#NephJC Summary
Tuesday, December 5th, 2023, at 9 pm Eastern
Wednesday, December 6th, 2023, at 9 pm Indian Standard Time and 3:30 GMT
Lancet. 2023 Nov 25;402(10416):2004-2017. doi: 10.1016/S0140-6736(23)02230-4. Epub 2023 Nov 3.
Zibotentan in combination with dapagliflozin compared with dapagliflozin in patients with chronic kidney disease (ZENITH-CKD): a multicentre, randomised, active-controlled, phase 2b, clinical trial
Hiddo J L Heerspink, Arihiro Kiyosue, David C Wheeler, Min Lin, Emma Wijkmark, Glenn Carlson, Anne-Kristina Mercier, Magnus Åstrand, Sebastian Ueckert, Peter J Greasley, Phil Ambery
PMID: 37931629
Introduction
Residual proteinuria despite maximum tolerated doses of RAS blockers and flozins (aka SGLT2 inhibitors) has led the nephrology community to look for new drugs to enhance their renoprotective effects over and above what we already have. A barrage of new drugs has been studied in proteinuric CKD in the past few years. As each tries to fight for its place to become the standard of care in CKD, will endothelin antagonists be able to find their place in this race?
Endothelin 1, a potent vasoconstrictor primarily released by the endothelium, acts via ET-A and ET-B receptors in antagonistic ways. ET-A receptor stimulation causes vasoconstriction, extracellular matrix deposition, and inflammation whereas ET-B stimulation causes vasodilatation, sodium and water excretion, antiproliferative effects and is also responsible for the metabolism as well as depletion of ET-1 (Kohan et al, KI 2014). Hence any beneficial effects of ET antagonism will be enhanced in drugs with selectivity for ET-A receptors and also be associated with fewer fluid related adverse effect.
Enter, stage left, zibotentan. With an ET-A : ET-B selectivity of 100,000 (Davenport AP et al, Pharmacol Rev 2016), zibotentan is the most selective of ET-A antagonists known. Its renal effects include (Kohan et al, KI 2014, Martinez-Diaz I et al, 2023, Int J Mol Sci):
Afferent arteriolar vasodilation causing increased renal plasma flow, decreased filtration fraction and decreased proteinuria
Podocyte cytoskeletal stabilization, decreased podocyte injury and shedding
Decreased mesangial cell contraction, proliferation and extracellular matrix deposition
Amelioration of of glomerular basement membrane glycocalyx structure and decreased glomerulosclerosis
Decreased tubulointerstitial fibrosis and inflammation
Whether these effects will translate into meaningful clinical benefits is yet to be known.
Figure from Kohan et al, KI 2014
The improvement in clinically relevant kidney outcomes with ET-A antagonists has been analyzed in several trials as shown in the table below.
ASCEND (Mann et al, JASN 2010) followed by RADAR (de Zeeuw et al, JASN 2014; NephJC summary) and SONAR (Heerspink et al , Lancet 2019, NephJC summary) have been previous attempts of endothelin antagonism in diabetic kidney disease, marred by safety (sodium retention) and early stoppage among other issues.
The recently published DUPLEX( Rheault et al, NEJM 2023, NephJC summary) and PROTECT (Rovin et al, Lancet 2023) trials used dual endothelin and angiotensin receptor blockade as first line agents against RAS inhibitor irbesartan in FSGS and IgAN respectively. These RCTs, which began some years ago, did not have flozins as part of the baseline standard of care - but after DAPA-CKD and EMPAKIDNEY, flozins now are standard of care. Indeed, the natriuretic and osmotic diuretic effect of flozins when combined with ET-A may counter fluid retention, simultaneously causing synergistic effects on proteinuria. The counterbalancing of zibotentan related fluid retention with dapagliflozin has been demonstrated earlier in animal and human studies (Veenit et al, NDT 2023; Heerspink et al, KI 2021). That was the rationale of the ZENITH-CKD trial, which we will discuss this week.
Figure from Heerspink et al. NDT 2023
The Study
Methods
The ZENITH-CKD was a Phase 2b, double-blind, randomized, active-controlled, parallel-group dose ranging study to assess the efficacy, safety and tolerability of zibotentan and dapagliflozin in participants with CKD. The trial was run across 18 countries in 170 centers. The study protocol was published earlier this year (Heerspink et al, NDT 2023).
