No, you're not just using ketones more efficiently

Last time I wrote about this topic, I buried the point inside a long article. So I'm going to spell it out more clearly.

Many people on Carnivore and other low carb diets have low serum β-hydroxybutyrate. I've heard many insist that that's just because they're "keto-adapted, so they're using it more efficiently". In other words, they are claiming that keto-adaptation allows you to get the benefits of ketosis at lower blood levels, because it's just getting used up so fast it isn't sticking around in the blood as long--but it's really still getting used at the same rate! This interpretation is exceedingly unlikely.

While it is true that at lower concentrations, your skeletal muscle uses more ketone bodies for energy, which is a kind of efficiency, this is not the kind of efficiency you would want, as it's unfortunately directly at the expense of the brain. The evidence strongly supports that if your blood levels are low, your brain is just actually using less. This is really important to understand if you want the neurological benefits of ketosis along with the benefits of low carb.

Contents

The primary determinant of brain ketone body uptake is circulating availability
Ketone body utilisation in the brain is primed in humans
Peripheral use competes with brain use until concentration is high; then not so much
Implausibility of the "more efficient" hypothesis
End-to-end citations

The primary determinant of brain ketone body uptake is circulating availability

In humans, the brain uses ketone bodies at a rate approximately proportionate to concentration in the blood. This has been demonstrated via infusions in keto-adapted [1] and merely overnight fasted people [2]. It is noted in these studies that transporters do not seem to get saturated. In other words, uptake just keeps increasing.

At the same time these infusions in the non-keto-adapted cause cerebral glucose metabolism to fall while oxygen consumption sustains [3], meaning that the brain is reducing its use of glucose when ketone bodies are highly available.

The reverse direction may also be true: falling glucose may increase the uptake of ketone bodies independently. There was a fascinating experiment from Drenick et al in 1972 in which insulin injections drove blood glucose to extremely low levels in subjects that had been fasted and were deeply ketogenic. In one case, glucose fell to 0.5 mM (9mg/dL), which would normally be expected to cause severe neurological impairment and even unconsciousness. But the subjects didn't even have any signs of hypoglycemia, let alone severe impairments. This was attributed to the high ketonemia. I have personally once noticed my own glucose drop precipitously during fasting to under 30mg/dL, with BOHB measurements over 5mM without any symptoms, so this phenomenon can occur even without pharmalogical inducement.

In a further group in the same study, brain uptake of ketone bodies was measured at much higher rates than before the insulin, despite lowered concentrations. In other words, while we know that uptake is proportionate to concentration, when glucose is driven lower pharmacologcally, uptake may further increase. Without further experimentation is it not yet clear whether glucose levels are an independent factor in brain uptake rates in a natural setting, but they might be.

Ketone body utilisation in the brain is primed in humans

The fact that this happens in humans even without a keto-adaptation period is considered noteworthy. It's noteworthy because it contrasts with what happens in rats, who do show progressive increases in transport capacity through adaptation. Rat brain ketone body transporters are saturatable. During starvation, transport capacity increases [4]. By contrast, in a paper arguing for therapeutic ketosis in aging and Alzheimer's, Cunnane and colleagues describe the human brain as "always prepared to burn ketones as soon as they are available; a situation totally consistent with their physiological role as its main back-up fuel." [5]

As you may know, I consider the "back-up" language to be, well, a misinterpretation. And I would even quote the preceding paragraph as supporting evidence:

Two observations indicate that ketone metabolism is a constitutive feature of brain function: (i) The amounts and activities of ketone-metabolizing enzymes in the brain are not changed by glucose status and always exceed the amount necessary to supply the brain’s total energy needs. (ii) During infancy, the brain has an obligatory requirement for ketones.

And the role of ketosis in humans is contrasted with the role of ketosis in other animals directly:

The ability of the brain to use ketones in place of glucose is better in humans than in other omnivores such as the rat, pig, or monkey. Dogs, sheep and pigs are unable to achieve significant ketosis, even during prolonged fasting perhaps because their relative brain size is not very large, thereby permitting gluconeogenesis and glycerol liberated during lipolysis to support brain function without resorting to an entirely different fuel.

These observations support a human brain that is not just ready to adapt to ketosis as a fallback when glucose is deficient, but ready to use it maximally at any time.

Peripheral use competes with brain use until concentration is high; then not so much

Because brain uptake of ketone bodies is dependent on circulating concentration, we can look to production and peripheral use to help determine how much will be left for the brain. Ketone bodies produced by the liver enter a shared pool. So at a given level of production (which is based largely on levels of glucose, fatty acids, and the insulin to glucagon ratio), more peripheral use lowers cerebral uptake.

The rate of removal of ketone bodies by muscle tissue is called metabolic clearance rate (MCR). Early evidence showing a concentration-dependent ketone body sparing effect on MCR comes from Owen and Reichard who measured forearm exchange in 1971 [6] Participants were tested after overnight, 3-day and 24-day fasts. After prolonged fasting, MCR went down, and there was actually a net release of β-hydroxybutyrate across the forearm in most participants. Fatty acids became the primary fuel extracted by the muscle. In other words, in a highly keto-adapted state, the muscles were not just taking up and using fewer ketone bodies, but in some cases actually producing β-hydroxybutyrate from acetoacetate and adding it to the bloodstream.

Exercise complicates the picture somewhat, because working muscle can increase its use of ketone bodies. Interestingly, though, and concordantly with the forearm study, excercise's stimulation of MCR diminishes as circulating concentration rises [7]. The broader picture this supports is that when ketosis reaches a high level, muscles rely more on fatty acids and less on ketone bodies for fuel, allowing more to be available to the brain. This makes sense because increased muscle consumption would diminish the adaptive value of ketogenesis in low glucose contexts.

It's not just concentration, and it's not saturation

Several people on first hearing this have said to me, "Oh, that's probably just saturation." Good thinking, but while it might contribute, that was ruled out as a full explanation. In a different study, Balasse compared fasting subjected with overnight-fasted people given ketone infusions [8]. He found 35% lower clearance in the fasting group at comparable concentrations. This means that something about the adapted state other than the concentrations alone is contributing to the reduction of MCR. Competition from fatty acids, and lower insulin were suggested as possible mechanisms. Different proportions of acetoacetate to β-hydroxybutyrate, and the alkalising effect of sodium acetate were ruled out by further experiments.

Implausibility of the "more efficient" hypothesis

The idea that low blood ketone levels in fat-adapted people reflects becoming "more efficient" elides the important consideration of which tissue would be taking up the slack. If it's muscle tissue, then not only does this indicate that concentrations were lower in the first place--because higher concentrations reduce MCR--but would be counterproductive to the entire point of ketosis, which is brain support. Meanwhile, to argue that it's the brain that's taking it up more efficiently would predict evidence that brain uptake is increased. The actual evidence supports the opposite conclusion: in normal physiological ranges, brain uptake continues to rise with concentrations in keto-adapted and non-keto-adapted people alike, with perhaps an added effect from very low glucose levels.

It's worth noting that the people who most often make the efficiency claim often have mildly elevated glucose, not reduced. The more parsimonious explanation is that in those cases of low ketone body concentrations with or without mildly elevated glucose, the brain is simply using glucose avidly and using only the small amounts of ketone bodies the blood makes available.

End-to-end citations

1. Hasselbalch, S. G., G. M. Knudsen, J. Jakobsen, L. P. Hageman, S. Holm, and O. B. Paulson. “Blood-Brain Barrier Permeability of Glucose and Ketone Bodies during Short-Term Starvation in Humans.” American Journal of Physiology-Endocrinology and Metabolism 268, no. 6 (1995): E1161–66. https://doi.org/10.1152/ajpendo.1995.268.6.E1161.

The blood-brain barrier (BBB) permeability for glucose and beta-hydroxybutyrate (beta-OHB) was studied by the intravenous double-indicator method in nine healthy subjects before and after 3.5 days of starvation. In fasting, mean arterial plasma glucose decreased and arterial concentration of beta-OHB increased, whereas cerebral blood flow remained unchanged. The permeability-surface area product for BBB glucose transport from blood to brain (PS1) increased by 55 +/- 31%, whereas no significant change in the permeability from brain back to blood (PS2) was found. PS1 for beta-OHB remained constant during starvation. The expected increase in PS1 due to the lower plasma glucose concentration was calculated to be 22% using previous estimates of maximal transport velocity and Michaelis-Menten affinity constant for glucose transport. The determined increase was thus 33% higher than the expected increase and can only be partially explained by the decrease in plasma glucose. It is concluded that a modest upregulation of glucose transport across the BBB takes place after starvation. Brain transport of beta-OHB did not decrease as expected from the largely increased beta-OHB arterial level. This might be interpreted as an increase in brain transport of beta-OHB, which could be caused by induction mechanisms, but the large nonsaturable component of beta-OHB transport makes such a conclusion difficult. However, beta-OHB blood concentration and beta-OHB influx into the brain increased by > 10 times. This implies that the influx of ketone bodies into the brain is largely determined by the amount of ketones present in the blood, and any condition in which ketonemia occurs will lead to an increased ketone influx.

