“Eating a fatty diet could reduce a man’s sperm count by 40%” said the Daily Mail – enough to put every man off his bacon & egg. The Globe and Mail warned similarly: “Eating too much saturated fat may decrease sperm counts.”
The headlines came from a study published in the American Journal of Clinical Nutrition.” High dietary intake of saturated fat is associated with reduced semen quality among 701 young Danish men from the general population.” Unfortunately only the abstract is on free view – I’ve got hold of the full article to see what it’s all about.
As the article title confirms – the study involved 701 Danish men who signed up for military training between April 2008 and June 2010. The men “delivered a semen sample, underwent a physical examination, and answered a questionnaire comprising a quantitative food-frequency questionnaire to assess food and nutrient intakes.” The food questionnaire was intended to review the three months prior to the military training sign up appointment.
The full article states (in the introduction) “We therefore examined the associations between dietary fat intakes and semen quality among 701 young Danish men from the general population, hypothesizing that a high intake of saturated fat is associated with reduced semen quality.” So, the researchers hypothesised that saturated fat intake is associated with reduced semen quality before doing the study. As Einstein said, if you know what you’re looking for, it ain’t research!
The conclusions of the study were: “…men in the highest quartile of saturated fat intake had a 38% (95% CI: 0.1%, 61%) lower sperm concentration and a 41% (95% CI: 4%, 64%) lower total sperm count than did men in the lowest quartile. No association between semen quality and intake of other types of fat was found.”
Table 1 in the full article shows the 701 men allocated into four groups – Quartile 1 (the lowest intake of saturated fat from the questionnaire) to Quartile 4 (the highest intake of saturated fat from the questionnaire). In Quartile 1, saturated fat intake was 11.2% of dietary energy intake or lower. In Quartile 4, saturated fat intake was 15.19% of dietary energy intake or higher. There were 174 men in Quartile 1, 179 in Quartile 2, 170 in Quartile 3 and 178 in Quartile 4.
For each quartile, there is information for a number of other parameters – the usual other things worth looking at – age,smoking, BMI, total energy intake, intake of other fats, physical activity, caffeine intake. There are also some factors particularly relevant to this study – the period of abstinence before the sperm sample was obtained and incidence of sexually transmitted diseases.
Here are the most relevant numbers from Table 1:
|Quartile 1||Quartile 2||Quartile 3||Quartile 4|
|Number in quartile||174||179||170||178|
|SFA as % of energy||<11.2||11.2-13.27||13.28-15.19||>15.19|
|Period of abstinence >96 hr||18||10||11||13|
|BMI <20 kg/m2||10||15||11||23|
|Alcohol intake >21 units/wk||23||25||32||29|
|Weekly current smoking||43||45||49||52|
|Age >20 yr||24||22||21||28|
|Sexually Transmitted Diseases||7||11||9||19|
|Total energy intake (MJ)||8.6||9.4||9.8||10.6|
|Total fat (% of energy)||24||29||33||38|
|MUFA as % of energy||4.2||4.8||5.2||5.5|
|PUFA as % of energy||8.6||10.8||12.3||13.8|
(NOTES: SFA is an abbreviation for saturated fat, MUFA is an abbreviation for monounsaturated fat and PUFA is an abbreviation for polyunsaturated fat. It is not always clear from the table what the numbers are and therefore looking at them relatively across the quartiles is more useful than trying to work out what they say absolutely. The BMI numbers are % of each quartile in each of the 3 BMI categories. So, in Quartile 1, for example 10% of people are <20 BMI, 70% are in the 20-24.9 range and 20% over 25 – this adds to 100% down the 3 rows – as do the other quartiles. The alcohol number seems to be the percentage of people in that quartile who consume more than 21 units a week – self reported. Smoking seems to be % who smoke and STD’s seems to be the percentage that confessed to having had a Sexually Transmitted Disease).
The above tells us the following:
Q4 had higher saturated fat intake (as a % of energy) than Q1 but it also had:
– higher energy (calorie) intake than Q1;
– higher total fat as a % of energy intake;
– higher monounsaturated fat as a % of energy intake;
– higher polyunsaturated fat as a % of energy intake;
– over twice as many people in the underweight BMI category (taking BMI of under 20 as underweight in this case);
– almost half as many in the overweight BMI category;
– higher incidence of higher alcohol unit intake;
– more smokers;
– more older people (over 20);
– almost three times the incidence of sexually transmitted diseases (STDs); and
– 50% fewer people who had abstained for more than 96 hours before the sperm sample.
Table 2 is interesting. This has semen volume, sperm concentration, total sperm count and motile sperm (the latter gives an indication of the quality of the sperm – their ability to move effectively towards an egg) against the different quartiles for total fat, SFA, MUFA, PUFA and even gets down to omega-3 and omega-6 essential fatty acids. Let’s just take the part of the table for SFA vs the sperm measurements:
|Semen volume||Sperm concentration||Total sperm count||Motile sperm|
|SFA >15.19% of energy||3.1||45||128||69|
This tells me that the ‘best’ intake of saturated fat for semen volume, sperm concentration, total sperm count and motile sperm is 11.2-13.27% of dietary energy (a bit precise, but that’s what it suggests). There’s barely any difference between the third and fourth quartiles and the second quartile is ‘better’ than the first. This is notwithstanding all the variables stacked against the quartiles as they go up from 1 to 4 for everything else – alcohol, smoking, STD’s, age, being underweight etc.
Table 3 tries to “take into account confounders”. However, it only tries to take into account BMI, alcohol consumption, smoking and the period of abstinence before the sample. It doesn’t appear to take into account the nearly three times higher incidence of STDs and there being more men in Q4 over 20 than in Q1. Even if all attempted confounders have been perfectly accounted for (and I can’t see how, from the SFA data in Table 2 being unremarkable and the different attributes in Table 1 being significant), surely the difference in age in Q4 and the highly significant difference in the incidence of STDs could alone explain any difference in sperm quality? (Again – not that the difference in sperm quality in Table 2 is much to get excited about).
The numbers in the media headlines come from Table 3, which gives “Adjusted differences in semen quality by percentage intake of total fats and major fatty acid groups from a multiple linear regression analysis.” The article doesn’t give enough detail for readers to be able to follow what has been done to get to this table, but you can see the 38% differential for sperm concentration and 41% differential for total sperm count at the top quartile for SFA intake as a % of total energy.
|Semen volume||Sperm concentration||Total sperm count|
|SFA >15.19% of energy||0.1||-38.0||-41.0|
This doesn’t make sense to me. Table 1 tells us that quartile 1 has every advantage over quartile 4. Hence, if this is appropriately accounted for, differences remaining attributable to saturated fat intake (as opposed to alcohol, smoking, period of abstinence etc) should be tiny.
Table 2 tells us that the raw data for the sperm concentration and quality aren’t that different and that the ‘best’ SFA intake for all measures is 11.2-13.27% of energy (not the lowest SFA intake possible). For sperm concentration, Quartile 2 is 6% higher than Quartile 1 and Quartile 4 is 10% lower than Quartile 1 – no where near a 38% differential and these should reduce after allowing for confounders. For total sperm count, Quartiles 3 and 4 are approximately 20% lower than Quartile 1. Quartile 1 is approximately 7% lower than Quartile 2. Also not in the 40% differential range and also should reduce following allowance for confounders.
Table 3 also no longer mentions motility – that’s the statistic to get excited about if you’re trying to conceive – why was this dropped? Did it not give ‘the right’ answer?
It also makes no sense to claim an association with saturated fat and not total fat or any other fat. Saturated fat cannot be eaten alone. Every single food on the planet that contains saturated fat also contains monounsaturated fat and polyunsaturated fat – there are no exceptions. The extracted numbers from Table 1 confirm that total fat, monounsaturated fat and polyunsaturated fat all increased from Quartile 1 to Quartile 4 and yet we are led to believe that only saturated fat is associated with sperm concentration and sperm count. Not only is this not plausible, no plausible mechanism is offered for any possible explanation for proposed association throughout the article. How can saturated fat intake (alone from other fat intake and total fat intake) impact sperm concentration and sperm count?
If the period of abstinence tells us anything, a much more interesting headline could have been “Men who eat more saturated fat have sex more frequently!”