Clin Chem Lab Med 2016; aop

EFLM Opinion Paper

Giuseppe Lippi, Michael P. Cornes, Kjell Grankvist, Mads Nybo and Ana-Maria Simundic*,
on behalf of the Working Group for Preanalytical Phase (WG-PRE), European Federation of
Clinical Chemistry and Laboratory Medicine (EFLM)

EFLM WG-Preanalytical phase opinion paper:

local validation of blood collection tubes in
clinical laboratories
DOI 10.1515/cclm-2015-1274 has drafted a consensus document aimed to provide a set
Accepted for publication January 2, 2016 of essential requisites, technical criteria (e.g. presence of
physical defects, malfunctioning, safety problems) and
Abstract: The selection or procurement of blood collection
clinical issues for supporting laboratory professionals in
devices in healthcare facilities is often an underestimated
organization blood collection tubes tenders and validat-
issue. This is probably due to different factors including
ing new devices before local routine implementation. The
the lack of knowledge of policymakers, hospital admin-
laboratory professionals should also make sure that the
istrators and even laboratory managers about the impor-
tenders accurately and strictly define the responsibili-
tance of preanalytical quality and phlebotomy process, as
ties for validation experiments and the potential conse-
well as to the absence of reliable guidelines or recommen-
quences in the case the validation outcome shows that
dations on how to precisely assess the quality of blood
tubes due not fulfill the expectations.
collection devices around the globe. With the awareness
that a gap remains between manufacturers’ and local vali- Keywords: blood collection; blood tubes; errors; preana-
dation of blood collection devices, the Working Group for lytical variability; venipuncture.
Preanalytical Phase (WG-PRE) of the European Federation
of Clinical Chemistry and Laboratory Medicine (EFLM)

Introduction
*Corresponding author: Prof. Ana-Maria Simundic, Department of Preanalytical variability plays a crucial role in laboratory
Medical Laboratory Diagnostics, University Hospital “Sveti Duh”, diagnostics [1]. Several lines of evidence, accumulated
Sveti Duh 64, 10,000 Zagreb, Croatia, Phone: +385 1 3712 021, over the past decades, attest that most errors through-
E-mail: am.simundic@gmail.com; and European Federation for out the testing process emerge from manually intensive
Clinical Chemistry and Laboratory Medicine (EFLM) Working Group
activities related to collection and management of bio-
for Preanalytical Phase (WG-PRE)
Giuseppe Lippi: European Federation for Clinical Chemistry and
logical samples [2]. The use of high quality blood collec-
Laboratory Medicine (EFLM) Working Group for Preanalytical Phase tion devices is an aspect of utmost importance in routine
(WG-PRE); and Section of Clinical Biochemistry, University of laboratory practice, wherein inappropriate or even differ-
Verona, Verona, Italy ent sample containers may be a source of preanalytical
Michael P. Cornes: European Federation for Clinical Chemistry and bias, which can ultimately impact results of testing both
Laboratory Medicine (EFLM) Working Group for Preanalytical Phase
in clinical and research settings [3]. It is also noteworthy
(WG-PRE); and Clinical Chemistry Department, New Cross Hospital,
Wolverhampton, United Kingdom that both the International Organization for Standardiza-
Kjell Grankvist: European Federation for Clinical Chemistry and tion (ISO) 9001:2008 and the ISO 15189:2012 certification
Laboratory Medicine (EFLM) Working Group for Preanalytical and accreditation procedures include standards encom-
Phase (WG-PRE); and Department of Medical Biosciences, Clinical passing all laboratory activities, including preanalytical
Chemistry, Umeå University, Umea, Sweden
procedures, which should be standardized and monitored
Mads Nybo: European Federation for Clinical Chemistry and
Laboratory Medicine (EFLM) Working Group for Preanalytical
according to evidence-based practices.
Phase (WG-PRE); and Department of Clinical Biochemistry and Despite accumulating evidence about pre­analytical
Pharmacology, Odense University Hospital, Odense, Denmark quality assurance, selection and procurement of blood

Brought to you by | Universita degli Studi di Padova

Authenticated
Download Date | 3/29/16 7:47 AM
2      Lippi et al.: EFLM WG-PRE: local validation of blood collection tubes in clinical laboratories

collection devices in healthcare facilities is often an Essential requisite for purchasing

underestimated issue. National, regional and local
tenders are frequently plagued by policies, guided pri- blood collection devices
marily by the price rather than by quality of devices. This
Blood collection systems are considered as integrated in
is probably due to different factors, including the lack of
vitro diagnostic (IVD) medical devices and are thereby
knowledge of policymakers, hospital administrators and
regulated by a number of national and supranational
even laboratory managers about the importance of pre-
bodies and organizations such as the European Com-
analytical quality and the phlebotomy process, as well as
munity (EC) or the US Food and Drug Administration
to the absence of reliable guidelines or recommendations
(FDA) [7]. A key characteristic, highlighted by virtually all
on how to precisely assess the quality of blood collection
regulatory documents, is that the whole blood collection
devices around the globe.
device (i.e. safety needle, butterfly needle, holder and
Validation studies are crucial activities for generat-
blood tube) must be regarded as an integrated system.
ing reliable evidence that a novel instrument, method,
Therefore, the combination of the different parts must
reagent or device is fit for purpose and satisfies the par-
be safe and should not impair the performance of the
ticular requirements for its specific intended use [4]. In
individual components [8]. Manufacturers are responsi-
2010, the Clinical and Laboratory Standards Institute
ble for assuring the full compatibility between the com-
(CLSI) released a specific GP-34A guideline, aiming to
ponents of the system, to subside the risk of impairing
detail the procedures for validation and verification of
the quality of testing and jeopardizing (both operator and
tubes for venous or capillary blood specimen collec-
patient) safety. Importantly, tenders allowing acquisition
tion [5]. However, this document is mainly orientated
of devices from different manufacturers may end up with
towards validation of blood collection tubes from a
combinations that are not validated for clinical use. The
­manufacturer’s perspective to ensure that design goals
manufacturers themselves also typically include specific
and performance claims are met. With the awareness
claims in their product datasheets, stating that “devices
that a gap remains between manufacturers validation
(needles, single-use holders, safety devices) are designed
and clinical laboratories implementation, the Working
to be used as a system of products, and the integration
Group for Preanalytical Phase (WG-PRE) of the European
of other manufacturer’s products is solely the respon-
Federation of Clinical Chemistry and Laboratory Medi-
sibility of the user”. However, according to the EFLM
cine (EFLM) has drafted the present consensus docu-
WG-PRE, it is outside the role and duty of laboratory
ment, which aims to provide a set of simple elements
professionals to perform a thoughtful validation study
and criteria specifically for laboratory professionals, to
to establish whether or not an integrated system is safe
verify whether the introduction of new blood collection
and does not impair the quality of testing. Therefore, the
tubes in clinical laboratories fulfills basic criteria of tech-
possibility of using separate parts of the blood collection
nical and clinical acceptability.
system obtained or purchased from different manufactur-
ers is strongly discouraged by the EFLM WG-PRE except
when the integration has been previously validated by
Operative definitions the manufacturer(s) or by national or supranational
­regulation bodies.
In agreement with the CLSI guideline GP-34A [5], the Apart from research and development, tube manufac-
“comparative tube” is defined as the blood collec- turers should be able to also demonstrate usability studies
tion tube currently used by the clinical laboratory, the of their products following study subject recruitment
“control tube” is defined as the blood collection tube that according to Good Clinical Practice/International Con-
is to be introduced and replace the current. The “desir- ference on Harmonisation of Technical Requirements for
able quality specifications for bias” are conventionally Registration of Pharmaceuticals for Human Use (GCP/IHC)
derived from biological variation. The “validation” is guidelines and independent Ethical Committee requests.
finally defined according to the current ISO 9000:2005 Demonstration of ease of use, sustained plasma quality,
specifications [6], as “confirmation, through the provi- perceived value compared to the comparative devices
sion of objective evidence that the main requirements and potential performance risk should be included. User
for a specific intended use or application have been (named sites) feedback on collection, transport (foot,
fulfilled”. pneumatic tube, courier, etc.), reception and analysis

Brought to you by | Universita degli Studi di Padova

Authenticated
Download Date | 3/29/16 7:47 AM
 Lippi et al.: EFLM WG-PRE: local validation of blood collection tubes in clinical laboratories      3

results are especially valuable. The tubes/devices should Table 1: Essential requisites for purchasing blood collection devices.
also be analyzed by the manufacturer in different clinical
settings, on major instrument platforms and investigated 1. Components of the blood collection system in use (i.e. safety
needle, butterfly needle, holder and blood tube) should be
for potential test result bias and/or imprecision of ana-
produced by the same manufacturer or else the combination/
lytes especially where this may be important for clinical integration of separate parts should be validated by accredited
decision making, including serum indices and analytes regulation organizations such as the European Community (EC)
with known instability over time. The reasons for failures or the US Food and Drug Administration (FDA).
including missing and excluded data should be stated. 2. Manufacturers should demonstrate performance studies of their
products following study subject recruitment.
For new suppliers evidence of factory capacity over time
3. Manufacturers should demonstrate ease of use, sustained
should also be supplied. The EFLM WG-PRE does however, plasma quality, perceived value compared to the comparative
also recommend that a laboratory performs a local valida- devices on the market and the risk associated with the use of
tion of all new blood collection tubes (i.e. control tubes) their product.
estimating the potential bias and imprecision of test 4. Failure rates per 10,000 tubes should be stated for each tube
results compared to the previously used material (i.e. type. The reasons for failures, including missing and excluded
data, should also be stated during usability studies.
comparative tubes) to verify the manufactures claims.
5. New suppliers should provide evidence of capacity to produce
This approach has been proven by the constantly growing the product over longer period of time (at least 2 years).
number of studies in all areas of diagnostic testing [9–12], 6. The cost for the appropriate number of tubes and reagents
including molecular biology [13, 14]. for local validation should be charged to the manufacturers
The costs attributable to using an appropriate participating in the tender.
7. The validation study should be submitted for Ethical committee
number of tubes for a local validation (as described in
and/or Institutional review board approval.
the following parts of this article) should be charged 8. The committee for a tender for blood collection devices should
to the manufacturers participating in the tender. More always include not less than one laboratory professional.
specifically, the details of the validation process should
be included in the tender specification, with a specific
request to the manufacturers to supplement the labora- manufacturer claims about structure, assembly, function-
tory with a number of tubes and cost of reagents that is ality and safety of the new (i.e. “control”) blood collection
sufficient to complete each part of the validation. Con- tubes are fulfilled, as verified by using local practices.
sideration should also be given to the process of sub- Preferably, the sample size should include not  < 240 blood
mitting the validation to an Ethical committee and/or collections randomized to both the control (n = 120) and
Institutional review board for approval (as for CLSI rec- the comparative (n = 120) blood tubes, as recommended by
ommendations) [5]. Last but not least, the commission of the CLSI guidelines EP28-A3 to meet the minimum require-
a tender for purchasing blood collection devices should ments for reliability and usefulness [5]. As an alternative,
always include not less than one laboratory professional the collection of two paired tubes from the same patient
among the members (Table 1). with the two different systems may be advisable for a more
stringent comparison, although not strictly necessary for
this technical validation. Patients who are difficult to bleed
Validation of blood tubes should be excluded, as they may skew the data. For the
technical validation, the EFLM WG-PRE supports recording
In the following sections of this article, the EFLM WG-PRE the following information:
suggest a consensus protocol and some pertinent indica- 1. Tubes with physical defects of manufacturing (calcu-
tors that may be used for the local validation, both tech- late percentage)
nical and clinical, of new (i.e. “control”) blood collection 2. Tubes with no vacuum or that fail to form a vacuum
tubes, to be compared with the current system in use by (calculate percentage)
the same laboratory (i.e. “comparative” blood tubes). 3. Tubes not properly fitting into the blood collection
device (calculate percentage)
4. Tubes under filling after blood collection (i.e. 10%
Local technical validation of blood collection lower than the nominal filling volume; calculate
tubes ­percentage) [15]
5. Leaking from tube caps (calculate percentage)
The local (user) technical validation of blood collec- 6. External surface contamination with blood at the end
tion tubes should be intended to verify whether the of venipuncture (calculate percentage)

Brought to you by | Universita degli Studi di Padova

Authenticated
Download Date | 3/29/16 7:47 AM
4      Lippi et al.: EFLM WG-PRE: local validation of blood collection tubes in clinical laboratories

7. Hemolyzed specimens, with significant hemolysis safety and quality problems should also be reported to
(e.g. 0.5 g/L) defined according to local practices (cal- the pertinent regulatory national or supranational agency
culate percentage) (i.e. FDA, EC or UK medicines and healthcare products
8. Undue clotting in (a) EDTA and (b) sodium citrate regulatory).
blood tubes (calculate percentage of undue clotting in The technical problems of blood tubes may also be
each type of blood tubes) investigated by means of objective approaches of risk
9. Tubes broken or spilling blood after manufacturer- analysis such as the Failure Mode and Effects Analysis
specific centrifugation (calculate percentage) (FMEA). This systematic technique was originally devel-
10. Inappropriate positioning of gel separator after manu- oped in the late 1950s to investigate problems emerging
facturer-specific centrifugation (calculate percentage) from military systems malfunctions. However, the FMEA
11. Serum tubes with incomplete clotting after manufac- approach may either be reliably used for identifying
turer-specific handling (i.e. time for clotting, centrifu- failure patterns of blood tubes, their causes and conse-
gation conditions; calculate percentage) quences, and registering the information in specific FMEA
worksheets [16]. Some previous experience in the field of
For the calculation of maximum allowable deviation, the preanalytical activities including blood collections have
EFLM WG-PRE is in support of estimating the percentage already been published [17].
of each indicator for 120 tubes of both the control and com-
parative blood tubes according to the formula reported in
Table 2. When the difference between the comparative and Local clinical validation of blood collection
control blood tubes is higher than the acceptability criteria tubes
consensually agreed by the EFLM WG-PRE for each of the
selected indicators (i.e. 1%), then consideration that the The local (user) clinical validation of blood collection
comparative blood tubes have failed to pass the valida- tubes should be intended to verify whether the new (i.e.
tion process should be raised. Importantly, the EC Direc- “control”) blood collection tubes may be a source of bias
tive 93/42/EEC appoints that any instrument, apparatus, in test results, as verified using local instrumentation and
appliance, material or other article, whether used alone or reagents. Therefore, the validation of new devices prior
in combination that is intended by the manufacturer to be to routine introduction should entail statistical analysis
used for human beings for the purpose of diagnosis, pre- of laboratory data obtained with the existing and locally
vention, monitoring, treatment or alleviation of disease, validated blood collection tubes. The sample size should
must be considered a medical device (MD). Therefore, include between 20 and 100 (the higher the better) paired
besides informing tube manufacturers about potential and sequential blood collections, using both the control
issues emerged during the validation process, the major and the comparative blood tube systems, by means of two

Table 2: Acceptability criteria for technical validation of new blood collection tubes.

Item   Acceptable difference

Tubes with physical defects of manufacturing    < 1%

Tubes with no vacuum or that fail to form a vacuum    < 1%
Tubes not properly fitting into the blood collection device    < 1%
Tubes under filling    < 1%
Tubes leaking from the cap before and after centrifugation    < 1%
Blood contamination of collection device    < 1%
Hemolyzed specimens    < 1%a
Undue clotting  
 EDTA blood tubes    < 1%
 Sodium citrate blood tubes    < 1%
Tubes broken or spilling blood after centrifugation    < 1%
Inappropriate positioning of gel separator    < 1%
Serum blood tubes with incomplete clotting    < 1%

number of comparative tubes  number of control tubes 

Difference:  ∗100−  ∗100. When causes other than the blood tube (e.g. blood
a
 120   120 
­collection device, phlebotomists, sample transportation or patient population) can be excluded.

Brought to you by | Universita degli Studi di Padova

Authenticated
Download Date | 3/29/16 7:47 AM
 Lippi et al.: EFLM WG-PRE: local validation of blood collection tubes in clinical laboratories      5

