Seeking Alpha

Komodo Global

 
View as an RSS Feed
View Komodo Global's Comments BY TICKER:
Latest  |  Highest rated
  • Affymax Sell-Off: Did Investors Overreact? [View article]
    Thanks. How frequently would they require a transfusion without any ESA use?
    Jul 14, 2010. 12:31 PM | Likes Like |Link to Comment
  • Affymax Sell-Off: Did Investors Overreact? [View article]
    > ngrunes, I think you should have written this article :)
    Time is a little tight for me now, but I intend to give your questions a thorough response.

    You're very kind, and I look forward to it. 10a-d are the specific outstanding questions for you.
    Jul 14, 2010. 10:24 AM | Likes Like |Link to Comment
  • Affymax Sell-Off: Did Investors Overreact? [View article]
    Thanks biotecher --

    The Ofsthun et al. study doesn't appear to indicate anything regarding ESAs. Apparently all patients were treated with ESAs. "Mean albumin was higher in the upper hemoglobin categories, as was transferrin saturation." Any number of health correlates could be associated with hemoglobin levels. This was a retrospective view of ESRD [end-stage renal disease] patients that didn't test treatment with ESA relative to placebo. ESRD patients with lower hemoglobin levels may simply have been sicker. I fail to see any conclusion that treatment with ESAs to attain that 11-13 blood level follows from the study, as all patients appear to have been in the midst of ESA treatment. If you look at Table 3 and Figure 1 on page 1911 of that study, comparing time to death and mean rHuEPO treatment, isn't one unfalsified hypothesis that simply higher ESA administration is associated with faster death?

    "The other question is - is it really hemoglobin that we should be concerned about?"

    Exactly. All recent studies of ESAs seem to suppose that ESAs impact hemoglobin levels and nothing else. Of course then some dosing of ESAs will be found preferable. And that's a ridiculous supposition. It's obviously a ridiculous supposition when ESA use to target 12-14 g/dL has been associated with higher deaths than ESA use to target 9-12 d/gL while normal levels are above 12, right? Isn't that ipso facto evidence that ESAs are potentially lethal for a reason aside from the hemoglobin level they engender?

    "The availability of the first ESA, Epoetin, was hailed as a major medical breakthrough in 1989 when, for the first time in history, anemia could be treated without blood transfusion. Dialysis patients used to have hemoglobin levels below 9 unless they received frequent transfusions; they were feeling weak, sick, and were dying quicker. ESAs improved their quality of life significantly without the need for a blood transfusion."

    The FDA suggested in that testimony, "Patients treated with Aranesp had a higher death rate and no reduction in the need for transfusions compared to those treated with placebo."

    The simple question -- regardless of what is generally believed as to ESA action, based on what large-scale studies do ESAs reduce blood transfusions significantly and prolong life materially relative to placebo? Surely when Epoetin was first approved, some such large-scale study was conducted?
    Jul 13, 2010. 11:20 PM | Likes Like |Link to Comment
  • Vivus, Arena and Orexigen Investors All Worried About Qnexa FDA Panel Review [View article]
    Materially worse than Contrave (as updated from initially overstated efficacy data) not based on endpoints in isolation but (given variable trial practices and only considering efficacy) relative to placebo?
    Jul 13, 2010. 10:26 AM | Likes Like |Link to Comment
  • Vivus, Arena and Orexigen Investors All Worried About Qnexa FDA Panel Review [View article]
    Excellent article up to the cursory discussion of Contrave and Lorcaserin at the end. And when the author has no position (or discloses his position) in companies mentioned and receives no compensation for the post, he shouldn't be criticized for offering an opinion. Here's my opinion, and I have no position in VVUS:

    Cognitive side effects of Topiramate render it grossly unfit for a weight loss drug. Contrave and Lorcaserin might be decent alternatives.

    Now my question:

    Cardiac issues appear precisely what Lorcaserin patients *don't* need to worry about. I'd enjoy a comprehensive comparison of the side effects of Contrave and Lorcaserin that focuses not at all on questions raised or trading sentiments, only on the empirical results of clinical studies.

    Well, I failed to state that in the form of a question :)
    Jul 13, 2010. 01:31 AM | Likes Like |Link to Comment
  • Affymax Sell-Off: Did Investors Overreact? [View article]
    Thanks Joseph and biotecher –

    "This is with regard to ngrunes' key question: "Would FDA approve it (Hematide)?"

