Haemoglobin response to senicapoc in patients with sickle cell disease: a re‐analysis of the Phase III trial

Abstract

In 2011, the British Journal of Haematology published the results of the Phase III study for senicapoc in sickle cell disease (SCD).1 Despite encouraging improvements of biomarkers indicating decreased haemolysis and improved anaemia, the study failed to show changes in acute painful crises, the primary clinical endpoint of the trial. In addition, a statistically significant increase in the crises rate was reported in the subgroup of patients treated with senicapoc versus placebo, not concomitantly taking hydroxycarbamide (HC). These results led to the discontinuation of the clinical research for senicapoc in SCD. On 25 November 2019, the United States Food and Drug Administration (FDA) granted accelerated approval for the use of voxelotor in adult and paediatric patients aged ≥12 years with SCD.2 In a paradigm shift, the FDA granted this approval based on increase in haemoglobin (Hb), contingent ‘upon verification and description of clinical benefit in confirmatory trial(s)’. As described in the published report for the trial,3 an increase in Hb of >10 g/l from baseline after 24 weeks of therapy was observed in 38% of the patients taking voxelotor 900 mg/day and in 59% of those taking 1500 mg/day. A voxelotor dose of 1500 mg resulted in significant reductions in indirect bilirubin and percentage reticulocyte values, while changes in absolute reticulocyte counts and lactate dehydrogenase (LDH) were not significant. The lower voxelotor dose failed to produce significant changes in biomarkers of haemolysis. The purpose of the present study was to highlight the changing endpoints required for regulatory agency drug approval in SCD since the senicapoc trial. Senicapoc is presently off patent and this analysis is unrelated to any efforts to repurpose the drug. We report here a re-analysis of the senicapoc Phase III data, similar to the one performed for the voxelotor study, namely highlighting the proportion of patients who responded to therapy with an increase in Hb of >10 g/l from baseline after 24 weeks of therapy. Although 145 patients were treated with senicapoc (61 without and 84 with HC), due to the premature termination of the trial, only 85 were treated for at least 24 weeks: Hb response was observed in 34% of the patients treated with senicapoc (29/85), with a 36% response in patients treated with senicapoc plus HU (21/59), 31% response in patients treated with senicapoc alone (eight of 26), whereas only 6% of the patients in the placebo group (five of 85) showed a Hb response. The Hb response was also evaluated in the voxelotor trial as the percentage of patients with Hb values of ≥100 g/l at week 24: using this definition, 42·4% (36/85) of all senicapoc-treated patients, 46·2% (12/26) of patients treated with senicapoc alone, and 40·7% (24/59) of patients treated with senicapoc plus HC showed a response. Among patients who did not receive senicapoc or HC, 18·8% (six of 32) had Hb values of ≥100 g/l at week 24. These data compare with 20% and 41% of patients having Hb values of ≥100 g/l at week 24 with 900 mg and 1500 mg voxelotor respectively. Supplementary Figure S1 shows a comparison of the Hb increase observed in the senicapoc and voxelotor trials. Supplementary Table S1 presents a comparison between the senicapoc and voxelotor effects on Hb and markers of haemolysis. These data highlight a similarity of Hb response between senicapoc and 900 mg voxelotor, with greater reductions in haemolysis markers observed following treatment with senicapoc. In the senicapoc trial, sickle cell painful crises (SCPC) were quantified as the SCPC rate, derived from the total number of adjudicated crises divided by the number of months of received treatment. In the original report, for patients not receiving HC, the mean (SEM) SCPC rate was 0·37 (0·05, n = 61) for senicapoc versus 0·29 (0·08, n = 65) for placebo (P = 0·037).1 Changes in SCPC were not statistically significant for all patients treated with senicapoc as well as those treated with senicapoc plus HC. Data for SCPC rates were re-examined for responders (Hb increase >10 g/l at week 24) versus non-responders. As shown in Table 1, for all the senicapoc responders (29/85), as well as for the subgroups of patients treated with either senicapoc alone (8/26) or senicapoc plus HC (21/59), the SCPC rates were not significantly different from those observed in the placebo group. Similar findings persisted when patients were subdivided into two strata, based on having a prior history of two to four crises or five or more crises in the prior 12 months. P Placebo vs. responders P Placebo vs. non-responders Senicapoc all N = 85 n = 29 0·33 (0·08) n = 56 0·36 (0·05) N = 144 0·31 (0·04) Senicapoc alone N = 26 n = 8 0·39 (0·18) n = 18 0·34 (0·11) N = 65 0·29 (0·08) Senicapoc plus HC N = 59 n = 21 0·31 (0·09) n = 38 0·37 (0·05) N = 79 0·33 (0·05) Senicapoc all N = 51 n = 20 0·18 (0·03) n = 31 0·24 (0·05) N = 93 0·21 (0·05) Senicapoc alone N = 17 n = 6 0·33 (0·08) n = 11 0·33 (0·08) N = 43 0·23 (0·10) Senicapoc plus HC N = 34 n = 14 0·17 (0·04) n = 20 0·22 (0·04) N = 50 0·19 (0·04) Senicapoc all N = 34 n = 9 0·67 (0·20) n = 25 0·50 (0·08) N = 51 0·50 (0·07) Senicapoc alone N = 9 n = 2 0·94 (0·64) n = 7 0·42 (0·17) N = 22 0·41 (0·11) Senicapoc plus HC N = 25 n = 7 0·59 (0·22) n = 18 0·53 (0·09) N = 29 0·56 (0·09) It is a potential concern, that by looking in this re-analysis of smaller sub-groups of each population, differences between groups need to be very large to be statistically significant. To increase the number of evaluable patients, we present in Supplementary Table S2, an analysis which defines Hb response as an increase >10 g/l from baseline for at least two consecutive monthly visits at any time during the trial. With this approach, 127 patients treated with senicapoc could be analysed: the values for SCPC rates in the Hb responders appeared to be lower than in the non-responders, without reaching statistical significance. In addition, they were not significantly different from placebo. These re-analyses of the senicapoc clinical trial data show a Hb response that is essentially similar to voxelotor 900 mg/day, with a more pronounced improvement of haemolytic markers and no significant changes in the crises rate for patients who responded with either an increase in Hb of >10 g/l after 24 weeks of treatment or reaching a Hb value of ≥100 g/l at week 24, or experiencing an increase in Hb of >10 g/l from baseline at two consecutive monthly intervals. While the aggregate data were not deemed promising enough to warrant FDA approval for the use of senicapoc in SCD in 2011, it is quite possible that the outcome might have been different with the current criteria emphasising Hb increase even in the absence of measurable reductions in painful episodes or modification of other clinical endpoints. The authors declare no competing financial interests. Carlo Brugnara and Steven Staffa designed and performed research, analysed the data, and wrote the paper. Kenneth Ataga and Jonathan Stocker analysed the data and wrote the paper. Figure S1. Panel A: Waterfall plot of change in haemoglobin level from baseline to week 24 for the 1500-mg dose of voxelotor, the 900-mg dose of voxelotor, and placebo in the per-protocol analysis that included observed data from the participants who completed 24 weeks of the assigned regimen and did not initiate hydroxyurea after randomisation and before week 24; vertical bars in the plot represent individual participants. The percentage of participants who had a haemoglobin (Hb) response (defined as an increase in the Hb level of >10 g/l from baseline at week 24) is reported for each trial group. Baseline values were calculated as the mean values of the data collected at screening and on the day of randomisation. Values for the response at week 24 were calculated as the mean values of the data collected at weeks 20 and 24. (Figure and legend reprinted from NEJM with permission). Panel B: Waterfall plot of change in Hb values from baseline to week 24 for the 88 patients treated with senicapoc, and placebo (n = 85). Panel C: Waterfall plot of change in Hb values from baseline to week 24 for the 59 patients treated with senicapoc and hydroxycarbamide (HC), the 26 patients treated with senicapoc alone, the 53 placebo patients treated with HC and the 32 placebo patients receiving no other treatment. Table S1. Effects of senicapoc and voxelotor on haemoglobin and haemolytic parameters at 24 weeks of treatment Table S2. Analysis of sickle cell painful crises (SCPC) rates for haemoglobin (Hb) responders. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts