Omalizumab for Chronic Idiopathic or Spontaneous Urticaria
Omalizumab for Chronic Idiopathic or Spontaneous Urticaria
When considering whether the authors took appropriate measures to minimize bias, we identified a few areas of concern.
Some of the eligibility criteria required are not adequately described in the paper; patients aged 12–75 years were considered eligible in all countries except Germany, where the minimum age was 18 years. There is no explanation as to why this was the case, raising concerns about possible safety profiling of omalizumab in adolescents in Germany.
Inclusion criteria included a history of CIU for 6 months and presence of hives associated with itch for at least eight consecutive weeks any time prior to enrolment, despite current use of H1-antihistamines. It is not clear whether this was assessed from retrospective clinical records or patient self-reporting. Similarly, the authors do not describe assessment of prior use of antihistamines, patient compliance and use of other treatments prior to enrolment.
Overall, the process of assessment of eligibility and the recruitment of participants are not clearly conveyed in the paper: it is not obvious whether patients were recruited from primary, secondary or tertiary centres; how many screening visits were conducted; the nature of the visits or the duration of the recruitment period. These factors are important as they are likely to have an impact on disease severity, and although at baseline there appears to be no difference between the groups, there are implications for generalization of these data as it is unclear which population the study would be applicable to.
Adequate randomization protects against selection bias. A hierarchical dynamic randomization scheme was used to achieve overall balance between treatment groups. Patients were randomly assigned using the Interactive Voice and Web Response System, in a 1:1:1:1 ratio. This type of randomization is generally considered advantageous over block randomization, as potential confounding factors should be more evenly distributed between the groups than chance alone might otherwise ensure. Despite these measures taken to establish randomization, analysis of the most frequent medication used by patients for their disease prior to enrolment in the study demonstrates lower usage of steroids, immunosuppressants and LTRAs in the 300-mg omalizumab arm at baseline compared with the placebo group (45·6% vs. 51·9%, 6·3% vs.11·4% and 19% vs. 26.6%, respectively). This suggests that there may have been degrees of inequality at baseline, with possibly less severe disease in the 300-mg arm. However, no statistically significant difference was reported, with similar baseline disease severity scores across all groups.
Although the study protocol states that both subject and investigator were blinded, this is not clearly reported by the authors, nor is there any detail on how this was achieved. The patients were apparently all treated identically apart from the injected substance administered through two injections at each visit. It is not stated what substance was actually used for placebo or whether the patient would have had any indication from appearance or symptoms that a placebo was being delivered as opposed to a treatment. Equally, doses of omalizumab or placebo were administered by a clinician who was not involved in the evaluation of patients' symptoms, and access to laboratory results 'that could potentially unblind treatment assignments was restricted'. However, it remains plausible that unblinded clinicians may have inadvertently informed the subjects which treatment they were receiving. The potential for systematic difference in the care provided to the participants in the intervention and placebo groups raises the issue of performance bias during the trial.
In an intention-to-treat analysis, all patients are followed until death or until the end of the trial, irrespective of whether they are taking the treatment to which they were randomized. In this study, the primary outcome was assessed as an intention-to-treat analysis, with missing data at week 12 imputed with the baseline score.
However, no explanations were provided for participants withdrawing from the study, or being withdrawn by a guardian or physician, and this information should have been clearly documented. For the analysis of 'angio-oedema-free days' an intention-to-treat analysis was not performed, as patients who withdrew before week 4 or those who had missing responses for > 40% of the daily diary entries were excluded. Furthermore, the analysis of mean weekly symptom scores was performed on a 'modified' intention-to-treat population; all patients who underwent randomization and received at least one dose of a study drug were included, omitting those who may have had other relevant reasons for not receiving the drug. Although this will not reflect the efficacy of the drug (as those omitted from analysis did not receive any treatment), it may reflect factors such as acceptability of the mode of drug administration.
To estimate the power to determine the efficacy of omalizumab, the authors assumed a mean change of nine points from baseline to week 12 in the itch severity score in the treatment group, and a mean change of 3·5 points in the placebo group. These figures were derived from phase 2 data and seem to provide an appropriate calculation.
