All therapies, both anticonvulsant medications and dietary manipulation, have either known or potential side effects. A great deal more has been learned the past few years, and will continue to be learned in the future, about what the unintended consequences of initiating and maintaining the ketogenic diet may turn out to be. Now in 2011, we are at the point of preventing side effects before they happen rather than treating them when they occur. The potential pitfalls and problems fall into four broad categories: pre-diet evaluation, issues during the initiation of the diet, short-term side effects while on the diet, and the long-term risks.
It is important to prevent side effects before they occur. One of the physicians on our ketoteam tries to schedule a face-to-face outpatient clinic visit with the candidate child and his parents/caregivers in almost every case. Nothing can substitute for the opportunity to personally interview parents, ask additional questions about the history that may not be available from the child’s medical records, and examine the child.
There are exceptions, of course, especially when geographic issues are hard to overcome or the wait for a clinic appointment is much longer than the wait to start the diet. In those cases, we rely heavily on receiving “primary source” records, such as hospital discharge summaries, EEG and MRI data, and laboratory reports.
There are a few screening tests that must be documented. Most critical is to confirm that each child is not at significant risk for certain known metabolic conditions. This can be understood by performing studies such as lactate, pyruvate, and ammonia levels; plasma amino acids; and urine organic acids, if they are not already documented as normal. In many cases, a spinal fluid analysis will be obtained.
A careful history regarding food allergies should be completed. Because the diet for the most part consists of “normal” foods, the likelihood of problems in this area is low but must be screened for. It is possible for a skilled dietitian to design a diet even for children who have food allergies, and we have done so successfully many times.
Simply stated, the ketoteam and the parents/caregivers of the child who are considering dietary therapy must communicate clearly and come to an agreement that the “cost/benefit” analysis is favorable enough to proceed.
Minor problems often arise during diet induction. That is why we still have all our patients in the hospital for this process over a 4-day admission. It should be noted that there are some centers that do start the diet as an outpatient procedure. We believe this can be done in very select families and with very careful monitoring (including daily visits to the outpatient clinic during the initiation period). In general, we advocate this primarily in situations where the risk of the child getting sick from another child in the hospital for something different is too high. We have also occasionally started the diet in children with well-established home nursing care where transport is problematic (e.g., children on ventilators or receiving many medications), but even in these situations there is constant daily contact.
There is controversy regarding the need to have children fast prior to starting to consume the diet. Over time, we have modified our protocol from a pre-diet 48 hour fast to only 24 hours. Exceptions are made for very young infants or children who are medically less stable, who do not fast at all. Fasting starts the night before admission so that it is more than half over by the time the child is on the hospital floor. Fasting clearly accelerates the onset of ketosis, and if carefully monitored in the hospital setting, does not put the child at significant risk. But studies show that there is no difference in long-term outcomes using fasting versus nonfasting protocols.
Hypoglycemia and acidosis
Fasting may lead to hypoglycemia. Blood sugars are monitored about every 6–8 hours during the first few days of the diet by dextrostix. We know that levels of 40–50 mg/dl are not uncommon during diet induction and virtually never require a medical response. When blood sugar falls below 30 mg/dl, the test is repeated in 1 hour. If the child becomes symptomatic—too sleepy or sweaty—30 ml of orange juice may be given and repeated in 30 minutes if necessary. We avoid doing studies that require any type of sedation during this time to avoid possible confusion with hypoglycemia.
Acidosis can occur as well when the diet is started, and sometimes does persist. This can manifest as vomiting or sleepiness. We expect a child’s bicarbonate level to fall when he or she becomes ketotic; this is typical when an acidosis condition is created in the body. If the bicarb level falls below 10 mEq/L or pH falls below 7.2, then vomiting often occurs. We don’t always check for this (requires a blood draw to the lab), but we presume it is happening based on clinical signs and will give extra fluids, sometimes including bicarbonate.
Dehydration
At the onset of ketosis after the fast, many children do not drink fluids at their usual rate. Therefore, we encourage “pushing” oral fluids that are carbohydrate free (water or diluted diet soda or ginger ale) during the fast and the entire hospitalization. We no longer measure and restrict fluid intake; there is no evidence for any benefit, and there is a slightly increased risk of dehydration. It is rare for a child to become so dehydrated that he requires a bolus of fluid to be given either intravenously (IV) or by nasogastric (NG) tube, but we do that on occasion if the vomiting is very severe.
Vomiting
After fasting for 24 hours, or even without it, the initial food intake, usually given as a “keto shake” or KetoCal®, may cause nausea and vomiting. This does not usually persist long and can be overcome by having the child take small sips of the shake over a relatively long period of time at the initial meal, as opposed to drinking it rapidly. We find that using the keto shake seems to be well-tolerated and less nauseating than if you went to straight ketogenic diet solid foods immediately. In some situations we will give Reglan®, Zofran®, or IV fluids. Children with reflux before the diet can have this more often on the diet, too, so be ready.
Refusal to eat
We introduce the total calorie load over 3 days, at a rate of 33% per day divided into three feedings. Thus, the first day, the allotment is one-third of calories, the next day two-thirds, and the last day 100%. Even though most of the children are very hungry after the fast, some respond to being ketotic by lacking a feeling of hunger. Some of them refuse to eat and need a lot of encouragement. It may require that the keto shake be sipped slowly over several hours, or frozen into “ice cream,” or microwaved into “scrambled eggs,” and eaten slowly. For other children, we’ll advance the diet to solid foods a bit more quickly—going straight to regular foods rather than the keto shake a day or two early. However, the downside to this is that vomiting may occur.
Changes in weight
We now know that achieving a child’s “ideal” body weight is no longer a crucial factor in the long-term success or failure of the diet. Optimal ketosis and seizure control are not usually weight dependent, and so you do not need to check weight frequently. However, if control is poor, or if control suddenly changes without a clear reason, knowing the child’s weight may be a factor in the decision-making process for making diet changes.
In some children, a slight weight loss may be healthy and intentional.
Illness and infections
All children suffer intermittent illnesses, and there is no clear evidence to suggest that children on the ketogenic diet are more susceptible to infections than others. Should a child have an illness that requires IV fluids, parents need to make physicians and emergency room personnel aware that their child is on the ketogenic diet and that in almost every case, the IV fluids used should be completely sugar- (dextrose) and carbohydrate-free. The exception would be if a child is so dehydrated and acidotic that he or she has a hypolygemic seizure, at which time, of course, the child must be given glucose by IV. A general rule is that whenever the safety of a child is a serious issue, and medical treatment requires action that will diminish or end ketosis, that takes precedence over maintaining the diet. Remember, after a crisis is over, ketosis can always be reestablished, and the benefits of the diet reestablished.
Vitamin and mineral deficiency
The ketogenic diet is deficient in vitamins and perhaps in minerals as well. Severe examples are beri beri and optic neuritis due to lack of thiamine. Major minerals such as calcium and trace elements such as selenium are examples of those that are frequently deficient. All centers prescribe carbohydrate-free multivitamins and mineral supplements, which are readily available. Vitamin D, calcium, zinc, and selenium are important to make sure are included in the supplements (see Chapter 10).
Kidney stones
Overall, the risk of kidney stones while on the diet is increased and occurs in 1 in 20 children. The most common association is documented family history of kidney stones in the parents or siblings of the child on the diet. The medication type called carbonic anhydrase inhibitors (the common ones are Topamax® and Zonegran®) have a slightly increased risk of kidney stone formation but do not have to be discontinued when the ketogenic diet is started or maintained. Data are a bit controversial whether these medications increase the risk over the diet alone.
Kidney stones are generally calcium or uric acid stones, and the first indicator is usually the presence of traces of blood in the child’s urine. Parents can check urine every other week for blood on the multistix test strip. Clinical symptoms of kidney stones may be as specific as frank blood in the urine, “gritty” urine, and/or lower back (so-called “flank’) pain. Nonspecific symptoms include low-grade fever, abdominal pain, poor appetite, and an increase in number of seizures.
If kidney stones are suspected, carbohydrate-free fluids should initially be pushed hard to flush the urinary tract. If symptoms persist, the primary care physician needs to be consulted so that this possible complication can be differentiated from the much more common gastrointestinal virus or flu. The use of diagnostic abdominal ultrasound is recommended in cases where stones cannot be ruled out.
We now prescribe Polycitra K® or Bicitra® as soon as the diet is started for all children, regardless of known risk factors. These compounds raise the pH (technically called alkalinization) of the urine, which lowers the likelihood of stone formation by 7-fold (see Chapter 10). In most cases where stones are documented but small, they can be flushed out by increasing urine flow through increased carbohydrate-free oral intake or IV fluids. In rare cases, stones must be broken up by lithotripsy (using vibration) or even at times by surgery.
Constipation
This is one of the most common issues for children on the ketogenic diet, likely caused by a combination of reduced bulk in the diet and decreased fluid intake. After the diet regimen is established, we expect children to have a bowel movement pattern similar to pre-diet. That should be a frequency of no less than once every other day and with normal stool consistency so there is no straining.
To help avoid digestive problems while the body gets used to the new diet, we recommend use of George’s Aloe Vera® or Miralax® crystals (starting with one capful—about 17 grams daily) and expect parents to make adjustments over time to “keep things moving.” Sometimes these products need to be used long term, and that is fine. Enemas can be used if necessary, but are rarely needed.
High cholesterol and other lipid abnormalities
We live in a society very concerned with total cholesterol levels, “good” and “bad cholesterol levels, and elevated triglycerides. Not surprisingly, when people learn that a child is being given a diet that is in excess of 80% fat, they are both surprised and concerned. Our study of lipid level changes on the ketogenic diet indicates that in about 30% of children there may be cholesterol and triglyceride levels “that exceed current recommendations for normal children,” but in most cases this is a transient finding. As the body (primarily the liver) adjusts to the greatly increased load of fat it must digest when the ketogenic diet is started, the levels begin to stabilize and then return very close to pre-diet levels after 6–12 months. In cases where children experience exceedingly high lipid levels, there may be a coincident genetic type of familial hyperlipemia that was exacerbated by the high fat diet.
