The whole event shocked me. I had no idea that the results of the Bardoxolone study had slipped beyond the geek fringes of nephrology. It reminded me of a story that Judah Folkman told. He came to Indiana University to collect an award and give a lecture, shortly after a NYTimes profile. In that front page story James Watson (yes that James Watson
Judah told the story that he was getting phone calls from strangers and friends asking for his miracle cure and was heart broken because he had nothing to offer them. At that stage his drug was only for mice.
![]() |
| That's Judah and me following the afore mentioned lecture in 1999. |
As I started my research I became concerned that patients randomized to bardoxolone developed increased albumniuria.
Some patients tripled their albuminuria! The drug increased GFR, but the increased albuminuria could not be fully accounted for by the improved function.
ASN Kidney Week fell 10 days prior to my Grand Rounds so I planned on grabbing some good ideas at the meeting. On Friday I went to Kidney Disease in Type 2 Diabetes: New Insights. There were four lectures. The last two were homeruns.
Dr. Bruce Perkins was perfect for my talk. He spoke of how albuminuria is not a great surrogate end-point for diabteic nephropathy studies. Bad outcomes often follow a reduction in proteinuria.
I used my iPhone to record the audio and took pictures of each slide with my Nikon (this was before I learned that ASN did not want attendees taking pics of the lectures. WTF). When I got home I grabbed the best thoughts from his lecture and made it the cornerstone of my talk on diabetic nephropathy, bardoxolone, and a more modern view of albuminuria.
Thanks Dr. Perkins.
The lecture was a little light, I finished in 45 minutes and used some filler from my Diabetic Nephropathy 2009 lecture. Before I use the lecture again I would add some of the points from Andrew Bomback's excellent lecture, "RAAS Blockade: More is better? Yes. No. It depends."
Here is my PowerPoint (58mb) and PDF (51mb).
Note to self: the Helvetica Neue UltraLight, didn't project so well.







I was initially surprised looking through your powerpoint slides to see the dramatic lessening of effect of Bardoxolone on eGFR between slide # 11 and slide # 12. This wasn't what I remembered from reading the study in the NEJM. When I went back and re-read the NEJM article I realized that this one-two punch of slides is pretty misleading, since you don't include the data at 52 weeks, which was much more robust. What you do show is the 56 week data, taken 4 weeks after withdrawal of the study medication. Is choosing to present it this way something you explain more fully in your lecture?
ReplyDeleteI would love to hear the entire lecture, since the slides don't seem explanatory enough on their own. Any chance that there is an audio mp3 of that floating around?
Yours,
Robert Leversee MD LMHCA
There is a video and audio capture. I will try to post it to the web.
ReplyDeleteThe point of the slide was to show there was some durable effect of the drug, that this was not merely some hemodynamic slight of hand. Given the modest durable affects I'm not sure how compelling the argument is.
Thanks for clearing that up... but I'm still curious - does the effect of an intervention need to hold up even after removal of the agent for the effect to be real? I'm thinking of ACE inhibitors - I realize that after removal of ACE or ARB the GFR temporarily goes up, but over time without ACE the GFR would continue to decline at an accelrated rate. I didn't think the effect of ACE was a permanent change to the kidney physiology, but rather that continuous treatment with ACE over time seems to lower intraglomerular BP and this results in slower decline of GFR. Couldn't the effect of Bardoxolone be similar, i.e. while the drug is present the kidney operates somehow better, and in the long run this preserves kidney function, but as soon as the drug is withdrawn the effect begins to lessen? What would be wrong with that model? Would appreciate your thoughts.
ReplyDeleteThanks in Advance
Your point is well taken. We don't require changes to be durable to be considered valid, however the end-point used in BEAM 2 was not a standard "hard-end point." They did not show a decrease in dialysis, hospitalizations or cardiovascular end-points. They only showed an increase in GFR. I wanted to point out that if that change in GFR was not durable it could be due to creatinine slight of hand, perhaps bardoxolone increased proximal tubule creatinine secretion or produced a hemodynamic boost in GFR as was seen in AASK with amlodipine. See here
ReplyDeleteRobert, I just posted on this. See if that helps.
ReplyDeleteThe changes absolutely should be durable. To be "renoprotective," you want to prevent renal fibrosis, right? Renal fibrosis won't reverse when you pull off the drug for a couple of weeks; hemodynamic changes will. Therefore, off-drug GFR (Cr) measurements at the end of the study provide the best non-invasive biological assessment of renoprotection.
ReplyDeleteOf course, the time horizon of the study must be sufficient to expect to see differences in renal parenchyma between groups (...and for those differences to translate into the surrogate outcome of serum creatinine....).