Thought experiment: Design your own Vitamin C trial

[godsilove]

If you had $1,500,000 to spend on a clinical trial evaluating Vitamin C for any cardiovascular condition(s) of your choice, how would you design the trial?

I've set up a crude excel template (downloadable here) to help determine the costs of the trial for this brainstorming exercise. It takes into account the number of patients that would have to be screened, the number of patients that will be enrolled, the total duration of the study, length of treatment, daily dose of vitamin C used, the number of follow-up visits, the tests used, and administrative costs.

Oh, and you have to give your trial a name:

Ascorbate in the Treatment of Essential Hypertension in Non-Diabetic Patients (ATTEND)

Condition: Primary hypertension in untreated, non-diabetic patients

Intervention: 1000 mg ascorbate 3x a day = 3000 mg/day
500 mg ascorbate 3x a day = 1500 mg/day

Inclusion criteria: Men and women aged 18-55. Systolic blood pressure between 120 and 160.

Exclusion criteria: Pregnancy or lactating. Diabetes. History of cancer. History of autoimmune conditions. History of cardiovascular events e.g. stroke,myocardial infarction, PCI, CABG. Concomitant NSAID use. Any surgical procedures within the last 12 months. Secondary hypertension. Abnormal creatinine clearance rate. Proteinuria. Renal disease. Abnormal renin levels. HIV. Current or prior use of antihypertensive medications e.g. ACE inhibitors or beta-blockers. History of trauma. Use of vitamin C supplements in previous six months. Excessive alcohol intake. Hyperkalemia. Current use of psychiatric medications. Participation in clinical trial in previous six months. Concomitant use of hormonal therapies.

Rationale: Previous randomized have shown a moderate, short-term effect of vitamin C supplementation on hypertension, although data is conflicting. Epidemiological studies also show an inverse relationship between plasma ascorbate levels and blood pressure in young populations.

Design: Double-blind, randomized, placebo-controlled study
Three treatment arms (placebo, 1500 mg/day, 3000mg/day)
36 patients in placebo arm, 72 patients in each experimental group (180 patients in total)
Patients will be enrolled for 12 months, with a follow-up visit every three months
At each visit, blood pressure, plasma ascorbate levels, cholesterol, Lp(a), C-reactive protein, blood glucose, creatinine clearance rate, urine electrolytes and urine albumin will be measured. Patients will also be required to bring in their pill bottles for a pill count to record compliance, and to complete a food frequency questionnaire.

Primary end-points: Mean change in daytime ambulatory systolic blood pressure and diastolic blood pressure from baseline

Secondary end-points: Mean changes in total cholesterol, LDL:HDL ratio, Lp(a), C-reactive protein, blood glucose, plasma ascorbate levels, and creatinine clearance rate from baseline

Number of patients: 180

Estimated length of trial: 30 months


Please feel free to comment.

[ofonorow]

This is a great idea. And I will participate, but we did something like this some time ago. The following is rather long, but it was an effort to design a Paulingtherapy study, that the Foundation could afford. People may want to "brush up" on the previous discussion, however we were constrained by reality! If we really had over a million dollars, it would be an entirely different ball game!

viewtopic.php?f=10&t=4749

[ofonorow]

See:

http://www.heartattackproof.com/resolving_cade.htm

The outcome measure of the ideal study should include before after images, such as the first A/B angiogram images contained in the above link, e.g. images that even casual observers can immediately understand. Images which make it is easy to see the improvement in health of the artery.

Recently, I have become satisfied that retinal imaging can make the same observations (as the angiogram images) and it turns out that there is a large body of work confirming this, i.e., that the findings in the eye carry over to the entire problem of CVD throughout the body. The big advantage of retinal images is that they are noninvasive. But in the case of having 1.5 million, we should have both - the angiogram to verify that the same conditions exist in the microscopic retinal arteries. (Depending on the outcome, in the future, retinal imaging might take the place of angiogram entirely. At that point, it certainly becomes easy for eye doctors to monitor patient progress and adjust dosages - with noninvasive diagnostics.)