[Note: the wording of this abstract is rather unfortunate. They seem to be using the word "transport" in different ways. When they say BOHB transport did not decrease, they are referring to PS--the permeability-surface area product. This measures transfer relative to concentration. So they expect relative transport to go down as transporters are saturated, but it doesn't. Then they suggest that maybe "transport" increase can explain the "transport" decrease, by which they mean transport capacity, but they think it's probably unlikely. Finally they note that influx which one might have thought "transport" would refer to, has increased more than tenfold.]

2. Blomqvist, G., M. Alvarsson, V. Grill, et al. “Effect of Acute Hyperketonemia on the Cerebral Uptake of Ketone Bodies in Nondiabetic Subjects and IDDM Patients.” American Journal of Physiology-Endocrinology and Metabolism 283, no. 1 (2002): E20–28. https://doi.org/10.1152/ajpendo.00294.2001.

With plasma concentration of -HB in the range 0.02–1.74 mol/ml, the brain tissue was found to react to increased availability of ketone bodies by an increased net utilization of these compounds. There was no sign of saturation of this process.

3. Hasselbalch, S. G., P. L. Madsen, L. P. Hageman, et al. “Changes in Cerebral Blood Flow and Carbohydrate Metabolism during Acute Hyperketonemia.” American Journal of Physiology-Endocrinology and Metabolism 270, no. 5 (1996): E746–51. https://doi.org/10.1152/ajpendo.1996.270.5.E746.

brain metabolism and CBF when the blood ketone level is acutely increased in humans. The main findings are as follows: 1) a fivefold increase in P-hydroxybutyrate metabolism, 2) a 33% reduction in glucose metabolism, 3) a 39% increase in CBF, and 4) a small change in the regional pattern of CMRglc.
...
The acute hyperketonemia-induced increase in CMRbohb was counterbalanced by a 33% reduction of CMRglc, but no significant alterations in CMRo, were observed (144 t 22 vs. 135 t 22 umol.100 g-l.min-l, not significant). Thus, without changing the level of oxidative metabolism, acute hyperketonemia induced a partial substitution of cerebral glucose metabolism with metabolism of ketone bodies. The present study thus indicates that the human brain can use ketone bodies during normoglycemia and emphasizes that brain ketone body uptake is regulated by the availability of ketone bodies and not by glucose deficiency.

4. Gjedde, A., and C. Crone. “Induction Processes in Blood-Brain Transfer of Ketone Bodies during Starvation.” American Journal of Physiology-Legacy Content 229, no. 5 (1975): 1165–69. https://doi.org/10.1152/ajplegacy.1975.229.5.1165.

Fed and starved rats were studied on successive days during a 5-day starvation period. The ability of ketone bodies to pass the blood-brain barrier was estimated by single common carotid injections of labeled ketone bodies and water, and results were expressed as the ratio between the normalized activities of tracers in tissue and blood, the brain uptake index (BUI). BUI of D-3-hydroxybutyrate and acetoacetate decreased as their total concentrations increased in the injectate bolus: BUI of D-3hydroxybutyrate decreased significantly from 8% at 0.2 mM to 3--4% at 20.2 mM in fed rats and from 11.570 at 0.2 mM to 6% at 20.2 mM in starved rats, indicating saturation of the uptake mechanism. The BUI of both ketone bodies increased significantly with increasing duration of starvation, indicating adaptation to ketonemia. Enzymatic kinetics explained the uptake behavior of o-3hydroxybutyrate in both fed and starved rats and involved a rise of Km and Vmax during starvation consistent with a doubling of the transport rate at the degree of ketonemia found in starved rats. The uptake of glucose was not influenced by starvation or ketonemia.

5. Cunnane, Stephen, Scott Nugent, Maggie Roy, et al. “Brain Fuel Metabolism, Aging, and Alzheimer’s Disease.” Nutrition 27, no. 1 (2011): 3–20. https://doi.org/10.1016/j.nut.2010.07.021.

