A. General ResultsDemographics

A total of 60 adult and pediatric subjects with genetically-confirmed PMD were studied. 38/60 (63.3%) were children ≤ 18 years of age, which included 2 infants (< 12 months), 15 children (1-9 years), and 21 adolescents (10–18 years) (Table 1).

Genetic etiologies were identified in mtDNA in 35 subjects (58.3%), and in nuclear DNA in 25 subjects (41.7%, Table e-1). The most frequent genetic etiology was single large-scale mtDNA deletions (SLSMD, 10/60, 16.7%). Eight subjects (13.3%) had a gastrostomy (G)-tube in place for supplemental feeds, including 2 adults and 6 children (Table 1). The two adults who presented with a G-tube, respectively, had SLSMD (n = 1) and a nuclear gene variant in TWNK (c.1110C > G: p.F370L) (n = 1). In the 6 pediatric subjects with G-tubes, 4/6 (66.7%) had mtDNA-disease etiologies including SLSMD (n = 2), MT-ND5 (n = 1), and MT-ND3 (n = 1), while 2/6 carried nuclear genetic etiologies (SURF1 (n = 1) and POLG (n = 1)).

Body Mass Index (BMI) & Weight Classification

Mean weight and height for adults (n = 22) was 58.7 ± 13.0 kg and 1.63 ± 0.1 m, respectively, and 29.0 ± 17.8 kg (mean z-score -1.7 ± 2.9; range -13.6 to 2.1) and 1.26 ± 0.29 m (mean z-score -1.1 ± 1.8; range -7.3 to 2.5) in the child subjects (n = 38) (Table 1). Mean BMI was 23.4 ± 1.4 in adults, and 16.7 ± 0.7 (mean z-score -1.04 ± 0.3; range -6.2 to 2.0) in child subjects. Weight classification was determined using BMI classification in adults and BMI percentile (%) in children [55, 56]. Of the 22 adults, 3/22 (13.6%) adults were underweight (BMI ≤ 18.5 kg/m2), 13/22 (59.1%) had appropriate weight (BMI 18.5 – 25 kg/m2), and 6/22 (27.3%) were overweight/obese (BMI ≥ 25 kg/m2). Of the 38 children, 12/38 (31.6%) were underweight (BMI percentile < 5th), 20/38 (52.6%) were appropriate (BMI percentile 5-84th), and 6/38 (15.8%) were overweight/obese (BMI percentile > 85th) (Table e-2, Fig. 2A). Of the six pediatric subjects with a G-tube, 4/6 (66.7%) were underweight, while the remaining were appropriate weight or overweight, 1/6 (16.7%) each.

Fig. 2

A BMI classification. BMI of subjects based on ASPEN criteria for adult and child subjects is displayed. Most subjects had appropriate BMI. 3/22 (13.6%) adults and 12/38 (31.6%) children were underweight. B Patient-reported gastrointestinal symptoms (GI). 47/60 (78.3%) PMD subjects reported at least one GI symptom. The most commonly reported GI symptoms were dysphagia (28/60, 46.7%), constipation (26/60,43.3%), GI dysmotility (15/60, 25%), and reflux (14/60, 23.3%). C Predicted and dietician (RDN)-estimated daily Kcal intake in adult and child subjects. The predicted Kcal goal (WHO Resting Energy Expenditure value x MOTIVATOR Activity Factor) is compared to the RDN-estimated Kcal intake in all subjects (n = 60) and stratified by age. Results show the significantly lower RDN-estimated cohort mean daily Kcal intake of 1,125 ± 54.4 kcal/day to the predicted Kcal intake values (86.3% predicted, n = 60, p < 0.0001****). In the adult subjects, the estimated Kcal intake was 1,143 ± 104.1 kcal/day (76.2% predicted, n = 22, p = 0.003**), while in the child subjects, the estimated Kcal intake was 1,114 ± 62.3 kcal/day (86.4% predicted, n = 38, p = 0.001***). D Macronutrient consumption/WHO-MOTIVATOR predicted Kcal goal in adult PMD subjects (%Kcal/day) (n = 22). In adult subjects, mean macronutrient consumption percentage of WHO-MOTIVATOR predicted Kcal intake for CHO was 42.1 ± 5.9% Kcal (76.5% predicted, 627.6 ± 81 kcal/day), 14.5 ± 1.4% Kcal (96.8% predicted, 219.2 ± 18.5 kcal/day) for PRO, and 20.5 ± 2.3% Kcal (68.2% predicted, 306.1 ± 29.6 kcal/day) for fat. E Macronutrient consumption/WHO-MOTIVATOR predicted Kcal goal in child PMD subjects (%Kcal/day) (n = 38). In child subjects, mean macronutrient consumption percentage of WHO-MOTIVATOR predicted Kcal intake for CHO was 44.1 ± 3.8% Kcal (89.7% predicted, 571.1 ± 40.7 kcal/day), 14.3 ± 1.2% Kcal (71.7% predicted, 190.0 ± 12.8 kcal/day) for PRO, and 29.0 ± 2.3% Kcal (94.2% predicted, 365.1 ± 23.4 kcal/day) for fat

