Effectiveness and Safety of Immunosuppressants and Biological Therapy for Chronic Spontaneous Urticaria: A Network Meta-Analysis

Chronic spontaneous urticaria (CSU) is the most common phenotype of chronic urticaria. We compared treatment effects and safety profiles of the medications in patients with CSU. We searched PubMed, MEDLINE, and Web of Science for randomized control trials (RCTs), from 1 January 2000 to 31 July 2021, which evaluated omalizumab and immunosuppressants. Network meta-analyses (NMAs) were performed with a frequentist approach. Outcome assessments considered the efficacy (Dermatology Life Quality Index (DLQI) and weekly urticaria activity score (UAS7)) and tolerability profiles with evaluations of study quality, inconsistencies, and heterogeneity. We identified 14 studies which we included in our direct and indirect quantitative analyses. Omalizumab demonstrated better efficacy in DLQI and UAS7 outcomes compared to a placebo, and UAS7 assessments also demonstrated better outcomes compared to cyclosporine. Alongside this, omalizumab demonstrated relatively lower incidences of safety concerns compared to the other immunosuppressants. Cyclosporin was also associated with higher odds of adverse events than other treatment options. Our findings indicate that omalizumab resulted in greater improvements in terms of the DLQI and UAS7 with good tolerability in CSU patients compared to the other immunosuppressants.


Search Strategy and Study Selection
We included RCTs from electronic databases, including PubMed, MEDLINE, and Web of Science, from 1 January 2000 to 31 July 2021. In addition, we hand-searched references from included papers and relevant systematic reviews for additional relevant trials to identify potential publications.
The characteristics of included studies should meet following criteria: (1) the study design should be RCTs of CSU/CIU treatments, (2) recruited patients over the age of 12; and (3) published in the English language. Multiple-arm studies with a placebo plus H1 antihistamine agents (standard or up to 4-fold doses) and add-on active comparators with different dosages were acceptable. Two investigators independently screened all titles, abstracts, and full papers, using the eligibility criteria below, with any disagreements resolved through discussion. Two reviewers critically appraised the methodologic quality of each included study using the Jadad scale and revised Cochrane tool [12,13]. According to the revised Cochrane tool, the overall risk of bias was then classified as low, some concern, or major concern. Furthermore, based on the Jadad scale, a study with a score of <3 represented a low-quality and high-bias risk study, and a study with a score exceeding 3 was considered a high-quality trial ( Figure S1). Overall assessments of imprecision, incoherence, and heterogeneity are summarized in Figure S2.

Eligibility Criteria
All included studies had to meet the following criteria: (1) symptomatic CSU patients despite being under H1 antihistamine treatment (standard or up to a 4-fold dose) and (2) older than 12 years of age. However, we excluded trials if (1) they were a non-human study, (2) age of recruited participants was less than 12 years old, and (3) the diagnosis was not CSU or CIU.

Intervention and Comparators
This study compared selected pharmacological management, including immunosuppressants such as methotrexate, cyclosporin, hydroxychloroquine, azathioprine, biological therapy, and omalizumab, and the leukotriene receptor antagonist, montelukast, in symptomatic CSU patients. We excluded studies with non-pharmacological treatment or studies that used other medications which did not fit our inclusion criteria. In terms of comparators, a placebo plus an antihistamine (a standard dose or 4-fold dose) or active comparators plus an antihistamine (a standard dose or 4-fold dose) could be included (Table S1).

Study Outcomes
The primary outcome was the mean change in DLQI scores from the baseline. Secondary outcomes included the mean change in the UAS7 from the baseline and the proportion of patients experiencing at least one adverse effect.

Statistical Approaches
This NMA was performed using the CINeMA (Confidence in Network Meta-Analysis), and we also evaluated the certainty of the final body of evidence [14]. We employed a random-effects model framework for both continuous and binary outcomes. To estimate continuous variables, results were demonstrated by the mean difference (MD) with the 95% confidence interval (CI); in terms of binary outcomes, the effects were depicted by the odds ratio (OR) with the 95% CI.
To produce an integrated dataset, we applied a statistical method to unify the data presentation for continuous variables as the mean ± standard deviation (SD). There were two scenarios: (1) interconversion between the mean with the 95% CI and mean ± SD and (2) interconversion between the median with the range and the mean ± SD [15,16].
continuous variables, results were demonstrated by the mean difference (MD) with 95% confidence interval (CI); in terms of binary outcomes, the effects were depicte the odds ratio (OR) with the 95% CI.
To produce an integrated dataset, we applied a statistical method to unify the presentation for continuous variables as the mean ± standard deviation (SD). There two scenarios: (1) interconversion between the mean with the 95% CI and mean ± SD (2) interconversion between the median with the range and the mean ± SD [15,16].

UAS7 Score Improvement
Mean changes in urticarial symptoms, including hives and pruritus, are presented in Table 3. The largest change was observed in those using omalizumab treatment with an MD of −12.0 (95% CI, −19.988 to −4.012) compared to a placebo. In addition, the cyclosporin group also showed a great reduction in the UAS7 with an MD of −10.4 (95% CI, −18.587 to −2.213). There was no significant difference in the montelukast group (MD, −0.366; 95% CI, −5.315 to 4.583), but it still demonstrated an improving trend. There was an interesting finding that the efficacy of the methotrexate group was weaker than a placebo. This result was also reported by a previous NMA publication [11].
In addition, there were similar results between the DLQI and UAS7, suggesting that omalizumab has a dose-dependent effect. Our results are consistent with reports of real-world evidence [44,45]. There was no significant difference between omalizumab at 300 mg/Q4W and cyclosporin, but it still indicated an improving trend in the group of omalizumab at 300 mg/Q4W. Network comparisons, on the other hand, suggested a significant improvement with omalizumab at 300 mg/Q4W compared to montelukast and methotrexate.

