Available online at www.sciencedirect.

com

European Journal of Pain 12 (2008) 4852

www.EuropeanJournalPain.com

Diagnostic criteria and follow-up parameters in complex regional

pain syndrome type I a Delphi survey
Florian Brunner a, Stephanie B. Lienhardt b, Rudolf O. Kissling
Lucas M. Bachmann b, Ulrich Weber a
b

a,*

a
Department of Physical Medicine and Rheumatology, Balgrist University Hospital, Forchstrasse 340, 8008 Zurich, Switzerland
Horten Centre for Patient Oriented Research, University of Zurich, Bolleystrasse 40, Postfach Nord, CH-8091 Zurich, Switzerland

Received 12 October 2006; received in revised form 1 February 2007; accepted 20 February 2007
Available online 2 April 2007

Abstract
Background: Although the current clinical guideline of diagnostic criteria for the complex regional pain syndrome I (CRPS I) is a
landmark endeavour to dene this complex condition it does not prioritise its most important clinical manifestations.
Aim: We set out to obtain an expert agreed priority list of diagnostic and follow-up parameters in the diagnosis and management of
CRPS I.
Methods: A two round Delphi survey: We asked international experts to list (rst round) and weight (second round) parameters
(scale 110) they believed to be relevant in diagnosis and follow-up. Median ratings and interquartile ranges (IQR) were calculated.
Rates P7 and IQR 63 depicted important and expert agreed parameters.
Results: Thirty-two diagnostic and 23 follow-up listings and ratings of 13 experts were available for analysis. In three domains (clinical presentation, further examinations and follow-up) experts agreed on the following parameters, pain (10; 910) with its subcategories hyperesthesia (7; 58) hyperalgesia (8; 88) and allodynia (8; 710), signs with oedema (9; 810) and colour change (8; 58)
and mobility with its categories motor change (7; 58) and decreased range of motion (8; 88). The experts agreed that no further
examinations were necessary for diagnosis (10; 810). The agreed important follow-up parameter was clinical course (10; 810) with
its categories decrease in pain (8; 89) and hyperalgesia (8; 68), decreased oedema (8; 710) and improvements in motor function
(10; 810) and strength (8; 69).
Conclusion: This expert survey conveys an agreed set of relevant diagnostic parameters of CRPS I and proposes that in follow-up
examinations treatment success should be based on restoration of those manifestations.
2007 European Federation of Chapters of the International Association for the Study of Pain. Published by Elsevier Ltd. All
rights reserved.
Keywords: Complex regional pain syndrome I; Diagnosis; Classication; Prognosis; Consensus

1. Introduction
The complex regional pain syndrome type I (CRPS I)
describes a challenging neuropathic pain state (StantonHicks et al., 1998; Janig and Baron, 2002). From a clinical
*

Corresponding author. Tel.: +41 44 386 35 08; fax: +41 44 386 35

09.
E-mail address: rudolf.kissling@balgrist.ch (R.O. Kissling).

perspective, CRPS I commonly presents with regional

pain and sensory changes following a noxious event (Stanton-Hicks et al., 1995). Along with the pain, abnormal skin
colour, temperature change, abnormal sudomotor activity
or oedema may also be present (Stanton-Hicks et al.,
1995).
Because of its broad clinical spectrum of manifestations, the International Association for the Study of
Pain (IASP) introduced a new set of diagnostic criteria

1090-3801/$32 2007 European Federation of Chapters of the International Association for the Study of Pain. Published by Elsevier Ltd. All rights
reserved.
doi:10.1016/j.ejpain.2007.02.003

F. Brunner et al. / European Journal of Pain 12 (2008) 4852

for CRPS I more than ten years ago (Merskey and Bogduk, 1994). However, according to a recent study a wide
variety of diagnostic criteria are still applied in CRPS I
research and only a minority of studies refer to the criteria as proposed by the IASP (van de Beek et al.,
2002). Many studies apply ad hoc clinical criteria, which
makes the results of those studies dicult to compare
(van de Beek et al., 2002).
We hypothesised, that the current diagnostic criteria
are under used, because the IASP criteria are clinically
dicult to adopt. Although, the guideline meticulously
denes the complex condition, it does not prioritise the
most important clinical parameters. Therefore, in this
study, we aimed at obtaining an expert agreed priority
list of diagnostic and follow-up parameters in the diagnosis and management of CRPS I using a two-round
Delphi survey among an international group of experts.

2. Methods
We conducted a two-round postal Delphi survey. At
initiation, one researcher (RK) generated a list of potential members of the expert panel, a convenience sample
of experienced clinicians in the eld of CRPS I who
where aliated at University hospitals or tertiary care
centres. Each nominee received a letter including information about the aim of the study and an invitation
to participate. We asked the participants to reply to
the letter with their approval to participate within two
weeks. In order to increase the response rate we sent
reminders to all experts.
Experts who agreed to participate received a next letter requesting them to prepare three separate lists of
parameters which in their personal opinion and experience are important clinical manifestations, further
examinations and follow-up parameters of CRPS I.
All stated parameters were compiled into separate
lists and were then sent again to the experts asking them
to weight each single parameter on all lists by assigning
a number between one (not important) and ten (very
important).
To identify the strongest diagnostic and follow-up
parameters the median of the attributed weights and
the corresponding inter-quartile ranges (IQR) were calculated. We arbitrarily dened the expert agreement if
the IQR of a parameter was 63. The optimal cut-o
value of the median attributed weights for a relevant
and agreed parameter was calculated drawing a ROC
curve of the medians against an IQR classication of
63 (Dincler et al., 2006). Based on this assessment we
estimated the optimal cut-o value for a relevant item
at a median attributed weight of P7.
Statistical analyses were performed using the STATA
9.2 statistical software package (Stata, College Station,
TX).

49

3. Results
The invitation letter was sent to twenty-eight experts,
nineteen agreed to participate, and fourteen returned the
rst questionnaire. Thirteen experts from the United
States of America (n = 4), Germany (n = 4), Switzerland
(n = 3), the Netherlands (n = 1) and Belgium (n = 1)
completed the survey and provided 32 diagnostic and
23 follow-up listings for analysis. In three domains (clinical presentation, further examinations and follow-up
parameters) experts agreed on the following parameters:
pain (10; 910) with its subcategories hyperesthesia (7;
58) hyperalgesia (8; 88) and allodynia (8; 710), signs
with oedema (9; 810) and colour change (8; 58) and
mobility with its categories motor change (7; 58),
decreased range of motion (8; 88) and strength (7; 6
8). The experts agreed that no further examinations were
necessary for diagnosis (10; 810). The agreed important
follow-up parameter was clinical course (10; 810) with
its categories decrease in pain (8; 89) and hyperalgesia
(8; 68), decreased oedema (8; 710) and improvements
in motor function (10; 810), and strength (8; 69).
The list of parameters with a median attributed
weight of seven or higher and an IQR 63 are shown
in Tables 1 (diagnostic criteria) and 2 (follow-up parameters). The remaining parameters are listed in the appendices. Temperature asymmetry for example fullled the
importance criterion (median weight eight), but had a
lower level of agreement (IQR 59). On the other hand
experts agreed that criteria such as skin movability from
the subcutis (median 3; IQR 24) or resolution of pain in
Table 1
Shows the set of diagnostic parameters of CRPS I fullling our
selection criteria (median attributed weight of P7 and an IQR 63)
Parameter

Subcategory

Median

IQR (2575)

Hyperesthesia
Hyperalgesia
Allodynia

10
7
8
8

910
58
88
710

Signs

Oedema
Colour change

9
8

810
58

Mobility

Motor change
Range of Motion (ROM)
Strength

7
8
7

58
88
68

Pain

Table 2
Shows the set of follow-up parameters of CRPS I fullling our
selection criteria (median attributed weight of P7 and an IQR 63)
Parameter

Subcategory

Median IQR
(2575)

Clinical course

10
Decrease in pain
8
Hyperalgesia
8
Decreased oedema
8
Improvements in motor function 10
Improvements in strength
8

810
89
68
710
810
69

50

F. Brunner et al. / European Journal of Pain 12 (2008) 4852

advanced CRPS I stages (median 2; IQR 14) were less

relevant for diagnosis (see Appendix 1 for details). In
respect to follow-up parameters, absence of sympathetic
maintained pain (SMP) reached a high median weight
(7) but failed to reach consensus (IQR 27). Conversely,
results of X-ray (median 1; IQR 12), parameters of
bone remodelling (2; 14) or three phase bone scintigraphy (2; 13) were deemed less relevant for follow-up
(see Appendix 2).

4. Discussion
In this Delphi survey, an international panel of
experts agreed that diagnosis of CRPS I can be based
on its clinical manifestations in the three domains pain,
signs and mobility. Experts also agreed that no additional exams such as X-Ray or scintigraphy were
required to conrm diagnosis. Finally, the results of this
study also indicate that the course of illness can be monitored taking change of its clinical manifestations into
account.
According to our knowledge this is the rst Delphi
survey exploring clinical and follow-up parameters in
CRPS I. The Delphi method has advantages compared
to other consensus methods. It is capable to achieve
agreement in a given area of uncertainty or lack of
empirical evidence (Murphy et al., 1998). It is swift,
inexpensive and allows combining the knowledge and
abilities of an expert group anonymously (Lindeman,
1975). Informal methods of reaching consensus, such
as committees, are recognised to be prone to domination
by powerful individuals, the biasing eects of personality traits, seniority and the fact that only one person
can speak at a time (Murphy et al., 1998). In group consensus meetings, the presence and actions of others may
inhibit creativity and the possibility of resolving ambiguous and conicting issues (Rowe et al., 1991).
Our ndings are in accordance with those published
in the ocial guidelines (Merskey and Bogduk, 1994;
Stanton-Hicks et al., 1995; Stanton-Hicks et al., 1998;
Harden et al., 2006), and other recent publications discussing salient issues of diagnosis and follow-up in
CRPS I (Burton et al., 2005; Harden and Bruehl,
2006). In addition to these eorts our study provides
an expert derived priority list of CRPS I specic manifestations, which allows weighing the importance of
each criterion compared to another. We believe that
the priority setting has its own practical clinical value.
The set of important and agreed parameters could
help clinicians to justify and standardise diagnosis of
CRPS I.
What are the limitations of this survey? It could be
argued that our set of thirteen international experts in
CRPS I of ve countries was an inappropriate sample
to represent experts worldwide and that thus our nd-

ings are poorly generalisable. Indeed, we cannot rule

out, that another sample of experts might have come
up with a dierent set of agreed parameters. However,
the high accordance with parameters stated in the guidelines makes us condent that selection bias was not a
major problem. Second, our questionnaire forced
experts to weight each of the proposed parameters of
the rst round independently ignoring potential patters
or manifestation clusters. We are aware that this somewhat articial reply mode might have distorted some of
the expert weightings. However, then again, this
approach also enabled us to quantify the attributed
weight of each parameter separately, which in response
allowed straightforward direct comparisons between
factors. Finally, it was felt, that motor change as
one of the dimensions of decreased mobility is a bit
vague. We hypothesize that this resulted from the fact,
that in daily clinical practice it is often not possible to
distinguish whether decreased ROM results from pain/
oedema/arthritic changes and actual motor changes.
We speculate that the experts summarized motor
changes including weakness, bradykinesia, dystonia,
myoclonus and tremor as suggested by a recent publication (Wilson et al., 2005).
We think that the implications for research are twofold. First we would like to encourage researchers planning studies in this clinical eld to consider our set of
diagnostic criteria for inclusion into studies. Beyond
this, we agree with a recent publication that careful clinical evaluation of sign and symptoms remains the mainstay of CRPS diagnosis (Harden and Bruehl, 2006).
Unfortunately, in the current the literature researchers
apply a broad spectrum of dierent diagnostic criteria,
which, particularly in trials can lead to a limited comparability of research ndings. Some critics even argue that
it could be questioned whether dierent authors are
really describing the same syndrome (Reinders et al.,
2002).
Second, in view, that many pathophysiological mechanisms of CRPS I remain unclear (Janig and Baron,
2003) we propose that a solid branch of research should
go into applied basic science (Baron and Janig, 2004).
From a practical point of view we suggest that clinical
research should emphasise prognostic rather than diagnostic studies. Diagnostic research requires a rm illness
denition and access to its dening elements (Bachmann
et al., 2005), which at this stage is not yet available in
CRPS I. In terms of prognostic research, besides randomised drug evaluations, we believe that large cohorts
of CRPS I patients evaluating current clinical care are
essential to increase our understanding of mechanisms
underlying health improvements in these patients.
A notable nding of this survey relates to the limited
clinical value of upstream diagnostic testing in suspected
CRPS I. This stays in contrast to the numerous new
diagnostic tests that evolved in this context recently

F. Brunner et al. / European Journal of Pain 12 (2008) 4852

(Baron and Janig, 2004). We argue that the imperfect

agreement among our experts in relation to this depends
on two mechanisms. First, in the rapidly growing eld of
diagnostic test evaluation, new promising candidates
have not yet succeeded to convince clinicians of their
additional value. Second, given that underlying mechanisms of CRPS I remain incompletely understood, diagnostic studies are dicult to perform. It can be disputed
that the lack of a reference or gold standard such as
radiography or ct-scanning in bone fractures, impedes
valid test evaluation research in CRPS I. Therefore,
physicians might prefer relying on a characteristic set
of clinical manifestations and its change over time as
suggested by a recent publication (Harden and Bruehl,
2006) rather than on information coming from lab testing or imaging.
In conclusion, this Delphi survey among an international panel of clinical experts conveys an agreed set
of relevant diagnostic parameters of CRPS I and proposes that in follow-up examinations treatment success
should be based on restoration of those manifestations.

Funding
We are indebted to the Wolfermann-Nageli Foundation and the Paul Schiller Foundation in Zurich, Switzerland for the generous nancial support with
unrestricted grants. Dr Bachmanns work (grants no.
3233B0-103182 and 3200B0-103183) was supported by
the Swiss National Science Foundation.

Competing interests

51

not reach a consensus in the second round (median

attributed weight parameters of <7 and an IQR <3)
Parameter

Median IQR (2575)

Temperature asymmetry
compared to unaected side
Pain of palpation
Redness
Perfusion
Cyanotic skin
Diuse painfulness
Sweating asymmetry compared
to the unaected side
Missing lesion of nerve radices
Hypertrichosis
Trauma
Trophic changes
Dierences in hair and nail
growth
Fibrosis of the aected joints,
tendons and ligaments
Stiness of the joints
Shiny cool skin
Skin atrophy
Motor changes
Dystonia
Increase in stiness (stages II
III)
Skin unmovable from the
subcutis
Not distinguishable from a
somatoform disorder
Tremor
Resolution of pain (stage IIIII)

59

6
6
6
6
6
6

68
57
4,58
48
37
49

6
5
5
5
5

28
46
58
47
47

36

5
5
4
4
4
4

48
37
36
37
37
38

24

14

3
2

37
14

None declared.

Acknowledgements
We thank the following experts of this survey for
their valuable contribution and accomplishment: Erich
Bar (Switzerland), Stephen P. Bruehl (USA), Jean-Pierre
Devogelaer (Belgium), Heini Gerber (Switzerland),
Georg Gradl (Germany), Michael Stanton-Hicks
(USA), Ralph-Thomas Kiefer (Germany), Rudolf Kissling (Switzerland), Christian Maihofner (Germany),
Srinivasa N. Raja (USA), Matthias Schurmann (Germany), Robert J. Schwartzman (USA) and Wim Weber
(The Netherlands).

