Burzynski: The CDA-2 Saga

By Keir Liddle

_63224405_prostatecancersplThere have been some interesting developments on the #Burzynski hashtag over on twitter related to a chemical solution known as CDA-2. Initially I dismissed much of the tweets from @PDiddymus as simple obfuscation of the issues, what-aboutery and the moving of goalposts over shifting sands. But than a strange thing happened unlike most of the Burzynski supporters who prefer to shout the same criticisms of skeptics and retread the same anecdotal evidence over and over again @PDiddymus adapts and changes their arguments. Developing them in response to skeptical counter arguments.

That sort of attempt at engagement, regardless of disagreement on motive, deserves the respect of having their claims explored in more detail than the 140 characters of twitter allows. Though regrettably the same cannot be said of their blog which is simply reference spam without critical analysis in support of or against Burzynski and the efficacy of ANPs. Annoyingly each of the links offered on the blog is not to a new reference but most often several links all take you to the same publication. Whether this is for completeness or a concious attempt to overstate the case I don’t know.

@PDiddymus central argument is that Burzynski critics and Skeptics have not done their research and are ignoring Asian research on the chemical solution known as CDA-2. A chemical solution that contains one of Burzynskis antineoplastons: AS2-5.

Antineoplasteon AS2-5 is not one of the antineoplasteons that Burzynski uses as a treatment for cancer so it’s mention seems a little dubious and disingenuous and this could be considered an attempt to muddy the waters. It is also only one component, albeit the main component, in the CDA-2 solution contains multiple active components, including phenylacetylglutamine (PG) (41%), benzoyl glycocoll (35%), peptides (MW 400–2800) (17%), 4-OH-phenylacetic acid (6%), and 5-OH-indoleacetic acid (1%).

As for Antineoplastons the active compound in A2 was found to be 3-phenylacetylamino-2,6-piperinedione, which was named antineoplaston A10 and from antineoplaston A10, three other compounds have been derived:

  • AS2-5, which is phenylacetylglutamine (PAG).
  • AS2-1, which is a 4:1 mixture of phenylacetic acid (PA) and PAG.
  • A5, which is PA and an aromatic fatty acid.

Thus one of the active components in CDA-2 is present in antineoplaston AS2-1. But it is important to note that CDA-2 is a chemical solution containing more than simply AS2-1 alone.

A compound which this paper (published in PLOS ONE) reporting animal research on mice suggests that it may have some promise in inhibiting lung tumour development. There are a number of other papers reporting results from studies of CDA-2 here are the examples quoted by @PDiddymus on their blog:

You may wonder why I have emphasised the origin of all these scientific papers. This is simply to show that research on CDA-2 seems to be solely focused with China and Taiwan (again I may have mismarked a Taiwainese study as Chinese when copying the links from @PDiddymus blog if this is the case please let me know and I will add a correction). This promising new cancer treatment does not seem to be an international research concern. Apologies if some are repeated but as stated above @PDiddymus has the annoying habit of linking to the same paper several times. Likewise apologies to @PDiddymus if I have missed any feel free to point out any abscences in the comments.

Similarly to my analysis of the Japanese research I have decided to focus on the human studies into CDA-2 of which @PDiddymus appears to have linked to one but I am willing to edit this post if they produce more.

Taking the paper at face value, ignoring for the minute that it has published dubious research in the past, it explores the effect of CDA-2 (labelled as a selective inhibitor of abnormal methylation enzymes in cancer cells,on the therapeutic efficacy of cytotoxic chemotherapy. So the first thing to note is that CDA-2 is not being tested as a chemotherapy agent on it’s own.

It is important to note this as Burzynski is only supposed to be using antineoplasteons in clinical trials where they are given alone without any other chemotherapy agents being administered. So if Burzynski has been following his own protocols in his “trials” then any evidence of CDA-2 working in conjunction with conventional chemotherapy agents is irrelevant. As evidence of a chemical solution which contains one ANP assisting the action of chemotherapy agents is not in any way evidence of ANP efficacy on their own.

If anything it suggests that Burzynski’s research and focus has been flawed for quite some time and any potential benefit suggested by the CDA-2 trials incompetently missed because of his belief in the efficacy of ANP alone.

Oddly, rather than focusing on one form of cancer, the study claims to be investigating the action of CDA-2 and conventional chemo on three cancers: 80 NSCLC (Non-small cell lung cancer) patients,188 primary hepatoma patients,and 186 breast cancer patients. Given the nature of cancer this seems a little odd on the face of it but we will go with it for now. There is also an odd turn of phrase in the paper that jumps out at me it claims that a “perfect” solution was used but again this could be a translation issue or it could be a misunderstanding on my part of another disciplines jargon.

83% of patients completed their course of treatment in the trial. In the reporting of the results there is an odd bit of imprecision where the authors report that about half of the participating patients were stage III patients and the other half stage IV. This personally strikes me as odd and I would welcome the opinion of those involved in cancer research as to whether this is an acceptable or standard way of reporting such things.

Most of the patients had previously been treated with chemotherapy, which makes it hard to disentangle the specific effects of CDA-2 from confounding variables such as the delayed effects of prior chemo. Another thing I note from the results section is that a total of 261 patients enrolled in the trial were recurrent patients who had been in remission for certain periods. This means that of the original 454 patients enrolled in the trial OVER HALF (57%) were in remission during the trial. Which surely must have some impact on the randomisation process making the study results questionable at best? As one of the trials arms would have a greater proportion of patients in remission than the other which has the potential to seriously bias the results of this study as far as I can see.

Indeed looking at table 2 we can see that there are 174 recurrent patients in the treatment arm and 87 in the control. Twice as many patients in remission were “randomized” to the treatment condition as opposed to the control. That looks a mite suspicious to me.

CORRECTION: someone has pointed out in this comments that this is not too much of a concern. Though they have also noted that the study does not report survival rates which is far odder.

Given the nature of the assignment I have little confidence in the conclusions arrived at. If anyone with more experience of oncological research would like to look over this paper and comment on this please feel free. But from my knowledge of trial design this study seems to be seriously flawed.

Perhaps we should consider the lessons learnt regarding the quality of publications that sometimes slip into the Chinese Journal of Clinical Oncology more than we should take on face value that CDA-2 is an effective addition to the cancer fighting arsenal.

I welcome corrections and expert commentary on the above article as it has taken me a little outside my comfort zone.

But regardless of the validity of these articles, and CDA-2 research in general, they don’t lend Burzynski the crucial support they are presented as.

Firstly CDA-2 has been presented alongside conventional Chemo. Burzynski doesn’t use ANPs and Chemo in combination.

Secondly they don’t change the fact Burzynski hasn’t published any trials results properly in a peer reviewed journal.

If ANP has s0me other role in cancer treatment other than as a direct chemotherapy agent than, outside of China, Burzynskis obstinant refusal to play the same game as every one else involved in medical research may have set patient treatment back years.

The Skeptics dichotomy remains intact: If ANP don’t work Burzynski has knowingly or unknowingly taken advantage of thousands of vulnerable and desperate patients but if they do work he has denied his treatment to thousand more.

Either way Burzynski is a medical scandal.

This entry was posted in Featured, headline, opinion, research, Scepticism, Science and tagged , , , , , , , , , . Bookmark the permalink.

6 Responses to Burzynski: The CDA-2 Saga

  1. Adam Jacobs says:

    Fascinating post, Keir. I am impressed with the hard work you’ve done looking into this.

    Unfortunately, your link to the paper of the “phase III trial” is broken, so I can’t check through your observations in detail. However, I did see a supposed “phase III trial” the other day, and it very much sounds like it’s the same one, so I can offer one or two thoughts.

    My main thought is that, despite the paper’s title proclaiming a phase III trial, in reality it looks far more like a number of phase II trials (each of which was done in a different cancer type) cobbled together. Perhaps the authors hope that by reporting the trials together they’ll have enough patients in total to make it look like a phase III trial.

    The reason why I say it’s not really a phase III trial is that it used treatment response as its primary outcome measure. It does not report survival. A valid phase III trial in cancer really needs to report on either progression-free survival or overall survival. This paper doesn’t. The use of response rates as an outcome measure is pretty standard in phase II.

