Body Composition Changes After Very-Low-Calorie Ketogenic Diet in Obesity Evaluated by 3 Standardized Methods

Abstract

Context:

Common concerns when using low-calorie diets as a treatment for obesity are the reduction in fat-free mass, mostly muscular mass, that occurs together with the fat mass (FM) loss, and determining the best methodologies to evaluate body composition changes.

Objective:

This study aimed to evaluate the very-low-calorie ketogenic (VLCK) diet-induced changes in body composition of obese patients and to compare 3 different methodologies used to evaluate those changes.

Design:

Twenty obese patients followed a VLCK diet for 4 months. Body composition assessment was performed by dual-energy X-ray absorptiometry (DXA), multifrequency bioelectrical impedance (MF-BIA), and air displacement plethysmography (ADP) techniques. Muscular strength was also assessed. Measurements were performed at 4 points matched with the ketotic phases (basal, maximum ketosis, ketosis declining, and out of ketosis).

Results:

After 4 months the VLCK diet induced a −20.2 ± 4.5 kg weight loss, at expenses of reductions in fat mass (FM) of −16.5 ± 5.1 kg (DXA), −18.2 ± 5.8 kg (MF-BIA), and −17.7 ± 9.9 kg (ADP). A substantial decrease was also observed in the visceral FM. The mild but marked reduction in fat-free mass occurred at maximum ketosis, primarily as a result of changes in total body water, and was recovered thereafter. No changes in muscle strength were observed. A strong correlation was evidenced between the 3 methods of assessing body composition.

Conclusion:

The VLCK diet-induced weight loss was mainly at the expense of FM and visceral mass; muscle mass and strength were preserved. Of the 3 body composition techniques used, the MF-BIA method seems more convenient in the clinical setting.

Excess fat mass (FM), especially visceral fat, is associated with a variety of pathological conditions such as hypertension, dyslipidemia, diabetes, and even cancer (1–4), as well as an increase in overall and cardiovascular mortality (2, 5). Very-low-calorie diets are commonly used as obesity treatments; however, a primary concern in using such diets is the amount of fat-free mass (FFM; i.e., muscular tissue) that is lost together with the FM (6). This could produce so-called sarcopenic obesity, that is, the coexistence of an excess of FM and a decline in FFM that yields a near-normal body weight (7). Sarcopenic obesity constitutes a double impact on the health of patients, because the reduction in muscle mass and muscle strength is also a cause of cardiometabolic disorders, such as myocardial infarction and stroke, and other adverse health outcomes (5, 7–9). For these reasons, it is of primary interest to find weight loss strategies that promote preferential loss of FM and preservation of muscle mass and its functional status (i.e., muscle strength) (10–12).

Some previous studies have suggested that very-low-calorie ketogenic (VLCK) diets may be effective tools to manage overweight and obesity (10, 11, 13). VLCK diets are a nutritional intervention that emulate fasting by restricting carbohydrates and fat with a relative increase in protein intake (6). The increased protein content may be partially responsible for the muscle mass preservation (12–14). Importantly, the weight-reducing action of these diets is rapid, and despite the fact that the ketosis state lasts only 60 to 90 days at the start of treatment, the weight reduction remains for up to 2 years (13). Therefore, VLCK diets operate by potent mechanisms to induce weight loss, and various body compartments might be altered. To the best of our knowledge, no studies have exhaustively assessed the changes in body composition associated with this type of diet, and variations in muscle strength have been only assessed in athletes (15).

Hence, the 2 main objectives of this study were to assess the changes in body composition and muscle strength promoted by a VLCK diet in the treatment of obese patients and to compare different methodologies used to evaluate body composition. To achieve this, body composition was evaluated by 3 potent and well-validated techniques: dual-energy X-ray absorptiometry (DXA), multifrequency bioelectrical impedance analysis (MF-BIA), and air displacement plethysmography (ADP) at different stages during the weight reduction process induced by a VLCK diet.

Materials and Methods

Study design

This study was a nutritional intervention clinical trial and was open, uncontrolled, prospective for 4 months, and performed in a single center.

Study population

The patients attending the Obesity Unit at the Complejo Hospitalario Universitario of Santiago de Compostela, Spain, for treatment of obesity were consecutively invited to participate in this study.

The inclusion criteria were ages 18 to 65 years, body mass index (BMI) ≥30 kg/m², stable body weight in the previous 3 months, a desire to lose weight, and a history of failed dietary efforts. The main exclusion criteria were diabetes mellitus, obesity induced by other endocrine disorders or by drugs, and participation in any active weight loss program in the previous 3 months. In addition, those patients with previous bariatric surgery, known or suspected abuse of narcotics or alcohol, severe depression or any other psychiatric disease, severe hepatic insufficiency, any type of renal insufficiency or gout episodes, nephrolithiasis, neoplasia, previous events of cardiovascular or cerebrovascular disease, uncontrolled hypertension, orthostatic hypotension, and hydroelectrolytic or electrocardiographic alterations were excluded. Females who were pregnant, breastfeeding, or intending to become pregnant, and those with child-bearing potential who were not using adequate contraceptive methods, were also excluded. Apart from obesity and metabolic syndrome, participants were generally healthy individuals.

The study protocol was in accordance with the Declaration of Helsinki and was approved by the Ethics Committee for Clinical Research of Galicia, Santiago de Compostela, Spain (registry 2010/119). Participants gave informed consent before any intervention related to the study. Participants received no monetary incentive.

Nutritional intervention

All of the patients followed a VLCK diet according to a commercial weight loss program (PNK Method), which includes lifestyle and behavioral modification support. The intervention included an evaluation by the specialist physician conducting the study and assessment by an expert dietician. All patients underwent a structured program of physical exercise with external supervision (16). This method is based on high-biological-value protein preparations obtained from cow milk, soya, avian eggs, green peas, and cereals. Each preparation contained 15 g protein, 4 g carbohydrates, 3 g fat, and 50 mg docosahexaenoic acid, and provided 90 to 100 kcal (16).

The weight loss program has 5 steps (Supplemental Fig. 1) and adheres to the most recent (2015) European Food Safety Authority guidelines on total carbohydrate intake (17). The first 3 steps consist of a VLCK diet (600 to 800 kcal/d), which is low in carbohydrates [< 50 g (26 to 30 g) per day from vegetables] and lipids (only 10 g of olive oil per day). The amount of high-biological-value proteins range from 0.8 to 1.2 g per kg of ideal body weight to ensure minimal body requirements are met and to prevent the loss of lean mass. In step 1, the patients consumed high-biological-value protein preparations 5 times per day, and vegetables with low glycemic indexes. In step 2, 1 of the protein servings was replaced by a natural protein (e.g., meat or fish) either at lunch or at dinner. In step 3, a second serving of low-fat natural protein replaced the second serving of biological protein. Throughout the ketogenic phases, supplements of vitamins and minerals, such as K, Na, Mg, Ca, and omega-3 fatty acids, were provided in accordance with international recommendations (18). These first 3 steps were maintained until the patient lost the target amount of weight, ideally 80%. Hence, the ketogenic steps were variable in time, depending on the individual and the weight loss target.

In steps 4 and 5, the ketogenic phases were ended by the physician in charge of the patient based on the amount of weight lost, and the patient started a low-calorie diet (800 to 1500 kcal/d). At this point, the patients underwent a progressive incorporation of different food groups and participated in a program of alimentary re-education to guarantee long-term maintenance of the weight loss. The maintenance diet consisted of an eating plan that was balanced with respect to carbohydrates, protein, and fat. Depending on the individual, the calories consumed ranged between 1500 and 2000 kcal/d, and the objective was to maintain the weight loss and promote a healthy lifestyle.

During the study, patients followed the different steps of the method until reaching the target weight or up to a maximum of 4 months of follow-up, although patients remained under medical supervision for the subsequent months.

Schedule of visits

Throughout the study, the patients completed a maximum of 10 visits with the research team (every 15 ± 2 days), of which 4 were for a complete physical, anthropometric, and biochemical assessment; the remaining visits were to manage adherence and evaluation of potential side effects. These 4 visits were made according to the evolution of each patient through the steps of ketosis as follows: visit C-1 (baseline), normal level of ketone bodies; visit C-2, maximum ketosis; visit C-3, reduction of ketotic approach because of partial reintroduction of normal nutrition; visit C-4, no ketosis (Supplemental Fig. 1). The total ketosis state lasted for 60 to 90 days only. In all of the visits, patients received dietary instructions, individual supportive counsel, and encouragement to exercise on a regular basis using a formal exercise program. Additionally, a program of telephone reinforcement calls was instituted, and a phone number was provided to all participants to address any concerns.

Anthropometric assessment

All anthropometric measurements were undertaken after an overnight fast (8 to 10 hours), under resting conditions, in duplicate, and performed by well-trained health workers. Participants’ body weights were measured to the nearest 0.1 kg on the same calibrated electronic device (Seca 220 scale; Seca North America, Chino, CA), in underwear and without shoes. BMI was calculated by dividing body weight by the height squared [BMI = weight (kg)/height² (m)]. Waist circumference was recorded with a standard flexible nonelastic metric tape at the middle point between the lower edge of the ribs and the iliac anterior spine. This measurement was made at the end of a normal expiration while the subject stood upright.

Handgrip strength

Muscular strength was measured with a Jamar handgrip dynamometer (Lafayette Instruments, Lafayette, IN). After a brief demonstration and verbal instructions, the test was performed in the standing position with the wrist in the neutral position and the elbow flexed to 90 degrees. Patients were given verbal encouragement to squeeze as hard as possible and to apply maximal effort for at least 3 seconds. Two trials were allowed in the dominant limb, and the highest score was recorded as peak grip strength (kg). Considering possible influences on the muscular strength of changes in body composition, handgrip strength (HG) was divided by appendicular lean mass (ALM) determined by DXA (HG/ALM) and by appendicular soft lean mass (ASLM) determined by MF-BIA (HG/ASLM).

Total body composition and selective visceral fat-mass analysis

DXA

Total body mass

Total body composition was first measured by iDXA (GE Healthcare Lunar, Madison, WI). Daily quality control scans were acquired during the study period. No hardware or software changes were made during the course of the trial. Subjects were scanned using standard imaging and positioning protocols, while wearing only light clothing. For the current study, bone mineral density, lean body mass, and FM values, which are directly measured by the GE Lunar Body Composition Software (GE Healthcare), were used. Some derivative values, such as bone mineral content, regional lean mass, ALM, FFM, and FM percentage (FM%), as well as android and gynoid fat (%), were also calculated. The android/gynoid ratio was automatically generated and analyzed using enCORE software, version 13.6 (GE Healthcare). For measuring android fat, a region of interest was automatically defined as the caudal limit placed at the top of the iliac crest, and its high was set to 20% of the distance from the top of the iliac crest to the base of the skull to define the cephalic limit of abdominal subcutaneous adipose tissue. The scanner was calibrated daily against the calibration block supplied by the manufacturer.

Visceral fat mass

Visceral fat was calculated using a newly developed software (Core Scan; GE Healthcare), which was validated against computed tomography in a patient population with a wide range of BMIs (19, 20). Visceral fat mass data from DXA were transformed into adipose tissue volume using a constant correction factor (0.94 g/cm³).

MF-BIA

MF-BIA (InBody 720; Biospace, Tokyo, Japan) was also used for determining body composition in terms of FM, FM%, FFM, total body water, intra- and extracellular water, skeletal muscle mass, and ASLM. This noninvasive technology employs 8 contact electrodes, which are positioned on the palm and thumb of each hand and on the front part of the feet and on the heels. In addition, MF-BIA uses the body’s electrical properties and the opposition to the flow of an electric current by different body tissues. The analyzer measured resistance at specific frequencies (1, 5, 50, 250, 500, and 1000 kHz) and reactance at specific frequencies (5, 50, and 250 kHz). The participants were examined while lightly dressed, and the examination took less than 2 minutes and required only a standing position. The validity of this technology has been documented in previous studies (21, 22). Visceral fat area values were also calculated in cm² by MF-BiA. Importantly, these values are significantly correlated with those generated by computed tomography (22, 23). The calculation of the different body compartments was performed according to the instructions of the manufacturer (Biospace).

ADP

The last technique used to determine body composition in the current study was ADP (BodPod; Life Measurements Instruments, Concord, Canada), which is accepted as a convenient alternative to the water immersion method for assessing body composition. The standard BodPod protocol was followed (24), and weekly quality control tests were performed during the study period; a second calibration was conducted immediately prior to the measurement of each participant. ADP determines body volume using Boyle’s law of the pressure/volume relationship. Therefore, body volume is equivalent to the decrease of volume in the chamber with the entrance of the patient under isothermal conditions. The participants were instructed to wear a swimming suit tight to the body and a swim cap during the test to diminish accumulated air and avoid volume discrepancies. Thoracic gas volume was measured by connecting the subject to a breathing circuit. The process was repeated until a consistent measurement was obtained. Body density was calculated as mass divided by volume and corrected for lung volume. The Siri formula was used to calculate FM, FM%, and FFM (24, 25).

Determination of levels of ketone bodies

Ketosis was determined by measuring ketone bodies, specifically B-hydroxybutyrate (B-OHB), in capillary blood using a portable meter (GlucoMen LX Sensor, A. Menarini Diagnostics, Neuss, Germany) before measurements of anthropometric parameters. As with anthropometric assessments, all of the determinations of capillary ketonemia were made after an overnight fast of 8 to 10 hours. These measurements were performed daily by each patient during the entire VLCK diet, and the corresponding values were reviewed on the machine’s memory by the research team for managing adherence. Additionally, B-OHB levels were determined at each complete visit by the physician in charge of the patient. The measurements reported as “low value” (≤ 0.2 mmol/L) by the meter were assumed to be zero for purposes of statistical analyses.

Statistical analysis

The data are presented as mean (standard deviation). All statistical analyses were carried out using Stata statistical software, version 12.0 (Stata, College Station, TX). A P < 0.05 was considered statistically significant. Changes in the different variables of interest from baseline and throughout the study visits were analyzed following a repeated measures design. A repeated measures analysis of variance test was used to evaluate differences between different measurement times, followed by post hoc analysis with Tukey’s adjustment for multiple comparisons. In addition, linear regression analyses were used to evaluate the accuracy of MF-BIA and ADP in comparison with DXA, because DXA is considered the reference technique in the estimation of body composition in clinical research (26). Finally, the Bland-Altman approach was also used to assess the accuracy of MF-BIA and ADP against DXA in the estimation of FM%.

