A Landmark Conference for Ketosis, Day 1

James McCarter
Aug 17, 2018 · 17 min read

UPDATE: VIDEO NOW AVAILABLE AT: https://blog.virtahealth.com/videos-conference-science-carbohydrate-restriction-ketosis/

The Ohio State University hosted “EMERGING SCIENCE OF CARBOHYDRATE RESTRICTION AND NUTRITIONAL KETOSIS”, August 16–17, 2018, will be seen as a watershed event in nutrition research — here is a first draft of history of the sold out event. (Apologies for typos.)

Note that with support from Virta Health, the entire conference is being captured on video. These will be posted in the coming months.

August 16, 2018

EMERGING SCIENCE OF CARBOHYDRATE RESTRICTION AND NUTRITIONAL KETOSIS

Scientific Sessions at the Ohio State University

Welcomes & Introduction

Dr. Ken Lee, OSU

Ken predicts a decrease in carbohydrate consumption and an increase in fat consumption leading to an improvement in overall health. Food Innovation Center.

Dr. Ken Lee, OSU

Dr. Jeff Volek, OSU & Virta Health

Expect major breakthroughs in the next 2 years. “Ketosis research has been living in the shadows for too long.” We want to inspire new research, collaborations, translation, and policy. Organization. Steve Phinney coined the term nutritional ketosis and ketoadaptation. Ken Ford and Ken Lee. Julie Manning — event organizer.

Dr. Jeff Volek, OSU and Virta Health

Dr. Morley Stone, OSU

OSU, Air Force, DARPA, AFRL, human performance for air & space, just joined OSU administration, Commitment — this topic of human nutritional enhancement WILL be a priority for OSU going forward, Peak soldier performance — nutritional approaches (e.g. ketone esters).

Dr. Morley Stone, OSU

Dr. Steve Phinney, Virta Health

Turn your chairs around and face forward — don’t crane your neck for 2 days. For the last decade, Jeff and Steve have been looking for a way to bring silos of ketosis research together. Thanks to food innovation center at OSU for providing the venue to create this conference. UC Davis connection to Dr. Jon Ramsey. Address mechanism of ketosis.

Dr. Steve Phinney, Virta Health

Dr. Jon Ramsey, UC Davis

[note: much of this is published in Cell Metabolism 2017]

Ketogenic Diet and Aging. Interest began with calorie restriction. Why does it extend lifespan in mice? Decreases enzymes of glycolysis in liver even when weight stable. Hagopian Exp Gerontol 2003. Persistent reduction in glycolysis. HK, PFK-1, PK. Poised to use fatty acids and ketones. Upregulation of ketogenesis. Is shift to lipid metabolism beneficial or not? Mammalian energy sensing pathways. BHB is a ligand for G-protein coupled receptors — HCAR2, FFAR3 to decrease metabolic rate and inflammation. BHB and histone acetylation. Histones influence gene transcription. Shimazu Science 2013. Relevance at physiological levels of fasting. BHB inhibits class 1 HDACs (1 & 3) downstream — Foxo2a, BDNF, FGF21, PGC1 alpha. BHB to Acetyl-CoA impacting acetylation (increase) and succinylation (decrease). Direct BHB attachment to proteins? BHB is a signaling molecule, not just a fuel.

Does a ketogenic diet alter longevity and markers of health in mice? Test without inducing obesity or weight loss. Diets with ketosis. Diets with fatty acid oxidation but non-ketosis. Diets — control, low carb, ketogenic. Dietary formulation is KEY!! Control is AIN93. Add lard to increase fat content in low carb and ketogenic. Male C57BI/6 mice. Ketogenic was 90% fat / 10% protein. Experimental design. Cohorts. Interest in healthspan not just lifespan. Measure substrate oxidation and body composition. Ketosis retain lean body mass. Old ketogenic animals retain bone mass and muscle mass in aging. 13.6% increase in lifespan with ketogenic diet. Out to 1,200 days. Lower incidence of histiocytic sarcomas in ketogenic group. C57Bl/6 are sarcoma susceptible. The ketogenic group had fewer tumors. Novel object test — test for memory in older mice. With ketosis, improved performance at 26 months. Did as well as 13 months. Now looking at brain in follow-on studies to understand possible mechanism. Muscle strength — grip strength is better in ketogenic group at 26 months. Hanging wire test is striking — older mice with ketosis retain more muscle endurance and strength. Rearing test too. Locotronic ladder run test too — ketogenic diet mice at 26 months are faster. Lower time to completion. Markers of inflammation — reduction with ketosis in IL-6, CXCL1 and TNF alpha. And reduced variability. No differences in lipids. With ketosis, there was impaired glucose tolerance but same insulin tolerance — seems like a physiological change due to absence of glucose in diet. Mechanism — likely to be multiple mechanisms of action (MOA). One may be inhibition of mTORC1 in liver. Another — a ketogenic diet increases protein acetylation in liver. Another — ketogenic diet alters mitochondrial function and content — variable by tissue.

Does ketogenesis have to be continuous? — Newman & Verdin paper cycled on and off ketogenic diet to avoid weight gain but still saw positive outcomes including gains in memory in old age. With caloric restriction, get cyclical ketosis — 12 hours a day. Key upcoming — fatty acid composition? Impact on age related diseases.

Conclusions — ketogenic diet CAN extend lifespan, health span. Energy intake matters.

Dr. Jon Ramsey, UC Davis

Dr. Jeff Volek, OSU & Virta Health

As a century and a half of the dark ages winds down, we enter the golden age of ketones. Ketones as a maligned metabolite. History of ketosis (see slide). Cahill 1960–70s — starvation studies of ketone biochemistry. Banting. Stefansson. Atkins. Phinney. Little research in 80s except Phinney. Negatives — ketoacidosis association, Ancel Keys, dietary guidelines, surge in obesity.

Jeff’s OSU group does … many things to study ketosis. Controlled feeding studies for many applications including cancer. See 6 posters. Terminology — concentration matters. SAD — <0.2mm BHB, <50g carbs you normalize your ketones to 0.5–4.0, ketoacidosis >10.0. Enhanced reliance on fatty acid oxidation. Doubles. “In ketosis, you double your rate of fat burning. Keto-adaptation is a perfectly normal. Intermittent ketosis was a perfectly normal human state until the agricultural revolution 10,000 years ago, so 98% of human history.” Ketosis — obese weight loss without calorie counting, improved lipid profile, saturated fat levels in blood decline or stay the same. The saturated fat fuel becomes CO2 and water, not staying in circulation. Consistent decline in inflammatory markers. IL-6, IL-8, TNF-a, MCP-1, I-CAM, E-selectin, PAI-1. The more carbs you eat, the more you see an insulin resistant phenotype. Keto-adapted phenotype is the opposite. The diet is anything from a sacrifice. It can be a pleasurable way to enjoy food.

Dr. Jeff Volek, OSU & Virta Health

Dr. Steve Phinney, Virta Health

what does a well formulated ketogenic diet look like? What would a mouse paleo diet look like? 3–5 servings of non-starchy vegetables, nuts & seeds, berry fruit, sauces, etc. adds up to <50g of carbs a day. Vast majority of calories from fat. Food is satisfying — less inclination to snack, less feeling of hunger. Why we talk about grams of carbs rather than macronutrient ratios. Fat calories can come from body fat stores or diet. 10 characteristics of WFKD. Natriuresis of fasting — restricted sodium in ketosis is reduced circulating volume. https://blog.virtahealth.com/well-formulated-ketogenic-diet/

Coach people on what to eat for a lifetime — not a crash diet. Short-term low calorie diets for weight losses have been disappointing — willpower doesn’t last. Diabetes — supposed to be the hardest population because on insulin resistance and hyperinsulinemia. Yet we see huge success at Virta Health in reversing diabetes and sustained weight loss.

A well formulated ketogenic diet — “A well-formulated ketogenic diet does not necessarily adhere to traditional dietary guidelines.” E.g. salt. Saturated fat. 150 grams of fat. Need saturated fat (and mono-unsaturated fat) for fuel. Ketones as a potent fuel and signaling molecule.

Dr. Steve Phinney, Virta Health

Dr. Dominic D’Agostino, USF, IHMC

Funders — department of defense, NASA, Navy, industry — Quest, Ketologic, etc. Concept of Therapeutic Ketosis. Some patented & licensed technology. Prevent oxygen toxicity seizures in closed circuit rebreathers in navy divers, hyperbaric oxygen therapy, NASA NEEMO — decompression chamber to denitrogenate with 100% oxygen. Challenge of oxygen toxicity seizures. Tools — atomic force microscopy, laser scanning confocal, electrophysio, radio telemetry, hyperbaric chambers. Oxygen free radicals as cause of seizures — study in hippocampal slice preparation. 20–100% oxygen environments. Ketogenic diet arose as an anti-seizure strategy. [Duke hyperbaric chamber with divers.] Divers has strict depth and exposure times to avoid seizures. Can only do 10 minutes at depth of 50 feet of sea water. Predisposition to seizure from sleep deprivation, medications. Antioxidants don’t work well in vivo but do work in tissue slices. Anti-seizure meds impair diver performance. Fasting ketosis can delay seizure 250%. Cahill work in 60s showed alternative fuel for brain — not just glucose. After 20 days of fasting, BHB is the primary fuel. “Nutritional ketosis is the only diet that can be defined by a biomarker!” Use of ketone esters as a tool to achieve a desired ketone level. Adenosine goes up with ketosis. GABA/Glutamate ratio goes up with ketosis. NLRP3 inflammasome is inhibited in ketosis. Kabuki syndrome — ketones as a potential therapy. Long list of potential therapeutic uses of ketosis — Weight loss management, T2D, inborn errors of metabolism, neurologic application, etc. see slides. Angelmann’s syndrome. Anxiety. Traumatic brain injury. CO2 effects in submarines & space living environments. Measurement — glucose / ketone index. Future devices — continuous monitoring of ketones. (some projects in the works). List of ongoing projects (see slides).

Dr. Dominic D’Agostino, USF, IHMC

AM Session Questions & Answers

ketones on gut microbiome and set-up point of weight? Ramsey — recent paper suggested microbiome change. Not well studied. Set-point is a phenomena for which we don’t have a good biomarker. Not well studied. Mice effects — aging vs. long-term administration? Haven’t looked yet at starting ketosis at different ages. Richard Feynman — huge progress on science. Little progress on policy. What can be done? How can we communicate?

Jeff Volek — Many diets can be used with success in humans. We are not proposing everyone should be on a ketogenic diet. Lots of people can eat carbohydrates. Not everyone can and age matters too.

Jose — Crossfit & weight lifting. Jeff Volek — acutely strength is not superior. Long-term, can help recovery and longevity in sport. E.g. wire hang test in mouse — equivalent in human. Jeff used it himself for powerlifting.

Dogs, Dogfood and Dogma. — Dogs, Dog Food, and Dogma: The Silent Epidemic Killing America’s Dogs and the New Science That Could Save Your Best Friend’s Life by Daniel Schulof

Link: http://a.co/gDNbndb, Iditarod race in ketosis — Steve.

Exogenous ketones — parallel to HDL drug trials, differences between endogenous and exogenous production, Dom — differential impact of BHB, Acetyl acetate and acetone.

Ravi — Application of ketones to wound care and sepsis? What is impact of anti-inflammatory environment? Forced retirement and depression and anxiety? — Any plans to use ketosis. FDA should be split. Shouldn’t regulate food and drugs together. Is anyone here from FDA?

Dom — rats in ketosis are easier to handle — less “anxiety”, less fear response, needs to be studied more.

Dr. Richard Feynman

Dr. Tim Noakes, Cape Town University

[Re-telling of story in The Lore of Nutrition]

Lunch speaker. I was already insulin resistant at age 28 (1980 paper). Father died of T2D in 10 years after diagnosis. After Waterlogged, restarted running, bad run, saw ad for the Atkins new book — angry — thought Steve, Jeff & Eric were sell outs, bought it, astonished — implemented, save his life. Backlash to writing about dietary changes. SAMJ article — case series on diabetes reversal. Real Meal Revolution Book. Proudfoot, Grier, Noakes, Creed. Dietitian backlash. Parliament. Attacked by faculty — University of Cape Town. Orwell — shutting down speech. Academic mobbing. The academic mob. Twitter charges. 7 words. Most expensive tweet ever. Ween baby to LCHF. Dangerous advice to “millions”. Ethics — never provided right to hear charges. Medical advice, doctor-patient relationship. Weening advice — meat, poultry, fish & eggs daily is in guidelines. Idea of charges preceded the tweet. Incriminating email chain. It was a witch hunt. Testimony from Nina. Zoe. Corrine Zinn. Gave 10 lectures to layout the data. All lectures are on Noakes Foundation website. Tim Noakes, Marika Sboros. Working on a new book that is simpler to lay out the dietary lies. In 1967–68, Tim was in high school in US. Dr. Martin Luther King quote. The dietary advice is wrong. We have to change it.

Dr. Tim Noakes, University of Cape Town

Dr. Andrew Mente, McMaster University

PURE Study, 135,335 people, 667 communities, 18 countries, 5 continents. Methods. Saturated fat vs. LCL-C, HDL-C, TC/HDL-C, ApoB/ApoA. With more saturated fat, LDL up but also improved TC/HDL-C ratio, improved ApoB/ApoA ratio. ApoB/A ratio is more important. If you focus on LDL, you give one piece of advice. If you look at other markers, you could give other advice. Carbs are the opposite as saturated fat. Meta-analysis of RCTs — saturated vs. polyunsaturated fat — no signal. Have to look at clinical events — prospective cohort data in PURE. 4,700+ CVD events, 5700+ mortality events. Most of carbs in PURE study are grains & sugar. Carbs >55% associated with higher mortality. Fats — lower risk. Saturated fat — lower risk of stroke. Salt. The salt hypothesis. <1.5g/day for high risk. <2.5g/day for everyone. Weak relationship between sodium and blood pressure. Intersalt. Etc. DASH trial NEJM 2001 proof of concept study. Meals over only 5 weeks. Not designed to look at mortality. 24-hour urine is the reference method for sodium but can’t do in large studies. Instead, did fasting morning urine. Valid method. 100,000 people — average sodium is 4.4 grams per day. US average is 3.5 grams. Recommendation of 2.3 gram — only 1% of people are doing that. China has a very HIGH sodium exposure- 5.6 grams. Sodium is essential! Risk of toxicity and deficiency at the extremes. J-shape curve. Risk if >6 or ❤. Sweet spot from 3–6 gram. If we reduced intake per guidelines, risk would rise. NEJM 2014. Risk of low sodium even in people with high blood pressure. Why low sodium harm? — Higher renin and aldosterone. New community paper in the Lancet. See increased risk in China in high sodium via stroke. Rest of world — inverse association. China would benefit from some sodium reduction. Rest of world would not. Low sodium is associated with higher mortality. 3–5 grams is fine — that’s where most people are.

Guidelines mistakes -

Salt: Too narrow a focus on blood pressure. Missed the big picture.

Saturated Fat: Too narrow a focus on LDL. Missed the big picture.

Dr. Andrew Mente, McMaster University

Dr. Ron Krauss, Children’s Hospital of Oakland Research Institute

Diet, Adiposity and Atherogenic Dyslipidemia. AHA & low fat diets early in his career. Atherogenic Dyslipidemia is high TG, Low HDL, increased apoB with small cholesterol-depleted LDL particles. Particle size. Small, dense sd-LDL are related to increased risk whereas the large buoyant are not. Hoogeveen ATVB 34:1069, 2014. Mora Circulation 132:2220, 2015 in JUPITER study. Very small LDL is worse. Pattern A and B. Atherogenic Dyslipidemia is people with pattern B. This is closely related to TG. Two pathways from liver that result in large and small LDL particles and remnants. Correlation with pattern B and BMI. Would low carb be beneficial? — Yes, can shift people away from phenotype B with just a modest carb reduction (from 54% to 39%). Quantum flip from phenotype B to phenotype A. Small LDL peaks with carbs. Has no response to change in saturated fat from 8 to 15%. More challenging to get weight loss with pattern B individuals vs. pattern A. Less weight loss per kcal reduction.

