64 replies to this topic

[Imaginarynumber1]

Little Fish, on 21 September 2012 - 09:36 AM, said:

i see a lot of protest about the strain of rat used.
are you all aware that it is the same strain used in other studies used to promote safety of gmo. the main difference in this study (which followed the same protocols as the gmo promoted studies as far as i can tell - read the actual paper for the details) is that the study was carried out for the full life term of the rats, not just 13 weeks as in the case for the existing safety studies such as hammond et al 2004.

the other main protest seems to be the sample size of the groups, again this is not dissimilar to the existing safety studies, so for those that want to dismiss this study based on sample size, you HAVE to dismiss the existing studies as well.

These rats have a very high rate of developing tumors. The longer they are alive the greater the chance is of them developing, too. To use 10 as opposed to 10 or 12 groups of 10 for a control group is ridiculous. It's statistically insignificant. Their "results" mean nothing.

Edit: Personally, I could care less about GMO's. I don't really have a stance on the matter. This is just very bad science.


Here is some information abour tumor development in this type of rat:

Quote

Unspayed female rats are prone to developing mammary tumors, pituitary tumors and other estrogen-dependent tumors as they age. At the time of menopause (450-540 days) mammary tumor incidence increases sharply and pituitary tumors begin to appear. A second increase in mammary and pituitary tumors occurs around the end of the second year (from 600-800 days) (Durbin 1966).
Most of the tumor susceptibility research in unspayed rats has been done on different strains of laboratory rats. Strains and populations differ in how susceptible the females are to developing tumors. Susceptibility may vary widely.
Below are some studies that examine the percentages of female rats of a particular strain that develop mammary tumors. Where reported, the percentage of benign vs. cancerous mammary tumors is included below:
Sprague-Dawley rats:

• 47% of female Sprague-Dawley rats developed mammary tumors. 12% of these tumors were malignant (Solleveld et al. 1986).
  
• 49% of female Sprague-Dawley rats developed mammary tumors (24 our of 49), 8.2% developed mammary carcinomas (4 out of 49) (Hotchkiss 1995)
  
• 71% of female Sprague-Dawley rats developed mammary tumors, of which 18% were carcinomas (Durbin et al. 1966)
  
• 76% of female Sprague-Dawley rats, most of which were benign fibroadenomas (Kaspareitt and Rittinghausen 1999)

For pituitary tumors, different studies have found:

Sprague-Dawley

• 39% of female Sprague-Dawley rats develop pituitary tumors (Kaspareitt and Rittinghausen 1999)

66% of female Sprague-Dawley rats develop pituitary tumors (Hotchkiss 1995)http://www.ratbehavi...umorSpaying.htm

So after the rats go through menopause (450-540 days), the rates of tumors increase and the increase AGAIN after the end of the second year (600-800 days). Hmm... right before their study ends.....
They did not accumulate enough data to rule out the randomness of tumors in these rats.

Bad science.

Edited by Imaginarynumber1, 21 September 2012 - 10:08 AM.

[Little Fish]

Imaginarynumber1, on 21 September 2012 - 09:49 AM, said:

Toxicity is quite different than allergic reactions and works by completely different methods.

it was an analogy, don't take it literally.

Quote

The point is that the rates of cancer and death were the same among the groups.

an animal which is genetically predisposed to a casual agent derived condition is likely to develop the condition independent of dose.

Quote

Not to mention the unmeasured amounts of food the rats were allowed to consume.

but this would also apply to the controls, so why didn't the controls develop illness at the higher rates as well?

Quote

It was the small size of the control that skewed the data towards the result they wanted.

the control groups were the same size as the tested groups. can you show why you think the control group was too small. the study seems to have been done within regulatory minimum guidelines, see table 1.
http://www.iatp.org/...icityreport.pdf

Quote

These rats have a very high rate of developing tumors. The longer they are alive the greater the chance is of them developing, too. To use 10 as opposed to 10 or 12 groups of 10 for a control group is ridiculous

as i understand it, that is the same basis done for the previous studies.
http://www.sciencedi...278691504000547

can you show any previous gmo studies done with substantially higher controls?

[Imaginarynumber1]

Little Fish, on 21 September 2012 - 10:34 AM, said:

it was an analogy, don't take it literally.