    It sounds like you’ve already identified a number of reasons to be skeptical of this paper. But even if the results could be taken at face value, then the complete absence of any data on whether CDA-2 improves survival makes the data of very limited use in drawing any conclusions about efficacy.

  2. Adam says:

    I’ve just had another look at this, and happy to report that your link to the paper is working now. Not sure whether you fixed it or whether their site was just down temporarily when I looked earlier, but no matter.

    So yes, this is the study I was thinking of, and to me the biggest flaw here is the complete lack of any survival data.

    There is one point, Keir, where I think your concerns are misplaced. You mention the imbalance in recurrent patients in table 2. This actually isn’t nearly as big a problem as you think: the trial had a 2:1 randomisation scheme, so having twice as many recurrent patients in the CDA-2 group as in the control group is exactly what we would expect. It doesn’t suggest any kind of problem.

    Something that does raise a bit of a red flag, however, is that quite a few patients were excluded from the analysis as being “ineligible”. Some of those patients were ineligible because their disease had progressed or they had died. That’s a really dreadful reason for excluding someone from an efficacy analysis, especially as more patients were excluded for progression or death from the CDA-2 group than from the control group. That introduces an important bias into the results. If you exclude some of the worse cases from your efficacy analysis, you’re going to make the treatment look better than it really is.

    But perhaps the most shocking thing about the whole paper is table 8. That table tells us the actual names of the patients. I was utterly gobsmacked when I saw that. No reputable journal would even consider printing something that breaches patient confidentiality in such a flagrant manner.

    If the investigators have such a catastrophic failure to grasp one of the most fundamental principles of medical ethics, then to me that puts a pretty big cloud over their entire claim to any kind of credibility.

  3. Acleron says:

    Adam, could you expand on the randomisation aspect? Especially on how whether the proportion of recurrent patients is deliberate or not, how this doesn’t abrogate the conclusions, thank you.

  4. Adam says:

    Sure, no problem. Sorry if my last post wasn’t clear.

    In a randomised controlled trial, it’s normal to randomise patients in a 1:1 ratio to the active and control groups. So if you have 100 patients, you’d expect about 50 in the active group and 50 in the control group.

    However, although a 1:1 randomisation ratio is the most commonly used ratio, it’s not the only possible way of randomising. Sometimes, patients may be randomised in unequal ratios. One fairly common choice is a 2:1 ratio. This means that any given patient is twice as likely to be randomised to the active group than to the control group. So now, if you were running a trial with 300 patients, you’d expect to have 200 patients in the active group and 100 in the control group.

    There are various advantages and disadvantages of equal vs unequal randomisation, but that’s probably a discussion for another day.

    In the trial we’re talking about here, they used a 2:1 randomisation scheme (it’s a little unusual, but there’s not necessarily anything wrong with it).

    So if you look at the total number of patients enrolled (Table 1), you see that there are 301 CDA-2 patients and 153 control patients. That’s near enough to a 2:1 ratio that it makes no difference, and is consistent with a perfectly valid randomisation process (you often don’t get exactly the nominal ratio, because there is an element of randomness).

    If we then look at the number of recurrent patients, we find 174 in the CDA-2 group and 87 in the control group. That’s exactly a 2:1 ratio, which is what we’d expect, given that the randomisation was done in a 2:1 ratio. To put it another way, we have 174/301 recurrent patients in the CDA-2 group, or 58%, and 87/153 patients in the control group, or 57%.

    So although table 2 may look at first glance to show some imbalance, that’s just a function of the unequal randomisation. When we have 58% recurrent patients in the CDA-2 group and 57% recurrent patients in the control group, it would be a bit harsh to accuse them of significant imbalance.

    Hope that makes sense!

  5. Les Rose (@Majikthyse) says:

    I was about to submit a comment about randomisation – just as well I read Adam’s excellent explanation first! I could not find anything in the paper to explain in adequate detail how response was assessed, other than to refer to WHO criteria. The study (or studies as Adam points out) was open label, and there is no indication that response was assessed by a blinded third party. Hence it was clearly open to observer bias.

    In my 40-year career I have never before seen patients’ names in a published paper. This can’t be considered a scholarly journal.

  6. I believe a follow up post with corrections is required.

    Also have more info to divulge.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s