Assessment of Validity
When considering whether the authors took appropriate measures to minimize bias, we identified a few areas of concern.
Eligibility Criteria and Recruitment
Some of the eligibility criteria required are not adequately described in the paper; patients aged 12–75 years were considered eligible in all countries except Germany, where the minimum age was 18 years. There is no explanation as to why this was the case, raising concerns about possible safety profiling of omalizumab in adolescents in Germany.
Inclusion criteria included a history of CIU for 6 months and presence of hives associated with itch for at least eight consecutive weeks any time prior to enrolment, despite current use of H1-antihistamines. It is not clear whether this was assessed from retrospective clinical records or patient self-reporting. Similarly, the authors do not describe assessment of prior use of antihistamines, patient compliance and use of other treatments prior to enrolment.
Overall, the process of assessment of eligibility and the recruitment of participants are not clearly conveyed in the paper: it is not obvious whether patients were recruited from primary, secondary or tertiary centres; how many screening visits were conducted; the nature of the visits or the duration of the recruitment period. These factors are important as they are likely to have an impact on disease severity, and although at baseline there appears to be no difference between the groups, there are implications for generalization of these data as it is unclear which population the study would be applicable to.
Randomization
Adequate randomization protects against selection bias. A hierarchical dynamic randomization scheme was used to achieve overall balance between treatment groups. Patients were randomly assigned using the Interactive Voice and Web Response System, in a 1:1:1:1 ratio. This type of randomization is generally considered advantageous over block randomization, as potential confounding factors should be more evenly distributed between the groups than chance alone might otherwise ensure. Despite these measures taken to establish randomization, analysis of the most frequent medication used by patients for their disease prior to enrolment in the study demonstrates lower usage of steroids, immunosuppressants and LTRAs in the 300-mg omalizumab arm at baseline compared with the placebo group (45·6% vs. 51·9%, 6·3% vs.11·4% and 19% vs. 26.6%, respectively). This suggests that there may have been degrees of inequality at baseline, with possibly less severe disease in the 300-mg arm. However, no statistically significant difference was reported, with similar baseline disease severity scores across all groups.
Blinding
Although the study protocol states that both subject and investigator were blinded, this is not clearly reported by the authors, nor is there any detail on how this was achieved. The patients were apparently all treated identically apart from the injected substance administered through two injections at each visit. It is not stated what substance was actually used for placebo or whether the patient would have had any indication from appearance or symptoms that a placebo was being delivered as opposed to a treatment. Equally, doses of omalizumab or placebo were administered by a clinician who was not involved in the evaluation of patients' symptoms, and access to laboratory results 'that could potentially unblind treatment assignments was restricted'. However, it remains plausible that unblinded clinicians may have inadvertently informed the subjects which treatment they were receiving. The potential for systematic difference in the care provided to the participants in the intervention and placebo groups raises the issue of performance bias during the trial.
Intention to Treat
In an intention-to-treat analysis, all patients are followed until death or until the end of the trial, irrespective of whether they are taking the treatment to which they were randomized. In this study, the primary outcome was assessed as an intention-to-treat analysis, with missing data at week 12 imputed with the baseline score.
However, no explanations were provided for participants withdrawing from the study, or being withdrawn by a guardian or physician, and this information should have been clearly documented. For the analysis of 'angio-oedema-free days' an intention-to-treat analysis was not performed, as patients who withdrew before week 4 or those who had missing responses for > 40% of the daily diary entries were excluded. Furthermore, the analysis of mean weekly symptom scores was performed on a 'modified' intention-to-treat population; all patients who underwent randomization and received at least one dose of a study drug were included, omitting those who may have had other relevant reasons for not receiving the drug. Although this will not reflect the efficacy of the drug (as those omitted from analysis did not receive any treatment), it may reflect factors such as acceptability of the mode of drug administration.
Power
To estimate the power to determine the efficacy of omalizumab, the authors assumed a mean change of nine points from baseline to week 12 in the itch severity score in the treatment group, and a mean change of 3·5 points in the placebo group. These figures were derived from phase 2 data and seem to provide an appropriate calculation.
Source...