Introduction
A ketogenic diet (KD) consists of low carbohydrate, high fat and adequate protein. This diet was designed to simulate the biochemical effects of fasting, which is used for the treatment of drug-refractory seizures (1,2). By limiting the intake of carbohydrates, it forces the body to switch to fatty acid oxidation as an energy source, leading to a ketogenic state. This KD guarantees sufficient calories for normal activity produced through ketogenesis, which occurs principally in the mitochondrial matrix of the liver (3,4). While KD is a classic therapy for refractory epilepsy, recent animal studies have demonstrated that it is also effective in the treatment of other neurodegenerative diseases, including Parkinson's disease, amyotrophic lateral sclerosis, Alzheimer's disease and traumatic brain injury (5–10).
Although KD has been used as a dietary regimen for managing pediatric refractory epilepsy over the past decades, its side effects remain controversial. Clinical trials have suggested that KD is safe and efficient (11,12); however, it also resulted in a progressive loss of bone mineral content in pediatric patients, particularly in those with a low body mass index (BMI) (13). A previous study by our group confirmed that KD resulted in a significant bone loss in the tibia and reduced mechanical function in the femur of mice (14); such side effects in cancellous-rich bones, e.g. the vertebrae, are worthy of further investigation. In addition, the association between bone loss and mechanical reduction under KD requires clarification.
Clinically, osteoporosis is associated with a reduced bone mineral density (BMD) and micro-architectural changes, resulting in diminished bone strength and an increased risk of fractures (15,16). Osteoporosis-associated fractures may lead to substantial long-term disability and decreased quality of life (17). Thus, it is necessary to identify the osteoporosis-inducing effects of KD in various parts of the skeleton, including reduced mechanical function of vertebral bones. Micro-finite element (FE), based on the 3-dimensional (3D) bone microarchitecture with high-resolution peripheral quantitative computed tomography (CT), is used to examine age-associated changes (18), effects of disease (19) and outcomes of various treatments on the bone microstructure (20). Analysis of bone mechanical performance using a micro-CT-based FE model, and correlation with microstructural changes and reduced mechanics of bones under a KD is essential.
The present study aimed to investigate the effects of a long-term KD on the microarchitecture of appendicular and axial bones, and the compressive stiffness and failure load based on micro-FE analysis in a rat model. The secondary objective was to assess the possible correlation between the microstructural changes and the mechanical properties of appendicular and axial bones.
Discussion
The present study demonstrated a significant decrease in the TBMD of rats fed KD for 12 weeks, with no difference in the serum calcium and phosphate concentration between the KD and control groups. Specifically, by scanning appendicular bones of the extremities (humerus and tibia) and L4 vertebra using micro-CT, it was observed that KD led to bone loss in cancellous and cortical bones, with insignificant changes in L4 vertebral bone. Furthermore, simulated compressive analysis using micro-FE analysis revealed that the compressive stiffness and strength of appendicular and axial bones were decreased under the KD, and were highly correlated with the microstructural parameters of cancellous and cortical bones.
The effects of KD bone loss in humans remain inconclusive. A previous study has indicated that maintaining a long-term KD did not cause any major negative effects on the bone mineral content and BMD in adults with glucose transporter-1 deficiency syndrome (21). The present study, which only included an animal experiment with a small sample, does not allow for any conclusions regarding the safety of KDs, and the results may not be comparable with previous studies focusing on children with intractable epilepsy treated with KD (13,22). However, the present study on a ketogenic rat model may enhance the understanding of bone loss due to KD, suggesting that this topic is worthy of further investigation.
Ketogenic animal models are typically established by feeding KD or by β-hydroxybutyrate administration (23,24); however, to the best of our knowledge, there is currently no unified ketogenic animal model. A previous study by Wu et al (14) has indicated that KD caused bone loss in mouse tibia, but the blood ketone levels in the KD group were only ~1.0 mmol/l. In the present study, elevated blood ketone levels (~2.0 mmol/l) were achieved with a customized KD (24,25), which indicated that the ketogenic model was efficient. In addition, the present study investigated the bone loss in L4 vertebra and humerus, providing a comprehensive overview of the bone loss in rats under the KD.
Body weight changes may be the first factor associated with bone loss under KD. In the clinic, KD is effective in suppressing the appetite, reducing lipogenesis and increasing lipolysis (26,27), and also results in weight loss (28). A low body weight or BMI is a well-documented risk factor for fracture, and a body weight increase is considered a protective factor against osteoporosis, as mechanical loading and the associated hypertrophy of muscles increases the bone mass (29). Thus, theoretically, rats with low body weight have a lower bone mass, which is in accordance with the results of the present study.
Furthermore, previous studies have reported that chronic ketoacidosis led to an increase in bone minerals to maintain the buffering capacity and decrease renal conversion of 25-(OH)-VD to 1,25-(OH)2-VD3 (30). Low 25-(OH)-VD levels, particularly those <20 ng/ml, were associated with an elevated fracture risk in women aged ≥50 years during a 15-year follow-up period (31). However, in the present study, the levels of serum calcium and phosphorus in the KD group were similar to those in the standard diet group; this trend was also observed in the total vitamin D and active vitamin D levels, which implied that neither the bone mineral content nor the conversion of vitamin D is a key factor for the bone loss under the KD.
