Comparing the accuracy of point-of-care with laboratory (capillary, venous, and arterial) blood glucose levels in critically ill patients with and without shock
BMC Research Notes volume 15, Article number: 372 (2022)
To compare the accuracy of point-of-care capillary and venous/arterial samples to laboratory testing of venous/arterial samples in critically sick shocked and non-shocked patients. This is a prospective case–control study including capillary, venous, and arterial blood samples from 268 critically ill patients. The King Fahd Military Medical Complex in Dhahran, Saudi Arabia, was the site of this investigation.
We were able to obtain data on 268 patients for this investigation. POCT and lab findings of venous and central blood did not differ significantly (P = 0.389 and 0.208), while POCT indicated somewhat higher results with venous glucose concentrations of 10.18 and 10.05 (POCT and lab tests respectively) and 9.18 and 9.54 (POCT and lab tests respectively). In addition, the mean differences between POC and laboratory analyses of venous, arterial, and central glucose were 0.13, − 1.75, and − 0.36 mmol/L for venous, arterial, and central glucose, respectively. Except for arterial blood glucose, we did not observe a significant difference between POCT and routine laboratory analysis of glucose concentrations in critically ill patients. Compared to laboratory blood analysis, the use of POCT is marginally accurate, with no difference between shocked and non-shocked patients.
Maintaining normoglycemia in patients with preexisting diabetes or stress-induced hyperglycemia is one of the most crucial components of intensive care. It has been determined that hyperglycemia has a deleterious effect on these patients [1, 2]. Appropriate management of hyperglycemia has been proven to have a considerable influence on lowering mortality and hospital length of stay, preventing acute renal injury, and facilitating a quicker weaning from mechanical breathing [1, 3, 4]. In addition, hypoglycemia is related with adverse effects  and has been identified as an independent predictor of mortality [6, 7] in a number of investigations.
Many hospitals utilize point-of-care (POC) glucose meters to monitor glycemic status in order to meet these aims. Patients requiring strict glycemic control, for whom waiting for central laboratory findings makes rapid modifications and management of glucose level in therapy difficult, benefit greatly from the mobility, simplicity of use, and immediate availability of results that POC glucose meters offer. In stable outpatients, the majority of glucometers were found to be reliable; however, among critically ill patients, a number of confounding factors, such as hematocrit, oxygenation, acid–base disturbance, temperature, and shock states, were reported to interfere with POC glucometers [10, 11].
Considering the presence of shock, several mechanisms have been proposed as possible explanations for its impact on the accuracy of POC glucometers, including peripheral vasoconstriction in hypoperfusion states, which could result in increased glucose extraction by tissues due to low capillary flow, leading to a falsely underestimated glucose measurement with capillary blood [12, 13]. Several papers on monitoring blood glucose levels in critically ill patients demonstrated significant variance between point-of-care and laboratory values, but did not distinguish between shocked and non-shocked patients [4, 5] [1, 14,15,16,17,18,19]. Previous studies conducted on critically sick patients contained a small number of measures taken from individuals in shock, resulting in a heterogeneous sample. In this study, we intended to assess the accuracy of point-of-care capillary and venous/arterial samples to venous/arterial samples analyzed in the laboratory in critically sick patients who were either shocked or not shocked.
Material and methodology
This is a prospective case–control study of 268 critically ill patients hospitalized to the King Fahd Military Medical Complex in Dhahran, Saudi Arabia. Inclusion criteria comprised any adult, non-pregnant patients aged 18 or older who were hospitalized to the hospital with or without diabetes millitus and whose key decision makers provided informed consent. Exclusion criteria included hypovolemic shock due to significant active bleeding, bleeding disorders, the use of substances that could interfere with POC glucose meter technology (such as icodextrin-containing solutions, intravenous immunoglobulins, abatacept, and maltose), and lack of consent.
In this study, demographic characteristics such as age, gender, and anthropometric measures, as well as previous data such as chronic comorbidities, were acquired from a record review in addition to baseline laboratory findings. Patients’ initial vital signs, the Glasgow Coma Scale score with derived verbal scores for intubated patients, the necessity for a ventilator or dialysis, and the presence of acute severe arrhythmias comprised their baseline clinical features. In the laboratory, arterial blood gas analysis, serum creatinine, albumin, and a complete blood count were recorded. The laboratory data obtained closest in time to the blood glucose measurement were recorded.
Three blood samples were taken from the venous, arterial, and central blood vessels, as well as the capillaries, in order to collect data on the findings of POC and lab analysis of blood for glucose concentrations. We gathered and utilized data regarding the outcomes of POC and laboratory analyses for our comparison. The tests were performed by the same bedside nurse at the request of the treating physician.
