The Journal of Pain, Vol 10, No 4 (April), 2009: pp 354-368

Available online at www.sciencedirect.com

Diagnostic Validity of Criteria for Sacroiliac Joint Pain:

A Systematic Review
Karolina M. Szadek,* Peter van der Wurff,y Maurits W. van Tulder,z
Wouter W. Zuurmond,x and Roberto S. G. M. Perezk
* Department of Anesthesiology, VU University Medical Center, and EMGO Institute for Research in Extramural
Medicine, VU University Medical Center Amsterdam, Amsterdam, The Netherlands.
y
Division of Perioperative Medicine and Emergency Care, Department of Anesthesiology and Pain Treatment,
University Medical Center, Utrecht, The Netherlands, and Department of Physiotherapy, Military Rehabilitation Centre
‘‘Aardenburg,’’ Doorn, The Netherlands.
z
Institute of Health Sciences, Faculty of Earth and Life Sciences, VU University Amsterdam and EMGO Institute, VU
University Medical Center, Amsterdam, The Netherlands.
x
Department of Anesthesiology, VU University Medical Center, Amsterdam, The Netherlands.
k
Department of Anesthesiology, VU University Medical Center, Amsterdam, The Netherlands, and EMGO Institute for
Research in Extramural Medicine, VU University Medical Center Amsterdam, The Netherlands.

Abstract: A systematic literature review was conducted to determine the diagnostic validity of the
criteria for sacroiliac (SI) joint pain as proposed by the International Association for the Study of Pain
(IASP). Databases were searched up to September 2007. Quality of the studies was assessed using
a Quality Assessment of Diagnostic Accuracy Studies (QUADAS) tool. Sensitivity, specificity, and diag-
nostic odds ratios (DOR) were calculated together with 95% confidence intervals (CI). Statistical pool-
ing was conducted for results of provocative tests. Eighteen studies were included. Five studies
examined the pattern of SI joint pain, whereas another 5 examined stressing test specific for SI joint
pain. None of the studies evaluated the diagnostic validity of the SI joint infiltration or the diagnostic
validity of the IASP criteria set as a whole. In all studies, the SI joint selective infiltration was used as
a gold standard; however, the technique, medications, and required pain relief after the infiltration
varied considerably between the studies. Taking the double infiltration technique as reference test,
the pooled data of the thigh thrust test (DOR, 18.461; CI, 5.82 to 58.53), compression test (DOR,
3.88; CI, 1.7 to 8.9), and 3 or more positive stressing tests (DOR, 17.16; CI, 7.6 to 39) showed discrim-
inative power for diagnosing SI joint pain.
Perspective: This review of clinical studies focused on the diagnostic validity of the IASP criteria for
diagnosing SI joint pain. A meta-analysis showed that the thigh thrust test, the compression test, and
3 or more positive stressing tests have discriminative power for diagnosing SI joint pain. Because
a gold standard for SI joint pain diagnosis is lacking, the diagnostic validity of tests related to the
IASP criteria for SI joint pain should be regarded with care.
ª 2009 by the American Pain Society
Key words: Sacroiliac joint pain, pain pattern, provocation test, joint infiltration, diagnostic vaidity,
systemic review.

S
acroiliac (SI) joint pain is considered to be a signifi- tion, varies from 15% to 30% in patients with low back
cant problem with a frequent occurrence in pa- pain.4,47,59 Despite this relatively high prevalence, it is
tients with low back pain. The prevalence of SI difficult to make a definite diagnosis, as presenting
joint pain, as established on the basis of clinical evalua- symptoms may be similar to those of other causes of
low back pain, and mimic, for instance, sciatica.4,8 To
make a clinical diagnosis, the International Association
Address reprint requests to Dr Karolina M. Szadek, Department of Anes- for the Study of Pain (IASP) has proposed a set of criteria
thesiology, VU University Medical Center, PO Box 7057, 1007 MB, Amster- for diagnosing SI joint pain that address mechanical dis-
dam, The Netherlands. E-mail: km.szadek@vumc.nl
1526-5900/$36.00 orders of this joint.52 According to these criteria, SI joint
ª 2009 by the American Pain Society pain refers to patients with pain in the area of the SI
doi:10.1016/j.jpain.2008.09.014 joint, which should be reproducible by performing

354
Szadek et al 355
specific pain provocation tests, or should be completely tion or language. Case series and case reports as well as
relieved by infiltration of the symptomatic SI joint with animal and cadaveric studies were excluded. Addition-
local anesthetics. These criteria, however, are not unam- ally, the reference sections of all articles selected for
biguous. Regarding the first IASP criterion, the literature the review were scanned for potentially relevant articles
suggests that SI joint pain, though originating from the that were not identified by the original search.
SI joint, can have a diffuse character and can refer to the
buttock, groin, or/and lower extremity.25,26,61,69,77 With Article Retrieval
respect to the second IASP criterion, there are several Eligibility of studies on the basis of title, key words,
pain provocation tests described that intend to stretch, and abstract was determined by 2 reviewers (K.S. and
compress, or contract certain tissue structures related P.v.d.W.), independently. If uncertainty remained, the
to the SI joint.60 However, the reliability (except for full text was reviewed. Differences in judgment were re-
the Gaenslen test and thigh thrust test) as well as valid- solved through a consensus procedure. Justifications for
ity of these tests in clinical practice is disputable.70,71 The excluding studies were noted and discrepancies dis-
third IASP criterion, the selective infiltration of the SI cussed. If no consensus was reached, a final decision
joint, whereby the local anesthetic is injected in the was made by a third reviewer (R.P.). The publications
joint cavity, plays a role of a reference stan- were included according to the following criteria: (1) pa-
dard.12,22,25,47,59,61,63 This technique, however, is criti- tients in a particular study were at least 18 years old, (2)
cized because of missing data about specificity and suffering from nonspecific, non–pregnancy-related low
sensitivity and therefore, considered not valid for diag- back pain with or without radiation to the lower extrem-
nosing SI joint pain.5 Moreover, according to the litera- ities or groin, (3) a diagnostic infiltration of the SI joint
ture, SI joint painful pathology appears to involve not was compared with another diagnostic test, or (4) any
only intra-articular structures but also periarticular diagnostic test was compared with 1 of the diagnostic
structures (for instance, ligaments and muscles).14 criteria for SI joint pain according to the IASP.
In the last few years, several reviews were published re-
lating to SI joint pain.5,10,11,14,20,23,24,31,35,36,51,60,65,70,71,80
Data Extraction
Results of these reviews are contradictory with respect to
The same 2 reviewers performed data extraction inde-
methodological quality of the evaluated studies and di-
pendently, using a standardized questionnaire devel-
agnostic value of tests. Three of the most recent reviews
oped for this study. One of the reviewers is first author
suggested utilization of multiple diagnostic tests for
of 2 of the included studies.68,69 He was not involved in
physical examination instead of relying on the result of
any decision regarding data extraction or quality assess-
a single test.35,60,65 The review by Hansen et al,37 how-
ment of these studies. Data extraction and quality assess-
ever, found that there is limited evidence for provocation
ment of these 2 studies were done by only 1 reviewer
tests and moderate evidence for diagnostic infiltration
(K.S.).
of the SI joint. On the other hand, Berthelot et al5 argued
The following data were extracted: author, year of
in their review that both the provocation tests and diag-
publication, country where the study was performed,
nostic infiltration are unreliable for diagnosing SI joint
the setting for patient recruitment, characteristics of
pain. Furthermore, none of these reviews considered
the study population (age, gender, duration of the com-
all of the criteria for SI joint pain as proposed by the
plaints), inclusion and exclusion criteria, the test(s) exam-
IASP. Therefore, with the purpose of assessing whether
ined, reference test used, duration of symptoms and
these criteria are capable of discriminating among the in-
study results, whether recruitment was consecutive and/
dividuals with and without SI joint pain, we performed
or data collection was performed prospectively (Table 1).
a systematic review of the literature. The evaluation
comprised the methodological assessment of the studies
and the summary of the diagnostic validity by sensitivity, Assessment of Methodological Quality
specificity, and diagnostic odds ratios (DORs) of diagnos- The methodological quality of the studies was assessed
tic tests that could be ascribed to the IASP criteria. by the 2 reviewers using the Quality Assessment of Diag-
nostic Accuracy Studies (QUADAS) tool.75 The QUADAS
items are defined as follows.
Methods 1. Was the spectrum of patients representative of the
patients who will receive the test in practice?
Study Identification This was considered positive if included subjects
To identify relevant literature, we conducted a compre- where older than 18 years, suffered from nonspecific,
hensive search in the following databases: PubMed, nonspondyloarthropathy, noninflammatory low back
EMBASE, and CINAHL, from the beginning of these data- pain presumably stemming from the SI joint; there
bases up to September 2007. The full search strategy was was no radicular radiation of pain to lower extremity
developed in collaboration with an experienced librar- along 1 or more dermatomes, accompanied by numb-
ian (I.R.) using a filter outlined by Deville et al.19 How- ness and tingling, muscle weakness and loss of specific
ever, using this filter in a preliminary search resulted in reflexes.
only a few hits in the area of SI joint pain. Therefore, 2. Were selection criteria clearly described?
we broadened our search, as outlined in Appendix 1. This was considered positive when both inclusion
We applied no restrictions with regard to year of publica- and exclusion criteria were clearly described, and it
 356
Table 1.Study Characteristics for the Individual Studies That Investigated Diagnostic Accuracy of Provocative Tests for Subjects With
Sacroiliac Joint Pain
RECRUITMENTS:
SAMPLE SIZE N = ALL SETTING (S),
1
AUTHOR, SUBJECTS, N = CASES ; CONSECUTIVE (C),
REFERENCE, AGE [MEAN (SD OR PROSPECTIVE INCLUSION CRITERIA TYPE OF REFERENCE TEST;
COUNTRY RANGE)]; GENDER % (F) (P); PERIOD; EXCLUSION CRITERIA OUTCOME MEASURE TYPE OF INDEX TEST

Broadhurst7 n = 40, n = 40 [F =36 s = university hospital, LBP below lumbosacral junction, associated Single SIJB with 4 mL Lidocaine Flexion abduction and external rotation
1998 years; range, 18-72 c = yes; p = yes; with groin pain and absence of lumbar 1% or NaCl 0.9% with image (Patrick’s sign), Posterior shear (thigh
Australia M =35 years; range, 2-year period symptoms; pain with full weight bearing intensification; 70% pain trust test), resisted abduction test.
25-53 - (No SD)]; on 1 leg; worsening going down hill reduction of pain on VAS
- 75 % F Previous LBP; fractures; infections; metastases;
pregnancy; use of major tranquilizers;
systemic disease
Dreyfuss22 n = 88, n = 85 s = university hospital, LBP below L5 Radiation 6Pain pattern Contrast-enhanced SIJB with 12 physical
1996 USA [45 years, (range, spine center; c = yes, consistent with SIJP 1.5 mL 2% lidocaine and examination tests
18-87)]; 72% F p = yes; period Unknown 0.5 mL corticosteroids;
unknown; 90% pain reduction on VAS
Fortin199425 n = 54, n = 16 s = Regional spinal LBP $2 weeks Pain provocation arthrography pain patterns25; pain
199727 USA [35 years (range diagnostic center; No preselection with 1-2.7 mL contrast; intra- area pointing (Fortin
21-45)]; 38% F p = yes; c = unknown; articular bupivacaine 0.75%; finger test)27
period unknown; positive if patient’s pain was
provoked
Fukui 2002 n = 28 [58 613 s = University hospital; Pain in the SIJ region 6referred pain; Pain Contrast-enhanced intra-articular pain patterns
Japan32 years]; gender c = unknown p = provocation by Patrick’s and Gaenslen’s injection of 2 mL 1%

Diagnostic Validity of Criteria for Sacroiliac Joint Pain

unknown unknown; 1994-2001 test; 80% pain relief after SIJB mepivacaine and 2 mg
dexamethasone;
Zygapophyseal joint pain; radicular pain not predefined
Laslett 200345 n = 62, n = 48; [42 s = Private radiology Buttock pain, 6lumbar or lower extremity Contrast enhanced intra-articular McKenzie evaluation,
200544 years (range, 20-79)]; practice; c = no; symptoms, able to tolerate full physical double infiltration of <1.5 mL SI joint tests, hip joint
Sweden 67 % F p = yes; 21 months examination lidocaine and bupivacaine assessment45; distraction
Unwilling to participate; Only midline or (corticosteroid); pain provocation test, thigh thrust,
symmetrical pain above L5; clear root during infiltration; required pain Gaenslen’s test,
compression signs; referred for special reduction 80% compression test, sacral thrust44
procedures except SI joint injections;
Maigne47 n = 67, n = 54; s = Public hospital, Chronic LBP >50 days, radiation 6VAS >40 mm, Contrast enhanced double SIJB with Clinical physical examination
1996 [median = 45.3 spine center; c = no, failure of epidural or facet joint injections, 2 mL l2% lidocaine and 0.5% tests
France years; IQR, p = yes; period of pain and tenderness over the region of SIJ, bupivacaine; positive if pain
34.5-56.8]; 61% F recruitment unknown 18-75 years old reduction >75% for at least
Previous spine surgery, disc herniation, 2 hours
chemonucleolysis, neurological changes,
discopathy, lumbar spinal stenosis,
spondylolisthesis
(Continued)
Maigne48 n = 39, n = 32; s = University hospital; Chronic low back pain, tenderness of the 1 mL contrast enhanced intra- Quantitative radionuclide

Szadek et al
1998 [53.4 618.5]; c = unknown; p = yes; SI joint line, VAS $4 cm, age >18 years articular SIJB with 2 mL lidocaine bone scan
France 56% F period of recruitment Surgery in history, lumbar disc narrowing, 2%; positive if pain relief 15 min
unknown spondylolisthesis, spinal stenosis, herniated after the injection was >75%
intervertebral disc, chemonucleolysis,
pregnancy
Maigne49 n = 61; n = 40; s = Public hospital, Chronic LBP $6 months, Unilateral pain, pain 1 mL contrast enhanced intra- sacroiliac bone scintigraphy
2005 [48 6 11 years]; spine center; c = not; distribution within SI joint pain pattern, articular SIJB with 2 mL
France 65% F p = yes; 1996-2002 sacral sulcus palpation tenderness lidocaine 2%; Positive if pain
Pain radiation below the knee, lumbar relief was 75%.
cause of pain, work-related injury,
lawsuit, psychiatric disorder
Manchikanti50 n = 120; [47.5 s = nonuniversity, Chronic LBP .6 months; age, 0.5-1 mL contrast enhanced Physical examination Single
2001 USA SEM = 1.16]; private practice; 18-90 years intra-articular double SIJB with versus double blockade
66% F c = unclear; p = yes; Neurological deficits, definite 0.5-1 mL lidocaine 2% or
period of recruitment diagnosis known, pain in the bupivacaine 0.5%; Positive
unknown sacral region, tenderness over effect undefined
SIJ, positive provocative maneuvers
Schwarzer59 n = 100; n = 43 s = University hospital; Pain below L5-S1, 18-80 years old 1 mL contrast enhanced intra- pain provocation during the
1995 [median = 32.8, c = yes; p = yes; April- Status post lumbar spinal surgery, articular SIJB with 1 mL lidocaine infiltration; pain patterns
Australia ICQ 28.7-40.9] October 1992 exhibition of neurological signs 2%; Positive if pain relief
51% F was >75%
Slipman62 n = 50, n = 50; s = spine center; c = yes; LBP, 6radiation to the lower extremity, sacral 0.5 mL contrast enhanced intra- Bone scan
1996 [range, 18-77]; p = yes; period of sulcus tenderness, 3 positive provocation articular SIJB with 2 mL lidocaine
USA 66% F recruitment unknown tests, no improvement after physical therapy 2% and 1 mL corticosteroids or
SA, urethritis, peripheral arthritis, psoriasis, 3.0 mL lidocaine 2% or 3 mL
early morning stiffness, inflammatory lidocaine 1%; Positive if pain
bowel disease, neurological deficit relief was $80%
Slipman63 n = 50, n = 50 s = University hospital, LBP, 6radiation to the lower limb 0.5 mL contrast enhanced SIJB physical examination Patrick’s test,
1998 [range, 18-77] spine center; c = yes; Previous spine surgery, SA, urethritis, peripheral with 2 mL lidocaine 2% en 1 mL pressure pain at the sacral sulcus;
USA 62 % F p = no; period of arthritis, psoriasis, inflammatory bowel betamethasone /or up to 3 mL shear test, standing extension;
recruitment unknown disease, neurological deficit lidocaine 2% or up to 3 mL Gaenslen’s test and Yeoman test
lidocaine 1%; positive if pain relief
was $80%
Slipman61 n = 50; n = 50 s = University hospital, LBP or buttock pain, 6radiation to the 0.5 mL contrast enhanced SIJB with pain referral zones
2000 [mean, 42.5; spine center; c = yes; lower limb, 3 positive provocation tests 2 mL lidocaine 2% Positive if pain
USA range, 20-75 p = no; period of SA, urethritis, peripheral arthritis, psoriasis, relief was at least 80 %
years] F 64% recruitment unknown inflammatory bowel disease, early morning
stiffness, neurological or muscular deficit,
spondylolisthesis, lumbar instability,
lumbosacral radiculopathy
(Continued)

357
 358
Table 1.Study Characteristics for the Individual Studies That Investigated Diagnostic Accuracy of Provocative Tests for Subjects With
Sacroiliac Joint Pain
RECRUITMENTS:
SAMPLE SIZE N = ALL SETTING (S),
1
AUTHOR, SUBJECTS, N = CASES ; CONSECUTIVE (C),
REFERENCE, AGE [MEAN (SD OR PROSPECTIVE INCLUSION CRITERIA TYPE OF REFERENCE TEST;
COUNTRY RANGE)]; GENDER % (F) (P); PERIOD; EXCLUSION CRITERIA OUTCOME MEASURE TYPE OF INDEX TEST