Prevalence of Gastrointestinal (GI) Symptoms

Patient and/or physician-reported GI symptoms were extracted from the electronic medical record (EMR) primary care physician and/or GI physician clinical evaluation letter. At least one GI symptom was reported in 28/38 (73.7%) of child subjects and 19/22 (86.4%) adults. Across the PMD cohort (n = 60), the most frequent GI symptoms were dysphagia (Human Phenotype Ontology, HPO ID 0002015) (28/60, 46.7%) and constipation (HPO ID 0002019) (26/60, 43.3%). Of those subjects with constipation, 10/26 required treatment for constipation, 10/26 were managed with diet, and 6/26 did not require treatment. In adults subjects, dysphagia was the most frequent GI symptom (13/22, 59.1%), while both dysphagia and constipation were most frequent in children (15/38, 39.5%). GI dysmotility (HPO ID 0002579) (15/60, 25%), vomiting (HPO ID 0002013) (14/60, 23.3%), gastroesophageal reflux (HPO ID 0002020) (14/60, 23.3%), failure to thrive (HPO ID 0001508) (12/60, 20%), and diarrhea (HPO ID 0002014) 10/60 (16.7%) were also reported. Nausea (HPO ID 0002018), abdominal pain (HPO ID 0002027), loss of appetite (HPO ID 0004396), poor weight gain (HPO ID0001508), and feeding difficulties (HPO 0011968) were each reported in < 15% of the cohort (Table e-3, Fig. 2B).

Among those who consumed 75.01–100% of their predicted daily Kcal needs (n = 14/60, 23.3% of the total cohort), 4/14 (28.6%) reported no GI symptoms, 1/14 (7.14%) reported 1 symptom, and 9/14 (64.3%) subjects reported having between 2 to 7 GI symptoms. The most common symptoms were dysphagia in 6/14 (42.9%) and constipation in 5/14 (35.7%). Of those who consumed ≤ 75% predicted Kcal/day (n = 29/60, 48.3% of the total cohort), which is consistent with malnutrition consumption levels [37, 40], 5/29 (17.2%) reported 1 or less GI symptoms, and 24/29 (82.8%) reported up to 7 GI symptoms. Constipation in 15/29 (51.7%) and dysphagia in 12/29 (41.4%) were the most common symptoms. Fisher’s exact test revealed no significant difference in the frequency of GI symptoms between the ≤ 75% and 75.01–100% Kcal intake groups, p = 0.08.

Of the underweight subjects based on BMI (n = 15/60, 25%), 12/15 subjects (80%) reported up to 6 GI symptoms, while 3/15 (20%) subjects reported no symptoms. The most common symptom was dysphagia in 10/12 (83.3%) subjects. There were 5/15 (33.3%) underweight subjects who consumed ≤ 75% predicted Kcal/day. Among this subset, 2/5 (40.0%) reported 2–6 GI symptoms.

Laboratory Markers

Biochemical measurements performed in adult and child PMD obtained ≤ 6 months before or immediately after the baseline nutrition evaluation included plasma CK, creatinine, albumin, cholesterol, triglycerides, pyruvate, and lactate (Table 1, mean ± standard error of mean (SEM). Of note, mean blood lactate was elevated at 2.8 ± 0.38 (normal range 0.8-2 mM) in 45/60 subjects. Blood lactate levels were significantly different in subjects with malnutrition, (5.0 ± 1.0, range 0.97–12.03 mM, n = 13/45), compared to those subjects without malnutrition, (1.9 ± 0.16, range 0.85–5.13 mM, n = 32/45), p = 0.0002. Plasma amino acid analysis performed in 25/60 subjects revealed that plasma alanine trended higher in subjects with malnutrition, 589.8 ± 93.2 μmol/g, n = 4, compared to those without malnutrition, 421.8 ± 27.2 μmol/g, n = 21, p = 0.1. Increased plasma alanine levels are indicative of chronic lactic acidosis or gluconeogenesis [54]. There was no significant difference in plasma CK, albumin, cholesterol, creatinine, or triglycerides levels in subjects with malnutrition (n = 16/60) and without malnutrition (n = 44/60). Plasma GDF15 levels were elevated at 1,567 ± 258.7 (mean ± SEM, range 183–6,000, normal ≤ 750 pg/ml) across the PMD cohort (n = 59), specifically 1,882 ± 425.6 in adult (n = 17/59) and 1,283 ± 312.2 in child PMD subjects (n = 32/59). GDF15 levels did not correlate to weight (r = 0.26, p = 0.07) or height (r = 0.25, p = 0.09) measurements. Linear regression analysis of GDF15 levels to BMI was not significant, R2 = 0.01, p = 0.61. There was no significant difference in GDF15 levels in subjects with malnutrition (2,338 ± 822.2 pg/ml), n = 10, and without malnutrition (1,274 ± 232.5 pg/ml), n = 44, p = 0.61.

B. Nutritional Evaluation ResultsEstimation of Total Energy Expenditure in PMD Subjects Using the WHO Predictive Equation and ASPEN/RDA Activity Factors (AF)

Eight three-day diet records and 52 diet interviews were obtained on 60 patients followed at the Children’s Hospital of Philadelphia Mitochondrial Medicine Frontier Program. Resting energy expenditure (REE) was estimated using the World Health Organization (WHO) REE equation [32] multiplied by an ASPEN/RDA activity factor (AF) based on physical activity levels [32]. Predicted total energy expenditure (TEE) is the product of REE x AF [55, 56].

As our center evaluates the full age spectrum of PMD patients, our objective was to apply the same predictive REE equation across the ages in this study. The WHO equation is routinely utilized in children, but in adults (≥ 19 years), the Harris-Benedict (HB) [36] and Mifflin St. Jeor (MSJ) [57] equations are commonly used. We compared WHO-calculated REE to Harris-Benedict and Mifflin St. Jeor-REE estimates in the adult subjects to show that WHO-REE can be utilized in our adult PMD subjects. In conjunction with the ASPEN/RDA AFs, the predicted TEE was not significantly different between the three prediction equations, at 1,936 ± 59.0 kcal by WHO-REE x AF, 1,938 ± 50.1 kcal by HB-REE x AF, and 1,827 ± 62.5 kcal by MSJ-REE x AF, p = 0.30 (ANOVA, Table e-4). These data confirm that the WHO-REE equation can be appropriately administered to predict TEE across the PMD age spectrum.

We initially applied the ASPEN/RDA AFs in the prediction of daily Kcal intake to our cohort (Table e-5). RDN estimated dietary intake revealed a mean Kcal intake (mean ± SEM) of 1,143 ± 104.1 kcal/day in adults (range 226 – 2,340 kcal/day, n = 22), which is 60.1 percent (%) predicted (RDN estimated Kcal intake/WHO-REE x ASPEN/RDA AF predicted Kcal intake) (Table e-5, Fig. 2C). Mean RDN estimated Kcal intake in child subjects was 1,114 ± 62.3 kcal/day (range 410—2,314 kcal/day, n = 38) or 76.5% of that predicted by the WHO-REE x ASPEN/RDA AFs (Table e-5). Our data reveals that adult and child PMD subjects consumed significantly decreased daily Kcal intake (p < 0.0001) at baseline nutritional assessment prior to dietary intervention.

