Energy expenditure and the accuracy of predictive equations in myotonic dystrophy type 1

Abstract

Background

Myotonic dystrophy type 1 (DM1) is associated with reduced physical activity, overweight and cardiovascular morbidity. Nutritional management requires accurate estimation of total daily energy expenditure (TEE), based on basal metabolic rate (BMR) and physical activity level (PAL). However, DM1-related changes in body composition may reduce the accuracy of commonly used BMR equations. This study evaluated BMR equation accuracy in DM1 versus controls, and assessed PAL and substrate oxidation.

Methods

In this secondary analysis of a prospective case-control study, 15 DM1 patients were compared with 15 age-, sex-, and BMI-matched controls. Body composition was measured using dual-energy X-ray absorptiometry. Overnight metabolic rate (OMR) was assessed by room calorimetry and compared with standard predictive equations (Harris-Benedict, WHO, Mifflin-St Jeor). Additionally, OMR was compared to body composition-based equations (Wang, Nelson, Sabounchi structures 4, 5, and 11). TEE was measured over 15 days using doubly labeled water. PAL was calculated as TEE/OMR, and substrate oxidation was assessed using the respiratory exchange ratio (RER).

Results

Standard predictive equations significantly overestimated metabolic rate in DM1, with median biases of +100 to +165 kcal/day (+7% to +12%, p<0.01), with no significant bias in controls. Structure 11 performed best in DM1 (+0.0%, p=1.000). PAL was lower in DM1 than in controls (1.42 vs. 1.69, p<0.001), whereas RER did not differ.

Conclusion

Common predictive equations overestimate energy requirements in DM1. Body composition-based approaches or correction factors may improve estimation, while low PAL should be considered when estimating TEE. These findings have direct implications for nutritional management in DM1.

Introduction

Myotonic dystrophy type 1 (DM1) is an autosomal dominant neuromuscular disease caused by a CTG repeat expansion within the DMPK gene. ¹ DM1 is characterized by progressive muscle weakness and myotonia, accompanied by metabolic manifestations including insulin resistance and dyslipidemia. ² In addition, overweight and obesity are common.^3–5

Body weight issues in this patient population likely reflect disease-related factors. First, muscle wasting is inherent to DM1 and will continue with disease progression. Second, motor impairments and diminished motivation or fatigue will decrease physical activity.^6,7 This might lead to increased cardiovascular morbidity and reduced quality of life. ⁸ In addition, insulin resistance and altered substrate metabolism may further contribute to overweight and obesity. ² Central nervous system involvement may also affect motivation and daily functioning, which could indirectly influence lifestyle behaviours, including dietary habits. ⁹ While a mismatch between energy uptake and energy expenditure can contribute to overweight and obesity, studies also report underweight as a clinically relevant feature, with cachexia observed in approximately 19.7% of patients.^3,7

Nutritional management can support a healthy body weight; an important component of nutritional management is the estimation of total daily energy expenditure (TEE) derived from basal metabolic rate (BMR) and physical activity level (PAL). ¹⁰ BMR, the main component of TEE, is largely determined by fat-free mass (FFM) and can be measured using indirect calorimetry. In clinical practice, BMR is commonly estimated using predictive equations such as the WHO, original Harris–Benedict (1918), revised Harris Benedict (2023) or Mifflin-St Jeor equation.^11–14 These equations rely on anthropometric variables, such as height and weight, and other physiological determinants of BMR, such as age and sex. Thereafter, PAL should be estimated, which is generally selected from standardized tables based on history taking by a nutritional expert. ¹¹

In DM1, disease-related factors may compromise correct estimation of TEE. Literature suggests that BMR may be lower in DM1-affected individuals than in healthy adults as a result of altered body composition.^5,15–17 Consequently, standard predictive equations for BMR, which are based on data from healthy participants, could systematically overestimate energy needs in the DM1 population.^18,19 Utilization of predictive equations incorporating body composition potentially represents a more valid approach.^20–22

In addition to BMR, PAL is the major determinant for TEE in adults. ¹¹ In patients with DM1, this factor remains difficult to approximate, as the amount of muscle weakness, fatigue, and cardiopulmonary involvement varies between and in patients over time, affecting the amount of physical activity. ²³

In addition to energy expenditure and energy requirements, better understanding how energy is utilized in the DM1 population would be of added value. Specifically, data on substrate oxidation and metabolic efficiency, estimated from the respiratory exchange ratio (RER) obtained via indirect calorimetry, could further optimize nutritional recommendations.

A previously performed case-control study by our research group examined body composition, physical activity, and energy expenditure using indirect calorimetry and the application of the doubly labeled water (DLW) approach in DM1 patients and healthy controls. ⁷ However, the individual components of TEE were not specifically addressed, and no guidance on the estimation of TEE in clinical practice for DM1-affected individuals was provided. Here, we extend our prior work through further analyses of previously acquired data from the same prospective case-control study. ⁷ Specifically, we focused on the accuracy of BMR predictive equations, PAL estimation, and substrate oxidation, which were not the primary focus of the original publication.

Methods

Study design, setting and participants

This study represents a retrospective secondary analysis of data previously acquired in a prospective case-control study conducted at Maastricht University Medical Centre+. ⁷ The cohort comprised 15 genetically confirmed adult-onset DM1 patients and 15 healthy controls, individually matched for sex, age, and BMI. DM1 patients were recruited through a study website, patient organisation, and treating physicians. Because only ambulatory adult-onset patients able to perform exercise tests were included, and those with severe muscle weakness, relevant comorbidities, or other factors that could interfere with study measurements were excluded, the cohort was relatively homogeneous. Neurological examination was performed by a trained neuromuscular neurologist to determine muscular impairment rating scale (MIRS) score (a disease-specific muscle impairment rating scale in myotonic dystrophy) in patients, and to confirm normal strength in healthy participants. ²⁴ DNA analysis took place at time of DM1 diagnosis. CTG-repeat lengths were determined by analyzing DNA extracted from peripheral blood samples through polymerase chain reaction (PCR), followed by fragment length analysis and Southern blot analysis. All participants provided written informed consent, and the study was approved by the institutional ethics committee (METC 182009).

The 24h-room calorimetry data were used to determine overnight metabolic rate (OMR), which can be used as a surrogate for BMR, as literature confirms that OMR closely approximates BMR. The 24h-room calorimetry protocol has been described in our previous publication. ⁷ OMR was calculated using calorimetry data from the 00:00–06:00 interval using the Weir equation and expressed as kcal/24 h for comparability. ²⁵

TEE was measured with the doubly labelled water (DLW) method. Participants ingested a bolus of ²H₂O and H₂¹⁸O, and urine samples were collected over the course of 14 days to determine isotope elimination, as part of the previously mentioned study. ⁷ CO₂ production and TEE were calculated using standard equations. ²⁶

Body composition was assessed by dual-energy X-ray absorptiometry (DEXA). Fat mass (FM) and fat-free mass (FFM) were extracted from DEXA scans and expressed in grams.

Height and body weight were measured using standardized procedures, with participants wearing no shoes. Height was recorded to the nearest cm using a stadiometer, and body weight was measured to the nearest kg using a calibrated scale.

Performance of predictive equations

OMR (as a surrogate for BMR) of each participant was compared with predicted BMR derived from WHO, Harris-Benedict (1918), Harris-Benedict (2023) and Mifflin-St Jeor equations to determine accuracy of the predictive equations.^11–14 In addition, OMR of each participant was compared with predicted BMR derived from equations incorporating body composition, including those described by Wang, Nelson, and Sabounchi (structures 4, 5, and 11).^20–22 The predictive equations for BMR used in this study are provided in the supplementary materials.

Accuracy of predictive equations was quantified as the median difference between predicted BMR and measured OMR (absolute bias, kcal/day) and as the relative bias (percentage error). To evaluate average performance across equations, predicted BMR, prediction bias, and relative prediction bias were averaged per participant separately across the four standard equations and across the five body composition-based equations.

Physical activity level

PAL was derived by dividing TEE (measured by DLW) by OMR. Since literature confirms that OMR closely approximates BMR, OMR was used as the denominator in PAL calculations.^27–31 PAL was computed individually for each participant. Mean PAL values were compared between groups and categorized according to Institute of Medicine (IOM) physical activity levels. ³²

Respiratory exchange ratio

RER was obtained from data collected by 24h-room calorimetry during the same 24-hour period used to measure OMR. Participants were provided with standardized meals (16% protein, 55% carbohydrate, 28% fat) and instructed to avoid exhaustive physical activity during the stay, to provide more reliable information on metabolic activity.

RER was calculated from the ratio of CO₂ production to O₂ consumption derived from the chamber’s gas-exchange measurements. A RER of approximately 0.7 indicates predominant fat oxidation, whereas a RER of 1.0 indicates predominant carbohydrate oxidation; values between 0.8 and 0.85 reflect mixed substrate utilization. ³³ Mean 24-hour RER was used to compare 24-hour substrate utilization between groups under standardized diet and low-activity conditions.

Statistical analysis

All data were analyzed using manually installed packages in RStudio. ³⁴ Continuous variables were assessed for normality using visual inspection of histograms and standardized normal probability plots. Normally distributed variables are presented as mean ± standard deviation, and skewed variables as median [interquartile range]. Categorical variables are reported as counts and percentages. There were no missing data for the variables included in the analyses. Baseline characteristics were compared between patients with DM1 and healthy controls using chi-square or Fisher’s exact tests for categorical variables, analysis of variance for normally distributed continuous variables, and Mann-Whitney U tests for skewed variables. The distribution of the continuous variables was visually displayed by raincloud plots.

OMR was compared between DM1 patients and controls using Mann-Whitney U tests. For each predictive equation, differences between predicted BMR and measured OMR were calculated and tested against zero with Wilcoxon signed-rank tests. Exploratory associations between the bias across the standard equations and body composition, physical activity, and total energy expenditure were assessed within the DM1 group using Spearman’s rank correlation coefficients. Examined variables included fat mass percentage, fat-free mass percentage, the fat mass-to-fat-free mass ratio, PAL, and TEE. Group differences in PAL and RER were evaluated with Mann-Whitney U tests. Because similar studies have not been conducted in this specific or comparable patient population, a formal sample size calculation could not be performed. Literature suggests that pilot studies should include at least 12 participants per group ³⁵ ; allowing for a potential drop-out rate of 10–20%, 15 participants per group were included. P values of <0.05 were considered statistically significant.

Results

Study population

15 DM1 patients and 15 age-, sex-, and BMI-matched controls were included in the study. Baseline characteristics of both groups are detailed in Table 1.

Table 1.

Baseline characteristics.

​	Myotonic dystrophy type 1 patients (n=15)	Healthy controls (n=15)	p-value
Age, median [IQR]	41 [30-50]	42 [30-50]	0.870
Male, n (%)	7 (47%)	7 (47%)	0.806
BMI, median [IQR]	26.5 [21.8-28.6]	26.2 [22.9-29.4]	1.000
BMI category, n (%)
18.5 - 24.9 kg/m 2	6 (40%)	6 (40%)	1.000
25 – 29.9 kg/m 2	7 (47%)	7 (47%)	1.000
30 – 35 kg/m 2	2 (13%)	2 (13%)	1.000
CTG repeat length, n (%)
100-150	3 (20%)	​	​
>150	6 (40%)	​	​
>200	4 (27%)	​	​
>350	2 (13%)	​	​
MIRS, n (%)
1	0 (0%)	​	​
2	2 (13%)	​	​
3	5 (33%)	​	​
4	8 (54%)	​	​
5	0 (0%)	​	​
Fat mass
Median kg [IQR]	29.4 [24.7-34.3]	24.4 [19.8-31.6]	0.250
Median % of bodyweight [IQR]	37.8 [32.3-40.3]	31.4 [27.8-35.7]	0.061
Fat-free mass
Median kg [IQR]	48.9 [47.4-56.6]	51.2 [48.1-60.0]	0.389
Median % of bodyweight [IQR]	62.2 [59.8-67.7]	68.6 [64.3-72.2]	0.061
Fat-/fat-free mass ratio, median [IQR]	0.61 [0.48-0.67]	0.46 [0.38-0.56]	0.061
Total energy expenditure, median, kcal/day [IQR]	2162 [1794-2494]	2814 [2424-3310]	0.027

IQR, inter quartile range; BMI, body mass index; MIRS, muscular impairment rating scale.

There were no statistical differences in fat- or fat-free mass between DM1 patients and matched controls.

Overnight metabolic rates

OMR did not differ significantly between groups (Mann-Whitney U test, p=0.149), with median values of 1517 and 1590 kcal/day in DM1 patients and controls, respectively (Supplemental Figure 1).

Accuracy of standard predictive equations

The performance of the WHO, Harris-Benedict (1918), Harris-Benedict (2023) and Mifflin-St Jeor equations for estimating BMR were evaluated in patients with DM1 and healthy controls (Tables 2a and 2b). Across all four equations, significant overestimation of metabolic rate was observed in the DM1 group with a median relative bias ranging from 6.6 to 11.5%. In healthy controls, predicted metabolic rates did not significantly differ from measured OMR. The absolute and relative bias of predicted BMR versus measured OMR are visualized in Figure 1.

Table 2a.

Measured OMR and predicted BMR using standard equations.

Measured OMR kcal/day, Median [IQR]		Predicted BMR kcal/day, Median [IQR]
		WHO	Harris-benedict (1918)	Harris-benedict (2023)	Mifflin-St jeor
DM1 patients	1517 [1413-1636]	1605 [1532-1846]	1627 [1527-1827]	1664 [1559–1877]	1617 [1476-1752]
Controls	1590 [1439-1849]	1663 [1549-1806]	1566 [1495-1726]	1605 [1571–1815]	1550 [1472-1690]

OMR, overnight metabolic rate; BMR, basal metabolic rate; IQR, inter quartile range; DM1, myotonic dystrophy type 1.

Table 2b.

Difference between predicted BMR and measured OMR using standard equations.

Equation	Group	Absolute bias in kcal/day, Median [IQR]	p-value	Relative bias in %, Median [IQR]	p-value
WHO	DM1	+165 [42 – 242]	0.001	+11.5 [2.4 – 17.0]	0.001
​	Controls	-8 [-36 – 81]	0.589	-0.4 [-2.0 – 5.0]	0.478
Harris-Benedict (1918)	DM1	+145 [79 – 190]	0.001	+10.4 [5.0 – 15.5]	<0.001
	Controls	-5 [-74 – 55]	0.755	-0.4 [-4.2 – 3.8]	0.842
Harris-Benedict (2023)	DM1	+155 [114–278]	<0.001	+9.7 [8.2–20.2]	<0.001
	Controls	+24 [-19–120]	0.244	+1.5 [-1.1–7.6]	0.244
Mifflin-St Jeor	DM1	+100 [33 – 161]	0.008	+6.6 [2.0 – 11.5]	0.005
	Controls	-78 [-145 – 0.2]	0.050	-4.4 [-7.9 – 0.0]	0.050

OMR, overnight metabolic rate; BMR, basal metabolic rate; IQR, inter quartile range; DM1, myotonic dystrophy type 1.

Table 2b: Values represent the median of participant-level differences (predicted BMR − measured OMR) and therefore do not necessarily equal the difference between the group medians shown in Table 2a.

Figure 1.

Absolute and relative bias of predicted BMR versus measured OMR in DM1 patients and matched controls. (a), (b) show results for the WHO equation, (c), (d) for the original Harris-Benedict equation (1918), (e), (f) for the revised Harris-Benedict equation (2023) and (g), (h) for the Mifflin-St Jeor equation.

For each participant, prediction bias was averaged across the four standard equations, which was +132.1 kcal/day in DM1 and -15.6 kcal/day in controls. The corresponding relative biases were +9.4% and -1.1%, respectively.

Exploratory correlation analyses were performed within the DM1 group. The mean bias across the standard equations was not significantly correlated with fat mass percentage (ρ=0.114, p=0.685), fat-free mass percentage (ρ=−0.114, p=0.685), the fat mass-to-fat-free mass ratio (ρ=0.114, p=0.685), or PAL (ρ=−0.257, p=0.355). A moderate inverse correlation was observed with TEE, but this did not reach statistical significance (ρ=−0.468, p=0.079).

Accuracy of predictive equations incorporating body composition

The performance of predictive equations for BMR incorporating body composition (Wang, Nelson, Structure 4, Structure 5 and Structure 11) in DM1 patients and healthy controls is shown in Tables 3a and 3b. The Wang and Nelson equations showed significant underestimation of metabolic rate in both DM1 patients and healthy controls. Structure 5 also tended to underestimate metabolic rate, reaching statistical significance in healthy controls and showing borderline significance in DM1 patients. Structure 4 performed relatively well in the DM1 group, showing no statistically significant absolute bias, although relative bias was borderline significant. In healthy controls, Structure 4 also showed no significant difference from measured OMR, suggesting better performance than Wang, Nelson, and Structure 5 in this group. Structure 11 performed best in the DM1 group, showing no significant difference from measured OMR for either absolute or relative bias. In healthy controls, however, Structure 11 significantly underestimated metabolic rate.

Table 3a.

Measured OMR and predicted BMR using body composition-based equations.

Measured OMR kcal/day, Median [IQR]		Predicted BMR kcal/day, Median [IQR]
		Wang	Nelson	Structure 4	Structure 5	Structure 11
DM1 patients	1517 [1413-1636]	1378 [1342-1567]	1412 [1335-1550]	1560 [1484-1661]	1439 [1399-1542]	1441 [1360-1700]
Controls	1590 [1439-1849]	1435 [1357-1650]	1419 [1343-1639]	1560 [1489-1738]	1484 [1417-1601]	1434 [1375-1745]

OMR, overnight metabolic rate; BMR, basal metabolic rate; IQR, inter quartile range; DM1, myotonic dystrophy type 1.

Table 3b.

Difference between predicted BMR and measured OMR using body composition-based equations.

Equation	Group	Absolute bias in kcal/day, Median [IQR]	p-value	Relative bias in %, Median [IQR]	p-value
Wang	DM1	-80 [-185 – -46]	0.003	-5.6 [-10.6 – -3.0]	0.003
​	Controls	-178 [-215 – -97]	0.002	-10.7 [-13.1 – -5.9]	0.003
Nelson	DM1	-80 [-169 – -43]	0.006	-5.6 [-10.0 – -2.8]	0.003
	Controls	-154 [-250 – -82]	0.002	-9.8 [-15.1 – -5.9]	0.002
Structure 4	DM1	+62 [-27 – 109]	0.074	4.1 [-1.7 – 7.9]	0.050
	Controls	-69 [-117 – -30]	0.148	-3.1 [-6.8 – -2.2]	0.201
Structure 5	DM1	-58 [-115 – -12]	0.050	-3.4 [-7.1 – -0.8]	0.050
​	Controls	-203 [-235 – -45]	0.008	-10.5 [-14.0 – -3.2]	0.010
Structure 11	DM1	-0.8 [-65 – 55]	1.000	+0.0 [-4.8 – 3.9]	1.000
​	Controls	-135 [-177 – -67]	0.006	-8.0 [-10.2 – -4.0]	0.010

OMR, overnight metabolic rate; BMR, basal metabolic rate; IQR, inter quartile range; DM1, myotonic dystrophy type 1.

Table 3b: Values represent the median of participant-level differences (predicted BMR − measured OMR) and therefore do not necessarily equal the difference between the group medians shown in Table 3a.

For each participant, prediction bias was averaged across the five body composition-based equations, which was -28.6 kcal/day in DM1 and -154.3 kcal/day in controls. The corresponding relative biases were -1.9% and -7.4%, respectively.

Physical activity level

PAL was statistically significantly lower in DM1 patients compared with matched controls (1.42 [1.38–1.49] vs 1.69 [1.55–1.80], p<0.001) (Figure 2(a)). Based on the IOM physical activity categories, DM1 patients were classified as low active, whereas matched controls were classified as active.

Figure 2.

Comparison of median (a) physical activity level (PAL) and (b) respiratory exchange ratio (RER) between DM1 patients and healthy matched controls.

Respiratory exchange ratio

RER did not differ significantly between DM1-affected individuals and healthy controls measured under standardized dietary and low-activity chamber conditions. Median RER was 0.82 [0.81–0.83] in DM1 patients and 0.83 [0.81–0.84] in healthy matched controls (p=0.740) (Figure 2(b)).

Discussion

The present study evaluated commonly used predictive equations to estimate metabolic needs in individuals affected by DM1. Equations based solely on anthropometric variables (WHO, Harris-Benedict (1918), Harris-Benedict (2023) and Mifflin-St Jeor) consistently overestimated OMR in DM1. When performance was averaged across the four standard equations, the relative bias was approximately +9%, whereas incorporation of body composition improved predictive accuracy. As anticipated, PAL was markedly reduced in patients compared with healthy controls, likely contributing to lower TEE. RER did not differ between individuals with DM1 and healthy controls, indicating comparable mixed substrate oxidation under standardized dietary and low-activity conditions in both groups.

The current analyses were based on data obtained from a previously performed cohort study designed to investigate metabolic characteristics in DM1. ⁷ OMR derived from indirect calorimetry was used as a proxy for BMR. Although OMR and BMR differ in measurement conditions, previously performed research demonstrates that OMR has a close agreement with BMR, with differences typically ≤5%.^27–31 This assumption is further strengthened by the fact that predicted BMR using anthropometric variables did not differ from measured OMR in included healthy controls. In contrast, the systematic overestimation observed in individuals with DM1 by WHO, Harris-Benedict (1918), Harris-Benedict (2023) and Mifflin-St Jeor equations may partly reflect disease-related differences in body composition and metabolically active tissue.^2,17 As lean tissue, particularly fat-free mass, is the principal determinant of BMR, ³⁶ it is plausible that anthropometric equations may inadequately capture the reduced proportion of metabolically active tissue in muscle disease patients. However, the exploratory analyses within the DM1 group did not demonstrate significant associations between prediction bias and fat mass percentage, fat-free mass percentage, or the fat mass-to-fat-free mass ratio.

Although the bias could therefore not be attributed to one specific body-composition measure, the combined incorporation of FFM and FM may still improve prediction accuracy. To examine this, several body composition-based equations were evaluated for estimating metabolic rate in DM1. The Nelson and Structure 5 equations are primarily FFM-based, whereas Structure 4 and Structure 11 include both FFM and FM.^20,22 The Wang equation represents an organ-based model derived from FFM. ²¹ Overall, the equations based mainly on FFM tended to underestimate metabolic rate, whereas the equations incorporating both FFM and FM showed more favourable, although still variable, performance. This suggests that selected body composition-based models may better reflect disease-specific metabolic characteristics in DM1, but also that inclusion of body composition alone does not necessarily ensure accurate prediction.

Evidence on basal or resting metabolic rate in DM1 is limited, but generally indicates reduced energy expenditure independent of body composition. ¹⁸ Comparable observations have been reported in Duchenne muscular dystrophy (DMD), where basic metabolic rate is frequently reported to be lower than predicted by standard equations, but differences often diminish after correction for FFM. ³⁷ At the same time, evidence from healthy populations suggests that body composition-based equations, including those based on FFM alone or on both FFM and FM, do not consistently improve prediction accuracy over traditional equations and may still substantially underestimate metabolic rate at the individual level. ³⁸ Against this background, the relatively good performance of the Structure 11 equation in DM1 may indicate that selected body composition-based models better reflect disease-specific metabolic characteristics, while still underscoring the need for condition-specific calibration.

Despite comparable OMR between groups, PAL was substantially lower in DM1 than in controls, corresponding to the low active category according to the Institute of Medicine physical activity classification, whereas control values corresponded to the active category. This difference amounted to an approximately 16% lower PAL in the DM1 group and is in line with previous reports demonstrating a 20-25% reduction in total energy expenditure.^3,6,39,40 This finding is also consistent with the habitual physical activity assessment reported in our previous publication from the same cohort, in which a step count measure also indicated reduced daily physical activity in DM1 patients compared with controls. ⁷

Importantly, DM1 is characterized by progressive muscle wasting. This indicates that reductions in muscle mass and consequently energy expenditure are likely to continue over time, increasing susceptibility to a positive energy balance as the disease progresses. ⁴¹ These findings underscore the importance of carefully balancing energy intake in DM1: avoiding overfeeding in the context of low physical activity, while at the same time supporting the preservation of lean mass. The use of reliable and accurate metrics is of paramount importance in this respect.

The RER did not differ significantly between groups under the standardized dietary and low-activity chamber conditions, indicating no clear evidence for altered resting substrate utilization in DM1, with values in both groups consistent with mixed substrate oxidation. ³³ Therefore, these findings do not support a disease-specific shift in substrate utilization at rest. Rather, dietary management in DM1 may need to focus on overall dietary quality and alignment of energy intake with reduced activity levels, while also supporting maintenance of muscle mass.

Several limitations warrant consideration, including the relatively small sample size, use of OMR as a surrogate for BMR and lack of dietary intake data to contextualize potential energy imbalance in individuals with DM1. In addition, no additional functional muscle strength tests or anthropometric measures, such as handgrip strength or skinfold thickness, were available to further characterize muscle function and body composition. Furthermore, the exclusion of non-ambulatory adult-onset patients able to perform exercise tests resulted in a relatively homogeneous cohort, which may limit generalizability to the broader DM1 population, particularly patients with more severe disease manifestations or comorbidities. Future studies should include substantially larger and more heterogeneous DM1 populations, including non-ambulatory patients and patients across the full spectrum of disease severity, to develop and independently validate a DM1-specific prediction equation. Such studies should also examine whether prediction accuracy differs according to CTG repeat length and clinical disease severity.

In conclusion, commonly used anthropometric predictive equations overestimate BMR in DM1, whereas selected body composition-based equations, including the Nelson, Wang, Structure 4, Structure 5 and Structure 11, provided more accurate estimates.

When standard equations are used, applying an approximate 9% downward correction may improve estimation of energy requirements. TEE calculations should account for low habitual physical activity, with PAL values in the range of 1.3-1.5 being most appropriate. These findings have direct implications for dietary management in DM1, as unadjusted estimates may lead to overfeeding and increased cardiometabolic risk, underscoring the importance of individualized and longitudinal nutritional assessment.

ORCID iDs

Wander J. A. Aarts https://orcid.org/0009-0000-4348-3573

Basil P. A. P. van Veen https://orcid.org/0009-0000-6350-671X

Catharina G. Faber https://orcid.org/0000-0002-2467-067X

Isis B. T. Joosten https://orcid.org/0000-0002-1110-6172

Ethical considerations

The study was approved by the institutional ethics committee (METC 182009), on 27-06-2018.

Consent to participate

All participants provided written informed consent.

Data Availability Statement

The data supporting the findings of this study are available from the corresponding author upon reasonable request.*