Vd Wurff68,69 n = 140; n = 60 s = General hospital, Chronic LBP $50 days, pain below L5 over contrast (1 mL) enhanced double pain mapping69; multitest regimen of
2006 [51 6 13] pain department; the posterior aspect of SI joint SIJB with 2 mL lidocaine 2% or 5 SI joint pain provocation tests68
The 78% F c = p = yes; January unilaterally, 6leg pain, VAS >45 mm, bupivacaine 0.25%; Positive if
Netherlands 2001-April 2002 age 18-80 years pain relief was at least 50% for
SA, leg-length discrepancies of >2 cm, 1 hour after lidocaine or 4 hours
Waddell score >2, tumors, recent after bupivacaine infiltration
lumbar spine fractures, disc abnormalities
with nerve root compression clinical signs,
osteoporosis, infection, clinically
symptomatic cox-arthrosis, radicular pain

Diagnostic Validity of Criteria for Sacroiliac Joint Pain

with neurological signs, pregnancy,
anticoagulants, liver and/or kidney failure
Young78 n = 102, n = 81 s = private radiology Chronic LBP single disc, facet and/or SI joint pain mapping
2003 USA [40.8; SD,12.1 practice, c = Unwilling to participate, signs of nerve root arthrography/pain provocation/
years] 60.5% F unclear; p = yes; compression, unable to tolerate clinical injections; <1.5 mL local anesthetic;
period of examinations pain provocation during infiltration
recruitment and pain reduction 80%
unknown

Abbreviations: F, female; M, male; SD, standard deviation; SEM, standard error of mean; LBP, low back pain; VAS, visual analog scale, 6, with or without; SIJB, SI joint blockade; SIJP, SI joint pain; SA, spondyloarthropathy.
NOTE. Publications are listed alphabetically by first author. Cases1 = number of subjects actually undergoing the index test.
Szadek et al 359
was clear whether the recruitment of subjects was con- 9. Was the execution of the reference standard
ducted prospectively and the inclusion of subjects was described in sufficient detail to permit its replica-
consecutive. tion?
3. Is the reference standard likely to correctly classify This was scored positive if the reference standard
the target condition? was sufficiently described to permit its reproduction.
In the absence of a gold standard in diagnosing This item would also be scored positive if a reference
primary SI joint pain, the double infiltration of the was provided to a full test description.
SI joint was considered the best available reference 10. Were the index test results interpreted without
test.6 Moreover, the reference test used in a specific knowledge of the results of the reference standard?
study was described without direct interpretation of 11. Were the reference standard results interpreted
its usefulness. Data from individual studies regard- without knowledge of the results of the index
ing this subject were extracted and recorded in test?
Table 1. These 2 items were assessed positive if the study
4. Is the time period between reference standard and clearly stated that the results of both the index test
index test short enough to be reasonably sure that and the reference standard, were interpreted in
the target condition did not change between the a blinded manner. If case of uncertainty, this item
2 tests? would be scored as ‘‘unclear’’ and negative by missing
This item was considered positive if the delay be- information.
tween the application of the index test and reference 12. Were the same clinical data available when test re-
standard used in the study was reported and was not sults were interpreted as would be available when
more than 7 days, or for SI joint infiltration with local the test is used in practice?
anesthetics not shorter than 24 hours (bupivacaine,
This was considered positive if the observer was
0.5% T0,5 = 1.5 to 5.5 hours, and lidocaine, 2% T0,5 =
aware of the characteristics mentioned in the inclusion
90 to 120 minutes).
and exclusion criteria. If the diagnosis of SI joint pain
5. Did the whole sample or a random selection of the was already known, this item was scored negative.
sample, receive verification using a reference stan- 13. Were uninterpretable/ intermediate test results
dard of diagnosis? reported?
This was considered positive if it was clearly described If it was clear that all test results, including uninter-
that all patients or a random selection of patients re- pretable/undetermined/ intermediate results were re-
ceived verification of their disease status with the refer- ported, this item was scored positive. Therefore, the
ence standard used in the study, regardless of the index number of included patients should match the number
test results. In the case of random selection, it would be of the subjects receiving the index test. If it was not
clear that the randomization took place before the im- a case and not further explanation was available this
plementation of the index test. was scored negative.
6. Did patients receive the same reference standard 14. Were withdrawals from the study explained?
regardless of the index test result? This was scored positive if it was clear what hap-
This was evaluated positive if it was clear that all pa- pened to all patients who entered the study, for exam-
tients received the same reference standard used in the ple if a flow diagram of study participants was
study regardless of the index test result. In the case of reported. If the data were not available, than the
a random selection, this item would be scored as posi- item would be scored negatively.
tive if it was clear that the randomization was The 2 reviewers, blind to each other assessment, scored
performed before applying both index and reference the criteria items: ‘‘yes’’ or ‘‘no’’ when studies satisfied or
test. failed to meet the criteria, respectively, and ‘‘unclear’’
7. Was the reference standard independent of the when information was lacking to decide whether the
index test? study satisfied or met that specific item. In the case of dis-
This was scored positive if it was clear that the refer- agreement, the 2 reviewers tried to reach a consensus on
ence test used in the study was independent of the in- each criterion, and in case of persisting disagreement,
dex test, positive/negative results of both tests were a third reviewer (R.P.) decided. Agreement between re-
predefined and the index test did not form part of viewers was quantified using the kappa (k) coefficient
the reference standard. If the method of the assess- with quadratic weighting (http://www.faculty.vassar.
ment of index test and reference test would not be ex- edu/lowry/kappa.html58). The strength of agreement
plained, this item would be scored as negative. was interpreted as poor (<0.20), fair (0.21 to 0.40), mod-
8. Was the execution of the index test described in erate (0.41 to 0.60), good (0.61 to 0.80) and very good
sufficient detail to permit replication of the test? (0.81 to 1.00).2
This item was scored positive if the index test was
sufficiently described to permit its reproduction, or
a reference to the adequate description of the test Pilot Study
was provided. Only mentioning the tests names would To improve agreement between the 2 reviewers, this
be scored negative, as some of the tests could be procedure was tested using 2 studies hat examined
named erroneously.43 the diagnostic accuracy of Lachman’s test in knee
360 Diagnostic Validity of Criteria for Sacroiliac Joint Pain
18,46
instability. The disagreement between the 2 re- 0.93 (95% CI, 0.87 to 0.97) for all criteria. Most disagree-
viewers was 18% (5/28); k was 0.77 (95% confidence in- ments regarded questions 6 and 7 and were principally
terval [CI], 0.58 to 0.96) for all criteria. due to either reading errors or differences in interpreta-
tion (‘‘yes’’ or ‘‘no’’ against ‘‘unclear’’). All disagreements
Data Analysis but 1 were resolved during a consensus meeting.
The diagnostic validity of a test was assessed by its abil-
ity to correctly discriminate between subjects with and Study Characteristics
without SI joint pain. The ideal diagnostic test would al- Study characteristics are presented in Table 1. Nine
ways be positive in subjects with the disease, and nega- studies were conducted in the 1990s,7,22,25,27,47,48,59,62,63
tive in those without the disease. In such a case, the whereas the remaining 9 studies were published after
sensitivity and specificity of the test would be 100%. 2000.32,44,45,49,50,61,68,69,78 Six studies evaluated pain
The effect size of sensitivity and specificity can be re- mapping, pain area, or pain referrals from the SI
ported as a diagnostic odds ratio (DOR). The DOR informs joint,22,25,32,59,61,69 which could be ascribed to the evalu-
how much greater the odds of having the disease are for ation of the first IASP criterion. Six studies assessed the
individuals with a positive test result than for individuals accuracy of pain provocation or stressing tests,7,22,44,47,
with a negative test result.33 Sensitivity and specificity 63,68
the second IASP criterion. One study evaluated the
were extracted from individual studies, and 2  2 contin- role of different contributors to low back pain, based
gency tables were reconstructed if possible. In 2 cases, on physical examination and positive outcome of various
this information was not available in the published pa- diagnostic infiltrations.50 Additionally, 10 studies evalu-
per and we requested the original data from the first au- ated other kinds of clinical examination as the value of
thors.68,78 For the statistical analysis, Meta-DiSc software clinical history and mobilization tests,22 mechanical ex-
was used.79 Confidence intervals for sensitivity, specific- amination of the lumbar spine,45,47,78 pain provocation
ity and DOR were calculated for each test and subse- arthrography,25,27,59 and bone scintigraphy.48,49,62
quently tested for heterogeneity. The heterogeneity The majority of the studies were carried out in univer-
and consistency of studies in the meta-analysis was as- sity hospitals and/or spine centers.7,22,25,27,32,47-49,59,61-63
sessed with c2 and I2 statistics, respectively. A heteroge- Patient recruitment was prospective in all but 3 stud-
neity result of P < .05 was considered significant, given ies32,61,63 and consecutive in a minority of stud-
the low power of the test. The I2 values have a continuous ies.7,22,59,61-63 In only 8 studies, the recruitment period
scale of 0% to 100%, with 0% defining no inconsistency was reported, which ranged from 7 months59 to 7 years.32
and 25%, 50%, and 75% were assigned as limits of low, All studies included patients with low back pain. In 7
medium and high inconsistency.38 Statistical pooling studies, inclusion criteria required only chronic low
was conducted only for results of provocative tests, and back pain complaints.47-50,68,69,78 The age of the patients
only for the data extracted from the prospective studies, ranged from 18 to 87 years. The overall involvement of
and 95% confidence intervals of the pooled DORs were women in the included studies was more than 60% and
calculated. Because of a limited number of studies, it ranged from 38% to78%.
was not possible to produce an ROC curve. Finally, a qual-
itative descriptive analysis was performed.
Methodological Quality Assessment
Results of the methodological quality assessment are
Results presented in Table 2. Overall, the majority of the studies
scored positive for at least 8 of 14 QUADAS items. All the
Results of the Search studies included representative individuals. However, in
In a PubMed database search, we identified 616 poten- 4 studies, the inclusion of patients was limited to pain be-
tially relevant articles. We excluded 587 studies on the low the level of L5-S122,59,68,69 and in another 2 to pain in
basis of their titles and abstracts. Subsequently, we re- the buttock.44,45 Furthermore, the majority of the studies
trieved and reviewed 29 full reports for possible inclu- also comprised patients with pain radiating to the groin
sion.7,9,12,13,21,22,25-27,40,42,44,45,47-50,55-57,59,61-63,67-69,73,78 or lower extremity, emphasizing its nonradicular charac-
Ten studies were excluded because they examined either ter.7,22,32,44,45,47,61-63,68,69 In 3 studies, the selection crite-
reliability or repeatability of the test.9,12,13,40,42,55-57,67,73 ria were not described25,27 or were not sufficiently
Two studies were excluded because they considered described.78 In 9 studies, the period between the index
asymptomatic individuals.21,26 This resulted in 17 studies and the reference test was either not reported or was
that fulfilled our inclusion criteria.7,22,25,27,44,45,47-50,59,61- not clearly described,25,27,32,48-50,59,63,78 which could
63,68,69,78
A comprehensive search in the EMBASE and lead to a disease progression bias (item 4). In 6 studies,
CINAHL databases resulted in 493 and 13 hits, respec- only a selected sample of patients was verified using a ref-
tively. After eliminating duplicates from PubMed, 305 erence test, and as the selection was not random this
references remained, of which only 1 article was could indicate a partial verification bias.25,27,44,45,50,78
included.32 Furthermore, in 10 studies, it was not clear whether or
not the patients received the same reference standard re-
Agreement Between Assessors gardless of the index test result25,32,44,45,47,50,61-63,78 (item
Agreement between the 2 reviewers was high (224/ 6). In 3 studies, it was not clear whether the reference
234, 95.7%), with a k (using quadratic weighting) of standard was independent of the index test,27,49,50 and,
Szadek et al 361
Table 2. Methodological Quality Assessment of Studies That Investigated the Diagnostic Accuracy
of Provocative Tests for Subjects With Sacroiliac Joint Pain
AUTHOR , YEAR1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 k2

Broadhurst7 Y Y N Y Y Y N Y Y U U Y Y Y ,82
Dreyfuss22 Y Y N Y Y Y Y Y Y U U Y Y Y 1
Fortin25 Y N N U N N N Y Y U U Y N N 1
Fortin27 Y N N U N Y U Y U Y N Y N N 1
Fukui32 Y Y N U Y U Y Y Y U U U N N ,88
Laslett44,45 Y Y Y Y N U Y Y Y U N Y Y Y ,82
Maigne47 Y Y Y Y Y U U N U U U Y N N ,64
Maigne48 Y Y N U Y Y Y Y Y Y Y Y Y Y 1
Maigne49 Y Y N N Y Y Y Y Y N N Y Y Y ,43
Manchikanti50 Y Y Y U N N U U Y U U Y Y Y 1
Schwarzer59 Y Y N U Y Y N U Y U U Y Y Y 1
Slipman62 Y Y N Y Y N Y Y U U U Y Y Y 1
Slipman63 Y Y N U Y U Y N U Y N Y Y Y 1
Slipman61 Y Y N Y Y U Y U Y U U U N N 1
Van der Wurff68 Y Y Y Y Y Y Y N Y U U Y N U ,85
Van der Wurff68 Y Y Y Y Y Y Y Y Y Y N Y Y Y 1
Young78 Y N Y U N N Y Y Y Y Y U N U 1

NOTE. 1Publications are listed alphabetically by author. Items were scored as follows: ‘‘Y’’ means adequate methods, ‘‘N’’ means inadequate methods, and ‘‘U’’ means
an item was inadequately described and therefore a decision could not be made whether it satisfied the criteria or not. 2The k score between the 2 reviewers (K.S. and
P.v.d.W.) over the 14 points of QUADAS. 3The observed proportion of overall agreement between the 2 reviewers considering particular QUADAS items.