Dr. Ron Krauss, CHORI

Dr. Jake Kushner, McNair Interests

Ketogenic diets for adult & childhood type 1 diabetes. T1D was a death sentence until discovery of insulin by Banting and Best in Canada. 1.25M Americans have T1D. challenging to manage with insulin. Frequent hypoglycemia — 1–2 episodes per week. Working on regrowing beta-cells but wasn’t clear it would get to a point of impact. Stem cell therapy is a long way off. Patient experience with T1D — cognitive load of glucose insulin roller-coaster. Looking for alternatives, found low carb. Why do we prescribe carbohydrates? ADA says it’s because Institute of Medicine says so. Dr. Bernstein — major success with applying low carb to T1D. Example patient. “I always figured I was going to die from type 1 diabetes. With low carb I realized I may be able to lead a normal life. CGM trace example. Adrian LxM developer for Android APS. Lennerz et al. paper in pediatrics — A1c average of 5.67%. Critique — patients believe in the approach. Really? Ludwig response — don’t suppress information. ISPAD guidelines — ‘consensus’ to not restrict carbs. ADA guidelines — has backed off — wishy-washy statement. Provides wiggle room. Resources — Adam Brown book. Berstein book. Needs. Low carb GRIT in Houston. Community!!

Dr. Jake Kushner

Dr. Sarah Hallberg, Virta Health, IUH.

Type 2 diabetes epidemic — huge health & economic impact. Type 2 diabetes is reversible!! Get that message out. There are 3 approaches — bariatric, caloric restriction, low carb. What doesn’t work — standard of care. Different macronutrients produce different insulin responses. “Healthy carb” — different responses in different individuals based on insulin sensitivity & carbohydrate tolerance. ADA — “the total amount of carbohydrate is predictive of glycemic response”. Case A1c 11.3, last A1c 6.1. On CGM, seeing flat glucose curves. CGM is going to be a game changer. Even people without diabetes have CGM. Is there not enough evidence? Actually there are 20 randomized trials, 5 meta-analyses and 10 other trials supporting low carb for diabetes. DASH, Med, plant-based. There are many more low carb studies.

Virta-IUH trial. N=465 including 262 with T2D in intervention arm. 116 prediabetes. 87 T2D usual care. Mean BMI 40. Recruited “all comers”. Mean years with diabetes 8.4 years. Longer than most other studies. Outcomes — T2D, met syndrome, Weight. At one year, A1c from 7.5 to 6.2% in intervention arm. And the medications are eliminated. 57% of diabetes-specific medications are discontinued including 100% of the sulfonylureas. 94% of insulin was reduced or eliminated. Metformin has indications outside of diabetes — tend to leave it in place at patient’s discretion. GLP-1 as a bridge off insulin. Reduce risk of hypoglycemic event. Cost of medications in reduced 46%. This is just medication costs. Engagement — 83% at one year. That is better than prescriptions. People adhere well. Weight loss 13.2% (by home scale), 12% by clinic scale. 80% + lost >5%. CVD risk — 22 of 26 factors improved in intervention vs. 0 of 26 in usual care. LDL-C up but ApoB and LDL-p do not increase. Reversing diabetes and improving CVD risk profile. 10 year ASCVD risk score improvement. Liver ALT and AST improve. Overall — 60% reversal at a year — meaning glycemic control off meds other than metformin. Did they actually eat a ketogenic diet? Rather than relying on food record we had BHB as a biomarker. No food journal. Study design — rather than randomized controlled trial. Problem with nutrition long-term, the patients have to be invested. People have to have choice, understanding and belief in order to adhere. Support behavior change — biomarker, resources, coach, physician, community. Advice that works. Key message — diabetes is reversible. Gives patient power. Patient choice. Right now they are not being given a choice.

