The Ice Cream Paradox: Scientists Are Perplexed

One of the strangest findings in nutrition research is what's called the ice cream paradox. It's the idea that people who eat ice cream regularly sometimes appear to have a lower risk of diabetes and heart disease. Which sounds completely backwards. Ice cream is considered to be junk food. It's supposed to be the food you remove from your diet if you want to improve your metabolic health. And yet this finding has shown up in multiple observational studies, not just one weird paper. So naturally this raises the question, is ice cream actually not as bad as people think it is? Or is it another example of how messy these studies can be? So in this video I wanted to go through the research myself and see, have I been missing out by not eating ice cream or is there an actual explanation underneath it all? The first time scientists faced this awkward situation of ice cream not being associated with worse health was back in the 80s, when Harvard researchers started tracking the diets of 41,000 middle-aged men and followed them up for 12 years. They saw that general dairy intake was associated with a lower risk of diabetes and specifically consuming ice cream was associated with a lower risk as well. Compared to less than once a month, consuming ice cream twice or more per week was linked to a 22% lower risk. That's not a very large decrease but there's a dose response. Consuming ice cream once a week was linked to a 15% lower risk and one to three times a month was 16% lower risk. Now the issue is that these findings were only borderline statistically significant with a p-value of 0.06. So it's just above the 0.05 cutoff point for statistical significance and if it's not statistically significant the association doesn't exist. For this study it is fair to say that there's a trend towards a lower risk of diabetes but it's only borderline statistically significant. Another 2006 study on 40,000 women saw the same results. Higher dairy intake was associated with a lower risk of diabetes. Those in the highest quintile of dairy intake had a 21% lower risk. well below the 0.05 cutoff point and with ice cream intake two or more times per week was linked to a 12% lower risk. The p-value for this finding however was 0.17 which is not statistically significant and even less so than the previous study on men. Yogurt however had a statistically significant effect which was driving the overall findings of the study. Those who consumed yogurt twice or more times per week had an 18% lower risk of diabetes with a p-value of 0.03 much lower than the 0.05 cutoff. Based on these studies it appears the relationship between dairy intake and a lower risk of diabetes is predominantly driven by yogurt intake and ice cream is just dragging along. This was confirmed by a 2014 meta-analysis of both of these studies plus a third one. They found that only yogurt intake not other forms of dairy and not dairy intake overall was associated with a lower risk of type 2 diabetes. Interestingly this meta-analysis did find ice cream's association with reduced diabetes to be statistically significant. An intake of two or more per week being associated with a 26% lower risk with a p-value of less than 0.001. However the authors of the study argued that this result could have been because of what's called reverse causation. It means that healthy people kept eating ice cream a few times a week and because they were healthy they didn't develop diabetes. Whereas people who developed high blood pressure, high cholesterol, high blood sugar or any other medical problem cut back on their ice cream intake because they were otherwise with worse health. So they stopped eating ice cream and therefore in the statistics it seems that ice cream intake is associated with a lower risk of diabetes whereas in reality most people who eat ice cream were healthy. Another 2016 meta-analysis of 22 cohort studies found that dairy intake especially yogurt was associated with a reduction in type 2 diabetes risk. The dose response analysis on ice cream showed that an ice cream intake of about 10 grams a day was associated with a 19% lower risk. This finding was also statistically significant. However because there was a lot of heterogeneity between the studies as you can see from these different lines which each represents the findings of different studies these results might be again more so to do with reverse causation as in the previous analysis. In addition to diabetes the ice cream paradox is found in heart disease papers as well. Ice cream intake being associated with a lower heart disease risk at an intake of two or more servings per week. These findings are found to be statistically significant when controlling for a healthy eating pattern, smoking, body weight, alcohol, physical activity and other comorbidities. Now epidemiology is very messy obviously. Could it be that ice cream is not as bad as people think or could it be because of the reverse causation that I mentioned? Healthy people eat ice cream and unhealthy people avoid ice cream in these studies. To eliminate the risk of reverse causation we can look at what does changing your dairy intake do to your future risk of diabetes. Then we could maybe see if unhealthy people stop eating ice cream. This 2019 analysis of three cohort studies looked at how does one's diabetes risk change with changing one's dairy intake over the next four years compared to looking at baseline dairy intake. The participants were followed up for up to 26 years and the authors linked four-year changes in dairy intake to type 2 diabetes risk in the subsequent four-year period. What they found was that increasing yogurt consumption was associated with a moderately lower risk of type 2 diabetes whereas increasing cheese consumption was associated with a moderately higher risk among US women and men. The yogurt finding supports previous findings but the cheese finding contradicts the previous findings. In terms of ice cream the participants who decreased ice cream consumption had an 8% higher risk of type 2 diabetes but those who increased ice cream consumption had a similar risk to those who kept their intake stable. So this study exactly answers the question we asked. Is the ice cream intake being linked to lower diabetes risk because of reverse causation or is it because ice cream being good for you? The answer is because of reverse causation mostly. As you can see people who cut back on their ice cream intake they had a higher risk of diabetes which doesn't make sense. If you remove the ice cream you see an increase in diabetes risk well that's because you remove the ice cream because you had an increased risk you are already with a poorer metabolic health whereas the people who kept their ice cream intake the same or even increased it slightly they didn't see an increased risk because their health was kind of okay they didn't have metabolic disease. Overall if you analyze the studies more deeply you can see that ice cream doesn't have a causal protective effect against diabetes or heart disease. It's not literally protecting you against heart disease or diabetes similar to other healthy foods. The more plausible explanation is reverse causation. People often reduce their ice cream intake when they get diagnosed with diabetes heart hypertension or heart disease whereas people who don't have these diseases don't reduce their ice cream intake. So it makes it seem that cutting back on the ice cream increases your risk but people reduce their ice cream because they have already a high risk. Reporting bias may also contribute because foods like ice cream are prone to underreporting or misreporting in observational nutritional studies but ice cream doesn't fall into the umbrella of healthy foods so health conscious people who exercise and eat a very clean diet are less likely to eat ice cream in the first place. So it actually counterbalances the idea that ice cream would be protective. So in my opinion the most likely explanation to the ice cream paradox is reverse causation. Reverse causation can be found in many other types of examples as well. In the example of certain biomarkers and certain lifestyles. For example in some studies being overweight is seen to be protective against certain diseases but that's because people who are malnourished drag the statistics down. They increase the risk making the scene that being lighter weight is bad for you but it's just because people who are super malnourished they have a higher risk of frailty. So obesity and being overweight is definitely not good for you it's reverse causation. However the evidence for yogurt is quite consistent that eating yogurt regularly specifically low-fat yogurt is associated with lower diabetes risk. The mechanisms for the benefits of yogurt are also quite interesting and plausible. Yogurt's probiotic bacteria are proposed to improve lipid profile antioxidant status and cholesterol metabolism which could help with insulin sensitivity and diabetes risk. The studies repeatedly point to fermented dairy and probiotic lactic acid bacteria as the main reason yogurt looks different from other dairy foods. People who eat yogurt also get less weight gain and better satiety because yogurt is high in protein and calcium that reduce fat absorption so you get less calories from that and because of the high protein you eat less of it. For ice cream the only potential is some protein and some calcium. It's not fermented which is why the effects for ice cream are often not statistically significant unlike for yogurt. The most charitable interpretation of this is that ice cream maybe it's not as bad as other types of comfort foods. Maybe it's a you know a little bit healthy but based on the studies it doesn't still provide a reasonable protective effect against diabetes or heart disease. Maybe it's just not as bad as people think it is. I'm not talking about Ben & Jerry's these super high fat very high sugar ice creams. I'm talking about mostly unflavored or without extra added sugar ice cream. If you want to learn more about how to improve your biological age and overall health span without spending hours at the gym without spending thousands on supplements and get yourself to top one percent of health without obsessing about it then check out my online community Youth Span Society. It's where you can learn more detailed programs and protocols from me. Link in the description.