in another 3, part of the index tests formed part of the provocation tests performed before and after infiltration
reference test.7,25,59 Eight studies provided insufficient was measured, and 70% pain reduction after the infiltra-
description of the index and/or reference test to permit tion was considered positive.7 In 4 studies, 75% postinfil-
its replication.27,47,50,59,61-63,69 Only 2 studies clearly noti- tration pain relief was required,47-49,59 whereas in 6
fied that both the index and reference test were assessed studies, 80% pain reduction was required44,45,61-63,78
blindly to each other’s results.48,78 Another interpreta- and in 1 study, 90% pain reduction was required.22 These
tion bias could concern 3 studies whereby we were uncer- variations in cut-offs for positive effect, however, do not
tain about the availability of the clinical data during the influence the test results, as shown in previous stud-
interpretation of test results by the examiners.32,61,78 ies.16,17 In the remaining 4 studies, a positive effect of
Finally, in 7 studies, we found bias associated with the the infiltration was not predefined.25,27,32,50
report of study results and withdrawals.25,27,32,47,61,69,78 Intra-articular infiltration was tested for its ability to
diagnose SI joint pain. Schwarzer et al59 hypothesized
that if the intra-articular infiltration would relieve
The Reference Test none of the patients’ pain, then the concept of SI joint
All of the included studies used contrast enhanced pain would be refuted. The authors estimated that the
intra-articular (intracavital) injections with local anes- prevalence of SI joint pain could be as high as 13% to
thetics as a reference test. The volume of contrast agent 30% of patients with low back pain. Maigne et al47 eval-
in the studies was usually very small and used only for uated the prevalence of SI joint pain in low back pain pa-
confirmation of the intra-articular position of the nee- tients, whereby instead of single infiltration, double
dle, whereas in 3 studies, SI joint arthrography was ob- infiltration was used. Obtaining comparable pain relief
tained.25,27,59 In 6 studies, double injections were used, of both injections would discriminate patients with SI
whereby on separate occasions lidocaine and bupiva- joint pain. The prevalence of SI joint pain in their study
caine were infiltrated.44,45,47,50,68,69 In addition, in 6 stud- was 18.5%. In a randomized, placebo-controlled study,7
ies corticosteroids were used together with local in contrast to lidocaine injections, none of the patients
anesthetics.22,32,44,45,62,63 The volume of injected medi- achieved 70% or more pain relief after placebo
cines varied between studies and ranged between injections.
1 mL59 and 4 mL,7 but in most of the studies 2 mL was
used. However, in 2 studies, the volume of infiltrated
medicines was adjusted to the volume of the joint itself, Diagnostic Value of the Tests
which was determined by a previous arthrography.25,27 From the results of 7 studies, we could not construct
A positive effect of the diagnostic infiltration was pre- 2  2 contingency tables because in the design of these
defined in most of the studies but varied considerably studies a control group was lacking.25,27,32,47,61,63,69 The
between studies. In 2 studies, 50% pain reduction was majority of these studies examined pain mapping or
considered positive if it was sustained for at least 1 or 4 pain referral patterns of the SI joint.25,27,32,61,69 Most fre-
hours after infiltration with lidocaine or bupivacaine, re- quently reported was the area that overlies the posterior
spectively.68,69 In 1 study, the level of pain produced by aspect of the SI joint, which is consistent with the first
362 Diagnostic Validity of Criteria for Sacroiliac Joint Pain
IASP criterion. However, the occurrence of radiation to Concerning the third IASP criterion, there are no data
the buttock, groin and lower limb was not considered about the diagnostic validity of the complete pain relief
uncommon.22,25,32,59,61,69 According to 1 study, the pres- after the selective infiltration of the SI joint. None of the
ence of pain in SI joint region or buttock characterizes included studies evaluated the IASP criteria set as a whole
a high sensitivity but very low specificity and DOR is for SI joint pain.
lower than 122 (see Table 3), indicating no diagnostic util-
ity for SI joint pain. Some investigators postulate that pa-
tients with presumed SI joint pain point out the area Discussion
adjacent to the superior posterior iliac spine.22,27 Based The purpose of the present review was to evaluate di-
on the results of 1 study, this clinical test has reasonable agnostic validity of tests that could be ascribed to the
specificity and sensitivity22 with a DOR of 2.745 (95% IASP criteria for diagnosing SI joint pain. The first crite-
CI, 0.99 to 7.63). Furthermore, based on the results of rion, the presence of pain in the SI joint region, tests
1 study, the most intense pain area in patients with SI such as pain mapping or pain referral patterns, have an
joint pain overlies the posterior margin of the SI joint.69 ability to correctly identify patients with SI joint pain.
This finding was consistent with the pain referral map, as However, they fail in discriminating patients without SI
determined by Fortin at al.25 joint pain. Furthermore, pain originating from the SI
With regard to the diagnostic validity of the second joint can extend to the buttock, groin, and even to the
IASP criterion, in 1 study the calculation of sensitivity lower extremity.22,25,32,59,61,69 In addition, it is doubtful
and specificity differed from our own calculation.7 Tak- whether the presence of pain in the region overlying
ing a positive response to a local anesthetic to be true the SI joint is exclusive for SI joint structures. It is well
positive, a recalculation for resisted abduction, Patrick’s known that other anatomical structures in the low
sign and thigh thrust test yielded specificities of 87%, back, such as the disc and facet joints, are also capable
77%, and 80%, respectively, and a sensitivity of 100% of producing referred pain in the buttock region.4,41 In
for all these tests. From the study of Dreyfuss et al,22 this respect, we are facing a major clinical problem to dif-
we extracted only the results in which 2 examiners (phy- ferentiate between SI joint pain and other pain sources
sician and chiropractor) agreed completely in their find- related to the lumbar and buttock region.
ings. For the study of Laslett et al44 the 2  2 contingency With regard to the second IASP criterion, 2 individual
table was reconstructed for the distraction, compression, pain provocation tests—the compression and thigh
Gaenslen, thigh thrust, and sacral thrust test. From the thrust test—are helpful in diagnosing SI joint pain. Pa-
study of Young et al,78 we retrieved data of noncentral- tients with a positive thigh thrust test or compression
ization of the pain and 3 or more positive provocation test are more likely to have SI joint pain. Subsequently,
tests. Because these data were not consistent with the studies validating a comprehensive set of stressing tests
number of positive (22) and negative (35) responders to proved good diagnostic validity of a threshold of 3 posi-
the SI joint infiltration as reported, we requested the tive tests for diagnosing SI joint pain. Using a threshold
raw data from the first author. Furthermore, we also re- of 3 or more positive stressing tests, the DOR of 3 positive
quested the raw data for individual tests results from the provocation test is high in patients with SI joint pain
study of van der Wurff et al.68 (DOR, 17.2). However, when applying pain provocation
We tested the results of 5 individual provocation tests: tests, it is nearly impossible to define which structures
compression, distraction, thigh thrust, Gaenslen’s test, actually are stressed.47,54 Even structures such as the
and Patrick’s sign. The analysis of heterogeneity and con- iliolumbar ligament or piriformis muscle cannot be ex-
sistency for the thigh thrust test was shown to be signif- cluded as potential source for this pain, since they are
icant and inconsistent. The result of the study of Dreyfuss functionally related.3,53 Consequently, it is very difficult
et al22 was proved to be an ‘‘outlier’’ during the plot in- to distinguish whether the provoked pain is exclusively
spection. A comprehensive analysis did not clarify this intra-articular, or related to capsular ligaments.
outlier. The only explanation that we found were varia- Concerning the third IASP criterion, all of the trials in-
tions in the protocol of the reference test. The differ- cluded in our review used the selective infiltration as
ences concern single versus double infiltrations, and a reference test; however, the diagnostic validity of this
the thresholds for a positive reference standard, ranging test has received a little attention. Although no particu-
from 50% to 90% pain relief. Consequently, we chose for lar injection technique is recommended by the IASP, it is
the subgroup analysis of the trial results concerning the generally accepted to perform an intra-articular infiltra-
thigh thrust test, which used the double SI joint infiltra- tion of a small volume of contrast medium to localize the
tions. The pooled sensitivity, specificity, and DOR of the joint, followed by a little amount of a local anesthetic.
thigh thrust test and the compression test showed that With this intra-articular technique, one has the ability
these tests have a discriminative power and are pre- to correctly determine symptomatic from asymptomatic
sented in Table 4. Results of Gaenslen’s test, distraction SI joint patients.7,47,59 It is surprising, however, that the
test, and Patrick’s sign could not be pooled due to the selective infiltration targets only the joint cavity. Taking
heterogeneity between the trials. Finally, the results of the basic anatomy of the SI joint into account, one may
4 studies that examined the accuracy of composition of assume that the complaints originating from the SI joint
provocation test for SI joint pain were pooled,44,45,68,78 could also involve neighboring SI joint ligaments. On the
showing good diagnostic validity and discriminative other hand, injecting even a very small volume of a local
power for SI joint pain (see Table 4). anesthetic into the joint cavity does not prevent leakage
Szadek et al 363
Table 3. Diagnostic Validity of Tests
CI CI CI

INDEX TEST AUTHOR (REF.) SEN LOW HIGH SP LOW HIGH DOR LOW HIGH

7
Broadhurst
Resisted abduction 1.0 0.81 1.0 0.87 0.66 0.97
Patrick’s sign 1.0 0.77 1.0 0.77 0.56 0.91
Thigh thrust test 1.0 0.78 1.0 0.80 0.59 0.93
Dreyfuss22
SI joint pain 0.85 0.81 0.92 0.08 0.03 0.15 0.5 0.13 1.99
Groin pain 0.19 0.11 0.29 0.63 0.55 0.73 0.4 0.14 1.13
Buttock pain 0.80 0.74 0.88 0.14 0.07 0.23 0.66 0.21 2.16
Sitting position 0.03 0.01 0.08 0.9 0.87 0.96 0.28 0.04 2.12
PSIS pointing 0.76 0.65 0.85 0.47 0.35 0.57 2.75 1.0 7.52
Gillet 0.43 0.28 0.57 0.68 0.56 0.8 1.59 0.49 5.21
Thigh thrust 0.36 0.26 0.47 0.50 0.39 0.62 0.56 0.22 1.46
Patrick’s 0.69 0.6 0.79 0.16 0.08 0.25 0.41 0.14 1.23
Gaenslen’s 0.71 0.62 0.81 0.26 0.16 0.36 0.87 0.31 2.44
Sacral thrust 0.53 0.44 0.64 0.29 0.17 0.44 0.47 0.16 1.41
Spring test 0.75 0.64 0.85 0.35 0.22 0.47 1.6 0.49 5.23
‘‘Sacral sulcus’’ tenderness 0.95 0.90 0.98 0.09 0.04 0.13 1.85 0.34 9.86
Laslett44
Distraction test 0.6 0.36 0.8 0.81 0.65 0.91 5.95 1.6 22.15
Compression 0.69 0.44 0.86 0.69 0.5 0.84 4.84 1.33 17.67
Thigh thrust test 0.88 0.64 0.97 0.69 0.5 0.84 15.4 2.93 80.95
Gaenslen’s test r 0.53 0.3 0.75 0.71 0.53 0.86 2.56 0.74 8.89
Gaenslen’s test l 0.5 0.27 0.73 0.77 0.6 0.89 3 0.85 10.63
Sacral thrust 0.63 0.39 0.82 0.75 0.58 0.87 5 1.38 18.17
1 or more positive tests 1.00 0.84 1.00 0.44 0.59 0.44
2 or more positive tests 0.94 0.76 0.99 0.66 0.57 0.68 28.36 4.11 187.18
3 or more positive tests 0.94 0.74 0.99 0.78 0.69 0.807 53.57 7.3 362.16
4 or more positive tests 0.60 0.40 0.76 0.81 0.72 0.89 6.5 1.72 24.71
5 or more positive tests 0.27 0.122 0.41 0.88 0.81 0.94 2.54 0.58 11.16
6 positive tests 0.07 0.012 0.201 0.88 0.85 0.94 0.5 0.07 3.78
Composition of 4 tests 1 or more 1 1.00 0.84 1.0 0.47 0.39 0.47
Composition of 4 tests 2 or more 1 0.88 0.69 0.96 0.78 0.69 0.82 25.0 4.92 121.5
Composition of 4 tests 3 or more 1 0.63 0.44 0.77 0.84 0.75 0.92 9.0 2.31 35.13
Composition of 4 tests 4 or more 1 0.80 0.41 0.96 0.84 0.78 0.87 21.6 2.50 170.54
Laslett45
3 or more positive tests n = 43 0.91 0.67 0.98 0.78 0.7 0.81 35.71 4.75 246.7
Positive clinical examination n = 34 0.91 0.69 0.98 0.87 0.76 0.91 66.67 7.3 540.23
Maigne48
Radionuclide bone scanning 0.46 0.19 0.75 0.95 0.74 1.0 15.43 1.56 152.35
Maigne49
Bone scintigraphy 0.43 0.24 0.62 0.65 0.55 0.76 1.42 0.39 5.22
Manchikanti50
Single versus double SIJB 1.0 0.16 1.0 0.78 0.52 0.94 16.11 0.65 401.34
Schwarzer59
Similar and exact pain reproduction 0.85 0.55 0.98 0.47 0.28 0.66 4.81 0.91 25.53
during infiltration
Slipman62
Radionuclide bone scan 0.13 0.36 2.98 1 0.82 1 6.38 0.33 125.5
Van der Wurff68
3 or more positive tests 0.85 0.73 0.93 0.79 0.68 0.85 21.36 5.7 79.44
Distraction test 0.26 0.11 0.46 0.73 0.54 0.87 0.93 0.29 2.95
Compression test 0.6 0.39 0.78 0.7 0.51 0.84 3.3 1.15 9.73
Thigh Thrust test 0.93 0.76 0.99 0.64 0.45 0.8 21.9 4.4 108.9
Gaenslen’s test 0.63 0.42 0.81 0.79 0.61 0.91 6.3 2.01 19.8
Patrick’s sign 0.63 0.42 0.81 0.76 0.58 0.89 5.31 1.74 16.2
Young78
No centralization of pain 0.9 0.79 0.97 0.2 0.13 0.25 2.59 0.54 12.06
3 or more positive tests 0.77 0.56 0.91 0.7 0.51 0.85 7.78 2.34 25.85
364 Diagnostic Validity of Criteria for Sacroiliac Joint Pain
Table 4. Pooled Results of the Diagnostic Validity of Tests
SENSITIVITY (CI) SPECIFICITY (CI) DOR (CI)

Compression test 0.628 (0.47-0.77) 0.692 (0.57-0.80) 3.885 (1.7-8.9)

Thigh thrust test 0.907 (0.78-0.97) 0.662 (0.53-0.77) 18.461 (5.82-58.53)
3 or more positive provocation tests 0.850 (0.75-0.92) 0.764 (0.68-0.84) 17.162 (7.6-39.0)

to the neighboring nerve structures and ligaments.30 test and the reference test should be chosen to limit
Consequently, there is a possibility that using this tech- the influence of confounders, such as additional therapy.
nique, more structures are targeted than the intra-syno- Consequently, we think that choosing a strict time inter-
vial space. Furthermore, according to the literature, val between both tests would make our results more
infiltration of interosseous ligament or the L4-S3 nerves, consistent.
have, respectively, diagnostic and prognostic ability in The question arises about the relevance of the results
patients with SI joint pain.15,77 At present, the use of SI of our review and whether they can contribute to the
joint injections for the treatment of chronic complaints improvement of the current practice. Included studies
of nonspecific origin from this joint is not recommended were performed mainly in university hospitals and/or
by the European guidelines because of limited evidence spine treatment units; their participants may differ
of their efficacy.1 However, the problem might be that from those in primary care, with respect to pain severity,
treatment effects depend on the correct differentiation chronicity and complexity of complaints and coexisting
between the structures contributing to SI joint pain pathology. However, reaching the threshold of 3 positive
and are likely to rely on blocking of pain-signaling struc- tests, it is possible to select a certain subgroup within
tures. Reports in the literature evaluating this subject are chronic low back pain patients, in which the diagnosis
contradictory. Early cadaveric studies reporting dorsal as of SI joint pain could be involved. Still, attention should
well as ventral innervation by the lumbar and sacral be paid to patients in whom singular test provokes famil-
nerves39,64 are not confirmed by a more recent report iar pain in SI joint, especially when using the compression
refuting ventral SI joint innervation.34 Subsequently, or thigh trust test. Positive response to the individual
(immuno-) histological studies of SI joint ligaments pain provocation tests for SI joint could indicate a need
showed the presence of sensory nerves in the ventral for further diagnostics, which could involve intra-
capsular ligament,28,66 the dorsal ligamentous tissue articular infiltrations with local anesthetics. However,
adjacent to the posterior superior iliac spine,29,72 and in using this technique, practitioners must be aware of its
the interosseous ligament.66 These findings, however, limitations.
only concern ligamentous structures, thus structures In view of the fact that a gold standard for SI joint pain
surrounding the space targeted by the diagnostic infil- diagnosis is lacking, the diagnostic validity of other tests
tration. Taking the limitations of the diagnostic infiltra- related to the IASP criteria for SI joint pain should be re-
tion into account, the diagnostic validity of other tests garded with care. Taking this limitation into account,
handled in our review is controversial. best evidence suggests that patients whereby at least
There are some discrepancies in the methodological 3 SI joint-selective stressing tests reproduce the patient’s
assessment of the trials between our systematic review pain could be regarded to have SI joint pain. Reproduc-
and other reviews in which the QUADAS tool has been tion of the patient’s pain with compression or thigh
used.35,37,60 Although the k of our quality assessment thrust test indicates the need for further diagnostics.
within review groups seems high, it is questionable if re- Relying only on the presence of pain in the SI joint region
liability across review groups is also high. The most ratio- could lead to over-diagnosing of SI joint pain and is
nal explanation for discrepancies between groups may therefore not recommended. Intra-articular injections
be the different interpretation of the QUADAS item appear to be valid, but their diagnostic validity should
list. Although the scoring of the QUADAS tool has been be studied further. Moreover, information with respect
widely described,74 it is necessary to incorporate the def- to the IASP criteria as a whole set is lacking, and further
inition of the items in the review.76 Also, the reliability of research is needed in this area. Finally, sources of SI
this instrument needs further testing. Furthermore, in joint pain other than intra-articular, for example, peri-
the present review, the 4th QUADAS item was consid- articular ligaments, should be considered for further
ered positive if the delay between the application of evaluation.
the index test and reference test used in the study was re-
ported and was not more than 7 days, and, additionally,
for SI joint infiltration with local anesthetics not shorter Acknowledgments
than 24 hours. For our definition, we considered the half- We express our greatest gratitude to Ingrid Riphagen
life of the anesthetics as the minimum period between (Medical Library, VU University Medical Center) for her
the first and second injections. The maximal period instructions and advice for the development of the liter-
between the index and the reference test is arbitrary ature search strategy. The work for this paper was per-
because to our knowledge there is no literature concern- formed in collaboration with the Pain Management
ing this point. However, in our opinion, for an individual and Research Center, University Hospital Maastricht,
study, a short-as-possible period between the index The Netherlands.
Szadek et al 365
References pain relief during diagnostic medial branch blocks: A multi-
center analysis. Spine J 8:498-504, 2008
1. Airaksinen O, Brox JI, Cedraschi C, Hildebrandt J, Klaber- 18. Cooperman JM, Riddle DL, Rothstein JM: Reliability and
Moffett J, Kovacs F, Mannion AF, Reis S, Staal JB, Ursin H, validity of judgments of the integrity of the anterior cruciate
Zanoli G: European guidelines for the management of ligament of the knee using the Lachman’s test. Phys Ther 70:
chronic nonspecific low back pain. Eur Spine J 15(Suppl 2): 225-233, 1990
S192-S300, 2006
19. Deville WLJM, Bezemer PD, Bouter LM: Publications on
2. Altman DG: Practical Statistics for medical research, New diagnostic test evaluation in family medicine journals: An
York, Chapman and Hall, 1999, pp 403-409. optimal search strategy. J Clin Epidemiol 53:65-69, 2000

3. Benzon HT, Katz JA, Benzon HA, Iqbal MS: Piriformis syn- 20. Dreyfuss P, Dreyer SJ, Cole A, Mayo K: Sacroiliac joint
drome: Anatomic considerations, a new injection technique, pain. J Am Acad Orthop Surg 12:255-265, 2004
and a review of the literature. Anesthesiology 98:1442-1448,
2003 21. Dreyfuss P, Dryer S, Griffin J, Hoffman J, Walsh N: Positive
sacroiliac screening tests in asymptomatic adults. Spine 19:
4. Bernard TN, Kirkaldy-Willis WH: Recognizing specific 1138-1143, 1994
characteristics of nonspecific low-back-pain. Clin Orthop
Rel Res 1987;266-280, 1987 22. Dreyfuss P, Michaelsen M, Pauza K, McLarty J,
Bogduk N: The value of medical history and physical exam-
5. Berthelot JM, Labat JJ, Le Goff B, Gouin F, Maugars Y: Pro- ination in diagnosing sacroiliac joint pain. Spine 21:
vocative sacroiliac joint maneuvers and sacroiliac joint block 2594-2602, 1996
are unreliable for diagnosing sacroiliac joint pain. Joint
Bone Spine 73:17-23, 2006 23. Foley BS, Buschbacher RM: Sacroiliac joint pain: Anat-
omy, biomechanics, diagnosis, and treatment. Am J Phys
6. Bogduk N: International Spinal Injection Society guide- Med Rehabil 85:997-1006, 2006
lines for the performance of spinal injection procedures,
I: Zygapophysial joint blocks. Clin J Pain 13:285-302, 1997 24. Forst SL, Wheeler MT, Fortin JD, Vilensky JA: The sacroil-
iac joint: Anatomy, physiology and clinical significance. Pain
7. Broadhurst NA, Bond MJ: Pain provocation tests for the Physician 9:61-67, 2006
assessment of sacroiliac joint dysfunction. J Spinal Disord
11:341-345, 1998 25. Fortin JD, Aprill CN, Ponthieux B, Pier J: Sacroiliac joint-
pain referral maps upon applying a new injection arthrogra-
8. Buijs E, Visser L, Groen G: Sciatica and the sacroiliac joint: phy technique: 2. Clinical evaluation. Spine 19:1483-1489,
a forgotten concept. Br J Anaesth 99:713-716, 2007 1994

