During the early days of the 2009 H1N1 influenza A pandemic, the popular herbal formula maxingshigan–yinqiaosan was used widely by TCM practitioners to reduce symptoms. (It’s hard to pronounce and spell, so I’ll refer to it as M-Y.) A new study was done to test whether M-Y worked and to compare it to the prescription drug oseltamivir. It showed that M-Y did not work for the purpose it was being used for: it did not reduce symptoms, although it did reduce the duration of one sign, fever, allowing researchers to claim they had proved that it works as well as oseltamivir.
“Oseltamivir Compared With the Chinese Traditional Therapy: Maxingshigan–Yinqiaosan in the Treatment of H1N1 Influenza” by Wang et al. was published in the Annals of Internal Medicine earlier this month. The study was done in China, which is notorious for only publishing positive studies. Even if it were an impeccable study, we would have to wonder if other studies with unfavorable results had been “file-drawered.” It’s not impeccable; it’s seriously peccable.
It was randomized, prospective, and controlled; but not placebo controlled, because they couldn’t figure out how to prepare an adequate placebo control. They considered that including a no treatment group compensated for not using a placebo control, and that objective temperature measurement could be expected to get around any bias. It might not: the nurses who took the temperatures were blinded to the study, but the patients were not. It’s possible that those who knew they were getting M-Y might have believed in it and their bias might have somehow subtly influenced data gathering so that M-Y appeared more equivalent to oseltamivir than it actually was.
There are other problems besides the lack of blinding.
Background
It’s always fun to read the introduction section of a CAM paper. It gives insight into the researchers’ beliefs and frequently includes footnote references to studies that don’t say what they seem to think they say. In their introduction, Wang et al. say
Traditional Chinese medicine has been used to treat seasonal influenza for thousands of years.
The article they reference to support that claim doesn’t support it. It is an opinion piece about “Chinese medicinal materials and their interface with Western medical concepts.” It’s not a historical study, and it doesn’t directly address influenza . Anyway, thousands of years ago TCM did not recognize influenza as a unique disease entity or have any way of diagnosing it accurately. And TCM does not equate to M-Y.
Next they cite a meta-analysis that they say showed the effectiveness and safety of TCM, but it was not about flu or about M-Y. It was about studies of common cold patients with wind-cold syndrome. It concluded
TCM new drugs developed in recent years for preventing and treating common cold have better therapeutic effects than the old ones. They can accelerate the onset time of lowering body temperature and improve the symptoms of common cold without any significant adverse reactions. Because of lacking of placebo-controlled and blank-controlled studies, further high-quality trials are still needed.
The abstract is too garbled to understand, but it indicates that if there was any significant effect of “TCM” it was very small in magnitude. Saying that a study of TCM for colds shows that M-Y works for influenza is like saying that a study of Western medicine for pneumonia proves that penicillin works for bronchitis.
Design
the decision about whether to initiate antiviral therapy is individualized on the basis of the clinician’s judgment and on what is known about the benefits of therapy. Therefore, it was ethically possible for us to include a control group that received no intervention.
If individual decisions are ethically acceptable, how can it be ethical to decide that a whole random group of patients will be consigned to treatment or no-treatment groups?
The new study had over 400 subjects and few dropouts. Patients were admitted to the hospital and followed closely. Viral studies confirmed H1N1 influenza. There were 4 arms: M-Y, oseltamivir, M-Y plus oseltamivir, and a no treatment group. Endpoints were both objective (temperature and viral shedding) and subjective (reported symptoms).
Results
Time to resolution of fever was significantly shorter in all treatment groups than in no-treatment control group, and was comparable in the different treatment groups. There were no differences between groups on symptoms or viral shedding. There was no difference in acetaminophen use between groups. Patients in the control group were more likely to be prescribed antibiotics, even though they acknowledge that physicians often respond inappropriately to persistent temperature elevations by administering antibiotics that are not called for in that situation. If they knew that, why did they allow it to happen?
Complications like pneumonia were rare and not significantly different between groups.
Side effects were incomprehensibly rare: Two patients reported nausea and vomiting while taking M-Y and there were no side effects in the other groups. This is puzzling, since studies of oseltamivir have shown a 9-10% incidence of nausea and vomiting with the drug and 3-6% with placebo.
They interpreted their findings as showing that maxingshigan–yinqiaosan does not act as an antiviral, whereas the benefit of oseltamivir derives entirely from its antiviral activity. Their data do not justify that interpretation.
The mixture
When herbs are mixed, I always wonder how they decided which ones to use. In this case, maxingshigan studies have shown immunomodulatory effects, but what about the other 11 ingredients?
M-Y is a mixture of 12 herbs. In this study the herbs were tested to rule out contamination with heavy metals, insecticides, and microbes. The need for such testing brings up a concern: when patients buy M-Y on the market, what are they getting?
The 12 ingredients, as far as I could determine from the information in the paper, correspond to ephedra, anemarrhena rhizome, artemisia (wormwood), gypsum, honeysuckle, skullcap, bitter apricot kernel, forsythia fruit, peppermint, fritillaria bulb, burdock, and licorice. Only about half of these are listed in the Natural Medicines Comprehensive Database. For most of these, flu is not listed as an indication (although allergies and respiratory infections are sometimes mentioned), and most say there is insufficient evidence to rate safety or to rate effectiveness for any indication. Some include warnings: ephedra is unsafe and is illegal in the US; apricot kernels are a source of cyanide. It is hard to imagine a rationale for combining these particular 12 ingredients for any purpose, much less for flu. Maybe I lack imagination.
An appendix that lists Chinese and English names for the ingredients includes this helpful explanation:
Chinese herbs are usually prescribed in formulas that contain “king” medicines, which provide the strongest therapeutic action; “minister” medicines, which assist the “king” medicine in its therapeutic actions; “assistant” medicines, which aid the “minister” medicine in treating a lesser aspect of the disease; and “ambassador” medicines that are intended to reduce the toxicity of the other medicines in the formula or guide the formula to the targeted organ or region of the body.
They offer no justification for which ingredient has which of these purposes or how they knew whether they actually work for those purposes. These terms are not scientifically meaningful and are unexpected in a journal like the Annals of Internal Medicine.
Conclusion: I’m not convinced
Most patients with influenza recover with no specific treatment; but oseltamivir is often used with the goals of faster recovery, reduced symptoms and fewer complications. It is also used to prevent influenza. But it is far from perfect. Studies have shown that it reduced the risk of pneumonia and of pneumonia, fever, and viral shedding. Its symptom-reducing effects are not very robust, and it can have some side effects. It would be nice if we had a safer, more effective remedy.
They concluded:
Maxingshigan–yinqiaosan can be used as an alternative treatment of H1N1 influenza A virus infection when oseltamivir is not available.
I can’t really argue with that, but why not try to make oseltamivir available? Why not try to ensure the purity and safety of available preparations of M-Y? More importantly, why not try to figure out what each of those 12 herbs does, whether all 12 are essential for its effectiveness, and whether worrisome ingredients like ephedra and cyanide could be eliminated?
It is admirable that TCM practitioners are trying to study their remedies scientifically; but this study doesn’t convince me that M-Y offers any advantage over oseltamivir, or even that it is a reasonable substitute.