Study Population
Inclusion criteria:
Trial included patients over 18 years old, diagnossed with CKD with a eGFR ≥20 mL/min, with urinary albumin-to-creatinine ratio (UACR) ≥150 and ≤5000 mg/g, on stable dose of ACEi/ARB/MRA for more than 4 weeks, not on SGLT2i and no immunosuppression. Notable, non-childbearing potential is well detailed in the supplemental appendix.
Key exclusion criteria:
The trial had a 36 point detailed exclusion criteria, key exclusion criteria are linked to previously known edothelin blockers side effects (heart failure, or underlining cardiac disease determining heart failure, anemia, hepatic impairment, patients on strong/moderate CYP34A4 inducers), or SGLT2 (type 1 DM, polycystic kidney disease), or both.
Interventions
The original study design consisted of two parts, part A and part B. Part A was designed to measure the effect of each agent (zibotentan or dapagliflozin) in isolation and that of their combination versus placebo. Part B, on the other hand, was designed to measure the effect of various doses of zibotentan in combination with dapagliflozin versus dapagliflozin alone. Eligible patients were either recruited to part A or part B.
In Part A, participants were randomized into the following four treatment arms:
zibotentan 5 mg once daily
dapagliflozin 10 mg once daily
zibotentan 5 mg +dapagliflozin 10 mg daily once daily
placebo once daily
In Part B, a separate cohort of eligible participants were randomized into six treatment arms: the same four treatment arms as Part A, and two additional treatment arms as below:
zibotentan 0.25 mg+ dapagliflozin 10 mg once daily
zibotentan 1.5 mg+ dapagliflozin 10 mg once daily
Figure from Heerspink et al, NDT 2023
In April 2022, an ad hoc safety review team advised to stop recruitment in the arms containing zibotentan 5 mg (alone or with dapagliflozin), owing to increased rate of fluid-retention events in these arms. Additionally, after the DAPA CKD trial (Heerspink et al NEJM 2020), flozins are recommended as standard of care in the management of CKD (NICE guidelines March 2022). This resulted in discontinuation of the placebo group. Thus, eligible candidates were randomized in a 2:1:2 ratio to the following 3 arms.
zibotentan 1.5 mg+ dapagliflozin 10 mg once daily
zibotentan 0.25 mg+ dapagliflozin 10 mg once daily
placebo + dapagliflozin 10 mg once daily
The ZENITH CKD study being discussed here reports the results of part B of the study. These participants were also stratified by eGFR and diabetes status. Recruitment was done in a way to ensure that between 30-50% of participants were non-diabetic CKD. The screening period was 4 weeks, total treatment duration 12 weeks followed by withdrawal of the trial drug and follow-up 2 weeks later to assess parameters after drug withdrawal.
Outcomes /Endpoints
The primary objective was to evaluate the effect of zibotentan 1.5 mg+ dapagliflozin 10 mg versus dapagliflozin alone on urine albumin creatinine ratio. The primary endpoint was the change in log-transformed UACR from baseline to Week 12
Secondary efficacy endpoints included:
the change in log-transformed UACR for Zibotentan 0.25 mg arm vs placebo (baseline to week 12)
the change in office systolic and diastolic blood pressure (all 3 arms)
the change from baseline in eGFR at Weeks 1, 12, and 14 (all 3 arms)
The change in eGFR in zibotentan 1.5 and 0.25 arms versus placebo (baseline to 12 weeks)
Exploratory endpoints included
Change in body weight (daily weight charting using digital scales provided)
Changes in cardiovascular biomarkers at all visits (BNP, nt-pro BNP)
Changes in total body water, extracellular water and intracellular water volumes using bioimpedance spectroscopy
Plasma concentrations of zibotentan metabolites, ET-1, c terminal pro-endothelin 1 and endothelin-like domain peptide
Safety
Safety analyses include assessment of adverse events (AEs), vital signs, clinical laboratory tests, 12-lead electrocardiogram, and other events of special interest, such as changes in fluid-related measures (body weight or BNP). Fluid retention was considered an adverse event of special interest in this trial and defined objectively. Any participant with a >3% increase in body weight (≥2.5% as total body water measured by bioimpedance) from baseline OR a 100% increase in BNP from baseline with BNP value >200 pg/mL without atrial fibrillation or BNP >400 pg/mL with atrial fibrillation was labeled to have “fluid retention”. Study drug was discontinued in patients with fluid retention.
Analysis
Assuming a 10% dropout rate, a sample size of 150 participants was required in each of the zibotentan 1.5 mg & placebo arms to detect at least a 25% change in UACR between the two arms with a statistical power of 80% and one-sided type 1 error of 5%. For the dose ranging study, 77 participants would be needed in zibotentan 0.25 mg to differentiate the dose response from zibotentan 1.5 mg arm assuming a maximum UACR reduction of 25% in zibotentan 1.5 mg vs zibotentan 0.25 mg and type 1 error rate of 5% with 78% power. Hence the 0.25 mg arm was half the other two arms. The data analysis for UACR was done using a mixed model for repeated measures. This statistical model is used to analyze data where multiple measurements are taken on the same subjects over time. This type of model is useful when dealing with repeated measurements because it can account for both fixed effects and random effects. The fixed effects in this study included stratification factors, study protocol, visit time, treatment and treatment by visit interaction. Variable effects included baseline log(UACR) and baseline log(UACR)-by-visit interaction. Unstructured covariance was used and missing data were not imputed but assumed to be missing at random. Every enrolled patient had to have at least 2 UACR readings including the baseline. A similar model was used for eGFR and blood pressure (secondary endpoints).
Funding Source
Astra Zeneca group (manufacturer of dapagliflozin and zibotentan) contributed to study design, data collection, and statistical analysis. It is not clear whether ‘contribution’ refers to complete data analysis, interpretation, and writing. Eight of the 11 authors are employees of the sponsor.
Results
Between April 2021 and January 2023, 1492 participants were screened and 525 included in the study. Part A of the study (with placebo, zibotentan 5 mg, and zibotentan 5 mg + dapagliflozin 10 mg) was discontinued after an increase in fluid-related events (total number recruited in part A= 62). The rest of the cohort was randomly divided into 3 arms: zibotentan 1.5 mg, zibotentan 0.25 mg, and placebo (n =180, 92 and 177 respectively). Details are as given below in Figure 1.
Figure 1 from Heerspink et al, Lancet 2023
Of note, 112 participants out of 449 (25%) in part B discontinued the intervention due to various reasons, most commonly adverse events (n =36) and protocol specified withdrawal criteria (n=27). Another 68 were withdrawn from the study, among which some more had adverse events (n = 15). This is higher than the anticipated 10% loss to follow up in the planned sample size. There were some more withdrawals by patient and doctor - which do not have any more details attached to them. There were some irregularities from two sites (2 participants from Poland, 12 participants from Japan) for whom the data were excluded in final analysis. It is not clear how that effects figure 1 above.
Baseline characteristics
Baseline characteristics of the patients are described in Table 1. Mean age of the study population was 62.8 years, 69% participants were male. Median UACR was 565.5 mg/g (63.85 mg/mmol) and median eGFR was 46.7 ml/min/1.73m². Roughly half the patients had diabetic kidney disease, and just over 10% had chronic glomerulonephritis.
Of note, numerically more patients with heart failure were present in the placebo (i.e. dapagliflozin alone) arm (10%) versus zibotentan arms (3% and 6% respectively, for 0.25 and 1.5 mg arms). 87% of patients were on ARB/ACEi. Although MRAs were allowed, data on percentage of patients on MRA and their distribution in the arms was not provided. About 40% of participants were on a diuretic at baseline.
Table 1. Baseline demographic and clinical characteristics from Heerspink et al, Lancet 2023
Primary endpoint
At week 12, the percentage change in UACR from baseline in zibotentan 1.5 mg arm was -52.5% (CI –59 to -44.9) versus -28.3% in placebo arm (CI -37.8 to -17.4) contributing to a -33.7% mean change from baseline as compared to placebo (CI -42.5 to -23.5, p< 0.0001). A keen eye might see the difference of 28.3 and 52.5 though is 24.2, not 33.7. However, the mean change isn't a simple math class reunion where everyone adds up. It's more like a party with a range – the confidence interval. It considers the wild side of data variability, making sure everyone's invited, even the uncertainties.
In the Zibotentan 0.25 mg arm, the percentage change in UACR from baseline was -47.7% (-55.7 to -38.2) amounting to a percentage change versus placebo arm of -27% (-38 to -13.6, p=0.022).
The percentage change in UACR was observed at week 3 and then maintained throughout the treatment duration. At 12 weeks, the drugs were withdrawn in all 3 arms and the UACR trended towards baseline in all 3 arms by week 14 without any difference in the 3 arms.