9. Carmichael JP: Inter- and intra-examiner reliability of pal- 26. Fortin JD, Dwyer AP, West S, Pier J: Sacroiliac joint-pain
pation for sacroiliac joint dysfunction. J Manip Physiol Ther referral maps upon applying a new injection arthrography
10:164-171, 1987 technique: 1. Asymptomatic volunteers. Spine 19:1475-1482,
1994
10. Cattley P, Winyard J, Trevaskis J, Eaton S: Validity and
reliability of clinical tests for the sacroiliac joint: A re- 27. Fortin JD, Falco FJ: The Fortin finger test: An indicator of
view of literature. Australas Chiropr Osteopathy 10:80, sacroiliac pain. Am J Orthop 26:477-480, 1997
2002
28. Fortin JD, Kissling RO, O’Connor BL, Vilensky JA: Sacroil-
11. Chen YC, Fredericson M, Smuck M: Sacroiliac joint pain iac joint innervation and pain. Am J Orthop 28:687-690, 1999
syndrome in active patients: A look behind the pain. Physi-
29. Fortin JD, Vilensky JA, Merkel GJ: Can the sacroiliac joint
cian Sports Med 30:30-37, 2002
cause sciatica? Pain Physician 6:269-271, 2003
12. Chou LH, Slipman CW, Bhagia SM, Tsaur L, Bhat AL, 30. Fortin JD, Washington WJ, Falco FJ: Three pathways be-
Isaac Z, Gilchrist R, El Abd OH, Lenrow DA: Inciting events ini- tween the sacroiliac joint and neural structures. AJNR Am J
tiating injection-proven sacroiliac joint syndrome. Pain Med Neuroradiol 20:1429-1434, 1999
5:26-32, 2004
31. Freburger JK, Riddle DL: Using published evidence to
13. Cibulka MT, Koldehoff R: Clinical usefulness of a cluster guide the examination of the sacroiliac joint region. Phys
of sacroiliac joint tests in patients with and without low back Ther 81:1135-1143, 2001
pain. J Orthop Sports Phys Ther 29:83-89, 1999
32. Fukui S, Nosaka S: Pain patterns originating from the sa-
14. Cohen SP: Sacroiliac joint pain: A comprehensive review croiliac joints. J Anesth 16:245-247, 2002
of anatomy, diagnosis, and treatment. Anesth Analg 101:
1440-1453, 2005 33. Glas AS, Lijmer JG, Prins MH, Bonsel GJ, Bossuyt PM: The
diagnostic odds ratio: A single indicator of test perfor-
15. Cohen SP, Abdi S: Lateral branch blocks as a treatment mance. J Clin Epidemiol 56:1129-1135, 2003
for sacroiliac joint pain: A pilot study. Reg Anesth Pain
Med 28:113-119, 2003 34. Grob KR, Neuhuber WL, Kissling RO: Innervation of
the human sacroiliacal joint. Z Rheumatol 54:117-122,
16. Cohen SP, Bajwa ZH, Kraemer JJ, Dragovich A, 199.
Williams KA, Stream J, Sireci A, McKnight G, Hurley RW: Fac-
tors predicting success and failure for cervical facet radiofre- 35. Hancock MJ, Maher CG, Latimer J, Spindler MF,
quency denervation: A multi-center analysis. Reg Anesth McAuley JH, Laslett M, Bogduk N: Systematic review of tests
Pain Med 32:495-503, 2007 to identify the disc, SIJ or facet joint as the source of low
back pain. Eur Spine J 10:165-184, 2007
17. Cohen SP, Stojanovic MP, Crooks M, Kim P, Schmidt RK,
Shields CH, Croll S, Hurley RW: Lumbar zygapophysial (facet) 36. Hansen HC, Helm S: Sacroiliac joint pain and dysfunc-
joint radiofrequency denervation success as a function of tion. Pain Physician 6:179-189, 2003
366 Diagnostic Validity of Criteria for Sacroiliac Joint Pain
37. Hansen HC, McKenzie-Brown AM, Cohen SP, 56. Riddle DL, Freburger JK: Evaluation of the presence of
Swicegood JR, Colson JD, Manchikanti L: Sacroiliac joint sacroiliac joint region dysfunction using a combination of
interventions: A systematic review. Pain Physician 10: tests: A multicenter intertester reliability study. Phys Ther
165-184, 2007 82:772-781, 2002

38. Higgins JP, Thompson SG, Deeks JJ, Altman DG: Measur- 57. Robinson HS, Brox JI, Robinson R, Bjelland E, Solem S,
ing inconsistency in meta-analyses. BMJ 327:557-560, 2003 Telje T: The reliability of selected motion- and pain provo-
cation tests for the sacroiliac joint. Manip Ther 12:72-79,
39. Ikeda R: Innervation of the sacroiliac joint: Macroscopi- 2007
cal and histological studies. Nippon Ika Daigaku Zasshi 58:
587-596, 1991 58. Rubinstein SM, Pool JJ, van Tulder MW, Riphagen II, de
Vet HC: A systematic review of the diagnostic accuracy of
40. Katz V, Schofferman J, Reynolds J: The sacroiliac joint: provocative tests of the neck for diagnosing cervical radicul-
A potential cause of pain after lumbar fusion to the sacrum. opathy. Eur Spine J 16:307-319, 2007
J Spinal Disord Tech 16:96-99, 2003
59. Schwarzer AC, Aprill CN, Bogduk N: The sacroiliac joint
41. Kirkaldy-Willis WH, Hill RJ: A more precise diagnosis for in chronic low-back-pain. Spine 20:31-37, 1995
low-back pain. Spine 4:102-109, 1979
60. Simpson R, Gemmell H: Accuracy of spinal orthopaedic
42. Kokmeyer DJ, van der Wurff P, Aufdemkampe G, tests: A systematic review. Chiropr Osteopat 14:26, 2006
Fickenscher TCM: The reliability of multitest regimens with
sacroiliac pain provocation tests. J Manip Physiol Ther 25: 61. Slipman CW, Jackson HB, Lipetz JS, Chan KT, Lenrow D,
42-48, 2002 Vresilovic EJ: Sacroiliac joint pain referral zones. Arch Phys
Med Rehabil 81:334-338, 2000
43. Laslett M: Pain provocation tests for diagnosis of sacro-
iliac joint pain. Aust J Physiother 52:229, 2006 62. Slipman CW, Sterenfeld EB, Chou LH, Herzog R,
Vresilovic E: The value of radionuclide imaging in the diag-
44. Laslett M, Aprill CN, McDonald B, Young SB: Diagnosis of nosis of sacroiliac joint syndrome. Spine 21:2251-2254, 1996
sacroiliac joint pain: Validity of individual provocation tests
and composites of tests. Manip Ther 10:207-218, 2005 63. Slipman CW, Sterenfeld EB, Chou LH, Herzog R,
Vresilovic E: The predictive value of provocative sacroiliac
45. Laslett M, Young SB, Aprill CN, McDonald B: Diagnosing joint stress maneuvers in the diagnosis of sacroiliac joint syn-
painful sacroiliac joints: A validity study of a McKenzie eval- drome. Arch Phys Med Rehabil 79:288-292, 1998
uation and sacroiliac provocation tests. Aust J Physiother 49:
89-97, 2003 64. Solonen KA: The sacroiliac joint in the light of anatomi-
cal, roentgenological and clinical studies. Acta Orthop
46. Lee JK, Yao L, Phelps CT, Wirth CR, Czajka J, Lozman J: An-
Scand 27:1-127, 1957
terior cruciate ligament tears: MR imaging compared with
arthroscopy and clinical tests. Radiology 166:861-864, 1988 65. Stuber KJ: Specificity, sensitivity, and predictive values of
47. Maigne JY, Aivaliklis A, Pfefer F: Results of sacroiliac clinical tests of the sacroiliac joint: A systematic review of the
joint double block and value of sacroiliac pain provocation literature. J Can Chiropr Assoc 51:30-41, 2007
tests in 54 patients with low back pain. Spine 21:
66. Szadek KM, Hoogland PV, Zuurmond WW, de Lange JJ,
1889-1892, 1996
Perez RS: Nociceptive nerve fibers in the sacroiliac joint in
48. Maigne JY, Boulahdour H, Chatellier G: Value of quanti- humans. Regul Anesth Pain Med 33:36-43, 2008
tative radionuclide bone scanning in the diagnosis of sacro-
iliac joint syndrome in 32 patients with low back pain. Eur 67. Toussaint R, Gawlik CS, Rehder U, Ruther W: Sacroiliac
Spine J 7:328-331, 1998 joint diagnostics in the Hamburg construction workers
study. J Manip Physiol Ther 22:139-143, 1999
49. Maigne JY, Planchon CA: Sacroiliac joint pain after lum-
bar fusion: A study with anesthetic blocks. Eur Spine J 14: 68. van der Wurff P, Buijs EJ, Groen GJ: A multitest regimen
654-658, 2005 of pain provocation tests as an aid to reduce unnecessary
minimally invasive sacroiliac joint procedures. Arch Phys
50. Manchikanti L, Singh V, Pampati V, Damron KS, Med Rehabil 87:10-14, 2006
Barnhill RC, Beyer C, Cash KA: Evaluation of the relative con-
tributions of various structures in chronic low back pain. 69. van der Wurff P, Buijs EJ, Groen GJ: Intensity mapping of
Pain Physician 4:308-316, 2001 pain referral areas in sacroiliac joint pain patients. J Manip
Physiol Ther 29:190-195, 2006
51. McKenzie-Brown AM, Shah RV, Sehgal N, Everett CR: A
systematic review of sacroiliac joint interventions. Pain Phy- 70. van der Wurff P, Hagmeijer RHM, Meyne W: Clinical tests
sician 8:115-125, 2005 of the sacroiliac joint: A systematic methodological review,
1: Reliability. Manip Ther 5:30-36, 2000
52. Merskey H, Bogduk N: Classification of chronic pain: De-
scriptions of chronic pain syndromes and definitions of pain 71. van der Wurff P, Meyne W, Hagmeijer RHM: Clinical tests
terms, Seattle, WA, IASP Press, 1994. pp 190-191 of the sacroiliac joint: A systematic methodological review,
2: Validity. Manip Ther 5:89-96, 2000
53. Pool-Goudzwaard AL, Kleinrensink GJ, Snijders CJ,
Entius C, Stoeckart R: The sacroiliac part of the iliolumbar 72. Vilensky JA, O’Connor BL, Fortin JD, Merkel GJ,
ligament. J Anat 199:457-463, 2001 Jimenez AM, Scofield BA, Kleiner JB: Histologic analysis of
neural elements in the human sacroiliac joint. Spine 27:
54. Potter NA, Rothstein JM: Intertester reliability for se- 1202-1207, 2002
lected clinical tests of the sacroiliac joint. Phys Ther 65:
1671-1675, 1985 73. Weiner DK, Sakamoto S, Perera S, Breuer P: Chronic low
back pain in older adults: Prevalence, reliability, and validity
55. Pulisetti D, Ebraheim NA: CT-guided sacroiliac joint in- of physical examination findings. J Am Geriatr Soc 54:11-20,
jections. J Spinal Disord 12:310-312, 1999 2006
Szadek et al 367
74. Whiting P, Rutjes AW, Dinnes J, Reitsma JB, Bossuyt PM, 77. Yin W, Willard F, Carreiro J, Dreyfuss P: Sensory stimula-
Kleijnen J: Development and validation of methods for as- tion-guided sacroiliac joint radiofrequency neurotomy:
sessing the quality of diagnostic accuracy studies. Health Technique based on neuroanatomy of the dorsal sacral
Technol Assess 8:1-234, 2004 plexus. Spine 28:2419-2425, 2003

