ARTICLE IN PRESS

Blood Glucose Determination:

Effect of Tube Additives
Giuseppe Lippi*,1,2, Mads Nybo†,2, Janne Cadamuro‡,2,
Joao T. Guimaraes§,¶,k,2, Edme
e van Dongen-Lases#,2,
,2
Ana-Maria Simundic**
*Section of Clinical Biochemistry, University of Verona, Verona, Italy
†
Clinical Biochemistry and Pharmacology, Odense University Hospital, Odense, Denmark
‡
Paracelsus Medical University, Salzburg, Austria
§
São João Hospital Center, University of Porto, Porto, Portugal
¶
Faculty of Medicine, University of Porto, Porto, Portugal
k
EPI Unit, Institute of Public Health, University of Porto, Porto, Portugal
#
Academic Medical Center, Amsterdam, The Netherlands
**University Hospital Sveti Duh, Zagreb, Croatia
1
Corresponding author: e-mail addresses: giuseppe.lippi@univr.it; ulippi@tin.it
2
On behalf of the European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) Working
Group for Preanalytical Phase (WG-PRE).

Contents
1. Introduction 2
2. Literature Search 5
3. Search Results 6
3.1 Impact of Citrate Buffer on Glucose Concentration and Stability 7
3.2 Impact of Citrate Buffer on Spurious Hemolysis 7
4. Discussion 19
5. Conclusions 20
Conflict of Interest 21
References 21

Abstract
The measurement of fasting plasma glucose may be biased by a time-dependent
decrease of glucose in blood tubes, mainly attributable to blood cell metabolism when
glycolysis is not rapidly inhibited or blood cells cannot be rapidly separated from
plasma. Although glycolysis inhibitors such as sodium fluoride (NaF) in combination
with potassium oxalate (KOx) are currently used for overcoming this drawback, their
efficacy for stabilizing blood glucose is seemingly limited, and probably lower than that
of newer additives such as the citrate buffer. Therefore, we performed a critical analysis
of the current scientific literature aimed to generate evidence-based information about
the advantages of using citrate buffer in blood tubes compared to the more conven-
tional NaF additive. The results of our systematic overview of the literature show that
citrate blood tubes represent a considerable step forward in achieving more accurate

Advances in Clinical Chemistry # 2018 Elsevier Inc. 1

ISSN 0065-2423 All rights reserved.
https://doi.org/10.1016/bs.acc.2017.12.003
 ARTICLE IN PRESS

2 Giuseppe Lippi et al.

and reliable plasma glucose measurements, thereby limiting the risk of under-
diagnosing diabetes due to spurious decrease of glucose concentration in
uncentrifuged blood specimens, ensuring higher stability of glucose levels over time,
while simultaneously producing less hemolysis compared to NaF blood tubes. There-
fore, we suggest that the use of this new mixture should be encouraged for achieving
a higher degree of accuracy and standardization of plasma glucose measurements.

ABBREVIATIONS
ADA American Diabetes Association
EDTA ethylenediaminetetraacetic acid
EFLM European Federation of Clinical Chemistry and Laboratory Medicine
FPG fasting plasma glucose
HbA1c hemoglobin A1c
KOx potassium oxalate
NaF sodium fluoride
OGTT oral glucose tolerance test
WG-PRE Working Group for Preanalytical Phase
WHO World Health Organization

1. INTRODUCTION
According to the most recent statistics from the World Health Orga-
nization (WHO), the worldwide prevalence of diabetes was 8.5% in 2014
among adults aged 18 years or older, with a proportion that has nearly dou-
bled since 1980 [1]. Notably, the prevalence of this condition has increased
more sharply in middle- and low-income countries than in so-called West-
ern countries. As regard the possible complications, diabetes is currently reg-
arded as a leading cause of blindness, renal failure, or cardiovascular disease,
and as many as 2.2 million worldwide deaths can be directly attributed to the
long-term effects of hyperglycemia, so that it is projected that diabetes will
become the seventh most frequent cause of death in 2030 [1].
The criteria for diagnosing and monitoring diabetes have evolved
remarkably over time [2,3]. According to the American Diabetes Associa-
tion (ADA), diabetes can now be diagnosed if just one out of four major
criteria is met: a hemoglobin A1c (HbA1c) value 48 mmol/mol (i.e.,
>6.5%), a fasting plasma glucose (FPG) 7.0 mmol/L (126 mg/dL), a
plasma glucose 11.1 mmol/L (200 mg/dL) during an oral glucose toler-
ance test (OGTT), or the presence of symptoms of hyperglycemia in
 ARTICLE IN PRESS

Tube Additives and Glucose Measurement 3

association with a random plasma glucose concentration 11.1 mmol/L

(200 mg/dL) [2]. Despite all four major criteria being eligible for diagnosing
diabetes, the measurement of FPG remains commonplace due to the fact
that this is probably the least expensive and most accessible approach world-
wide, including low-income countries [4], while gestational diabetes can
only be diagnosed with glucose measurement [2]. Unlike HbA1c, however,
the measurement of FPG is plagued by a well-known preanalytical issue due
to a time-dependent decrease of glucose in uncentrifuged blood tubes,
which is attributable to in vitro blood cell (especially erythrocyte) metabo-
lism when glycolysis is not rapidly inhibited or blood cells cannot be rapidly
separated from plasma [5]. To overcome this crucial preanalytical issue,
which can obviously lead to underdiagnosing diabetes in blood tubes in
which a spurious decrease of glucose concentration has occurred, some
potential solutions have been envisaged.
The first breakthrough for stabilizing glucose concentration in the blood
tube was the introduction of sodium fluoride (NaF) in combination with the
anticoagulant potassium oxalate (KOx) [6,7]. The mechanism of fluoride
action is based on the inhibition of the enzyme enolase (E.C. 4.2.1), which
acts late in the glycolytic pathway (Fig. 1). As such, the activity of glycolytic
enzymes located upstream of enolase is not significantly affected, so they
remain active and continuously metabolize glucose. This clearly explains
why the complete effect of fluoride on inhibition of glycolysis can take as
much as 4 h, during which time the concentration of glucose in the blood
tube may decrease considerably, especially when the sample is stored at room
temperature or even warmer [5,6]. Although the initial decay of glucose
concentration can be limited by rapid centrifugation or storage of blood
tubes in an ice/water slurry [8], this is often impractical or unfeasible, espe-
cially when blood is drawn relatively far from the laboratory, such as in dis-
tant wards or in peripheral phlebotomy centers [9], or more simply because
the recommended protocols are frequently overlooked [10]. Moreover, this
process may lead to patient identification errors, since tube label may
become unreadable when blood tubes are placed into ice water.
Another advance in the stabilization of blood glucose in uncentrifuged
blood tubes followed the discovery that the hexokinase enzyme (E.C.
2.7.1), located at the beginning of the glycolytic pathway (Fig. 1), is only
active at pH 5.9 or higher [5,7]. Therefore, acidification of blood by means
of a citrate buffer would prevent glucose catabolism due to ongoing glycol-
ysis at a much earlier step, and hence more rapidly than using fluoride. As for
this important finding, a new generation of blood tubes has been designed,
 ARTICLE IN PRESS

4 Giuseppe Lippi et al.

Fig. 1 Glycolysis and its inhibitors. KOx, potassium oxalate; NaF, sodium fluoride.

which contain three different types of additives, one acting to reduce the pH
of blood and thus inhibiting the enzyme hexokinase (i.e., citrate), the second
directly inhibiting the enzyme enolase (i.e., NaF), and the third irreversibly
inhibiting blood coagulation (i.e., ethylenediaminetetraacetic acid; EDTA).
Some of the major manufacturers of blood-collecting systems have devel-
oped and commercialized these blood tubes with different size and aspiration
volume, all containing this new mixture in different physical states (i.e., liq-
uid or granular) (Table 1) [11]. Since the use of this innovative mixture is
rapidly increasing around the world [11,12,13], the European Federation
of Clinical Chemistry and Laboratory Medicine (EFLM) Working Group
for Preanalytical Phase (WG-PRE) felt being responsible to perform a crit-
ical analysis of the current scientific literature on the effectiveness of citrate
buffer to stabilize glucose in blood tubes compared to the more conventional
NaF additive. The aim was to generate evidence-based information about
the use of these tubes for measurement of plasma glucose.
 ARTICLE IN PRESS

Tube Additives and Glucose Measurement 5

Table 1 Citrate Blood Tubes for Plasma Glucose Measurement

Tube Size and Dilution
Manufacturer Tube Name Draw Volume Mixture Factor
Sarstedt S-Monovette® 11  66 mm; Liquid; citrate/ 1:16
GlucoEXACT 3.1 mL fluoride
Terumo Venosafe FC-Mixture 13  75 mm; Granular; citrate/ No
2.0/3.0 mL fluoride/EDTA
Greiner Vacuette Glucomedicsa 13  75 mm; Liquid; citrate/ 1:16
2.0 mL fluoride/EDTA
Greiner Vacuette FC-Mix 13  75 mm; Granular; citrate/ No
2.0/3.0 mL fluoride/EDTA
a
No longer commercialized.

2. LITERATURE SEARCH
An electronic search was carried out in Medline (interface PubMed),
Scopus, and Web of Science, using the keywords “glucose” AND “tube(s)”
AND “citrate” OR “Glucomedics” OR “GlucoEXACT” OR “Venosafe”
OR “FC-Mix,” with the aim of identifying original studies which investi-
gated the impact of the new glycolysis inhibitor mixture containing citrate
buffer on baseline plasma glucose measurement, on the stability of glucose in
blood tubes over time, and on the risk of producing spurious hemolysis. No
language or date restrictions were applied. Two independent reviewers
(G.L. and M.N.) screened all items retrieved by the electronic literature sea-
rch for eligibility by reading the title, abstract, and (when necessary) the full
text. Potential disagreement between reviewers was resolved through con-
sensus discussion. The reference list of all pertinent items was then accurately
checked for identifying additional studies (cross-reference search). The fol-
lowing information, when available, was reported for each eligible study:
(a) date of publication; (b) study population and sample size; (c) sample
matrix; (d) difference of baseline glucose concentration between NaF and
citrate buffer plasma or other types of samples; (e) stability over time
(between 30 min and 24 h) of glucose concentration in different blood
tubes; and (f ) hemolysis rate in blood tubes, method used for detecting
hemolysis, and the corresponding cutoff.
 ARTICLE IN PRESS

6 Giuseppe Lippi et al.

3. SEARCH RESULTS
In total, 92 articles were identified according to the search criteria.
Seventy-one of these were excluded because they did not show data about
blood tubes containing citrate buffer (n ¼ 55), were animal studies (n ¼ 9),
did not provide a comparison with paired samples collected at baseline
and afterward (n ¼ 4), or were not experimental studies (n ¼ 3) (Fig. 2).
Overall, 21 articles were finally included in our critical analysis of the
literature: 19 of these evaluating stability of glucose concentration in
uncentrifuged tubes [14–32], 1 evaluating stability of glucose concentration
in both uncentrifuged and centrifuged tubes [33], and the remaining study
evaluating stability of glucose concentrations in centrifuged tubes [34].