It's a poor analogy. It's comparing two completely different things.

Quote

an animal which is genetically predisposed to a casual agent derived condition is likely to develop the condition independent of dose.

Which they did. These rats get tumors all the time. The more they eat the higher there rates of tumors. The longer they live the higher their rates of tumors.

Quote

but this would also apply to the controls, so why didn't the controls develop illness at the higher rates as well?

180 rats is a much larger group than 20 rats (the controls were 1 pair for each sex in the 10 groups of 20, giving you 20 control animals. It does not specify if this was in addition or part of the 200 divided into groups) Either way, the control group is 9 or 10 times smaller that the testing group. I would expect the larger group to have higher rates of tumors if they were just pets.

Quote

the control groups were the same size as the tested groups. can you show why you think the control group was too small. the study seems to have been done within regulatory minimum guidelines, see table 1.
http://www.iatp.org/...icityreport.pdf

The control was 1 pair of each sex in the 10 groups of 20 rats. Again, it does not specify if these controls were part of the 200 or in addition to the 200.

Quote

as i understand it, that is the same basis done for the previous studies.
http://www.sciencedi...278691504000547

can you show any previous gmo studies done with substantially higher controls?

I can't pull up the data for the control size for this study without paying for the article.......  But they do say that they used 400 rats divided into 10 groups of 20, so I can only assume that there were 200 that were tested on and 200 used for control. Again, I do not know for sure since I cannot access the data.



With such low numbers in the control group with the new study, 20 control animals, the results are not in any way statistically significant enough to account for natural tumor development and not significant to draw any conclusions other than these rats get tumors.


Again, I don't really care about GMO's one way or another, so I'm not arguing if they are good or bad, just that this study is not indicative one way or the other.
Apologizes if any of my responses are muddled or I have massive typos. I've been up all night working on a research paper and have yet to sleep.

Edited by Imaginarynumber1, 21 September 2012 - 11:15 AM.

[bmk1245]

Holy crap... It went for 4 pages...

[Little Fish]

Imaginarynumber1, on 21 September 2012 - 10:59 AM, said:

With such low numbers in the control group with the new study, 20 control animals, the results are not in any way statically significant enough to account for natural tumor development and not significant to draw any conclusions other than these rats get tumors.

"Further criticism centred on the size of the control groups. Anthony Trewavas, professor of cell biology at the University of Edinburgh, said the control group was inadequate to make any deduction, and were of no value without knowing the degree of variation in a control group of 90 or 180 rodents.

"These figures for normal appearance of tumours in these rodent lines are surely available and using a line which is very susceptible to tumours can easily bias any result," he warned. "To be frank, it looks like random variation to me in a rodent line likely to develop tumours anyway."

But Antoniou said the two-year experiment followed international OECD guidelines. The rodents in the experiment were divided into 10 groups of 20 animals, with nine of those groups exposed to Roundup or NK603. "Standard practice is for the control group to be matched in size to the experimental groups. The experimental groups were 20 animals and therefore the control group should be 20 animals.

"Trewavas's statement is not correct. From the 20-animal control, you can get a measure of tumour frequency in the control group. You don't need to look at hundreds of animals. If he believes this, then he should also agree that the studies done by others – including industry – are also invalid.

"The key thing is that there are big differences between the tumour frequencies in the control and the experimental groups. Claims that the results are just the result of random variation in a rat line that has a high frequency of tumours is not valid.

"The evidence for this is that the differences between the groups are much larger than the standard deviations of the two groups. In Seralini’s study, the differences are so large that it is not necessary to use a statistical test. It is obvious.

"This study used more rats in test groups, for a far longer duration than any previous investigation employed by industry to obtain approval for this and other GM crop products."

furthermore from the article
"Dr Michael Antoniou, a reader in molecular genetics and member of Criigen – the Committee of Research & Independent Information on Genetic Engineering – has vigorously refuted questions raised by fellow scientists about the robustness of the study..... "The key is that there were both quantitative and qualitative differences in the tumours arising in control and test groups. In the former they appeared much later and at most there was one tumour per animal, if at all.

"In the latter case, the tumours began to be detected much earlier (4 months in males; 7 months in females), grew much faster and many animals had two or even three tumours.

