Does alternative medicine kill?

Interesting blog:

http://scienceblogs.com/insolence/2009/03/does_alternative_medicine_use_result_in.php

Does alternative medicine use result in worse outcomes in breast cancer?

Has it really been that long?

More than two years ago, I wrote a post entitled Death by Alternative Medicine: Who's to Blame? The topic of the post was a case report that I had heard while visiting the tumor board of an affiliate of my former cancer center describing a young woman who had rejected conventional therapy for an eminently treatable breast cancer and then returned two or three years later with a large, nasty tumor that was much more difficult to treat and possibly metastatic to the bone, which, if ture, would have made it no longer even possibly curable. My discussion centered on what the obligation of physicians are to such patients who utterly refuse the science_ and evidence_based medicine that we know to be able to cure them of a potentially fatal disease and centered on the reaction of one oncologist who stated that it is the collected "our fault" (as in all the physicians who saw her) that she ended up this way because "we" had failed to persuade her that what she was doing would at the very least delay her treatment and make it much harder to treat her when her tumor did progress to the point where she realized that whatever woo she had chose was not working and at the very worse cost her her life. Even more compelling and sad, this patient was young, in her late 30s, and had three young children. Her husband had even filed for divorce and custody of the children on the quite reasonable, but no doubt painful to come to, basis that she would soon no longer be able to take care of them and had rebuffed all attempts to persuade her to accept science_ and evidence_based medical care that could still possibly save her life.

Since that post, I've always been meaning to take a look at what, exactly, the effect of choosing "alternative" medicine over "conventional" medicine is on the odds of survival for breast cancer patients. Even though intuitively one would hypothesize that refusing scientific medicine and relying on placebo medicine instead would have a detrimental effect on survival, it turns out that this question is not as easy to answer as one might think. For example, if you do a search on PubMed using terms like "alternative medicine," "breast cancer," and "survival," the vast majority of the hits will be studies of complementary and alternative medicine (CAM) and breast cancer with little reference to what possible effect these therapies might have on overall survival and breast cancer_related mortality. I can speculate about several posisble reasons for this, although I don't know how valid my guesses are as to why. The first reason may be that__thankfully__relatively few women use alternative medicine primarily or exclusively to treat their breast cancers. It can't be overstated how this is a very good thing. Also, those who do use only alternative medicine probably drop off the radar screen of their "conventinal" doctors, and it is difficult, if not impossible, to capture data regarding their outcomes. All too often they stick with their alternative healers until the end, going from one to another as their tumors eat through their skin and ravage their bodies. True, they may pop up again in their surgeon's or primary care doctor's office with huge, fungating tumors, only to be told that they have to undergo chemotherapy to shrink the tumor before any surgery is possible, after which they will often disappear again.

Another important reason is that the natural history of breast cancer is extremely variable, from nasty, aggressive tumors that kill within months to indolent, slow_growing tumors that, even when metastatic, women can survive with for several years. (It is, of course, these women who usually show up in "alternative medicine" testimonials, because they can survive a long time with little or no treatment before their tumors progress.) Because it's important to understand the natural history of breast cancer, I'll reference a classic study examining the natural history of untreated breast cancer. It was published in 1962 by H. J. G. Bloom, W. W. Richardson, and E. J. Harries, and examined data from Middlesex Hospital from 1805 to 1933 where 250 cases of untreated breast cancer were identified and studied. They calculated survival as the period of time from onset of symptoms to death. What they found was that 18% of the 250 patients survived five years; 3.6% survived 10 years; and 0.8% survived 15 years. Of note, it was 19 years before all patients were dead. Overall, the median survival was 2.7 years. A survival graph from this classic paper is below:

It should be noted that all of these tumors were detected as (at the very minimum) lumps in the breast, given that there was no other way of detecting them at the time. However, the reason we go back to this study time and time again is because, at least in developed countries, it is the rare woman with breast cancer who does not undergo treatment of some kind for it. These days, most tumors are detected at far less advanced stages; indeed, most are detected by mammography. What that means is that, if such a study could be done today, it is very likely that lead time bias would significantly increase the apparent median survival, because increasingly tiny tumors are being found. It is also possible that a significant number of very small tumors might spontaneously regress, which further complicates the issue today, not to mention making it easier to find women who have rejected some or all of "conventional" medicine to treat their cancers and survived significant lengths of time to produce alternative medicine testimonials.