Dr. Sarah Hallberg, Virta Health and IUH

Questions

What about a higher protein approach? Often these people won’t be in ketosis — low carb, moderate protein, fat to satiety and measuring ketones works to get people into ketosis. T1D — 50 year old endocrinologist concerned about high LDL even though all other markers looking good — how do you proceed? — Jake is starting to measure remnant cholesterol per Dave Feldman — looking very good in his T1D patients even with a high LDL. Sarah — most of the patients don’t want to be on a statin. Get coronary calcium score and a CRP. If calcium score is not zero, discuss low dose statin. There are unknowns. Jay — all the studies on risk have been done on high carb populations. Short-chain fatty acids and fiber? Recommending 5 servings of vegetables and high BHB may substitute for butyrate in feeding colonocytes. Medical education? — Sarah — Nutrition Coalition is focused on evidence-based guidelines. Trickle-down into education. Even grade school kids “I had to lie again Mom.” Noakes Foundation course. UK course. — Online modules. Physicians fear ketoacidosis — it is an issue of T1D if not getting insulin. But nutritional ketosis doesn’t pre-dispose you to insulin deficiency. There is a biomarker tracking concern. CGM can circumvent this issue. Nutritional ketosis 0.5 mM? SGLT-2 inhibitors show CVD improvement. Could be due to low levels of BHB 0.2–0.3 mM. Decrease heart failure. What level is required for benefit is unknown. Ketoacidosis — SGLT-2 — get off of it before starting ketosis because of risk of DKA from the SGLT-2. Jay — similar risk of SGLT-2 with fasting. Richard Feynman — benefit of carb restriction occurs even in the absence of sustained weight loss. With ketosis, what should sodium intake be? Given that there is diuretic effect of ketosis. Andrew Mente — there remains increased risk at the low end. Unknown how high may be safe. 3–5 grams is fine. Unknown if 5, 6, 7 is harmful or beneficial. Sarah Hallberg — can increase is symptomatic. Use pre-exercise or in summer. PURE — fruits and vegetables? Lots of heterogenesity. Uncooked vegetables looked more beneficial than cooked vegetables. Haven’t looked at specific fruits and vegetables. Legumes category excluded nuts. Nuts do seem to have protective effect. Hospitals? Terrible what they feed people with diabetes. Problem is systemic. Dr. Mark Cucuzzella is making progress — sugar out of hospital. Ignoring the food and giving insulin. People who don’t respond to a ketogenic diet? CRP up? Sarah — not seeing this. People who adhere do well. Measuring BHB is hugely helpful. One group familial hyperchylomicronemia — about 1 in 1M. Naturopath — Iodine? Mente — fortified salt is important and iodine-deficiency is still occurring. Selenium? Tim Noakes on cholesterol — associational studies with cholesterol and heart disease. Hazard ratio is 1.2. Anything less than 2 is in the noise. Insulin resistance predicts CVD outcomes, not cholesterol. Familial hypercholesterolemia — treat for 50% reduction in cholesterol, 0.46% reduction in CVD. Cholesterol and statins doesn’t hold up. Ron Krauss — on thin ice. Recent consensus statement that there is LDL effect on CVD. PCSK-9 effect. Agree to disagree on this. Victoria (surgeon) Ketosis and microbiome in mice. How to deal with NPO fasting pre- or post-surgery with sugar in fluids. Could do ketoadaptation before surgery. Could do a trial to test this. Sarah — at IUH-Arnett, promotes carb loading pre-op. May extend hospital stays in T2D patients. Graham — are statins a scam? Nutrition in hospitals? It’s bad. Antagonistic toward low carbohydrate nutrition. Coca Cola machines on every floor with long term contracts. Mark Cucuzzella — making progress. Nurses and RDs are held to standards. Took administrative intervention in order to skip sliding scale insulin and order low carb meals. Once nurses had personal experience with low carb culture began to change then cafeteria started to change. Had to get a lawyer to take on Coca Cola contract. Substituted non-sugar versions of Coca Cola. Haven’t lost money on beverage sales. It will eventually spread in West Virginia. Geisinger is making similar progress in Pennsylvania. Dave Feldman — loves Ron and Tim. We are in unknown territory. Let’s not get antagonistic. For instance, lean mass hyper-responders with high LDL-C and LCL-P where every other biomarker is great. Jay — Shakespeare — “Men will wish that they were here” quote from Henry the 5th.

Panel — Wortman, Hallberg, Kushner, Krauss, Mente

Nina Teicholz, The Nutrition Coalition

[Re-telling of seed oils story in The Big Fat Surprise]

Dinner speaker. Not taking notes — briefly …

What would progress with the 2020 Dietary Guidelines ideally be …

  1. Allow low carb eating pattern as an option
  2. Drop the caps on saturated fat consumption, no evidence to support current recommendations
  3. Drop the restrictions on salt, no evidence to support current recommendations
Dr. Jeff Volek introducing Nina Teicholz
Nina Teicholz
Audience at dinner. OSU Ketosis Conference.

James McCarter

Written by

@JPMcCarter, MD PhD scientist & entrepreneur, @BioGeneratorSTL EIR, WashU Med Prof, former head of research at @Virtahealth, founder of Divergence & @YSPwustl

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade