Baseline features and outcomes of anti-MDA5 positive patients
A total of 90 anti-MDA5 positive patients were included in our study. The demographic features, antibody variables and treatment regimens on admission, and the clinical outcomes were recorded in Table1. The average ageat disease onset was51.9 ± 12.1 years, and female accounted for 63.3% of the cohort. This cohort consisted of 70(77.8%) CADM,19(21.1%) DMand 1(1.1%) PM cases, among whom the CADM patients accounted for the majority.All the patients were negative for other MSAs except anti-MDA5 antibody. Meanwhile, 64(71.1%) patients were also positive for anti-Ro52 antibody, indicatingthe necessity to testMSAs in patients positive for anti-Ro52.
Although almost all patients were diagnosed in early stage at a median timeof 3.7months, 81(90%) patients already presented ILD on admission and 35(38.9%) patients developed RP-ILD subsequently. During the first six months follow-up,a high mortality rate of 24.4% was observed in our cohort. Patients all died of respiratory failure caused by RP-ILD, some of whom werecomplicated with infections in the end stage,atan average time of 6.6 ± 5.9 weeks(range 1-24weeks)after diagnosis.
During the first six months follow-up, malignancieswere found inthreeCADM patients:one withesophageal cancer, one with lung adenocarcinoma andonewith thyroid carcinoma. However, no patient died of malignancy in the first six months.
Treatment Regimens
All patients were treated with high-doseGCat diagnosis, of whom 58 cases were initially combined with CNIand/orCYC and 32 cases were step-up treated (Table 1). In our cohort, CNI was a preferable choice than CYCbut without significant difference(41.1% vs.31.1%, P=0.16). Only seven patients(7.8%) were treated with a triple combination of high-dose GC with CYC and CNI, due to the concern of serious infections. Additional IVIG(0.4g/kg daily) was administered in 43 cases based on physician’s decision after comprehensive assessment of patients’ condition.
Comparison of clinical features between survivors and non-survivors
No significant difference was observed in other demographic features, exceptthenon-survivorswere significantly older at theage of disease onset than the survivors (Table 2). Clinically,asignificant higher incidence of fever anda lower incidence of DM rashes,such asheliotrope sign and neck V sign,were observedin the death group. As to the laboratory data, the death group presented a significant higher initial serum levels of ferritin, KL-6, LDH, ESR and CRP,suggesting a more intensiveinflammation inside, than the survival group. The non-survivors had a significant higher incidence of anti-Ro52 positivity than the survivors, reminding physicians to pay attention to the patients double-positivefor anti-MDA5 and anti-Ro52. Not surprisingly, the non-survivors had significant worse HRCT presentations on admission, both in total CT fibrosis score and total GGO score. Initial combination therapy was more common in the survivors, butthe use of IVIG was of no significant difference between the two groups.
Initial prognostic factors in anti-MDA5 positive patients
To identify the prognostic factors associated with death in anti-MDA5 positive patients, we performed univariate analysis using all the initial variables. The cut-off points were determined using the ROC analysis, and then all continuous variables were converted to dichotomous variables for analysis.In the univariate analysis, several variables, including demographic, clinical, laboratory and image parameters, were found significantly related to death, while initial combination related to survival (Table 3).Variables with P< 0.1 in the univariate analysis were sequentially included in the multivariate Cox regression analysis. After adjusting covariates, we identifiedtotal CT GGO score≥4(HR 4.8, 95%CI 1.3-17.9, P=0.020), initial KL-6>1600U/ml (HR3.7, 95%CI 1.5-9.1,P=0.004), CRP>5.8mg/L(HR 3.7,95%CI 1.0-12.8,P=0.044)as independent risk factors for death, and initial combination(HR 0.3, 95%CI 0.1-0.8,P=0.019)as independentrisk factor for survival.
