JAMA . 2024 Mar 5;331(9):740-749. doi: 10.1001/jama.2024.0728.
RNA Interference With Zilebesiran for Mild to Moderate Hypertension: The KARDIA-1 Randomized Clinical Trial
George L Bakris, Manish Saxena, Anil Gupta, Fadi Chalhoub, Jongtae Lee, Daniel Stiglitz, Nune Makarova, Nitender Goyal, Weinong Guo, Dion Zappe, Akshay S Desai; KARDIA-1 Study Group
PMID: 38363577
This was but one of several studies, manuscripts, and projects in hypertension and diabetic kidney disease that the late George Bakris led or was involved in during his long career. You can read some of the obituaries here (Kidney Int; ASN website). He was also a guest on the podcast; listen to him here (Freely Filtered on the FIDELIO trial).
Why was this study done?
This was a phase 2 study, conducted by a drug company (Alnylam Inc., MA - riffing off the Alnilam in Orion’s belt), so this is merely a step towards regulatory approval, so profits may flow. So, rather, why are we talking about this study? One reason is that anything involving the RAAS is fair play for nephrology. And though ACEi, then ARBs, preceded by MRAs, then resurgence with nsMRA and ASIs has progressed over the last several decades, we haven’t discussed the origins of the RAASi. Renin inhibitors never took off - but could targeting angiotensinogen more efficiently cut off the RAAS (or at least the RAS) pathology in hypertension? The second reason is the cool technology involved. RNA interference got a Nobel for Fire and Mello in 2006. This is a natural process identified by these scientists, which is important for the regulation of gene expression, defense against viral infections, and keeping jumping genes under control. It uses double-stranded RNA (yes, you read that right) to disrupt mRNA, which is on its way to ribosomes for polypeptide/protein synthesis. Thus, even if the nucleus sends a signal to build a certain protein (eg, angiotensinogen), the cytoplasmic ribosomal factory does not get the message. Check out this cool video for a niftier explanation.
There is, of course, much more to this. Developing a targeted siRNA (small interfering RNA) that disrupts the mRNA of interest is one thing, but delivering it to the specific cells or organ where the peptide is being synthesized is no easy task. One also has to hope it does not disrupt the synthesis of some other gene-peptide pathway (or provide you with 5G signalling properties). After the initial enthusiasm, most other big Pharma has walked away, leaving Alnylam as one of the only big players. For zilebesiran in particular, an initial phase 1 study (Desai et al, NEJM 2023) demonstrated BP lowering with no fatal consequences (unlike with revusiran), which led to this phase 2 study. Zilebesiran itself is an siRNA which targets hepatic synthesis of angiotensinogen. It targets hepatocytes using a conjugate with GalNAc, or N-acetylgalactosamine, which is expressed on liver cells.
Image from Alnylam website
How was the trial done, and what did it show?
This was a double blind RCT, with 4 different doses of zilebesiran and one arm of placebo, with ambulatory BP being the outcome. The trial was powered for a 10 mm Hg BP difference at 3 months, thus including 394 participants. As an aside, 16 patients were enrolled in wartorn Ukraine and had to be excluded from the analysis as outcome data was not available. Being a pharma phase 2 trial, it was the opposite of a pragmatic trial, with many exclusions (including GFR < 30 and K > 5). The main eligibility was having hypertension and requiring up to 2 BP-lowering meds. These were then withdrawn in the 2-4 week washout period, and the subsequent ABPM daytime had to be between 135 and 160 for randomization. Thus, the placebo group got no BP medications at all for 3 months to allow assessment of BP lowering with zilebesiran. BP meds were allowed at physician discretion for months 4 and 5, but not the last month before the 6-month BP outcome ascertainment. This is somewhat similar to the double blind period for the aprocitentan PRECISION trial, where a limited period of untreated hypertension was permitted (Schlaich et al, Lancet 2022 | NephJC Summary).
eFigure 1. Study design, from Bakris et al, JAMA, 2024
Notable in Table 1 is ~ 45% women and 25% Black participants. GFR was pretty high with a mean of ~ 80 ml/min/1.73m2. The BP drop difference compared to placebo was an impressive ~ 11 to 14 mmHg, with no impressive dose response. Interestingly, the BP increased in the placebo by about 5 mm Hg (despite the 2-4 week washout), showing how long the effect of BP medications might potentially last.
eFigure 4. Hourly mean SBP at 3 months, from Bakris et al, JAMA, 2024
There was an impressive decline in serum angiotensinogen, which lasted for 6 months in all groups.
eFigure 3. Change in serum angiotensinogen, from Bakris et al, JAMA, 2024
As expected, adverse effects of hypotension, hyperkalemia, and ‘AKI’ were more common with zilebesiran. Overall, 17 patients had a serum K > 5.5, and 3 even received a K binding agent.
What are the implications?
Zilebesiran lowers to a similar extent as seen in the original captopril trial (VA Co-op study group, JAMA IM, 1984). It has the usual side effects one could expect in a RASi agent, with no worrying off-target effects noted in this sample of ~ 300 who received this (an open-label extension where everyone gets zilebesiran is ongoing).
While we await a phase 3 trial, which will lead to regulatory approval, one has to ponder on the utility of an agent when we already have potent, safe, and effective agents targeting ACE, AT receptors, and downstream aldosterone pathways. One purported advantage is that of adherence. A once-in-six-months injection (which one could pair with a PCSK9i for BP + cholesterol lowering) may sound especially attractive, especially to those who do not like the pill burden that builds up in chronic diseases. But shouldn’t we be worried about safety? After all, one can stop or hold the RAASi in acute illnesses (even if the evidence for sick day rules does not exist) or with volume contraction episodes causing severe hypotension (no trials needed here). But one cannot take out the siRNA, which is deep within hepatocytes, disrupting angiotensinogen synthesis. It seems these episodes are saline responsive - zilebesiran does not cause refractory shock. But this safety aspect of irreversible (at least for 6 months) RASi needs to be properly confirmed in longer studies, alongside the phase 3 trial that is ongoing.