6. Owen, O. E., and George A. Reichard. “Human Forearm Metabolism during Progressive Starvation.” Journal of Clinical Investigation 50, no. 7 (1971): 1536–45. https://doi.org/10.1172/JCI106639.

Acetoacetate and β-hydroxybutyrate consumption after an overnight fast was low. At 3 days of fasting with increased arterial concentrations of acetoactate and P-hydroxybutyrate, consumption of these substrates rose dramatically. At 24 days of fasting, despite further elevation of arterial levels of acetoacetate and P-hydroxybutyrate, the utilization of acetoacetate did not increase further and if anything decreased, while five out of eight subjects released P-hydroxybutyrate across the forearm. Acetoacetate was preferentially extracted over β-hydroxybutyrate. At 24 days of starvation, free fatty acids were the principal fuels extracted by forearm muscle; at this time there was a decreased glucose and also ketone-body consumption by skeletal muscle.

Note that the preferential use of acetoacetate and the further conversion of some acetoacetate to β-hydroxybutyrate, supports the interpretation of β-hydroxybutyrate as a storage, brain-directed form of ketone bodies as I discussed in the article linked at the beginning of this one. Acetoacetate is less stable, and plateaus in the bloodstream at much lower levels than β-hydroxybutyrate does.

7. Fery, F., and E. O. Balasse. “Response of Ketone Body Metabolism to Exercise during Transition from Postabsorptive to Fasted State.” American Journal of Physiology-Endocrinology and Metabolism 250, no. 5 (1986): E495–501. https://doi.org/10.1152/ajpendo.1986.250.5.E495.

This study examines the effects of a 2-h exercise of moderate intensity (50% of VO2max) on the tracer-determined turnover rate of ketone bodies (KB) in 21 normal subjects fasted for 16 h, 5 days, whose basal ketonemia ranged between 0.09 and 6.16 mM. The KB response observed at the end of exercise is a function of the initial degree of ketosis. When basal ketonemia is below 0.6 mM, exercise enhances ketogenesis (Ra), the amplitude of this process being positively correlated with KB level. There is a concomitant acceleration of the metabolic clearance rate (MCR) of KB attaining 40-50%. When ketonemia exceeds 2.5 mM, the stimulatory effects of exercise on R, and on MCR become less marked as basal ketonemia rises and are completely abolished or even reversed when initial KB level is higher than 3-4 mM. The pattern of changes in the concentration and in the overall disposal rate of KB were similar to that of Ra. It is suggested that the parallel inhibition of the stimulatory effect of work on hepatic ketogenesis and on muscular extraction of ketones associated with increasing degrees of fasting hyperketonemia has two physiological implications: it maintains the preferential utilization of KB by nonmuscular tissues (presumably the brain) and prevents the development of uncontrolled hyperketonemia, despite the intense catabolic situation created by the combination of exercise and starvation.

8. Balasse, E. O. “Kinetics of Ketone Body Metabolism in Fasting Humans.” Metabolism 28, no. 1 (1979): 41–50. https://doi.org/10.1016/0026-0495(79)90166-5.

The comparison between the fasted and the control subjects at ketone concentrations of 3-10 ~mole/ml showed that on an average, starvation is associated with a 35% decrease in the metabolic clearance rate of ketones. These data suggest that fasting is associated with an impairment of mechanisms for utilizing ketones, this defect contributing to the hyperketonemia of food deprivation.
...
The mechanism of this impairment is unknown. It has been suggested on the basis of forearm studies’ that it might be related to a progressive exclusion of muscle as site of uptake of ketones during starvation. However, the fact that a removal defect is already fully apparent after a few days of fast at a time when muscle consumes appreciable amounts of ketones,’ suggests that other mechanisms are also operative. Another possibility is that the high FFA levels associated with fasting compete with ketones as a fuel. In this respect, it should be noticed that the infusion of AcAc to our control subjects reduced their plasma FFA levels below their preinfusion value, this being possibly responsible for an enhancement of their ketone utilization rate. Finally, the reduction in ketone MCR associated with fasting could result from the decrease in plasma IRI levels that is known to occur under these conditions; it has been shown indeed that insulin promotes the uptake and the oxidation of ketones by the whole animal’ and by the voluntary muscle preparation in vitro.” At first glance, this possibility seems to be ruled out in this study by the fact that the IRI concentration was not significantly lower in the fasting group than in the controls. It could be argued, however, that the well known hypoinsulinemic effect of fasting might have been masked in this study by the wide individual variations in insulin levels observed in both groups.