There are other measures of course, when cost is no object, such as a good Lp(a) and Cholesterol test. (We had planned to use the VAP test because they routinely segregate Lp(a) into 5 subgroups. (It is known that the smaller, denser Lp(a) is more atherogenic (dangerous) so that Lp(a) level per se (just volume or just mass) is not sufficient to ascertain risk.) Atherotech can also provide all cholesterol numbers as well as homocysteine, etc. using the same blood.

I also like the idea of video tape interviews showing the patients before/after, such as their ability to walk, their color, their breathing, etc.

A very simple test we had planned to also employ tests for arterial stiffness (ASI) using a blood pressure cuff fed into a computer. The unit measures the ability of arteries to "bounce back" after being squeezed - or stiffness. The more calcium, generally the higher the stiffness index. The unit is called CardioVision.

Finally, I would want to measure serum vitamin C levels, and other levels if possible. (Would be a check on whether patients were taking their Pauling therapy "medicine") and that brings us to the issue of a Placebo group.

Because of the high amounts required, it is almost impossible to have a true placebo. In other words, it would be almost impossible for patients not to know they are in the placebo (control) group. The longer the study goes, the more of a problem this becomes.

This is why we invented an entirely new paradigm for our study. The idea is to start by running all tests on everyone - for as long as the skeptics think necessary - we thought 2 or 3 months (depending on the illness, they may not survive that long, or much longer) and you take and record all measurements. This becomes the baseline. The control.

Then in all the subjects you introduce the Pauling therapy. If the group is small, you give them the same formula. And the study can then continue as long as necessary, every month comparing the results to baseline.

If the group is large, different variations can be added to subgroups. All of which should get vitamin C/E and lysine, but perhaps a group gets proline. Another group might get extra vitamin A, and a group should get the entire Tower Ascorsine-9 formula, etc.

[godsilove]

With respect to a placebo group, it does not have to be patients not receiving treatment or getting a dummy pill. You could have a control arm that is receiving standard medical therapy (i.e. statins, aspirin, ACE inhibitors, etc) and the treatment group receiving either Pauling therapy alone or PT + standard medical therapy. Obviously, patients need not be blinded to treatment in this case. I think a head-to-head study would be pretty interesting - and if PT turns out to be more effective, then you'll want the trial to be as well designed as possible in order to ensure that critics can't easily dismiss the results.

Placebos are not only in place to ensure that patients do not know whether they are receiving treatment - it also allows for double-blinding, which removes potential bias from investigators. But as you say, in certain scenarios it is difficult to have true double-blinding, especially when the intervention has a distinctive taste or peculiar side effects. I still think it is possible to have blinding with PT (unless you're planning to use Cardio-C, in which case coming up with a "dummy" drink would be difficult). The main side effects with high-dose vitamin C, i.e. bloating and diarrhea, are often commonly associated with placebos in addition to nausea, headaches, etc. Not every patient complaining about bloating will necessarily be getting vitamin C, and so I think it would still be possible to maintain a sufficient amount of blinding. Perhaps the control arm could get all the components of PT other than ascorbate, whereas the treatment group would get PT. I don't think it would be prudent to breakdown all the components of PT and have several treatment arms, as that would then require quite a large study sample in order to ensure that statistical significance can be reached.

I think a single-arm study is fine when doing a pilot study with a handful of patients or a Phase I study that is primarily evaluating the safety or pharmacokinetics of the experimental agent. However, I think that a control arm makes the study far more rigorous. In oncology, even though Phase I studies also aim to evaluate efficacy e.g. tumor shrinkage or response rates, such studies alone are not sufficient enough to determine how effective a treatment is. Even if the agent shows remarkable efficacy, Phase II and III studies comparing the treatment to the standard of care or best supportive care are still necessary.

[ofonorow]

Well, since there haven't been ANY studies with dosages even close to Pauling's recommended dosages other than the "lost" Kale Kenton study, and perhaps Matthias Rath's on his own products, (however, his products were quite low in dosages) shouldn't the aim should be to have it run and published (rather than worrying about whether it will be widely accepted?)

There are also ethical reasons behind wanting to change the placebo paradigm. Many alternative doctors would consider it unethical to deprive patients of a treatment they know works. But the central problem is the general availability of vitamin C at any drug store. If I am a typical subject in a long term study, and I determine that I am not in the vitamin C group, the heck with the study. I'll take my own vitamin C.