Results

Baseline characteristics of participants

Initially, 23 participants were recruited into the study, but 3 dropped out voluntarily during the first week of the intervention for reasons unrelated to diet, and therefore were excluded from analysis. The 20 patients who completed the study exhibited the following baseline characteristics: mean age, 47.2 ± 10.2 years; BMI, 35.5 ± 4.4; and waist circumference, 109.4 ± 12.8 cm; 12 (60%) were women (Supplemental Table 1). Other baseline characteristics and their corresponding changes during the study are presented in Table 1.

Table 1.

Evolution of Body Composition During the Entire Study as Determined by DXA, MF-BIA, and ADP

Variable	Visit C-1	Visit C-2	Visit C-3	Visit C-4	P
Diet duration, d	0	39.2 (8.4)	89.7 (19.1)	123.3 (17.6)
B-hydroxybutyrate, mmol/L	0.0 (0.1)	1.0 (0.6)a	0.7 (0.5)a,b	0.2 (0.1)b,c	< 0.001
Anthropometric measurements
Weight, kg	95.9 (16.3)	84.2 (13.0)a	76.6 (11.1)a,b	75.1 (11.8)a,b	< 0.001
BMI, kg/m2	35.5 (4.4)	31.2 (3.3)a	28.4 (2.6)a,b	27.8 (2.9)a,b	< 0.001
Waist circumference, cm	109.4 (12.8)	98.0 (11.0)a	90.0 (9.3)a,b	88.6 (10.1)a,b	< 0.001
DXA
FM, kg	42.2 (9.1)	35.0 (7.8)a	27.8 (5.7)a,b	25.7 (5.8)a,b,c	< 0.001
FM percentage, %	45.6 (5.4)	43.1 (6.1)a	37.9 (5.8)a,b	35.7 (5.9)a,b,c	< 0.001
Visceral FM, kg	2.18 (1.28)	1.46 (0.86)a	1.01 (0.58)a,b	0.95 (0.62)a,b	< 0.001
Visceral fat volume, cm3	2313.6 (1358.0)	1551.8 (912.3)a	1062.9 (612.9)a,b	1008.4 (659.2)a,b	< 0.001
FFM, kg	52.8 (10.2)	48.5 (9.2)a	48.3 (9.1)a	49.0 (9.7)a	< 0.001
Lean body mass, kg	50.2 (9.9)	46.0 (8.8)a	45.8 (8.7)a	46.5 (9.3)a	< 0.001
Bone mineral content, kg	2.59 (0.49)	2.59 (0.48)	2.57 (0.49)a,b	2.55 (0.48)a,b	< 0.001
Bone mineral density, g/cm2	1.24 (0.13)	1.24 (0.13)	1.24 (0.12)	1.24 (0.13)	0.663
ALM, kg	23.5 (5.7)	21.4 (5.0)a	20.8 (4.6)a,b	21.0 (4.8)a	< 0.001
MF-BIA
FM, kg	40.0 (10.0)	31.4 (8.4)a	23.5 (6.8)a,b	21.7 (6.6)a,b	< 0.001
FM percentage, %	41.6 (6.4)	37.3 (7.5)a	30.8 (7.5)a,b	29.0 (7.4)a,b,c	< 0.001
Visceral fat area, cm2	154.1 (31.9)	126.8 (30.9)a	99.3 (26.7)a,b	93.3 (28.2)a,b	< 0.001
FFM, kg	55.8 (11.0)	52.7 (10.6)a	53.0 (10.5)a	53.4 (11.0)a	< 0.001
Total body water, kg	41.0 (8.1)	38.7 (7.8)a	38.9 (7.7)a	39.1 (8.1)a	< 0.001
Intracellular water, kg	25.5 (5.1)	24.0 (4.9)a	24.1 (4.8)a	24.2 (5.1)a	< 0.001
Extracellular water, kg	15.5 (3.0)	14.6 (2.9)a	14.8 (2.9)a	14.9 (3.0)a	< 0.001
Skeletal muscle mass, kg	31.2 (6.6)	29.3 (6.4)a	29.4 (6.3)a	29.6 (6.6)a	< 0.001
ASLM, kg	23.0 (5.0)	21.7 (4.7)a	21.4 (4.7)a	21.3 (4.9)a	< 0.001
ADP
FM, kg	44.0 (11.1)	36.1 (11.4)a	27.9 (7.4)a,b	26.3 (8.5)a,b	< 0.001
FM percentage, %	45.9 (8.5)	42.6 (9.4)	36.9 (10.3)a	34.7 (8.3)a,b	< 0.001
FFM, kg	51.8 (12.3)	48.0 (10.8)	48.6 (12.1)	48.8 (9.5)	0.124
Muscle strength
HG, kg	35.0 (9.9)	35.5 (8.9)	34.3 (11.2)	34.1 (10.5)	0.417
HG/ALM	1.4 (0.2)	1.6 (0.2)a	1.6 (0.3)a	1.6 (0.2)a	0.001
HG/ASLM	1.5 (0.2)	1.6 (0.2)	1.5 (0.3)	1.5 (0.2)	0.056