A study shows that in children on the diet for longer than 6 years, most of them had cholesterol levels in the normal range. When the diet is discontinued, and the child returns to a diet with a presumed “normal” fat intake, lipids almost always return to normal. We do not have any evidence that there are long-term effects of a temporary increase in lipids as a child.
What if the cholesterol won’t come down? Reducing the diet ratio, increasing the percentage of polyunsaturated fats, substituting medium chain triglycerides, and adding carnitine are the most common approaches we take. However, a study here showed that although 60% of children have at least a 20% reduction in cholesterol by making some (or all) of these changes, about 40% had their cholesterol decrease by at least 20% with just observation alone. In other words, it may be best to just repeat the labs after 1–2 months and leave the diet alone.
We have had virtually no child discontinue the diet exclusively for lipid abnormalities, and have never yet treated a child on the diet who has markedly elevated cholesterol with statin medication.
Carnitine deficiency
Many children on the diet for several months, especially if on valproate (Depakote®) at the same time, will have a decrease in their free carnitine. According to the international consensus statement, these children should be then treated with Carnitor®. At our center, we only do that if the levels are low and the child is symptomatic (e.g., fatigue, low energy, low ketosis). There is more information on carnitine in Chapter 10.
Studies of children on the diet for over 6 years have shown that seizure control and cholesterol are not adversely affected over time. However, the bad news is that bone fractures, kidney stones, and height disturbance are more of a problem long term. This doesn’t mean the diet has to be stopped, but it does mean that ketoteams need to be very aware of this and try to prevent problems before they occur.
There are definitely exceptions. A man in his 20s with tuberous sclerosis who has been on the ketogenic diet for nearly his entire adult life identified himself and presented to our clinic several years ago. He had no history of appetite or growth issues, kidney stones, constipation, or evidence of acidotic episodes. His bone density (DEXA) scan and carotid ultrasound were normal for his age. He remains on the diet, seizure free to this day.
Bone metabolism
Very debilitated, nonambulatory children who are on the ketogenic diet have a high incidence of bone changes that can be documented on radiologic studies and DEXA scans. We also know that vitamin D levels decrease over time (after initially going up on the diet due to the added supplements). We supplement 400–800 International Units per day of vitamin D when levels are demonstrably low Until they can be normalized and maintained at a lower dose. We do not routinely obtain skeletal X-rays or DEXA scans on our patients in the short term, but if they are on the diet over 2–3 years it is worth considering.
Changes in height
We expect children to grow normally in height while on the diet, although younger children may have a slight initial drop off of their growth rate. Studies have shown that the problems with height may be related to ketosis itself, so could happen as well with the modified Atkins diet, not just the ketogenic diet. Most children grow fine, but if they don’t, there is evidence that there is growth “catch up” when the diet is discontinued.
When we see our patients periodically in follow-up clinic, we of course measure and chart their weight and height. We want to see weight increase and linear height growth over time, in context of the child being able to have as much seizure control as possible, continue everyday activities, and all the while remaining in good health. If there is a problem, we will often lower the ratio to increase protein, even to the extent of the modified Atkins diet. Some children on the diet around the world have been started on growth hormone, with early reported good results (no studies yet).
Included in this category are bleeding disorders, increased bruising, and hepatitis, which are all liver-related problems; pancreatitis; iron deficiency anemia; prolonged QT intervals (heart); and alteration in immunoglobulin levels and function leading to possible increased occurrence of infections. Most of these problems are easily recognizable with regular clinical and laboratory monitoring and are correctable when diagnosed using the accepted treatments currently available.
Death
There are rare case reports of deaths occurring that may be attributable to the ketogenic diet. These are due to cardiomyopathy, selenium deficiency leading to cardiac arrhythmia, secondary to aspiration of fatty food contents (usually liquid formulas) causing lipoid pneumonia. One of our patients died due to recurrence of a cardiomyopathy of unknown cause, which had been identified well before the diet was started. Some children sadly may die as well of SUDEP (sudden unexpected death in epilepsy patients), which is more common in those with frequent, generalized-tonic-clonic seizures, especially receiving many anticonvulsants.
The ketogenic diet was not originally intended for long-term use as an epilepsy treatment, but because it has successfully reduced seizure frequency and improved general quality of life for so many patients, including infants, children, and adults, its usage has frequently been extended for many consecutive years. It is not alternative or, therefore, free of side effects. Thus, when patients, caregivers, and the ketogenic diet team embark on the difficult course that this therapy requires, we have accumulated enough experience and real data to help guide us through many of the pitfalls and issues that will likely confront us. And as time goes on, we will learn more, and hopefully optimize treatment and minimize the complications discussed in this chapter.