All data were entered, manipulated, and analyzed using SPSS version 26. Frequency and percentage were utilized to characterize categorical variables, whereas mean and standard deviation were employed to characterize continuous variables. T-paired test was used to analyze the potential difference between glucose measurements obtained from POC and lab analysis of blood. All statements with P values less than or equal to 0.05 are deemed significant.
We were able to collect data on 268 patients admitted to King fahad military Medical Complex, Dhahran, Saudi Arabia, for this study. 26.1% of these patients with shock were male, whereas 18.3% were female. The average age is 63, 01 years. 20% of the included stunned patients were in the heart department, whereas only 2.5% were in the neurosurgery department. In addition, 27.7% of the patients with shock were diabetic, 25.3% required hemodialysis, and 59.3% were ventilated (Table 1).
Moreover, 78.5% of hospitalized patients require insulin infusion. In addition, we discovered that 55.6% of the patients were not shocked, while 44.4% were. In evaluation of patient baseline characteristics, all relevant data were listed in Table 2. In the previous 24 h, the mean Urine Output for the shocked patients was 2195.28 ml compared to 2064.9 ml for the non-shocked patients. The average weight of patients in the shocked group was 74.9 kg compared to 77.3 kg in the non-shocked group. The mean temperature of hospitalized patients was 36.82 degrees Celsius, and their average heart rate was 91.2 beats per minute. The Sequential Organ Failure Assessment (SOFA) has an average score of 9.41. On the basis of laboratory and clinical data, the SOFA score forecasts ICU mortality. The score for the Acute Physiology and Chronic Health Evaluation (APACHE 2) was 26.3. APACHE 2 assesses ICU mortality based on a variety of laboratory data and patient symptoms, taking into consideration both acute and chronic diseases.
In Table 2, we compared the venous/arterial glucose readings of critically sick patients who had POC analysis against lab analysis. Point of Care Testing (POCT) and lab results of both venous and central blood showed no significant difference (P = 0.389 and 0.208), although POCT showed slightly higher results with venous glucose concentrations of 10.18 and 10.05 (POCT and lab tests respectively) and 9.18 and 9.54 (POCT and lab tests respectively). The main difference between POCT and lab analysis of glucose concentrations was seen in arterial glucose concentrations, where lab analysis revealed significantly higher glucose concentrations (P = 0.029) with 11.41 against 9.66 in POCT analysis (Table 2).
In addition, we compared the findings of POCT and Lab analysis of glucose concentrations in shocked and non-shocked patients in Table 3. The results demonstrated that there is no substantial difference between the results of POCT and lab analysis between patients who were shocked and those who were not, with only a tiny discrepancy between POCT and lab analysis. The POCT revealed that the venous glucose concentration was greater in non-shocked patients and lower in the case of capillary analysis; however, laboratory analysis revealed that the venous glucose concentration was greater in shocked patients and somewhat lower in the case of central analysis.
Patients on a strict glycemic protocol and at increased risk of hypoglycemic episodes must have their glucose levels measured accurately. This is generally the case in this study’s population, as many of these patients are unable to interact with physicians or nurses, and their hypoglycemia symptoms are not readily available. POCT has a number of advantages over conventional blood glucose testing, including the availability of glucose values to the nurse within two minutes and immediate visibility in the hospital information system. In addition, POCT devices require a negligible volume of blood and the risk of blood spillage from the syringe or the device is minimal .
In order to implement a protocol for glucose regulation, it is necessary to measure blood glucose levels rapidly and accurately [21, 22]. The application of these protocols increases the nurse’s burden, hence it must be feasible [12, 13, 23]. This implies that not the most exact equipment, but one that is the most practical and provides reasonably accurate glucose analysis would be chosen for this process. In critically ill patients, however, hypoglycemia is critical, and its warning signals are missing; hence, these devices must also be highly reliable in the low range [14,15,16,17,18,19].
When we compared POCT results with lab analysis of glucose concentrations, we discovered that there is no difference between venous and central glucose, however arterial glucose concentrations differ significantly. In addition, we discovered that the POCT exaggerated the venous and central glucose concentrations, while underestimating the arterial glucose concentration. Petersen J et al.  and Boyd et al.  and Critchell et al.  similarly found that glucose meters overstated blood glucose levels in arterial, central, venous, and capillary samples relative to reference standard concentrations. In a different study conducted by Clarke et al. the authors found that the subcutaneous CGMS was accurate in the euglycemic range  and in a study conducted by Goldberg et al. they discovered that the POCT had a pearson correlation coefficient of 0.88 with 98.7% of patients falling within the clinically acceptable zones . In a study conducted by Cook et al. the authors discovered that lab glucose values for blood from catheter in critically ill patients were significantly different from POC values for blood from catheter (P = 0.001) and fingerstick (P = 0.001) . In addition, a second study revealed that the clinical agreement between POCT and laboratory analysis is greater in central blood analysis than arterial blood analysis, and in the case of hypoglycemia, only 26.3% of patients with capillary blood analysis demonstrated clinical agreement .
