Point-of-care laboratory halves door-to-therapy-decision time in acute stroke.

BRIEF COMMUNICATIONS
Point-of-Care Laboratory
Halves Door-to-TherapyDecision Time in Acute Stroke
Silke Walter, MD,1 Panagiotis Kostopoulos, MD,1
Anton Haass, MD,1 Martin Lesmeister,1
Mihaela Grasu,1 Iris Grunwald, MD,2
Isabel Keller, MD,1 Stephan Helwig,1
Carmen Becker, MD,1 Juergen Geisel, MD,3
Thomas Bertsch, MD,4 Sarah Kaffiné,1
Annika Leingärtner,1 Panagiotis Papanagiotou,
MD,5 Christian Roth, MD,5 Yang Liu, MD,1
Wolfgang Reith, MD,5 and Klaus Fassbender, MD1
Currently, stroke laboratory examinations are usually
performed in the centralized hospital laboratory, but often planned thrombolysis is given before all results are
available, to minimize delay. In this study, we examined
the feasibility of gaining valuable time by transferring
the complete stroke laboratory workup required by
stroke guidelines to a point-of-care laboratory system,
that is, placed at a stroke treatment room contiguous to
the computed tomography, where the patients are
admitted and where they obtain neurological, laboratory, and imaging examinations and treatment by the
same dedicated team. Our results showed that reconfiguration of the entire stroke laboratory analysis to a
point-of-care system was feasible for 200 consecutively
admitted patients. This strategy reduced the door-totherapy-decision times from 84 6 26 to 40 6 24 min (p
< 0.001). Results of most laboratory tests (except activated partial thromboplastin time and international normalized ratio) revealed close agreement with results
from a standard centralized hospital laboratory. These
findings may offer a new solution for the integration of
laboratory workup into routine hyperacute stroke management.
ANN NEUROL 2011;69:581–586
S
troke is a major cause for chronic disability and
death.1 Recanalization of occluded arteries by systemic
thrombolysis with recombinant tissue plasminogen activator (rt-PA) within 3 hours after onset of ischemic stroke is
the only effective and approved therapy.2–4 However,
because time-consuming diagnostic workup is required
before administration of rt-PA,3–6 implementation of rt-PA
therapy is difficult to achieve, and <2% of eligible stroke
patients receive this treatment in most countries.7
A major obstacle in hyperacute stroke management is
performance of laboratory examinations in a centralized hos-
pital laboratory (CL). Current guidelines for thrombolysis
recommend that specific laboratory tests (ie, platelet count,
leukocyte count, erythrocyte count, hemoglobin, activated
partial thromboplastin time [aPTT], international normalized ratio [INR], c-glutamyltransferase, p-amylase, and glucose) be performed in patients with suspected stroke to identify conditions that mimic stroke or that limit therapeutic
options.3,4 However, in actual practice, results of this timeconsuming diagnostic procedure are often not awaited.
Indeed, major guidelines even recommend that ‘‘although it
is desirable to know the results of these tests before giving rtPA, thrombolytic therapy should not be delayed while awaiting the results unless (1) there is clinical suspicion of a bleeding abnormality or thrombocytopenia, (2) the patient has
received heparin or warfarin or (3) use of anticoagulants is
not known.’’3 This recommendation might lead to an
increased risk of overlooking stroke mimics or patients with
contraindications for thrombolysis.
A point-of-care (POC) platform consists of mobile
laboratory devices that are, in contrast to the CL, located
directly at the site where the patients are treated. The POC
laboratory tests are performed by the same personnel who
treat the patient, thus potentially reducing interface times
and examination times. To bridge the gap between time
pressure issues and safety concerns in hyperacute stroke
management, we asked whether the whole stroke laboratory
workup could be transferred to a POC system and whether
this strategy might shorten the critical time until therapy
decision.
Subjects and Methods
Subjects and Study Design
In this prospective monocenter study, 200 patients (93 females,
107 males), aged between 26 and 91 (median, 73) years
From the 1Department of Neurology, University of the Saarland,
Homburg, Germany; 2Biomedical Research Centre, Acute Vascular
Imaging Centre, John Radcliffe Hospital, Oxford, England; 3Department
of Clinical Chemistry, University of the Saarland, Homburg, Germany;
4
Department of Clinical Chemistry, Nürnberg Hospital, Nürnberg,
Germany; 5Department of Neuroradiology, University of the Saarland,
Homburg, Germany.
Address correspondence to Dr Fassbender, Department of Neurology,
Saarland University, Kirrberger Str. D-66421 Homburg, Germany.
E-mail: [email protected]
Received Jun 15, 2010, and in revised form Nov 23, 2010. Accepted for
publication Dec 6, 2010.
View this article online at wileyonlinelibrary.com. DOI: 10.1002/
ana.22355
C 2011 American Neurological Association
V
581
ANNALS
of Neurology
admitted between May 2009 and April 2010 to the Department of Neurology of the University Clinic of the Saarland
were evaluated in regard to determinants of acute stroke diagnostic workup. Inclusion criteria for study participants were (1)
differential diagnosis of acute stroke and (2) potential eligibility
for recanalizing therapies, based on time 6 hours from onset
of symptoms. Decisions for thrombolysis in the prospective
study group were based on POC results within the reference
ranges (in addition to the current guidelines), and in the case
of abnormal POC values, on awaited CL laboratory results.
Further, a historical control group, admitted to the same
department and consisting of 200 patients (98 females, 102
males), aged between 18 and 99 (median, 75) years treated
from March 2006 to December 2007 was retrospectively analyzed in regard to the same variables. Informed consent was
obtained from the patients, and the study was approved by the
Regional Ethics Committee of the Medical Association of the
Saarland, Germany.
Study Variables
The primary variable used to compare POC-based stroke management with CL-based stroke management was door-to-therapy-decision time, defined as the time between arrival at the
hospital until the end of diagnostic procedures relevant for therapy decision (computed tomography [CT], laboratory examination). Other variables that were assessed included door-to-startof-laboratory-analysis time, door-to-end-of-centrifugation time,
door-to-analyses-end-and-result-transmission time, the final
diagnoses and the number of thrombolyses. Times for ‘‘door’’
were determined from reports of the emergency team and from
hospital records. The studied times for laboratory and CT performance were retrieved from the hospital information system.
Blood Sampling and Asservation
Blood was drawn in ethylenediaminetetraacetic acid-containing
tubes (Sarstedt, Nümbrecht, Germany) for hematological analysis and in heparin-containing tubes (Sarstedt) for clinical chemistry analysis. For coagulation assays, blood was drawn in sodium citrate-containing tubes (Sarstedt). Blood samples were
drawn simultaneously for POC- and CL- based laboratory analyses, and start of POC analysis and transfer of blood samples
to CL (via a fast pneumatic air tube system) occurred in
parallel.
POC Laboratory Analysis
Hematological variables (platelet count, leukocyte count, erythrocyte count, hemoglobin) were measured with the PocH-100i
analyzer (Sysmex, Hamburg, Germany). Clinical chemistry parameters (c-glutamyltransferase, p-amylase, and glucose) were
measured with the Reflotron plus analyzer (Roche Diagnostics,
Mannheim, Germany). Coagulation parameters (aPTT, INR)
were performed on the Hemochron Junior (ITC, Edison, NJ).
Quality control was carried out as recommended by the manufacturer once per week with 3 (PocH-100i) or 2 (Reflotron,
Hemochron) different levels of control material. Additionally,
daily controls with the clean and check control strips (Reflo-
582
tron) or daily temperature and electronic quality controls with
control cartridges (Hemochron) were performed. External quality control determinations were performed every 3 months.
CL Laboratory Analysis
Hematological variables were measured on the Sysmex XE5000 analyzer (Norderstedt, Germany). Clinical chemistry parameters were determined according to the manufacturer’s
instructions (Roche) on the P 800 module of the Modular Analytic analyzer. Coagulation parameters were assayed with the
BCS hemostasis testing system by Siemens (Eschborn, Germany). Innovin (Siemens) was used for measurement of INR
and Actin FS (Siemens) for aPPT. Quality control procedures
were performed according to the guidelines for medical laboratory diagnostics of the Federal Chamber of Physicians of
Germany.
Statistics
The agreement between POC- and CL-based results was
assessed using Bland-Altman analysis,8 for which the paired differences are plotted against the paired means of POC and CL
results. The limits of agreement represent the 95% reference
range of comparable measurements and are defined as 61.96
standard deviation. The paired t test and the t test for independent samples were used.
