Preliminary Review by Arpad Pusztai of the 1998 Russian rat feeding study
using Monsanto GM potatoes
PRELIMINARY REVIEW
of the GM Potato rat feeding study:
"Medical-biological investigations of transgenic potatoes, resistant to
the Colorado beetle (under agreement with Monsanto Co.)
Russian Academy of Medical Sciences, Institute of Nutrition, Moscow, 1998.
Signed off by VA Tutelian, Deputy Director.
Physiological, biochemical and morphological investigations in rats. Full Report
61 pp,
excluding raw data.
by Arpad Pusztai
I have briefly examined both the Russian and English language versions of this report. It was not a happy reading! This is not a peer-reviewed study, and if it had been peer reviewed it would not have been accepted. I am not going into much detail because if I can believe what the Tables tell me (I am still in shock) that the rats' diet only contained 4% protein regardless of whether they were eating control or test diets, I would have wasted my time. Unbelievable! Let me give you some "tasting" notes!
Tables 1 & 2. It is not clear whether the protein, etc. content is expressed as g/100 g dry weight or wet weight? There are no error bars!
Throughout the tables it is not clear whether they give the mean of the results
with standard error or standard deviation. But in any case it was a
lousy potato with 2.5% (or the even worse 2.3% for the GM) protein. If
it was for wet weight, without the water content of the potatoes, we
cannot convert it to true protein and also cannot compare the GM and
non-GM because these two potatoes may not have the same water content.
By the way, if these two potatoes were substantially equivalent I would
eat my hat! Looking at Table 6 just confirms my feelings. We only have Monsanto's
word for it that the two potatoes are related.
Where are Tables 3,4 and 5? (They are in the Russian version of the text, but were omitted in error from the English version.)
Table 8. Total protein content in control 1 diet is just over 4% and dry weigh content some 43%. What and where is the rest?
Table 9. The composition of control 2 diet is not much better. But where is the composition of the test diet?
ALL DIETS ARE DEFICIENT IN PROTEIN, FAT, CALORIES ETC. ACCORDING TO THE DATA IN THE STUDY THE RATS WERE KEPT UNDER NON-PHYSIOLOGICAL CONDITIONS AND THEREFORE NO VALID CONCLUSIONS CAN BE DRAWN FROM THE RESULTS. OR TO USE THE RS' FAVOURITE PHRASES: THE DESIGN AND EXECUTION OF THESE EXPERIMENTS AND THE CONCLUSIONS WERE ALL DEEPLY FLAWED.
Now comes even more:
Tables 10 & 11. No feeding rate is given for control 1 diet without
which one cannot calculate the feed conversion ratio for this control
diet. This is particularly serious because the growth of these control
rats is far superior to rats in control 2 and to the test groups.
Someone ought to have told these Russian nutritionists that they cannot
compare the weight of the rats in the control and test groups when the
animals' feed intake is different. It has to be given as feed conversion
ratios (i.e. g weight accretion per g feed eaten) and when one compares
these the feed conversion ratios of the performance of the Test (GM)
group IS HIGHLY SIGNIFICANTLY INFERIOR TO THAT OF THE
CONTROL 2 GROUP (0.7g vs 1.0g per 1 g diet for the first week or 0.73g vs 1.03g
for the 4th week for the GM vs non-GM groups, respectively). It is also notable
that the feed conversion ratios fluctuate from week to week drastically, for
example it drops to 0.16g vs 0.18 g for the second week, GM vs non-GM, respectively.
This is a lousy piece of work!
The same goes for the absolute organ weights in Table 12. Incidentally, the error bars for some of the organ weights are huge. This in fact applies to the whole study. In several instances when it is stated in the study that the differences were statistically not adequate (?), I think it must mean significant, by looking at some of the results even roughly I am not so sure (even without primary data). For example in Tables 19 and 20 some of the difference between data for the GM and non-GM groups (haemoglobin, etc) looks perilously close to being significant.
Without knowing how well the histology was done and actually seeing the sections I have grave reservations for taking the authors' reassurance seriously that there were no differences between the groups.
After the one month data I am not going to go on giving my views on the 6 months results because the same criticisms should also apply to them as I described above. Although I had a look at the six months figures, the diet was so poor that even in a relatively short period of four weeks any prolongation of the non-physiological conditions in the trial rats makes all the results not better than the tales of Hans Andersen. Should the effects visible after one month be perpetrated or accentuated over longer periods? I just simply don't know what to expect in the long run. You get figures but what they mean is beyond my speculation.
However, I must make one more general point. Means, particularly with large error bars are meaningless. In a true design (as I previously expounded this in the evaluation of the MON 863 Monsanto study) every rat has to be followed through the experiment individually and the progression of the results will have to be compared because means can hide a lot of highly meaningful differences. Moreover, although it is not described in the paper, but in this study the rats were probably not housed individually and therefore it would have been impossible to follow their progress individually.
CONCLUSION: This Monsanto GM potato is causing problems after one
month even at this stupidly low, 12%, incorporation in the diet.
For comparison in our 1998 experiments (at least in two) we had nearly 7O%
incorporation. To show up genuine and highly significant differences when you
have so little of the GM potato in the diet must make us really to sit up and
take notice. Its effect on growth retardation is very marked and highly significant.
You also have to take into account the very low voluntary intake of these diets.
This is almost certainly due to the low nutritional value (low protein, fat,
etc). For comparison in our long term experiment with cooked potatoes the voluntary
intake of our rats was around 2O g/day/rat.
So one cannot be surprised that the growth rate here was poor. However, I don't
think that people who are not nutritionists/physiologists can truly appreciate
that from a non-physiologically conducted experiment one cannot draw valid
conclusions.
I feel that I wasted a lot of my time looking at the first part of this report, and I don't want to waste any more. Let me just sum it up like this:
(1) Even with the poor design and execution the Russian researchers did not
make a proper and reliable case out to declare that these GM potatoes could
be used instead of the non-GM potatoes in the diet without significant health
risks for the consumers; and
(2) the Russian regulators were not right to allow these GM potatoes into the
human food chain in Russia (or anywhere else for that matter!).
Arpad Pusztai
28th February 2007