Discussion
In this retrospective cohort study, we assessed the effect of HCQ on platelet counts in children with chronic ITP, in the context that HCQ was usually used in patients with positive ANA and in combination with other treatments. Our findings supported that HCQ was associated with increased platelet counts in children with chronic ITP and suggested that the baseline ANA positivity may be not associated with the efficacy of HCQ.
Data from our study showed potential benefits from HCQ in treating children with chronic ITP. Compared with patients who don’t receive HCQ during the course, patients who received HCQ had a higher response rate at the latest follow-up (69.0% vs. 48.3%; odds ratio, 2.39; 95% CI, 1.15 to 4.95). In our study, at 3 months from HCQ initiation, 21 (56.8%) patients responded, and 15 (40.5%) achieved CR/R at 1 year from HCQ initiation. Previously, a study from Roche et al. reported a 1-year response rate of 56.5% in children with ITP10, and a study from Brik-Simon et al. reported a higher response rate of 82.4% (14/17) in children with SLE-related ITP at 1-year after HCQ initiation 9, 15. For adults, a study from Khellaf et al. reported an overall response rate of 60% (24/40) with a median follow-up of 64 months in ITP patients with ANA titer ≥ 1:16016. Of note, results from above studies showed that patients with manifestation associated with autoimmune diseases may have a better response to HCQ, which was usually explained by the effect of HCQ on immunomodulatory and the previously successful experience of treating SLE patients with HCQ. However, our results showed that there was no difference in symptoms and markers associated with autoimmune diseases between chronic ITP patients with response to HCQ and those without, indicating other potential mechanism of HCQ in treating patients with ITP such as its effect on the autophagy and apoptosis of megakaryocytes17, 18.
In patients who received monotherapy of HCQ, 10 of 14 (71.4%) had no response to the treatment at 3 months after initiation, while in patients who received HCQ combined with corticosteroids, 14 of 18 (77.8%) achieved response. Considering the failure of corticosteroids in maintaining platelet counts before receiving HCQ and the low response rate of HCQ used as a monotherapy in our study, we speculate that the effectiveness of HCQ needs the facilitation of corticosteroids, which was also noted by Mohammadpour et al. whose review noticed the role of HCQ as a steroid-sparing agent in the treatment of ITP19. Furthermore, to assess the effect of HCQ in treating ITP in the context that some patients receive HCQ combined with other drugs, we conducted mixed-effect models and results revealed that HCQ could facilitate increasing platelet counts after considering the potential influence of concomitant treatments, sex, age, and ANA level on platelet counts (mean difference between HCQ and non-HCQ groups: 23.82×109/L; 95% CI: 7.44 ~ 40.21).
Previously, researchers paid more attention to patients with positive ANA and few studies analyzed the association between ANA and the effect of HCQ in patients with ITP. In this study, we did not find interaction effect of initial ANA level on the treatment efficacy of HCQ through mixed-effect models. And there was no difference in ANA positivity between patients with response to HCQ and those without. These findings remind us that ANA positivity is not a completely reliable indicator for the introduction of HCQ in patients with chronic ITP. And this might support the use of HCQ in patients with ANA at a titer of less than 1:160 because they may also benefit from the treatment. However, the insignificant interaction result should be interpreted with caution because this study was not powered to find a significant interaction effect. Larger sample studies are needed to explore the efficacy of HCQ in ANA-negative ITP children and discuss its mechanism.
HCQ associated side effects were recorded in 6 patients (6/42, 14.3%), and most of them (4/6,66.7%) received HCQ more than 2 years. Thus, we should take caution when using HCQ for long-term treatment of children with ITP. In children with proliferative lupus nephritis who were treated with HCQ, Gheet et al. found that 3.3% of patients experienced mucocutaneous alopecia, 10% of patients experienced hyperpigmentation, and 6.7% of patients experienced retinopathy.20Cutaneous side effects due to HCQ can disappear within few weeks of discontinuation,21 while retinopathy due to HCQ can be irreversible.22, 23 In this study, we noticed that one patient had no ophthalmologic monitoring after the discontinuation of HCQ, reminding clinicians of the importance of patient education and long-term follow-up. Besides, searching for effective monitoring means of side effects will be helpful for the use of HCQ in a safe and rational way.
This study had some limitations. Firstly, the dosage of drugs was not taken into consideration in mixed-effects models due to missing data, which was mainly restricted by its retrospective nature. Also, the missing value of platelet counts at certain time made it difficult to evaluate the response time of HCQ. In addition, the dosage of HCQ was not consistent in each patient, which may lead to a potential bias when we assessed the effect of HCQ on platelet counts.