different venipunctures, preferably on the opposite arms, Although the CLSI document GP34-A is an useful tool
as recommended by the CLSI guidelines EP28-A3 to meet for verifying tubes for venous and capillary blood drawing,
the minimum requirements for reliability and usefulness the real impact of blood collection tubes on local quality
[18]. For the clinical validation, the EFLM WG-PRE sup- and safety of testing is often overlooked, and laboratory
ports the paired measurement of all laboratory param- professionals often fail to recognize the need to accurately
eters for which the comparative blood tubes are to be assess the reliability of new devices or perform continu-
implemented. ous monitoring of ongoing performance [25]. Due to the
For calculation of the maximum allowable deviation, fact that it is unfeasible for manufacturers to establish the
the EFLM WG-PRE recommends comparing and analyz- impact of their devices on all instruments and reagents,
ing results obtained with the two different tube systems the EFLM WG-PRE has drafted this consensus document
by Passing and Bablok regression (and/or Deming fit) with the aim of supporting laboratory professionals plan-
and Bland and Altman plots, using values obtained with ning blood collection tube tenders and validating the
the control blood tubes as reference. When the regression devices before routine implementation. The laboratory
is not acceptable and the mean percentage bias between professionals should also make sure that the tenders
the two blood tube systems is found to be greater than accurately and strictly define the responsibilities for vali-
the previously defined desirable quality specifications dation experiments and the potential consequences in
for bias for each of the analyte tested, then the EFLM the case the validation outcome shows that tubes due not
WG-PRE suggests that either (i) previous blood collection fulfill the expectations.
tube system is kept in use, or (ii) the laboratory imple-
ments new tubes, but modifies local reference ranges Author contributions: All the authors have accepted
for parameters for which there is a clinically significant responsibility for the entire content of this submitted
difference between old and new tubes. Quality specifi- manuscript and approved submission.
cations for validation experiments should be defined Research funding: None declared.
taking into consideration the Milan EFLM Strategic Employment or leadership: None declared.
conference hierarchy [19]. The final evaluation should Honorarium: None declared.
remain dependent upon the clinical decisions the results Competing interests: The funding organization(s) played
are used for, differences between health and disease and no role in the study design; in the collection, analysis, and
biological variation. interpretation of data; in the writing of the report; or in the
decision to submit the report for publication.

Conclusions
References
The validation of new laboratory equipment, instru-
mentation, methods and IVD devices by manufacturers 1. Lippi G, Banfi G, Church S, Cornes M, De Carli G, Grankvist K,
and local users is a necessary part of clinical laboratory et al. Preanalytical quality improvement. In pursuit of
accreditation. Failure to comply with good manufactur- ­harmony, on behalf of European Federation for Clinical Chem-
istry and Laboratory Medicine (EFLM) Working group for Prean-
ing or good laboratory practices may have adverse conse-
alytical Phase (WG-PRE). Clin Chem Lab Med 2015;53:357–70.
quences on operator and patient safety. Several lines of 2. Simundic AM, Lippi G. Preanalytical phase – a continuous
evidence now attest that the implementation of new blood challenge for laboratory professionals. Biochem Med (Zagreb)
collection devices, including blood tubes, may modify 2012;22:145–9.
local practices and also influence the measured concen- 3. Simundic AM, Cornes MP, Grankvist K, Lippi G, Nybo M, ­Ceriotti F,
tration of the analytes. Indeed, the selection and acquisi- et al. Colour coding for blood collection tube closures – a call for
harmonisation. Clin Chem Lab Med 2015;53:371–6.
tion of systems for blood collection should be considered
4. Stankovic AK, Silvestri J, Mails M, Najork C. Total quality in labo-
a critical aspect for assuring quality, safety and efficiency ratory diagnostics: the role of commercial companies. Biochem
of the preanalytical phase of laboratory diagnostics and, Med 2010;20:207–14.
therefore, of total testing process. This issue is expected 5. Clinical and Laboratory Standards Institute (CLSI). Validation and
to become even more important as long as innovative Verification of Tubes for Venous and Capillary Blood Specimen
Collection; Approved Guideline. Wayne, PA: Clinical and Labora-
molecular biomarkers make it through the translational
tory Standards Institute, 2010. CLSI document GP34-A.
process, and are introduced into routine clinical practice 6. International Organization for Standardization. Quality manage-
[20, 21], as this emerging arena is particularly vulnerable ment systems – Fundamentals and vocabulary. ISO 9000:2005.
to preanalytical issues [22–24]. ISO, Geneva, Switzerland, 2005.

