This article is part of our October 2020 special issue. Download the full issue here.
Klement RJ, Champ CE, Kämmerer U, et al. Impact of a ketogenic diet intervention during radiotherapy on body composition: III—final results of the KETOCOMP study for breast cancer patients. Breast Cancer Res. 2020;22:94.
A prospective comparison study, nonblinded, at a single outpatient clinic in Germany
To assess the feasibility and effects on body composition of a ketogenic diet versus a standard “prudent” diet during adjuvant radiotherapy for breast cancer
All participants were women with nonmetastatic breast cancer (aged 25-78 years) who were scheduled to receive adjunctive radiotherapy (from 16-35 days). Allocation of participants was done consecutively at a single clinic (Leopoldina Hospital Schweinfurt, Germany) in 3 batches: For Batch 1, 5 consecutive participants were placed into the ketogenic diet (KD) group to assess whether compliance was likely; for Batch 2, the standard diet (SD) group was recruited (38 women were asked, 30 enrolled, and 1 dropped out before analysis); and for batch 3, the rest of the KD group was filled (total of 52 women were asked, 19 declined, 32 enrolled, and 3 dropped out). Participants were given the opportunity to self-select KD or SD group if they felt strongly, and this resulted in 1 woman electing to participate in the SD group. Total participants in each group at the study’s end point were KD=29 and SD=30.
Despite nonrandomization, there were no significant differences in baseline characteristics between participants, except the average phase angle per bioimpedance analysis (BIA) was slightly higher in the KD group (4.98 vs 4.65, P=0.05) at baseline.
Karnofsky Performance Scale Index <70, body mass index (BMI) <18 kg/m2, metallic implants (due to interference with BIA measurements), pregnancy, inability to understand German, any condition that made the KD a contraindication.
The SD group received no specific dietary recommendations or advice. However, 4 patients in the SD group requested dietary guidance and received standard guidelines of the German Nutrition Society. This included consuming mostly “unrefined foods of plant origin (in particular whole grains, vegetables, and fruits), and limiting fats to 30-35% daily energy intake, with an emphasis on reducing fat from animal origin.”
Participants in the KD group were briefly counseled by their medical oncologist (5 minutes). The day of their baseline metabolic measurements, they also received instructions from a registered dietician experienced in KD implementation. This included handouts with food choices and cooking recipes. KD participants were instructed to replace carbohydrates with fat, specifically consuming 75% to 80% of calories from fat and limiting carbohydrates to 50 grams daily or 10 grams per meal. There was no caloric restriction. High-quality animal protein and nutrient-dense foods were emphasized along with avoidance of grains, legumes, processed foods, and vegetable oils (except olive and coconut). Dairy was allowed in moderation, with instructions to favor butter, cheese, and fermented dairy products. Medium-chain triglyceride (MCT) oil was provided for free. Lastly, they were instructed to begin the KD at least 2 days prior to initiating radiotherapy.
Note: Of the 29 participants in the KD group, 15 were given a 10-gram amino acid supplement on the days of radiotherapy (Master Amino Acid Pattern, or MAP), which they consumed right after receiving radiation. (Contrary to what may be expected, this did not modify body composition trends, likely due to adequate high-quality protein in the KD group as a whole.)
Baseline measurements, which were taken at the time of radiotherapy planning (about 1 week before radiotherapy) and included weight, BIA, a validated quality-of-life (QOL) questionnaire (EORTC QLC-C30), and a blood draw for baseline measurements of complete blood count CBC, metabolic parameters, liver enzymes, insulin, insulin-like growth factor 1 (IGF-1), and free T3 and T4. BIA and weight measurements were repeated weekly throughout the study. The QOL questionnaire and blood analysis were repeated once during therapy and during the final week of radiotherapy.
In the KD group, compliance measures included: a food diary for 2 days, measures of acetoacetate in urine daily (using ketone strips at home), blood ketones and glucose weekly using finger stick testing after BIA measurement.
The primary outcome measures included: 1) dropout rate in the ketogenic diet group, 2) changes in body composition parameters from baseline to final week of radiation: body weight (BW), fat-free mass (FFM), skeletal muscle mass (SMM), extracellular water (ECW), total body water (TBW), and intracellular water (ICW=TBW – ECW), and 3) changes in bioimpedance phase angle (PA) at 50 kHz.
The implementation of the KD was feasible, with only 3 women dropping out of the KD group (3/32), while 1 woman dropped out of the SD group (1/31).
As expected, the KD group reached ketosis, with mean and median beta hydroxybutyrate (BHB) concentrations of 0.72 and 0.49 mmol/L (range 0.06–4.9). This was significantly higher than the SD group (P<2.2x10-16). In the SD group there was a small increase in total body weight and fat mass and a decrease in fat-free mass, none of which reached statistical significance.
In the KD group, there was a rapid loss of body weight and fat-free and skeletal muscle mass, which was attributed mostly to rapid water loss. After this rapid loss, there was no further loss of fat-free mass or skeletal mass, but there was a slow decrease of 0.4 kg (0.9 lb) of body weight and fat mass per week (P<0.0001). In addition, the KD group had decreased free T3 levels (0.06 pg/mL/week) (P=6.3x10-5).
