Friday, August 17, 2012

Meat-Ology: A Brief Glance at the Latest Data on The Link Between Red Meat, Cooking Techniques & Prostate Cancer

Image 1: The mass media is already sharpening its knifes - finally some sensational headlines on the negative side effects of red meat! But is your prostate really at risk? Should you already tell your urologist to sharpen his surgical knife and schedule your operation? I have compiled the data for you, now it's up to you to decide.
I guess most of you will have noticed the upheaval around the latest egg-bashing study by Spence et al. that made the rounds a couple of days ago. In view of (1) the authors' outright statement that their intention was to go against the hazardous disregard for the potential dangers of cholesterol laden eggs that was spreading so rapidly (sounds more like pamphlet than like a study, right?), (2) the fact the authors present data with standard deviations that are larger than the actual differences in carotid plaque areas between the different quantiles of weekly egg intake and (3) the superior (lower total and LDL and higher HDL) cholesterol levels in those participants who ate the most of the oh-so-unhealthy eggs, I decided to leave it to a brief comment on the SuppVersity Facebook Wall, instead of a full blogpost (Mark from Mark's daily apple was not as lazy as me, so if you are interested check out his critique). When I hit onto a ScienceDaily article on its "red meat counterpart" today, though, I felt compelled not to leave you hanging with the sensational "red meat causes prostate cancer" reports of which I am pretty sure that they will soon pop up on all mass media outlets.

What's that about red meats, burgers and steaks?

Notes on the methodology: The scientists evaluated the raw data for their two cohorts, Los Angeles County site (LA), and San Francisco Bay Area (SFBA), individually.

The scientists excluded over- (>6,000kcal) and undereaters (<600kcal) and applied two different models with...
  • model 1 adjusting for age (years), BMI (<25, 25-29, ≥30), total calorie intake (kcal/ day) and family history of PCA (yes/ no) and  
  • model 2 adjusting for age (years), BMI (<25, 25-29, ≥30), total calorie intake(kcal/ day), family history of PCA (yes/ no), total fat intake (g/ day), alcohol consumption (g/ day), cigarette smoking (pack-years),total fruit consumption (g/ day), total vegetable consumption (g/ day)
Data was acquired by "professional interviewers" in home interviews.

Dietary data was queried by food frequency questionnaires and commonly used cooking module that would elicit the degree of doneness and browning (participants were given photos to pick from).
According to what I bet you have already or will soon read (and probably even see on TV), researchers from the UCLA have finally identified why red meat is so bad for you - it's the frying! Strangely, though, it's only bad if you fry hamburgers and what's more, you better be Hispanic if you want to make sure it will give you prostate cancer (see image 3)... if we take the ScienceDaily article on the matter as a reference, we will yet find this information only, if we are not satisfied with the "gist" the author gives us as an introduction:
"Research from the University of Southern California (USC) and Cancer Prevention Institute of California (CPIC) found that cooking red meats at high temperatures, especially pan-fried red meats, may increase the risk of advanced prostate cancer by as much as 40 percent."
And you bet that 90% of the readers will do just that: They read "40% increase in prostate cancer risk" and tell themselves either"I will finally go vegan, like Steve jobs" or "F*** it, I love my red meat!" The 10% who are still undecided will then go on and find a few neat quotes like
"The observations from this study alone are not enough to make any health recommendations, but given the few modifiable risk factors known for prostate cancer, the understanding of dietary factors and cooking methods are of high public health relevance"
- Mariana Stern, associate professor of preventive medicine at the Keck School of Medicine of USC
Which show us that (contrary to the mass media) the scientists are well aware of the exploratary nature of their study, so take the following elaborations of the study not as a critique of Stern, Joshi, Corral, et al. whose study is exemplary in the way they are finally going beyond the "all red meats are created equal" approach that is so prevalent in epidemiological studies, but as an example of why third-hand information (even if this is my hand ;-) is never absolutely reliable.

What statistical significant findings were there actually?

