I have explained in various posts why this idea does not appear to be correct (1, 2, 3), and why, after extensive research, the insulin hypothesis of obesity lost steam by the late 1980s. However, I recently came across two experiments that tested the hypothesis as directly as it can be tested-- by chronically increasing circulating insulin in animals and measuring food intake and body weight and/or body fatness. If the hypothesis is correct, these animals should gain fat, and perhaps eat more as well.
In the first study, published in 1980 by Dr. Dennis A. Vanderweele and colleagues, rats were implanted with mini-pumps delivering insulin at a steady rate throughout the day and night for 7 days (4). They tested four different doses: 0, 1, 2 and 6 units per day, and measured food intake and body weight. This is a model of chronically elevated insulin reminiscent of what is seen in insulin-resistant people.
They found that all doses of insulin reduced body weight gain, but only the 2 unit dose was statistically significant. They speculated that the 6 unit dose was less effective at reducing weight gain because it was sufficiently high to cause hypoglycemia (low blood sugar), which is a potent trigger for food intake. Total food intake was suppressed at all doses by insulin as well, but this was apparently not due to illness. The authors conclude:
1) insulin limits meal size when blood levels are modestly elevated for prolonged periods of time in the rat, 2) this decrease in meal size is not compensated for by an increase in meal frequency and, hence, total daily food ingestion and body weight gain are reduced, and 3) this effect appears to be a heightening of satiety rather than an induction of illness.The second study, published in 2002 in Nature Medicine by Dr. Ellen L. Air and colleagues, took a different approach (5). The authors used a drug called Cpd2, which mimics the effect of insulin on the insulin receptor, but can be taken orally rather than by injection (6).
They placed mice on a fattening diet (HFD) with or without Cpd2 added to the food, then measured food intake, body weight and body fatness over a 6 week period. Here is what they observed when they measured body fatness:
Consistent with the previous study, Cpd2 opposed weight and fat gain, and suppressed food intake as well, without causing illness. Glucose tolerance and blood glucose levels were also improved in the Cpd2-treated group.
It is worthwhile to note that there are also studies showing that daily injections of long-acting insulin lead to body fat gain in rodents over time (6a, 6b, 6c). However, these studies used high doses of insulin that resulted in hypoglycemia, which as I mentioned above is a potent trigger for food intake and causes a major hormonal response (the "counter-regulatory response"). That's the unique thing about the studies discussed above-- they investigated the impact of insulin signaling on body fatness without involving hypoglycemia, a situation that is more akin to what is observed in people with elevated fasting insulin due to insulin resistance.
The Final Nail
Contrary to what the insulin hypothesis of obesity predicts, chronically elevated insulin if anything seems to oppose weight and fat accumulation in animal models. This likely involves insulin's action in the brain to constrain body fat accumulation.
The evidence suggests that:
- Experimentally preventing the increase in circulating insulin that occurs on fattening diets does not alter the course of fat gain in rodents or dogs (7).
- Experimentally elevating circulating insulin by creating liver insulin resistance does not lead to fat gain in rodents (8).
- Experimentally increasing circulating insulin by infusing it directly into the blood does not cause fat gain in rodents, but instead makes them leaner (above).
- Roughly a quarter of obese humans have normal circulating insulin and normal insulin sensitivity ("metabolically healthy" obese) (9).
- In Pima (Akimel O'odham) native Americans, one of the most obesity-prone populations in the world, and certain other populations, insulin resistance and higher insulin secretion consistently predict less body fat gain over time (10, 11, 12, 12a, 12b). This has not been observed in all populations, but the fact that it occurs in some casts further doubt on the idea that elevated insulin is a central contributor to fat gain.
We need a better hypothesis to explain why low-carbohydrate diets cause fat loss in many obese people, and we definitely need a better hypothesis to explain why obesity arises in the first place. An alternative hypothesis, that insulin opposes fat gain through its action in the brain, is plausible and has received support from a variety of lines of evidence. Insulin resistance in the brain, leading to reduced insulin signaling in neurons, has been documented in animal models of obesity, and likely contributes to fat accumulation. Therefore, the 'insulin hypothesis of obesity' that makes sense is that a reduced insulin (and particularly leptin) signal in the brain contributes to fat gain.