Risk factors of significant relapse and appropriate maintenance therapy strategy in SLE-associated immune thrombocytopenia

Abstract

Background:

Systemic lupus erythematosus–associated immune thrombocytopenia (SLE-ITP) is characterized by relapse. The risk factors of relapse and appropriate maintenance therapy strategy deserve further exploration.

Objectives:

To determine the risk factors for relapse and appropriate maintenance therapy in significant SLE-ITP patients (a platelet count ⩽30 × 10⁹/l) after the first complete response.

Design:

Retrospective cohort study using the medical records of 105 patients diagnosed as significant SLE-ITP in Fujian Medical University Union Hospital during December 2012 to March 2021. Patients were followed through a call for observations in January 2022.

Methods:

Data including demographics, initial clinical feature, induction and maintenance therapy, and outcome at the end of follow-up were analyzed. Risk factors for significant relapse were analyzed using multivariate logistic regression models. The cumulative hazard of significant relapse and the duration of response were estimated, and the differences in outcome between groups were compared using the Cox regression analysis.

Results:

A total of 65 significant SLE-ITP patients were eligible for the final analysis. Median [interquartile range (IQR)] follow-up duration and median [IQR] duration of response were 62.2 [41.0–79.6] months and 43.4 [20.3–68.7] months, respectively. After the first complete response, 19/65 (29.2%) had a significant relapse. Compared with sustained clinical remission (SCR) + sustained response (SR) group, significant relapse group had a higher proportion of discontinued patients (47.4% versus 8.7%, p = 0.001). Among the 13 discontinued patients, the duration of maintenance therapy of the patients in significant relapse group was significantly shorter than that of the patients in SCR + SR group (months, median [IQR], 43.1 [32.0–62.4] versus 12.0 [5.1–22.0], p = 0.009). Multivariate logistic regression analysis showed that drug withdrawal was an independent risk factor for significant relapse [odds ratio (OR) = 10.4, confidence interval (CI) 95% 2.2–47.8, p = 0.003]. There was no significant difference between glucocorticoids (GCs) + hydroxychloroquine (HCQ) group and GCs + HCQ + immunosuppressive agents (ISAs) group in significant relapse rate (26.7% versus 22.2%, p > 0.05). The two SR curves of GCs + HCQ and GCs + HCQ+ ISA group basically coincided by the Cox regression analysis, demonstrating comparable long-term outcomes (p > 0.05).

Conclusion:

Drug withdrawal, especially abrupt withdrawal with insufficient duration of maintenance therapy, is an independent risk factor for significant relapse of SLE-ITP. HCQ combined with GCs is expected to be the first choice of the maintenance therapy for SLE-ITP patients.

Keywords

hydroxychloroquine immunosuppressive agents maintenance therapy SLE-associated immune thrombocytopenia

Introduction

Systemic lupus erythematosus–associated immune thrombocytopenia (SLE-ITP) is one of the main manifestations of systemic lupus erythematosus (SLE), which can occur alone or with other organ damage. Approximately 70% of SLE patients follow a relapsing-remitting course.¹ Similarly, SLE-ITP often relapses during glucocorticoids (GCs) tapering.² The accompanying thrombocytopenia at the onset of SLE might be a risk factor for SLE flare.³ Then, what are the risk factors for SLE-ITP relapse? How to choose the maintenance therapy for SLE-ITP at different stages (first flare or relapse)? All these issues deserve further exploration. According to the 2019 European Alliance of Associations for Rheumatology (EULAR) recommendations for the management of SLE, high-dose GCs (including pulses of intravenous methylprednisolone) is an option for induction therapy of significant lupus thrombocytopenia (platelet count below 30,000/mm³) in the acute phase. Intravenous immunoglobulin (IVIG) is also an option for induction therapy, in cases of inadequate response to high-dose GCs or to avoid GC-related infectious complications.⁴ At the same time, the recommendations also state that treatment of thrombocytopenia is typically lengthy and often characterized by relapses during GCs tapering.⁴ Relapses increase the risk of bleeding, which requires an increase in drugs, especially high-dose GCs. Therefore, it is not only the disease activity itself but also the repeated use of high-dose GCs that can aggravate organ damage when the disease relapses. This highlights the importance of maintenance therapy for significant SLE-ITP both after the first treatment and after relapse. Immunosuppressive agents (ISAs) such as mycophenolate (MMF), azathioprine (AZA), or cyclosporine A (CsA) are recommended as the first-line maintenance therapy. In addition, rituximab (RTX), cyclophosphamide (CTX), thrombopoietin agonist, or splenectomy is considered one-by-one options for patients with no response to GCs or relapses.⁴ There is, however, often a lack of strong evidence to support robust recommendations for various management approaches.⁵

The aims of this study were to determine (1) the risk factors for relapse of significant SLE-ITP (a platelet count ⩽30 × 10⁹/l) and (2) preferred maintenance therapy after the first complete response.

Methods

Study cohort and patients

This was a retrospective case–control study on the Chinese Han population. We retrospectively reviewed the medical records of 617 hospitalized patients with newly diagnosed SLE from Fujian Medical University Union Hospital during December 2012 to March 2021. Patients with significant SLE-ITP were included and followed through a call for observations in January 2022. All selected patients fulfilled the Systemic Lupus International Collaborating Clinics (SLICC) 2012 SLE classification criteria.⁶ The exclusion criteria were defined as follows: (1) initially treated before December 2012 or at another hospital; (2) with other important organ damage requiring combined treatment of GCs and ISA such as lupus nephritis (LN), neuropsychiatric lupus (NPSLE), myocarditis, lupus pneumonitis, myositis, and severe vasculitis; (3) with complications affecting diagnosis and treatment; (4) concomitant antiphospholipid syndrome; (5) missing important data; (6) other disease other than SLE known to be associated with thrombocytopenia, such as genetic diseases, bone marrow diseases, malnutrition, infections, tumors, drugs, thrombotic microvascular diseases, poisoning, hypersplenism, etc.; (7) patients without complete response after induction therapy (platelet count < 100 × 10⁹/l); and (8) lost to follow-up.