Since this is a thought experiment, why not bat around the idea of the initial baseline. (I am assuming that all patients would be either stable, or declining from all measures during this initial measuring period. ) If, as I predict, all patients begin improving after the introduction of Pauling's protocol, perhaps at different rates, I don't see why you couldn't run statistical analysis against the baseline. This would also provide more knowledge, as every patient would have different dosage requirements, and thus would be predicted to improve at a different rate from the others.

I don't see a downside from making all subjects their own control.

[godsilove]

Well, since there haven't been ANY studies with dosages even close to Pauling's recommended dosages other than the "lost" Kale Kenton study, and perhaps Matthias Rath's on his own products, (however, his products were quite low in dosages) shouldn't the aim should be to have it run and published (rather than worrying about whether it will be widely accepted?)

I definitely think a small study of PT should be run before any larger studies in the real world. But for the purposes of this hypothetical exercise, you have a lot more imaginary money to work with. In fact, even if this were a real-world exercise - why would you want to spend $1.5m on a study only to have results that are ambiguous because of the study design? This is not to say that future trials will not be possible.

There are also ethical reasons behind wanting to change the placebo paradigm. Many alternative doctors would consider it unethical to deprive patients of a treatment they know works.

I agree that there are ethical considerations to the use of placebo. This is why most cancer trials (and in other therapeutic areas as well) will give patients in the placebo arm the standard of care, except in trials for certain advanced cancers where treatment options are limited. It is also why many trials have prespecified interim analyses to decide whether a trial should continue.

I think that there are a number of factors that need to be considered. For instance, placebos might be considered more unethical in a cancer trial than in a trial for a cosmetic agent. Furthermore, you could have conditions where patients do not normally receive treatment - for instance, many HIV patients do not get initiated on antiretroviral therapy when their CD4 counts are above a certain level - some doctors prefer to wait until the levels drop before starting treatment. In light of this, I would not consider it unethical to run a short-term, placebo-controlled trial on newly diagnosed HIV patients with CD4 counts above 350.

But the central problem is the general availability of vitamin C at any drug store. If I am a typical subject in a long term study, and I determine that I am not in the vitamin C group, the heck with the study. I'll take my own vitamin C.

I'm not sure if that's necessarily a description of a "typical" subject. Subjects would obviously be informed of certain restrictions, e.g. you're allowed to take a multivitamin of your choice, but no vitamin C supplements. After a point, you're basically relying on the integrity of the participants. You can't always ensure that patients in the treatment arm take their medication as instructed either. Nonetheless, I think this would only be a concern in a minority of patients - which is another plus point for having a larger trial.

Since this is a thought experiment, why not bat around the idea of the initial baseline. (I am assuming that all patients would be either stable, or declining from all measures during this initial measuring period. ) If, as I predict, all patients begin improving after the introduction of Pauling's protocol, perhaps at different rates, I don't see why you couldn't run statistical analysis against the baseline. This would also provide more knowledge, as every patient would have different dosage requirements, and thus would be predicted to improve at a different rate from the others.

I don't see a downside from making all subjects their own control.

A cross-over study design could be used to test the effect of placebo, and have all patients receiving treatment at least for part of the study. You could have some patients receiving treatment and some receiving a matching placebo - and then halfway through the study have them switch. The patients could still be blinded to treatment. (An interesting side project would be to survey the respondents at the midway point and after the trial to see what % are able to correctly guess when they received treatment). Depending on which condition is being treated, this could be one way of ensuring that all patients get treated.

For some conditions, depending on baseline measures alone without having a control arm can be problematic. Some conditions are more susceptible to the placebo effect especially psychiatric conditions where there are few biological markers that can be objectively tested. And certain baseline measures don't always get worse. For instance, if you were measuring LDL-cholesterol, some patients will have lower than baseline levels even without being on any treatment.


With respect to retinal imaging - it sounds like an interesting and cost-friendly approach. However, are there are studies that establish its prognostic value? For instance, a study that correlates measures from this method with carotid IMT, or even with long term morbidity or mortality rates?

[ofonorow]

[ofonorow]