Survival curves of anti-MDA5 positive patients in different groups divided by independent risk factors
To further testify the prognostic value of the aforementioned risk factors, we examined survival curves of patients divided by these independent risk factors. As shown in Fig. 1, the survival curve ofpatients with total CT GGO score≥4was significantly worse than those with GGO<4 (P<0.001) (Fig.1A). Similarly, significant differences were observed in the survival curves between patients with KL-6>1600U/ml(P<0.001) (Fig.1B), CRP>5.8mg/L (P<0.001) (Fig.1C) than those without respectively. In addition, compared to the patients treated with step-up regimens, those treated with initial combination had a significant superior survival curve (P<0.001)(Fig.1D).
Table 1
Baseline data and treatment regimens of 90 anti-MDA5 positive patients
Variables(n=90)
|
Value
|
Demographic features
|
|
Female, n (%)
|
57 (63.3)
|
Age, yrs, mean±SD
|
51.9±12.1
|
Disease duration, months, median (IQR)
|
2.0 (1.0, 3.8)
|
CADM, n (%)
|
70 (77.8)
|
Antibody variables
|
|
Anti-Ro52, n (%)
|
64 (71.1)
|
Anti-MDA5, U/ml, median (IQR)
|
188.5 (158.5, 206.3)
|
ANA≥1:320, n (%)
|
15 (16.7)
|
Initial treatment regimens
|
|
High dose GC, n (%)
|
90(100)
|
CNI, n (%)
|
37(41.1)
|
CYC, n (%)
|
28(31.1)
|
Initial combination therapy, n (%)
|
58(64.4)
|
IVIG, n (%)
|
43(47.8)
|
Malignancy (within 6 months), n (%)
|
3(3.3)
|
Prognosis
|
|
Death, n (%)
|
22(24.4)
|
CADM clinically amyopathic dermatomyositis; anti-MDA5 anti-melanoma differentiation-associated protein-5; ANA anti-nuclearantibody; CNI calcineurin inhibitors; CYC cyclophosphamide; IVIG intravenous immunoglobulin; SD standard deviation; IQR interquartile rang |
Table 2
Comparison of clinical features and treatment regimens between the survival group and the death group
|
Survival group (n = 68)
|
Death group (n = 22)
|
P value
|
Demographic features
|
|
|
|
Female, n (%)
|
43 (63.2)
|
14 (63.6)
|
0.973
|
Age, yrs, mean ± SD
|
49.5 ± 11.4
|
59.2 ± 11.6
|
0.001*
|
Disease duration, months, median (IQR)
|
2.0 (1.0, 4.0)
|
1.0 (1.0, 2.8)
|
0.086
|
Clinical features
|
|
|
|
CADM, n (%)
|
51 (75.0)
|
19 (86.4)
|
0.380
|
Fever, n (%)
|
33 (48.5)
|
19 (86.4)
|
0.002*
|
Raynaud’s phenomenon, n (%)
|
7(10.3)
|
2(9.1)
|
1.000
|
Neck V sign, n (%)
|
18(26.5)
|
1(4.5)
|
0.035*
|
Gottron’s papules, n (%)
|
46(67.6)
|
15(68.2)
|
1.000
|
Heliotrope sign, n (%)
|
47 (69.1)
|
10 (45.5)
|
0.045*
|
Skin ulcer, n (%)
|
1 (1.5)
|
1 (4.5)
|
0.431
|
Laboratory test
|
|
|
|
Serum ferritin, ng/ml, median (IQR)
|
734.5 (426.2, 1255.5)
|
2198.5 (1008.8, 2923.5)
|
<0.001*
|
KL-6, U/ml, median (IQR)
|
857.0 (640.8, 1202.2)
|
1736.0 (830.0, 2958.5)
|
0.002*
|
LDH, U/L, median (IQR)
|
314.0 (268.2, 427.8)
|
491.5 (320.0, 715.5)
|
0.009*
|
CK, U/L, median (IQR)
|
64.5 (34.0, 114.5)
|
72.5 (41.8, 132.2)
|
0.508
|
CRP, mg/L, median (IQR)
|
4.4 (1.7, 16.5)
|
23.0 (11.2, 44.7)
|
<0.001*
|
ESR, mm/h, median (IQR)
|
30.5 (19.8, 45.0)
|
37.0 (27.2, 71.8)
|
0.043*
|
Lymphocyte, ×109/L, median (IQR)
|
0.8 (0.6, 1.2)
|
0.6 (0.4, 0.9)
|
0.014*
|
Anti-Ro52, n (%)
|
44 (64.7)
|
20 (90.9)
|
0.018*
|
Anti-MDA5, U/ml, median (IQR)
|
187.5 (155.5, 206.0)
|
192.3 (165.8, 205.0)
|
0.333
|
ANA ≥ 1:320, n (%)
|
11 (16.2)
|
4 (18.2)
|
1.000
|
HRCT findings
|
|
|
|
Total CT GGO score, median (IQR)
|
0.5 (0.0, 4.0)
|
11.0 (5.0, 14.8)
|
<0.001*
|
Total CT fibrosis score, median (IQR)
|
2.0 (0.0, 4.0)
|
6.0 (2.2, 12.0)
|
0.001*
|
Treatment
|
|
|
|
Initial combination therapy, n (%)
|
53 (77.9)
|
5 (22.7)
|
<0.001*
|
IVIG, n (%)
|
31 (45.6)
|
12 (54.5)
|
0.465
|
CADM: clinically amyopathic dermatomyositis; KL-6: Krebs von den Lungen-6; LDH: lactate dehydrogenase; CK: creatine kinase; CRP: C-reactive protein; ESR:erythrocyte sedimentation rate; anti-MDA5: anti-melanoma differentiation-associated protein-5; ANA: anti-nuclear antibody; HRCT: high-resolution CT; GGO:ground-glass opacity; CNI: calcineurin inhibitors; CYC: cyclophosphamide; IVIG: intravenous immunoglobulin; SD: standard deviation; IQR: interquartile range; * significant |
Table 3
Initial parametersassociated with death significantly using a Cox regression model
|
Univariate analysis
|
|
Multivariate analysis
|
|
HR
|
95% CI
|
Pvalue
|
|
HR
|
95% CI
|
P value
|
Age>55yrs
|
5.1
|
1.9-13.9
|
0.001*
|
|
-
|
-
|
-
|
Fever
|
5.5
|
1.6-18.8
|
0.006*
|
|
-
|
-
|
-
|
Heliotrope sign
|
0.4
|
0.2-1.0
|
0.051
|
|
-
|
-
|
-
|
Serum ferritin>2000ng/ml
|
6.7
|
2.8-16.1
|
<0.001*
|
|
-
|
-
|
-
|
KL-6>1600U/ml
|
6.8
|
2.9-15.8
|
<0.001*
|
|
3.7
|
1.5-9.1
|
0.004*
|
LDH>500U/L
|
4.7
|
2.0-10.9
|
<0.001*
|
|
-
|
-
|
-
|
CRP>5.8mg/L
|
7.1
|
2.1-24.1
|
0.002*
|
|
3.7
|
1.0-12.8
|
0.044*
|
ESR>23mm/h
|
9.6
|
1.3-71.0
|
0.028*
|
|
-
|
-
|
-
|
Lymphocyte<0.6×109/L
|
0.4
|
0.2-1.0
|
0.057
|
|
-
|
-
|
-
|
Anti-Ro52
|
4.7
|
1.1-19.9
|
0.038*
|
|
-
|
-
|
-
|
Total CT GGO score≥4
|
12.2
|
3.6-41.2
|
<0.001*
|
|
4.8
|
1.3-17.9
|
0.020*
|
Total CT fibrosis score≥10
|
7.0
|
2.9-16.4
|
<0.001*
|
|
-
|
-
|
-
|
Initial combined treatment
|
0.1
|
0.04-0.3
|
<0.001*
|
|
0.3
|
0.1-0.8
|
0.019*
|
KL-6: Krebs von den Lungen-6; LDH: lactate dehydrogenase; CRP: C-reactive protein; ESR:erythrocyte sedimentation rate; GGO:ground-glass opacity; * significant |
Figure 1.Survival curves of anti-MDA5 positive patients in each group based on initial KL-6, CRP level, total GGOscore and initial combined treatment.
KL-6: Krebs von den Lungen-6; CRP: C-reactive protein; GGO:ground-glass opacity