Keywords:
Electronic search - “glucose”
- AND “tube(s)”
- AND “citrate”
OR “Glucomedics”
OR “GlucoEXACT”
OR “Venosafe”
OR “FC-Mix

92 articles identified
Excluded for title, abstract, or full text (n = 71)
after exclusion of replicates
- No data on acid citrate blood tubes (n = 55)
- Animal studies (n = 9)
- No paired samples comparison (n = 4)
- No experimental study (n = 3)

21 articles included

1 article using
19 articles using 1 article using
centrifuged and
uncentrifuged tubes centrifuged tubes
uncentrifuged tubes

Fig. 2 Summary of the electronic literature search.

 ARTICLE IN PRESS

Tube Additives and Glucose Measurement 7

The agreement for eligibility between the two reviewers was 100%.
The main characteristics, type of additives, stability of glucose concentration
over time, and potential impact on hemolysis are shown in Tables 2 and 3.

3.1 Impact of Citrate Buffer on Glucose Concentration

and Stability
One of the first important issues emerging from the available published stud-
ies is that the concentration of glucose in blood tubes containing citrate
plasma seems to be consistently higher compared to that measured in paired
specimens of fluoride plasma (mean difference, +4.2%; 95% confidence
interval [95% CI], +6.2% to +2.3%), lithium-heparin plasma (mean differ-
ence, +4.2%; 95% CI, +8.3% to +0.1%), or serum (mean difference, 9.8%;
95% CI, +14.3% to +5.2%) (Table 2). These variations were quite consistent
in the articles analyzed.
The variation of glucose concentration in uncentrifuged samples stored
at room temperature for different time points seemed lower when blood was
collected using tubes containing citrate buffer, intermediate for those con-
taining NaF/KOx, whereas a virtually linear decay could be observed in
both serum and lithium-heparin plasma. A succinct representation of data
obtained in samples stored uncentrifuged for 0.5–4 h at room temperature
is shown in Fig. 3. Since the figure has been constructed by pooling infor-
mation from studies with different designs and populations, these data must
be interpreted cautiously. Nevertheless, it is quite informative regarding the
gradual decrease of glucose concentration in specimens collected with diffe-
rent additives and maintained uncentrifuged at room temperature for differ-
ent time intervals. Notably, the average decrease of glucose in uncentrifuged
blood tubes containing citrate buffer was approximately 0.9% and 1.1% at 2
and 4 h, respectively, whereas the average decay of plasma glucose in blood
tubes containing NaF/KOx was higher (i.e., approximately 6.1% and 8.1%
at 2 and 4 h, respectively). Interestingly, in the two studies investigating the
stability of glucose in blood tubes immediately centrifuged after collection,
virtually identical data were found irrespective of the samples matrix
(Table 2) [33,34], with a mean variation of glucose concentration between
0.7% and 1.3% up to 4 h of storage at room temperature.

3.2 Impact of Citrate Buffer on Spurious Hemolysis

Spurious hemolysis is an important issue in laboratory testing, wherein the
injury of blood cells (especially erythrocytes) is associated with release of
Table 2 Summary of Available Studies Evaluating the Effect of Different Preservatives on Plasma Glucose Measurement
Variation
Authors Samples Sample Matrix Baseline 0–30 min 0–1 h 0–2 h 0–3 h 0 24 h 0–24 h
Uncentrifuged
Gambino Samples from Citrate buffer acid NA 0.3% 1.2%
et al. [14] nonfasting employees plasma (Venosafe) (37°C) (37°C)
(n ¼ 30–90)
Fluoride/KOx NA 4.6% 7.0%
plasma (37°C) (37°C)

ARTICLE IN PRESS
Lithium-heparin NA 0.9%
plasma (RT)
Serum separated NA +0.1%
(37°C)
Yagmur Samples from healthy Citrate buffer acid Reference +1.6% 2.5% 2.9% 1.6% 2.5%
et al. [15] volunteers (n ¼ 10) plasma (RT) (RT) (RT) (RT) (RT)
(GlucoEXACT)
Fluoride/EDTA 5.2% +2.8% 1.6% 0.9% 0.6% 3.3%
plasma (RT) (RT) (RT) (RT) (RT)
del Pino Routine laboratory Citrate buffer acid NA +0.2% 0.1%
et al. [16] samples (n ¼ 79) plasma (Venosafe) (RT) (RT)
Fluoride/KOx NA 3.8% 4.6%
plasma (RT) (RT)
Longini Samples from healthy Citrate buffer acid Reference 6.9%
et al. [17] volunteers (n ¼ 25) plasma (homemade) (RT)
Fluoride/EDTA 5.8% 6.7%
plasma (RT)
Serum separated 3.7% 25.6%
(RT)

ARTICLE IN PRESS
Szoke Samples from healthy Citrate buffer acid Reference
et al. [18] subjects (n ¼ 15) and plasma (Venosafe)
diabetics (n ¼ 35)
Fluoride/KOx 6.7%
plasma
Fobker Samples from healthy Citrate buffer acid NA 2.2%
[19] volunteers (n ¼ 60) plasma (Venosafe) (RT)
Citrate buffer acid NA +0.8%
plasma (RT)
(GlucoEXACT)
Fluoride/KOx NA 3.8%
plasma (RT)
K2EDTA NA 85.2%
(RT)
Continued
Table 2 Summary of Available Studies Evaluating the Effect of Different Preservatives on Plasma Glucose Measurement—cont’d
Variation
Authors Samples Sample Matrix Baseline 0–30 min 0–1 h 0–2 h 0–3 h 0 24 h 0–24 h
Dimeski Samples from healthy Citrate buffer acid Reference 0.8% 0.9% 1.5%
et al. [20] volunteers (n ¼ 42) plasma (RT) (RT) (RT)
(Glucomedics)
Fluoride/KOx 7.0% 5.4% 7.0% 8.5%
plasma (RT) (RT) (RT)
7.4% 4.2% 10.5% 18.5%

ARTICLE IN PRESS
Lithium-heparin
plasma (RT) (RT) (RT)
Serum 10.5% 3.1% 5.5% 10.1%
(RT) (RT) (RT)
EDTA 5.3% 3.1% 7.1% 13.0%
(RT) (RT) (RT)
van den Samples from healthy Citrate buffer acid Reference +0.6% 0.0% 1.0%
Berg et al. employees (n ¼ 22) plasma (Venosafe) (RT) (RT) (RT)
[21]
Fluoride/KOx 0.0% 1.8% 3.0% 4.7%
plasma (ice slurry) (RT) (RT) (RT)
Fluoride/EDTA 1.7% 1.9% 3.7% 5.0%
plasma (ice slurry) (RT) (RT) (RT)
Lithium-heparin 3.4% 1.3% 3.2% 5.7%
plasma (ice slurry) (RT) (RT) (RT)
van den Samples from pregnant Citrate buffer acid NA 1.0%
Berg et al. women (n ¼ 30–50) plasma (Venosafe) (RT)
[22]
Bonetti Samples from fasting Citrate buffer acid NA +0.2% 0.1% 0.2%
et al. [23] and nonfasting plasma (Venosafe) (RT) (RT) (RT)
employees (n ¼ 100)
Lithium-heparin NA 5.7% 9.2% 13.8%
plasma (RT) (RT) (RT)
Fluoride/heparin NA 6.0% 7.8% 8.8%

ARTICLE IN PRESS
plasma (RT) (RT) (RT)
Daly et al. Samples from pregnant Citrate buffer acid Reference
[24] women (n ¼ 121) plasma (no
information on
brand)
Fluoride/EDTA 8.2%
plasma (RT)
Fluoride/EDTA +2.0%
plasma (ice slurry)
Juricic Samples from Citrate buffer acid Reference 0.9% 1.4% 1.6%
et al. [25] nonfasting healthy plasma (RT) (RT) (RT)
volunteers (n ¼ 80) (Glucomedics)
Fluoride/KOx 8.3% 5.6% 7.4% 8.0%
plasma (RT) (RT) (RT)
Serum + separator 11.7% 6.0% 9.4% 12.0%
(RT) (RT) (RT)
Continued
Table 2 Summary of Available Studies Evaluating the Effect of Different Preservatives on Plasma Glucose Measurement—cont’d
Variation
Authors Samples Sample Matrix Baseline 0–30 min 0–1 h 0–2 h 0–3 h 0 24 h 0–24 h
Carey Samples from healthy Citrate buffer acid Reference 0.4% 2.0%
et al. [26] subjects (n ¼ 17) and plasma (Venosafe) (RT) (RT)
diabetics (n ¼ 33)
Fluoride/KOx 3.6% 5.3% 6.2%
plasma (ice slurry) (RT) (RT)
Lithium-heparin 4.6% 6.1% 51.2%

ARTICLE IN PRESS
plasma (ice slurry) (RT) (RT)
Bonetti Samples from healthy Citrate buffer acid NA +1.0% +1.4% +0.9%
et al. [27] volunteers (n ¼ 49) plasma (Venosafe) (RT) (RT) (RT)
Fluoride/EDTA NA 8.1% 10.6% 11.9%
plasma (RT) (RT) (RT)
Lithium-heparin NA 6.7% 11.6% 20.2%
plasma (RT) (RT) (RT)
Serum + separator NA 12.1% 19.9% 32.0%
(RT) (RT) (RT)
Gupta Samples from healthy Citrate buffer acid Reference 2.1% 5.4%
et al. [28] employees (n ¼ 44) plasma (homemade) (RT) (RT)
Fluoride/KOx 5.9% 3.0% 7.1%
plasma (RT) (RT)
Juricic Samples from Citrate buffer acid Reference
et al. [29] nonfasting healthy plasma
volunteers (n ¼ 40) (Glucomedics)
Citrate buffer acid 3.4%
plasma (Venosafe)
Zhang Samples from healthy Citrate buffer acid Reference +0.4% 0.2% 0.4% 1.6%
et al. [30] volunteers (n ¼ 58) plasma (RT) (RT) (RT) (RT)

ARTICLE IN PRESS
(GlucoEXACT)
Serum 5.5% 0.2% 1.0% 1.3% 3.5%
(RT) (RT) (RT) (RT)
van der Samples from pregnant Citrate buffer acid Reference
Hagen woman (n ¼ 43) plasma (Vacuette
et al. [31] FC-Mix)
Fluoride/EDTA +0.1%
plasma (ice slurry)
Stapleton Samples from pregnant Citrate buffer acid Reference 0.3% (RT, 2.5 h)
et al. [32] woman (n ¼ 200) plasma
(GlucoEXACT)
Fluoride/EDTA 1.0% 7.6% (RT, 2.5 h)
plasma (ice slurry)
Lithium-heparin 1.2% 14.5% (RT,
plasma (ice slurry) 2.5 h)
Continued
Table 2 Summary of Available Studies Evaluating the Effect of Different Preservatives on Plasma Glucose Measurement—cont’d
Variation
Authors Samples Sample Matrix Baseline 0–30 min 0–1 h 0–2 h 0–3 h 0 24 h 0–24 h
Dimeski Samples from healthy Citrate buffer acid Reference 0% 0.3% 0.3% 0.6%
et al. [33] subjects (n ¼ 2) plasma (Vacuette (RT) (RT) (RT) (RT)
FC-Mix)
Fluoride/KOx 2.6% 3.7% 5.8% 7.5% 7.8%
plasma (RT) (RT) (RT) (RT)