Results
Setting of POC-Based Stroke Laboratory
Workup and Clinical Characteristics
We placed the POC based laboratory workup at a stroke
treatment room directly at the site of the CT, where the
patients were admitted; obtained neurological, laboratory,
and imaging examinations; and, if indicated, were treated
by the same team. Diagnostic workup confirmed the initial differential diagnosis of acute stroke in 126 patients.
Of these, 6 patients had hemorrhagic stroke, and 120
patients had ischemic stroke (large-artery atherosclerosis,
44%; cardioembolic stroke, 29%; lacunar stroke, 22%;
unknown or other, 5%). Thirty-two patients with ischemic stroke received thrombolysis, with door-to-needle
times of 59 6 37 minutes.
The profile of the prospective patient group was
similar to that of the historical control group, consisting
of 91 patients with cerebral ischemia (large-artery atherosclerosis, 33%; cardioembolic, 40%; lacunar, 25%;
unknown or other, 2%). Thirteen patients had hemorrhagic stroke. Sixteen of the ischemic stroke patients
received thrombolysis.
Relevance of POC Laboratory for Therapy
Decisions
In the 120 patients with acute cerebral ischemia, 64
POC laboratory abnormalities were found. Abnormalities
consisted of values outside the normal range that were
Volume 69, No. 3
Walter et al: Point-of-Care Laboratory
not contraindications for immediate POC-based thrombolysis (limit range values, n ¼ 37), and values that represented absolute contraindications (n ¼ 27). In the latter case, CL results had to be awaited as the gold
standard. Of the 64 POC laboratory abnormalities, 38
affected coagulation parameters. Nineteen of these 38
values exceeded the cutoff values used for exclusion from
immediate POC-based thrombolysis (absolute contraindications). Of these exclusionary abnormalities, 8 were confirmed by CL. The remaining 26 abnormalities affected
noncoagulation parameters. Eight abnormal values
exceeded the cutoff values for exclusion from immediate
POC-based thrombolysis, and 6 of these exclusions were
confirmed by CL.
In terms of patients, 59 of the 120 stroke patients
had abnormal POC values. Of these, 32 patients
showed limit range laboratory abnormalities, and 27
had absolute laboratory contraindications for immediate
thrombolysis. Of the 27 patients with POC-based absolute contraindications, 14 were confirmed by CL, and
13 could not be confirmed. Of the latter group, 2
patients received thrombolysis, and the remaining 11
patients were not treated due to other conventional
exclusion criteria. Thus, POC identified 93 patients
(88%) as possible thrombolysis candidates. Importantly,
POC did not miss any patient with absolute laboratory
thrombolysis contraindications. Considering the 32
patients who actually received thrombolysis, 30 (94%)
were initially identified as thrombolysis candidates by
POC, with the consequence of immediate thrombolysis.
Only 2 of the treated patients obtained thrombolysis
after awaiting CL.
Acceleration of Door-to-Therapy-Decision Times
by POC-Based Stroke Laboratory Workup
Figure 1 shows the results of the door-to-decision times,
subdivided into relevant laboratory examination management intervals (door-to-start-of-laboratory-analyses time,
door-to-end-of-centrifugation time, door-to-analyses-endand-result-transmission time). Transfer of laboratory analyses from a CL to a POC system resulted in an approximately 50% reduction in the door-to-therapy-decision
times (84 6 26 vs 40 6 24 minutes, respectively; p <
0.001). The door-to-start-of-laboratory-analyses times
were lower in the POC system compared to the CL system (21 6 18 vs 32 6 22 minutes; p < 0.001), as were
laboratory analysis times (11.7 6 5.1 vs 51 6 15
minutes; p < 0.001) and door-to-analyses-end-andresult-transmission times (33 6 19 vs 83 6 25 minutes;
p < 0.001).
There were no significant differences between the
prospective and historical study groups in door-to-therMarch 2011
FIGURE 1: Comparison of delay of acute stroke management caused by point-of-care (POC)-based and centralized
hospital laboratory (CL)-based laboratory workup. The timing of laboratory management of a historical control group
(CL hist.) is also presented. The door-to-therapy-decision
time is broken down into the following subintervals: doorto-start-of-laboratory-analysis (black bars), until end-of-centrifugation (grey bars), until analysis-end-and-result-transmission (bars with diagonal lines), and until therapy decision
(white bars).
apy-decision times (84 6 26 vs 86 6 54 minutes; nonsignificant [NS]), laboratory analysis times (51 6 15 vs
48 6 19 minutes; NS), or door-to-analyses-end-andresult-transmission times (83 6 25 vs 80 6 48 minutes;
NS). The door-to-start-of-laboratory-analyses times with
CL-based diagnostic workup were higher in the prospective study group compared to the historical study group
(39 6 26 vs 32 6 22 minutes; p ¼ 0.022).
Agreement of POC-Based Laboratory Workup
with CL Results
Spearman correlation and Bland-Altman analysis showed
a close agreement between laboratory results obtained by
POC- or CL-based laboratory analysis, that is, platelet
count, leukocyte count, erythrocyte count, hemoglobin,
c-glutamyltransferase, p-amylase, and glucose (Fig 2). In
contrast, agreement was lower for aPTT and INR,
although false-negative rates for both variables within the
normal range (aPTT, <42; INR, <1.5) were 0%. The
sensitivity and specificity of aPTT were 100.0 and
61.9%, respectively and of INR, 100.0 and 98.4 %,
respectively.
Discussion
Although laboratory workup is generally considered an
integral part of acute stroke management, in clinical
practice the results are often not awaited to avoid delay
of planned thrombolysis. Indeed, major guidelines
583
ANNALS
of Neurology
FIGURE 2: Bland-Altman analysis of the agreement between results derived from point-of-care (POC) laboratory and from
standard centralized hospital laboratory (CL). Means and standard deviations of percentage pair-wise differences were 1 6
11% (platelet count), 4.8 6 9.2% (leukocyte count), 3.4 6 8.5% (erythrocyte count), 2.0 6 7.4% (hemoglobin), 52 6 30% (activated partial thromboplastin time [aPTT]), -5 6 21% (international normalized ratio [INR]), 23% 6 65% (c-glutamyltransferase
[c-GT]), 14 6 33% (p-amylase), and 25 6 14% (glucose).
recommend not waiting for laboratory test results under
such time pressure.3 This study, however, shows that
transfer on a POC system markedly decreased delay by
stroke laboratory workup, offering a solution to the
problem of routinely integrating laboratory analysis in
hyperacute stroke management.
In the past few years, we9,10 and others11–13 have proposed the use of POC as a strategy to accelerate acute stroke
management. Green et al. were the first to describe POCbased INR analysis in stroke workup.11 Here we show the
feasibility of transferring the entire laboratory diagnostic
workup requested by current stroke guidelines to a POC platform, that is, placed at the site of the CT, where the patients
584
were admitted; obtained neurological, laboratory, and imaging examinations; and were treated by the same team.
Complete transfer of laboratory workup to a POC
platform approximately halved the door-to-therapy-decision time. The comparison with the historical control
group shows that there were no major differences in
regard to the stroke management times before and after
the trial, arguing for a main role of the POC method
rather than unrelated factors in reduction of door-totherapy-decision times. Use of a POC platform offers
possibilities to accelerate stroke management not only
directly, by reduction of times for transport, analyses, or
transmission of results, but also indirectly, by increased
Volume 69, No. 3
Walter et al: Point-of-Care Laboratory
efficiency in interfaces between different health care professionals. The resulting door-to-therapy-decision times
by use of POC were below those encountered in daily
clinical practice (>60 minutes).14 Pathophysiological and
clinical evidence suggests that saved time will likely translate to better outcomes.15–17
Results obtained by POC and by CL showed close
agreement, with the exception of aPTT and INR, as their
quantification revealed room for improvement. The reasons for the discrepancy between these 2 methods and
the relatively high variation of coagulation values, despite
their being already on the market, are unclear and have
been described before, that is, in acutely ill patients.18
Potential reasons that could explain our observations
include differences in techniques of detection, reagents,
and calibrators and lack of international standardization,
as well as different methods of asservation, longer transport times to the CL, and differences in processing methods, such as centrifugation. Newer and more accurate
POC devices are reaching the market, thereby further
increasing the relevance of this concept of completely
POC-based stroke management.
Although we cannot rule out the possibility that
the POC values were the more precise values, CL-based
analysis is currently considered as standard, as this is still
the generally used laboratory method.
However, even at present, normal values of aPTT
and INR are useful, as their false-negative rates were 0%.