Patients were grouped as follows: (1) based on the first significant relapse after initial complete response, patients were divided into two groups: sustained clinical remission (SCR) and sustained response (SR) group (SCR + SR group), and significant relapse group and (2) based on maintenance therapy after initial complete response, patients were divided into four groups: GCs, GCs + hydroxychloroquine (HCQ), GCs + HCQ + ISA, and GCs + ISA group.

Definition of clinical data

We obtained the demographic and clinical data by means of a review of electronic medical records. We analyzed the following parameters: age at disease diagnosis, sex, autoantibodies, clinical manifestations, laboratory examinations, treatment data, and outcomes.

At the time of initial induction, complete response was defined as a platelet count >100 × 10⁹/l and the absence of bleeding and new other symptoms. Clinical remission (CR) was defined according to the 2021 DORIS definition of remission in SLE.⁷ SCR was defined as the absence of any deterioration in CR throughout the follow-up period. Response was defined as a platelet count >30 × 10⁹/l and at least twofold increase of the baseline count and absence of bleeding and new damage to important organs (renal, central nervous system, cardiopulmonary, vasculitis, fever, active hemolytic anemia, or gastrointestinal activity); SR as the absence of any deterioration in response throughout the follow-up period and the maintenance dose of prednisone (or equivalent) ⩽7.5 mg daily (an increase in prednisone to ⩽0.5 mg/kg/day was allowed if manifestations required, but it must be reduced to 7.5 mg/day or less within 3 months); significant relapse as a platelet count ⩽30 × 10⁹/l or less than twofold increase of the baseline platelet count or bleeding or new damage to important organs or manifestations requiring an increase in prednisone (or equivalent) to >0.5 mg/kg/day. The first complete response was confirmed by platelet counts on two separate occasions at least 7 days apart, and significant relapse was confirmed by platelet counts on two separate occasions within 2 weeks or relevant clinical manifestations. High-dose GCs treatment was defined as a dose of prednisone (or equivalent) ⩾1 mg/kg/day; super-high-dose GCs treatment as intravenous dexamethasone 40 mg daily for 4 days or intravenous methylprednisolone 250 mg or more daily for continuous 3–5 days; time to complete response (TTCR) as the duration from initiation of high-dose GCs treatment to the achievement of complete response; disease duration as the duration from symptom onset to the diagnosis of SLE; duration of GCs maintenance therapy as the duration from the achievement of complete response to discontinuation of GCs or to the first significant relapse (significant relapse group) or to the end of follow-up (SCR + SR group); duration of maintenance therapy as the duration from the achievement of complete response to drug withdrawal or to the first significant relapse (significant relapse group) or to the end of follow-up (SCR + SR group). The follow-up duration was measured from the initial clinic visit to the end of follow-up. Duration of response was measured from the achievement of complete response to the first significant relapse (significant relapse group) or to the end of follow-up (SCR + SR group).

Patients in the SCR + SR group were in CR or response until the end of follow-up, and there was no significant relapse. Patients in the significant relapse group had one or more significant relapses during the follow-up.

Before drug withdrawal, the first significant relapse, or the end of follow-up, the maintenance medication status of each group was as follows: patients in GCs group were treated with GCs alone; patients in GCs + HCQ group had been treated with GCs and HCQ with no or little ISA; patients in GCs + HCQ + ISA group had been treated with GCs, HCQ, and ISA (one or more of AZA, MMF, CsA/tacrolimus, CTX, or RTX) simultaneously for a period of time; patients in GCs + ISA group had been treated with GCs and ISA (HCQ may have been used during treatment, but not in combination with ISA).

Statistical analysis

According to the number of samples we included finally and the significant relapse rate, the power permitting to detect a difference in the significant relapse rate between maintenance therapy and drug withdrawal was 0.91 with a one-sided type I error rate of 2.5%. Categorical variables were described as numbers (percentages) and compared using Chi-square test or Fisher’s exact test. Continuous variables were presented as medians [interquartile range (IQR) or range] and compared using the Mann–Whitney nonparametric U test. Multivariate logistic regression models, with odds ratios (ORs) and 95% confidence intervals (CIs), were used to determine the variables associated with significant relapse. The cumulative hazard of significant relapse stratified by drug withdrawal was estimated using the Cox regression analysis. The duration of response was estimated and difference in outcome between the four maintenance therapy groups was compared using the Cox regression analysis. The validity of the proportional hazards assumption was assessed by log-minus-log-survival function and confirmed by time-dependent covariate analysis. All reported p-values are two-sided, with statistical significance level being set to p < 0.05. All statistical analyses were performed using SPSS 25.

Results

Study population and characteristics

Of 617 hospitalized patients with newly diagnosed SLE, a total of 238 had thrombocytopenia, among which there were 105 patients with significant thrombocytopenia. According to the second exclusion criteria, patients combined with LN, NPSLE, myocarditis, lupus pneumonitis, myositis, and severe vasculitis were excluded because the involvement of these organs itself required the combined treatment of GCs and ISA. Finally, after 40 patients who fulfilled the exclusion criteria were removed, the 65 remaining patients with significant SLE-ITP were included in the analysis (Figure 1).

Figure 1.

Selection and inclusion of patients in the study.

The baseline demographic, clinical, and immunologic characteristics were shown in Table 1. Overall, there were 59 (90.8%) female patients. The median [IQR] ages at SLE diagnosis were 32 [19–45] years. The median [IQR] disease duration was 1 [0.4–9.1] months. The median [IQR] follow-up duration was 62.2 [41.0–79.6] months. In addition to thrombocytopenia, some patients had fever (7/65), rash (9/65), nonscarring alopecia (5/65), arthritis (8/65), serositis (10/65), hemolytic anemia (9/65), and leukopenia (23/65). Table 1 provided additional details.