    My key question is whether the FDA *should* approve it. By "should" I don't mean whether we expect the FDA to approve it, or any probability associated with that expectation; I mean what we think should happen for patients' benefit. I'll break down that question here.

    "Your primary argument against Hematide approval is not the fundamentals of safety or efficacy; you seem less concerned with those parameters; your concern is how Hematide will appear in light of the FDA sentiment about ESAs changing."

    My concern is precisely with safety, efficacy and value. I see I didn't well communicate what I mean by worst(a,b). Let me clarify that. Continuing the numbering for convenient reference, I'll briefly address (4) valuing biotechs, (5) worst(a,b), (6) my key question, (7) Affymax's trial results, (8) other ESA trial results, (9) what I suggest the FDA do, and (10) the questions for you.

    4. Binary outcomes are no impediment to valuation if we reasonably handicap the outcomes. For example, and these numbers are wholly made up just for example, suppose we reasonably conclude based on global averages and clinical trial results that some drug has around a 60% probability of approval and will generate 30% pretax cash earnings on annual sales of $250-500 million for eight years. That's a straightforward expected value. Of course, what management will do with that probabilistic cash is a huge valuation input, as is development cost. Still, if we can estimate (i) probability of approval and (ii) resulting sales and margins even within a large range, the binary outcome per se presents no impediment to fundamental valuation.

    5. From your responses I can see I explained worst(a,b) poorly. By worst(a,b) I mean the worst of (a) or (b) (written like it's a spreadsheet formula). In using worst(a,b) for investing, I mean to use the worst of (a) or (b) where:

    (a) = what we expect to happen, period. That expectation may incorporate probabilities, such as binary outcomes. But it is simply what we expect as dispassionate observers of facts. This would include our estimate of the probability of any FDA action.

    (b) = what we would like to see happen if we're both financially disinterested in the matter and benevolent leaders who control the world. Really.

    Using worst(a,b) means we use (a) when we expect a company to do less well than we think it ought to do in an enlightened world (for example, due to corruption or dumb rules impeding it), and, and perhaps more often, we use (b) when we think a company isn't really helping people or regulators are foolishly asleep.

    Critical point #1 here: (a) is absolutely irrelevant to (b). B is what we think should happen in an ideal world. That's totally separate from what we think will happen.

    Critical point #2 here: defining (b) can be very helpful notwithstanding it entails setting aside reality. It can avert incentives on our part to impede good policy. It can also avert adverse financial outcomes insofar as regulators *or* customers wake up. It's a pleasant way to invest, to exclusively own companies that we think deserve to do well (that's the b), beyond our analysis that we think they will do well probabilistically (that's the a). To invest on the basis of (a) means we think a company will do well. To invest on the basis of (b) means we think a company deserves to do well.

    Critical point #3 here: if we use worst(a,b) consistently, then we need to specify (a) and (b) as defined above per investment case.

    6. And (b) was my key question: what should happen with Hematide for patients' benefit?

    7. We can tackle that question by reviewing salient facts from the trial results that Joseph well summarized and a few other significant trials that caught the FDA's attention. First, the Hematide trials (www.investors.affymax....):

    7a. Hematide bulls may say that Hematide was found bioequivalent and non-inferior to Epoetin in the Emerald 1 & 2 trials of dialysis patients.

    7b. Hematide apparently was found similarly efficacious to Darbepoetin, but caused more Serious Adverse Events (SAEs) and particularly adverse cardiovascular events than Darbepoetin, in the Pearl 1 & 2 trials of non-dialysis patients.

    7c. Given just those statements, one might conclude that Hematide should be approved for dialysis patients and not, at least not yet, approved for non-dialysis patients. I do not conclude that yet, because I see, just at a glance, three problems with the trial results:

    7d1. The difference between the structure of the Emerald and Pearl trials (they give clinical trials of potentially lethal drugs such wonderful names, don't they?) was not just dialysis patients vs. non-dialysis patients. Emerald trials of dialysis patients exclusively compared Hematide with Epoetin, whereas Pearl trials of non-dialysis patients exclusively compared Hematide with Darbepoetin. One possibility is that Hematide is non-inferior to any ESA for dialysis patients. But another possibility is that Hematide is inferior to Darbepoetin for patients whether dialysis or non-dialysis. That possibility can't be ruled out from the trial results, as Hematide was found inferior to Darbepoetin in all trials in which they were compared (Pearl 1 & 2). And Darbepoetin is used for patients both on dialysis and not (source: aranesp.com). I don't mean to do Amgen (no position, ever) any favors here. That's just a fact that Hematide was not found non-inferior to Darbepoetin for any indication.

    7d2. It wasn't just the headline cardiovascular events that made Hematide inferior to Darbepoetin; Hematide patients also received more red blood cell transfusions (10-11%) than Darbepoetin patients (5%) and had more overall serious adverse events. Hematide wasn't shown to be non-inferior to Darbepoetin for any indication in any endpoint. And Hematide wasn't shown to be efficacious for any indication that Darbepoetin isn't already used to treat.

    7d3. I gather the sole purpose of these drugs is to avoid red blood cell transfusions, and yet around 10% of Hematide patients across the trials received RBC transfusions. Roughly what percentage of patients would have received transfusions without Hematide? The trials absurdly don't answer that question, because in no trial was there a control group of no-ESA patients. In other words, the trials don't indicate whether these drugs serve any purpose whatsoever.

    8. That could be alright if Epoetin or Darbepoetin was already demonstrated to be convincingly helpful. On first glance, they haven't been. Widespread use does not constitute any such demonstration. Testimony in the House addressing ESAs (fda.gov/NewsEvents/Tes...) just over three years ago noted:

    "In February 2007, FDA notified health care professionals of the results from a large clinical trial evaluating the use of an ESA to treat anemia in cancer patients not receiving chemotherapy. In this trial, patients received either Aranesp [Darbepoetin] according to the approved dosing regimen or a placebo. Patients treated with Aranesp had a higher death rate and no reduction in the need for transfusions compared to those treated with placebo."

    Well, isn't the sole purpose of ESAs to reduce transfusions? Regardless of whatever is believed as to their action, if they empirically do not, then why are they approved at all?

    Now, perhaps that's an erroneous summary of the trial and "placebo" actually means treatment with a lesser amount of ESAs. The significant trials to which I referred in the prior post compared only one dosing of ESAs with another dosing of ESAs, never with no-ESA-at-all.

    And that's been the damn ridiculous feature of every significant ESA trial that I saw at a glance. Not one of these trials had a control group of no-ESA-treated patients.

    9. Now if I ran the FDA, I would be mildly annoyed at pissants like me questioning our diligence, and yet feel like like an utter fool, and apologize profusely to the public, for tolerating these trials of multi-billion-dollar drugs that compare only one dose of the drug with another dose of the drug, with no control group ever. I would be inclined to invite the public to spank me. Then I would fix it.

    Hematide, Epoetin and Darbepoetin should be tested on dialysis and non-dialysis patients with an absolute control group that does not use any ESA before or during the trial. That no-ESA-treatment control group should be a significant size.

    Before we have any such trial, we can't know empirically whether these drugs are worth a damn. Practitioners who prescribe these drugs might guess that they reduce need for blood transfusions significantly and that risks attending them are no worse than risks attending transfusions, even less. But we seem to have no data, from any trial on a significant scale, that indicates that.

    We do know that higher doses of ESAs are associated with higher deaths. Consistently.

    In not requiring control groups of no-ESA-treated patients alongside ESA patients in large scale trials, it appears the FDA has been derelict. And for a multi-billion-dollar class of drugs, it would appear inexcusable to remain derelict.

    10. My questions for you remain these few:

    10a. What are your guesses, with explanation of the rationale of your guesses, of (i)-(vi) at 2e above that break down expected Affymax cash at Hematide PDUFA date before capital raises?

    10b. What are the weblinks to large-scale trials that demonstrate non-inferiority of any ESA -- Epoetin (Epogen/Procrit), Darbepoetin (Aranesp) or Hematide -- to treatment with no-ESA-at-all on the endpoint of serious adverse events?

    10c. If you find no such trial, and if your sources find no such trial, are you compelled to agree that the FDA has been derelict in overseeing this class of drugs?

