Alternative Medicine and Cancer
Back to Index
The Hoffer/ Pauling Studies on the Orthomolecular Treatment of Cancer
Below is a letter I sent to Dr Abram Hoffer, one of the main proponents of the Orthomolecular (essentially megavitamin) therapy of cancer. It is self-explanatory, if you understand a little about clinical studies. It refers to Dr Linus Pauling's retrospective study of Hoffer's patients, wherein cancer patients who adequately applied Dr Hoffer's methods were compared with those who for any reason did not. Dramatic benefits were claimed.
Although Dr Hoffer is still referring to this work in his books, and says he is still following the apparently cured patients up, I was unable to obtain from him the simplest of clinical details regarding ANY of the patients, even when I supplied his own patient record numbers for those of interest. This was despite quite a lengthy correspondence.
The critical issue was his claim that ----
>The only difference between the control and the treated group
> was that the controls had not been on the treatment program for at least
> the two months I had laid down as essential.
My response was --
" I accept that this is your honest understanding. But what may seem to you to be reasonable assumptions are carrying disproportionately heavy implications for the significance of the work. If the ONLY difference between the groups is truly the treatment received, as Pauling also implied, you have demonstrated the effectiveness of orthomolecular treatment without question and regardless of how the data is evaluated. The clinical information offered thus far is, however, insufficient to establish he truth of the contention upon which everything hangs. I have pointed out some matters which raise doubts. "
No luck, even with this. In my initial letter (the one below) , I also pointed out that if the above were true he must now have a remarkable series of thirty or forty well-documented remissions in patients with proven terminal cancer. Where are they all?
Dr Hoffer is aging and Linus Pauling is dead. I felt someone should make an earnest attempt to get at the truth of these matters before this work inevitably passes into ralternative" medical legend. Recent better controlled studies strongly suggest that oral megavitamin therapy is NOT helpful in treating cancer <1> but data often has a hard time when pitted against legend and a need to believe.
One of my interests here was trying to understand the mind of Linus Pauling. None of his quite extravagant claims about orthomolecular treatment have been sustained, despite extensive later research. How could such a distinguished scientist have left normal scientific caution behind when looking at ralternative" claims?
Dr A Hoffer MD PhD FRCP (C)
Suite 3-2727 Quadra St
Dear Dr Hoffer,
I am writing a piece on the available research into orthomolecular treatment of cancer and would
value input from you. I have the paper published by you and Linus Pauling on the survival of
the cancer patients you treated ( " Hardin Jones Biostatistical Analysis of Mortality Data for
Cohorts of Cancer Patients -----". Journal Of Orth molecular Medicine 5(3):143-154,1990); it
concludes that most cancer patients will survive 21 times longer with large doses of vitamin C
and other vitamins.
I am also aware of the more recent study of your patients with non-metastatic breast cancer <1>
which has less favourable findings. It would help me weigh the significance of all the available
research if you could provide more information about your own study and I would value any
other comments you may have.
I am especially curious as to why Dr. Pauling engaged in a complex, difficult to understand, and
rarely adopted Hardin Jones analysis of your data, when it is clear by any standard that the
patients you treated with vitamins fared very much better than those who were not.
Of the 11 untreated patients with female cancers, only one (with an extremely rare Fallopian
tube cancer) survived beyond twelve months; only two survived beyond six months. Of the 40
orthomolecular-treated patients, only one died before six months, 33 of the 40 lived for longer
than twelve months, and many lived over five years. Such differences have a vanishing small
likelihood of being due to chance.
The results with 61 non-gynaecological cancers are equally striking.
Pauling may have contributed to mainstream disinterest in this work by obscuring the obvious
and not addressing adequately the single critical question the study arouses: "Were the patients in the treated and untreated groups truly comparable?" Pages are devoted to rather abstruse statistical discussion, but a mere sentence or two to the clinical characteristics of the patients. The ostensible aim was to obtain meaningful mean survival times for groups with very different numbers of surviving patients, when that finding already spoke far more loudly than any such contrived figures would. I suggest a possible explanation for this peculiarity later.
With respect to the adequacy of the controls, I fully accept that there was no conscious attempt
to select or reject patients for orthomolecular treatment, and that the groups had similar ages and more or less similar usage of other treatments. But did they have truly comparable disease status?
We are simply told, presumably in Pauling's words, --
"The patients had cancer in an advanced stage, mainly untreatable (with little hope
by the physician that any conventional treatment would have more than a palliative
--- but we are not given biopsy results or information about the extent of cancer in any case,
even though patients are individually detailed in other respects.
Nevertheless, there is no doubt about the seriousness of the cancers in the untreated (control)
patients. The short life spans of these poor folk attest to widespread or otherwise very advanced terminal cancer.
It is with the treated patients that niggling concerns arise, even from the scant information
For example, a patient who had a spinal cord tumour of unspecified type previously treated by
surgery is included. He apparently saw you for orthomolecular treatment a "long" time later,
suggesting the late recurrence of a low-grade neurological or meningeal tumour. One wonders
what current findings made this patient comparable to more obviously terminal patients, and
what type of malignancy, if it was malignant, it was.
You also treated three patients with leukaemia, who had an "N" for no treatment in the other
treatment column. They were accepted for treatment immediately after diagnosis and died 9, 25, and 75 months later. It seems that they had no conventional treatment at all. Apart from the
concern that these patients might have fared better with standard chemotherapy, there are four
patients with leukemia in the treated group, but none at all among the controls. How was it
adjudged that these leukemia patients were in the same desperate straits as the controls with
solid cancer, especially if several had never had any prior treatment, and it was not considered
important to specify the type of leukemia, with its marked effect upon prognosis?
The most impressive of your findings is that nearly half the treated patients are described as
"well" at periods ranging from two and a half to over eight years. Does "well" mean that they
were now cancer-free? If so, and these 101 patients initially had cancer as advanced as it
obviously was in the controls, you should by now have a remarkable series of at least thirty or
forty patients whose lung, liver or bone secondaries, or other advanced cancer masses melted
away with your treatment. More than anything, such a "best case" series would prove the
effectiveness of orthomolecular treatment. No other treatment even approaches such success
with assorted advanced cancers.
Do you have such a series of cases? Or did cancers "in an advanced stage" somehow stay
stationary without ill effects for up to eight years? While unlikely, such a finding would be
fascinating, and equally easy to document.
A remaining possibility is that despite Pauling's assurances, some of your patients did not have
clear-cut clinical or radiological evidence of advanced cancer. Might patient, relative, or
general practitioner opinion on the seriousness of the disease, sometimes mistaken in my
experience, have been occasionally accepted without objective verification? Might some
patients with favourable prognoses have self-selected for your treatment, as insurance against
cancer recurrence? Were patients who were referred primarily for psychiatric care included in
the study if found to have had prior treatment for cancer of uncertain current status? I can
understand you would not be excluding such patients from orthomolecular treatment, but were
they also carefully excluded from the study?
I dare to raise these somewhat provocative questions because a nervousness about this aspect of the subject matter might explain Pauling's approach. For in a strict "Hardin Jones Biostatistical Analysis of Mortality Data", one which accepts his basic premises, it would not necessarily affect mortality rates significantly if patients with indeterminate cancer status or at different stages of the disease were lumped together.
Hardin Jones thought death rates from cancer were unaltered by either the passage of time or the medical treatments of his day (indeed, probably worsened by them). He based his methods on the belief that --
"the death rate is the same at all times after establishment of the cancer state.
--------------. The death rates early in the disease are strikingly
similar to death rates in the middle course or at the end of the
disease, so the patient with cancer, while he has a high death risk,
does not decrease his chance of survival by already lived some time with
his disease" <3>.
That may have been approximately true for most of the studies that Hardin Jones had available
to him. He was working with mainly short term studies containing the very high mortality rates
of the more advanced cancers of early last century <mostly published in the 1930s- PJM>. His most detailed statistics related to breast cancer, and that was possibly the only cancer for which he had figures extending beyond five years of follow-up, But breast cancer is atypical among common cancers in that deaths are still regularly occurring at ten years. With many other cancers, e.g. of colon, death rates revert back to that of the normal population by five years, indicating that any survivors are to all intents and purposes then cured. Hardin Jones actually came across just such statistics, suggesting cure, e.g. of cancer of the cervix,, but dismissed them by suggesting that it they were due to "cases of a milder disease than is generally reported" <4>
Pauling, nevertheless, accepted Hardin Jones' premises, stating:
" Hardin Jones reported (1956) that published mortality data for cohorts of
similar cancer patients indicate the death rate to be constant: -----".
He also seemed to be acknowledging the need to deal with some heterogeneity of subjects in the following ---
"The analysis of mortality data for a cohort with many survivors is more difficult
than that for a cohort with few survivors, especially if the cohort is heterogeneous <my
emphasis PM> ; it is, however, possible to carry out such an an analysis by application of
the Hardin Jones principle."
The implication is that Pauling realised, without actually saying so, that the clinical material
was too variegated to stand up to direct comparisons of the untreated and treated groups such as I have made above. He presumably thought that Hardin Jones formulae could still dissect out
trustworthy results, while reducing the usual burden in clinical research of demonstrating
equivalence of comparison groups. Hardin Jones would have approved, although I suggest that
that statistician would himself insist that the mixes of cancer types in the groups be more uniform and that the groups be somewhat larger.
If such thoughts were even merely lurking at the back of Pauling's mind , he would be
overlooking an important qualification from Hardin Jones himself -----
"There is abundant evidence, however, for a terminal phase of cancer, and the
death rate of this phase is in the range of 1000 deaths per 1000 individuals per
-----which is uncomfortably close to the death rates in your two control groups. Regardless of
what Pauling may have thought, Hardin Jones' formulae could not possibly compensate for
controls weighted towards terminal patients. Nor could they adjust for a treated group
containing a preponderance of patients with better starting prognoses, whether due to prior
treatment or more favorable pathology.
A somewhat weaker concern about the study is that nearly all the treated patients seem to be
classed as either "well" or "deceased", without the number of patients unwell with active disease that might be expected in a series of 101 cancer patients with at least a fifty per cent mortality rate. Nevertheless I can well understand that this was not your primary employment and that surveillance of these patients would not be as thorough as it might be in any oncological study.
I need to know what you do know about these patients. I am sure others will have asked similar
questions. Is the information I seek available anywhere, or can you answer the questions I have
raised directly? I am also interested in any light you can shine upon Pauling's thoughts at this
most important point in the history of alternative cancer treatments..
I am placing this letter on public record.
Peter Moran MB, BS, BSc(Med), FraCS, FRCS(Eng)
1. Lesperance ML, Olivotto IA, Forde N, Zhao Y, Speers C, Foster H, Tsao M, MacPherson N,
Hoffer A. Mega-dose vitamins and minerals in the treatment of non-metastatic breast cancer: an
historical cohort study. Breast Cancer Res Treat. 2002 Nov;76(2):137-43. Medline
2. Page 314 of the paper "Demographic Consideration of the Cancer Problem" presented to the
Section of Biology of the New York Academy of Science on January 9, 1956.
4. P323 Ibid.
Back to Index