Table 1.

Comparison of the main clinical features between groups and multivariate analysis of the influencing factors of significant relapse.

	All (n = 65)	SCR + SR group (n = 46)	Significant relapse group (n = 19)	p-value	Multivariate logistic analysis
	All (n = 65)	SCR + SR group (n = 46)	Significant relapse group (n = 19)	p-value	OR	95% CI	p-value
Demographic
Sex, F (%)	59 (90.8)	44 (95.7)	15 (78.9)	0.100	–	–	–
Age at SLE diagnosis, year, median [IQR]	32 [19–45]	34.5 [18–45.5]	27 [22–38]	0.554	–	–	–
Disease duration, months, median [IQR]	1 [0.4–9.1]	1 [0.4–7.65]	1 [0.3–12]	0.812	–	–	–
Follow-up duration, months, median [IQR]	62.2 [41.0–79.6]	55.4 [40.0–78.3]	74.9 [49.5–92.5]	0.160	–	–	–
Clinical
Fever, n (%)	7 (10.8)	7 (15.2)	0	0.174	–	–	–
Cutaneous, n (%)	9 (13.8)	7 (15.2)	2 (10.5)	0.918	–	–	–
Oral/nasal ulcers, n (%)	0	0	0		–	–	–
Nonscarring alopecia, n (%)	5 (7.7)	3 (6.5)	2 (10.5)	0.969	–	–	–
Arthritis, n (%)	8 (12.3)	7 (15.2)	1 (5.3)	0.486	–	–	–
Serositis, n (%)	10 (15.4)	7 (15.2)	3 (15.8)	1.000	–	–	–
Hemolytic anemia, n (%)	9 (13.8)	7 (15.2)	2 (10.5)	0.918	–	–	–
WBC (×10⁹/l), median [IQR]	4.9 [3.5–6.8]	5.09 [3.41–6.77]	4.55 [3.57–7.93]	0.931	–	–	–
Leukopenia, n (%)	23 (35.4)	16 (34.8)	7 (36.8)	0.875	–	–	–
HGB (×10⁹/l), median [IQR]	100 [76–125]	101 [68.5–125]	100 [85–127]	0.681	–	–	–
PLT (×10⁹/l), median [IQR]	11 [4–22]	12.5 [4–21.25]	9 [4–30]	0.660	–	–	–
Immunologic
Anti-SSA/Ro, n (%)	39/64 (60.9)	26/45 (57.8)	13/19 (68.4)	0.761	–	–	–
Anti-SSB/La, n (%)	16/64 (25.0)	11/45 (24.4)	5/19 (26.3)	0.667	–	–	–
Positive anti-dsDNA, n (%)	46/64 (71.9)	31/45 (68.9)	15/19 (78.9)	0.414	–	–	–
Direct Coombs test, n (%)	13/55 (23.6)	7/38 (18.4)	6/17 (35.3)	0.309	–	–	–
Medium–high titer aCL, n (%)	15/63 (23.8)	13/44 (29.5)	2/19 (10.5)	0.192	–	–	–
Increased anti-ß2GPI, n (%)	18/58 (31.0)	12/40 (30.0)	6/18 (33.3)	0.800	–	–	–
IgG (g/l), median [IQR]	20.6 [14.7–27.8]	21.3 [15.2–29.0]	19.2 [14.1–26.4]	0.582	–	–	–
Low C3 or C4, n (%)	60 (92.3)	43 (93.5)	17 (89.5)	0.969	–	–	–
Platelet antibodies, n (%)	21/40 (52.5)	14/27 (51.9)	7/13 (53.8)	0.906	–	–	–
SLEDAI-2K, median [IQR]	5 [5–6.5]	5 [5–7]	5 [3–5]	0.213	–	–	–

aCL, anticardiolipin antibody; anti-ß2GPI, anti-ß2glycoproteinI; anti-dsDNA, anti double-stranded DNA; CI, confidence interval; disease duration, the duration from symptom onset to the diagnosis of SLE; follow-up duration, the duration from the initial clinic visit to the end of follow-up; HGB, hemoglobin; IgG, immunoglobulin G; IQR, interquartile range; NA, not applicable; OR, odds ratio; PLT, platelet; SCR, sustained clinical remission; SLE, systemic lupus erythematosus; SLEDAI-2K, Systemic Lupus Erythematosus Disease Activity Index 2000; SR, sustained response; WBC, white blood cell count.

Treatment and outcomes

In induction therapy, all patients received high-dose GCs treatment, of which 20.0% (13/65) had received super-high-dose GCs treatment, and 52.3% (34/65) also received IVIG pulse therapy (0.4 g/kg/day for continuous 3–5 days). For all patients, the median [IQR] TTCR was 8 [5.5–11.5] days. The median [IQR] duration of response was 43.4 [20.3–68.7] months. After the first complete response, 70.8% (46/65) of patients sustained CR or response, and 29.2% (19/65) had a significant relapse (Table 2).

Table 2.

Comparison of induction and maintenance therapy between groups and multivariate analysis of the influencing factors of significant relapse.

	All (n = 65)	SCR + SR group (n = 46)	Significant relapse group (n = 19)	p-value	Multivariate logistic analysis
	All (n = 65)	SCR + SR group (n = 46)	Significant relapse group (n = 19)	p-value	OR	95% CI	p-value
Induction therapy
Super-high-dose GCs, n (%)	13 (20.0)	10 (21.7)	3 (15.8)	0.838	–	–	–
Cumulative dose of GCs during induction, g, median [IQR]	1.0 [0.6–1.70]	1.04 [0.6–1.73]	0.97 [0.4–1.45]	0.554	–	–	–
IVIG pulse therapy, n (%)	34 (52.3)	23 (50.0)	11 (57.9)	0.562	–	–	–
RHTPO/IL-11, n (%)	9 (13.8)	6 (13.0)	3 (15.8)	1.000	–	–	–
VCR, n (%)	9 (13.8)	5 (10.9)	4 (21.1)	0.493	–	–	–
Danazol, n (%)	7 (10.8)	5 (10.9)	2 (10.5)	1.000	–	–	–
TTCR, days, median [IQR]	8 [5.5–11.5]	9 (6–13)	7 [5–10]	0.322	–	–	–
Maintenance therapy				0.048			0.614
GCs, n (%)	3 (4.6)	0	3 (15.8)	–	–	–	–
GCs + HCQ, n (%)	30 (46.2)	22 (47.8)	8 (42.1)	–	–	–	–
GCs + ISA + HCQ, n (%)	27 (41.5)	21 (45.7)	6 (31.6)	–	–	–	–
GCs + ISA, n (%)	5 (7.7)	3 (6.5)	2 (10.5)	–	–	–	–
Drug withdrawal, n (%)	13 (20)	4 (8.7)	9 (47.4)	0.001	10.4	2.2–47.8	0.003
Duration of maintenance therapy, months, median [IQR]	36.0 [12.3–62.3]	49.5 [21.9–67.0]	7.6 [3.9–26]	0.000	NA	NA	NA
Duration of response, months, median [IQR]	43.4 [20.3–68.7]	55.1 [38.8–76.8]	18 [7–32]	0.001	NA	NA	NA

CI, confidence interval; duration of maintenance therapy, from the achievement of complete response to drug withdrawal or to the first significant relapse (significant relapse group) or to the end of follow-up (SCR + SR group); duration of response, from the achievement of complete response to significant relapse (significant relapse group) or to the end of follow-up (SCR + SR group); GCs, glucocorticoids; HCQ, hydroxychloroquine; IL-11, interleukin-11; IQR, interquartile range; ISA, immunosuppressive agent; IVIG, intravenous immunoglobulin; NA, not applicable; OR, odds ratio; RHTPO, recombinant human thrombopoietin; SCR, sustained clinical remission; SR, sustained response; super-high-dose GCs, dexamethasone was administered intravenously at 40 mg daily for four consecutive days or methylprednisolone was administered intravenously at 250 mg or more daily for three to five consecutive days; TTCR, time to complete response; VCR, vincristine.

After a period of maintenance, a total of 13 patients stopped taking drugs and 9 significantly relapsed. Among them, some patients (10 out of 13) stopped taking drugs by themselves because they were worried about side effects or felt good about themselves, and had a high significant relapse rate (9 out of 10). The remaining three patients gradually stopped taking drugs under the recommendation of doctors after maintaining remission for a long time, and there was no significant relapse in these patients. Among the 13 discontinued patients, the duration of maintenance therapy of patients in significant relapse group was significantly shorter than that of patients in SCR + SR group (months, median [IQR] = 43.1 [32.0–62.4] versus 12.0 [5.1–22.0], p = 0.009) (Table 3).

Table 3.

Comparison of characteristics between discontinued patients in SCR + SR group and those in significant relapse group.

	Discontinued patients (n = 13)	SCR + SR group (n = 4)	Significant relapse group (n = 9)	p-value
Reasons for drug withdrawal				0.014
Patient’s self-discontinuation of drugs	10	1	9
Discontinuation of drugs recommended by doctors	3	3	0
Dose of GCs at withdrawal, mg, median [IQR]	5.0 [3.8–8.8]	3.5 [0.5–5]	5.0 [5.0–13.8]	0.065
Dose of HCQ at withdrawal, g, median [IQR]	0.2 [0.2–0.2]	0.2 [0.2–0.2]	0.2 [0.1–0.3]	1.000
Duration of maintenance therapy, months, median [IQR]	20.0 [6.6–35.9]	43.1 [32.0–62.4]	12.0 [5.1–22.0]	0.009

Duration of maintenance therapy, from the achievement of complete response to drug withdrawal; GCs, glucocorticoids; HCQ, hydroxychloroquine; IQR, interquartile range; SCR, sustained clinical remission; SR, sustained response.

The details of maintenance therapy and outcomes are shown in Table 4.

Table 4.

Comparison of treatment and outcome in different maintenance therapy groups.

