# Comparison of glucose concentrations in simultaneously collected plasma and serum samples from outpatients in a routine laboratory setting

## TL;DR

This study compares glucose levels in plasma and serum samples from outpatients to see if they lead to different diabetes diagnoses.

## Contribution

The study provides empirical evidence on glucose concentration differences in plasma and serum samples under routine outpatient conditions.

## Key findings

- Glucose differences between serum and plasma are minimal above 100 mg/dL but increase below this threshold.
- Diabetes classifications based on plasma and serum glucose agree in 97.8% of cases using a 126 mg/dL cutoff.
- Neither serum nor non-acidified plasma fully inhibits glycolysis, making both suboptimal for definitive diagnosis.

## Abstract

Although international and national German guidelines recommend measuring glucose in venous plasma, serum remains the most commonly used sample type for glucose measurement in routine outpatient settings in Germany. The aim of the study was to compare glucose concentrations in simultaneously collected serum and plasma samples from outpatients under routine conditions, and to assess whether any observed differences could lead to divergent diagnostic outcomes for diabetes mellitus.

Between 03/01/2022 and 20/12/2024, a total of 27,864 simultaneously collected sodium fluoride plasma (NaF plasma) and serum sample pairs were submitted by 200 different medical practices to MVZ Labor Ravensburg for glucose measurement. Serum samples, all containing separator gel, were centrifuged at the medical practice after complete coagulation. In contrast, NaF plasma samples were transported uncentrifuged. Upon arrival at the laboratory, all samples were centrifuged immediately and analyzed within two hours. The median of the differences between serum and plasma glucose was calculated for each quintile of the mean glucose concentration. A contingency table was used to analyze the classification outcomes based on plasma and serum glucose concentrations.

In the concentration range above 100 mg/dL, the difference between serum and plasma glucose levels was minimal (4th quintile: 0.54 mg/dL; 5th quintile: 1.13 mg/dL). Below 100 mg/dL, the difference increased linearly with decreasing glucose concentration (3rd quintile: −1.69 mg/dL; 2nd quintile: −5.20 mg/dL; 1st quintile: −11.05 mg/dL). Using 126 mg/dL as the glucose cutoff, positive and negative diabetes classifications from NaF plasma and serum glucose agreed in 97.8% of cases. Overall, only 0.3% more cases of diabetes mellitus were classified based on NaF plasma glucose compared to serum glucose.

In real-world outpatient settings, glucose measurement in on-site centrifuged serum with separator gel does not appear to present a clinically relevant disadvantage for diabetes mellitus screening compared to on-site uncentrifuged NaF plasma. However, neither serum nor non-acidified NaF plasma ensures complete inhibition of in vitro glycolysis and must therefore be regarded as suboptimal for definitive diagnostic purposes.