With this background, I have found a couple of studies that can help answer the question. The first one was published in 2005 in the Annals of Surgery by a group in from Geneva University Hospitals. This study involved a search of Switzerland's database between 1975 and 2000 and included 5,339 patients diagnosed with nonmetastatic breast cancer. The strength of this study is that the Geneva Cancer Registry includes data from all patients from the Geneva canton who underwent treatment and allowed the investigators to compare the outcomes of the women who refused to undergo surgery with curative intent with those who underwent surgery. In the Registry, there were identified 70 patients (1.3%) who refused surgery and concluded:

These women [those who refused surgery] were older, more frequently single, and had larger tumors. Overall, 37 (53%) women had no treatment, 25 (36%) hormone_therapy alone, and 8 (11%) other adjuvant treatments alone or in combination. Five_year specific breast cancer survival of women who refused surgery was lower than that of those who accepted (72%, 95% confidence interval, 60%_84% versus 87%, 95% confidence interval, 86%_88%, respectively). After accounting for other prognostic factors including tumor characteristics and stage, women who refused surgery had a 2.1_fold (95% confidence interval, 1.5_3.1) increased risk to die of breast cancer compared with operated women.

It is true that this is not a randomized study; rather, it is a retrospective study. Consequently, it is impossible to rule out selection bias, but, as the authors point out, this is one case where doing a randomized study is unethical. Moreover, half the women accepted some form of other standard, effective treatm,ent, such as hormonal therapy alone. In any case, what this study shows is that women with no surgery can still live a long time, but are far more likely to die of their cancer than women who do undergo surgical extirpation.

But what about alternative medicine?

As far as I can find, there is one study that specifically looked at the question of what happens to women who opt for alternative medicine instead of scientific medicine. This study, like the one I just cited, was published in the surgical literature, specifically the American Journal of Surgery. Given the nature of the question it was seeking to answer, its design is single_armed and retrospective, using prognosis estimated by Adjuvant! Online, an online tool into which clinicians can enter prognostic factors of a breast cancer at the time of presentation and come up with an estimate of chances of survival and recurrence with and without treatment. This, of course, is a weakness, but, again, randomizing patients to scientific medicine or alternative medicine would be completely unethical. In the case of such questions, we scientists have to make do with whatever methodology we can; i.e., do the best we can with what we have. Unfortunately, the study was also small, only 33 patients. Even so, given the huge difficulties involved in undertaking such a study, the investigators, who, as private practitioners operating a community practice in Eugene, OR, went above and beyond the call by trying to look at their data and answer this question. That their study has a number of shortcomings is not their fault; they appear to haved done the best they could with what they had, which includes patients who underwent a panoply of alternative therapies, including coral calcium, herbal therapy, mushrooms, high dose vitamins, whey, chelation therapy, hemlock, and coenzyme Q10.

So what were their findings?

They're summed up in the following table:

The authors comment:

We found that the overwhelming majority of the patients who initially refused surgical treatment for breast cancer developed disease progression. Five of these patients ultimately underwent surgical resection. Of the other 6 patients, 5 had developed metastatic disease that precluded benefit from surgery. Furthermore, the disease progression caused by the delay in surgery was associated with an increase in the estimated 10_year mortality rate.

Patients who declined chemotherapy or hormone therapy faired slightly better. Optimism for this strategy should be severely tempered by the fact that the length of follow_up evaluation in these patients was relatively short, and these patients had early stage (I or II) disease. By software estimates, the 10_year mortality rate for these patients is still expected to be more than 50% higher than it would have been if the patients had taken their recommended therapy.

A number of patients who expressed their intention to pursue alternative therapies did not return for follow_up evaluation. Attempts were made to contact these patients. Those for whom follow_up evaluation was unavailable were excluded from this study. Although their omission may introduce a selection bias in the results, the effect of this bias is expected to be small because relatively few patients (14 of 47) were in this category.

I find two points important about this study. First, it confirms once again the importance of surgery as a primary therapeutic modality for breast cancer, especially early stage. Second, and more importantly, it strongly suggests that foregoing or delaying surgery or chemotherapy is at the very least associated with a significantly decreased chance of recurrence_free survival. The authors do note that it is impossible to tell whether this increase in mortality was solely due to delay or refusal of effective therapy or whether the modalities chosen were deleterious. My guess is that it was almost certainly due to the ineffectiveness of the alternative therapies chosen.