	GCs (n = 3)	GCs + HCQ (n = 30)	GCs + HCQ + ISA (n = 27)	GCs + ISA (n = 5)
SLEDAI-2K before treatment, median [IQR]	5 [5–5]	5 [5–5.3]^b	5 [4–7]	3 [3–7.5]
TTCR, days, median [IQR]	5 [3–9]	8 [5–11]^b	9 [6–14]	10 [5–69]
Induction therapy
Super-high-dose GCs, n (%)	0	6 (20.0)^b	6 (22.2)	1 [20.0]
Cumulative dose of GCs during induction, g, median [IQR]	NA	1.0 [0.6–1.7]^b	1.2 [0.5–2.2]	0.9 [0.4–5.2]
IVIG pulse therapy, n (%)	1 (33.3)	19 (63.3)^b	12 (44.4)	2 [40]
Maintenance therapy
Dose of GCs for maintenance therapy, mg/day, median [IQR]	NA	2.5 [0–3.1]^b	2.5 [0–5.0]	5.0 [2.5–10]
Duration of GCs maintenance therapy, months, median [IQR]	NA	28.9 [15.5–52.9]^b	35.8 [9.0–66.5]	54.2 [12.2–70.6]
Dose of HCQ for maintenance therapy, g/day, median [IQR]	0	0.2 [0–0.2]^b	0.1 [0–0.2]	0
Duration of maintenance therapy, months, median [IQR]	NA	34.0 [18.4–54.7]^b	40.0 [9.0–68.4]	54.2 [12.2–70.6]
ISA for maintenance therapy
Cyclophosphamide, n (%)	NA	NA	11 (40.7)	0
Rituximab, n (%)	NA	NA	3 (11.1)	0
Mycophenolate mofetil, n (%)	NA	NA	10 (37.0)	0
Cyclosporine or tacrolimus, n (%)	NA	NA	18 (66.7)	2 (40.0)
Azathioprine, n (%)	NA	NA	3 (11.1)	3 (60.0)
Duration of simultaneous use of GCs, HCQ, and ISA, months, median [IQR]	NA	NA	34.9 [12.6–63.7]	NA
Long-term outcomes during follow-up
SCR + SR, n (%)	0	22 (73.3)^b	21 (77.8)	3 (60.0)
SCR, n (%)	0	16 (53.3)^b	15 (55.6)	2 (40.0)
SR, n (%)	0	6 (20.0)^b	6 (22.2)	1 (20.0)
Significant relapse, n (%)	3 (100%)^a	8 (26.7)^b	6 (22.2)	2 (40.0)
Significant relapse after drug withdrawal, n (%)	2 (66.7)	6 (20)^b	1 (3.7)	0
Duration of response, months, median [IQR]	18 [7–43.7]	37.5 [20.3–57.5]^b	43.4 [18.2–77.3]	54.2 [12.2–70.6]
Outcomes at the end of follow-up
CR, n (%)	2 (66.7)	22 (73.3)	20 (74.1)	3 (60.0)
CR off drug, n (%)	0	2 (6.7)	1 (3.7)	1 (20.0)
CR off GCs, n (%)	0	3 (10)	3 (11.1)	0
CR on GCs + HCQ, n (%)	2 (66.7)	16 (53.3)	5 (18.5)	0
CR on GCs + HCQ + ISA, n (%)	0	1 (3.3)	10 (37.0)	0
CR on GCs + ISA, n (%)	0	0	1 (3.7)	2 (40.0)
Response, n (%)	1 (33.3)	8 (26.7)	6 (22.2)	1 (20.0)
Renal failure, n (%)	0	0	1 (3.7)	0
significant relapse	0	0	0	1 (20.0)

CR, clinical remission; duration of GCs maintenance therapy, from the achievement of complete response to discontinuation of GCs or to the first significant relapse (significant relapse group) or to the end of follow-up (SCR + SR group); duration of maintenance therapy, from the achievement of complete response to drug withdrawal or to the first significant relapse (significant relapse group) or to the end of follow-up (SCR + SR group); duration of response, from the achievement of complete response to significant relapse (significant relapse group) or to the end of follow-up (SCR + SR group); GCs, glucocorticoids; HCQ, hydroxychloroquine; IQR, interquartile range; ISA, immunosuppressive agent; IVIG, intravenous immunoglobulin; NA, not applicable; SCR, sustained clinical remission; SLEDAI-2K, Systemic Lupus Erythematosus Disease Activity Index 2000; SR, sustained response; super-high-dose GCs, dexamethasone was administered intravenously at 40 mg daily for four consecutive days or methylprednisolone was administered intravenously at 250 mg or more daily for three to five consecutive days; TTCR, time to complete response.

Compared with GCs + HCQ and GCs + HCQ + ISA group (p < 0.05).

Compared with GCs + HCQ + ISA group (p > 0.05).

After the first complete response, maintenance therapy with GCs alone was used in three patients. Two patients relapsed significantly after GCs withdrawal, one presenting with significant thrombocytopenia (a platelet count <30 × 10⁹/l) and the other with a platelet count <30 × 10⁹/l, skin lesions and myocardial involvement, and are currently in clinical remission under GCs and HCQ maintenance therapy. The third patient relapsed significantly during GCs maintenance therapy, presenting with skin lesions, nephritis, fever, severe anemia, and mild thrombocytopenia (<100 × 10⁹/l), and is currently in response under GCs, HCQ, MMF, and AZA maintenance therapy.

In GCs + HCQ group, one patient was treated with CsA at the beginning of treatment. After 9 days, it was changed to MMF due to liver injury. After another 10 days, MMF was stopped due to lung infection. Later, the patient remained in CR under long-term GCs and HCQ maintenance therapy, so the patient was classified into this group. Of the 30 patients in this group, the median [IQR] dose for maintenance therapy was 2.5 [0–3.1] mg/day for prednisone (or equivalent) and 0.2 [0–0.2] g/day for HCQ. In this group, 16 patients (53.3%) sustained CR, 6 (20.0%) SR (4 patients presenting with a platelet count <100 × 10⁹/l, 1 patient with arthritis, and 1 with leukopenia), and 8 (26.7%) relapsed significantly (7 patients presenting with a platelet count <30 × 10⁹/l, and 1 patient with mesenteric vasculitis repeatedly), of which 6 relapsed significantly after drug withdrawal. After repeated induction therapy and maintenance therapy, a total of 22 (73.3%) patients are now in CR, including 2 off drug and 3 off GCs, and 8 (26.7%) are in response.

Of the 27 patients in GCs + HCQ + ISA group, the median [IQR] dose for maintenance therapy was 2.5 [0–5.0] mg/day for prednisone (or equivalent) and 0.1 [0–0.2] g/day for HCQ. The median [IQR] duration of simultaneous use of GCs, HCQ and ISA was 34.9 [12.6–63.7] months. CsA or Tacrolimus (18/27), CTX (11/27), and MMF (10/27) were the most commonly used ISA. The dose of ISA was as follows: MMF 0.25–2.0 g daily, CsA 25–200 mg daily, tacrolimus 1–2 mg daily, and AZA 50–100 mg daily. Eleven patients received CTX with a median [IQR] cumulative dose of 7.6 [5–8.4] g. Three patients received RTX (0.1 g each time) four, five, and eight times, respectively (not shown in tables). ISA were added at the beginning of treatment or after complete response in 24 patients, and 14 days or more after treatment began in the other 3 patients because they did not achieve complete response to initial therapy. In this group, 15 patients (55.6%) sustained CR, 6 (22.2%) SR (4 patients presenting with a platelet count <100 × 10⁹/l and 2 patient with skin erythema), and 6 (22.2%) relapsed significantly (5 patients presenting with a platelet count <30 × 10⁹/l and 1 patient with LN and irreversible renal failure), of which 1 relapsed significantly after drug withdrawal. After repeated induction therapy and maintenance therapy, a total of 20 (74.1%) patients are now in CR, including 1 off drug and 3 off GCs, and 6 (22.2%) are in response. One patient developed LN and irreversible renal failure due to irregular medication. Currently, she is undergoing maintenance hemodialysis.

Of the five patients in GCs + ISA group, three were treated with GCs and AZA because of contraindications to HCQ, including two patients with SCR and one patient with SR (leukopenia). The fourth patient was treated with GCs and CsA, and had significant relapse (a platelet count <30 × 10⁹/l) after 1.8 months of maintenance therapy. After the second induction therapy and maintenance therapy, he is now in CR off drug. The last patient received GCs and CsA after 53 months of treatment with GCs and HCQ. After 21 months of maintenance therapy with GCs and CsA, she is now experiencing significant relapse (a platelet count <30 × 10⁹/l).

During the follow-up, in the GCs + HCQ group, there were one case of cervical squamous cell carcinoma (stage Ia), one case of acute appendicitis, one case of hypertension, and one case of abnormal liver function and pulmonary infection (while using CsA and MMF). In GCs + HCQ + ISA group, there were one case of herpes zoster, one case of severe respiratory infection, one case of right vestibular gland abscess, one case of necrosis of femoral head, and one case of blurred vision. The remaining patients had no documented adverse events.

Independent risk factors for significant relapse

Compared with SCR + SR group, significant relapse group had a higher proportion of drug withdrawal patients (47.4% versus 8.7%, p = 0.001) (Table 2). There were also statistical differences between the four groups in terms of maintenance therapy (p = 0.048, Table 2). When the two factors were further entered into multivariable logistic regression model, only drug withdrawal (OR = 10.4, CI 95% = 2.2–47.8, p = 0.003) was identified as an independent risk factor for significant relapse (Table 2). After the proportional hazards assumption was confirmed to hold, we further estimated the cumulative hazard of significant relapse stratified by drug withdrawal. The cumulative hazard of significant relapse was significantly higher among drug withdrawal patients by the Cox regression analysis [hazard ratio (HR) = 3.7, CI 95% = 1.3–10.6, p = 0.016] (Figure 2).

Figure 2.

The Cox regression analysis estimates of the cumulative hazard of significant relapse stratified by drug withdrawal.

Effects of different maintenance therapies on significant relapse

All three patients in GCs group had significant relapse, which was statistically significant in significant relapse rate compared with GCs + HCQ group and GCs + HCQ + ISA group (100% versus 26.7% and 22.2%, respectively, p < 0.05), while there was no statistically significant difference between GCs + HCQ group and GCs + HCQ + ISA group (26.7% versus 22.2%, p > 0.05) (Table 4). Further multivariate logistic regression analysis showed that there were no statistically significant differences of the four maintenance therapies in significant relapse rate (Table 2). There was no statistically significant difference between GCs + HCQ group and GCs + HCQ + ISA group in Systemic Lupus Erythematosus Disease Activity Index 2000 (SLEDAI-2K) before treatment (median [IQR] = 5 [5–5.3] versus 5 [4–7]), TTCR (days, median [IQR] = 8 [5–11] versus 9 [6–14]), cumulative dose of GCs during induction (g, median [IQR] = 1.0 [0.6–1.7] versus 1.2 [0.5–2.2]), proportion of patients receiving IVIG pulse therapy (63.3% versus 44.4%), dose of GCs for maintenance therapy (mg/day, median [IQR] = 2.5 [0–3.1] versus 2.5 [0–5.0]), duration of GCs maintenance therapy (months, median [IQR] = 28.9 [15.5–52.9] versus 35.8 [9.0–66.5]), dose of HCQ for maintenance therapy (g/day, median [IQR] = 0.2 [0–0.2] versus 0.1 [0–0.2]), duration of maintenance therapy (months, median [IQR] = 34.0 [18.4–54.7] versus 40.0 [9.0–68.4]), duration of response (months, median [IQR] = 37.5 [20.3–57.5] versus 43.4 [18.2–77.3]), SCR rate (53.3% versus 55.6%), and SR rate (20.0% versus 22.2%) (Table 4; p > 0.05). The two SR curves of GCs + HCQ and GCs + HCQ + ISA group basically coincided by the Cox regression analysis, demonstrating comparable long-term outcomes (Figure 3; p > 0.05).

Figure 3.

The Cox regression analysis estimates of the duration of response.

Discussion

The incidence of SLE-associated thrombocytopenia is different in reports due to disparity in race and region.^8,9 In general, there is a relatively high incidence of SLE-associated thrombocytopenia, mainly immune-mediated (SLE-ITP), which is an independent predictor of adverse SLE outcome.¹⁰

Although the list of available drugs in the treatment of SLE-ITP is growing, there remain several unsatisfactory aspects.¹¹ Like the overall SLE population, SLE-ITP is also characterized by relapse.² Both the disease itself and the repeated use of high doses of drugs, especially GCs, can cause cumulative damage to the organs when the disease relapses. Therefore, it is particularly important to explore the factors of relapse and appropriate maintenance therapy strategies to prevent relapse for SLE-ITP. Although several studies have compared the efficacy of medications for primary and secondary ITP, there is no consensus as to the most effective maintenance medication.¹² Clinical treatment of SLE-ITP is more based on the experience and habits of clinicians, which inevitably leads to safety and economic problems caused by overtreatment, or disease activities caused by undertreatment.