Development of PMD-Specific Activity Factors (MOTIVATOR) in the Estimation of Total Energy Expenditure (TEE)

Currently, guidance on AF selection in PMD does not exist. ASPEN/RDA AFs, which reflect the physical capacity of healthy individuals of average body composition [32, 55, 56], uses muscle strength as the main indicator of physical activity. Similarly, guidance on estimation of AFs in cerebral palsy (CP) [65] is based on muscle strength, as muscle weakness and/or spasticity is the main phenotype in CP. By contrast, there is a high prevalence of muscle fatigue and exercise intolerance with or without muscle weakness in PMD [2] that significantly impacts activity levels and thus energy expenditure. Individuals with PMD likely have reduced muscle mass [58], which leads to decreased energy needs. In addition, prevalence of low muscle tone and involuntary movements in PMD [59] should also be considered for estimation of energy expenditure.

To ensure a standardized approach to assigning AFs in predicting PMD energy needs, we established and propose the broad implementation of customized PMD-specific AFs, Mitochondrial Activity Factors (MOTIVATOR, Tables e-6–8). With the benefit of extensive clinical experience in characterizing PMD phenotype using a validated objective assessment of Mitochondrial Myopathy (MM), the MM-Composite Assessment Tool (MM-COAST) [28], and review of existing literature on AFs and gross motor function classifications [32, 34,35,36, 55], we defined clinically relevant ranges of AFs for child (Table e-6–7) and adult (Table e-6, e-8) PMD subjects that correspond to their motor function and fatigue levels, in order to facilitate a more personalized and rigorous estimation of TEE. In this study, the lower limit of the MOTIVATOR-AF range (Tables e-6–8) was consistently applied unless the upper limit was deemed more suitable based on clinical assessment. Routine RDN practice to consider growth patterns, muscle mass, acute illness, medications, involuntary movements [59], and mechanical ventilation (Table e-6) to estimate the AF should proceed as usual and be added to the selected MOTIVATOR-AF (Tables e-7–8). Determination of MOTIVATOR-AFs should be based on clinician assessments of mobility and fatigue levels and not by PMD subject self-report, which may be less accurate.

Estimation of TEE Using PMD-Specific Activity Factors (MOTIVATOR)

We anticipated that ASPEN/RDA AFs [31] would overestimate TEE as it overlooks the impact of exercise intolerance and fatigue on PMD mobility. In adult subjects, predicted TEE (WHO-REE x ASPEN/RDA AFs) was 1,936 ± 59.0 kcal/day, compared to 1,554 ± 58.0 kcal/day with WHO-REE x MOTIVATOR AFs, p < 0.0001. Thus, RDN estimated dietary intake in PMD subjects as percent (%) predicted of the TEE was 60.1% using ASPEN/RDA AFs compared to 76.2% with MOTIVATOR AFs, p < 0.0001 (t-test, Table e-5, Fig. 2C). In child subjects, predicted TEE was 1,663 ± 94.5 kcal/day (76.5% predicted dietary intake) with ASPEN/RDA AFs compared to 1,444 ± 84.9 kcal/day (86.4% predicted) with MOTIVATOR AFs, p < 0.0001 (t-test, Table e-5, Fig. 2C).

These data confirm that the MOTIVATOR predicts lower energy needs in PMD, which we consider to be more precise since MOTIVATOR-AFs were designed for the PMD phenotype, compared to ASPEN/RDA AFs that do not discriminate the various mobility levels of a PMD subject. Indeed, ASPEN/RDA AFs overestimated Kcal goals by 382 kcal (19.7%) in adults and 189 kcal (11.6%) in child subjects (Table e-5). Subsequent results presented in this manuscript are based on WHO-MOTIVATOR predicted total Kcal requirements.

Daily Macronutrient Consumption a. Altered Macronutrient Consumption in PMD

Having confirmed inadequate daily Kcal intake across the PMD cohort, we sought to characterize their daily macronutrient distribution. Macronutrient consumption expressed as a fraction (percentage, %) of daily Kcal intake revealed mean carbohydrate (CHO) consumption to be 52.9% (96.2% predicted when compared to the Recommended Daily Allowances (RDA) [32], 20.2% protein (PRO) (134.7% predicted), and 27.4% fat (91.5% predicted) in adult PMD subjects (Table e-9). In child subjects, mean CHO calorie consumption was 50.1% (101.3% predicted), 17.5% PRO (87.5% predicted), and 33.5% fat (109.7% predicted) (Table e-9). Of note, 17.5% PRO calorie consumption in the child subjects remains within the acceptable macronutrient distribution range (AMDR) range of 15–30% [31]. These results reveal an apparently normal distribution of macronutrient consumption. However, total Kcal intake is essentially a measure of total macronutrient intake. Therefore, in the context of significantly low daily Kcal intake in our PMD cohort in whom overall nutritional needs are not being met, the macronutrient distribution would be expected to be deficient.

The adult RDA macronutrient distribution is based on a healthy adult ~ 2,000 kcal diet [32], of CHO 1,110 kcal/day (55%), PRO 300 kcal/day (15%), and fat 600 kcal/day (30%). Our adult PMD cohort mean consumption of each individual macronutrient was distinctly below these stated levels at 627.6 ± 81.0 kcal CHO, 219.2 ± 18.5 kcal PRO, and 306.1 ± 29.6 kcal fat (Table e-9), with an overall daily Kcal intake at 76.2% predicted. Thus, it is apparent that analysis of macronutrient distribution as a fraction of daily Kcal consumption is less informative when the estimated daily Kcal consumption is decreased, as was observed in our PMD cohort. We propose that when the estimated daily Kcal consumption is low, macronutrient distribution analysis would be more meaningful if interpreted in the context of the predicted amount of Kcal that should be consumed.

Indeed, when the macronutrient distribution was expressed as a fraction of predicted Kcal intake goals (WHO-REE x MOTIVATOR AFs), at a goal of 1,554 ± 58 kcal for adults (n = 22) and 1,444 ± 84.9 in children (n = 38) subjects, macronutrient distribution in adult subjects was 42.1% CHO calorie consumption (76.5% predicted in comparison to the RDA goal of 55% [32], [AMDR of 45–65%], 14.5% PRO (RDA goal 15%, [AMDR of 10–35%], 96.8% predicted), and 20.5% fat (RDA goal 30%, [AMDR 20–35%], 68.2% predicted). In child subjects, the macronutrient distribution was 44.1% CHO calorie consumption (RDA goal 50%, [AMDR 45–65%], 89.7% predicted), 14.3% PRO (20%, [15–30%], 71.7% predicted), and 29.0% fat (30%, fat [25–35%], 94.2% predicted) (Table e-9, Fig. 2D–E). These results indicate altered distribution of macronutrient consumption, in alignment with the inadequate daily Kcal intake in PMD subjects (Fig. 2D–E). Thus, macronutrient consumption should be routinely evaluated in PMD, in the context of estimated as well as predicted calorie goals, particularly if the estimated Kcal intake is inadequate. Subsequent analyses reported here express macronutrient consumption as a fraction of WHO-MOTIVATOR predicted Kcal intake goals.

b. Significantly Decreased Fat Intake in PMD Subjects

We proceeded to further characterize PMD macronutrient distribution by defining absolute intake in grams/day (g/day). We compared estimated CHO and PRO intake to DRI goals [32], and fat intake to population data published in the DRI [60], and reviewed in the context of recent National Health and Nutrition Examination Survey (NHANES) references (2017–2018) [61]. PMD adults were found to consume a mean of 156.9 ± 20.2 g (g)/day of CHO (DRI goal 130 g, 120.7% predicted), 54.8 ± 4.6 g/day of PRO (DRI goal 56 g/day for males, 46 g/day for females, 111.9% predicted), and 34.01 ± 3.3 g/day of fat (population mean [60] 75.4 ± 3.3 g/day, 47.2% predicted, Table e-10). This confirms that adult PMD fat intake was significantly decreased, p < 0.0001. In the child cohort, mean CHO intake was 142.8 ± 10.2 g/day (DRI goal 130 g/day, 111.2% predicted), 47.5 ± 3.4 g of PRO (age-dependent DRI, (Table e-10), overall 186.1% predicted), and 40.6 ± 2.6 g of fat (72.2 ± 2.8 g/day [60], 58.2% predicted). Thus, in the combined adult and child PMD cohort, fat intake was significantly decreased at 30.2 ± 2.1 g/day (54.2% predicted).

In the adult (n = 13/22) and child (n = 15/38) PMD subjects with decreased fat calorie consumption (n = 28, defined as < 30% proportion of daily Kcal intake), GI symptoms were common. Dysphagia was reported in 13/28 (46.4%) subjects, constipation in 16/28 (57.1%), reflux in 10/28 (35.7%), and GI dysmotility in 8/28 (28.6%). A total of 11/28 (39.3%) subjects with decreased fat calorie consumption consumed > 60% of their daily calories as CHO, and 12/28 (42.9%) consumed > 20% of their total calories from PRO. Of the 13 subjects who reported dysphagia, decreased fat calorie consumption with corresponding high CHO calorie consumption was observed in 7/13 (53.8%), suggesting that GI symptoms are likely associated with altered macronutrient consumption in PMD.

Zweers et al. reported low PRO intake (1.1 ± 0.34 g/kg) in their adult PMD cohort compared to their corresponding population mean [25, 26]. In our study cohort, PRO intake was 2.0 ± 0.2 g/kg (217% predicted) and 0.60 ± 0.1 g/kg (118.5% predicted) in child and adult PMD subjects, respectively. Of note, % predicted PRO intake (g/day) decreases with age as PRO goals increase with age [32]. Indeed, males 14–18 years old in this cohort (n = 4) consumed only 55.9% predicted, while PRO intake in the younger age groups either met or exceeded the DRI PRO goals (Table e-10). The observation of high PRO intake in young children in the healthy general population is common, due to high consumption of dairy products [62, 63].

c. Altered Macronutrient Distribution Among Subjects Who Consumed ≤ 75% Predicted Calorie Intake

To assess whether the altered macronutrient intake was concordant with energy intake, we classified PMD subjects as having either ‘excess’ % predicted Kcal intake (≥ 100.01% Kcal, n = 17), ‘sufficient’ (90.01- 100% Kcal, n = 2), ‘low’ (75.01 – 90%, n = 12), or ‘insufficient’ intake (≤ 75% Kcal, n = 29) (Table e-11, Fig. 3A). Based on WHO-MOTIVATOR predicted energy intake, ‘insufficient’ energy intake (consumption of ≤ 75% predicted) was observed in 50% of adult PMD subjects (11/22). Of those, 4/11 (36.3%) consumed 50–70% of their predicted energy intake, while 7/11 (63.6%) consumed < 50% predicted. In child PMD subjects, ‘insufficient’ energy intake (≤ 75% predicted) was observed in 18/38 subjects (47.4%); 10/18 (55.6%) consumed 50–70% of their predicted energy intake; and 8/18 (44.4%) consumed < 50% predicted. These results emphasize the remarkably high prevalence of inadequate Kcal intake in our PMD cohort.

Fig. 3

A Macronutrient consumption (Kcal/day) based on calorie intake classification. Subjects were classified by their daily Kcal intake, as either ‘Excess’ intake (≥ 100.1% predicted, n = 17), ‘Sufficient’ intake (90.01- 100% predicted, n = 2), ‘Low’ intake (75.01–90% predicted, n = 12), and ‘Insufficient’ intake (≤ 75% predicted, n = 29). Results show that the macronutrient consumption goal is met in the ‘Excess’ and ‘Sufficient’ Kcal groups, but is significantly reduced in the ‘Insufficient’ Kcal group, at 441.9 ± 38 kcal (51. 5% mean predicted values) for CHO, 184.5 ± 15.7 kcal for PRO (61.1% predicted), and 266.2 ± 20.2 kcal for fat (54.5% predicted) when compared to the other Kcal groups, p < 0.0001. B Macronutrient intake (g/day). Macronutrient consumption by subjects in the ‘Insufficient’ Kcal group (≤ 75.01% predicted intake) trends lower compared to the other Kcal groups, with significantly lower fat intake at 41.3% predicted, p < 0.0001 (ANOVA). C Percent (%) predicted fat calorie consumption. Fat calorie consumption (% predicted Kcal/day) was lowest in the ‘Insufficient’ Kcal intake group compared to the other groups, ****p < 0.0001(ANOVA, F(3,56) = 21.65). D Percent (%) predicted fat intake (g/day). Absolute fat intake (g/day) was significant decreased in the ‘Insufficient’ Kcal group (n = 30), ****p < 0.0001(ANOVA, F(3,56) = 9.298)