Figure 2. Mean change in UACR from Heerspink et al, Lancet 2023
(A) Bar graph of the percentage mean change in UACR from baseline to week 12 in the dapagliflozin 10 mg plus placebo, zibotentan 1·5 mg plus dapagliflozin 10 mg, and zibotentan 0·25 mg plus dapagliflozin 10 mg groups. (B) UACR trajectory over time in the three treatment groups. Vertical bars indicate the 90% CIs of the mean at given timepoints. UACR=urinary albumin-to-creatinine ratio.
The subgroup analysis showed a consistent benefit of zibotentan (both low and high dose) in decreasing UACR irrespective of diabetic status, eGFR and baseline SBP. No subgroup analysis on the basis of proteinuria was available.
Suppl Figure 1: Forest plot showing adjusted percentage mean change in UACR from baseline to week 12 (90%CI) overall, by type 2 diabetes status, by baseline eGFR and by baseline blood pressure for zibotentan 0.25 mg/dapagliflozin 10 mg and zibotentan 1·5 mg/dapagliflozin 10 mg.
Secondary endpoints
With respect to kidney function, an acute drop of eGFR consistent with a hemodynamic effect of the drugs was seen at week 1 in all the arms, the largest effect being in the zibotentan 1.5 mg arm. The percentage change in eGFR in zibotentan 1.5 mg arm vs dapagliflozin only arm at week 1 was –0.8 mL/min per 1.73 m² (90% CI –2.1 to 0·5) and at week 12 was –1.1 mL/min per 1.73 m² (–2.5 to 0.3). The corresponding difference in the zibotentan 0.25 mg arm at week 1 was 1.1 mL/min per 1·73 m² (–0.5 to 2.6) and at week 12 was -1.2 mL/min per 1.73 m² (–2.8 to 0.5). The eGFR returned to baseline in arms zibotentan 0.25 and placebo arms two weeks after the discontinuation of the drugs. However, the eGFR remained stable at the 12 week nadir in the zibotentan 1.5 mg arm.
Figure 3. Mean change from baseline in systolic blood pressure, diastolic blood pressure, and eGFR from Heerspink et al, Lancet 2023
Effect on BP, body weight and body water content
An SBP difference of -7.6 mmHg and -3.6 mm Hg along with a DBP difference of -5.4 mmHg and -3.0 mmHg was seen in zibotentan 1.5 and 0.25 mg arms as compared to dapagliflozin only arm. There was however, no correlation of change in UACR with change in SBP from baseline.
There was a reduction in mean body weight by -0.85 kg, -0.79 kg and -1.19 kg in Zibotentan 1.5, 0.25 mg and placebo arms respectively. There was an acute increase in weight in the Zibotentan 1.5 mg arm at week 1 which gradually came down over the period of 12 weeks. Weight was seen to decrease in both the zibotentan arms after drug withdrawal whereas it increased slightly in the placebo arm post drug withdrawal.
There was an increase in mean percentage change from baseline in extracellular fluid of 1·5% (SD 6·9) in the zibotentan 1·5 mg arm, whereas negligible change in zibotentan 0.25 arm and a -0.8% decrease in the dapagliflozin arm.A numerically larger reduction in LDL cholesterol and HbA1c is found in the zibotentan arms vs those on dapagliflozin alone.
Safety Analyses
The percentage of adverse events in zibotentan 1.5 mg, 0.25 mg, and placebo arms were 47%, 49%, and 37%, respectively. This led to discontinuation in 12%, 12%, and 4% respectively, which adds up to 40. This number is different than those who discontinued intervention or withdrawn for adverse events in figure 1 (36 + 15 making 51). Serious adverse events occurred in 5%, 2%, and 2% respectively - though which SAEs these were is not specified.
Table 2. Number of participants with adverse events from Heerspink et al, Lancet 2023
Fluid retention was seen in 18% participants in the zibotentan 1.5 mg arm as compared to 9% and 8% in zibotentan 0.25 mg and dapagliflozin only groups. The proportion of overall participants with fluid-related events was 12% (55/447). Six cases of heart failure were reported: four in the zibotentan 1·5 mg and two in the zibotentan 0·25 mg group and none in the dapagliflozin only group. Three of these cases were reported as serious adverse events. An additional 6%, 7% and 5% of patients in zibotentan 1.5 mg, 0,25 mg, and placebo were started on diuretics during the trial period.
Data on drug-induced liver injury and hepatotoxicity was not available in detail although the authors state that the liver function tests were not clinically significantly different compared with dapagliflozin plus placebo (apparently figure 5C which is hematocrit, not liver tests).