In this study, the risk of significant relapse in patients who stopped medication was 3.7 times higher than that in those who maintained therapy by the Cox regression analysis (Figure 2). In fact, long-term remission with no use of GCs and immunosuppressant is uncommon in SLE patients.¹³ Indeed, GC use is independently associated with the accrual of damage in SLE, including in non-GC-related domains.^14,15 GCs discontinuation or merely a reduction of dose, however, is also often associated with a flare. A randomized controlled trial compared the efficacy to prevent flares of maintenance versus withdrawal of 5 mg/day prednisone in clinically quiescent SLE with a mean duration of corticosteroids greater than 130 months.¹⁶ The result showed that proportion of patients experiencing a flare at 52 weeks was significantly higher in the withdrawal group as compared with the maintenance group [17/63 patients versus 4/61; relative risk (RR) = 0.2 (95% CI = 0.1–0.7), p = 0.003].¹⁶ Therefore, the benefits of GCs withdrawal must be balanced against the potential risks. At present, there is a consensus that the risk of harm is low for the majority of patients at long-term doses of ⩽5 mg prednisone equivalent per day.¹⁷ Despite this, the major target after disease control is also reach the lowest GC dose or its withdrawn all the time. A recent study showed that from the prospective of different tapering strategies embodied as duration from prednisone 5 mg/day to complete discontinuation, slower tapering strategy (12–24 months) significantly reduced the risk of flare compared with faster tapering strategy (<3 months).¹⁸ In contrast, in the study by Mathian et al.,¹⁶ withdrawal of 5 mg of prednisone was relatively abrupt, thus showing a higher flare rate in the withdrawal group. Moreover, patients included in several studies of GCs withdrawal^16,18 had maintained long-term remission before study enrollment. The study by Fasano et al.¹⁹ showed that ⩾5-year remission at GCs withdrawal was protective factors for disease flare. This study also showed that the duration of maintenance therapy before drug withdrawal was significantly shorter in patients with significant relapse after drug withdrawal than in patients without significant relapse, suggesting that a long period of remission under maintenance therapy before drug withdrawal is a necessary condition to reduce relapse after drug withdrawal. Therefore, similar to the whole SLE population, the strategy of GCs use in SLE-ITP maintenance therapy should be slow GCs tapering to withdrawal after a long-term maintenance remission.

Relapse-remission is the nature of SLE.²⁰ Considering the potential risks of long-term use of GCs, GC-sparing agents are indispensable in the treatment of serious SLE. ISA such as MMF, AZA or CsA is recommended as the first choice for maintenance treatment.^1,4 This study, however, showed that the addition of the ISA mentioned above in the context of GCs + HCQ did not benefit patients with the first complete response. No difference was found between GCs + HCQ and GCs + HCQ + ISA group in SR. Were patients in GCs + HCQ + ISA group given ISA due to more severe illness or poorer response to initial induction therapy? There was no difference between the two groups in SLEDAI-2K before treatment and TTCR (Table 4), suggesting similar baseline and induction response. The treatment data of patients also showed that the vast majority of patients (24/27) were given ISA at the beginning of treatment or after complete response. The addition of ISA, with a relatively high proportion (24 out of 65), seems to be consistent with the current guidelines,^1,4 not due to more severe illness or poorer response. Admittedly, the value of this clinical finding is greatly limited by the small number of patients enrolled in this study. The SR curves by the Cox regression analysis were, however, highly consistent, at least reflecting the trend of similar efficacy of maintenance therapy in the two groups. In fact, several large studies^18,19 have shown that HCQ is a factor that reduces disease flare after GCs withdrawal, whereas ISA is not necessarily.²¹ Therefore, there is no doubt about the role of HCQ in reducing SLE relapse. The role of ISA, however, is still controversial, perhaps due to differences of the dose and type of ISA, as well as differences in the phenotype of patients enrolled in different studies. Overall, the effect of ISA on reducing SLE relapse is influenced by many factors and needs to be confirmed by more targeted studies. HCQ, as a safe and effective drug, is currently recognized as a background treatment for SLE and is recommended by various guidelines. At present, there are few studies on HCQ in SLE-ITP. Two previous small sample study (including 11 and 12 SLE-ITP patients, respectively)^4,22 showed that HCQ was a safe and effective treatment for SLE-ITP patients showing failed long-lasting response to prednisone alone or with second-line treatment.^23,24 Therefore, although it is difficult to confirm that the combination of GCs and HCQ can be as effective as additional ISA in combination with GCs and HCQ in reducing significant relapse of SLE-ITP due to the limited sample size, our findings also suggest a new idea that is worthy of further observation and verification in clinical practice.

IVIG pulse therapy cannot produce SR.²⁴ In this study, there was no difference in the proportion of patients treated with IVIG pulse therapy between SCR + SR group and significant relapse group (50.0% versus 57.9%) (Table 2, p > 0.05). Therefore, combined with literature report, we ruled out the influence of IVIG pulse therapy on the long-term outcomes of patients in this study cohort.

Several limitations should be noted in the interpretation of our results. First, although a small number of patients in our cohort achieved discontinuation or GC-free remission, it is not known whether discontinuation or discontinuation of GCs is appropriate for more SLE-ITP patients due to the high heterogeneity of SLE. Second, although we believe that long-term maintenance therapy is a prerequisite for discontinuation, it is difficult to determine the specific duration of maintenance therapy. Third, a few previous studies focused on administration of HCQ for SLE-ITP conditions,²⁵ and it is recognized that race and ethnicity significantly affect SLE incidence, prevalence, and mortality.²⁶ Therefore, it is unclear whether our results are applicable to other ethnic groups and races, and this should be taken into account when interpreting our results.

Conclusion

Drug withdrawal, especially abrupt withdrawal with insufficient duration of maintenance therapy, is an independent risk factor for SLE-ITP relapse. The strategy of GCs use in SLE-ITP maintenance therapy should be slow GCs tapering to withdrawal after a long-term maintenance remission. HCQ in combination with GCs is effective in reducing significant relapse of SLE-ITP, which may be similar to that of additional combined ISA. After first complete response, HCQ combined with GCs is expected to be the first choice of the maintenance therapy for SLE-ITP patients who are not accompanied by other important organ damage.

Footnotes

Acknowledgements

None.

Declarations

ORCID iD

He-Jun Li

References

Fanouriakis

Tziolos

Bertsias

, et al. Update on the diagnosis and management of systemic lupus erythematosus. Ann Rheum Dis 2021; 80: 14–25.

Jung

Soh

Ahn

, et al. Thrombocytopenia in systemic lupus erythematosus: clinical manifestations, treatment, and prognosis in 230 patients. Medicine (Baltimore) 2016; 95: e2818.

Zeng

Zheng

Rui

, et al. Risk factors for the flare of systemic lupus erythematosus and its influence on prognosis: a single-center retrospective analysis. Adv Rheumatol 2021; 61: 43.

Fanouriakis

Kostopoulou

Alunno

, et al. 2019 update of the EULAR recommendations for the management of systemic lupus erythematosus. Ann Rheum Dis 2019; 78: 736–745.

Neunert

Terrell

Arnold

, et al. American society of hematology 2019 guidelines for immune thrombocytopenia. Blood Adv 2019; 3: 3829–3866. Erratum in: Blood Adv 2020; 4: 252.

Petri

Orbai

Alarcón

, et al. Derivation and validation of the systemic lupus international collaborating clinics classification criteria for systemic lupus erythematosus. Arthritis Rheum 2012; 64: 2677–2686.

van Vollenhoven

Bertsias

Doria

, et al. 2021 DORIS definition of remission in SLE: final recommendations from an international task force. Lupus Sci Med 2021; 8: e000538.

Koh

Park

Lee

, et al. Clinical characteristics and survival of 413 patients with systemic lupus erythematosus in southeastern areas of South Korea: a multicenter retrospective cohort study. Int J Rheum Dis 2020; 23: 92–100.

Adwan

. Clinical and serologic characteristics of systemic lupus erythematosus in the Arab world: a pooled analysis of 3,273 patients. Arch Rheumatol 2018; 33: 455–463.

10.

Ziakas

Dafni

Giannouli

, et al. Thrombocytopenia in lupus as a marker of adverse outcome – seeking Ariadne’s thread. Rheumatology (Oxford) 2006; 45: 1261–1265.

11.

Depré

Aboud

Mayer

, et al. Efficacy and tolerability of old and new drugs used in the treatment of immune thrombocytopenia: results from a long-term observation in clinical practice. PLoS ONE 2018; 13: e0198184.

12.

Kado

McCune

. Treatment of primary and secondary immune thrombocytopenia. Curr Opin Rheumatol 2019; 31: 213–222.

13.

Medina-Quiñones

Ramos-Merino

Ruiz-Sada

, et al. Analysis of complete remission in systemic lupus erythematosus patients over a 32-year period. Arthritis Care Res (Hoboken) 2016; 68: 981–987.

14.

Apostolopoulos

Kandane-Rathnayake

Raghunath

, et al. Independent association of glucocorticoids with damage accrual in SLE. Lupus Sci Med 2016; 3: e000157.

15.

Apostolopoulos

Morand

. It hasn’t gone away: the problem of glucocorticoid use in lupus remains. Rheumatology 2017; 56: i114–i122.

16.

Mathian

Pha

Haroche

, et al. Withdrawal of low-dose prednisone in SLE patients with a clinically quiescent disease for more than 1 year: a randomised clinical trial. Ann Rheum Dis 2020; 79: 339–346.

17.

Strehl

Bijlsma

de Wit

, et al. Defining conditions where long-term glucocorticoid treatment has an acceptably low level of harm to facilitate implementation of existing recommendations: viewpoints from an EULAR task force. Ann Rheum Dis 2016; 75: 952–957.

18.

Gao

Hao

, et al. Low-dose glucocorticoids withdrawn in systemic lupus erythematosus: a desirable and attainable goal. Rheumatology (Oxford) 2022; 62: 181–189.

19.

Fasano

Coscia

Pierro

, et al. Which patients with systemic lupus erythematosus in remission can withdraw low dose steroids? Results from a single inception cohort study. Lupus 2021; 30: 991–997.

20.

Wilhelm

Magder

Petri

. Remission in systemic lupus erythematosus: durable remission is rare. Ann Rheum Dis 2017; 76: 547–553.

21.

Xie

Fasano

, et al. Risk factors of flare in patients with systemic lupus erythematosus after glucocorticoids withdrawal. A systematic review and meta-analysis. Lupus Sci Med 2022; 9: e000603.

22.

Kidney Disease: Improving Global Outcomes (KDIGO) Glomerular Diseases Work Group. KDIGO 2021 clinical practice guideline for the management of glomerular diseases. Kidney Int 2021; 100(4S): S1–S276.

23.

Khellaf

Chabrol

Mahevas

, et al. Hydroxychloroquine is a good second-line treatment for adults with immune thrombocytopenia and positive antinuclear antibodies. Am J Hematol 2014; 89: 194–198.

24.

Arnal

Piette

Léone

, et al. Treatment of severe immune thrombocytopenia associated with systemic lupus erythematosus: 59 cases. J Rheumatol 2002; 29: 75–83.

25.

Mohammadpour

Kargar

Hadjibabaie

. The role of hydroxychloroquine as a steroid-sparing agent in the treatment of immune thrombocytopenia: a review of the literature. J Res Pharm Pract 2018; 7: 4–12.

26.

Stojan

Petri

. Epidemiology of systemic lupus erythematosus: an update. Curr Opin Rheumatol 2018; 30: 144–150.