    10d. Let's focus on the (b) in worst(a,b). Please review 7d1-7d3. After you reflect on those facts in 7d1-7d3 above, what do you think should happen with Hematide for patients' benefit: For patients' benefit, should it be approved for dialysis patients without requirement of further trials? Or for dialysis patients' benefit, should it be required to be tested for non-inferiority vs. Darbepoetin and a no-ESA control group in dialysis patients before approval, as Hematide has not been found non-inferior to either?
    Jul 13, 2010. 01:16 AM | Likes Like |Link to Comment
  • Affymax Sell-Off: Did Investors Overreact? [View article]
    Thanks Joseph --

    To reply to your keen comments above:

    "Let me summarize your points as I understand them: "The market's response is to the safety concerns from the trial data is appropriate because...."

    Your first paragraph there is an excellent summary of one line of reasoning, and it's not quite mine. Let me summarize (1) what I'm looking for in terms of analytical endpoints, (2) AFFY net cash value at a glance and (3) the key question.

    1a. I'm not interested in the market response. I view Mr. Market like Ben Graham -- as a manic-depressive fellow -- though less charitably: as an intoxicated manic-depressive fellow.

    1b. Wild price swings like this can reasonably pique interest that Mr. Market is acting especially nutty, but that's it: the question then, whatever the idea source, is investment merit.

    1c. Investment merit boils down to three endpoints of analysis: risk bucket, sensible expected value estimate (V), and price (seekingalpha.com/symbol/p). Risk bucket guides portfolio allocation. Price divided by sensible expected value estimate (P/V) determines investment merit.

    1d. I mean by V: Expected Value to one who owns the investment for keeps. That must be the definition of value if the investor is to own securities without dependence on Mr. Market for satisfaction. Then the investor can trade with Mr. Market opportunistically and without concern of his mood swings. (Long/short and leveraged market participants, if their lenders require frequent measurement of marketable prices, have no such luxury.)

    1e. "Price is what you pay, value is what you get." Price is readily ascertainable. Value is what we analyze. Price is irrelevant to an investment case until we have an expected value estimate.

    1f. When we have an expected value estimate, we won't think of the stock in terms of $6 or $23. Simply, we'll view Price divided by Value. We know the Price is around $150 million.

    1g. Any price action -- even a 99.9% drop -- cannot inform our Expected Value estimate per 1d above. An investment case boils down to risk bucket, expected value, and price. Expected Value if held for keeps is exogenous of price. Expected Value if held for keeps is the primary endpoint of investment analysis.

    1h. I have a further requirement that guides my questions, with which most investors might disagree: by Value I mean the worst of (a) the reasonably expected and (b) our provisional view of what would transpire with enlightened behavior. Others constrict analysis to (a) alone. By using worst(a,b), we accomplish two things. First, we avert incentives to impede what we consider beneficent public policy. I think that's sensible to do, because investors influence much of the world and I want to shape a world for my family that I like. Second, we avert many adverse outcomes. For example, if, a few years ago, we knew that one oil major had a persistently terrible safety record relative to the others, and concluded (a) it'll likely generate $10-15b free cash for several decades and yet (b) in an enlightened world would be fined out of existence, then we would have used worst(a,b)=a for bearish position evaluation and worst(a,b)=b for ownership evaluation. Therefore we wouldn't have owned it and would have avoided cost of a big spill. Similarly, the first question in evaluating Phase III biotechs may be (a) is the FDA likely to approve its candidate and (b) in an enlightened world, would the FDA approve it -- should, for patients' benefit, it be approved and commercialized based on its safety, efficacy and cost relative to best alternatives. Evaluating (a) isn't enough. (b) is my question. Any sensible answer to that question must be based not on general practices or marketplace sentiment but on verifiable facts.

    2a. Cash position: at the last reported quarter (Mar10), AFFY had $168m cash+investments+recei... and $128m liabilities of which $58m remained deferred revenue. Typically, you know, deferred revenue is associated with future costs: it can't be held in liabilities simply for tax-timing or income-smoothing purposes. In line with that general verity, you see deferred revenue recognition reduced AFFY cash $14m in the Mar10 quarter beyond P&L net loss. Indeed however the cash appears unrestricted, and so we can evaluate the company's future burn and cash balance separately. $168-128+58 = $98m "net cash" (unrestricted cash, government securities, certificates of deposit, apparently insured auction rate securities, and receivables net of all liabilities payable) at Mar10. Let's also give the company credit for $13m prepaid expenses and deferred tax assets given accrued expenses in excess of that, and call it $111m "net cash".

    2b. Investments are both pictures and movies, which is to say: that cash balance is one frame in an ongoing film, which is to ask: what will it be in the foreseeable future.

    2c. Affymax received $30m in Jun10 for Phase 3 trials database lock, and, based on trailing quarters, appears likely to have spent near $41m in the June quarter. At June 30, 2010, then: subtract $11m leaving $100m "net cash".

    2d. Affymax will receive another $20m if&when FDA accepts Hematide NDA, and we can guess spend $20m/quarter before additional clinical trials or commercialization costs. (Quarterly G&A of $9m may be unlikely to decline amidst NDA activity, while quarterly R&D of $33m of which $13m were outside clinical trial expenses may unlikely be cut materially below $10m. I would direct the company to spend a fraction of that, but my preference is irrelevant to typical biotech operations. In evaluating expenses for long purposes, worst(a,b)=a: not what we would prefer, only what we expect.) Say the NDA is at least accepted for consideration. Unless you expect it to be decided before June 2011 -- and it hasn't even been filed yet -- then fast-forwarding to June 30, 2011 we can envision +$20m -$20m*4 = -$60m cash movement, taking around $100m "net cash" at Jun10 to $40m *before* any additional clinical trial costs. That's just under $1.7/share. And if (as) that estimate is wrong, we might expect burn to more likely than not exceed that absent credible guidance to the contrary.

    2e. If you have a higher view of its probable net cash at PDUFA decision date excluding any future capital raise, how do you arrive at it? Let it start with ~ $100m "net cash" at Jun10; add $20m upon NDA admission; and suggest NDA submission at D1 date, PDUFA at D2 date, and average quarterly cash expenses of $Q until then (in estimating $Q, note the company forecasted quarterly G&A to rise from $9.4m for NDA and commercialization activities):

    (i) Your estimate of D1 date:
    (ii) Your estimate of D2 date:
    (iii) Your estimate of $Q:
    (iv) Number of quarters from June 30, 2010 to D2 date:
    (v) iii * iv = foreseeable estimated burn
    (vi) $100m + $20m - (v) = AFFY "net cash" estimate at Hematide PDUFA date before future capital raises or clinical trials.

    A snapshot of the current cash position wouldn't answer this question as to Affymax's net cash in the foreseeable future. To decline to view the company as an unfolding movie would be to fight reality. If you take a different view than my $1-2/share net cash estimate of (vi), what are your quick estimates of (i) - (vi)?

    3a. The key question is (b) from 1h above: in an enlightened world, would the FDA approve it.

    3b. You're very kind in imputing to me any knowledge, but I didn't even know what anemia is until I read your post and associated material yesterday. My only thoughts on this key question derive from reading the company's recent SEC filings until that risk factor noting FDA review of the class of drugs, then googling to the few links noted above. I'm about the last person in the world you would want to query for medical knowledge. What I can do is evaluate elementary reasoning without regard to prevailing sentiment.

    3c. The elementary reasoning surrounding this class of drugs ("ESAs" for erythropoiesis-stimula... agents, wherein the E word means producing red blood cells) appeared astounding.

    That FDA backgrounder from the FDA's 2007 reconsideration of ESAs (<www.fda.gov/ohrms/dock...;;) included this helpful summary:

    "Erythropoiesis-stimul... agents (ESAs) were originally developed to replace the deficiency of erythropoiein that frequently develops in patients with CRF [chronic renal failure, renal Latin for kidneys because it would be too easy to call it kidney failure]. In these CRF patients, ESA administration was shown to increase the red blood cell count (as measured by blood hemoglobin or hematocrit values) and reduce the need for red blood cell transfusion.

    Two ESA products are licensed in U.S. and one of these products is marketed under two names, Epogen/Procrit. The other ESA, darbepoetin alfa, is marketed solely under the proprietary name, Aranesp.

    Following the initial approval of ESAs, randomized, controlled clinical trials showed that the use of ESAs to attain higher, compared to lower, hemoglobin/hematocrit levels was associated with an increased risk for mortality and serious cardiovascular risks." That means death.

    That's the FDA writing in 2007. In other words ESAs were approved only because they reduced need for blood transfusions, and subsequent studies found them to be dangerous. Indeed, all major studies referenced found larger ESA administration to be associated with more deaths.

    Further study seems to have confirmed that. See, e.g., that <content.nejm.org/cgi/c...;;.

    Three doctors writing from the FDA confirm that in May 2010: "These trials compared ESA dosing strategies that achieved higher hemoglobin targets with those that achieved lower ones,1,2,3 [references] and in each trial, the higher target was associated with a worse outcome. However, these higher targets exceeded recommendations in ESA labeling. The recommended hemoglobin target has never been compared with placebo in a trial of sufficient size to assess cardiovascular outcomes [deaths]." (That's from <content.nejm.org/cgi/c...;; page bottom.)

    What's astounding is the sentence that follows that: "Given the plausible symptomatic benefits of higher hemoglobin levels and some evidence of better outcomes, it still appears that randomized trials are needed that compare lower targets, lower ESA doses, and less aggressive dosing strategies."

    Given clear and consistent evidence that higher doses of ESAs are associated with more deaths, and immediately after stating that ESAs have never been materially tested relative to placebo, the FDA authors suggest conducting tests not of ESAs relative to placebo but of different dosing strategies of ESAs.

    This was the astounding feature of every significant trial in that FDA backgrounder: every study compared not ESAs relative to treatment without them, but some dose of ESAs relative to another dose of ESAs. Of course the outcome of every such study must be a prescription for one dose of ESAs or another dose. Such damn ridiculous studies are a drug company's dream.

    3d. Here, you see, is my question with regard to the critical (b) from 1h above (that is, whether the FDA *should* approve it):

    All significant studies of ESAs, including of Hematide, appear to have tested -- and only tested -- non-inferiority of some dosing of ESAs relative to another dosing of ESAs. Even that "placebo" study used ESAs rather than blood transfusion when patient hemoglobin dipped below a certain level.

    Where is the documentation of large-scale studies of chronic renal failure patients, on dialysis and pre-dialysis, that demonstrate non-inferiority of treatment with ESAs -- any of them (Epogen, Aranesp, Hematide) -- relative to treatment without any ESA?

    Those are the studies that should undergird this class of drugs. If they exist: please supply the weblinks. If they don't exist, then -- and regardless of what such studies will find -- the FDA has gotten approval of this class of drugs wrong for two decades.

    3e. If and when we see those large-scale studies that should undergird use of ESAs for patients with kidney failure, what are the overwhelming quantified benefits of ESAs relative to their cost?

    3f. That brings us back to the endpoints of investment analysis: risk bucket, sensible expected value estimate, and price. Value Estimate is the primary endpoint of investment analysis. For value estimate we use worst(a,b), where (a) = the reasonably expected and (b) = our provisional view of what would transpire in an enlightened world. For Affymax, (a) hinges on the reasonably expected probability of FDA approval and commercialization of Hematide, while (b) hinges on the answers to 3d and 3e. At a glance, there appears to be no reasonable investment case for AFFY before we have answers to 3d and 3e. Surely someone who benefits from selling several billion dollars of these drugs per year can answer those questions instantly.
    Jul 9, 2010. 06:49 AM | 1 Like Like |Link to Comment
  • Affymax Sell-Off: Did Investors Overreact? [View article]
    Thanks Joseph --

    "My main argument is that AFFY was $23 two weeks ago and under $6 now; what changed, and is it reasonable?"

    Right, that doesn't do it for me, because I don't consider prices necessarily sensible before any over- or under-reaction. Equity, rather than cash viewed without regard to liabilities (that will cause greater cash use than accrual losses in due course), would appear to be the relevant indicator of AFFY value excluding Hematide, and that's around $2/share, not $6. With further burn: $1/share. That's over an 80% loss if Hematide is a no-go, without comment on the probability of its approval.

    "In much of your response you are attacking ESAs in general..this class of drugs has come under fire because of the side effects and their effiacy criticized as you pointed out. You have provided some very significant data to this end. But the FDA and general medical community still believes them to be effective for their intended purpose of maintaining Hb levels, with the intention of reducing morbitity and mortality."

    Right, and you see the FDA aimed to revisit that later this year, and any bullish case that they won't be curtailed ought to be based on facts of cost and efficacy relative to alternative treatment. Where are those facts?

    The FDA seems to nail the missing facts here (May 2010):

    content.nejm.org/cgi/c...

    "The applicability of the results of major ESA outcome trials to the approved use of ESAs is not straightforward. These trials compared ESA dosing strategies that achieved higher hemoglobin targets with those that achieved lower ones,1,2,3 and in each trial, the higher target was associated with a worse outcome. However, these higher targets exceeded recommendations in ESA labeling. The recommended hemoglobin target has never been compared with placebo in a trial of sufficient size to assess cardiovascular outcomes. Given the plausible symptomatic benefits of higher hemoglobin levels and some evidence of better outcomes, it still appears that randomized trials are needed that compare lower targets, lower ESA doses, and less aggressive dosing strategies.

    The intersection of ESA dose, dosing strategy, ESA responsiveness, and hemoglobin target is complex, and we plan to hold a public discussion of these issues at an advisory committee meeting. We must get this right; the public health stakes are extraordinarily high."

    Those are FDA employees' comments in May 2010. Boom: "The recommended hemoglobin target has never been compared with placebo...." How the heck has a multibillion/year drug been sustained without comparison to no-drug, to transfusions alone, in any significant study?

    And here's <wiki.noblood.org/Eryth...(;wiki.noblood.org/Ery...
    "generally EPO is still quite expensive (about twice as costly) relative to the cost of blood transfusion, for which EPO is an alternative treatment."

    Here's one of the damning studies:
    content.nejm.org/cgi/c...

    And here's the only intelligent comment I've seen on it (<content.nejm.org/cgi/c...;;;: same thread as above):

    "When a drug is found to be toxic as compared with placebo in a large clinical trial, the logical response is to discontinue the drug in clinical practice or to lower the dose as much as possible."

    To recap: AFFY may be worth around $1/share if Hematide is a no-go; the FDA may require more studies (i.e., even one decent study) of the safety of this class of drugs relative to transfusions for anemic patients; higher concentrations of this class of drugs appear to cause higher deaths consistently; transfusions alone sound cheaper than these drugs absent evidence to the contrary; and you have the little issue of those extra deaths on Hematide, which in my mind are almost beside the point that the value and safety of this class of drugs appear unmeasured relative to blood transfusions alone. If they are indeed unmeasured (and this would be a surprise to me, as I know nothing about blood transfusions and just started looking at it), then the FDA has dropped the ball for two decades and might reasonably use any excuse to make AFFY test Hematide, before it's approved, not relative to Amgen's similar drugs, but relative to placebo with no similar drug administration.
    Jul 8, 2010. 02:42 PM | Likes Like |Link to Comment
  • Affymax Sell-Off: Did Investors Overreact? [View article]
    (Parenthetically:

    Here's a little background on this class of drugs: www.fda.gov/ohrms/dock...

    Discussed studies appear both clever and annoying, in that they evaluate Epoetin (Epogen/Procrit approved beginning in 1989) and Darbepoetin (Aranesp approved beginning in 2001) only in terms of different dosing, not relative to blood transfusions alone or whatever other treatment predated Epoetin. The obvious objection is: all such studies are pathetically favorable to the drug company, for they all must conclude to take more or less of the given drug.

    Moreover they horrendously assume that the differently adverse outcomes result from different targeted blood levels rather than from side effects of the drug. Indeed, persistent findings of higher deaths accompanying higher drug concentrations may simply reflect a drug that's dangerous for reasons irrelevant to the targeted blood level. Aranesp appears to have been approved in 2001 based on its non-inferiority to Epoetin approved in 1989, which appears to have been approved based on flimsy studies of not much. Why in the world does the FDA accept all these studies comparing more or less use of the same drug rather than its use relative to other treatments?

    The simple question remains: what are the alternatives to Epogen/Aranesp/Hematide treatment -- just more frequent blood transfusions? -- and what is a solid estimate of $X and $Y above.)
    Jul 8, 2010. 12:21 PM | Likes Like |Link to Comment
  • Affymax Sell-Off: Did Investors Overreact? [View article]
    Looking at this for the first time this morning, because I enjoyed some of your earlier write-ups. Two issues:

    1. "This situation poises Hematide to have the long-lasting erythropoietin market in the US essentially cornered." AFFY 8-K filed 6/21/2010 notes "Hematide was dosed once every four weeks while ... darbepoetin was dosed every two weeks," not one week. In itself, is 4-week vs. 2-week dosing a significant advantage?

    Moreover: "Aranesp [darbepoetin] is approved for once-monthly dosing for treatment of anemia in pre-dialysis patients in Europe. In the U.S., Amgen is reportedly in the process of seeking approval for once-monthly dosing of Aranesp for treatment of anemia in pre-dialysis patients." (AFFY 10-Q filed 5/6/2010)

    2. I stopped reading the AFFY 10-Q filed 5/6/2010 at the following excerpt, because it suggests the class of drugs is inferior to taking nothing in all critical respects except the requirement for blood transfusions, and "The FDA recently announced that it anticipates convening a committee meeting later in 2010 to re-evaluate the use of ESAs...." Consequently I imagine that any bullish case for this class of drugs should quantify their apparent cost benefit vs. blood transfusions along the lines: administering Aranesp [darbepoetin] to 1,000 patients costs $X and saves $Y in blood transfusions. What are your estimates of $X and $Y, and on which facts do you base those estimates?

    "In late 2009, Amgen Inc., or Amgen, announced the results of its large, randomized, double-blind, placebo-controlled Phase 3 study of patients with chronic kidney disease (CKD) (not requiring dialysis), anemia and type-2 diabetes (TREAT). In this study, treatment of anemia with Aranesp to a target hemoglobin of 13 g/dL, which is higher than the 10 g/dL - 12 g/dL range approved by the FDA in the current label, reportedly failed to show benefit compared to placebo with regard to composite of time to all-cause mortality or cardiovascular morbidity (including heart failure, heart attack, stroke, or hospitalization for myocardial ischemia) and a composite of time to all-cause mortality or chronic renal replacement. In addition, higher rates of stroke were reported amongst patients treated with Aranesp compared to the placebo group. Further, among a subgroup of patients with a history of cancer at baseline, a statistically significant increase in deaths from cancer was observed in the Aranesp treated patients compared to placebo treated patients. However, Aranesp treatment reportedly was associated with a statistically significant reduction in blood transfusions and a modest improvement in patient reported fatigue.

    "In January 2010, FDA officials recently published an editorial in the New England Journal of Medicine entitled “Erythropoiesis-Stimul... Agents — Time for a Reevaluation” and announced that it anticipates convening a public advisory committee meeting later this year to evaluate the use of ESAs in the treatment of anemia due to chronic kidney disease. The editorial noted that a number of randomized trials, including TREAT, have attempted to show that using ESAs to raise hemoglobin concentrations to higher targets improves clinical outcomes but rather have suggested the opposite...."

    Your bullish case appears to rest on the infrequent dosing advantage of Hematide -- questioned on a quick view of the facts per #1 -- and its general non-inferiority relative to a class of drugs that the above excerpt suggests are dangerous and worthless in all critical respects beyond reducing blood transfusions.

    Consequently that key question: administering Aranesp [darbepoetin] to 1,000 patients costs $X and saves $Y in blood transfusions. What are your estimates of $X and $Y, and on which facts do you base those estimates?
    Jul 8, 2010. 11:00 AM | 1 Like Like |Link to Comment
  • First Line Treatment for Obesity – The Holy Grail of Biotech  [View instapost]
    "To demonstrate the selective nature of Lorcaserin, ARNA undertook one of the largest echo monitoring studies ever done including more then 7,000 patients, clearly demonstrating that there is no increased risk of cardiac valvulopathy. Safety – Check."

    Where are the logicians? That doesn't mean targeting 2c is long-term safe.
    May 31, 2010. 06:01 AM | Likes Like |Link to Comment
  • Mannkind: Overlooked Biotech With Excellent Prospects (Part I) [View article]
    Mannkind's 10K compares "Post-meal insulin levels in healthy subjects" with "Post-meal insulin levels with Technosphere Insulin in patients with type 2 diabetes": "This [graphical] comparison illustrates the degree to which our Technosphere System approximates the insulin profile of healthy individuals following meal onset."

    But it appears to do nothing of the sort. The first graph peaks at 60 minutes, whereas TI appears to peak within 15. Scales of the two graphs are even dissimilar. Does anyone actually read the 10K? What am I missing in this graphical comparison?
    Jul 10, 2008. 02:34 AM | Likes Like |Link to Comment
COMMENTS STATS
87 Comments
74 Likes