Proof of Concept Study to Assess the Influence of Oxygen Partial Pressure in Capillary Blood on SMBG Measurements

Abstract

Background:

Measurement results provided by blood glucose monitoring systems (BGMS) can be affected by various influencing factors. For some BGMS using glucose oxidase (GOx)-based test strips, one of these factors is the oxygen partial pressure (pO₂) of the applied blood sample. Because assessing the potential influence of pO₂ when measuring capillary blood samples is not straight-forward, we performed a proof of concept study.

Method:

Influence of pO₂ was investigated for two GOx-based BGMS (BGMS A and B). Measurement results of the GOx-based BGMS were compared with measurement results from a pO₂-independent BGMS (BGMS C). A total of 119 samples from 60 subjects were measured, twice with BGMS C, then 6 times each with BGMS A and BGMS B or vice versa, and again twice with BGMS C. Immediately afterward, pO₂ was determined. Linear regression analysis based on relative differences between results from BGMS A or BGMS B and results from BGMS C was performed to estimate the degree of pO₂ influence.

Results:

The relative bias between the lowest and highest pO₂ values differed by 14.3% for BGMS A, indicating a pO₂ influence that might be clinically relevant, and by 9.7% for BGMS B, indicating that pO₂ influence may be too small to be reliably detected because of the BGMS’ imprecision.

Conclusions:

This proof of concept study showed that with the procedures used, a potentially clinically relevant influence of pO₂ in capillary blood samples on GOx-based BGMS could be detected. Further larger-scale studies are needed to verify this influence.

Keywords

blood glucose monitoring system glucose oxidase partial pressure of oxygen self-monitoring of blood glucose

Self-monitoring of blood glucose (SMBG) is of continuing relevance for people with diabetes. As therapeutic decisions regarding insulin dosing and carbohydrate intake are often made based on measured glucose concentrations, the analytical performance of the blood glucose monitoring systems (BGMS) has an impact on the therapeutic outcome.

BGMS measurements are based on chemical reactions which, depending on the respective composition of enzymes, mediators and further components, can be influenced by various factors, for example, interfering substances present in the blood sample. Whereas glucose dehydrogenase (GDH), one of the two enzymes commonly utilized in BGMS’ reagent systems, is not affected by the partial pressure of oxygen (pO₂) present in the applied blood sample, glucose oxidase (GOx) is known to be prone to oxygen interference, that is, elevated or decreased pO₂ levels may cause measurement bias.^1-4 The oxygen dependence of GOx-based BGMS can be minimized combining the enzyme with suitable mediators.^1,3 Nevertheless part of the GOx-based BGMS available on the market is evidently influenced by pO₂.^5-9 ISO 15197:2013 stipulates that an interference effect exceeding 10 mg/dl or 10% at glucose concentrations <100 mg/dl or ≥100 mg/dl, respectively, shall be described in the instructions for use.¹⁰ In previous studies, we investigated the influence of different pO₂ levels on BG measurements with GOx-based BGMS in laboratory settings using venous blood.^8,9 Part of the tested BGMS showed considerable measurement bias at decreased or elevated pO₂, irrespective of being labeled as pO₂-dependent or not. The majority of BGMS labeled to be influenced by pO₂ only refer to elevated oxygen levels, for example, as expected in patients undergoing oxygen therapy.⁷ However, the results of the studies mentioned above indicate that the effect of decreased pO₂ levels might even be more relevant than the effect observed with elevated levels. Whereas high oxygen levels can lead to an underestimation of glucose concentrations, affected BGMS may overestimate glucose concentrations at decreased pO₂. Low pO₂ levels can be expected, for example, in elderly people or in patients with respiratory diseases but also during long distance flights or when staying at high altitude.^11-18

In order to assess whether the effects observed in the laboratory studies using venous blood are relevant also for SMBG measurements, pO₂ influence has to be investigated in capillary blood samples measured directly from the skin puncture site. Such studies, however, have to take into account a number of issues, like the limited volume of capillary blood that may be obtained from one skin puncture or the availability of subjects with extreme pO₂. We therefore set up a proof of concept study with the objective to investigate the possible interfering effect of pO₂ on GOx-based BGMS in general, irrespective of the effective magnitude of measured pO₂ values.

Methods

The study was performed at the Institut für Diabetes-Technologie, Forschungs- und Entwicklungsgesellschaft mbH an der Universität Ulm (IDT) in Germany in March 2018 in compliance with the German Medical Devices Act and with requirements of Good Clinical Practice (DIN EN ISO 14155:2012). The study was approved by the responsible Ethics Committee and exempted from approval by the German Federal Institute for Drugs and Medical Devices. The study was registered at DRKS—Deutsches Register Klinischer Studien (DRKS-ID: DRKS00014229), an approved Primary Register in the World Health Organization’s network of clinical trial registries.

Study Population

Sixty subjects were included in the evaluation (28 men, 32 women; mean age, 61.6 [range 23-82] years; 15 people with diabetes type 1, 29 with diabetes type 2, 16 people without diabetes). In order to cover a reasonably wide pO₂ range, subjects with respiratory diseases or elderly people were preferred because of suspected lower capillary pO₂ levels. All participants signed informed consent forms prior to the study procedures. A physician reviewed the subjects’ anamnesis and medication and checked for interfering substances indicated in the BGMS’ instructions for use. The subjects’ hematocrit values (35.0-51.0%) were verified to be within the ranges given for each BGMS.

Blood Glucose Monitoring Systems

In this study, influence of pO₂ on measurement results was investigated for two GOx-based BGMS (BGMS A and B) that had shown marked pO₂ influence in a previous study⁹ with one reagent system lot each. These two BGMS are only labeled to be sensitive to increased blood oxygen content (eg, during oxygen therapy). A GDH-based BGMS (BMGS C) was included in the investigation as pO₂-independent comparison BGMS.

The BGMS were purchased from pharmacies. All systems were stored, used and maintained as stipulated in the respective manufacturer’s instructions. Control measurements were performed on each study day to ensure the proper function of each test meter and the integrity of each test strip vial used on that day.

Study Procedures

The evaluation was performed by trained study personnel in a laboratory setting in which the room temperature and humidity were checked to be within the range indicated in the manufacturer’s labelling (temperature: 21.0-23.9°C, humidity: 33.6-49.6%).

All measurements were performed on capillary blood obtained from the subject’s finger tips by skin puncture.

Measurement Procedure for Each Subject

Study personnel punctured a finger from the subject’s left hand and performed two measurements using two meters of the GDH-based comparison BGMS C, followed by six measurements with the GOx-based BGMS A (using six meters), six measurements with the GOx-based BGMS B (using six meters) and two measurements again with two meters of BGMS C. Immediately after measurements with the BGMS, one sample was taken for the determination of pO₂. The pO₂ was measured using a blood gas analyzer (OPTI™ CCA-TS Analyzer, OPTI Medical Systems, Inc, Roswell, GA, USA).

Subsequently two samples were collected for glucose measurements with a hexokinase-based laboratory analyzer (Cobas Integra® 400 plus, Roche Instrument Center, Rotkreuz, Switzerland).

Samples for all blood glucose (BG) measurements with the BGMS and for the pO₂ measurements had to be obtained from a single skin puncture.

For each subject, this measurement procedure was repeated using a finger from the right hand. The sequence of the GOx-based BGMS was changed between left and right hands of each subject and from subject to subject.

Data Analysis

As the results of left and right hands of one subject did not show systematic differences, data sets from both hands were included in the evaluation as independent data sets if data from both hands were found to be valid. Thus, in total 119 data sets were included in the evaluation. Fifty-nine subjects provided data sets from both hands. One subject provided only one valid data set; data from the left hand had to be excluded because more than one finger prick was required for the measurements with the BGMS and blood gas analyzer.

Data analysis was performed in mg/dl. For BGMS displaying results in mmol/l, values were converted (1 mmol/l = 18.02 mg/dl).