Despite the fact that hypoperfusion during shock is recognized to be a factor in the underestimating of glucose levels with capillary sampling [19, 21, 22], it is not observed to be a significant concern in this investigation. Considering glucose concentrations, for instance, there is no significant difference between shocked and non-shocked patients using POCT or laboratory analysis. This conclusion is comparable to the outcomes of earlier investigations [12, 13, 23].
We relied on critically sick patients in this study because we wanted to verify the reliability of POCT under certain situations, such as shock. Under high conditions of pH, temperature, electrolyte abnormalities, and hypoglycemia, there are consequently few data points from which to draw conclusions regarding the dependability of specific analyzers.
Limitation of the study
The sample size was rather modest. This is a single-centre study. Blood was collected by various nurses. Bias induced by the design, manufacture, or use of a monitor.
In conclusion, Except for arterial blood glucose, the results of POCT and standard laboratory analysis of glucose concentrations in critically sick patients did not differ significantly in this investigation. Compared to laboratory blood analysis, the use of POCT is marginally accurate, with no difference between shocked and non-shocked patients.
Availability of data and materials
The data provided by the investigators was anonymized and processed. The following data, models, or code developed or used during the study are proprietary or private and may only be shared under certain conditions (e.g. anonymized data).
Point of care
Intensive care unit
King Fahd Military Medical Complex
Clement S, Braithwaite S, Magee MF, et al. Management of diabetes and hyperglycemia in hospitals. Diabetes Care. 2004;27(2):553–91. https://doi.org/10.2337/diacare.27.2.553.
Malmberg K, Norhammar A, Wedel H, Rydén L. Glycometabolic state at admission: an important risk marker for mortality in patients with diabetes mellitus and acute myocardial infarction treated conventionally. Circulation. 1999;99(20):2626–32. https://doi.org/10.1161/01.CIR.99.20.2626.
Van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in critically ill patients. N Engl J Med. 2001;345(19):1359–67. https://doi.org/10.1056/NEJMoa011300.
Van den Berghe G, Wilmer A, Milants I, et al. Intensive insulin therapy in surgical and medical intensive care units. Diabetes. 2006;55(11):3151–9. https://doi.org/10.2337/db06-0855.
Melmed S, Polonsky KS, Larsen PR, Kronenberg HM. Williams’ Textbook of Endocrinology. 12th edn. Saunders; 2011.
Griesdale DEG, de Souza RJ, van Dam RM, et al. Intensive insulin therapy with mortality in critically sick patients: a meta-analysis using data from the NICE-SUGAR study. CMAJ. 2009;180(8):821–7. https://doi.org/10.1503/cmaj.090206.
Krinsley JS, Grover A. Severe hypoglycemia in critically ill patients: risk factors and outcomes*. Crit Care Med. 2007;35(10):2262–7. https://doi.org/10.1097/01.CCM.0000282073.98414.4B.
AD. Standards of medical care of diabetes 2012. Diabetes Care. 2012;35:S11–63.
Dellinger RP, Levy MM, Carlet JM, et al. International recommendations for the management of severe sepsis and septic shock. Crit Care Med. 2008;36(1):296–327. https://doi.org/10.1097/01.CCM.0000298158.12101.41.
Heinemann Lutz. Quality of glucose measurement with blood glucose meters at the point-of-care relevance of interfering factors. Diabetes Technol Ther. 2010;12(11):847–57. https://doi.org/10.1089/dia.2010.0076.
Dungan K, Chapman J, Braithwaite SS, Buse J. Glucose measurement: confounding issues in setting inpatient management targets. Diabetes Care. 2007;30(2):403–9. https://doi.org/10.2337/dc06-1679.
Juneja D, Pandey R, Singh O. Comparison of arterial and capillary blood glucose monitoring in shock patients. Eur J Intern Med. 2011;22(3):241–4. https://doi.org/10.1016/j.ejim.2011.01.004.
Atkin SH. Fingerstick glucose determination in shock. Ann Intern Med. 1991;114(12):1020. https://doi.org/10.7326/0003-4819-114-12-1020.