The present study used DEXA to evaluate the TBMD as a primary endpoint in rats fed with the KD. DEXA is the gold standard for the diagnosis of osteoporosis, and the T-score is highly correlated with the fracture risk (32). While the present results indicated an obvious decrease in the TBMD in the KD group, the bone loss within certain specific bones required further evaluation. In order to determine microstructural changes of different bones in detail, micro-CT analysis was performed on the bones of the upper limbs, lower limbs and axial skeleton.
KD led to cancellous and cortical bone loss in the tibia and humerus, which was consistent with findings of tibia bone loss in a previous study by our group (14). A previous study by our group demonstrated that KD specifically caused cortical bone loss of tibiae in mice when compared with ovariectomized mice (14). The present study also investigated the bone loss in middle of L4 vertebra and the humerus, providing a comprehensive overview of the bone loss under KD. Furthermore, the present study revealed that the most severe trabecular bone loss and structural deterioration occurred in the metaphyseal area of the tibia, and was sensitive to the location selected. By contrast, similar cortical bone loss occurred in the tibia and humerus, while a lesser decrease was apparent in appendicular bones. In spite of greater microstructural changes in cancellous bone of L4 vertebral bodies in the KD group, the present study failed to detect significant changes, potentially due to the small sample size. However, the thickness of trabecular bone remained unchanged between the KD rats and the control group, which indicated a decrease in trabecular bone density associated with the bone loss.
The present study further verified the biomechanical changes under the KD, and a correlation between the bone morphological changes with the bone function was identified. Previous studies have estimated bone strength using high-resolution peripheral quantitative CT and the FE method, and provided important insight into fracture risk in patients (33,34). The present study used a micro-FE model based on the micro-CT images, and evaluated the compressive stiffness and strength of appendicular and axial bones under the KD. The micro-FE analysis is suitable in cases where the sample size is small and the region of interest unavailable for actual mechanical tests. A significant decrease in biomechanical performance of appendicular bones was detected by micro-FE analysis. The non-significant decrease in compressive stiffness and strength of L4 vertebrae in KD group was potentially due to the small sample size (n=4). These biomechanical features were consistent with the microstructural changes of the appendicular and axial bones under KD. The results of the present study suggest that KD therapy for ~12 weeks increases the risk of fracture.
Mechanical properties may be correlated with microstructural parameters; however, the correlation coefficient between the compressive stiffness and a single microstructural parameter was dissatisfactory in the present study. BMD, as the gold standard for evaluating bone strength, is associated with bulk modulus through a power law association in bone (cm scale). By contrast, the association is weak at the tissue level (µm-mm scale), likely due to the effect of microstructural features at small-length scales (35). In vitro studies have demonstrated that the combination of the trabecular microarchitecture and BMD improved the prediction of trabecular bone mechanical behavior and vertebral strength (36,37). In the present study, the combination of the BMD of trabecular bone with the Barea of cortical bone produced a satisfactory prediction of bone mechanical properties, with the adjusted R2 ranging from 0.81 to 0.91.
The effects of diet on bone microstructure may be achieved by reduced mineral absorption or increased bone turnover. The present study confirmed that KD did not affect mineral absorption, suggesting that the KD affected the bone metabolism. In the present study, the KD successfully converted glycometabolism to ketone body metabolism, as ketone levels were high and blood glucose levels were low (38). Although no increase in bone turnover induced by a low-carbohydrate diet was observed in humans (39), our previous studies have confirmed that KD and ovariectomy resulted in a similar promotion of bone absorption via activation of osteoclasts, rather than inhibition of bone formation mediated by osteoblasts (14). In addition, certain studies have reported that ketone bodies are important endogenous factors that may regulate bone metabolism under physiological and pathological conditions, as the ketone bodies modulate osteoblast functions bi-directionally (40).
The effects of KD on bone loss have been rarely reported (14,22). The present study demonstrated more general bone loss under a KD for the first time, to the best of our knowledge, and evaluated the mechanical performance using a micro-FE model. These results establish a link between the microstructure and the function of bones. However, the dynamical changes in bone microstructure should also be evaluated at multiple time-points. For instance, the dynamic changes in bone microstructure and the effects of KD on cortical bone may be evaluated using in vivo micro-CT (41). In addition, the molecular mechanisms of bone loss caused by KD, which is expected to differ from the effects of a high-fat diet, was not determined in the present study; the effect of a KD on bone metabolism has been assessed by a previous study (40), but it should be further elucidated in the future.
In conclusion, the present study demonstrated that a KD led to bone loss and biomechanical reduction in appendicular bones, with a lesser impact on axial bones. Thickness was reduced in the cortical bone, and trabecular density was decreased in the cancellous bone in rats under KD, also leading to decreased compressive stiffness and strength of appendicular bones. The present study demonstrated a significant correlation between the compressive properties and microstructural parameters of cancellous and cortical bones, respectively.