The included samples had glucose concentrations between 84.6 and 340.1 mg/dl, measured pO₂ values covered a range of 52.0-85.0 mmHg (Table 1).

Table 1.

BG Concentrations (Determined With Laboratory Analyzer) and pO₂ Values Measured in 119 Included Data Sets.

Parameter	Mean	SD	Minimum	1st quartile	Median	3rd quartile	Maximum
pO₂ (mmHg)	67.1	7.5	52.0	61.5	67.0	72.5	85.0
Glucose (mg/dl)	139.2	48.6	84.6	105.8	127.7	156.4	340.1

For the GOx-based BGMS A and B, mean values of the 6 measurements were calculated, whereas for the comparison BGMS C mean values were calculated from 4 measurements (2 measurements performed before and after the measurements with BGMS A and B). Relative biases were determined between mean values obtained with BGMS A and C and between mean values obtained with B and C to estimate pO₂ influence in BGMS A and BGMS B, respectively. As a linear relationship was observed between these biases and the measured pO₂ levels, linear regression equations were applied in order to estimate the extent of pO₂ influence calculating the absolute value of the difference in bias between the samples with the most extreme pO₂ values.

Statistical Considerations

The relevance of the pO₂ influence was assessed based on the following statistical considerations: Regarding other influencing factors, ISO 15197:2013¹⁰ stipulates that influence quantities should be labeled in the instruction for use if the induced change in glucose measurement results exceeds 10 mg/dl or 10% (for BG concentration <100 mg/dl or ≥100 mg/dl). With respect to hematocrit, ISO 15197:2013¹⁰ stipulates that hematocrit influence should be labeled if hematocrit’s effect on glucose measurement results exceeds 10 mg/dl or 10% (for BG concentration <100 mg/dl or ≥100 mg/dl) with respect to a specific midlevel hematocrit (substitute for an average hematocrit).

Therefore, if the systematic measurement difference between the most extreme pO₂ values in the study exceeds 20%, at least one pO₂ value can be found for which the systematic measurement differences between glucose measurements at that pO₂ value and glucose measurements at both ends of the pO₂ range will be at least 10% at the same time (corresponding to hematocrit criteria). In addition, the BGMS’ analytical imprecision has to be accounted for. In this study, the 97.5% quantiles for the coefficient of variation calculated from 6 (systems A and B) or 2 (system C) replicate measurements were found to be 5.69% for system A, 6.48% for system B, and 3.99% for system C. Thus, a difference in bias between the most extreme pO₂ values exceeding 10%, is highly likely not to be found due to random error. In contrast, a difference in bias below 10% would indicate an effect of pO₂ influence too small to be reliably detected.

Results

In order to estimate the relevance of pO₂ influence for GOx-based BGMS, the absolute difference in relative bias (GOx-based BGMS versus GDH-based comparison BGMS) between the most extreme pO₂ values obtained in the study was calculated using linear regression.

For BGMS A the relative bias at the lowest and highest pO₂ value as determined from the regression equation was 11.7% and −2.6%, respectively (Figure 1A). According to the considerations described above, the resulting absolute difference in relative bias of 14.3% indicated a detectable pO₂ influence that might be clinically relevant. Relative biases at the lowest and highest pO² values calculated from the regression equation of BGMS B were 5.4% and −4.3%, respectively (Figure 1B). The difference in relative bias was 9.7%, indicating a pO₂ influence, which is possibly too small to be reliably detected. Although it is close to the borderline of 10% difference in biases set for this study, the clinical relevance could not be finally assessed.

Figure 1.

Relative biases between BGMS A (A) and B (B) and GDH-based BGMS C plotted versus pO₂. The regression line is represented as blue line. The mean bias is indicated at the lowest and highest pO₂ values, respectively.

In order to visualize a possible effect of varying glucose concentrations on pO₂ influences, samples were categorized by 3 ranges of different pO₂ levels (<65 mmHg, 65-80 mmHg, >80 mmHg) and represented with different colors in difference plots (Figure 2). For both GOx-based BGMS low glucose concentration samples seem to show stronger pO₂ influence in terms of a larger bias. However, more data are required, especially at low and high glucose concentrations, to suitably assess this possible dependence of the pO₂ influence from glucose concentrations.

Figure 2.

Difference plots showing biases between mean glucose concentrations measured with GOx-based BGMS A (A) and B (B) and pO₂-independent comparison BGMS C plotted versus glucose concentrations measured with BGMS C. Data points are colored depending on the category in which the sample’s pO₂ value fell (<65 mmHg, 65-80 mmHg, >80 mmHg). The black lines indicating differences from the comparison measurements of ±15 mg/dl at glucose concentrations <100 mg/dl and ±15% at glucose concentrations ≥100 mg/dl are included for visual guidance.

Discussion

In this proof of concept study, we investigated the possible influence of pO₂ in capillary blood samples on measurements with GOx-based BGMS. Two BGMS were selected for the study that had shown marked pO₂ influence in a previous study using venous blood.⁹

The pO₂ influence was estimated comparing the differences in relative bias between the two GOx-based BGMS and a pO₂-independent comparison BGMS at the lowest and highest pO₂ levels observed during the study. The resulting differences were 14.3% for BGMS A and 9.7% for BGMS B. According to the statistical considerations made for the study, the result for BGMS A indicates a detectable pO₂ influence which might be clinically relevant. In BGMS B, pO₂ influence is close to the borderline of 10% difference in biases set for this study and is therefore possibly too small to be reliably detected.

Dependence of pO₂ influence from glucose concentrations may be suspected, because low glucose concentration samples seemingly showed slightly larger pO₂ influence for both GOx-based BGMS in this study.

Even in populations without acute serious diseases, capillary blood samples can cover a broad range of pO₂ values.¹⁹ In addition, certain groups of patients may commonly show more extreme pO₂ levels. Elevated pO₂ may occur, for example, in patients undergoing oxygen therapy, whereas decreased pO₂ values can be expected, for example, in patients with respiratory diseases, in elderly people or also at high altitude and during long distance flights.^7,11-18 Considering that chronic obstructive pulmonary disease (COPD) is described as being associated with type 2 diabetes and considering that the still increasing prevalence of diabetes mainly affects elderly people, BG measurements of a significant number of patients using SMBG might be affected by measurement deviations caused by decreased pO₂ when using oxygen-sensitive BGMS.^20,21 When measuring blood samples with low pO₂, these BGMS tend to overestimate BG concentration and therefore patients might be at risk of not detecting hypoglycemia, especially if pO₂ influence was more pronounced at low glucose levels, as suspected from results obtained in this study.

This investigation was designed as a proof of concept study with a limited number of subjects to assess whether the methodical approach was suitable for investigation of pO₂ influence on BGMS. Although this proof of concept study indicated that pO₂ influence might be relevant for oxygen sensitive GOx-based BGMS, the study entails several limitations that have to be taken into account when interpreting this study’s results. On the one hand the subjects’ BG concentrations did neither adequately cover the BGMS’ measuring ranges nor the low / hypoglycemic range which could be particularly relevant with affected BGMS overestimating BG at decreased pO₂.^8,9 As the study population consisted of subjects without severe acute or chronic diseases (besides diabetes), the range of pO₂ values obtained during the study was probably tighter than the range typically found among users of BGMS.

Another limitation that should be kept in mind is that the study setting only allows the investigation of a potential influence of pO₂ in general. The relevant pO₂ value in the blood sample for a patient’s SMBG measurement cannot be assessed due to methodical reasons, as the blood volume required for pO₂ determination in this study was substantially larger than the volume needed for SMBG (≥60 µl compared to usually <1 µl). The pO₂ was obtained after the measurements with the BGMS to minimize the blood volume removed before the individual BGMS measurements, and, therefore, to perform BGMS measurements in conditions as similar as possible to patient SMBG measurements. In addition, wiping blood off the fingertip between measurements with different meters of a BGMS was avoided (whenever possible). For the same reason, samples measured with the GDH-based BGMS immediately before and after the measurements with the GOx-based BGMS were used for comparison measurements to enable accounting for glucose changes, and the samples for the laboratory analyzer, which require a comparably large blood volume, were taken at the end of the measurement series. Therefore, as the study focused on the relevance of low pO₂ values, the pO₂ used in this study is expected to be a conservative estimate for the pO₂ relevant for each of the BGMS measurements included in the evaluation.

Finally, potential interfering effects of decreased or increased pO₂ levels do not relate to all GOx-based BGMS, but are limited to the subgroup of oxygen-sensitive systems. In addition, it has to be mentioned that numerous other factors (eg, interfering drugs, hematocrit, ambient temperature) can affect SMBG measurements, irrespective of the BGMS being based on GOx or GDH.

Besides indicating a possible relevance of pO₂ influence in SMBG measurements, this study confirmed adequacy of the applied procedures for an assessment of pO₂ influences on BGMS using capillary blood samples in a straightforward setting. However, taking into account the possible clinical impact of pO₂ influence on SMBG, a further investigation of the indicated relevance of pO₂ influences on oxygen-sensitive BGMS, including the objective of dependence of pO₂ influence from glucose concentrations, should be considered. For this purpose, larger scale studies are needed to generate valid conclusions. These evaluations should be performed on a sufficiently large number of subjects and focus on representative distributions of glucose concentration and pO₂ values. Because of the suspected relevance of decreased pO₂ levels, preferably including subjects with respiratory diseases, for example, COPD should be considered. In case the relevance of the suspected pO₂ influence on oxygen-sensitive GOx-based BGMS will be verified, further studies might also address the impact of varying pO₂ on continuous glucose monitoring (CGM) systems if their sensors are based on GOx.²²

Conclusion

The results of this proof of concept study indicate a detectable influence of pO₂ on oxygen-sensitive GOx-based BGMS. As this interference effect might be clinically relevant, additional studies should be performed in order to further characterize the pO₂ influence. Relevant information regarding pO₂ influence, as well as information about all other possibly interfering factors, should be included in the respective BGMS’ instructions for use, to enable patients and health care providers to choose adequate BGMS for specific health conditions.

Footnotes

Acknowledgements

The authors would like to thank the subjects who participated in the study as well as Manuela Link, MD, Eva Zschornack, MD, Martina Tesar, Natalie Neuburger, Tuba Alkan, and other IDT staff who contributed to the conduct of the study. The study was registered at DRKS—Deutsches Register Klinischer Studien (DRKS-ID: DRKS00014229).

Abbreviations

BG, blood glucose; BGMS, blood glucose monitoring system; CGM, continuous glucose monitoring; COPD, chronic obstructive pulmonary disease; GDH, glucose dehydrogenase; GOx, glucose oxidase; ISO, International Organization for Standardization; pO₂, partial pressure of oxygen; SD, standard deviation; SMBG, self-monitoring of blood glucose.

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: AB, SP, NJ, CL, and CH are employees of the IDT. RH is employee of Roche Diabetes Care GmbH, Mannheim, Germany. GF is general manager of the IDT (Institut für Diabetes-Technologie Forschungs- und Entwicklungsgesellschaft mbH an der Universität Ulm, Ulm, Germany), which carries out clinical studies on the evaluation of BG meters and medical devices for diabetes therapy on its own initiative and on behalf of various companies. GF/IDT have received speakers’ honoraria or consulting fees from Abbott, Ascensia, Bayer, Dexcom, LifeScan, Menarini Diagnostics, Metronom Health, Novo Nordisk, Roche, Sanofi, Sensile and Ypsomed.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The performance of this investigator-initiated study was supported by Roche Diabetes Care GmbH with a grant.

ORCID iDs

Stefan Pleus

Guido Freckmann

References

Hönes

Müller

Surridge

The technology behind glucose meters: test strips. Diabetes Technol Ther. 2008;10:10-26.