In contrast, POC-based aPTT and INR values in the
pathological ranges should at the present technical state
be confirmed by CL. Finally, because maintenance and
quality management of POC devices is not less complex
than that of CL devices, supervision by professionals
trained in clinical chemistry is important to ensure constant quality.
The laboratory analysis time of 51 minutes for the
CL seems to be very long, until one considers that it is
composed of 2 parts: first, time for registration, centrifugation, analyses, and validation; and second, time for
electronic data processing and transmission. Times
needed for the first component are approximately 10
minutes for hematology, 30 to 40 minutes for coagulation tests, and 35 to 45 minutes for enzyme testing. The
times for electronic data processing and transmission
were approximately 15 minutes. Times also vary under
real world conditions, for example, in relation to the
actual workload.
Fewer hemorrhagic strokes than expected were seen,
and the reasons are unclear. A contributing factor was
the admission of at least 9 patients with hemorrhagic
stroke (within the 6-hour temporal window) with acute
March 2011
deterioration and/or ventilation directly admitted to the
intensive care unit of our department, bypassing the
emergency ward, where the study took place.
In summary, despite the limitations associated with a
single-center study, we showed for the first time that consequent transfer of entire stroke laboratory workup to a
POC system is feasible in clinical practice and contributes
to a halving of the door-to-therapy-decision times. After
technical improvement in quantification, that is, of aPTT
and INR, and after further validation, POC-based stroke
laboratory analysis may offer a solution to the problem of
how to perform necessary laboratory assessments within a
critical time frame. In the future, a POC system placed at
the stroke treatment room could be an effective strategy in
routine management of hyperacute stroke patients.
Acknowledgments
This study was funded by grants from the Ministry of
Health of the Saarland, Germany; Jackstädt-Foundation,
Germany; and Else Kröner-Fresenius Foundation.
Authorship
S.W., P.K., and A.H. contributed equally to this study.
Potential Conflicts of Interest
Nothing to report.
References
1.
Rothwell PM, Coull AJ, Siver LE, et al. Population-based study of
event-rate, incidence, case fatality, and morality for all acute
vascular events in all arterial territories (Oxford Vascular Study).
Lancet 2005;366:1773–1783.
2.
National Institute of Neurological Disorders and Stroke rt-PA
Stroke Study Group. Tissue plasminogen activator for acute ischemic stroke. N Engl J Med 1995;333:1581–1587.
3.
Adams HP Jr, del Zoppo G, Alberts MJ, et al. Guidelines for the
early management of adults with ischemic stroke: a guideline from
the American Heart Association/American Stroke Association
Stroke Council, Clinical Cardiology Council, Cardiovascular Radiology and Intervention Council, and the Atherosclerotic Peripheral
Vascular Disease and Quality of Care Outcomes in Research Interdisciplinary Working Groups: The American Academy of Neurology affirms the value of this guideline as an educational tool for
neurologists. Circulation 2007;115:e478–e534.
4.
European Stroke Organisation (ESO) Executive Committee; ESO
Writing Committee. Guidelines for management of ischaemic
stroke and transient ischaemic attack 2008. Cerebrovasc Dis 2008;
25:457–507.
5.
Broderick J, Connolly S, Feldmann E, et al. Guidelines for the
Management of Spontaneous Intracerebral Hemorrhage in Adults,
2007 Update: a guideline from the American Heart Association/
American Stroke Association Stroke Council, High Blood Pressure
Research Council, and the Quality of Care and Outcomes in
Research Interdisciplinary Working Group. The American Academy
of Neurology affirms the value of this guideline as an educational
tool for neurologists. Circulation 2007;116;e391–e413.
585
ANNALS
of Neurology
6.
Boehringer, Summary of Product Characteristics http://www.medicines.org.uk/emc/medicine/308/SPC/Actilyse/ last updated on the
eMC: December 4, 2009.
7.
Albers GW, Olivot JM. Intravenous altepase for ischemic stroke.
Lancet 2007;369:249–250.
8.
Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet
1986;1:307–310.
9.
Fassbender K, Walter S, Liu Y, et al. ‘‘Mobile stroke unit’’ for
hyperacute stroke treatment. Stroke 2003;34:e44.
10.
Walter S, Kostpopoulos P, Haass A, et al. Bringing the hospital to
the patient: first treatment of stroke patients at the emergency
site. PLoS One 2010;29;5:e13758.
11.
Green TL, Mansoor A, Newcommon N, et al. Reliability of pointof-care testing of INR in acute stroke. Can J Neurol Sci 2008;35:
348–351.
12.
Rizos T, Herweh C, Jenetzky E, et al. Point-of-Care International
Normalized Ratio Testing accelerates thrombolysis in patients with
acute ischemic stroke using oral anticoagulants. Stroke 2009;40:
3547–3551.
13.
Rizos T, Jenetzky E, Herweh C, et al. Point-of-care reversal treatment in phenprocoumon-related intracerebral hemorrhage. Ann
Neurol 2010;67:788–793.
14.
Price CI, Clement F, Gray J, et al. Systematic review of stroke
thrombolysis service configuration. Expert Rev Neurother 2009;9:
211–233.
15.
Hossmann KA. Viability thresholds and the penumbra of focal ischemia. Ann Neurol 1994;36:557–565.
16.
Marler JR, Tilley BC, Lu M, et al. Early stroke treatment associated
with better outcome: the NINDS rt-PA study. Neurology 2000;55:
1649–1655.
17.
Lees KR, Bluhmki E, von Kummer R, et al. Time to treatment with
intravenous alteplase and outcome in stroke: an updated pooled
analysis of ECASS, ATLANTIS, NINDS, and EPITHET trials. Lancet
2010;375:1695–1703.
18.
Ferring M, Reber G, de Moerloose P, et al. Point of care and central laboratory determinations of the aPTT are not interchangeable
in surgical intensive care patients. Can J Anaesth 2001;48:
1155–1160.
Bevacizumab Does Not
Increase the Risk of Remote
Relapse in Malignant Glioma
Antje Wick, MD,1 Nils Dörner, MD,2
Navina Schäfer, CandMed,1 Silvia Hofer, MD,3
Sabine Heiland, PhD,2,4
Daniela Schemmer, RN,1 Michael Platten, MD,1
Michael Weller, MD,3 Martin Bendszus, MD,2
and Wolfgang Wick, MD1
Preclinical evidence and uncontrolled clinical studies
suggest an increased risk for distant spread and development of a gliomatosislike phenotype at recurrence or
progression of malignant glioma patients treated with
bevacizumab (BEV), an antibody to vascular endothelial
growth factor (VEGF). Here we asked whether BEV
treatment of recurrent malignant glioma increases the
risk of distant or diffuse tumor spread at further recurrence. BEV-treated patients were compared with
matched pairs of patients treated without anti-VEGF regimens. T1 contrast-enhanced (T1þc) and fluid-attenuated
inversion recovery (FLAIR) images were analyzed using a
novel automated tool of image analysis. At the start of
the study, 20.5% of BEV-treated and 22.7% of non–BEVtreated patients had displayed distant or diffuse recurrence. Distant or diffuse recurrences were observed in
22% (BEV) and 18% (non-BEV) on T1þc and in 25% and
18% on FLAIR (p > 0.05). The correlation between
changes on T1þc and FLAIR at progression was high.
The risk of distant or diffuse recurrence at the time of failure of BEV-containing treatments was not higher than
with anti-VEGF–free regimens, arguing against a specific
property of BEV that promotes distant tumor growth or
a gliomatosislike phenotype at recurrence.
ANN NEUROL 2011;69:586–592
wo uncontrolled phase II studies1,2 were the basis for
the approval of the vascular endothelial growth factor
(VEGF) antibody bevacizumab (BEV) for patients with
T
From the 1 Department of Neuro-oncology and 2 Department of
Neuroradiology, University Clinic Heidelberg, Heidelberg, Germany;
3
Department of Neurology, University Hospital Zurich, Zurich,
Switzerland; and 4Division Experimental Radiology, University Clinic
Heidelberg, Heidelberg, Germany.
Address correspondence to Dr W. Wick, Department of Neuro-oncology,
Neurology Clinic and National Center for Tumor Diseases, University of
Heidelberg, Im Neuenheimer Feld 400, D-69120 Heidelberg, Germany.
E-mail: [email protected]
Additional Supporting Information can be found in the online version of
this article.
Received Sep 8, 2010, and in revised form Oct 26, 2010. Accepted for
publication Nov 5, 2010.
View this article online at wileyonlinelibrary.com. DOI: 10.1002/
ana.22336
586
Volume 69, No. 3