Appendix 1
Presents the diagnostic parameters of CRPS I which
were proposed by experts in the rst round but did

Appendix 2
Presents the follow-up parameters of CRPS I which
were proposed by experts in the rst round, but did
not reach a consensus in the second round (median
attributed weight parameters of <7 and an IQR <3)
Parameter

Median IQR (2575)

Improvements in range of
motion (ROM)
Absence of SMP (sympathetic
maintained pain if previously
demonstrated)
Psychometric measures (Pain
disability index of SF36)
Quantity of analgesics required
by the patient
Decreased dystonia
Decreased redness

610

27

17

57

6
6

47
37

52

F. Brunner et al. / European Journal of Pain 12 (2008) 4852

Appendix 2 (continued )
Parameter

Median

IQR (2575)

Normalisation in temperature
asymmetry
Report by physical therapist
Normalisation of sweating
Decreased tremor
Return of sympathetic tone
Course of QST (quantitative
sensory testing)
Regeneration of the
movability between skin and
subcutis
Three phase bone scintigraphy
Parameters of bone
remodelling
Decreased hypertrichosis
X-Ray

37

5
5
4
4
3

26
17
26
27
18

13

2
2

13
14

2
1

15
12

References
Bachmann LM, Juni P, Reichenbach S, Ziswiler HR, Kessels AG,
Vogelin E. Consequences of dierent diagnostic gold standards
in test accuracy research: carpal tunnel syndrome as an example.
Int J Epidemiol 2005;34:9535.
Baron R, Janig W. Complex regional pain syndromeshow do we
escape the diagnostic trap? Lancet 2004;364:173941.
Burton AW, Bruehl S, Harden RN. Current diagnosis and therapy of
complex regional pain syndrome: rening diagnostic criteria and
therapeutic options. Expert Rev Neurother 2005;5:64351.

Dincler S, Bachmann LM, Buchmann P, Steurer J. Predictors of intraand postoperative complications in laparoscopic colorectal surgery:
results of an expert survey. Dig Surg 2006;23:1104.
Harden RN, Bruehl SP. Diagnosis of complex regional pain syndrome:
signs, symptoms, and new empirically derived diagnostic criteria.
Clin J Pain 2006;22:4159.
Harden RN, Swan M, King A, Costa B, Barthel J. Treatment of
complex regional pain syndrome: functional restoration. Clin J
Pain 2006;22:4204.
Janig W, Baron R. Complex regional pain syndrome is a disease of the
central nervous system. Clin Auto Res 2002;12:15064.
Janig W, Baron R. Complex regional pain syndrome: mystery
explained? Lancet Neurol 2003;2:68797.
Lindeman CA. Delphi survey of priorities in clinical nursing research.
Nurs Res 1975;24:43441.
Merskey H, Bogduk N, editors. Classication of chronic pain:
descriptions of chronic pain syndromes and denition of pain
terms. second ed. Seattle: IASP Press; 1994.
Murphy MK, Black NA, Lamping DL, McKee CM, Sanderson CF,
Askham J, et al. Consensus development methods, and their use in
clinical guideline development. Health Technol Assess 1998;2:i88.
Reinders MF, Geertzen JH, Dijkstra PU. Complex regional pain
syndrome type I: use of the international association for the study of
pain diagnostic criteria dened in 1994. Clin J Pain 2002;18:20715.
Rowe G, Wright G, Bolger F. Delphi a reevaluation of research and
theory. Technol Forecast Social Change 1991;39:23551.
Stanton-Hicks M, Janig W, Hassenbusch S, Haddox JD, Boas R,
Wilson P. Reex sympathetic dystrophy: changing concepts and
taxonomy. Pain 1995;63:12733.
Stanton-Hicks M, Baron R, Boas R, Gordh T, Harden N, Hendler N,
et al. Complex regional pain syndromes: guidelines for therapy.
Clin J Pain 1998;14:15566.
van de Beek WJ, Schwartzman RJ, van Nes SI, Delhaas EM, van
Hilten JJ. Diagnostic criteria used in studies of reex sympathetic
dystrophy. Neurology 2002;58:5226.
Wilson M, Stanton-Hicks M, Harden RN, editors. CRPS: current
diagnosis and therapy; progress in pain research and management. Seattle, USA: IASP Press; 2005, ISBN 0-931092-55-8.