Many people consider me an expert on ketogenic and carnivore diets, and have asked me my opinion on their relevance to COVID-19.

I have been studying emerging evidence and ideas from the crisis. In general, this evidence has been interpreted by many low carb diet advocates to strengthen their view that metabolic syndrome is the root of all modern disease and that treating it with a low carb diet will thereby reduce disease burden of all kinds.

While I do think that treating metabolic syndrome will reduce disease burden, and that a low carb diet is currently the most effective treatment for metabolic syndrome, I also think this focus is blinding us to some important considerations. I think the standard carbohydrate-insulin hypothesis (that high carbohydrate intake causes metabolic syndrome) is inadequate as an explanation, and some of the evidence emerging during the crisis is consistent with other ideas.

If we merely see metabolic syndrome as the battle of the day, then it looks like susceptibility to COVID-19 is just another tragic outcome of the world not legitimising low carb diets as a treatment. However, if we look deeper, and recognise that metabolic syndrome may have other causes, then we can actually learn something from COVID-19.

Hemoglobin

There is a new hypothesis that COVID-19 might not kill by inducing a respiratory disorder that then lowers oxygen, but rather that it attacks hemoglobin, damaging its ability to carry oxygen and carbon dioxide. This then puts pressure on the lungs that results in the observed lung inflammation. If this is true, then the cause of death is tissue hypoxia, and lung inflammation is secondary.

See this Twitter thread for the paper this idea is based on, and a description for layfolk:

If this hypothesis is true, then it may explain several associations in a different way from how most low carb diet researchers are conceiving of it. To explore this, I will focus on:

  • metabolic syndrome comorbidities
  • immune system dynamics
  • respiration

I am not so much trying to assert that this new hypothesis is true, only that we can learn more by considering more hypotheses.

Comorbidities

Among those who are dying from COVID-19, a large number had some pre-existing evidence of metabolic syndrome. For example, a report from Italy said that over a third had diabetes, over a third had heart disease or stroke, and over three quarters had hypertension. Similarly, in France, it was reported that more than half had heart disease and more than a third had diabetes.

Because there is ample evidence that low carb diets can treat metabolic syndrome, many presume that this will transitively reduce the likelihood of death due to COVID-19. This has been the main focus of discussion among advocates of low carb diet theories.

However, this view downplays the potential importance of the information that the association between metabolic syndrome and COVID-19 death rates may be due to compromised hemoglobin.

One important marker for diabetes is high levels of HbA1c — glycated hemoglobin. While many in the low carb community see high levels of HbA1c as a direct result of uncontrolled hyperglycemia (which it may sometimes be), it's also possible that hemoglobin is more glycated because it is forced to stay overlong in the bloodstream. This kind of result happens, for example, in anemia, where HbA1c is elevated [Kim2010], [Sin2012], [Ade2014], [Mas2019] perhaps because there isn't enough new hemoglobin in circulation (as stated in https://www.timeofcare.com/falsely-lowered-a1c-and-falsely-elevated-a1c/).

It's possible that HbA1c is elevated in diabetes because of an underlying tissue hypoxia, as described, for example by Norouzirad et al. [Nor2017]. Insofar as this is the case, the focus on blood sugar as a cause of diabetes is a red herring.

In terms of the hemoglobin hypothesis, higher HbA1c could be creating susceptibility because of the glycation damage or indicating a pre-existing strain on RBCs, or both.

We will return to the implications of this below.

Immune system

Many people who are dying of COVID-19 are experiencing dramatic levels of inflammation in the lungs. As I mentioned above, the hemoglobin hypothesis reverses the presumed causality. It says that impaired hemoglobin makes it impossible to get enough oxygen and carbon dioxide exchange, which is putting stress on the lungs.

If, as most presume, lung inflammation is causing inability to attain enough oxygen to live, then the right approach would be to dampen this inflammation. So low carb advocates may focus on emerging evidence that ketogenic diets might have anti-inflammatory properties. If the hemoglobin hypothesis is true, this would be irrelevant.

I already find this "excess inflammation" explanation problematic. Permit me a short defence of the immune system to explain why.

Inflammation: good or bad?

This may sound too simple to be true, but a good rule of thumb is to remember that inflammation and other immune activity is a response to infection or damage. As such, when you see it, you should assume that it is there for a reason, and that reason should be the target of your search to improve health outcomes, not the inflammation itself.

People routinely attribute disease to inflammation, instead of what the inflammation is responding to. Often inflammation is responsible for uncomfortable symptoms; healing can be painful, and if prolonged can be its own source of damage. Blocking it can make you think you're better off. Indeed, it can be a justifiable tradeoff to slow healing in the interest of reducing pain, but we should do it with awareness.

To add to the confusion, it has become common in some widely accepted medical theories, to attribute disease states to an immune system that is "overactive", or inflammation that is "out of control". I don't find these interpretations plausible, and this puts me, unfortunately, in a contrarian position. I hypothesise that in these cases the immune system is responding to some source of damage that has been overlooked, rather than randomly, erroneously responding to nothing at all. In a subsequent post, I will look more deeply at the "cytokine storm" idea, for example as described here: https://www.nytimes.com/2020/04/01/health/coronavirus-cytokine-storm-immune-system.html, and explain why I’m uncomfortable with it. My discomfort may be merely a lack of understanding on my part, but perhaps by writing about it, I’ll gain clarity.

Regardless, when we are looking for the effects of some intervention on inflammation or other immune system markers, we ought to be very careful about the context and how we interpret the results.

For example, if we see inflammatory markers go down, it's important to know whether that's because we have prevented the need for inflammation, or if the need remained and we've merely blunted our own defenses. Likewise, if inflammation markers go up, we should distinguish between an increase in damage that warranted higher inflammation, or an augmentation of the immune system's inflammatory response.

As an example of the former kind of error, it has now been acknowledged that preventing the inflammation that causes discomfort after exercise, also prevents the beneficial adaptation to exercise. It is the healing of exercise-induced damage that makes you stronger.

An example of the latter kind of error is addressed in an article by researcher Brianna Stubbs. It bears on the question of ketosis and the immune system.

How does ketosis affect inflammation and immune function?

In "New Study Claims Exogenous Ketones Increase Inflammation - Why the Title Is Misleading", Dr. Stubbs explains that when ketone esters are given in the presence of a bacterial toxic load, inflammation is enhanced, whereas in the absence of that toxic load, it is not. Similarly, Lussier et al. [Lus2016] show that mice on a ketogenic diet have increased immune responses to gliomas. This immune system enhancement seems to be exactly what one would want.

One the other hand, there are studies suggesting that ketogenic diets can reduce inflammation in a beneficial way. Take, for example the work of Kim et al. [Kim2012], where they used a common animal model of multiple sclerosis. They induced encephalomyelitis in mice by injecting them with toxins. The resulting inflammatory response has been implicated in compromised memory function. Mice who were treated with a ketogenic diet had less memory decline. The authors attribute this to the lower inflammation, but we should be asking: was the inflammation lower because ketosis downregulates inflammatory processes, or does it improve healing so the inflammation is less necessary? The result on memory would be the same, but the overall result would be very different.

Ketosis is neuroprotective, particularly in hypoxic conditions. This has been known for decades. See, for example, [Kir1980], [Woo2015], [Yan2017], [Zha2017]. So in this case, it seems that lower inflammation is likely a result of lower damage; also exactly what one would want.

In other words, the work that makes the best case for ketogenic diets being anti-inflammatory probably only shows that it prevents damage, not that it blocks inflammation. If that's right, it could mean that ketogenic diets are not actually "anti-inflammatory", as some believe. But I would consider this absence of inflammation suppression from ketogenic diets to be a feature, not a bug.

Respiration

COVID-19 appears to kill by causing pneumonia and acute lung inflammation so that people cannot get enough oxygen. Because ketogenic diets are useful in hypoxia, a low carb researcher might well focus on this use of low carb nutrition as an intervention. And it wouldn't be wrong!

Ketogenic diets greatly reduce respiratory quotient (RQ). RQ is a clever way to infer what fuels are being used on a whole system basis. You can get an estimate of how much your cells are relying on glucose metabolism as opposed to fat metabolism, because the chemical reactions involved have different ratios of oxygen taken up to carbon dioxide released. Fat oxidation produces less carbon dioxide per oxygen consumed than glucose.

It's been experimentally shown that artificially ventilated patients with respiratory failure need less ventilation on low carb diets [alS1989]. Lowering RQ can likewise extend the ability of COPD patients to walk [Eft1992]. This kind of result seems practical and useful in the context of the crisis right now.

However, if the hemoglobin hypothesis is correct, then this intervention would be symptomatic treatment, and while it may help delay the consequences of impaired oxygen allowing the body more time to fight, it is not directly solving the problem in the way it would otherwise appear. Instead the question becomes, how do we cope with RBC compromise?

Having robust RBCs in the first place may help. Reducing metabolic syndrome seems like it could be helpful for that, in accordance with either reduced glycation damage or by reduced hypoxia whatever the causal link.

But when we think about it in this light, it points back to directions of causality that we might not have considered before. What if, as suggested above, metabolic syndrome is a disease of hypoxia due to impaired hemoglobin? This would explain its connection, for example with high ferritin and iron overload. What if ketogenic diets help metabolic syndrome because they help replenish RBCs?

There is another way that a ketogenic diet might actually be of help here. There is some evidence that fasting (and by extension a ketogenic metabolism which is in many ways metabolically identical) actually protects RBC replenishment [Che2014]. This has been used to explain the benefits of fasting during chemotherapy.

More ideas

The hemoglobin hypothesis has more explanatory power that might be overlooked if we focus only on how a ketogenic diet might help.

It may explain the emerging positive results with chloroquine, which has effects on red blood cells. Incidentally, chloroquine also reduces HbA1c, although I'm not clear on how quickly or by what mechanism. See, e.g., [Rek2010], [Hag2014].

It may explain recent reports that African Americans are contracting the virus and dying from it at higher rates. The linked article suggests that there is a cultural behaviour of not properly social distancing, and I've seen others suggest that African Americans are just more likely to have metabolic syndrome, but what if instead it is related to sickle cell traits?

There has been an association reported between ABO blood type that may be relevant [Zha2020].

Finally, the same Twitter account as I've quoted above, has also pointed out that "HAPE", high altitude pulmonary edema, has similar symptoms to COVID-19, which is consistent with the idea that the virus is inducing a shortage of red blood cells:

Bottom Line

We don't and can't yet know whether a ketogenic diet will help mitigate the risks associated with COVID-19. If you are using one of those diets to improve your health, there is no reason I know of that it would cause extra risk. To the contrary, it seems that it may help with respiratory function. Certainly it seems prudent to do whatever you can to reduce metabolic syndrome if that's an issue for you. If COVID-19 acts through attacking RBCs, there may also potentially be some replenishment benefit from a ketogenic metabolism. All of that is true.

However, we can learn a lot more about COVID-19, about metabolic syndrome, and maybe even about the mechanisms of low carb diets, if we let go of our assumptions about how they work and follow the evidence.

Acknowledgements

Thank you to Zooko for helping me to write this and providing me with many of the interesting sources of evidence here. Mistakes, of course, are mine.

References

[Ade2014]Adeoye, Segun, Sherly Abraham, Irina Erlikh, Sylvester Sarfraz, Tomas Borda, and Lap Yeung. “Anaemia and Haemoglobin A1c Level: Is There a Case for Redefining Reference Ranges and Therapeutic Goals?” 7, no. 1 (2014): 5.
[alS1989]al-Saady, N. M., C. M. Blackmore, and E. D. Bennett. “High Fat, Low Carbohydrate, Enteral Feeding Lowers PaCO2 and Reduces the Period of Ventilation in Artificially Ventilated Patients.” Intensive Care Medicine 15, no. 5 (1989): 290–95. https://doi.org/10.1007/bf00263863.
[Che2014]Cheng, Chia-Wei, Gregor B. Adams, Laura Perin, Min Wei, Xiaoying Zhou, Ben S. Lam, Stefano Da Sacco, et al. “Prolonged Fasting Reduces IGF-1/PKA to Promote Hematopoietic Stem Cell-Based Regeneration and Reverse Immunosuppression.” Cell Stem Cell 14, no. 6 (June 5, 2014): 810–23. https://doi.org/10.1016/j.stem.2014.04.014.
[Eft1992]Efthimiou, J, P J Mounsey, D N Benson, R Madgwick, S J Coles, and M K Benson. “Effect of Carbohydrate Rich versus Fat Rich Loads on Gas Exchange and Walking Performance in Patients with Chronic Obstructive Lung Disease.” Thorax 47, no. 6 (June 1, 1992): 451–56. https://doi.org/10.1136/thx.47.6.451.
[Gol2019]Goldberg, Emily L., Ryan D. Molony, Eriko Kudo, Sviatoslav Sidorov, Yong Kong, Vishwa Deep Dixit, and Akiko Iwasaki. “Ketogenic Diet Activates Protective Γδ T Cell Responses against Influenza Virus Infection.” Science Immunology 4, no. 41 (15 2019). https://doi.org/10.1126/sciimmunol.aav2026.
[Hag2014]Hage, Mirella P., Marwa R. Al-Badri, and Sami T. Azar. “A Favorable Effect of Hydroxychloroquine on Glucose and Lipid Metabolism beyond Its Anti-Inflammatory Role.” Therapeutic Advances in Endocrinology and Metabolism 5, no. 4 (August 2014): 77–85. https://doi.org/10.1177/2042018814547204.
[Kim2010]Kim, Catherine, Kai McKeever Bullard, William H. Herman, and Gloria L. Beckles. “Association Between Iron Deficiency and A1C Levels Among Adults Without Diabetes in the National Health and Nutrition Examination Survey, 1999–2006.” Diabetes Care 33, no. 4 (April 1, 2010): 780–85. https://doi.org/10.2337/dc09-0836.
[Kim2012]Kim, Do Young, Junwei Hao, Ruolan Liu, Gregory Turner, Fu-Dong Shi, and Jong M. Rho. “Inflammation-Mediated Memory Dysfunction and Effects of a Ketogenic Diet in a Murine Model of Multiple Sclerosis.” Edited by Pablo Villoslada. PLoS ONE 7, no. 5 (May 2, 2012): e35476. https://doi.org/10.1371/journal.pone.0035476.
[Kir1980]Kirsch, J R, L G D’Alecy, and P B Mongroo. “Butanediol Induced Ketosis Increases Tolerance to Hypoxia in the Mouse.” Stroke 11, no. 5 (September 1980): 506–13. https://doi.org/10.1161/01.STR.11.5.506.
[Lus2016]Lussier, Danielle M., Eric C. Woolf, John L. Johnson, Kenneth S. Brooks, Joseph N. Blattman, and Adrienne C. Scheck. “Enhanced Immunity in a Mouse Model of Malignant Glioma Is Mediated by a Therapeutic Ketogenic Diet.” BMC Cancer 16, no. 1 (May 13, 2016): 310. https://doi.org/10.1186/s12885-016-2337-7.
[Mas2019]

Masuch, Annette, Nele Friedrich, Johannes Roth, Matthias Nauck, Ulrich Alfons Müller, and Astrid Petersmann. “Preventing Misdiagnosis of Diabetes in the Elderly: Age-Dependent HbA1c Reference Intervals Derived from Two Population-Based Study Cohorts.” BMC Endocrine Disorders 19, no. 1 (February 12, 2019): 20. https://doi.org/10.1186/s12902-019-0338-7.

[RBC clearance might be impaired by whatever impairs macrophages.]

[Nor2017]Norouzirad, Reza, Pedro González-Muniesa, and Asghar Ghasemi. “Hypoxia in Obesity and Diabetes: Potential Therapeutic Effects of Hyperoxia and Nitrate.” Oxidative Medicine and Cellular Longevity 2017 (2017). https://doi.org/10.1155/2017/5350267.
[Rek2010]Rekedal, Laura R., Elena Massarotti, Rajesh Garg, Radhika Bhatia, Timothy Gleeson, Bing Lu, and Daniel H. Solomon. “Changes in Glycated Hemoglobin after Initiation of Hydroxychloroquine or Methotrexate in Diabetic Patients with Rheumatologic Diseases.” Arthritis and Rheumatism 62, no. 12 (December 2010): 3569–73. https://doi.org/10.1002/art.27703.
[Sin2012]

Sinha, Nitin, T.K. Mishra, Tejinder Singh, and Naresh Gupta. “Effect of Iron Deficiency Anemia on Hemoglobin A1c Levels.” Annals of Laboratory Medicine 32, no. 1 (January 2012): 17–22. https://doi.org/10.3343/alm.2012.32.1.17.

"Further studies showed that HbA1c levels were higher in patients with iron deficiency anemia and decreased significantly upon treatment with iron"

"However, in stark contrast to previous studies, in our study, the HbA1c levels were found to be significantly lower in patients with iron deficiency anemia than in the controls. Moreover, the HbA1c levels increased after treatment, which had not been reported in any previous study."

[Woo2015]Woolf, Eric C., Kara L. Curley, Qingwei Liu, Gregory H. Turner, Julie A. Charlton, Mark C. Preul, and Adrienne C. Scheck. “The Ketogenic Diet Alters the Hypoxic Response and Affects Expression of Proteins Associated with Angiogenesis, Invasive Potential and Vascular Permeability in a Mouse Glioma Model.” PLoS ONE 10, no. 6 (June 17, 2015). https://doi.org/10.1371/journal.pone.0130357.
[Yan2017]Yang, Qi, Min Guo, Xun Wang, Yanxin Zhao, Qi Zhao, Hongyan Ding, Qiang Dong, and Mei Cui. “Ischemic Preconditioning with a Ketogenic Diet Improves Brain Ischemic Tolerance through Increased Extracellular Adenosine Levels and Hypoxia-Inducible Factors.” Brain Research 1667 (July 2017): 11–18. https://doi.org/10.1016/j.brainres.2017.04.010.
[Zha2017]Zhao, Ming, Xin Huang, Xiang Cheng, Xiao Lin, Tong Zhao, Liying Wu, Xiaodan Yu, Kuiwu Wu, Ming Fan, and Lingling Zhu. “Ketogenic Diet Improves the Spatial Memory Impairment Caused by Exposure to Hypobaric Hypoxia through Increased Acetylation of Histones in Rats.” PLOS ONE 12, no. 3 (March 29, 2017): e0174477. https://doi.org/10.1371/journal.pone.0174477.
[Zha2020]Zhao, Jiao, Yan Yang, Hanping Huang, Dong Li, Dongfeng Gu, Xiangfeng Lu, Zheng Zhang, et al. “Relationship between the ABO Blood Group and the COVID-19 Susceptibility.” MedRxiv, March 27, 2020, 2020.03.11.20031096. https://doi.org/10.1101/2020.03.11.20031096.