"Many animals in the test groups had to be euthanised for welfare legal reasons due to the massive size of the tumours; none of the control animals had to be euthanised but died in their own time. One should not ignore these biological facts.""

the other objections are also addressed in the article.
http://www.gmwatch.o...gm-cancer-trial

[Imaginarynumber1]

Little Fish, on 21 September 2012 - 11:20 AM, said:

"Further criticism centred on the size of the control groups. Anthony Trewavas, professor of cell biology at the University of Edinburgh, said the control group was inadequate to make any deduction, and were of no value without knowing the degree of variation in a control group of 90 or 180 rodents.

"These figures for normal appearance of tumours in these rodent lines are surely available and using a line which is very susceptible to tumours can easily bias any result," he warned. "To be frank, it looks like random variation to me in a rodent line likely to develop tumours anyway."

But Antoniou said the two-year experiment followed international OECD guidelines. The rodents in the experiment were divided into 10 groups of 20 animals, with nine of those groups exposed to Roundup or NK603. "Standard practice is for the control group to be matched in size to the experimental groups. The experimental groups were 20 animals and therefore the control group should be 20 animals.

"Trewavas's statement is not correct. From the 20-animal control, you can get a measure of tumour frequency in the control group. You don't need to look at hundreds of animals. If he believes this, then he should also agree that the studies done by others – including industry – are also invalid.

"The key thing is that there are big differences between the tumour frequencies in the control and the experimental groups. Claims that the results are just the result of random variation in a rat line that has a high frequency of tumours is not valid.

"The evidence for this is that the differences between the groups are much larger than the standard deviations of the two groups. In Seralini's study, the differences are so large that it is not necessary to use a statistical test. It is obvious.

"This study used more rats in test groups, for a far longer duration than any previous investigation employed by industry to obtain approval for this and other GM crop products."

furthermore from the article
"Dr Michael Antoniou, a reader in molecular genetics and member of Criigen – the Committee of Research & Independent Information on Genetic Engineering – has vigorously refuted questions raised by fellow scientists about the robustness of the study..... "The key is that there were both quantitative and qualitative differences in the tumours arising in control and test groups. In the former they appeared much later and at most there was one tumour per animal, if at all.

"In the latter case, the tumours began to be detected much earlier (4 months in males; 7 months in females), grew much faster and many animals had two or even three tumours.

"Many animals in the test groups had to be euthanised for welfare legal reasons due to the massive size of the tumours; none of the control animals had to be euthanised but died in their own time. One should not ignore these biological facts.""

the other objections are also addressed in the article.
http://www.gmwatch.o...gm-cancer-trial

Even if the rats in the test groups got tumors sooner, it's still statistically irrelevant because of the size of the control group. If I did an experiment like this for class, I would fail because I didn't have a large enough control group get enough data. Of course a group of 180-200 rats are going to have a higher rate of tumors than a group of 20. There's 10 times as much variation there,They also haven't released data for the diet the control group was on. In fact, they released hardly any info on the control group at all.

I understand what they're saying about "if this is the standard an the other tests use it then they have to be thrown out, too". I would agree, but it appears (and I could be wrong) that the Monsanto test used 200 control animals for 200 test animals.

(Edit: I have not looked into any other studies so I cannot comment on them.)

This study did not acquire enough data to determine anything other than these rats get tumors. It still appears to be bad science.
This is way too reminiscent of Wakefield and his crap about the MMR vaccine causing autism so he could sell his vaccine. This Serilini guy even has a book about GMO's coming out soon.

Edited by Imaginarynumber1, 21 September 2012 - 11:40 AM.

[Little Fish]

Imaginarynumber1, on 21 September 2012 - 11:34 AM, said:

Even if the rats in the test groups got tumors sooner, it's still statistically irrelevant because of the size of the control group. If I did an experiment like this for class, I would fail because I didn't have a large enough control group get enough data. Of course a group of 180-200 rats are going to have a higher rate of tumors than a group of 20. There's 10 times as much variation there,They also haven't released data for the diet the control group was on. In fact, they released hardly any info on the control group at all.

I understand what they're saying about "if this is the standard an the other tests use it then they have to be thrown out, too". I would agree, but it appears (and I could be wrong) that the Monsanto test used 200 control animals for 200 test animals.

(Edit: I have not looked into any other studies so I cannot comment on them.)