On comparing the mean % predicted value for each macronutrient consumption [Estimated macronutrient consumption (Kcal/day)/WHO-MOTIVATOR predicted Kcal intake (Kcal/day)/RDA predicted macronutrient distribution goal (%)], we observed a significant and gradual decline from the ‘Excess’ through to the ‘Insufficient’ Kcal groups for CHO (R2 = 0.56, p < 0.0001, n = 60), PRO (R2 = 0.32, p < 0.0001, n = 60), and fat (R2 = 0.58, p =  < 0.0001, n = 60). Macronutrient consumption goal was met in the ‘Excess’ and ‘Sufficient’ Kcal groups but was significantly inadequate in the ‘Insufficient’ Kcal group, at 441.9 ± 38.0 kcal for CHO (51.5% mean predicted values), 184.5 ± 15.7 kcal for PRO (61.1% predicted), and 266.2 ± 20.2 kcal for fat (54.5% predicted) when compared to the other Kcal groups, p < 0.0001 (ANOVA, Table e-11, Fig. 3A and C). Thus, the decreased macronutrient consumption likely accounts for the overall low energy intake in the ‘Insufficient’ Kcal group.

A similar observation was made when macronutrients were expressed as g/day and compared across the four Kcal intake groups, for CHO (R2 = 0.24, p < 0.0001, n = 60), PRO (R2 = 0.24, p = 0.0001, n = 60), and fat (R2 = 0.41, p < 0.0001, n = 60). Most notably, fat intake was lowest in the ‘Insufficient’ Kcal group when compared to the other groups, at 41.3 ± 3.3% predicted (29.6 ± 2.2 g/day, Table e-12, Fig. 3B and D, p < 0.0001 (ANOVA).

Daily Calorie Consumption in BMI Groups is Not Predictable

Healthy adults with either very low (< 801 kcal/day) or low energy intake (≤ 1,000 kcal/day), tend to suffer medical complications, including a wide host of concerns, such as poor bone health, decubitus ulcers, intensifying gastroparesis, and vitamin and mineral deficiencies [55].

Mean daily Kcal intake in the underweight (n = 3), appropriate (n = 13), and overweight (n = 6) adult subjects (n = 22) was not significantly different, p = 0.99, ANOVA (Fig. 4A). Patient-reported GI symptoms were prevalent in all 3 BMI groups. Of the PMD adult subjects with appropriate BMI (n = 13/22, 59.1%), 2 subjects (15.4%) consumed < 800 (‘very low’) Kcal/day, 3 (23.1%) consumed 801–1,000 (‘low') Kcal/day, 4 (30.8%) consumed 1,001- 1,200 (‘moderate’) Kcal/day, and 4 (30.8%) consumed > 1,200 (‘appropriate’) Kcal (Table e-13, Fig. 4A). Underweight PMD adults (n = 3/22, 13.6%) consumed ‘low’ (n = 1) or ‘appropriate’ (n = 2) Kcal/day consumption. Of the 6 adults classified as overweight/obese, 1 consumed ‘very low’ Kcal/day, 2 consumed ‘low’ Kcal/day, and the remaining 3 adults consumed > 1,200 kcal/day, at a mean of 1,595 ± 96.5 kcal/day, which is within a calorie range (1,400–1,700 kcal/day) known to contribute to weight loss in obese adults [55] (Table e-13, Fig. 4A). These data demonstrate the lack of association between calorie consumption and BMI in our PMD cohort, and that malnutrition (≤ 75% predicted Kcal intake) can be present in the context of appropriate BMI.

Fig. 4

A Comparison of BMI and daily Kcal intake in adult subjects only (n = 22). Mean daily Kcal intake in the underweight (1,179 ± 227.7 kcal/day, n = 3), appropriate (1,134 ± 145.4 kcal/day, n = 13), and overweight/obese (1,143 ± 215.5 kcal/day, n = 6) adult subjects (n = 22) was not significantly different, p = 0.99, ANOVA. Subjects with low Kcal intake (801–1000 kcal/day) are indicated in orange, while subjects with very low Kcal intake (< 800 kcal/day) are indicated in purple. B Percent weight change (%) over mean interim period of 6.4 ± 1.3 months in adult subjects (n = 22). There were a total of 9/22 (40.9%) adult subjects who lost weight and 13/22 (59.1%) who gained weight. There were 2 subjects who met ASPEN and/or GLIM malnutrition criteria, one in the appropriate BMI group and one in the overweight BMI group, subjects are indicated in purple. C Percent weight change (%) over mean interim period of 11.9 ± 2.8 months in child subjects (n = 35)*. There were a total of 7/35 (20%) child subjects who lost weight and 28/35 subjects (80%) who gained weight. There were 14 subjects who met ASPEN malnutrition criteria, subjects are indicated in purple. *There were n = 3/38 subjects with no weight change observed in the interim period between prior clinic visits and RDN baseline assessment. D Malnutrition classification in adult PMD subjects (n = 22) and daily Kcal intake. There were 2/22 (9.1%) adult subjects who met malnutrition criteria, both subjects meeting ASPEN and GLIM criteria, subjects with malnutrition are indicated in purple. Mean daily Kcal intake in the underweight (1,179 ± 227.7 kcal/day, n = 3), appropriate (1,134 ± 145.4 kcal/day, n = 13), and overweight/obese (1,143 ± 215.5 kcal/day, n = 6) adult subjects (n = 22) were not significantly different, p = 0.99, ANOVA. E Malnutrition classification in child PMD subjects (n = 38) and daily Kcal intake. There were 14/38 (36.8%) child subjects who met ASPEN malnutrition criteria. Malnutrition is present in each BMI category, indicated in purple. Mean RDN-estimated Kcal intake (Kcal/day) in underweight child subjects (n = 12) was 1,010 ± 109.8 (SEM) Kcal/day, 1,240 + 86.4 kcal/day in subjects with appropriate BMI (n = 20), and 905.1 ± 103.3 kcal/day in overweight/obese (n = 6) subjects. Estimated Kcal intake was not significantly different between the three BMI groups, p = 0.08, ANOVA

Across the total adult and children PMD cohort, 11/60 (18.3%) subjects were overweight/obese. Of those, 6/11 (54.5%) gained an average of 5.7 ± 3.4 kg (range, 0.2- 22.2 kg) across a mean interim period of 11.6 months, despite their low mean Kcal intake of 909.8 ± 84.3 kcal/day (Fig. 4B–C). In summary, these results highlight that BMI should not be considered as the sole indicator of nutritional status in PMD. However, as this was a cross-sectional study, monitoring longitudinal BMI and growth in a future PMD study to characterize the nutritional status of PMD subjects would be critical.

Weight Loss was Observed in 16/60 (26.7%) PMD Subjects at Baseline RDN Visit

Having recognized that insufficient energy consumption occurred across our PMD cohort, we next explored weight trajectory, by comparing weight measured at the RDN baseline evaluation to a prior weight measurement documented in the electronic medical record at a physician visit up to ~ 12 months prior. We found a total of 16 subjects, consisting of 9 adults and 7 pediatric subjects, who presented with weight loss. At a mean interim period of 8.8 ± 1.5 months, mean weight loss was observed to be -2.4 ± 0.5 kg (mean ± SD, -4.8 ± 0.8%, range, -7.7 to -0.2 kg). When analyzed by age, mean weight loss for adults (n = 9) was -2.8 ± 0.8 kg (-4.3 ± 1.2%, range, -7.7 to -0.5 kg, mean interval period, 6.4 ± 1.3 months) and for children (n = 7) was -1.8 ± 0.5 kg (-5.4 ± 1.1%, -3.4 to -0.2 kg, 11.9 ± 2.8 months) (Table e-14, Fig. 4B–C). These child subjects had a preceding weight for age z-score of -2.7 ± 4.1 (range, -10.5 to 1.8, median, -1.2), height for age z-score of -1.2 ± 2.2 (ran