Cook A, Laughlin D, Moore M, et al. Differences between point-of-care glucose meters and laboratory analysis in critically ill patients. Am J Crit Care. 2009;18(1):65–72. https://doi.org/10.4037/ajcc2009626.
Desachy A, Vuagnat AC, Ghazali AD, et al. Accuracy of bedside glucometry in critically ill patients: influence of clinical characteristics and perfusion index. Mayo Clin Proc. 2008;83(4):400–5. https://doi.org/10.4065/83.4.400.
Kanji S, Buffie J, Hutton B, et al. Reliability of glucose measurement at the point-of-care in critically sick adults*. Crit Care Med. 2005;33(12):2778–85. https://doi.org/10.1097/01.CCM.0000189939.10881.60.
Lonjaret L, Claverie V, Berard E, et al. In critically ill patients, the relative accuracy of arterial and capillary glucose meter values was evaluated. Diabetes Metab. 2012;38(3):230–5. https://doi.org/10.1016/j.diabet.2011.12.003.
Shearer A, Boehmer M, Closs M, et al. Comparison of point-of-care and laboratory values for glucose in critically ill patients. Am J Crit Care. 2009;18(3):224–30. https://doi.org/10.4037/ajcc2009448.
Critchell CD, Savarese V, Callahan A, Aboud C, Jabbour S, Marik P. Blood glucose measurement accuracy at the bedside in critically ill patients. Intensive Care Med. 2007;33(12):2079–84. https://doi.org/10.1007/s00134-007-0835-4.
Corstjens AM, Ligtenberg JJM, van der Horst ICC, et al. Accuracy and practicability of continuous and point-of-care blood glucose analysis in ICU patients with critical illness. Crit Care. 2006;10(5):R135. https://doi.org/10.1186/cc5048.
Krinsley JS. Effect of an intensive glucose management protocol on adult patients with critical illness mortality. Mayo Clin Proc. 2004;79(8):992–1000. https://doi.org/10.4065/79.8.99222.
Maser RE, Butler MA, Decherney GS. Utilization of arterial blood with glucose reflectance meters at the bedside in an intensive care unit. Crit Care Med. 1994;22(4):595–9. https://doi.org/10.1097/00003246-199404000-00014.
Pereira AJ, Corrêa TD, de Almeida FP, et al. A cross-sectional analysis of the accuracy of venous point-of-care glucose measurements in critically ill patients. PLoS ONE. 2015;10(6):e0129568. https://doi.org/10.1371/journal.pone.0129568.
Petersen JR, Graves DF, Tacker DH, Okorodudu AO, Mohammad AA, Cardenas VJ. Using arterial, capillary, and venous blood samples from ICU patients on a strict glycemic protocol, a comparison of POCT and central laboratory blood glucose findings was performed. Clin Chim Acta. 2008;396(1–2):10–3. https://doi.org/10.1016/j.cca.
Boyd R. Capillary versus venous blood glucose estimates at the bedside. Emerg Med J. 2005;22(3):177–9. https://doi.org/10.1136/emj.2003.011619.
Clarke WL, Anderson S, Farhy L, et al. Using continuous glucose-error grid analysis to evaluate the clinical accuracy of two continuous glucose sensors. Diabetes Care. 2005;28(10):2412–7. https://doi.org/10.2337/diacare.28.10.241227.
Goldberg PA, Siegel MD, Russell RR, et al. Intensive care unit experience with the continuous glucose monitoring system®. Diabetes Technol Ther. 2004;6(3):339–47. https://doi.org/10.1089/152091504774198034.
There are no acknowledgements.
No funding was received for this research.
Ethical approval and consent to participate
The King Fahd Military Medical Complex, Dhahran, Saudi Arabia, Institutional Review Board (IRB) approved this prospective study. The authorization number is REC-2018-0005 DATE: 01/30/2019. The approval was granted for the duration of the study, which was 1 year. Each participant/next of kin enrolled in this study was informed and a complete a written consent form was completed.
Consent for publication
The signed written consent collected from each participant / next of kin granted permission to participate in this study as well as permission to publish the results anonymously. In addition, the IRB approval letter authorized the publication of the data because they complied with the “Ethical Research Practice Guidelines“.
All authors report that there are no conflicts of interest.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Alshaer, A., Badgheish, B.A., Alsadah, Z.H. et al. Comparing the accuracy of point-of-care with laboratory (capillary, venous, and arterial) blood glucose levels in critically ill patients with and without shock. BMC Res Notes 15, 372 (2022). https://doi.org/10.1186/s13104-022-06256-0
- Blood glucose
- Intensive care unit
- Patients in shock