Session Information
Session Type: Abstract Submissions (ACR)
Background/Purpose:
Berlin criteria are a diagnostic algorithm based on calculation of the likelihood ratio (LR) product of currently available diagnostic test for spondyloarthritis (SpA). However, these probabilities were calculated on a longstanding population, with different features from those of an early SpA cohort. ESPeranza (ESP) is a national health care multicentre program for the early diagnosis and treatment of SpA. ESP patient characteristics were collected from baseline to onset, so it permits to explore Bayesian probabilities of each feature and reproduce Berlin algorithm. OBJECTIVEs: 1. To evaluate, in an early SpA, the LR of different symptoms and signs included in Berlin criteria. 2. To evaluate the validity of Berlin algorithm for SpA diagnosis.
Methods:
Patients were referred from general practitioners according with this criteria: 1) age <45years old; 2) symptom duration 3-24 months; 3) at least one of the following: inflammatory back pain (IBP), asymmetrical arthritis, SpA features associated to back pain or arthralgias (psoriasis, inflammatory bowel disease, anterior uveitis, radiographic sacroiliitis, HLA B27 positive or a family history of SpA). Patients with axial symptoms were selected (those with only arthritis were excluded) and categorized according to experts clinical diagnosis (SpA or not SpA). Data from ESP included: demographics (gender, age), clinical features (IBP characteristics, sacroiliac syndrome, enthesitis, arthritis, dactilitis, psoriasis, inflammatory bowel disease, diarrhea, urethritis, cervicitis, prostatitis, positive family history for SpA or good response to NSAIDs), HLA B27 or imaging data (sacroiliitis in x-ray or MRI, according to Omeract). Descriptive analysis was performed. Sensitivity and specificity, and LR positive were calculated. Classification according to Berlin LR product method was performed. In case of an LR product ≥200, a classification of axial SpA is made.
Results:
From the 1179 patients referred in ESP program, only 422 fulfilled inclusion criteria with axial symptoms. 316 were diagnosed as SpA. Insidious onset and > 3months onset were the most frequent symptoms (73% and 84%, respectively). The most useful feature among IBP was the improvement with exercise (LR+ 2.2), more frequent than alternating buttock pain. Arthritis was the most useful symptom for the diagnosis (LR+ 7), apart from MRI, following by enthesitis, but heel pain (LR 5.7). Dactiyitis and uveitis and other features were less helpful, (see table). Finally, MRI was the most beneficial feature (LR+ 16), although it was not always available. Berlin criteria applied to this early SpA population showed a lower sensitivity (65%) and higher specificity (98%) than Berlin group.
|
Table |
Frequent (n) |
Percentage (%) |
SENS % |
SPE % |
PPV |
PNV |
LR+ |
LR- |
|
Inflammatory back pain |
247 |
58.5 |
63.9 |
57.5 |
81.8 |
34.9 |
1.5 |
0.6 |
|
>3 month onset |
356 |
84.4 |
83.2 |
12.3 |
73.9 |
19.7 |
0.9 |
1.4 |
|
Morning stiffness ³30min |
241 |
57.1 |
59.2 |
49.1 |
77.6 |
28.7 |
1.2 |
0.8 |
|
Improvement with exercise |
230 |
54.5 |
63.0 |
70.8 |
86.5 |
39.1 |
2.2 |
0.5 |
|
No improvement with rest |
262 |
62.1 |
67.7 |
54.7 |
81.7 |
36.3 |
1.5 |
0.6 |
|
Pain at night |
195 |
46.2 |
50.9 |
67.9 |
82.6 |
31.7 |
1.6 |
0.7 |
|
Alternating buttock pain |
126 |
29.9 |
33.9 |
82.1 |
84.9 |
29.4 |
1.9 |
0.8 |
|
Insidious onset |
309 |
73.2 |
74.4 |
30.2 |
76.1 |
28.3 |
1.1 |
0.8 |
|
Arthritis |
87 |
20.6 |
26.3 |
96.2 |
95.4 |
30.4 |
7.0 |
0.8 |
|
Heel pain |
75 |
17.8 |
21.2 |
92.5 |
89.3 |
28.2 |
2.8 |
0.9 |
|
Enthesitis, but heel pain |
18 |
4.3 |
5.4 |
99.1 |
94.4 |
26.0 |
5.7 |
1.0 |
|
Iritis |
17 |
4.1 |
4.4 |
97.2 |
82.4 |
25.4 |
1.6 |
1.0 |
|
Dactylitis |
18 |
4.3 |
5.7 |
100.0 |
100.0 |
26.2 |
. |
0.9 |
|
Inflammatory bowel disease |
16 |
3.8 |
4.4 |
98.1 |
87.5 |
25.6 |
2.4 |
1.0 |
|
Psoriasis |
35 |
8.3 |
9.2 |
94.3 |
82.9 |
25.8 |
1.6 |
1.0 |
|
Uretritis/cervicitis/diarrhea |
7 |
1.7 |
2.2 |
100.0 |
100.0 |
25.5 |
. |
1.0 |
|
Raised CRP |
97 |
24.6 |
29.8 |
92.4 |
92.8 |
28.6 |
3.9 |
0.8 |
|
HLA B27 positive |
208 |
49.3 |
59.2 |
80.2 |
89.9 |
39.7 |
3.0 |
0.5 |
|
Family history |
99 |
23.5 |
27.5 |
88.7 |
87.9 |
29.1 |
2.4 |
0.8 |
|
Sacroiliitis X-ray |
203 |
48.1 |
36.8 |
100.0 |
100.0 |
30.0 |
. |
0.6 |
|
Sacroiliitis RMI |
99 |
23.5 |
63.0 |
96.2 |
98.0 |
46.6 |
16.7 |
0.4 |
|
Good response to NSAID |
252 |
59.7 |
63.6 |
51.9 |
79.8 |
32.4 |
1.3 |
0.7 |
|
Berlin Criteria |
210 |
49.7 |
65.2 |
96.2 |
98.1 |
48.1 |
17.3 |
0.4 |
Conclusion:
Our data shows that MRI is a key symptom in early SpA, while HLA B27 was less useful than results from Berlin group. Berlin diagnostic algorithm should be revisited according to these data, at least in early SpA patients.
Disclosure:
B. E. Joven,
None;
M. Gobbo,
None;
M. A. Descalzo,
None;
E. De Miguel,
None;
ACR Meeting Abstracts - https://acrabstracts.org/abstract/performance-of-berlin-criteria-in-patients-with-early-spondyloarthritis/