Wang

Electrochemical glucose biosensors. Chem Rev. 2008;108:814-825.

Yoo

Lee

SY.

Glucose biosensors: an overview of use in clinical practice. Sensors (Basel). 2010;10:4558-4576.

Chaubey

Malhotra

BD.

Mediated biosensors. Biosens Bioelectron. 2002;17:441-456.

Tang

Louie

Payes

Chang

Kost

GJ.

Oxygen effects on glucose measurements with a reference analyzer and three handheld meters. Diabetes Technol Ther. 2000;2:349-362.

Louie

Tang

Sutton

Lee

Kost

GJ.

Point-of-care glucose testing: effects of critical care variables, influence of reference instruments, and a modular glucose meter design. Arch Pathol Lab Med. 2000;124:257-266.

Tang

Louie

Lee

Miller

Kost

GJ.

Oxygen effects on glucose meter measurements with glucose dehydrogenase- and oxidase-based test strips for point-of-care testing. Crit Care Med. 2001;29:1062-1070.

Baumstark

Schmid

Pleus

Haug

Freckmann

Influence of partial pressure of oxygen in blood samples on measurement performance in glucose-oxidase-based systems for self-monitoring of blood glucose. J Diabetes Sci Technol. 2013;7:1513-1521.

Schmid

Baumstark

Pleus

Haug

Tesar

Freckmann

Impact of partial pressure of oxygen in blood samples on the performance of systems for self-monitoring of blood glucose. Diabetes Technol Ther. 2014;16:156-165.

10.

International Organization for Standardization. In vitro diagnostic test systems—requirements for blood-glucose monitoring systems for self-testing in managing diabetes mellitus. ISO 15197:2013.

11.

Blom

Mulder

Verweij

Arterial oxygen tension and saturation in hospital patients: effect of age and activity. BMJ. 1988;297:720-721.

12.

Garcia

Borderias

Casanova

, et al. Air travel and respiratory diseases. Arch Bronconeumol. 2007;43:101-125.

13.

O’Connor

Barry

Jahangir

Finn

Buckley

El-Gammal

Comparison of arterial and venous blood gases and the effects of analysis delay and air contamination on arterial samples in patients with chronic obstructive pulmonary disease and healthy controls. Respiration. 2011;81:18-25.

14.

Coker

Partridge

MR.

What happens to patients with respiratory disease when they fly?

Thorax. 2004;59:919-920.

15.

Seccombe

Kelly

Wong

Rogers

Lim

Peters

MJ.

Effect of simulated commercial flight on oxygenation in patients with interstitial lung disease and chronic obstructive pulmonary disease. Thorax. 2004;59:966-970.

16.

Mays

EE.

An arterial blood gas diagram for clinical use. Chest. 1973;63:793-800.

17.

Ogun

Bayir

Kayis

Koylu

Prediction of arterial blood gas values from venous blood gas values in patients with acute exacerbation of chronic obstructive pulmonary disease. Tohoku J Exp Med. 2006;210:285-290.

18.

deMol

Krabbe

de Vries

, et al. Accuracy of handheld blood glucose meters at high altitude. PLoS One. 2010;5(11):e15485.

19.

Freckmann

Schmid

Baumstark

Pleus

Link

Haug

Partial pressure of oxygen in capillary blood samples from the fingertip. J Diabetes Sci Technol. 2013;7:1648-1649.

20.

Rana

Mittleman

Sheikh

, et al. Chronic obstructive pulmonary disease, asthma, and risk of type 2 diabetes in women. Diabetes Care. 2004;27:2478-2484.

21.

Mirrakhimov

AE.

Chronic obstructive pulmonary disease and glucose metabolism: a bitter sweet symphony. Cardiovasc Diabetol. 2012;11:132.

22.

Cappon

Acciaroli

Vettoretti

Facchinetti

Sparacino

Wearable continuous glucose monitoring sensors: a revolution in diabetes treatment. Electronics. 2017;6:1-16.