Data are presented as mean (standard deviation). ALM was measured by DXA. ASLM was measured by MF-BIA. P is from repeated measures analysis of variance for differences between measurement times (visits C-1, C-2, C-3, and C-4). Tukey’s adjustment was used for multiple comparisons.

^a

P < 0.05 in comparison with visit C-1.

^b

P < 0.05 in comparison with visit C-2.

^c

P < 0.05 in comparison with visit C-3.

Changes in anthropometric parameters and body composition

Although the patients underwent a total of 10 visits, the complete body composition analyses were synchronized with the ketone levels in 4 visits (Table 1; Fig. 1). Visit C-1 was the baseline visit, before starting the diet, with no ketosis (0.0 ± 0.1 mmol/L) and a body weight of 95.9 ± 16.3 kg. Visit C-2 was at the time of maximum level of ketosis (1.0 ± 0.6 mmol/L) with a body weight of 84.2 ± 18.0 kg. At visit C-3 (after 89.7 ± 19.1 days of VLCK), patients began the return to a normal diet and showed a reduction in ketone levels (0.7 ± 0.5 mmol/L) and a body weight of 76.6 ± 11.1 kg. Finally, at visit C-4, the patients were out of ketosis (0.2 ± 0.1 mmol/L) and showed a body weight of 75.1 ± 11.8 kg. All weights were statistically different from baseline levels (P < 0.05; Table 1; Fig. 1).

Figure 1.

Changes in total body weight and their relationship with levels of ketone bodies.

The severe reduction in body weight was mainly a result of FM reduction, as assessed by DXA scan; the −20.2 kg of weight reduction at the end of the study was in large part due to the −16.5 kg reduction in FM. When the FM compartment was assessed by MF-BIA, the result was very similar (−18.2 kg) and was further corroborated by the ADP analysis [−17.7 kg; Fig. 2(A)], without statistical differences among the results. It was remarkable that 3 methods of evaluating body composition, which operate through different principles, yielded such similar results. FM loss represents nearly 85% of the total weight loss achieved across the study.

Figure 2.

(A) Changes in total FM in comparison with baseline as determined by DXA, MF-BIA, and ADP. (B) and (C) Changes in visceral fat tissue in comparison with baseline as determined by DXA and MF-BIA, respectively.

Because visceral fat is physiologically and clinically more relevant than total FM, special emphasis was placed on its analysis. The VLCK diet led to a significant reduction in visceral fat that can be seen in assessment by either new DXA software (−1.2 ± 0.7 kg) or by MF-BIA [−60.8 ± 20.7 cm²; Fig. 2(B) and 2(C)]. Therefore, when evaluated by different methods, the VLCK diet induced a significant body weight reduction by targeting total FM and visceral FM [Fig. 2(A–C)].

The concern regarding the possible loss in muscular mass was addressed by a strict analysis of the FFM component using DXA, MF-BIA, and ADP, and mild but significant reductions in relation to baseline were observed [Fig. 3(A)]. The greatest decrease in FFM occurred at the visit of maximum ketosis (visit C-2), with partial recovery thereafter [Fig. 3(A)]. The FFM losses determined by the different techniques at this point were −4.2 ± 1.8 kg (DXA), −3.1 ± 1.5 kg (MF-BIA), and −3.7 ± 7.6 kg (ADP). At the end of the study (visit C-4), these FFM losses had diminished to −3.7 ± 2.0 kg, −2.4 ± 1.8 kg, and −3.0 ± 8.0 kg, respectively.

Figure 3.

(A) Changes in total FFM in comparison with baseline determined by DXA, MF-BIA, and ADP. (B) Changes in body water evaluated by MF-BIA. (C) Changes in muscle strength evaluated by handgrip dynamometer.

By discriminating the components of the FFM as determined by DXA, it was observed that these variations were mainly due to changes in lean mass, whereas bone mineral content remained unchanged from baseline (0.003 ± 0.066 kg at visit C-2; −0.018 ± 0.066 kg at visit C-3; and −0.028 ± 0.066 kg at visit C-4; P > 0.05). Given that DXA technique is unable to discriminate the composition of lean mass, the question was raised as to whether the observed reductions in lean mass were at the expense of muscle mass or body water content. Therefore, further analysis was performed by MF-BIA , which is able to discriminate these 2 variables. Remarkably, the measurements performed by MF-BIA showed that the initial loss of FFM at visit C-2 (−3.1 ± 1.5 kg) was mostly due to total body water loss (−2.3 ± 1.1 kg), both intra- (−1.5 ± 0.7 kg) and extracellular [−0.8 ± 0.5 kg; Fig. 3(B)], probably because of the intense diuresis that occurs in the first phase of any VLCK diet. In subsequent visits, a slight recovery of intra- and extracellular water was observed, similar to the recovery observed with total FFM. This means that reductions attributable to muscle mass were, depending on the method used, around 1 kg throughout the 4-month study; only 5% of the total 20.2 kg of weight lost was FFM.

The observation that the VLCK diet severely reduced FM while preserving muscle mass was reinforced by the maintenance of its physiological action (i.e., muscle strength). Despite a slight reduction in ALM and ASLM, as determined by DXA and MF-BIA, respectively, crude HG remained unchanged during the study (Table 1). Moreover, HG/ALM and HG/ASLM showed a moderate increase in comparison with baseline [Fig. 3(C)].

Estimation of body composition among different methods

Finally, the accuracy of MF-BIA and ADP in the estimation of body composition was studied in relation to DXA. As shown in Table 2, the unadjusted regression coefficients for FM, FM%, and FFM were consistently higher with MF-BIA in comparison with ADP throughout the study. Specifically, regression coefficients for MF-BIA were high (r² > 0.8) for FM and FFM, whereas those regression coefficients for FM% were slightly lower (r² > 0.7). However, most of the regression coefficients using ADP were lower (r² < 0.7) for FM, FM%, and FFM. A similar pattern was observed when adjusting for age and sex. The regression coefficients for both MF-BIA and ADP decreased with weight loss.

Table 2.

Body Composition Measured During the Entire Study: Comparison of MF-BIA and ADP With DXA

Method and Variable	Visit C-1			Visit C-2			Visit C-3			Visit C-4
	r2	95% CI	P	r2	95% CI	P	r2	95% CI	P	r2	95% CI	P
Model 1: unadjusted
MF-BIA
FM kg	0.89	0.80–0.99	< 0.001	0.88	0.76–1.01	< 0.001	0.79	0.64–0.93	< 0.001	0.82	0.68–0.97	< 0.001
FM percentage, %	0.79	0.64–0.94	< 0.001	0.76	0.63–0.89	< 0.001	0.73	0.61–0.85	< 0.001	0.75	0.63–0.87	< 0.001
FFM, kg	0.91	0.80–1.01	< 0.001	0.84	0.75–0.94	< 0.001	0.84	0.75–0.93	< 0.001	0.86	0.79–0.94	< 0.001
ADP
FM, kg	0.74	0.56–0.92	< 0.001	0.61	0.47–0.76	< 0.001	0.52	0.24–0.80	0.001	0.49	0.26–0.72	< 0.001
FM percentage, %	0.50	0.31–0.69	< 0.001	0.49	0.28–0.70	< 0.001	0.34	0.12–0.56	0.004	0.37	0.07–0.66	0.017
FFM, kg	0.76	0.60–0.93	< 0.001	0.72	0.50–0.95	< 0.001	0.67	0.51–0.84	< 0.001	0.82	0.52–1.12	< 0.001
Model 2: adjusted by age and sex
MF-BIA
FM, kg	0.90	0.81–1.00	< 0.001	0.95	0.84–1.05	< 0.001	0.86	0.73–0.99	< 0.001	0.86	0.74–0.98	< 0.001
FM percentage, %	0.77	0.57–0.96	< 0.001	0.81	0.62–1.00	< 0.001	0.73	0.57–0.89	< 0.001	0.76	0.59–0.93	< 0.001
FFM, kg	0.89	0.71–1.07	< 0.001	0.84	0.66–1.02	< 0.001	0.79	0.63–0.94	< 0.001	0.82	0.67–0.97	< 0.001
ADP
FM, kg	0.75	0.55–0.94	< 0.001	0.61	0.46–0.76	< 0.001	0.59	0.26–0.91	0.001	0.55	0.27–0.82	0.001
FM percentage, %	0.40	0.20–0.61	0.001	0.35	0.15–0.56	0.002	0.15	−0.12–0.42	0.271	0.36	0.08–0.63	0.013
FFM, kg	0.57	0.37–0.76	< 0.001	0.37	0.07–0.66	0.016	0.42	0.17–0.66	0.002	0.45	0.19–0.71	0.002

Correlation coefficients (r²), confidence intervals (95% CIs), and P values are given for each of the methods compared with DXA.

The results of the Bland-Altman approach in regard to the FM% are shown in Fig. 4. MF-BIA underestimates the FM% during all visits, although with increasing body fat there is a trend toward better agreement [Fig. 4(A)]. This negative slope was significant in visits C2 (P = 0.015), C3 (P = 0.003), and C4 (P = 0.005). Importantly, MF-BIA had a consistent variability of about 5% in determining FM% when compared with DXA. However, the concordance between DXA and ADP is shown in Fig. 4(B). In visits C1 (P = 0.005), C2 (P = 0.010), and C3 (P = 0.004) significant negative slopes were observed, indicating underestimation of ADP at lower levels of FM%, but ADP seemed to overestimate FM% with increasing body fat. During visit C-4, a similar pattern was observed, although the slope did not reach statistical significance (P = 0.093). During all visits there was a high variability in the FM% determined by ADP, reaching values of up to 20% in comparison with DXA.

Figure 4.

Bland-Altman analysis plotted from FM percentage in comparison with DXA. (A) Agreement between DXA and MF-BIA. (B) Agreement between DXA and ADP.

Discussion

In this study, the effects on body composition and muscle strength induced by a VLCK diet (PNK Method) in obese patients during an intervention period of up to 4 months was determined. This work assessed body composition during and after severe weight loss by using 3 different, highly sophisticated, and widely validated techniques (DXA, MF-BIA, and ADP), which allowed an accurate evaluation of the body changes during dieting. The main findings of the present work were: (1) there was significant weight loss throughout the entire study, which was mostly explained by reductions in total FM and visceral fat tissue; (2) there was a mild initial loss of FFM followed by a partial subsequent recovery of FFM, which was principally a result of changes in body water; (3) adequate muscle strength was preserved during the course of the diet; and (4) the less expensive and more convenient technique of MF-BIA showed an acceptable agreement with DXA in estimating body composition.

The VLCK diet was used because of its ability to produce a rapid and well-tolerated weight loss with a ketogenic phase that lasts 60-90 days and a final result of 20 kg of weight reduction at 4 months. The rapid reduction in weight is the probable explanation of the positive effects of this dieting approach, which are evident 1 and 2 years later (12, 13). Four different stages occurred with the VLCK diet used: a basal stage with obese body weight and no ketosis, a second stage with extreme ketosis and marked body weight loss, a third stage with body weight loss and declining ketosis, and a fourth stage with weight loss and no ketosis. Body composition was studied with the 3 techniques at each of these stages.

The accurate measurement of body composition changes is relevant to assess the contribution of the diet intervention, not only to total body weight but to the changes produced in FM, FFM, visceral fat tissue, and total body water (25, 27). To obtain such information, multicompartmental models that integrate information obtained from a single measurement (body density, total mineral mass, total body water) may be used to reduce the number of assumptions made on the stability of body characteristics (28). These models are of limited application in clinical practice, because they do not provide immediate results, are expensive, and require advanced analytical expertise (29, 30). For such reasons the 3 more widely used body composition analysis techniques were used in the present work. DXA is the most validated and commonly used technique to analyze body composition in obese patients and is based on the attenuation of a low-energy X-ray beam, depending on the tissue density and chemical composition. DXA is considered the gold standard technique by most groups working with body composition and was used as the reference method in the present work. Bioelectrical impedance techniques are low cost and readily available and rely on the use of population-specific equations to assess intracellular and extracellular water distribution. The MF-BIA system used in this study is a recently developed version that is not based on statistical population data and is capable of accurately assessing subjects with different body shapes and also obese subjects. Finally, ADP measures body density and is used more easily than other more complex systems for measuring body density, such as underwater weighing, and provides comparable results for obese subjects. Therefore, the use of 3 validated methods that use different principles was relevant for evaluating patients in different stages of a body weight reduction program.

Previous studies have shown that ketogenic diets preferably reduce the total FM in obese patients (10–13). However, the precise distribution of these losses has not been determined. In this study we confirmed that the diet reduces total FM and specifically visceral adipose tissue, which has a greater impact in predicting cardiometabolic complications associated with obesity than does the total volume of body adiposity (2, 31).

However, maintaining muscle mass and its functionality (i.e., muscle strength) has an important role in preventing weight regain, maintaining physical functionality, improving cardiometabolic risk factors, and reducing cardiovascular outcomes (5, 7–9, 32, 33). It is commonly assumed, and stated in several textbooks on obesity, that weight loss is associated with an important loss of muscle mass that evolves in parallel with the fat reduction. Some dietary guidelines have even suggested that diets that induce rapid weight loss, such as VLCK diets, create a greater energy deficit and contain lower amounts of protein, and therefore increase the risk of reductions in muscle mass compared with other interventions with more gradual weight loss (34, 35). In this article it was shown that the reductions observed in lean mass were mostly a result of body water loss, both intra- and extracellular. The combined information from the DXA and MF-BIA methods allows for such differentiation (28, 36–38). The loss attributable to muscle mass was minimal (∼1 kg), and an absolute preservation of HG strength was observed, a remarkable fact considering that the patients have experienced a weight reduction of ∼20 kg.

Various mechanisms may explain the variations in body water. For example, glycogen depletion induced by VLCK diets could cause a marked increase in diuresis, given that glycogen is usually stored together with water (39, 40). Water loss might also be associated with ketonuria, because ketone bodies increase the renal sodium and water loss as a result (39, 41). These assumptions seem reasonable considering that the peak water loss coincides with the phase of maximum ketosis. However, the mechanisms explaining the diuresis observed with VLCK and with most hypocaloric diets are not known at present (30). Contrary to previous observations (42, 43), DXA analysis evidenced a maintenance in bone mineral density in the current study.

Although in most clinical settings, BMI and waist circumference are used because they are inexpensive and convenient, it is evident that they are not able to precisely determine excess fat mass and its loss during treatment (44). More precise techniques to assess body composition are needed in specialized clinical settings and for research purposes. Therefore, another target of this work was to compare the accuracy of the information provided by the more expensive and less convenient DXA, currently considered the gold standard, with the less expensive and more convenient MF-BIA, as well as with ADP, which is only used in highly specialized centers because of its high cost (45). The results obtained showed that MF-BIA correlates very well with DXA, although with a tendency to slightly underestimate the FM%. These results are consistent with previous work that found that MF-BIA may overestimate the FFM, and thus produce an underestimation of the FM and FM% (45). MF-BIA provided highly relevant information about the water component during dieting. On the other hand, the ADP instrument showed a lower correlation with DXA and a greater variability in estimating the FM%. Compared with DXA, ADP underestimates the FM% in thinner patients, and overestimates the FM% in those patients with a higher body fat. The 3 techniques correlated remarkably well, although the less expensive MF-BIA performed with high precision.

An important strength of this study was the use of 3 different techniques for determining body composition in different settings, i.e., obesity and no ketosis, marked reduction in body weight with high ketosis, and finally, substantial reduction in body weight without ketosis. The tight control of adherence by daily measurement of B-OHB is another relevant strength of this work. A potential limitation of our study could be the sample size; however, because each subject underwent 4 evaluations, enabling each individual subject’s own results to be compared, this adds statistical power to the study and a real difference between the experimental points.

In conclusion, this study exhaustively assessed body composition changes during the weight reduction process induced by a VLCK diet. This assessment was performed using 3 complementary techniques: DXA, MF-BIA, and ADP. The VLCK diet induced a marked weight loss that was achieved mainly at the expense of total FM, and also visceral fat, with a maximum conservation of muscle mass and muscle strength. Additionally, MF-BIA was demonstrated to be an effective and convenient alternative for measuring body composition in clinical practice due to its minimal burden for the patient, ease of operation, low cost, and high accuracy.

Abbreviations:

 
• ADP

  air displacement plethysmography

 
• ALM

  appendicular lean mass

 
• ASLM

  appendicular soft lean mass

 
• BMI

  body mass index

 
• B-OHB

  B-hydroxybutyrate

 
• DXA

  dual-energy X-ray absorptiometry

 
• FFM

  fat-free mass

 
• FM

  fat mass

 
• HG

  handgrip strength

 
• MF-BIA

  multifrequency bioelectrical impedance

 
• VLCK

  very-low-calorie ketogenic

Acknowledgments

We acknowledge the PronoKal Group® for providing the high-biological-value protein preparations for all of the patients free of charge. In addition, we thank A. Menarini Diagnostics, Spain, for providing the portable meters for all of the patients free of charge.

This work was supported by the PronoKal Group^® and by grants from the Fondo de Investigacion Sanitaria (PE13/00024 and PI14/01012 research projects) and CIBERobn (CB06/003), Instituto de Salud Carlos III–Subdireccion General de Evaluacion y Fomento de la Investigación; Fondo Europeo de Desarrollo Regional, and the Health Department of the Xunta de Galicia, Spain. D.G.A. is grateful to the Colombian Department of Science, Technology and Innovation – COLCIENCIAS as a recipient of their predoctoral scholarship to support his work. The funding sources had no involvement in the study design, recruitment of patients, study interventions, data collection, or interpretation of the results. A PronoKal representative (I.S.) was involved in the study design and revised the final version of the manuscript, without intervention in the analysis of data, statistical evaluation, or final interpretation of the results of this study.

Disclosure Summary: D.B. and F.F.C. received advisory board fees and/or research grants from PronoKal Protein Supplies Spain. I.S. is Medical Director of PronoKal Spain. The remaining authors have nothing to disclose.

Author notes

Address all correspondence and requests for reprints to: Felipe F. Casanueva, MD, PhD, Division of Endocrinology, Department of Medicine, Complejo Hospitalario Universitario de Santiago, Travesia da Choupana Street s/n, 15706 Santiago de Compostela, La Coruña, Spain. E-mail: [email protected].