ARTICLE IN PRESS
Citrate buffer acid NA 1.0% 0.7% 1.5% 1.5%
plasma (Vacuette (2–8°C) (2–8°C) (2–8°C) (2–8°C)
FC-Mix)
Fluoride/KOx 1.8% 2.1% 2.8% 3.8%
plasma (2–8°C) (2–8°C) (2–8°C) (2–8°C)
Citrate buffer acid NA 0.1% 0.8% 0.1% 0.6%
plasma (Vacuette (37°C) (37°C) (37°C) (37°C)
FC-Mix)
Fluoride/KOx 5.5% 7.1% 7.3% 7.3%
plasma (37°C) (37°C) (37°C) (37°C)
Centrifuged
Juricic Samples from Citrate buffer acid Reference 1.0% 1.3%
et al. [34] nonfasting volunteers plasma (RT) (RT)
(n ¼ 40) (Glucomedics)
Fluoride/KOx 9.7% 0.9% 1.1%
plasma (RT) (RT)
11.3%

ARTICLE IN PRESS
Serum + separator 0.9% 1.1%
(RT) (RT)
Dimeski Samples from healthy Citrate buffer acid NA 0.2% +0.5%
et al. [33] subjects (n ¼ 35) and plasma (Vacuette (RT) (RT)
diabetics (n ¼ 6) FC-Mix)
Fluoride/KOx NA 0.7% 0.5%
plasma (RT) (RT)
EDTA, ethylenediaminetetraacetic acid; KOx, potassium oxalate; NA, not available; RT, room temperature.
Table 3 Summary of Available Studies Evaluating the Risk of Spurious Hemolysis of Different Preservatives Used for Plasma
Glucose Measurement
HI
Authors Samples Sample Matrix Method Cutoff Maximum Hemolysis
del Pino et al. Routine laboratory samples Citrate buffer acid plasma NA >1 g/L <3%–19%
[16] (n ¼ 79) (Venosafe)
Fluoride/KOx plasma NA >1 g/L 54%–94%
van den Berg Samples from healthy employees Citrate buffer acid plasma HIL on Roche 60 μmol/ +5.4-fold compared to
et al. [21] (n ¼ 22) (Venosafe) Cobas C501 L lithium-heparin plasma

ARTICLE IN PRESS
Fluoride/KOx plasma HIL on Roche 60 μmol/ +4.9-fold compared to
(ice slurry) Cobas C501 L lithium-heparin plasma
Fluoride/EDTA plasma HIL on Roche 60 μmol/ +1.1-fold compared to
(ice slurry) Cobas C501 L lithium-heparin plasma
Lithium-heparin plasma HIL on Roche 60 μmol/ 11 μmol/L
(ice slurry) Cobas C501 L
Juricic et al. Samples from nonfasting healthy Citrate buffer acid plasma Harboe method >0.3 g/L 0%
[25] volunteers (n ¼ 80) (Glucomedics)
Fluoride/KOx plasma Harboe method >0.3 g/L 13%–50%
Serum + separator Harboe method >0.3 g/L 0%
Carey et al. Samples from healthy subjects Citrate buffer acid plasma HIL on Abbott HI >100 Unvaried up to 24 h
[26] (n ¼ 17) and diabetics (n ¼ 33) (Venosafe) C16000
Fluoride/KOx plasma HIL on Abbott HI >100 +1.9- and +6.5-fold after
(ice slurry) C16000 2 and 24 h
Lithium-heparin plasma HIL on Abbott HI >100 +2.2- and +3.5-fold after
(ice slurry) C16000 2 and 24 h
Bonetti et al. Samples from healthy volunteers Citrate buffer acid plasma HIL on >0.5 g/L 2%
[27] (n ¼ 49) (Venosafe) Siemens
Dimension
Vista
Fluoride/EDTA plasma HIL on >0.5 g/L 88%
Siemens
Dimension
Vista
Lithium-heparin plasma HIL on >0.5 g/L 2%

ARTICLE IN PRESS
Siemens
Dimension
Vista
Serum + separator HIL on >0.5 g/L 41%
Siemens
Dimension
Vista
van der Samples from pregnant woman Citrate buffer acid plasma HIL on Roche NA 0%
Hagen et al. (n ¼ 43) (Vacuette FC-Mix) C8000
[31]
Fluoride/EDTA plasma HIL on Roche NA 0%
(ice slurry) C8000
EDTA, ethylenediaminetetraacetic acid; HI(L), hemolysis index (level); KOx, potassium oxalate; NA, not available.
 ARTICLE IN PRESS

18 Giuseppe Lippi et al.

5%
Percentage decrease from baseline

0%

-5%

-10%

Citrate buffer
-15%
NaF/Kox
Lithium-heparin
Serum
-20%
0.0 1.0 2.0 3.0 4.0
Time (h)
Fig. 3 Variation of glucose concentration in samples stored uncentrifuged for 0.5–4 h at
room temperature. KOx, potassium oxalate; NaF, sodium fluoride.

many intracellular compounds in the surrounding plasma or serum, thus

introducing a clinically significant bias in test results of many clinical chem-
istry analytes, including glucose [35,36]. Despite the fact that specimens col-
lected for measuring FPG are not traditionally used for testing other
biochemical analytes, the occurrence of hemolysis in these samples cannot
be simply neglected. The impact on spurious hemolysis of different blood
tube additives used for glucose testing is shown in Table 3. The use of citrate
buffer does not seemingly generate a higher degree of hemolysis compared
to lithium-heparin plasma. In a single study the concentration of cell-free
hemoglobin has increased over fivefold compared to lithium-heparin plasma
after storing uncentrifuged samples for 120 min at room temperature [22].
However, the maximum value of hemolysis was still within the limit of sig-
nificant bias set in that study. Unlike samples collected with citrate buffer,
those collected with NaF were frequently plagued by spurious hemolysis,
with a frequency of clinically significant hemolysis (i.e., >1.0 g/L) that
was found in one study to be as high as 94% of all uncentrifuged samples
stored for 120 min at room temperature, compared to 19% of all samples
collected with citrate buffer stored under identical conditions [16].
 ARTICLE IN PRESS

Tube Additives and Glucose Measurement 19

4. DISCUSSION
A precise measurement of plasma glucose concentration is the main-
stay for an accurate diagnosis of diabetes, in order to limit the risk of false-
positive (i.e., overdiagnosis) and false-negative (i.e., underdiagnosis) results
[37,38].
The progressive decrease of glucose concentration in uncentrifuged
blood tubes has been well known for decades [39], so leading to the addition
of a number of glycolysis inhibitors to the blood tubes, the most common of
which is indeed NaF/KOx [5,38,40]. Although this additive enables a better
stabilization of glucose than using conventional serum and lithium-heparin
blood tubes, its efficacy remains inadequate. This is clearly demonstrated by
data from our critical analysis of the current scientific literature, wherein
plasma glucose concentration may decrease by over 6% in uncentrifuged
NaF/KOx samples stored at room temperature for 2 h (Fig. 3). After this
period, the rate of plasma glucose consumption slows down, reaching a
maximum 8% reduction after 4 h. Another important disadvantage of
NaF is its impact on erythrocyte integrity, as shown by data on cell-free
hemoglobin (Table 3). Overall, the percentage of significant hemolysis using
this additive can be as high as 94% of all samples.
Unlike NaF, the use of citrate buffer, a strong inhibitor of hexokinase, an
enzyme located upstream in the glycolytic pathway (Fig. 1), seems to pro-
duce more reliable measurements of plasma glucose in samples left
uncentrifuged for many hours. According to our analysis of the scientific lit-
erature, the average decay of glucose concentration in blood tubes left
uncentrifuged for up to 24 h was below 1.7% and thus always within the
current limits of imprecision and inaccuracy calculated from data on
within-subject and between-subject biologic variation (2.3% and 1.8%,
respectively) [41]. Notably, citrate buffer has also a lower impact in produc-
ing spurious hemolysis, as shown in Table 3. This may also partially explain
the finding that the glucose concentration measured in samples containing
citrate buffer is constantly higher than that measured in paired samples con-
taining other additives (Table 2), since the release of hexokinase in plasma
due to hemolysis may spuriously decrease glucose concentration measured
using the reference hexokinase assay [42–44]. According to this hypothesis,
the assessment of plasma glucose in samples measured in tubes containing
citrate buffer may hence more reliably reflect the real (in vivo) concentration
 ARTICLE IN PRESS

20 Giuseppe Lippi et al.

of this analyte. Irrespective of the mechanism underlying the difference of

plasma glucose in sample collected with citrate buffer and other additives,
such a meaningful variation raises important questions about the current
diagnostic thresholds, which have been identified and validated using differ-
ent additives [2]. This aspect has been clearly highlighted in recent publica-
tions showing that the shift from NaF to citrate buffer tubes has contributed
to a considerable increase in the rate of glucose measurements above the cur-
rent decision limits for diabetes [34,45,46], warranting further research to
define whether reference intervals and diagnostic thresholds may necessitate
reevaluation or reassessment using citrate blood tubes.
Interestingly, the glucose concentration was found stable (variation
always <1.3% up to 3 h of storage) in blood tubes immediately centrifuged
after collection regardless of the sample matrix, thus highlighting that the
major advantage of citrate buffer is enhancing glucose stability when blood
tubes are left uncentrifuged for over 30 min, but not when the appropriate
procedure (i.e., immediate centrifugation) is followed.
Some critical aspects of blood tubes containing citrate buffer should also
be highlighted. These typically entail a minimum blood tube filling, which is
critical using liquid additives (Table 1), along with the need of immediate
mix of blood tube after collection to enable an appropriate contact between
blood and additives [20]. The use of lyophilized additive may be advisable,
since this will help overcoming the need to adjust glucose results for the dilu-
tion factor, thus ultimately decreasing the risk of reporting inaccurate data.
Importantly, the current thresholds for diagnosing glucose have not been
defined using citrate blood tubes, in which glucose concentration is clearly
higher than in serum or other types of plasma.

5. CONCLUSIONS
The currently available evidence comprehensively reviewed in this
article suggests that citrate blood tubes represent a unique opportunity
and a considerable step forward in achieving more accurate and reliable
plasma glucose measurements, thereby limiting the risk of underdiagnosing
diabetes due to spurious decrease of glucose concentration in uncentrifuged
blood specimens [47], ensuring higher stability of glucose levels over the
time, while simultaneously producing less hemolysis compared to NaF
tubes. Therefore, we suggest that the use of this new mixture should be
encouraged for achieving a higher degree of accuracy and standardization
of plasma glucose measurements and further clinical evidence is required
 ARTICLE IN PRESS

Tube Additives and Glucose Measurement 21

to assure that the currently used cutoffs, derived from a different sample
matrix, can be applied straightforwardly to these new tubes, with no risk
of overdiagnosing diabetes. Finally, we make urgent appeal to associations
and international organizations such as the ADA and the WHO for addi-
tional research and establishment of new cutoffs obtained with the rec-
ommended additive.

CONFLICT OF INTEREST
All authors declare no conflict of interest.