78. Young S, Aprill C, Laslett M: Correlation of clinical exam-

75. Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, ination characteristics with three sources of chronic low
Kleijnen J: The development of QUADAS: A tool for the back pain. Spine J 3:460-465, 2003
quality assessment of studies of diagnostic accuracy in-
cluded in systematic reviews. BMC Med Res Methodol 10:
25, 2003 79. Zamora J, Abraira V, Muriel A, Khan K, Coomarasamy A:
Meta-DiSc: A software for meta-analysis of test accuracy
data. BMC Med Res Methodol 12:31, 2006
76. Whiting PF, Weswood ME, Rutjes AW, Reitsma JB,
Bossuyt PN, Kleijnen J: Evaluation of QUADAS, a tool for 80. Zelle BA, Gruen GS, Brown S, George S: Sacroiliac joint
the quality assessment of diagnostic accuracy studies. BMC dysfunction: Evaluation and management. Clin J Pain 21:
Med Res Methodol 6:9, 2006 446-455, 2005
368 Diagnostic Validity of Criteria for Sacroiliac Joint Pain
Appendix 1. Search Strategy Used to Identify Studies on Diagnostic Tests for SI Joint Pain
PubMed Backache/exp OR Pain-assessment/exp OR Pain-Clinics/
exp OR pelvis-pain-syndrome/de OR Neuralgia/exp OR

S
earch conducted on August 29, 2007. (‘‘Sacroiliac
Joint’’[MeSH] OR ‘‘Lumbosacral Region’’[MeSH] OR Arthralgia/exp OR musculoskeletal-pain/de OR pain:ti,ab
sij[tiab] OR ((si[tiab] OR sacroiliac[tiab]) AND OR arthralgia:ti,ab OR neuralgia:ti,ab OR backache:ti,ab
joint*[tiab]) OR (lumbosacral[tiab] AND region[tiab])) OR somatosensory-disorder/exp OR ’somatoform disor-
AND der’/de OR ’psychosomatic disorder’/de OR psychogenic-
(‘‘Pain’’[MeSH] OR ‘‘Pain Measurement’’[MeSH] OR ‘‘Pain pain/de
Clinics’’[MeSH] OR ‘‘Low Back Pain’’[MeSH] OR ‘‘Back AND
Pain’’[MeSH] OR ‘‘Myofascial Pain Syndromes’’[MeSH] OR injection/exp OR ’intraarticular drug administration’/
‘‘Pain, Referred’’[MeSH] OR ‘‘Pain Threshold’’[MeSH] OR exp OR inject*:ti,ab OR preinjection*:ti,ab OR infiltrat*:-
‘‘Pain, Intractable’’[MeSH] OR ‘‘Pelvic Pain’’[MeSH] ti,ab OR (pain:ti,ab AND (pattern*:ti,ab OR mapping*:-
OR ‘‘Pain, Postoperative’’[MeSH] OR ‘‘Somatosensory Dis- ti,ab)) OR algometr*:ti,ab OR ’Pain assessment’/exp OR
orders’’[MeSH] OR ‘‘Neuralgia’’[MeSH] OR ‘‘Arthralgia’’ ’sensory system examination’/de OR provocation:ti,ab
[MeSH] OR ‘‘Somatoform Disorders’’[MeSH] OR pain[tiab]) OR block*:ti,ab OR test:ti,ab OR tests:ti,ab OR standard:-
AND ti,ab OR standards*:ti,ab OR ’nerve block’/de OR ’physi-
(‘‘Injections’’[Mesh:NoExp] OR ‘‘Injections, Intra-Artic- cal examination’/de
ular’’[Mesh] OR ‘‘Injections, Spinal’’[Mesh:NoExp] OR in- AND
ject*[tiab] OR preinjection*[tiab] OR infiltrat*[tiab] OR Diagnos* OR etiolog* OR Diagnosis/de OR ’diagnostic
(pain[tiab] AND (pattern*[tiab] OR mapping*[tiab])) accuracy’/de OR ’diagnostic value’/de OR ’differential di-
OR algometr*[tiab] OR Pain measurement[mesh] OR agnosis’/de OR ’diagnostic test’/de OR ’physical examina-
provocation[tw] OR block*[tw] OR test[tiab] OR test- tion’/de OR ’Pain assessment’/exp OR ’sensory system
s[tiab] OR standard[tiab] OR standards*[tiab]) examination’/de OR ’gold standard’/exp
AND AND
(Diagnosis[mesh] OR Diagnosis[sh] OR Etiology[sh] OR ’human’/de.
diagnos*[tw] OR specificity[Title/Abstract] OR ‘‘Sensitiv-
ity and Specificity’’[MeSH]) CINAHL
NOT Search conducted on September 9, 2007.
(animals[mesh] NOT humans[mesh]) [(sacroiliac joint) or (MM ‘‘Sacroiliac Joint’’)] and [(low
NOT back pain) or (MM ‘‘Low Back Pain’’) or (MM ‘‘Back Pain’’)
case reports[pt]. or (arthralgia) or (MH ‘‘Arthralgia1’’)] and [(injection) or
(MM ‘‘Injections, Intraarticular’’) or (block) or (MH ‘‘Nerve
EMBASE Block’’) or (MM ‘‘Pain Measurement’’) or (MM ‘‘Algome-
Search conducted on September 9, 2007. try’’) or (diagnostic test) or (MH ‘‘Diagnostic Tests, Rou-
Sacroiliac-joint/de OR Lumbosacral-spine/de OR tine’’)] and [(diagnosis) or (MM ‘‘Diagnosis’’) or (MM
sij:ti,ab OR ((si:ti,ab OR sacroiliac:ti,ab) AND joint*:ti,ab) ‘‘Diagnosis, Differential’’) or (MM ‘‘Sensitivity and
OR (lumbosacral:ti,ab AND region:ti,ab) Specificity/MT/ST’’)] not (MM ‘‘Animals’’) not (MM ‘‘Case
AND Studies’’).