Characteristics of the study population
A total of 246 patients with pSS who fulfilled the AECG or ACR criteria were included (female/male ratio 239/7; age 53.16 (SD 12.13) years; symptom duration 4 (range 0.1–31) years). The control group comprised a total of 167 subjects: 113 non-SS patients with various rheumatic diseases (29 with undifferentiated connective tissue disease, 32 with connective tissue disease, 2 with mixed connective tissue disease, 28 with rheumatoid arthritis, 13 with systemic lupus erythematosus, 1 with sclerosis, 1 with systemic vasculitis, 7 with primary biliary cirrhosis), 27 with idiopathic Sicca syndrome and 27 healthy subjects (female/male ratio 156/11; age 51.21 (S.D. 13.72) years). The two groups were matched by age and the sex ratio (P = 0.139 and 0.068, respectively). The characteristics of the pSS and non-pSS subjects are shown in Table 1.
Table 1
Characteristics of the study population
Characteristics
|
pSS patients
(n = 246)
|
non-pSS subjects (n = 167)
|
Healthy subjects
(n = 27)
|
non-pSS patients
(n = 140)
|
Age, mean (S.D.), years
|
53.16 (12.13)
|
51.21 (13.72)
|
Female/male
|
239/7
|
156/11
|
Symptom duration
Median (range), years
|
4 (0.1–31)*
|
0
|
2 (0.1–30)
|
Parotid enlargement
|
75 (30.5)*
|
0
|
3 (2.1)
|
Dental loss
|
95 (38.6)*
|
0
|
4 (2.9)
|
Schirmer < 5 mma
|
128/129*
|
0
|
17/33
|
Anti-SSAa
|
176/223*
|
0
|
44/130
|
Anti-SSBa
|
89/224*
|
0
|
15/125
|
MSG biopsya
|
37/52*
|
0
|
0/1
|
Except where indicated otherwise, the values are presented as the mean (S.D.), median (range) or number (%). aValues of objective tests given as rates of positive results (positive/total). The P values were determined using the Mann-Whitney test or chi-square test, as appropriate; *P < 0.001: statistically significant (regarding the difference between pSS patients and non-pSS subjects). MSG: minor salivary gland. |
Features of SGUS in pSS
All patients and control group underwent SGUS scores which displayed no difference between bilateral parotid glands (P = 0.798), bilateral submandibular glands (P = 0.842) followed by the combination of left parotid and submandibular glands compared with the right parotid and submandibular glands (P = 0.966; Fig. 1A). The US scores of the right parotid and submandibular glands, the scores of the left parotid and submandibular glands and the sum of the scores of all four glands in the pSS group were significantly higher than those in the non-pSS group (P < 0.001, overall) (Fig. 1B).
Subsequently, to further assess the associations between the SGUS scores and disease duration, the pSS patients were further stratified according to symptom duration (symptom duration: first stage ≤ 5 years (n = 151), second stage 5–10 years (n = 35) and third stage ≥ 10 years (n = 60). After the analysis based on one side among diverse groups, data showed a trend to have significant differences between the first and second stages and between the first and third stages (p = 0.042, 0.002, respectively) but not between the second and third stages (P = 0.345) (Fig. 2). Besides, when connecting SGUS scores to age or gender, there also witnessed no statistical difference among all the subjects in our study.
SGUS diagnostic accuracy for pSS
The diagnostic accuracy of the US scores of all four glands in diagnosing pSS was excellent (AUC = 0.908, 95% CI 0.879, 0.938). Accuracy was comparable when the scores of only the right parotid and submandibular glands (AUC = 0.910, 95% CI 0.881, 0.939) or the scores of only the left parotid and submandibular glands (AUC = 0.904, 95% CI 0.875, 0.934) were included (P > 0.05, overall) (Table 2). A US cut-off value of 7 provided maximal sensitivity (78%) and specificity (91.6%) for the scores of all four glands, while a US cut-off value of 4 showed maximal sensitivity (77.2% and 75.6%, respectively) and specificity (92.2% and 91.6%, respectively) for the scores of the parotid and submandibular glands on the right or left side (Table 2), so these values were thought optimal.
Table 2
Diagnostic accuracy of the OMERACT scoring system for pSS
|
Area under the curve
(95% CI)
|
Cut-off value
|
SEN, %
|
SPE,%
|
PPV,%
|
NPV,%
|
right PG
|
0.90
(0.869–0.930)
|
2
|
77.6
|
88.6
|
91.4
|
72.9
|
right SMG
|
0.879
(0.845–0.912)
|
2
|
82.9
|
79.6
|
85.7
|
76
|
left PG
|
0.893
(0.861–0.924)
|
2
|
76.0
|
89.8
|
91.7
|
71.8
|
left SMG
|
0.874
(0.84–0.908)
|
2
|
83.3
|
77.8
|
84.7
|
76.0
|
right PG + SMGa
|
0.910
(0.881–0.939)
|
4
|
77.2
|
92.2
|
93.6
|
73.3
|
left PG + SMGb
|
0.904
(0.875–0.934)
|
4
|
75.6
|
91.6
|
93
|
71.8
|
total four glandsc
|
0.908
(0.879–0.938)
|
7
|
78
|
91.6
|
93.2
|
73.9
|
P < 0.05 determined by the McNemar test was considered statistically significant. a vs b, P = 0.508; b vs c, P = 0.062, a vs c, p = 0.625. PG: parotid gland; SMG: submandibular gland; SEN: sensitivity; SPE: specificity; PPV: positive predictive value; NPV: negative predictive value. |
SGUS in clinical stratification
The cut-off value of the SGUS scores of the parotid and submandibular glands on one side allowed for the classification of patients with pSS: the positive SGUS group (SGUS score ≥ 4) and the negative SGUS group (SGUS score < 4). The results suggested that 190 (77.2%) patients had positive SGUS findings, and 56 (22.8%) patients had negative SGUS findings. It was observed that the correlation between the positive and negative groups with respect to systemic complications including cutaneous vasculitis, Interstitial lung disease and other extra-glandular manifestations, did not reach significance (P > 0.05, overall). However, one patient in the SGUS-positive group had cryoglobulinemia and amyloidosis. Parotid swelling and dental loss were found to occur more frequently in patients with positive scores than in those with negative scores (P = 0.008, 0.001 respectively) (Table 3).
Table 3
Characteristics of pSS patients with positive and negative SGUS results
|
Negative
(SGUS < 4)
|
Positive
(SGUS ≥ 4)
|
P
|
N (%)
|
56 (22.8)
|
190 (77.2)
|
-
|
Age (S.D.), years
|
53.16 (12.99)
|
53.16 (11.91)
|
0.92
|
Female/male
|
54/2
|
185/5
|
0.71
|
Symptom duration
median (range), years
|
2 (0.1–27)
|
5 (0.1–31)
|
< 0.001
|
Parotid swelling
|
9 (16.1)
|
66 (34.7)
|
0.008
|
Dental loss
|
11 (19.6)
|
84 (44.2)
|
0.001
|
ANA ≥ 1:320
|
25 (48.1)
|
105 (61.8)
|
0.08
|
Anti-SSA positivity
|
33 (66.0)
|
143 (82.7)
|
0.011
|
Anti-SSB positivity
|
12 (24.0)
|
77 (44.3)
|
0.01
|
Positive RF
|
29 (58)
|
130 (81.8)
|
0.001
|
IgG, g/l
|
17.1 (6.6)
|
21.37 (8.34)
|
0.001
|
γ-globulin,%
|
21.74 (6.84)
|
25.62 (6.47)
|
0.01
|
C3, median (range), g/l
|
1.15 (0.88–1.49)
|
0.96 (0.64–1.66)
|
0.07
|
C4, median (range), g/l
|
0.18 (0.12–0.38)
|
0.17 (0.12–0.34)
|
0.951
|
Systemic complications
|
Cutaneous vasculitis
|
6 (10.7)
|
25 (13.2)
|
0.628
|
Interstitial lung disease
|
14 (25)
|
50 (26.3)
|
0.844
|
Renal involvement
|
3 (5.4)
|
20 (10.5)
|
0.243
|
Nervous system involvement
|
3 (5.4)
|
10 (5.3)
|
0.978
|
Leukopenia
|
14 (25)
|
59 (31.1)
|
0.384
|
Cryoglobulinemia
|
0
|
1 (0.5)a
|
0.586
|
Except where indicated otherwise, the values are presented as the mean (S.D.), median (range) or number (%). The P values were determined using the Mann-Whitney test or chi-square test, as appropriate. aOne patient had cryoglobulinemia and amyloidosis. PG: parotid gland; SMG: submandibular gland; ANA: antinuclear antibodies; RF: rheumatoid factor; MSG: minor salivary gland. |
Associations between SGUS and serological parameters
In light of the specific value of serological characteristics in diagnosing autoimmune diseases, the pSS patients with positive scores expressed a preference for autoantibodies to SSA and SSB and obviously higher levels of RF, IgG and γ-globulin% (P < 0.05, overall) (Table 3). However, the ANA results along with the level of complement in the positive SGUS group shared the similar information with the negative SGUS group (P > 0.05) (Table 3).
Given that the presence of autoantibodies is available for the classification of pSS and daily clinical practice, 239 individuals were evaluated with serum levels of anti-SSA or anti-SSB. Data showed that patients with any of antibodies were significantly involved in the higher SGUS scores as compared to patients with negative outcomes (P<0.05) (Fig. 3A). Additionally, all the samples with positive antibodies were distributed into three groups including only anti-SSA, only anti-SSB antibodies and both positive. Our analysis revealed no significant difference among them (P = 0.982) (Fig. 3B).
SGUS as an alternative item in 2012 classification criteria
In our study, fifty-two patients in the pSS group underwent labial gland biopsies. Among the 52 patients, SGUS score ≥ 4 of the parotid and submandibular glands on one side and positive biopsy findings were recorded in 38/52 (73.1%) and 37/52 (71.2%) patients, respectively (P > 0.05). The absolute agreement between SGUS scores and labial gland biopsy was 67.3% (35/52). However, as invasive lip biopsy was difficult to apply as regular examination techniques, we next tested whether substituting US for lip biopsy of the three ACR classification items influenced diagnostic accuracy. Because not all patients were tested for all three ACR classification items, we only analysed 224 pSS patients and 130 non-pSS patients. Incorporating the US criteria (SGUS ≥ 4 on one side) as an alternative to lip biopsy of the ACR classification items achieved 91.1% sensitivity and 88.4% specificity, which was comparable to that of the original ACR classification (Table 4).
Table 4
Diagnostic ability of SGUS as an alternative to labial biopsy in ACR classification
|
pSS patients
(n = 224)
|
non-pSS patients
(n = 130)
|
positive RF plus ANA ≥ 1:320
|
89
|
10
|
Positive Anti-SSA
|
176
|
44
|
Positive Anti-SSB
|
89
|
15
|
OSS > 4
|
110
|
12
|
SGUS ≥ 4 on either side
|
173
|
9
|
lip biopsy replaced with SGUS in ACR
|
204
|
15
|
Diagnostic ability
|
SEN 91.1%
|
SPE 88.4%
|
OSS: ocular staining score; SEN: sensitivity; SPE: specificity. |