Ketoacidosis - A Complication of Atkins' Diet?

Today's headline in Bloomberg caught my attention - Low-Carb Atkins Diet Isn't Safe for Losing Weight, Doctors Say - sounds like the same old, same old warning, huh?

The opening sentence is certainly alarming: The low-carbohydrate Atkins diet, which experienced a peak of popularity two years ago, isn't safe and shouldn't be recommended for weight loss, according to doctors writing in The Lancet.

And once you get past the first few paragraphs you find the cause for this 'sounding the alarm bells' - In a case report, doctors including Klaus-Dieter Lessnau, clinical professor of medicine at New York University School of Medicine, described a 40-year-old white female patient from February 2004 who was vomiting as often as six times daily and had difficulty breathing after strictly following the Atkins diet for a month. She reported a weight loss of 9 kilograms (20 pounds) while on the diet.

I really had to go read the case study for myself since the media reports, found in literally hundreds of sources today - from the Chicago Sun Times to Forbes, from ABC Online to USA Today - are carrying the story with scant details.

The details from the media include: the woman was 40-years old, obese, vomiting six times a day, and the Atkins Diet caused her to develop ketoacidosis.

The case study was published in the Lancet. Fortunately for all the alarmists, it's within the confine of "premium content" so it's not publically accessible without paying to read it. I'll just say here that limited access must go by the way-side if we're going to advance evidence-based medicine...full open-access is the way to do this, but that's an entirely different subject!

So, what did the case study tell us that the media isn't? More importantly, does the information contained within justify the frenzy of alarm today?

Let's tackle the first question first. Within the case report are the following details: the woman was 40-years old, obese with a BMI of 41.6, who'd followed the 1972 version of the Atkins diet for one-month before she lost her appetite and started to feel nausea. During that time before she started feeling ill, she was eating meat, cheese and salads daily, was taking an assortment of nutritional supplements and lost about 20-pounds in the month. Five days prior to her emergency room visit for shortness of breath, she'd vomited four to six times a day.

The report notes that other than the mild distress, clinical examination was "unremarkable" with normal vital signs. Blood tests showed a high level of lipase (indicative of pancreatitis) with normal amalyse, which confuses things since both are elevated with pancreatitis - sometimes this type of reading is actually suggestive of something more sinister happening, specifically cancer. But we find no mention of that possibility in the case report.

The report does state, clearly, that a CT scan of her pancreas on admission was normal. Interesting, at least to me, is that with ketoacidosis her glucose would be high (at a level greater than 14mmol/l - hers was 4.2mmol/l) and her sodium would be abnormal, yet it was normal according to the report. But, hey, what do I know?

The diagnosis of ketoacidosis in this woman was based on "The differential diagnosis of high-anion-gap metabolic acidosis includes ingestion of methanol, ethylene glycol, or salicylate, L- or D-lactate acidosis, and ketoacidosis due to diabetes mellitus, alcohol, or starvation." It's clearly stated that "Our patient denied alcohol use; her serum osmolar gap was 0, which excludes the presence of unmeasured osmotic agents such as methanol or ethylene glycol; L-lactate concentration was normal; and salicylate was undetectable. D-lactate acidosis was unlikely without antibiotic use or bowel surgery. Serum was positive for acetone, and ß-hydroxybutyrate was high at 390 µg/mL (normal 0–44 µg/mL), consistent with ketoacidosis."

Quite frankly, the level of ketones in her urine are not alarming for someone who is not diabetic or someone properly following a low-carb diet...but let's keep moving forward and get through this report. The author states that in trials where children are fed a ketogenic diet for epilepsy ketoacidosis is a complication.

Unfortunately, the reference he cites as his evidence, did not have ketoacidosis due to diet as a complication. Nice try though - Fanconi's renal tubular acidosis, reported as a complication in one patient in his citation, is a disorder that may be inherited as a primary disorder or may be one symptom of a disease that affects many parts of the body. Researchers have now discovered the abnormal gene responsible for the inherited form. More often, however, classic distal renal tubular acidosis is a complication of diseases that affect many organ systems (systemic diseases), like the autoimmune disorders Sjögren's syndrome and lupus.

Gotta give him credit for trying! I check referenced citations specifically because I see, too often, a study referenced and when you read through the data, the data doesn't support the use of the citation in context to support an assertion.

Here's the main problems with the case report.

The symptoms of acute methanol poisoning are shortness of breath, vomiting, headache, with metabolic acidosis occuring often. The course of treatment for her was dextrose and sodium bicarbonate infusion - the course of treatement for methanol poisoning.

She'd been vomiting for days - that alone will upset her electrolyte levels and cause dehydration. Five days of vomiting - well, you can imagine just what state her body was in by that time. The vomiting and dehydration are glossed over as inconsequential to make the case it was the diet - and only the diet.

So what's that now? Two possibilities besides the diet?

How about another - maybe she did drink alcohol and that disturbed her metabolism leading to the acidosis state. Yes, alcohol can do this to a person, regardless of the diet they consume!

Lastly, she may have actually had undiagnoised diabetes and/or a cancerous tumor that would have also served as the catalyst for her symptoms.

We just don't know since none of the other possibilities were explored - the diet was the cause in the mind of her physician, so there was no further investigation.

That troubles me - we have thousands of participants from hundreds of studies to date and not one incident of ketoacidosis. Is it possible this woman's ketoacidosis was a result of her diet? I would say it's not impossible, but very highly improbable given the reams of data from clinical trials to date that have not found ketoacidosis as a complication, even in diabetic patients following a low-carb diet.

The reason it isn't a complication is that dietary ketosis, in and of itself, does not cause ketoacidosis. More likely this woman experienced ketoacidosis as a result of something else and unfortunately her diet placed an obstacle in front of a complete investigation as to its cause.

That said, I cannot state strongly enough that one person is not evidence to indict a low-carb diet as dangerous. Which leads to my answering the second question above - does the information contained within justify the frenzy of alarm today?

We do not see this frenzied alarm when someone in a statin trial experiences complications which may or may not be related to the drug nor do we see such alarm when someone participating in a low-fat diet trial experiences a complication which may or may not be due to their diet.

In the real world, one person in a case study presenting information about one person is considered to be the "n of 1" - meaning there is no other person to compare them to - their symptoms and course of treatment are in isolation to a greater population and highly dependent on the investigation, or lack thereof, of causes related to the symptoms.

While this particular case study was interesting, it remains an "n of 1" since there are no other reports of this type of complication from any other clinician in the world, and the physician did not, in my opinion, explore all avenues for cause since he believed it could only be her diet - short-sighted and a headline grabber to be sure.

Good science? No.

Correction: In the originial publication of this article I noted the woman's blood pH was normal. In reviewing the case report again, I realize I read through the details too quickly and was incorrect - her pH was lower than normal, suggestive of acidosis. The sentence was edited out for accuracy.