In integrative oncology there are 2 diets that are sweeping both professional conferences and popular online searches: the ketogenic diet and fasting diets of various forms. The earliest studies on ketogenic diets that showed therapeutic promise were specifically for primary brain tumors, and these were calorically restricted ketogenic diets, making them difficult to maintain long term.1,2 In addition, a study on 16 people with metastatic cancers found that many were unable to reach ketosis, and 14 of the participants progressed while the 2 who achieved stability did not have high ketones in circulation.3 Add to this that while “ketogenic” diets in medicine are well-established therapeutic diets that reliably induce ketosis, much of what is dubbed “ketogenic” in nonmedical sources fails to adhere to the strict guidelines that a therapeutic ketogenic diet requires. Many patients who believe they are eating a ketogenic diet are not doing so in the strict sense of producing ketones. This is all to say that the term “ketogenic diet” must be looked at closely when considering its effect on the patient, their treatment, and their outcome over time.
Therapeutic ketogenic diets, while promising in conjunction with other treatments, are not a stand-alone means of controlling cancer growth. In the current study under review, a medically prescribed ketogenic diet was studied for its feasibility and metabolic effects. This study was not designed to assess outcomes, although the authors indicate that this will be reported in time. Rather, this study is instructive of the short-term metabolic effects of a non-calorie-restricted KD in a real-world clinical setting.
Many patients who believe they are eating a ketogenic diet are not doing so in the strict sense of producing ketones.
In this study, all of the participants were receiving radiotherapy 5 out of 7 days per week, and the median duration was 35 days. The women were able to ask questions and get clarifications regarding the KD protocol at will. Daily oversight may improve compliance with any program, but this likely affected the compliance with the KD given its rigorous standards to achieve ketosis. Daily measurement of urinary ketones at home also apprised the women of their status, perhaps prompting them to improve compliance when needed. In addition, participants were under close oversight, including weekly glucose and ketone measurements along with weekly BIA assessment. All of this resulted in a very low dropout rate, which was 1 of the primary outcome measures of this study.
Klement and colleagues included the use of bioimpedance analysis to track precisely what type of weight is being lost. The KD often leads to rapid weight loss in the first weeks of institution, and that held true in this study. They noted that most of the changes occurred by the second measurement in the KD group. The rapid loss of weight was attributed mostly to water loss, per the BIA data. The authors contend this is expected due to rapid depletion of glycogen stores, which require water for glycogen storage at a 3:1 ratio (every gram of glycogen holds 3 grams of water in the liver or muscle). Ketogenic diets also lead to lower insulin levels, as was demonstrated in this study. Insulin increases the reabsorption of sodium in the kidneys, and the authors speculate that this diuretic effect may also have led to a more rapid water loss at the start of the KD.
The KD not only led to rapid weight loss at the start of the study, but it also led to a more gradual decrease of fat mass with preservation of muscle mass. This is in keeping with prior data that suggested a KD can help preserve muscle mass.4 In comparison, while not statistically significant, the SD group had small gains in total body weight and fat mass (0.04 and 0.08 kg/wk) along with small decreases in fat-free mass/muscle (–0.05 kg/wk). This was highly variable between participants, but prior studies have suggested women undergoing radiotherapy for breast cancer often gain fat, lose muscle, or both.5
The authors predicted that a KD would influence free T3, insulin, and IGF-1 levels. While there was a decrease in insulin and IGF-1 in the KD group versus the SD group, this was a small difference. The authors postulated that if the trends continued over a longer time horizon, there may be a significant difference eventually. Overall, higher BMI was significantly associated with higher insulin levels, and higher age was significantly associated with lower IGF-1 levels. There was a highly significant average drop in free T3 (0.06 pg/mL/week) in the KD group, which was present whether the analysis was total KD group versus SD group or limited to only those not taking thyroid medication (KD=22 and SD=21 participants). The authors did not speculate on the meaning of this finding.
There are several limitations of this study. It is limited to a single institution with highly trained dieticians and staff familiar with the KD. There is no standardization of the KD apart from generalized guidelines. The study design was dramatically altered from the original. The original study was a larger study that included several types of cancers, including breast, head and neck, and rectal cancers, and had a prepackaged ketogenic drink and amino-acid supplements versus a whole-food KD with MAP. However, before any breast cancer patients enrolled into that study, over half of the participants undergoing radiotherapy for head and neck or rectal cancers dropped out due to inability to tolerate the maximum target doses of the keto drink and amino-acid supplements. This clearly indicated that a program with large doses of MCT and amino acids was not feasible for those with rectal or head and neck cancers. This study essentially took arm 2 (the whole-food KD ) of the original study, limited recruitment to women with breast cancer only, and created a SD group for comparison.
We eagerly await future results of this study, since there are theoretical reasons for a therapeutic benefit of combining KD and radiation.6 The close oversight and high compliance, along with documentation of ketosis throughout this study for the KD group, should provide some insight into whether the KD affects outcomes of recurrence or breast cancer–related mortality. For now, we know that the KD is feasible and may be useful for those women who have excess adipose, which is considered a risk factor in itself for recurrence of breast cancer.