This study did not acquire enough data to determine anything other than these rats get tumors. It still appears to be bad science.

you are ignoring the statements in the post you responded to.

"the two-year experiment followed international OECD guidelines"

"From the 20-animal control, you can get a measure of tumour frequency in the control group. You don't need to look at hundreds of animals"

"The key is that there were both quantitative and qualitative differences in the tumours"

Quote

Of course a group of 180-200 rats are going to have a higher rate of tumors than a group of 20

no thats no correct, the numbers would be expected to be higher, but not the rates.

[Imaginarynumber1]

Little Fish, on 21 September 2012 - 11:41 AM, said:

you are ignoring the statements in the post you responded to.

"the two-year experiment followed international OECD guidelines"

"From the 20-animal control, you can get a measure of tumour frequency in the control group. You don't need to look at hundreds of animals"

"The key is that there were both quantitative and qualitative differences in the tumours"

I'm not questioning anything about they types of tumors, just the frequency.  It would appear from the limited amount of info released about the control group, they were only kept for 13 weeks compared to 24 months for the test groups.

For representative data to eliminate anomalous or extraneous variables you absolutely have to look at a control group that is at least equal to the test group.

Quote

no thats no correct, the numbers would be expected to be higher, but not the rates.

Higher number, not rate. I'm tired.

[Little Fish]

Imaginarynumber1, on 21 September 2012 - 11:53 AM, said:

I'm not questioning anything about they types of tumors, just the frequency.  It would appear from the limited amount of info released about the control group, they were only kept for 13 weeks compared to 24 months for the test groups.

not true, read the study. previous studies used to pass gmo for safe consumption used the same methodology as this new study but only ran for 13 weeks. this study (including the controls) ran for 2 years.

Quote

For representative data to eliminate anomalous or extraneous variables you absolutely have to look at a control group that is at least equal to the test group.

Higher number, not rate. I'm tired.

the control group was the same size as the different test groups. the control group is used to determine the "normal rate", then used to compare to the different groups. the study followed OECD guidelines. you don't need to run the control 10 times. you don't need to run a different control to compare to each test. if the control is big enough to determine natural rate, then you know the rate. is the control big enough? - "From the 20-animal control, you can get a measure of tumour frequency in the control group. You don't need to look at hundreds of animals"

Edited by Little Fish, 21 September 2012 - 12:07 PM.

[Imaginarynumber1]

Little Fish, on 21 September 2012 - 12:04 PM, said:

not true, read the study. previous studies used to pass gmo for safe consumption used the same methodology as this new study but only ran for 13 weeks. this study (including the controls) ran for 2 years.

the control group was the same size as the different test groups. the control group is used to determine the "normal rate", then used to compare to the different groups. the study followed OECD guidelines. you don't need to run the control 10 times. you don't need to run a different control to compare to each test. if the control is big enough to determine natural rate, then you know the rate. is the control big enough? - "From the 20-animal control, you can get a measure of tumour frequency in the control group. You don't need to look at hundreds of animals"

If they were running 20 rats at a time, no, they wouldn't need to do more than 20 in a control. But if your doing 180 rats a the same time, you need at least close180 in your control. You cannot compare your two groups with any statistical significance otherwise.

You need a control at least as large as your test group to eliminate as many variables as you can. I don't know any other way to put this. Their study is flawed.

Edited by Imaginarynumber1, 21 September 2012 - 12:16 PM.

[Little Fish]

Imaginarynumber1, on 21 September 2012 - 12:15 PM, said:

If they were running 20 rats at a time, no, they wouldn't need to do more than 20 in a control. But if your doing 180 rats total, you need 180 in your control. You cannot compare your two groups with any statistical significance otherwise.

You need a control at least as large as your test group to eliminate as many variables as you can. I don't know any other way to put this. Their study is flawed.

again you ignored what you responded to.
there were several tests in the study, read the study.
the test groups were 20 rats. the control group was 20 rats. the control group was as large as the test group. what you have done is add all the different test group numbers up to get a number bigger than the control. if you still don;t get it, then think of the study as several different studies each using the same control group.
this study follows and is within OECD guidelines (i think they state 10 as a minimum).
this follows the same methodology as the previous gmo studies used to pass gmo as safe.

you do have a stake in this despite what you say since you don't know whether you are eating it or not, not to mention your kids and/or parents and friends, you can't control what they eat.

Edited by Little Fish, 21 September 2012 - 12:25 PM.

[Imaginarynumber1]

Little Fish, on 21 September 2012 - 12:23 PM, said:

again you ignored what you responded to.
there were several tests in the study, read the study.
the test groups were 20 rats. the control group was 20 rats. the control group was as large as the test group. what you have done is add all the different test group numbers up to get a number bigger than the control. if you still don;t get it, then think of the study as several different studies each using the same control group.

You can't run the total tests against the one control for significant results.

Quote

this study follows and is within OECD guidelines (i think they state 10 as a minimum).
this follows the same methodology as the previous gmo studies used to pass gmo as safe.

you do have a stake in this despite what you say since you don't know whether you are eating it or not, not to mention your kids and/or parents and friends, you can't control what they eat.

If I have 10 groups of 20 rats and I want to use the TOTAL results from those 10 groups, then I need a control as large as that TOTAL or there are too many variables for my results to be statistically significant in any way.

And if there were SEVERAL test in the study, I cannot run a control of 20 and claim it to be the norm for x y or z amount of test subjects. The data is not significant.

If this was the methodology used in the other studies then they are flawed as well.

This study is bad science.

I do not have stake in it. I do not care if the food I eat is genetically modified.

[Little Fish]

Imaginarynumber1, on 21 September 2012 - 12:39 PM, said:

You can't run the total tests against the one control for significant results.

the purpose of the control is to establish the baseline of "normal" disease to compare the tests with. once you have estblished that baseline, its pointless to do it again each time you run a different test. if you disagree with this, then you are throwing away all the previous research used to pass gmo as safe, throwing away the standard practice and disagreeing with Dr Michael Antoniou, do you think he is lying or unqualified?, why don't you email him - http://www.criigen.o...=276&Itemid=105

Quote

If I have 10 groups of 20 rats and I want to use the TOTAL results from those 10 groups, then I need a control as large as that TOTAL or there are too many variables for my results to be statistically significant in any way.

the control is to measure the baseline. there is no need to measure the same thing again and again, you are not changing any variables in the control. the variables are changed from individual test group to test group, not the control group. the individual test group is the same size as the control group. I don't understand why you don't get this.

Quote

And if there were SEVERAL test in the study, I cannot run a control of 20 and claim it to be the norm for x y or z amount of test subjects. The data is not significant.

is your complaint that the study claims a control of 20 represents a normal baseline for 200? because the study does not do that. the study tests several groups of 20 each group under different conditions, then compares each test group of 20 to the 20 baseline control. i don't understand why you don't get that.

or is your complaint that 20 is not enough to establish a baseline?
"From the 20-animal control, you can get a measure of tumour frequency in the control group. You don't need to look at hundreds of animals"-  Dr Michael Antoniou

[Karlis]

I have not read all posts in this thread, but the following four-paragraph excerpt is from the beginning of an article posted by Imaginarynumber1. What is stated there is clear and unambiguous.

Opponents of genetically modified crops have jumped on the results of a new study, which claims to have linked the consumption of GM maize with the development of tumors in rats -- despite widespread criticism of the research from independent scientists around the world.

The paper, published in the journal Food and Chemical Toxicology, has generated highly tweetable headlines, including the Global Post's alarming announcement that: "Monsanto corn causes tumors in rats, new study finds."

Far below the horrifying pictures of mice with enormous tumors, some of the most frightening articles embed caveats. But, experts say, the criticisms should trump the results of the study, which used questionable methods, tumor-prone rats, and poor statistical techniques.

The research also came out of a French research lab, whose anti-GM studies have been harshly criticized in the past.

Has reliably sourced information contradicting the above been published? Are those sources in this thread?

Thanks in advance -- and please pardon my not having read all the posts here, which would have answered my question.

Karlis

Edited by Karlis, 21 September 2012 - 04:31 PM.

[Little Fish]

Karlis, on 21 September 2012 - 04:06 PM, said:

Has reliably sourced information contradicting the above been published? Are those sources in this thread?

yes, post#50
here again
http://www.gmwatch.o...gm-cancer-trial

Edited by Little Fish, 21 September 2012 - 04:38 PM.