Adverse Events
We found no significant difference among all treatment comparisons in terms of at least one adverse event. Nevertheless, we noticed some interesting trends. Cyclosporin and montelukast, for instance, were associated with higher odds of adverse events than other treatment options. The NMA league table of adverse events is shown in Table 4.

Discussion
There are now several approved medications available for CSU, including H1 antihistamines and omalizumab. In clinical settings, physicians may use other immunosuppressants, including cyclosporin, methotrexate, azathioprine, and hydroxychloroquine. In the absence of head-to-head studies, we conducted this NMA to evaluate the comparative efficacy and tolerability of biotherapeutic agents and immunosuppressants in symptomatic CSU patients in addition to standard treatment.
Compared to immunosuppressants, omalizumab was generally associated with greater reductions in DLQI and UAS7 scores, while also demonstrating a low incidence of adverse effects. Cyclosporin was recommended as fourth-line treatment for uncontrolled CSU patients by EAACI⁄GA2LEN⁄EDF⁄WAO guidelines. In clinical settings, physicians may try to manage refractory CSU patients using cyclosporin or other immunosuppressants due to economic considerations. Indeed, there are few studies that reviewed the safety and efficacy of cyclosporin. Kulthanan et al. [46] conducted a systematic review and meta-analysis demonstrating that cyclosporin was an effective treatment option for CSU patients. However, adverse events occurred in more than half of patients treated with a moderate dose. Our analysis indicated similar results, that cyclosporin could lead to increased incidences of adverse effects compared to omalizumab. After reviewing the clinical evidence available to date, we conclude that cyclosporin can provide benefits to patients with CSU, despite a higher incidence of adverse effects.
Our analysis suggests that the effectiveness of omalizumab in CSU management is dose dependent. This observation is supported by an RCT, by clinical evidence, and also by some up-dosing studies in a real-world setting [44,45]. The pathophysiology of CSU is associated with immunoglobulin E (IgE), IgG, or related autoreactivity factors which activate mast cells and basophils, leading to CSU symptoms [47]. Mast cells and basophils can be stabilized through neutralizing the serum free form of IgE and downregulating the high-affinity IgE receptor. The most important pharmacological effect of omalizumab is IgE receptor downregulation, and this action occurs by neutralizing the serum-free form of IgE to reduce the level to near zero [48]. Our results and the scientific theory suggest that omalizumab confers superior efficacy and a good safety profile compared to other immunosuppressants. On the other hand, Turk et al. suggested that CSU patients who are partial responders to omalizumab may benefit from up-dosing of omalizumab or shortening the treatment interval in different conditions [49]. This may be explained by Chang et al.'s study, which demonstrated that an adequate anti-IgE dose can maintain the serum-free level of IgE to near zero, leading to subsequent pharmacological effects to achieve symptom improvement [48]. Therefore, CSU patients who respond to an increasing dose might have higher blood IgE levels or body weight than CSU patients who are responders to the approved dose of omalizumab.
Another intriguing finding was that the efficacy of methotrexate was found to be weaker than a placebo in CSU management. In fact, this result was also supported by a previous systematic review and meta-analysis, which found that there was no significant benefit of adding methotrexate to an antihistamine in refractory urticaria management [50]. Due to limited adequate references of methotrexate in CSU management, there was only one clinical trial that met our inclusion criteria for inclusion in our quantitative analysis. That was the same circumstance we encountered in the group of omalizumab 600 mg/Q4W; there was only one arm in one study which tested this specific regimen. The small sample size limits the validity of these results.

Strengths and Limitations
To our knowledge, this is the first NMA study not only to analyze a questionnaire of the QoL, which is the focus of the DLQI, as the primary endpoint, but also to focus on immunosuppressants and biologics on the market. We conducted a comprehensive and robust systematic review and NMA to produce some clinical insights.
There are some limitations to our review and NMA. First, statistical calculations were implemented to maintain consistency in the data presentation when conducting the analysis. This may have over-or under-estimated the effects of related medications. Second, smaller sample sizes in individual studies might have led to a high level of heterogeneity by the large variation in the magnitude of the effect across all included studies. Finally, we only included RCT studies to conduct the NMA. As a result, our findings may not represent all refractory CSU patients due to low external validity.

Conclusions
In conclusion, our results suggest that biological treatment with omalizumab produced a greater improvement in the DLQI and UAS7 in CSU patients compared to other immunosuppressants. The safety profile of omalizumab found in this NMA was similar or superior to those of other immunosuppressants. These NMA results on treatment options in CSU can help guide our clinical practice and serve as evidence for policymakers for revising treatment recommendations.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/biomedicines10092152/s1. Table S1. Study inclusion and exclusion criteria. Table S2. Characteristics of excluded studies. Table S3. Summary of the risk of bias assessment. Figure S1. Risk of bias bar chart of direct and indirect comparisons between omalizumab and other immunosuppressants. Figure S2. Summary of overall assessments of imprecision, incoherence, and heterogeneity of indirect and direct comparisons between omalizumab and other immunosuppressants. Figure