Sunday, February 16, 2014

The Diet Heart Hypothesis: Part 2

The Difference between Men and Women 

Most of the trials were done in men probably because CVD is thought to be the ‘fat, middle-aged man’s disease’, but CHD is the number one cause of death in women too [1] [2] and yet only two of the trials included women (FMHS and MCS) 

Supportive of DHH for CHD Events
Supportive of DHH for CHD Mortality
Supportive of DHH for Total Mortality

In these two trials there’s a fairly consistent pattern: men do better than women/women do worse than men (depending on how you look at it).  Since the FMHS had more confounders in favour of the experimental group and the MCS appeared to be well controlled, this suggests that seed oils less likely to confer benefits to women and are more likely to cause harm.  This is particularly unfortunate as women tend to be health conscious and so may end up using more seed oils than men. 

The Difference between CHD Events and Mortality 

The two meta-analyses that supported the DHH both found that replacing SFA with PUFA reduced CHD events, but not total mortality (this minor detail should be quite embarrassing for those who use either meta-analysis as evidence to replace SFA with PUFA) 

Supportive of DHH for CHD Events
Supportive of DHH for CHD Mortality
Supportive of DHH for Total Mortality
Hooper (Cochrane)
RR = 0.82
CI = 0.66 to 1.02
RR = 0.92
CI = 0.73 to 1.15
RR = 1.02
CI = 0.88 to 1.18
(+5% PUFA)
RR = 0.81
CI = 0.70–0.95
RR = 0.80
CI = 0.65–0.98
RR = 0.98
CI = 0.89–1.08
* Probably significant with more trials/numbers

Anyway, this particular pattern of results (reduced CHD events, but no difference in either CHD mortality or total mortality) was found in some of the trials, including: ACC (which actually found both increased CHD mortality and total mortality); LA Vets, MRC, FMHS (women only) and MCS (men only and just a 10.5% difference in CHD events).  Two main explanations for this pattern of results could be: 

1) A negative effect of either SFA, TFA or junk food on CHD events but either a less negative, neutral or positive effect on mortality; or a positive effect of PUFA (omega 6, omega 3 or both) on CHD events, but either a less positive, neural or negative effect on mortality.  At least in DART, ‘fish advice’ lowered CHD mortality to a greater extent than CHD events (in fact there were more non-fatal CHD events in the ‘fish advice’ group), suggesting that this pattern of results may not be due to fish or omega 3 [3].  Besides that I’m not sure, but my main guesses would be: (1) the decrease in TFA in the experimental group, which would be expected to reduce blood coagulability and thrombosis, probably preventing minor, non-fatal heart attacks and angina; (2) a difference in post-prandial angina from a meal higher in refined foods and lower in beneficial nutrients (like vitamin E and phytonutrients) 

2) A lack of blinding and bias in the assessment of CHD events and/or CHD mortality.  Bias can occur where the person who assesses CHD events and CHD mortality isn’t blinded and then, whether deliberately or unconsciously, incorrectly assesses CHD events and/or deaths from CHD in favour of a particular group.  In LA vets and MCS the assessors were blinded, but in the ACC, MRC and FMHS they weren’t.  In MRC there was no difference between ‘definite’ CHD events and CHD mortality, but a decrease in ‘probable’ and ‘possible’ CHD events in the experimental group, suggesting the possibility that bias may have influenced the results.  Also, the investigators mentioned that: “one possibility is that the men in the control group themselves may have felt that they were having less active treatment, and might have been psychologically motivated both to complain more and more readily of cardiac symptoms” [4] 

* I’m not saying that replacing SFA with PUFA reduces CHD events because ACC, LA Vets and FMHS had several confounders in favour of the experimental group, the 10.5% decrease in events in MCS was not significant even with 4393 men and there are some issues with the MRC (see above)

PUFA and Non-CVD Morbidity and Mortality 

Mozaffarian, et al and people that use their meta-analysis should be asking: ‘why is it that replacing SFA with PUFA (apparently) reduces CHD events and CHD mortality, but not total mortality*?’  In most of the trials there were only a small number of non-CVD deaths (~0-5), which is to be expected as most of them were secondary prevention trials in middle-aged men.  There were only 5 trials where there were more than 10 deaths from non-CVD causes and probably not a coincidence that 4/5 were primary or primary + secondary prevention and these were the larger trials as well 

Non-CHD Mortality
Experimental Group
Control Group
Anti-Coronary Club
Los Angeles Veterans
Finnish Mental Hospital Study (M/F)
Minnesota Coronary Survey (M/F)**
Diet and Reinfarction Trial
* Deaths per 1000 person years
** Minnesota had a high turnover rate and I don’t know what the per 1000 person year figures would be so the figures may not be an accurate representation 

Given the results of Mozaffarian, et al and that they included all of these except the ACC, it shouldn’t be surprising that on average, deaths from non-CHD causes were higher in the experimental group.  The trials generally don’t have good information on the cause of the non-CHD deaths.  However, in the LA Vets there was a near statistical significant increase in cancer in the experimental group (31 vs. 17, p = 0.06), and MCS has a detailed listing on the causes of death, but no clear pattern emerged. 

Chris Masterjohn points out a few things about the LA Vets and potential problems with high PUFA diets in general (highly recommended reading) [5]:
  • There isn’t much long term data on very high PUFA diets and no societies eat very high PUFA diets, except the traditional Inuit (a point also made by the LA Vet investigators)
  • High PUFA diets seem to increase the vitamin E requirement and perhaps more so over time with further accumulation of linoleic acid in adipose
  • The increase in cancer deaths and non-CHD mortality in the experimental group only took off after 2 and 7 years respectively
All together this suggests that potential harmful effects of high PUFA diets may take at least several years to develop, so we can’t really extrapolate long term health effects from these short term trials. 

Finally, what about non-CHD morbidity and quality of life?  Unfortunately there’s almost no data on this.  In LA Vets there was an increase in gallstone formation in the experimental group (which was correlated with adherence to the experimental diet within the experimental group) [6], but FMHS found no difference in gallstone formation [7].  So, does replacing SFA with PUFA increase gallstones?  I’m not sure, as I don’t much about gallstones, but I do know that the FMHS was less well controlled than the LA Vets.

1 comment:

  1. Does a high-PUFA diet cause gallstones? This could be a result of linoleic acid uncoupling the expression of LDL receptors from the regulation of cholesterol synthesis. Discussed here: