Abstract
Background
Indirect calorimetry (IC) is the gold standard for determining energy requirements. Resting energy expenditure (REE) equations could noninvasively estimate energy requirements in healthy individuals. Whether the published equations could accurately predict the REE of adults with nonalcoholic fatty liver disease (NAFLD) and type 2 diabetes mellitus (T2DM) remains unclear. Here, we aimed to investigate the accuracy of predicted REE in patients with NAFLD and T2DM using different equations.
Methods
REE was measured using IC and compared with ten predictive equations in 88 adult patients aged 30-53 years. The agreement between the measured REE with IC and REE, as predicted by equations, was assessed using the Bland-Altman method.
Results
The analysis showed that the FAO/WHO/UNU (version using weight) equation had the smallest average bias (10.2 kcal/d; 95% confidence interval [CI], –57.4 to 78) and the highest accuracy (46.5%); meanwhile, the Thumb equation had the greatest average bias (-402.2 kcal/d; 95% CI, –477.3 to -327.1) and the lowest accuracy (20.4%). The highest values of overestimating and underestimating were related to the Thumb and Owen equations, respectively. The 95% limits of agreement were found to be smaller in the equations by Muller (version using fat-free mass), compared to others.
Conclusion
The FAO/WHO/UNU (weight) equation, thus, performed the best in predicting energy expenditure when applied to patients with NAFLD and T2DM. However, all the prediction equations provided REE estimates within 10% of the measured ones in less than 50% of the cases.
Introduction
Nonalcoholic fatty liver disease (NAFLD) is the leading cause of liver disease worldwide, with the global prevalence rate of 25% in the general population and an even higher prevalence in patients with type 2 diabetes mellitus (T2DM) (55%) [1]. Lifestyle modifications including weight loss and nutritional recommendations are considered the first line of treatment for NAFLD [2]. The knowledge of energy requirements is essential for nutritional recommendations in those with metabolic disorders like T2DM or NAFLD [3].
Total energy expenditure (TEE) is the amount of energy burned to maintain body functions. TEE can be proportioned into resting energy expenditure (REE), diet-induced thermogenesis (energy expended to digest, absorb, and store food), and physical activity (spontaneous and other physical activity of daily living) [4, 5]. REE is the largest component (up about 2/3) that refers to the energy measured after an overnight fast at rest, in a thermo-neutral condition [6]. REE can be measured by the indirect calorimetric (IC) technique, which relies on the measurement of oxygen consumption (VO2) and carbon dioxide production (VCO2) [7]. Although IC represents the gold standard for the estimation of energy expenditure, it has limitations in terms of higher prices and less availability [8]. Accordingly, efforts have been made to develop prediction equations for REE by using sex, age, and anthropometric measures. Different equations including Mifflin-St. Jeor (known as Mifflin), and Harris-Benedict have been used to calculate energy expenditure. Typically, these predictive equations have been developed and validated in healthy populations [8]. However, some conditions like obesity or diabetes may influence the accuracy of estimated REE [9]. Previous studies suggested that patients with NAFLD or T2DM may have higher levels of REE [10–13]. Moreover, prediction equations may underestimate or overestimate the REE measured in conditions like T2DM or NAFLD [14, 15]. Therefore, using these equations in the subjects with T2DM and NAFLD may lead to inaccurate estimation of REE.
In this study, we aimed to assess the validity and accuracy of commonly used predictive equations for the predicted REE, as compared with measured REE using IC, in patients with NAFLD and T2DM.
Methods
Study population
The present study is a cross-sectional analysis of the data collected between September 2017, and March 2018, from patients who were diagnosed with T2DM accompanied with NAFLD. The study was conducted at the Endocrinology and Metabolism Research Center of Tehran University of Medical Sciences, in Tehran, Iran, and the protocol is available at IRCT.IR (IRCT201707024010N21). The study was performed in accordance with the Helsinki Declaration and received the approval of the National Institute for Medical Research Developments Ethics Committee (IR.TUMS.MEDICINE.REC.1399.1226). Informed consent was obtained from all recruited patients. Individuals with confirmed moderate or severe NAFLD and T2DM, aged 30–53 years, were included in the current study. The diagnostic criteria for T2DM include any of the following: [1] fasting plasma glucose level of 126 mg/dL or higher [2], 2-hour plasma glucose level of 200 mg/dL or higher after a 75 g oral glucose load [3], HbA1c ≥ 6.5% [4], a random glucose level of at or above 200 mg/dL in the presence of classic hyperglycemia symptoms or a hyperglycemic crisis [16]. Patients were considered to have moderate and severe NAFLD based on FibroScan controlled attenuation parameter (CAP) score. Moderate fatty liver (S2) is defined by a CAP value of 260–290 dB/m, while severe fatty liver (S3) is identified by a CAP value >290 dB/m [17].
CAP scores range from 100 to 400 dB/m and measure fatty content in the liver, independently from the presence of fibrosis. CAP score measures the gradual decrease of intensity and amplitude of the ultrasound waves in the liver with steatotic liver and compares it with the attenuation in normal liver [18].
Eligible patients were negative for the presence of viral hepatitis (hepatitis B surface antigen, anti-hepatitis C virus antibody) and any other liver diseases. Patients were excluded if they were receiving insulin therapy, were lactating or pregnant, had a history of current or passive alcohol consumption, or had cardiovascular diseases, thyroid disorders, kidney diseases, cancer, mental disorders, or any other acute or chronic diseases except for obesity and hypertension.
Measurement of resting energy expenditure
REE was measured using Fitmate Calorimeter (Cosmed, Rome, Italy), while the patients were awake and lying in the supine position, quiet and motionless. All testing was conducted in the morning (between 7:00 AM and 10:00 AM), in a sound-controlled and temperature-controlled room, after at least 30 min of resting. Period gas and flow calibration were performed before each measurement. VO2 (L/min) and VCO2 (L/min) were measured continuously for 20 min, discarding the first 5 min of measurement. VCO2 and VO2 were converted to REE by using Weir’s equation [19]. Participants were instructed not to consume caffeine or tobacco products, and to restrict moderate- or high-intensity physical activity during the 24 h prior to the test. Moreover, the subjects were asked to fast for 8 to 12 h before REE measurement (water freely allowed).
Predictive equations
REE was estimated in our study by 10 commonly used prediction equations in our institution, as summarized in Table 1. The equations of interest were as follows: The Food and Agriculture Organization/World Health Organization/United Nations University (FAO/WHO/UNU) equations were developed in 1985, based on a large population from previous studies, and categorized into different age groups [20], which were later revised and height was also included as an independent variable [21]. Harris-Benedict equations represent the principle of generating equations and were based on a sample size of 239 individuals aged 16–63 years [22]. The first revised version of Harris-Benedict was then developed by Roza and Shizgal, based on an older sample population with nearly equally distributed gender (169 women and 168 men) [23]. Livingston & Kohlstadt generated a formula with weight alone as an independent variable [24]. Muller predictive equations were established based on data obtained from 2528 subjects aged 5–91 years. Two sets of BMI group–specific prediction equations were developed using weight, height, sex, and age in Muller (weight) or Fat-free mass (FFM), fat mass (FM), sex, and age in Muller (FFM) [25]. The Mifflin equation was another revision of the Harris-Benedict equation, based on a larger sample population (n = 498), with a wider age range (19–78 years) [26]. Owen, in two consecutive years, established new equations using weight as the only variable. In 1986, the Owen equation for women was presented based on a sample population of 44 women aged 18–65 with body weight ranging from 43 to 171 kg [27]. A year later, the equation for men was developed using data from a sample of 60 men aged 18 to 62 years, with body weights ranging from 60 to 171 kg [28]. We also included a simple equation, which estimates REE by multiplying weight (kg) by 25 (known as Thumb) [29]. Then, the measured REE obtained by IC was compared with the predicted REE values.
Table 1.
Selected predictive equations for estimating resting energy expenditure
| Equation name | Unit | Equation |
|---|---|---|
|
FAO/WHO/UNU (weight) [20] |
kcal/d |
Male: (Age: 18–30 y): 15.3 × WT (Kg) + 679 (Age: 30–60 y): 11.6 × WT (Kg) + 879 Female: (Age: 18–30 y): 14.7 × WT (Kg) + 496 (Age: 30–60 y): 8.7 × WT (Kg) + 829 |
|
FAO/WHO/UNU (weight and height) [21] |
kcal/d |
Male: (Age: 18–30 y): 15.4 × WT (Kg) − 27× HT (m) + 717 (Age: 30–60 y): 11.3 ×WT (Kg) − 16 × HT (m) + 901 Female: (Age: 18–30): 13.3 × WT(Kg) + 334 × HT(m) + 35 (Age: 30–60 y): 8.7 × WT(Kg) − 25 × HT(m) + 865 |
| Harris-Benedict [22] | kcal/d |
Male: WT (Kg) × 13.7516 + HT (cm) × 5.0033 − Age × 6.755 + 66.473 Female: WT (Kg) × 9.5634 + HT (cm) × 1.8496 − Age × 4.6756 + 655.0955 |
|
Harris-Benedict (Revised) [23] |
kcal/d |
Male: WT (Kg) × 13.397 + HT (cm) × 4.799 − 5.677 × Age + 88.362 Female: WT (Kg) × 9.247 + HT (cm) × 3.098 – Age × 4.33 + 477.593 |
| Livingston [24] | kcal/d |
Male: 293 × WT (Kg)0.433 −5.92 × Age Female: 248 × WT (Kg)0.4356 −5.09 × Age |
| Müller (weight) [25] | MJ/d |
Male: (BMI: 18.5–25): 0.02219 × WT (Kg) + 0.02118 × HT (cm) + 0.884 × 1-0.01191 × Age + 1233 (BMI: 25–30): 0.04507 × WT (Kg) + 1.006 × 1- 0.01553 × Age + 3.407 (BMI: ≥ 30): 0.05 × WT(Kg) + 1.103 × 1- 0.01586 × Age + 2.924 Female: (BMI: 18.5–25): 0.02219 × WT (Kg) + 0.02118 × HT (cm) − 0.01191 × Age + 1233 (BMI: 25–30): 0.04507 × WT (Kg) − 0.01553 × Age + 3.407 (BMI: ≥ 30): 0.05 × WT (Kg) − 0.01586 × Age + 2.924 |
| Müller (FFM) [25] | MJ/d |
(BMI: 18.5–25): 0.0455 × FFM (Kg) + 0.0278 × FM (Kg) + 0.879 × Sex – 0.01291 × Age + 3.634 (BMI: 25–30): 0.03776 × FFM (Kg) + 0.03013 × FM (Kg) + 0.93 × Sex – 0.01196 × Age + 3.928 (BMI: ≥ 30): 0.05685 × FFM (Kg) + 0.04022 × FM (Kg) + 0.808 × Sex – 0.01402 × Age + 2.818 Sex (Male = 1, Female = 0) |
| Mifflin-St Jeor [26] | kcal/d |
Male: (9.99 × WT [kg]) + (6.25 × HT [cm]) – (4.92 × Age) + 5 Female: (9.99 × WT [kg]) + (6.25 × HT [cm]) – (4.92 × Age) – 161 |
| Owen [27, 28] | kcal/d |
Male: WT (Kg) × 10.2 + 879 Female: WT (Kg) × 7.18 + 795 |
| Thumb [29] | kcal/d | 25 × WT (Kg) |
FAO/WHO/UNU Food and Agriculture Organization/World Health Organization/United Nations University, WT weight (kg), HT height, BMI body mass index, FFM fat-free mass, FM fat mass
Clinical and laboratory data
Weight, with a precision of 0.1 kg, and body composition FFM, FM, and total body water (TBW) were estimated by bioelectrical impendence analysis (BIA); this was done using the Tanita scale (BC 418 MA Segmental Body Composition Analyzer, Tanita, Japan), with the subjects in light clothes and without shoes. Height was measured to the nearest 0.1 cm. Body mass index (BMI) was calculated as weight (in kilograms) divided by the square of height (in meters). Waist circumference (WC), systolic blood pressure (SBP), and diastolic blood pressure (DBP) were obtained using standardized procedures. The current smoker referred to someone who had smoked within the past 6 months.
Venous blood samples were obtained after a 10–12 h fasting. Fasting glucose and hemoglobin A1c (HbA1c) levels were determined by a glucose oxidase method on an autoanalyser (Cobas c 311, Roche) and high-performance liquid chromatography analyzer (Tosoh), respectively. Serum cholesterol, high-density lipoprotein (HDL), and triglycerides were measured using ELISA kit (Roche, Germany). The low-density lipoprotein (LDL) cholesterol value was calculated using the Friedewald equation.
The homeostasis model assessment of insulin resistance (HOMA-IR) was calculated as previously described [30]. Fasting serum insulin, C-peptide, and thyroid-stimulating hormone (TSH) levels were determined by the ELISA kit (Monobind; USA). Aspartate aminotransferase (AST), alanine aminotransferase (ALT), gamma-glutamyl transferase (GGT), high-sensitivity C-reactive protein (hs-CRP), and creatinine were measured using the ELISA method (Roche, Germany). The serum levels of adiponectin, cytokeratin-18 (CK-18), and total antioxidant capacity (TAC) were also assessed using the ELISA method (ZellBio GmbH, Ulm, Germany). Tumor necrosis factor-α (TNF-α) levels were then determined by the ELISA kit (Diaclone, France).
All patients were examined using FibroScan® 502 machine (EchoSense, Paris, France, 5 MHz) by an experienced investigator. Patients were examined after at least 3 h of fasting with either an M probe or an XL one. The cutoff for the detection of liver steatosis was set as ≥ 260 dB/m [17].
Statistical analyses
Bias, as a measure of accuracy at group level, was calculated as mean percentage error differences between predicted REE by equations and measured REE by IC. Bias was considered to be acceptable if it fall within ± 5% [31]. The proportion of the patients with a predicted REE within ± 10% of the measured one was determined as a measure of accuracy at the individual level [32]. Therefore, values lower than 90% or higher than 110% were classified as underestimation or overestimation, respectively. The mean difference, 95% confidence intervals of differences, and the 95% limits of agreements (mean differences ± 1.96 standard deviation [SD] of difference) were calculated. Bland-Altman plots were also applied to evaluate the limits of agreement between the predicted REE with various equations and the measured one. P values < 0.05 were considered statistically significant.
Results
A total of 88 eligible patients were included in the current study. The general and clinical characteristics of patients are summarized in Table 2. The mean ± SD age of the patients was 44.4 ± 5.5 years, with 58% being male. BMI values ranged from 24.0 to 45 kg/m2 with a mean ± SD of 30.5 ± 3.7. According to BMI values, 52% (n = 46) of the patients were obese (BMI ≥ 30) and 45% (n = 40) were overweight (BMI = 25–30). The mean ± SD of FFM and FM were 61.7 ± 11 and 27 ± 8.3 Kg, respectively.
Table 2.
Baseline characteristics of included patients
| Variable | Total (n = 88) |
|---|---|
| Age (year) | 44.4 ± 5.5 |
| Sex (male) | 51 (58) |
| Current smoker (yes) | 16 (18.2) |
| Duration of diabetes (year) | 4.5 ± 4.1 |
| Weight (kg) | 89.2 ± 12.5 |
| Height (cm) | 170.8 ± 9.1 |
| BMI (kg/m2) | 30.5 ± 3.7 |
| Waist (cm) | 107.7 ± 9 |
| Fat mass (Kg) | 27 ± 8.3 |
| Fat-free mass (Kg) | 61.7 ± 11 |
| TBW (kg) | 45.1 ± 8 |
| CAP score (dB/m) | 318.3 ± 37.2 |
| Fibrosis score (kPa) | 6.2 ± 1.9 |
| ALT (U/L) | 22.1 ± 9.8 |
| AST (U/L) | 25.5 ± 10.3 |
| GGT (U/L) | 40 ± 30.2 |
| HOMA-IR score | 4.3 ± 2.3 |
| HbA1C (%) | 7.8 ± 1.6 |
| Fasting glucose (mg/dL) | 142.5 ± 52.8 |
| Insulin (uIU/mL) | 11.9 ± 4.8 |
| C-peptide (ng/mL) | 1.7 ± 0.6 |
| Triglyceride (mg/dL) | 212 ± 212.5 |
| LDL (mg/dL) | 89.8 ± 31.7 |
| HDL (mg/dL) | 35.7 ± 9.4 |
| Cholesterol (mg/dL) | 166.9 ± 36.7 |
| hs-CRP (mg/L) | 3.9 ± 6.5 |
| TAC (mmol/L) | 0.3 ± 0 |
| TNF-α (pg/mL) | 28.4 ± 22.5 |
| Adiponectin (mg/L) | 7.3 ± 4.1 |
| CK-18 fragments (U/L) | 1.2 ± 0.3 |
| TSH (mIU/L) | 2 ± 1.4 |
| Creatinine (mg/dL) | 0.9 ± 0.1 |
| Physical activity (METs/h) | 31.3 ± 4.2 |
| Systolic blood pressure (mmHg) | 122.2 ± 13.7 |
| Diastolic blood pressure (mmHg) | 83.9 ± 10.4 |
Data are presented as mean ± SD or number (%)
SD standard deviation, BMI body mass index, TBW total body water; CAP controlled attenuation parameter, ALT alanine aminotransferase, AST aspartate aminotransferase, GGT gamma-glutamyl transferase, HOMA-IR homeostasis model assessment of insulin resistance, HbA1c hemoglobin A1c, LDL low-density lipoprotein, HDL high-density lipoprotein, hs-CRP high-sensitivity C-reactive protein, TAC total antioxidant capacity, TNF-α tumor necrosis factor-α, CK-18 cytokeratin-18, TSH thyroid stimulating hormone, MET metabolic equivalent of task
Table 3 represents the mean differences and bias percentages between the measured and predicted REE values obtained from the equations. Of all the equations studied, the FAO/WHO/UNU (weight) equation had the smallest bias percentage (0.5%). The bias percentages were acceptable for FAO/WHO/UNU (weight) (0.5%), Harris-Benedict (revised version) (1.7%), FAO/WHO/UNU (weight and height) (1.8%), and Muller (weight) (3.2%) equations.
Table 3.
Comparisons between measured and predicted resting energy expenditure values
| Equation | Mean difference1 (average bias; 95% CI) | Bias % | 95% limits of agreement2 (range of difference) |
|---|---|---|---|
| FAO/WHO/UNU (weight) | 10.2 (−57.4 to 78) | 0.5 | −1045.8 to 949.2 |
| FAO/WHO/UNU (weight and height) | 33.9 (−33.1 to 101.1) | 1.8 | −1005.4 to 954.7 |
| Harris-Benedict | 36.9 (−31.2 to 105.1) | 5.6 | −1036 to 933.1 |
| Harris-Benedict (revised) | 32.0 (−35.7 to 99.7) | 1.7 | −1034.2 to 912.5 |
| Livingston | 144.1 (76.9 to 211.4) | 8.5 | −835 to 1032.9 |
| Mifflin | 142.6 (75.3 to 209.8) | 8.4 | −861.2 to 106.4 |
| Muller (weight) | 57.0 (−10.1 to 124.2) | 3.2 | −981.7 to 960.5 |
| Muller (FFM) | 102.1 (38.4 to 165.8) | 5.8 | −621.9 to 966.2 |
| Owen | 151.5 (83.4 to 219.7) | 9 | −887.6 to 1089.9 |
| Thumb | −402.2 (−477.3 to −327.1) | −18 | −1881 to 634 |
REE resting energy expenditure, CI Confidence interval, FAO/WHO/UNU Food and Agriculture Organization/World Health Organization/United Nations University, FFM fat-free mass; SD standard deviation
1measured REE – estimated REE (kcal)
2mean difference ± 1.96 SD of the difference
The range of the 95% limits of agreement was smaller in the equation by Muller (FFM) (−621.9 to 966.2 kcal/day), as compared to other ones. Figure 1 also shows the Bland-Altman plots including mean difference and limits of agreements between predicted and measured REE values.
Fig. 1 .
Bland-Altman plots of the various predictive equations compared with measured resting energy expenditure with indirect calorimetry. The upper and lower lines represent 2 standard deviations from the mean (limits of agreement)
Abbreviations: mREE, measured resting energy expenditure; HB, Harris-Benedict; FAO/WHO/UNU, Food and Agriculture Organization/World Health Organization/United Nations University; FFM, fat-free mass
As shown in Table 4, the accuracy at the individual level (the percentage of subjects with a predicted REE within ± 10% of the measured REE) was lower for all selected equations. The FAO/WHO/UNU (weight) equation, with an accuracy of 46.5%, was more accurate than other equations. On the other hand, the Thumb equation showed the least accuracy (20.4%). The highest values of overestimating and underestimating, relative to the measured REE, were related to the Thumb (overestimation = 75%) and Owen (underestimation = 48.8%) equations, respectively (Table 4).
Table 4.
Accuracy of prediction equations
| Equations | Overestimation (measured REE/predicted REE < 0.9), % | Accurate estimation (measured REE/predicted REE 0.9–1.1), % | Underestimation (measured REE/predicted REE > 1.1), % |
|---|---|---|---|
| FAO/WHO/UNU (weight) | 26.1 | 46.5 | 27.2 |
| FAO/WHO/UNU (weight and height) | 25 | 44.3 | 30.6 |
| Harris-Benedict | 23.8 | 44.3 | 31.8 |
| Harris-Benedict (revised) | 23.8 | 44.3 | 31.8 |
| Livingston | 13.6 | 39.7 | 46.5 |
| Mifflin | 14.7 | 40.9 | 44.3 |
| Muller (weight) | 23.8 | 43.1 | 32.9 |
| Muller (FFM) | 17.0 | 44.3 | 37.5 |
| Owen | 14.7 | 36.3 | 48.8 |
| Thumb | 75.0 | 20.4 | 4.5 |
REE resting energy expenditure, FAO/WHO/UNU Food and Agriculture Organization/World Health Organization/United Nations University, FFM fat-free mass
We also re-evaluated the prediction equations using a higher CAP score cut-off (>290 dB/m) for the diagnosis of NAFLD [33]. The accuracy of the prediction equations did not change significantly when applying this stricter criterion for defining NAFLD (data not shown).
Discussion
In our study, the highest accuracy was attained by the FAO/WHO/UNU (weight) equation, with 0.5% bias (almost 47% of accurate predictions), while the Thumb equation, with 18% bias (20% of accurate predictions), had the lowest accuracy. Our study, thus, indicated that out of ten energy estimation equations, only the FAO/WHO/UNU (weight) equation was suitable for estimating REE with bias < 1%, but its estimation accuracy was less than 50%. The results also suggested that none of the ten included equations could reliably estimate the individual differences in this particular population, as they all had an accuracy of less than 50% estimates. Therefore, efforts should be made to develop novel disease-specific equations with higher accuracy, specifically for metabolic conditions like T2DM and NAFLD.
Given the characteristics of NAFLD, diabetic patients are more prone to develop NAFLD and liver fibrosis [34]; additionally, the main recommendation for NAFLD is weight reduction [35]. Accordingly, it is important to correctly estimate their energy requirements. Recently, the validity of REE predictive equations has been evaluated for patients with T2DM [15, 36–39] or those with NAFLD [14, 40]. However, to the best of our knowledge, this is the first study specifically targeting patients with T2DM and NAFLD.
In our study, the most accuracy was found by applying the FAO/WHO/UNU (weight) equation, which was in agreement with the previous studies on diabetic patients. Steemburgo et al. evaluated the REE in diabetic patients aged 48–70, with BMI of 29.4 kg/m2 (20.2–37.4) and reported that the FAO/WHO/UNU (weight) equation performed the best (bias% = 5.6%) [36]. Another study in Brazil also showed that the FAO/WHO/UNU equation performed better in the females with T2DM (bias% = 1.8%) than male counterparts (bias% = 2.4%) [37], indicating that sex differences might be a factor that should be considered in the accuracy of predictive equations. However, some studies found controversial results. Figueiredo Ferreira et al. reported that the Owen equation provided an accurate prediction of REE (bias% = −0.5%) in 28 adult women with T2DM [15], while the study of 283 newly diagnosed T2DM patients found that the Mifflin equation had the best accuracy (accurate prediction of 48.8% for ± 10% of measured REE [39].
To date, two studies have compared energy estimation formulas in NAFLD patients. Endo et al. conducted a study of 77 Japanese patients with NAFLD, showing that the Kyto equation, designed for the Japanese population, had the highest estimation accuracy to predict REE (accurate prediction of 71.4% for ± 10% of measured REE). They also indicated that the accuracy of FAO/WHO/UNU (weight), FAO/WHO/UNU (weight and height), and Schofield equations were less than 50% of Japanese patients with NAFLD [40]. The study of 67 patients with NAFLD also found that the Mifflin equation was not suitable for accurate REE estimation (bias% = 14.3%) in Brazilian patients with NAFLD, and Schofield (bias% = 3.2%) equation outperformed Harris-Benedict (bias% = −10.4%), and FAO/WHO/UNU (bias% = 4.2%) equations [14]. These conflicting findings could be due to variations in the body composition of different populations.
In line with our results, a study of 427 obese Saudi participants suggested that none of the energy estimation equations was suitable for estimating REE, in the absence of IC. They also showed that FAO/WHO/UNU and Schofield equations had a high percentage of estimation accuracy for obese Saudi women (about 55%), with a bias of less than 1% [41]. Altogether, the disparities among studies could be related to the differences in the population characteristics, methodological design, and disease-specific factors.
Several studies confirmed that NAFLD [13, 42] or T2DM [10–12] are associated with higher REE. The higher level of REE in NAFLD individuals might be attributed to chronic low-grade inflammation [43]. Inflammation is considerably related to increased energy expenditure [44]. This can be justified by some explanations. Inflammation is related to increased VO2 [45], increased concentration of catabolic hormones, higher lipolysis, and extensive protein catabolism [13, 46].
Moreover, it was found that metabolic rates are higher in morbidly obese NAFLD individuals with metabolic syndrome, compared with NAFLD individuals with similar weight without metabolic syndrome [43]. In conditions like NAFLD, diabetes, obesity, and metabolic syndrome increased reactive oxygen species and cytokine production may lead to mitochondrial disturbances and finally excess energy consumption and higher levels of metabolism [47].
Additionally, in diabetes, the elevated REE may be related to increased gluconeogenesis, which is an energy-costly metabolic pathway responsible for increased fasting plasma glucose [48–50]. This hypothesis is supported by the research that observed decreased REE in diabetes after administration of intravenous insulin, which is a known suppressor of gluconeogenesis [51] Increased protein turnover as an energy-consuming process is frequently observed in poorly controlled diabetic patients and may be another cause of higher REE [50]. Hyperglycemia can increase the loss of glucose in urine by 30–80 g/day, which is in other words energy loss of about 120–320 kcal/day [52, 53].
Finally, this study has limitations to be considered. The research was conducted on a limited sample size. The body composition was measured by the bioimpedance method and the results might be affected by factors like hydration status, size shape, or electrode positioning [54].
Conclusion
None of the common equations was suitable for estimating energy expenditure in the patients with NAFLD and T2DM, in our sample population. Even the best of the equations, the FAO/WHO/UNU (weight) equation, was accurate only in less than 50% of the patients. Given that nutritional support is very important in NAFLD patients and requires accurate energy estimation to prevent the complications of the disease, novel equations should be developed to calculate REE.
Acknowledgements
We are thankful to the patients who participated in the study.
Authors’ contributions
A.M. Study conception and design, Acquisition of data, Analysis and interpretation of data, Drafting of manuscript, Critical revision, S.E.M. Analysis and interpretation of data, Drafting of manuscript, A.H. Drafting of manuscript, A.H. Critical revision, M.Q. Analysis and interpretation of data, H.G. Acquisition of data, M.M. Critical revision, M.R.M.T. Critical revision, F.B. Acquisition of data, S.H.M. Critical revision, S.M.S.J. Study conception and design, Analysis and interpretation of data, Drafting of manuscript, Critical revision.
Funding
No funding to declare.
Data availability
The datasets used and/or analysed during the current study are available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
The study was performed in accordance with the Helsinki Declaration and received the approval of the Tehran University of Medical Sciences. Ethics Committee (IR.TUMS.MEDICINE.REC.1399.1226). Informed consent was obtained from all patients recruited.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Bergram M, Nasr P, Iredahl F, Kechagias S, Rådholm K, Ekstedt M. Low awareness of non-alcoholic fatty liver disease in patients with type 2 diabetes in Swedish primary health care. Scand J Gastroenterol. 2022;57(1):60–9. [DOI] [PubMed] [Google Scholar]
- 2.Kwak M-S, Kim D. Non-alcoholic fatty liver disease and lifestyle modifications, focusing on physical activity. Korean J Intern Med. 2018;33(1):64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Cioffi I, Marra M, Pasanisi F, Scalfi L. Prediction of resting energy expenditure in healthy older adults: a systematic review. Clin Nutr. 2021;40(5):3094–103. [DOI] [PubMed] [Google Scholar]
- 4.Ravussin E, Bogardus C. Relationship of genetics, age, and physical fitness to daily energy expenditure and fuel utilization. Am J Clin Nutr. 1989;49(5):968–75. [DOI] [PubMed] [Google Scholar]
- 5.Kreymann G, Adolph M, Mueller MJ. Energy expenditure and energy intake - Guidelines on parenteral nutrition, chap. 3. Ger Med Sci. 2009;7:Doc25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ndahimana D, Kim E-K. Measurement methods for physical activity and energy expenditure: a review. Clin Nutr Res. 2017;6(2):68–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Popp CJ, Tisch JJ, Sakarcan KE, Bridges WC, Jesch ED. Approximate time to steady-state resting energy expenditure using indirect calorimetry in young, healthy adults. Front Nutr. 2016;3:49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Martincevic I, Mouzaki M. Resting energy expenditure of children and adolescents with nonalcoholic fatty liver disease. J Parenter Enter Nutr. 2017;41(7):1195–201. [DOI] [PubMed] [Google Scholar]
- 9.Wang Z, Heshka S, Gallagher D, Boozer CN, Kotler DP, Heymsfield SB. Resting energy expenditure-fat-free mass relationship: new insights provided by body composition modeling. Am J Physiology-Endocrinology Metabolism. 2000;279(3):E539-45. [DOI] [PubMed] [Google Scholar]
- 10.Gougeon R, Lamarche M, Yale JF, Venuta T. The prediction of resting energy expenditure in type 2 diabetes mellitus is improved by factoring for glycemia. Int J Obes. 2002;26(12):1547–52. [DOI] [PubMed] [Google Scholar]
- 11.Ikeda K, Fujimoto S, Goto M, Yamada C, Hamasaki A, Shide K, et al. Impact of endogenous and exogenous insulin on basal energy expenditure in patients with type 2 diabetes under standard treatment123. Am J Clin Nutr. 2011;94(6):1513–8. [DOI] [PubMed] [Google Scholar]
- 12.Miyake R, Ohkawara K, Ishikawa-Takata K, Morita A, Watanabe S, Tanaka S. Obese Japanese adults with type 2 diabetes have higher basal metabolic rates than non-diabetic adults. J Nutri Sci Vitaminol. 2011;57(5):348–54. [DOI] [PubMed] [Google Scholar]
- 13.Reddavide R, Cisternino AM, Inguaggiato R, Rotolo O, Zinzi I, Veronese N, et al. Non-alcoholic fatty liver disease is associated with higher metabolic expenditure in overweight and obese subjects: a case-control study. Nutrients. 2019;11(8):1830. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Oliveira A, Fernandes SA, Carteri RB, Tovo CV. Evaluation of rest energy expenditure in patients with non alcoholic fatty liver disease. Arq Gastroenterol. 2021;58:157–63. [DOI] [PubMed] [Google Scholar]
- 15.de Figueiredo Ferreira M, Detrano F, Coelho GMO, Barros ME, Serrão Lanzillotti R, Firmino Nogueira Neto J et al. Body composition and basal metabolic rate in women with type 2 diabetes mellitus. J Nutri Metabol. 2014;2014:574057. [DOI] [PMC free article] [PubMed]
- 16.Organization WH. Definition, diagnosis and classification of diabetes mellitus and its complications: report of a WHO consultation. Part 1, Diagnosis and classification of diabetes mellitus. World health organization; 1999. [Google Scholar]
- 17.Huang Z, Ng K, Chen H, Deng W, Li Y. Validation of controlled Attenuation parameter measured by fibroscan as a novel surrogate marker for the evaluation of metabolic derangement. Front Endocrinol. 2021;12:739875. [DOI] [PMC free article] [PubMed]
- 18.Eddowes PJ, Sasso M, Allison M, Tsochatzis E, Anstee QM, Sheridan D, et al. Accuracy of fibroscan controlled attenuation parameter and liver stiffness measurement in assessing steatosis and fibrosis in patients with nonalcoholic fatty liver disease. Gastroenterology. 2019;156(6):1717–30. [DOI] [PubMed] [Google Scholar]
- 19.Weir JB. New methods for calculating metabolic rate with special reference to protein metabolism. J Physiol. 1949;109(1–2):1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.;Joint FAO/WHO/UNU Expert Consultation Report. Energy and Protein Requirements. Technical report series 724. Geneva: WHO; 1985. [PubMed]
- 21. FAO/WHO/UNU: Human energy requirements. Report of a Joint FAO/WHO/UNU Expert Consultation, Rome, 17–24 October 2001. 2004.
- 22.Harris JA, Benedict FG. A biometric study of human basal metabolism. Proc Natl Acad Sci U S A. 1918;4(12):370. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Roza AM, Shizgal HM. The Harris Benedict equation reevaluated: resting energy requirements and the body cell mass. Am J Clin Nutr. 1984;40(1):168–82. [DOI] [PubMed] [Google Scholar]
- 24.Livingston EH, Kohlstadt I. Simplified resting metabolic rate—predicting formulas for normal-sized and obese individuals. Obes Res. 2005;13(7):1255–62. [DOI] [PubMed] [Google Scholar]
- 25.Müller MJ, Bosy-Westphal A, Klaus S, Kreymann G, Lührmann PM, Neuhäuser-Berthold M, et al. World health organization equations have shortcomings for predicting resting energy expenditure in persons from a modern, affluent population: generation of a new reference standard from a retrospective analysis of a German database of resting energy expenditure. Am J Clin Nutr. 2004;80(5):1379–90. [DOI] [PubMed] [Google Scholar]
- 26.Mifflin MD, St Jeor ST, Hill LA, Scott BJ, Daugherty SA, Koh YO. A new predictive equation for resting energy expenditure in healthy individuals. Am J Clin Nutr. 1990;51(2):241–7. [DOI] [PubMed] [Google Scholar]
- 27.Owen OE, Kavle E, Owen RS, Polansky M, Caprio S, Mozzoli MA, et al. A reappraisal of caloric requirements in healthy women. Am J Clin Nutr. 1986;44(1):1–19. [DOI] [PubMed] [Google Scholar]
- 28.Owen OE, Holup JL, D’Alessio DA, Craig ES, Polansky M, Smalley KJ, et al. A reappraisal of the caloric requirements of men. Am J Clin Nutr. 1987;46(6):875–85. [DOI] [PubMed] [Google Scholar]
- 29.Rousseau A-F, Losser M-R, Ichai C, Berger MM. ESPEN endorsed recommendations: nutritional therapy in major burns. Clin Nutr. 2013;32(4):497–502. [DOI] [PubMed] [Google Scholar]
- 30.Matthews DR, Hosker J, Rudenski A, Naylor B, Treacher D, Turner R. Homeostasis model assessment: insulin resistance and β-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28(7):412–9. [DOI] [PubMed] [Google Scholar]
- 31.Frankenfield D, Roth-Yousey L, Compher C, Group EAW. Comparison of predictive equations for resting metabolic rate in healthy nonobese and obese adults: a systematic review. J Am Diet Assoc. 2005;105(5):775–89. [DOI] [PubMed] [Google Scholar]
- 32.Carpenter A, Pencharz P, Mouzaki M. Accurate estimation of energy requirements of young patients. J Pediatr Gastroenterol Nutr. 2015;60(1):4–10. [DOI] [PubMed] [Google Scholar]
- 33.Eskridge W, Vierling JM, Gosbee W, Wan GA, Hyunh ML, Chang HE. Screening for undiagnosed non-alcoholic fatty liver disease (NAFLD) and non-alcoholic steatohepatitis (NASH): a population-based risk factor assessment using vibration controlled transient elastography (VCTE). PLoS ONE. 2021;16(11):e0260320. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Tomah S, Alkhouri N, Hamdy O. Nonalcoholic fatty liver disease and type 2 diabetes: where do diabetologists stand? Clin Diabetes Endocrinol. 2020;6(1):1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Holmer M, Lindqvist C, Petersson S, Moshtaghi-Svensson J, Tillander V, Brismar TB, et al. Treatment of NAFLD with intermittent calorie restriction or low-carb high-fat diet–a randomised controlled trial. JHEP Rep. 2021;3(3):100256. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Steemburgo T, Lazzari C, Farinha JB, Paula TPd, Viana LV, Oliveira ARd, et al. Basal metabolic rate in Brazilian patients with type 2 diabetes: comparison between measured and estimated values. Arch Endocrinol Metab. 2019;63:53–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Grassi T, Boeno FP, de Freitas MM, de Paula TP, Viana LV, de Oliveira AR, et al. Predictive equations for evaluation for resting energy expenditure in Brazilian patients with type 2 diabetes: what can we use? BMC Nutr. 2020;6(1):1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Tabata S, Kinoshita N, Yamada S, Matsumoto H. Accuracy of basal metabolic rate estimated by predictive equations in Japanese with type 2 diabetes. Asia Pac J Clin Nutr. 2018;27(4):763–9. [DOI] [PubMed] [Google Scholar]
- 39.Doros R, Lixandru D, Petcu L, Tudosoiu J, Mitu M, Picu A, et al. Resting metabolic rate in type 2 diabetes–accuracy of predictive equations. Rom Biotechnol Lett. 2016;22:12033–8. [Google Scholar]
- 40.Endo K, Kakisaka K, Oikawa K, Endo R, Takikawa Y. Comparison of predicted energy expenditure in Japanese patients with non-alcoholic fatty liver disease to establish a suitable nutrition intervention. J Nutr Sci Vitaminol. 2016;62(2):108–15. [DOI] [PubMed] [Google Scholar]
- 41.Almajwal AM, Abulmeaty M. New predictive equations for resting energy expenditure in normal to overweight and obese population. Int J Endocrinol. 2019;2019:5727496. [DOI] [PMC free article] [PubMed]
- 42.Ye Q, Yan J, Xiao HJ, Han T. Value of visceral fat area and resting energy expenditure in assessment of metabolic characteristics in obese and lean nonalcoholic fatty liver disease. Turk J Gastroenterol. 2021;32(2):116–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Tarantino G, Marra M, Contaldo F, Pasanisi F. Basal metabolic rate in morbidly obese patients with non-alcoholic fatty liver disease. Clin Invest Med. 2008;31(1):E24-9. [DOI] [PubMed] [Google Scholar]
- 44.Utaka S, Avesani CM, Draibe SA, Kamimura MA, Andreoni S, Cuppari L. Inflammation is associated with increased energy expenditure in patients with chronic kidney disease. Am J Clin Nutr. 2005;82(4):801–5. [DOI] [PubMed] [Google Scholar]
- 45.Starnes H, Warren RS, Jeevanandam M, Gabrilove JL, Larchian W, Oettgen H, et al. Tumor necrosis factor and the acute metabolic response to tissue injury in man. J Clin Invest. 1988;82(4):1321–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Buttgereit F, Burmester G-R, Brand MD. Bioenergetics of immune functions: fundamental and therapeutic aspects. Immunol Today. 2000;21(4):194–9. [DOI] [PubMed] [Google Scholar]
- 47.Esser N, Legrand-Poels S, Piette J, Scheen AJ, Paquot N. Inflammation as a link between obesity, metabolic syndrome and type 2 diabetes. Diabetes Res Clin Pract. 2014;105(2):141–50. [DOI] [PubMed] [Google Scholar]
- 48.Consoli A, Nurjhan N, Capani F, Gerich J. Predominant role of gluconeogenesis in increased hepatic glucose production in NIDDM. Diabetes. 1989;38(5):550–7. [DOI] [PubMed] [Google Scholar]
- 49.Landau BR, Wahren J, Chandramouli V, Schumann WC, Ekberg K, Kalhan SC. Contributions of gluconeogenesis to glucose production in the fasted state. J Clin Invest. 1996;98(2):378–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Buscemi S, Donatelli M, Grosso G, Vasto S, Galvano F, Costa F, et al. Resting energy expenditure in type 2 diabetic patients and the effect of insulin bolus. Diabetes Res Clin Pract. 2014;106(3):605–10. [DOI] [PubMed] [Google Scholar]
- 51.Prager R, Wallace P, Olefsky JM. Direct and indirect effects of insulin to inhibit hepatic glucose output in obese subjects. Diabetes. 1987;36(5):607–11. [DOI] [PubMed] [Google Scholar]
- 52.Caron N, Peyrot N, Caderby T, Verkindt C, Dalleau G. Energy expenditure in people with diabetes mellitus: A review. Front Nutr. 2016;3:56. [DOI] [PMC free article] [PubMed]
- 53.Piaggi P, Thearle MS, Bogardus C, Krakoff J. Fasting hyperglycemia predicts lower rates of weight gain by increased energy expenditure and fat oxidation rate. J Clin Endocrinol Metab. 2015;100(3):1078–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Ryan AS, Elahi D. Body: Composition, Weight, Height, and Build. In: Birren JE, editor. Encyclopedia of Gerontology (Second Edition). New York: Elsevier; 2007. pp. 177 – 86.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets used and/or analysed during the current study are available from the corresponding author upon reasonable request.