Discussion
The combination of flozins with endothelin antagonists is a new avenue being explored for nephroprotection. The primary endpoint was a change in UACR which was significantly improved in the zibotentan 1.5 mg arm as compared to dapagliflozin alone.
There was also a statistically significant improvement in the systolic and diastolic BP in the zibotentan arms compared to dapagliflozin arms. The change in UACR, however, did not correlate with the change in BP, suggesting that the renoprotective effects of zibotentan may not be dependent on the enhanced BP control alone.
The effect of zibotentan on eGFR was a secondary outcome. eGFR showed an acute drop in at week 1 in all 3 arms, the effect being numerically largest (although not statistically significant) in the zibotentan 1.5 mg arm. Considering that there was no placebo arm and it was a short study of 12 weeks only, the long term effects on eGFR and composite kidney outcomes is difficult to glean from this result but most nephrologists would not be worried about a miniscule acute drop in eGFR if long term beneficial effects on clinical kidney outcomes are proven. The authors also talk about improvements in LDL cholesterol and HbA1C in zibotentan arms versus dapagliflozin alone. In the absence of data on other cholesterol lowering agents and anti-diabetic drug dosage changes, we can not use this information to state that zibotentan has cardiovascular benefits over and above dapagliflozin.
“Every rose has its thorn” - this sentence was used in relation to avosentan in the ASCEND trial (Ritz et al, JASN 2021, Mann et al, JASN 2021). Zibotentan is no exception to this aphorism. Akin to all previous trials with ET-A antagonists in CKD, whether diabetic or non-diabetic, there was a tendency for fluid retention, more so in the zibotentan 1.5 mg arm than 0.25 mg arm. This was seen to peak at week 1 with an increase in body weight and extracellular water in zibotentan 1.5 mg. It was subsequently countered by the effect of dapagliflozin and diuretics, with the 12 week mean body weight being lower than the baseline due to a decrease in intracellular water percentage but still showing an increase in percentage of extracellular water and total body water. Six cases of heart failure were also reported in the zibotentan arms, two of which were serious adverse events. 12% patients in each of the zibotentan arms discontinued the drug owing to treatment emergent adverse effects compared to 4% in the dapagliflozin arm. The overall trend of body weight remained favorable in the zibotentan 0.25 mg arm along with an almost stable level of extracellular and intracellular water percentages. At discontinuation, the body weight remained stable in contrast to an increase in those with dapagliflozin.
The trial selected CKD patients with fewer comorbidities and absence of anemia or significant cardiac dysfunction. BNP was used as a criterion to screen and follow-up patients. In real world practice, CKD patients often have cardiac comorbidities leading to heart failure, anemia, and fluid retention are often troublesome complications (especially at lower eGFRs) and BNP monitoring is not practical to decide on drug continuation at every visit. Hence, we need to carefully select which patients are likely to tolerate this drug and benefit from it (also evident from Figure 1, where only 525 of the 1492 patients fit the stringent inclusion criteria). Despite this, more than 10% of participants had to discontinue in the short 12 weeks of follow-up for adverse events, suggesting this drug class remains a double-edged sword, whose sharpness for side effects is only dulled a bit by flozins.
Despite these limitations, the trial is a welcome sign as it includes a large heterogeneous population of proteinuric CKD patients and shows the efficacy of zibotentan in proteinuria reduction in contrast to the specific subgroups studied in PROTECT and DUPLEX studies. The antiproteinuric effect was significant irrespective of diabetes status, eGFR, and blood pressure control.
If the ZENITH high proteinuria phase III trial set to enroll 1500 patients randomized to different doses of zibotentan based on CKD stage shows a significant benefit of zibotentan on eGFR slope and clinical kidney outcomes, nephrologists will soon be faced with the happy dilemma of which drug to choose third line for nephroprotection (after RASi and SGLT2I) for which patient from amongst all the new drugs on the horizon (Mineralcorticoid antagonists/Aldosterone synthase inhibitors, GLP-1 RAs, DEARA, zibotentan) Till then, keep flozinating and hope for the best.
Conclusion
Zibotentan enhances the antiproteinuric effect of dapagliflozin in CKD. The effect on eGFR and long-term kidney outcomes is yet to be ascertained. Combining dapagliflozin to low doses of zibotentan may help in ameliorate - though not completely prevent - the fluid retention related adverse effects in patients with CKD.