📄 medicine

Agreement and bias of three capillary blood glucose meters compared with venous plasma glucose across clinically relevant glycemic ranges in primary care: a multicenter cross-sectional study

This multicenter study found that while three common capillary blood glucose meters show high correlation with venous plasma glucose, they exhibit significant systematic biases—particularly at lower glucose levels—highlighting that high correlation does not ensure clinical interchangeability and underscoring the need for consistent device use in primary care.

Original authors: Raúl Porras-Benjumea, Sandra Arco-Rodríguez, Jaume Català-Coll, Silvia López-Marchena, Lorena Martín-Callejón, Susanna Martínez-Gálvez, Esther Moral-Ramírez, María Jesús López-Piqueres, Maite Álvarez-

Published 2026-08-10

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Raúl Porras-Benjumea, Sandra Arco-Rodríguez, Jaume Català-Coll, Silvia López-Marchena, Lorena Martín-Callejón, Susanna Martínez-Gálvez, Esther Moral-Ramírez, María Jesús López-Piqueres, Maite Álvarez-Fernández, Àngels Fumàs-Coma, Xavier Bayona-Huguet

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to navigate a city using a map. In the world of diabetes, that map is your blood sugar level. For millions of people, knowing exactly where they stand on this map is the difference between feeling great and feeling very sick. To read this map, they use tiny, handheld devices called blood glucose meters. You prick your finger, put a drop of blood on a test strip, and the machine gives you a number. It's a bit like checking the fuel gauge in a car; if the gauge says you have plenty of gas but you actually don't, you might get stranded. If it says you're empty when you're full, you might panic and waste money refueling unnecessarily.

But here is the tricky part: these meters aren't perfect. They are like different brands of thermometers. If you stick three different thermometers in the same cup of hot water, they might all agree that the water is "hot," but one might say 100 degrees, another 102, and a third 98. In medicine, those small differences can be huge. A reading that is just a little too high might make a doctor think a patient is safe, when they are actually in danger of a low blood sugar crash (hypoglycemia). A reading that is too low might make someone eat too much sugar, sending their levels soaring. So, the big question for doctors is: Do these different meters tell the same story, or are they telling different versions of the truth?

This is exactly what a team of researchers in Spain wanted to find out. They didn't just look at the numbers on a screen; they went into real doctor's offices to see how three popular blood sugar meters behaved when compared to the "gold standard" lab test. They wanted to know if these devices could be used interchangeably or if switching from one brand to another was like switching from a reliable map to a slightly distorted one.

The Great Meter Showdown

The researchers set up a real-world experiment across seven different primary care centers. They invited patients with diabetes who were already there for their routine blood work. While the patients were waiting, the medical team took a tiny drop of blood from their finger and tested it on three different meters at the same time: the Contour, the GlucoMen, and the FreeStyle. At the exact same moment, they also took a sample of blood from the patient's arm (venous blood) and sent it to the lab to get the "true" reference number.

Think of it like a taste test. The lab result is the chef's perfect recipe. The three meters are three different tasters trying to guess the exact amount of salt in the soup. The researchers wanted to see if the tasters agreed with the chef and if they agreed with each other.

The Results: High Scores, But Different Flavors

When the researchers crunched the numbers, they found something fascinating. First, all three meters were very good at tracking changes. If the patient's blood sugar went up, all three meters went up. If it went down, they all went down. Statistically, they were all "best friends," with a correlation score between 0.953 and 0.970 (where 1.0 is perfect agreement). It's like three friends who always agree on the plot of a movie; they all remember the story happening in the same order.

However, the researchers discovered that agreement on the story doesn't mean agreement on the details. This is where the meters started to drift apart.

The FreeStyle Meter: This device was the most accurate "taster." It was very close to the lab's perfect recipe. On average, it was only off by about -2.67 mg/dL. In the tricky "low sugar" zone (below 100 mg/dL), it was incredibly reliable, getting the right answer 96.6% of the time.
The Contour Meter: This one had a habit of being overly optimistic. It consistently reported numbers that were higher than the lab result. On average, it was +10.73 mg/dL too high. In the low sugar zone, it was even worse, overestimating by +14.10 mg/dL and only getting the answer right 55.2% of the time.
The GlucoMen Meter: This device also tended to be a bit too optimistic, overestimating by +6.85 mg/dL on average.

The Danger Zone: When "Close Enough" Isn't Good Enough

The most important discovery happened in the low blood sugar range (below 100 mg/dL). This is the danger zone where a patient might be at risk of hypoglycemia.

Imagine a patient whose blood sugar is actually 80 mg/dL (which is getting low).

The FreeStyle meter might read it as 83 mg/dL. Close enough to be safe.
The Contour meter, however, might read it as 94 mg/dL (80 + 14).

That difference of 14 points is huge. A reading of 94 might make the patient think, "Phew, I'm fine," and skip a snack. But their actual level is 80, and they could be sliding into a dangerous crash. The study found that in this low range, the Contour meter only met the safety tolerance criteria 55.2% of the time, whereas the FreeStyle met them 96.6% of the time.

The researchers also compared the meters against each other. The difference between the Contour and the FreeStyle was the widest, with an average gap of 13.40 mg/dL. This means if a patient switched from a FreeStyle to a Contour, their numbers could suddenly jump up by more than 10 points, even though their actual blood sugar hadn't changed.

What This Means for Patients

The study concludes that just because two devices are highly correlated (they move together), it doesn't mean they are interchangeable. You can't just swap one meter for another and expect the numbers to stay the same.

The researchers suggest that patients should stick to the same brand of meter for as long as possible. If a doctor or insurance company forces a switch to a new device, the patient shouldn't just start using it immediately. Instead, they should use both the old and the new meter side-by-side for a while to see how the numbers compare. This "parallel testing" helps the patient understand the new device's "personality" and prevents dangerous mistakes in treatment.

In short, while all three meters are generally good at telling the story of your blood sugar, they tell it with different accents. In the high-stakes world of diabetes, especially when sugar levels are low, those accents matter. The FreeStyle meter came out as the most consistent narrator in this study, while the Contour and GlucoMen tended to tell a slightly more optimistic (and potentially risky) version of the story. The key takeaway is that consistency is king: stick with one device, and if you must change, do it with caution.

Technical Summary: Agreement and Bias of Three Capillary Blood Glucose Meters in Primary Care

Problem Statement
Despite the increasing adoption of continuous glucose monitoring (CGM) systems, capillary blood glucose meters (BGMs) remain a fundamental tool for diabetes management in primary care, particularly for patients not using CGM or for confirming sensor readings during rapid glycemic changes. While international standards (e.g., ISO 15197:2013) define analytical accuracy requirements, there is a critical gap in understanding how these devices perform under routine, real-world clinical conditions. A specific concern is that high statistical correlation between devices does not necessarily imply clinical interchangeability. Systematic biases, particularly at lower glucose concentrations, can lead to misinterpretation of hypoglycemia risk and inappropriate therapeutic decisions, such as incorrect insulin dosing. Previous evaluations often occur under controlled conditions that may not reflect the variability of strip lots, hematocrit levels, and environmental factors found in everyday practice.

Methodology
The authors conducted a multicenter, cross-sectional, observational study across seven primary care centers within Badalona Serveis Assistencials (BSA) in Catalonia, Spain.

Participants: 142 measurement episodes were recorded from patients with diabetes attending routine blood testing appointments. The final analytical sample comprised 131 episodes with complete data from all methods.
Devices Evaluated: Three commercially available capillary blood glucose meters were tested: Contour® (Ascensia Diabetes Care), GlucoMen® (A. Menarini Diagnostics), and FreeStyle® (Abbott Diabetes Care).
Comparator: Venous plasma glucose (VPG) obtained from the same clinical encounter and analyzed by the reference laboratory served as the clinical comparator.
Procedure: Capillary measurements were taken consecutively by healthcare professionals following standard primary care procedures.
Statistical Analysis: Agreement was assessed using:
- Pearson's correlation coefficient ( $r$ ) for linear association.
- Bland–Altman analysis to determine mean bias and 95% limits of agreement (LoA).
- Mean Absolute Difference (MAD).
- An ISO-inspired tolerance criterion: $\pm 15$ mg/dL for VPG $< 100$ mg/dL and $\pm 15\%$ for VPG $\ge 100$ mg/dL.
Stratification: Analyses were stratified into three glycemic ranges: $< 100$ mg/dL, $100–180$ mg/dL, and $\ge 180$ mg/dL to evaluate performance across clinically relevant thresholds.

Key Results

Correlation vs. Agreement: All devices demonstrated high Pearson correlation coefficients with venous plasma glucose ( $r = 0.953–0.970$ ) and with each other ( $r = 0.945–0.959$ ). However, high correlation did not translate to clinical interchangeability due to significant systematic biases.
Overall Performance:
- FreeStyle® showed the closest agreement with VPG, with a mean bias of $-2.67$ mg/dL, a MAD of $8.08$ mg/dL, and $96.9\%$ of measurements falling within the predefined tolerance limits.
- Contour® systematically overestimated glucose values with a mean bias of $+10.73$ mg/dL and $80.2\%$ within tolerance.
- GlucoMen® also overestimated values with a mean bias of $+6.85$ mg/dL and $82.4\%$ within tolerance.
- The widest discrepancy was between Contour® and FreeStyle®, with a mean bias of $+13.40$ mg/dL and only $71.0\%$ of measurements within tolerance.
Performance by Glycemic Range:
- Hypoglycemia Range ( $< 100$ mg/dL): Discrepancies were most pronounced here. FreeStyle® maintained a low bias ( $+3.34$ mg/dL) and $96.6\%$ tolerance compliance. In contrast, Contour® showed a substantial positive bias of $+14.10$ mg/dL, with only $55.2\%$ of measurements within tolerance. GlucoMen® also overestimated by $+10.83$ mg/dL ( $69.0\%$ within tolerance).
- Intermediate ( $100–180$ mg/dL) and High ( $\ge 180$ mg/dL) Ranges: FreeStyle® maintained low bias and high tolerance compliance. Contour® and GlucoMen® continued to show positive biases, though tolerance compliance improved in the highest range.

Key Contributions

Real-World Evidence: The study provides data on device performance under routine primary care conditions rather than highly controlled laboratory settings, capturing the variability inherent in clinical practice.
Distinction Between Correlation and Interchangeability: The research empirically demonstrates that strong statistical correlation between BGMs and reference methods (or between devices) does not guarantee that the devices can be used interchangeably for clinical decision-making.
Stratified Analysis: By stratifying results by glycemic range, the study highlights that accuracy issues are not uniform; they are most critical in the hypoglycemia-prone range ( $< 100$ mg/dL), where systematic overestimation by certain devices could mask low glucose levels.
Device-Specific Performance: The study identifies specific performance profiles for three common devices, noting that FreeStyle® outperformed the others in terms of agreement with venous plasma glucose in this specific cohort.

Significance and Claims
The paper claims that the findings have direct implications for patient safety and therapeutic management in primary care. The authors assert that:

Clinical Caution is Required: Systematic biases, particularly the overestimation of glucose by Contour® and GlucoMen® in the lower range, may lead to delayed recognition of hypoglycemia and inappropriate therapeutic adjustments.
Consistency is Preferable: Given the lack of interchangeability, healthcare providers should encourage the consistent use of the same device for individual patients.
Transition Protocols: When switching devices is necessary (due to procurement or preference), a period of parallel measurements should be implemented to facilitate interpretation and mitigate the risk of clinically significant discrepancies.
Evaluation Metrics: The study reinforces the need to evaluate blood glucose meters using agreement analyses (Bland–Altman) and stratified performance metrics, rather than relying solely on correlation coefficients or regulatory compliance in isolation.

The authors acknowledge limitations, including the use of venous plasma glucose as a comparator rather than an ISO-traceable reference method, the lack of systematic recording of confounding factors (e.g., hematocrit, strip lot), and the opportunistic recruitment from a single healthcare organization. Consequently, the study positions itself as providing evidence for real-world practice rather than a formal regulatory validation of device accuracy.

The Great Meter Showdown

The Results: High Scores, But Different Flavors

The Danger Zone: When "Close Enough" Isn't Good Enough

What This Means for Patients

Technical Summary: Agreement and Bias of Three Capillary Blood Glucose Meters in Primary Care

More like this