Brought to you by | Universita degli Studi di Padova

Authenticated
Download Date | 3/29/16 7:47 AM
6      Lippi et al.: EFLM WG-PRE: local validation of blood collection tubes in clinical laboratories

7. Favaloro EJ, Plebani M, Lippi G. Regulation of in vitro diagnos- 16. Plebani M, Lippi G. Closing the brain-to-brain loop in laboratory
tics (IVDs) for use in clinical diagnostic laboratories: towards testing. Clin Chem Lab Med 2011;49:1131–3.
the light or dark in clinical laboratory testing? Clin Chem Lab 17. Jiang Y, Jiang H, Ding S, Liu Q. Application of failure mode and
Med 2011;49:1965–73. effects analysis in a clinical chemistry laboratory. Clin Chim Acta
8. Plebani M, Caputo M, Giavarina D, Lippi G. Methodological 2015 25;448:80–5.
notes on acquisition and use of close evacuated systems for 18. Clinical and Laboratory Standards Institute (CLSI). Defining,
collection, handling and storage of venous blood samples for Establishing, and Verifying Reference Intervals in the ­Clinical
laboratory diagnostics. Biochim Clin 2013;37:244–52. Laboratory; Approved Guideline – Third Edition; Approved
9. Lima-Oliveira G, Lippi G, Salvagno GL, Montagnana M, Guideline. Wayne, PA: Clinical and Laboratory Standards
Picheth G, Guidi GC. Preanalytical management: serum vacuum ­Institute; 2010. CLSI document CLSI EP28-A3.
tubes validation for routine clinical chemistry. Biochem Med 19. Sandberg S, Fraser CG, Horvath AR, Jansen R, Jones G,
(Zagreb) 2012;22:180–6. ­Oosterhuis W, et al. Defining analytical performance specifica-
10. Lima-Oliveira G, Lippi G, Salvagno GL, Montagnana M, tions: Consensus Statement from the 1st Strategic Conference
Picheth G, Guidi GC. Sodium citrate vacuum tubes validation: of the European Federation of Clinical Chemistry and Laboratory
preventing preanalytical variability in routine coagulation test- Medicine. Clin Chem Lab Med 2015;53:833–5.
ing. Blood Coagul Fibrinolysis 2013;24:252–5. 20. Yao NA. Health Services & Policy Research in translational medi-
11. Lima-Oliveira G, Lippi G, Salvagno GL, Montagnana M, Poli G, cine. Ann Transl Med 2014;2:72.
Solero GP, et al. K(3)EDTA Vacuum Tubes Validation for Routine 21. Lippi G, Plebani M. Personalized medicine: moving
Hematological Testing. ISRN Hematol 2012;2012:875357. from ­simple theory to daily practice. Clin Chem Lab Med
12. Zungun C, Yılmaz FM, Boru EG, Topcuoglu C. Comparison of 2015;53:959–60.
Improvacuter™ tubes with BD Vacutainer™ tubes for v­ arious 22. Tuck MK, Chan DW, Chia D, Godwin AK, Grizzle WE, Krueger KE
hormones in the aspects of stability and influence of gel et al. Standard operating procedures for serum and plasma
­separators. Clin Chem Lab Med 2015;53:231–8. collection: early detection research network consensus state-
13. Häntzsch M, Tolios A, Beutner F, Nagel D, Thiery J, Teupser D, ment standard operating procedure integration working group.
et al. Comparison of whole blood RNA preservation tubes and J Proteome Res 2009;8:113–7.
novel generation RNA extraction kits for analysis of mRNA and 23. El Messaoudi S, Rolet F, Mouliere F, Thierry AR. Circulating
MiRNA profiles. PLoS One 2014;9:e113298. cell free DNA: Preanalytical considerations. Clin Chim Acta
14. Zhang H, Korenková V, Sjöback R, Švec D, Björkman J, Kruhøffer M, 2013;424:222–30.
et al. Biomarkers for monitoring pre-analytical quality variation of 24. Channavajjhala SK, Rossato M, Morandini F, Castagna A,
mRNA in blood samples. PLoS One 2014;9:e111644. Pizzolo F, Bazzoni F, et al. Optimizing the purification and
15. Clinical and Laboratory Standards Institute (CLSI). Tubes and analysis of miRNAs from urinary exosomes. Clin Chem Lab Med
Additives for Venous and Capillary Blood Specimen Collec- 2014;52:345–54.
tion; Approved Standard – Sixth Edition; Approved Guideline. 25. Bowen RA, Remaley AT. Interferences from blood collection
Wayne, PA: Clinical and Laboratory Standards Institute; 2010. tube components on clinical chemistry assays. Biochem Med
CLSI document GP39-A6. (Zagreb) 2014;24:31–44.

Brought to you by | Universita degli Studi di Padova

Authenticated
Download Date | 3/29/16 7:47 AM