The list of side effects is often used to dismiss ketogenic therapy. Look at the mechanisms and almost every one of them turns out to arise from how the diet is composed and supervised.
🩺 Written by K.Y.J.A.M. Ho, MD PhD, medical specialist · reviewed by C. Pleiter, medical specialist · published 27 August 2026 · last medically reviewed 27 August 2026 · sources & method
Among 313 children on a ketogenic diet, 10.5 per cent developed symptomatic kidney stones without preventive potassium citrate, against 0.9 per cent with it. One cheap supplement, an elevenfold difference. The same pattern repeats across nearly every documented side effect: there is a known mechanism and a known fix. Electrolytes prevent keto flu, fibre and magnesium the constipation, supplementation the selenium and carnitine deficiencies, and more protein the effect on growth in children. The exception is the cholesterol response; that is partly individual and the last word has not been said. If you use medication for blood sugar, blood pressure or fluid removal, discuss any carbohydrate reduction with your own doctor first, because your dose shifts with it.
Do the check (1 minute) or jump straight to the science
Moderate Potassium citrate cut kidney stones from 10.5 to 0.9 per cent Strong The chain from acidosis to stone is standard renal physiology Emerging The cholesterol response is partly individual and still unsettled
This is educational information based on averages, not personal medical advice. Take these points to your own doctor and discuss what fits your situation. Never change medication on your own.
This check gives an indication only. It is not a medical test and makes no diagnosis. Discuss any doubt or symptoms with your doctor.
Search for ketogenic diet side effects and you find a list: kidney stones, constipation, keto flu, slowed growth in children, selenium and carnitine deficiencies, lower bone density. That list is accurate. What rarely accompanies it is that nearly every item on it has a known mechanism and a known fix.
The clearest example sits in a cohort of 313 children with difficult-to-treat epilepsy. In the group without preventive potassium citrate, 10.5 per cent developed symptomatic kidney stones. In the group that received it, that figure was 0.9 per cent: one child out of 106.
What stands out is what did not happen. The supplement did not make the blood less acidic; the effect sat almost entirely in the urine. Ketosis simply continued, and only the consequence nobody wanted disappeared. The stone was not a price you pay for ketosis, but the result of ketosis without its accompanying support.
The same pattern holds for most of the list. The keto flu of the first week comes from sodium loss and resolves with salt and electrolytes. Constipation comes from low volume and fluid. The selenium and carnitine deficiencies come from a narrow menu without supplementation. And the slowed growth in children relates to the protein content of the classic 4:1 diet, chosen in the 1920s to suppress seizures rather than to let children grow.
One point does not fit that pattern, and it belongs here: in a subset of people LDL cholesterol rises sharply, and that is not only a matter of composition but also of predisposition and build. The science layer below goes into that.
What this means in practice: the known side effects are an argument for supervision and measurement, not against the approach. Anyone eating ketogenic without electrolytes, without citrate and without following blood values runs into exactly the problems described in the literature.
Without support it does raise the risk. In a cohort of 313 children on a ketogenic diet, 10.5 per cent of the group without preventive potassium citrate developed symptomatic kidney stones, against 0.9 per cent of the group that received it. The cause is a chain of known effects: mild acidosis lowers urine pH, mobilises calcium from bone and lowers urinary citrate, while low-carb menus are often oxalate-rich and fluid intake tends to be low. Potassium citrate, enough fluid and spreading oxalate address that chain.
It has been described with the classic 4:1 diet, but the cause most likely lies in the protein content. That ratio was chosen in the 1920s to suppress epileptic seizures and is therefore deliberately low in protein, because protein is partly converted to glucose. On variants with a lower ratio and more protein, such as the modified Atkins protocol, the effect on growth is considerably smaller. It is a design choice made for another purpose, not a property of ketosis.
This is the one side effect that does not fit the pattern neatly, because the response is partly individual. It is clearest in people who are lean, insulin-sensitive and often athletic. One proposed explanation is that with little glycogen the liver exports more triglycerides to fuel muscle, leaving more LDL particles behind. In a study of one hundred people with this profile, ApoB and cumulative LDL exposure did not predict who developed plaque growth; the amount of plaque at baseline did. That is not a free pass, because plaque increased on average and there was no control group. Practically: have a calcium score done rather than steering on that single number, and swap part of your saturated fat for olive oil. Discuss this with your own doctor.
Evidence labels: Strong = consistent randomised trials or meta-analyses matching the population and intervention · Moderate = randomised or solid observational evidence with limitations · Emerging = early, small or animal studies, not yet confirmed in humans.
Why those stones form. Ketosis comes with mild, chronic metabolic acidosis. That single shift sets five things in motion which all push the same way. Urine becomes more acidic, making uric acid poorly soluble. The body buffers the acidosis partly with bone, sending calcium into the urine: in this cohort 52 per cent of children had hypercalciuria. Under acidosis the kidney reabsorbs more citrate, while citrate is precisely the main natural stone inhibitor. Low-carb menus also lean on oxalate-rich products such as almond flour, spinach and cocoa. And fluid intake is structurally too low in the first weeks. Strong
What potassium citrate does about it. It addresses three of those five links at once. Urine pH in this cohort rose from 6.2 to 6.8, back to pre-diet levels. Urinary citrate goes up. And calcium stays in solution, including in the children who continued to excrete too much of it. The researchers conclude that the effect runs almost entirely through the urine rather than through blood acidity. Moderate
Why the classic diet is low in protein. The classic ketogenic diet works with four parts fat to one part protein plus carbohydrate combined. It is deliberately low in protein, because protein is partly converted to glucose and dampens ketosis. That ratio was chosen to suppress epileptic seizures at a time when there were barely any alternatives. On variants with a lower ratio and more protein, the effect on growth is considerably smaller. This is therefore a design choice made for another purpose, not a property of ketosis itself. Moderate
Keto flu is a sodium matter. As soon as insulin falls, the kidney retains less sodium. That loss explains most of the headache, flatness, dizziness on standing and muscle cramp of the first week. It is not a detox and not an adaptation you have to sit out, but an electrolyte shortage with a direct fix. With blood pressure medication or diuretics, the amount of salt should be agreed with your own doctor. Strong
The exception: the cholesterol response. In a subset of people LDL cholesterol rises sharply on a carbohydrate-restricted diet. This relates to genetic predisposition and build and is clearest in people who are lean, insulin-sensitive and often athletic. One explanatory model holds that with little glycogen the liver exports more triglycerides in VLDL particles to fuel muscle, leaving more LDL particles behind; high LDL would then reflect high lipid throughput. In a study of one hundred people with this profile, followed with CT angiography for a year at a mean LDL of 242 mg/dL, neither ApoB nor cumulative LDL exposure predicted who developed plaque growth; the amount of plaque at baseline did. Emerging
Why that is not a free pass. In that same study non-calcified plaque increased by 37 per cent in relative terms on average. There was no control group, it covered one year and one selected profile. What the research shows is that the usual lipid markers did not predict who progressed in this group, which is different from demonstrating that they do not matter. Practically it follows that a coronary calcium score gives more information in this situation than chasing a single number, and that partly replacing saturated fat with monounsaturated fat lowers the response in many people. Emerging
This is general medical information, not a diagnosis or a treatment. For advice about your own situation, and before changing anything about your medication, always talk to your own doctor.
Explore more topics in our knowledge base →