REFERENCES
[1] World Health Organization, Global report on diabetes, World Health Organization,
Geneva, 2016.
[2] American Diabetes Association, Standards of medical care in diabetes—2017, Diabetes
Care 40 (Suppl. 1) (2017) S1–S135.
[3] S.N. Narla, M. Jones, K.L. Hermayer, Y. Zhu, Critical care glucose point-of-care test-
ing, Adv. Clin. Chem. 76 (2016) 97–121.
[4] G. Lippi, G. Targher, Haemoglobin A1c and diagnosis of diabetes, not ready for the
prime time? Ann. Clin. Biochem. 49 (2012) 508.
[5] L.M. Mikesh, D.E. Bruns, Stabilization of glucose in blood specimens: mechanism of
delay in fluoride inhibition of glycolysis, Clin. Chem. 54 (2008) 930–932.
[6] M.J. Peake, D.E. Bruns, D.B. Sacks, A.R. Horvath, It’s time for a better blood collec-
tion tube to improve the reliability of glucose results, Diabetes Care 36 (2013) e2.
[7] M. Montagnana, G. Lippi, Overcoming preanalytical issues for diagnosing diabetes with
fasting plasma glucose, Ann. Transl. Med. 5 (2017) 257.
[8] D.B. Sacks, M. Arnold, G.L. Bakris, D.E. Bruns, A.R. Horvath, M.S. Kirkman, et al.,
Executive summary: guidelines and recommendations for laboratory analysis in the
diagnosis and management of diabetes mellitus, Clin. Chem. 57 (2011) 793–798.
[9] G. Lippi, A.M. Simundic, Laboratory networking and sample quality: a still relevant
issue for patient safety, Clin. Chem. Lab. Med. 50 (2012) 1703–1705.
[10] S.A.A. van den Berg, M.H.M. Thelen, K.J. Boonen, Inventarisatie van de (pre)analyt-
ische aspecten van de glucose bepaling in nederlandse laboratoria; tijd voor uniformiteit,
Ned. Tijdschr. Klin. Chem. Labgeneesk 40 (2015) 68–71.
[11] S. Pasqualetti, D. Szőke, S. Birindelli, A. Dolci, M. Panteghini, Optimal collection
tubes for plasma glucose determination: confusion reigns supreme, Clin. Chem. Lab.
Med. 54 (2016) e281–283.
[12] D.E. Bruns, Are fluoride-containing blood tubes still needed for glucose testing? Clin.
Biochem. 46 (2013) 289–290.
[13] R. Gambino, Sodium fluoride: an ineffective inhibitor of glycolysis, Ann. Clin. Bio-
chem. 50 (2013) 3–5.
[14] R. Gambino, J. Piscitelli, T.A. Ackattupathil, J.L. Theriault, R.D. Andrin,
M.L. Sanfilippo, et al., Acidification of blood is superior to sodium fluoride alone as
an inhibitor of glycolysis, Clin. Chem. 55 (2009) 1019–1021.
[15] E. Yagmur, J. van Helden, A. Koch, J. Jadem, F. Tacke, C. Trautwein, Effective inhi-
bition of glycolysis in venous whole blood and plasma samples, J. Lab. Med. 36 (2012)
169–177.
[16] I.G. del Pino, I. Constanso, L.V. Mourı́n, C.B. Safont, P.R. Vázquez, Citric/citrate
buffer: an effective antiglycolytic agent, Clin. Chem. Lab. Med. 51 (2013) 1943–1949.
 ARTICLE IN PRESS

22 Giuseppe Lippi et al.

[17] M. Longini, E. Belvisi, R. Leoncini, R. Guerranti, C. Scapellato, Management of

blood glucose in the clinical analysis laboratory: evaluation of old and new inhibitors
of glycolysis. Preliminary data, Riv. Ital. Med. Lab. 10 (2014) 44–49.
[18] D. Szoke, C. Valente, M. Panteghini, Better blood collection tubes for plasma glucose:
ready for prime time? Clin. Chem. Lab. Med. 52 (2014) e87–89.
[19] M. Fobker, Stability of glucose in plasma with different anticoagulants, Clin. Chem.
Lab. Med. 52 (2014) 1057–1060.
[20] G. Dimeski, K.S. Yow, N.N. Brown, What is the most suitable blood collection tube
for glucose estimation? Ann. Clin. Biochem. 52 (2015) 270–275.
[21] S.A. van den Berg, M.H. Thelen, L.P. Salden, S.W. van Thiel, K.J. Boonen, It takes
acid, rather than ice, to freeze glucose, Sci. Rep. 5 (2015) 8875.
[22] S.A. van den Berg, M.J. de Groot, L.P. Salden, P.J. Draad, I.M. Dijkstra, S. Lunshof,
et al., Pregnancy diabetes: a comparison of diagnostic protocols based on point-of-care,
routine and optimized laboratory condition, Sci. Rep. 5 (2015) 16302.
[23] G. Bonetti, M. Carta, M. Montagnana, C. Lo Cascio, A.R. Bonfigli, A. Mosca, et al.,
Effectiveness of citrate buffer-fluoride mixture in Terumo tubes as an inhibitor of
in vitro glycolysis, Biochem. Med. (Zagreb) 26 (2016) 68–76.
[24] N. Daly, I. Flynn, C. Carroll, M. Stapleton, R. O’Kelly, M.J. Turner, Comparison
of citrate-fluoride-EDTA with fluoride-EDTA additives to stabilize plasma glucose
measurements in women being screened during pregnancy with an oral glucose
tolerance test: a prospective observational study, Clin. Chem. 62 (2016) 886–887.
[25] G. Juricic, A. Bakliza, A. Saracevic, L.M. Kopcinovic, S. Dobrijevic, S. Drmic, et al.,
Glucose is stable during prolonged storage in un-centrifuged Greiner tubes with liquid
citrate buffer, but not in serum and NaF/KOx tubes, Clin. Chem. Lab. Med. 54 (2016)
411–418.
[26] R. Carey, H. Lunt, H.F. Heenan, C.M. Frampton, C.M. Florkowski, Collection tubes
containing citrate stabiliser over-estimate plasma glucose, when compared to other sam-
ples undergoing immediate plasma separation, Clin. Biochem. 49 (2016) 1406–1411.
[27] G. Bonetti, V. Cancelli, G. Coccoli, G. Piccinelli, D. Brugnoni, L. Caimi, et al., Which
sample tube should be used for routine glucose determination? Prim. Care Diabetes
10 (2016) 227–232.
[28] S. Gupta, A.K. Gupta, M. Verma, K. Singh, A. Kaur, B. Chopra, et al., A study to com-
pare the plasma glucose levels obtained in sodium fluoride and citrate buffer tubes at a
tertiary care hospital in Punjab, Int. J. Appl. Basic Med. Res. 6 (2016) 50–53.
[29] G. Juricic, A. Saracevic, L.M. Kopcinovic, A. Bakliza, A.M. Simundic, The evidence
for clinically significant bias in plasma glucose between liquid and lyophilized citrate
buffer additive, Clin. Biochem. 49 (2016) 1402–1405.
[30] X. Zhang, C. Rong, H. Li, X. Qin, S. Li, Z. Lai, et al., Glucose stability study: NaF/
citrate plasma versus serum, Clin. Lab. 62 (2016) 389–393.
[31] E.A. van der Hagen, M.J. Fokkert, A.M. Kleefman, M.H. Thelen, S.A. van den Berg,
R.J. Slingerland, Technical and clinical validation of the Greiner FC-Mix glycaemia
tube, Clin. Chem. Lab. Med. 55 (2017) 1530–1536, https://doi.org/10.1515/cclm-
2016-0944.
[32] M. Stapleton, N. Daly, R. O’Kelly, M.J. Turner, Time and temperature affect gly-
colysis in blood samples regardless of fluoride-based preservatives: a potential under-
estimation of diabetes, Ann. Clin. Biochem. 54 (2017) 671–676, https://doi.org/
10.1177/0004563216682978.
[33] G. Juricic, L.M. Kopcinovic, A. Saracevic, A. Bakliza, A.M. Simundic, Liquid citrate
acidification introduces significant glucose bias and leads to misclassification of patients
with diabetes, Clin. Chem. Lab. Med. 54 (2016) 363–371.
[34] G. Dimeski, K.S. Yow, N.N. Brown, Evaluation of the accuracy of the Greiner
Bio-One FC Mix Glucose tube, Clin. Chem. Lab. Med. 55 (2017) e96–98.
 ARTICLE IN PRESS

Tube Additives and Glucose Measurement 23

[35] A.M. Simundic, E. Topic, N. Nikolac, G. Lippi, Hemolysis detection and management
of hemolysed specimens, Biochem. Med. (Zagreb) 20 (2010) 154–159.
[36] A.M. Simundic, N. Nikolac, W.G. Guder, Preanalytical variation and preexamination
processes, in: N. Rifai, R. Horvath, C. Wittwer (Eds.), Tietz Textbook of Clinical
Chemistry and Molecular Diagnostics, sixth ed., Elsevier, 2017, pp. 81–120.
[37] S. Selph, T. Dana, I. Blazina, C. Bougatsos, H. Patel, R. Chou, Screening for type 2
diabetes mellitus: a systematic review for the U.S. Preventive Services Task Force,
Ann. Intern. Med. 162 (2015) 765–776.
[38] D.E. Bruns, W.C. Knowler, Stabilization of glucose in blood samples: why it matters,
Clin. Chem. 55 (2009) 850–852.
[39] A.Y.W. Chan, R. Swaminanthan, C.S. Cockram, Effectiveness of sodium fluoride as a
preservative of glucose in blood, Clin. Chem. 35 (1989) 315–317.
[40] R.H. Major, Potassium fluoride as a preservative for blood, JAMA 81 (1923) 1952.
[41] C. Ricos, V. Alvarez, F. Cava, J.V. Garcia-Lario, A. Hernandez, C.V. Jimenez, et al.,
Current databases on biologic variation: pros, cons and progress, Scand. J. Clin. Lab.
Invest. 59 (1999) 491–500.
[42] D. Grafmeyer, M. Bondon, M. Manchon, P. Levillain, The influence of bilirubin,
haemolysis and turbidity on 20 analytical tests performed on automatic analysers. Results
of an interlaboratory study, Eur. J. Clin. Chem. Clin. Biochem. 33 (1995) 31–52.
[43] G. Lippi, G.L. Salvagno, M. Montagnana, G. Brocco, G.C. Guidi, Influence of hemo-
lysis on routine clinical chemistry testing, Clin. Chem. Lab. Med. 44 (2006) 311–316.
[44] M. Koseoglu, A. Hur, A. Atay, S. Cuhadar, Effects of hemolysis interference on routine
biochemistry parameters, Biochem. Med. (Zagreb) 21 (2011) 79–85.
[45] P. Ridefelt, T. Åkerfeldt, J. Helmersson-Karlqvist, Increased plasma glucose levels after
change of recommendation from NaF to citrate blood collection tubes, Clin. Biochem.
47 (2014) 625–628.
[46] M. Norman, I. Jones, The shift from fluoride/oxalate to acid citrate/fluoride blood col-
lection tubes for glucose testing—the impact upon patient results, Clin. Biochem.
47 (2014) 683–685.
[47] R. Gambino, D.E. Bruns, Stabilization of glucose in blood samples: out with the old, in
with the new, Clin. Chem. Lab. Med. 51 (2013) 1883–1885.