrossRef] [PubMed] [PubMed Central] [Google Scholar]