More evidence of the uselessness of "alternative" medicine in breast cancer was published two years ago by Edzard Ernst, author of Healing, Hype or Harm? A Critical Analysis of Complementary or Alternative Medicine. Ernst, as you might recall, was formerly a CAM advocate. In 2006, he wrote a review for the Breast Journal along with Katja Schmidt, MSc, C Psychol, and Michael Baum, MD, ChM, a review entitled Complementary/Alternative Therapies for the Treatment of Breast Cancer. A Systematic Review of Randomized Clinical Trials and a Critique of Current Terminology. The objective of the study was to examine all studies randomized clinical trials (RCTs) for "alternative cancer cures" (ACCs). Treatments examined included various methods of psychosocial support such as group support therapy, cognitive behavioral therapy cognitive existential group therapy, a combination of muscle relaxation training and guided imagery, the Chinese herbal remedy Shi Quan Da Bu Tang, thymus extract, transfer factor, melatonin, and factor AF2.

The first finding was that the methodological quality of the studies was, by and large, pretty low. The most common deficiencies included: lack of power sample calculation; small sample size; lack of adequate randomization and/or (patient and assessor or only assessor) blinding; and insufficient follow_up periods. It was noted that only one trial applied an intention to treat analysis. From the 15 studies Ernst examined, this is what he concluded:

The totality of the data fails to show a single intervention that would be demonstrably effective as an ACC. The paucity and the often_low methodological quality of the RCTs are as unexpected to us as they are disappointing. Most trails had small sample sizes; thus a type II error is conceivable. But even if this were true, one would be correct in stating that to date, no effective ACC has been identified.

A lot of this is, of course, true based on discussions of prior plausibility alone. One could argue that, given the poor quality of the studies examined by Ernst, there might be an effect that was missed. However, if an effect were missed, it would have to be small, and small effects are not what is claimed for many of these ACCs. What is often claimed is a near_miraculous "cure" for cancer, which, if it were true, would be relatively easy to detect. As I've often argued about, for example, the Gonzalez regimen for pancreatic cancer, if such ACCs really were cures, it would actually be fairly easy to show. In the case of pancreatic cancer, for instance, just producing well documented case reports of a few five year survivors among patients with documented metastatic adenocarcinoma of the pancreas would, I daresay, make even me sit up and take notice. Somehow, we never see this. Of course, what makes the question in breast cancer more difficult to answer is its highly variable natural history. Few people with metastatic pancreatic cancer survive more than one year (Patrick Swayze is one of those lucky few), fewer still longer than two; lots of women with even metastatic breast cancer do.

Putting it all together, even given the relative paucity of studies, I conclude that there is nonetheless no compelling evidence for a significant survival benefit due to any "alternative" therapy, nor is there even good evidence for significant treatment effects. The studies that do purport to show an effect are virtually all plagued with methodological difficulties and tend to show effects that are barely above background noise. The vast majority of them are retrospective and difficult to interpret, and what evidence is out there is that alternative medicine use among breast cancer patients is associated with an increased risk of dying from cancer, particularly when conventional therapy is eschewed. Taken together, these data make it very hard not to conclude that at best the vast majority of alternative therapies are either useless, no more than placebos, or that some of them might even be harmful. That is why they have no role in science_based medicine at present.

As Edzard Ernst put it:

The idea of an "alternative cancer cure" assumes that conventional oncology would not adopt a cancer treatment simply because it originates from an area outside of mainstream medicine. We feel that, should such a cure one day emerge, it would be investigated without delay by oncologists and adopted into routine care as soon as the data supporting it are convincing. Plant_based cancer medications such as Vincristin and Vinblastin (both extracted from the plant Vinca rosea) or Taxol (Taxus baccata) could be employed to back up this theory. It follows that the term ACC is and most likely will always be a contradiction in terms.

As I frequently put it: There is no such thing as "alternative" medicine. There is medicine that is effective, medicine that is not, and medicine that has not been tested yet. Nearly all of so_called "alternative" medicine falls into one of the latter two categories, and those that have not been tested yet nearly all fall into the category of being so wildly improbable that testing them without more positive evidence makes no sense. In any case, as a cancer surgeon, I don't care where a therapy came from. I really don't. If someone could show me that reiki or homeopathy cures cancer, I'd use either. In the meantime, I will continue to argue that the very concept of "alternative" medicine is a false dichotomy. Unfortunately, it's a false dichotomy that can kill.