I won't bore you any longer with additional moralizing and try my very best to refrain from too much interpretation... ah, let's just take a look at what we have
  • age is a factor for LA residents, not for San Francisco residents (older = more cancer)
  • cigarette smoking makes a significant difference only in LA
  • race is highly predictive in both LA and the San Francisco bay area
  • high dietary fat intake increases the risk significantly in LA and borderline significantly in SFBA
  • alcohol intake, fruit, veggies, and dairy don't make a difference
So far for the baseline characteristics, now what about the hazard ratios for meat consumption? What's going to make your prostate grow? Well, let's start with what won't make it grow, or I should say, "parameters that did not show significant increases in hazard ratios (HR)":
Image 2: The total amount of red meat (irrespective of how it was prepared) did not stat. significant increase prostate cancer risk
  • the absolute amounts of red meat, hamburgers, steaks, poultry, processed meat, bacon, sausages and homemade gravy regardless of how it's processed
  • red meat, when it is grilled, broiled or baked
  • hamburgers, when they are well done
  • steaks, when they are well done or high temperature cooked
  • poultry, when it is grilled, oven broiled, pan-fried, baked, high temperature cooked or well-done
Similarly, in a non-genomic analysis across all ethnicities, none of the purported meat mutagens, i.e. HCA 29amino919methyl1969phenylimidazo[4,59b]pyridine (PhIP), PAH Benzo9a9pyrene (BaP), HA 2-amino-3, 8-dimethylimidazo[4, 5-fquinoxaline (MelQx), DiMelQx or the total mutagenic activity had statistically significant influence on the the hazard risks (they did for G/C genotype and C/C genotype, but the results are inconclusive, because median intakes of were associated with increases of up to 50%, while very high intakes were associated with decreases in of up to 40% specifically in the C/C genotype of the PTGS2 SNP (rs20417), a gene that controls inflammatory processes and has been implicated in the etiology of various cancers)

So where are those 40% increases then? I see only statistically insignificant stuff!

You do in fact have to get your glasses out to find p-values with p < 0.05, which mean that the chances that the given increase in the risk of developing prostate cancer are just coincidence are below 5% and therefore "statistically significant"... now I did just that for you, got my glasses and hope I did not overlook one of the few that are actually in there - and since this is what the laypress considers the most important, I even drew a graph for ya ;-)
Figure 1: There are exactly three items in the study, for which the scientists observed statistical significant increases in advanced (an only advanced not total!) prostate cancer risk - red meat cooked at high temperatures (low = 0.0-9.4, medium = 9.4-16.9, high = 16.9+), hamburgers cooked at high temperatures (low = 0.0-4.9, medium = 4.4-7.9, high = 7.9+) and red meat low = 0.0-5,0, medium = 5.0-9.8, high = 9.8+ that was fried in a pan; intakes in g/1,000kcal diet (data based on model 2 from Joshi. 2012)
At first it my appear schizophrenic that red mead cooked at high temperatures, which obviously includes pan-fried red meat, has it's own category and is associated with a higher increase in hazard ratios than it's pan-fried counterpart . If you do yet take a closer look at the quantiles you will realize that the +40% increase in advanced cancer risk in the high red meat group is brought about by intake of more than 16.0g/1,000kcal, the +30% increase from pan-fried meat, on the other hand, is the result of only 9.6g/1,000kcal per day of pan-fried red meat.

Image 3 (getty images): The hispanic guy in the back could increase his  PC risk by 40, 80 and 160% (low, medium, high intakes), almost 2x more than non-hispanic whites; African Americans are "immune" against high-temperature cooked burgers -20, -10 and +10% nonsignificant changes in HR (p = .725)
Without knowing exactly how the scientists processed their data, which is by the way based on assessments of photographs of the food the study participants ate (make up your own mind about that!?), I will just treat it as a statistical artifact that the model yields a +40% increase in advanced cancer risk for total "high temperature-cooked meat", when two of the individual sub-categories, grilled meat and oven broiled meat did not yield a statistical significant association and the third one, i.e. pan-fried meat had a very clear association, albeit for a lower increase in hazard risk. Be that as it may, since I am about to conclude this mini-overview with the benefits that were spit out by the very same statistics, and artifacts like that are actually one of the reasons it is never prudent to think of statistical associations or correlations as if they were causations, I won't dig further into it.

Things you won't hear in the news: "Bad meat" that will reduce your prostate cancer risk

If we thus put faith in the negative news, we should yet also appreciate the positive ones of which you probably won't read anywhere else but at the SuppVersity. Actually quite a pitty, as even among the "bad red meats" there appear to be some that are not so bad after all - at least if you prepare them correctly:
  • poultry - 10-30% reduced hazard ratios (HR) for advanced prostate cancer for Q2, Q3 and Q4 intakes of poultry with the highest (=30% reduction) for Q4, and a daily intake of more than 35.7g / 1,000kcal
  • home made gravy - 30% reduced HR for local prostate cancer risk (borderline significant p = 0.06) after adjustment for age, BMI, calorie intake, family history of prostate cancer, total fat intake, alcohol, smoking, fruit and veggie consumption.
I guess by now you will have read enough to see the shades of grey and even the little white there is, so I'll leave it up to you to decide, whether or not you are from now on living on poultry and home made gravy, only. As funny as it may sound: If we took at the results for a "T" that and not veganism would namely be the "ultima ratio" ;-)
Current Age 10 Years 20 Years 30 Years
30 0.01 0.32 2.49
40 0.31 2.52 8.30
50 2.30 8.30 14.40
60 6.62 13.36 16.11
70 8.50 11.97 N/A
Table 1: Percent of U.S. Men Who Develop Prostate Cancer over 10-, 20-, and 30-Year Intervals According to Their Current Age, 2005–2007 (according to Altekruse. 2010)
Putting things into perspective: I know that I said, I would refrain from overt commenting and just give you the facts, but I won't let you get away without at least some quantitative considerations that will help you to put data like +30% advanced prostate cancer risk from the consumption of more than 9.6g / 1,000kcal energy intake per day into perspective. I mean, let's be honest 9.6g/1,000kcal per day? If we assume that the average piece of meat you would eat has ~120g that would mean that even if you had a piece of pan fried red meat every 12 days, only, you would still end up in the medium intake category and have a +20% increase in cancer risk.

In view of the fact that the average omnivore male human being probably eats way more meat it may no seem as if 90%+ of us were doomed to develop advanced prostate cancer and that may well be the case if we would not be talking about  if we were not talking about increases in hazard risk from a certain baseline. If we take the 2010 study by Altekruse et al. as a baseline (see table 1) our +30% increase in hazard risk would thus mean that a man who is now 40 years old would have a +0.1% increase of developing cancer in 10 years, a +0.8% increase to suffer from prostate cancer when he is 60 and a +2.5% greater risk with 70 (please remember that this data is at best of qualitative nature as calculations like this are outright unscientific and that's not just because the "average American" who is captured in the SEER data from 1975-2007 will definitely be a red meat eater ;-)
  • Altekruse SF, Kosary CL, Krapcho M, Neyman N, Aminou R, Waldron W, Ruhl J, Howlader N, Tatalovich Z, Cho H, Mariotto A, Eisner MP, Lewis DR, Cronin K, Chen HS, Feuer EJ, Stinchcomb DG, Edwards BK (eds). SEER Cancer Statistics Review, 1975–2007, National Cancer Institute. Bethesda, MD, based on November 2009 SEER data submission, posted to the SEER Web site, 2010.
  • Joshi AD, Corral R, Catsburg C, Lewinger JP, Koo J, John EM, Ingles S, Stern MC. Red meat and poultry, cooking practices, genetic susceptibility and risk of prostate cancer: results from the California Collaborative Prostate Cancer Study. Carcinogenesis. 2012 Jul 20.
  • Spence JD, Jenkins DJ, Davignon J. Egg yolk consumption and carotid plaque. Atherosclerosis. 2012 Aug 1. 
  • University of Southern California - Health Sciences (2012, August 16). Pan-fried meat increases risk of prostate cancer, new study finds. ScienceDaily. Retrieved August 17, 2012, from­ /releases/2012/08/120816170404.htm#.UC3I26LCdzU.facebook