Measures of Dietary Fat and Energy Absorption in Healthy Adults

Abstract

Objectives:

Existing reference ranges for stool fat and energy absorption were developed using subjects in controlled environments on precise diets. This study measured energy and fat absorption in healthy, community-dwelling adults eating a moderate-to-high fat American diet via stool and serum-based methods.

Methods:

This was a secondary analysis of healthy subjects recruited as the comparison group in a chronic pancreatitis study. Subjects recorded dietary intake and collected stool over 3-day periods. Stool was analyzed for fat content using coefficient of fat absorption and for energy content using bomb calorimetry. The malabsorption blood test (MBT) was used to measure dietary fat absorption.

Results:

Nineteen subjects had mean daily stool measures of 143 g wet weight, 4.1 g of fat, and 178 kcal. Mean coefficients of fat and energy absorption were 96% and 93%, respectively. The mean MBT area under the curve cut point was >8 mg^xh/dL.

Conclusions:

This study confirms the historical reference range for coefficient of fat absorption in contemporary healthy, community-dwelling adults on a moderate-to-high fat diet. The study contributes to the development of reference range values for multiple bomb calorimetry-based outcomes of stool energy losses and to serum-based MBT as a promising method for measuring fat absorption.

INTRODUCTION

For more than a century, scientists have recognized that people lose energy from ingested foods in the stool.¹ In those who suffer from intestinal malabsorption, stool energy losses can be significant and increase the risk for malnutrition. Fat malabsorption is a common finding in numerous disease states, including pancreatic, liver, and intestinal diseases, and many studies focus on the amount of fat lost in the stool in these conditions. Given the higher caloric density of fat (9 kcal/g) compared to protein (4 kcal/g) or carbohydrate (4 kcal/g), stool fat losses waste more calories and have a substantial impact on overall energy balance.² Energy losses from protein and carbohydrate, however, might be significant depending upon diagnosed severity and thus total stool energy loss, including energy from fat, carbohydrate, and protein, should be considered when determining the presence and extent of energy malabsorption.³

Stool energy and macronutrient losses can be estimated with balance studies. Traditionally, energy and macronutrient balance studies were conducted in hospital-based metabolic research units where participants consumed precisely constructed meals of known caloric, micronutrient and macronutrient composition.⁴ While some investigators continue to utilize such metabolic units for modern nutritional balance studies, these units are costly to operate and of limited availability.⁵ Recently, energy balance studies have been conducted in both non-research hospital units and community-based settings allowing for assessment of stool and energy losses under varying conditions.^6–8

Reference ranges for stool total energy and macronutrient absorption in healthy people were determined in controlled metabolic unit environments, and thus it is unclear if these ranges should be applied to healthy participants in studies conducted in other settings. Evaluations of measures of stool energy and macronutrient absorption in community-dwelling study populations will support clinical care and research for both healthy people and for those with potential malabsorption symptoms.

This analysis reports stool fat and total energy absorption to evaluate current reference ranges for stool fat and energy absorption in healthy, community-dwelling adults consuming a moderate-to-high fat American diet. In addition, a serum-based method to determine fat absorption was tested in the healthy population. This serum-based test assesses fat absorption independent of dietary intake and stool output, is easy to administer, and may be more reliable than current dietary energy absorption study techniques.

MATERIALS AND METHODS

This study is a secondary analysis of healthy participants in a cohort study comparing dietary fat and total energy absorption in healthy subjects and subjects with chronic pancreatitis.⁹ The trial was approved by the Children’s Hospital of Philadelphia (CHOP) Institutional Review Board, conducted at the CHOP Center for Human Phenomic Science (CHPS), and registered with clinicaltrials.gov (NCT02849704). Three different methods for determining dietary absorption were evaluated. The coefficient of fat absorption (CFA) and the malabsorption blood test (MBT) evaluated dietary fat absorption, and stool bomb calorimetry (BC) evaluated total energy absorption. The primary outcomes for this study were the percentage of fat absorbed as determined by CFA as well as multiple measures of total stool energy loss determined by BC: daily stool energy loss, dry stool caloric density, and coefficient of energy absorption (CEA). The secondary outcomes for this study were triglyceride absorption as measured by the MBT as well as fat-soluble vitamin concentrations.

Study Subjects

Subjects were recruited as healthy participants for the primary study between August 2016 and November 2017. The CHOP Recruitment Enhancement Core provided assistance with recruitment from the region and internal communication resources. Healthy subjects were considered for inclusion in the study if they were between 30–75 years of age, had a body mass index (BMI) of 14–35 kg/m², were in their usual state of good health (no change in medications, diet or weight in the previous two weeks), and were able to consume a moderate-to-high fat diet. Subjects were excluded if they were pregnant or breastfeeding, had evidence of fat malabsorption, or had a chronic illness or were taking medications that could affect fat absorption or intestinal transit.

Study Visit

Subjects underwent a 7-day study protocol (Fig. 1) including a study visit where they provided a complete health history and demographic information. Anthropometric measurements (weight, height) were obtained in triplicate by CHPS research staff according to standardized techniques.¹⁰ Weight (to 0.1 kg) was measured using a digital scale (Seca, Munich, Germany); standing height (to 0.1 cm) was measured using a stadiometer (Holtain, Crymych, UK). The mean of the three measurements was used for analysis and BMI (kg/m²) calculation.¹¹ Fasting serum was collected for fat-soluble vitamin concentrations and for the MBT. Subjects were instructed on how to accurately record prospective, weighed dietary intake by a CHPS research dietitian and taught how to perform a complete 3-day stool collection.

FIGURE 1.

Study protocol.

Blood Samples

Blood samples were obtained after a minimum 12-hour fast. Concentrations of 25-hydroxyvitamin D (25(OH)D) were measured (Clinical Laboratories, CHOP, Philadelphia, Penn). Serum retinol and α- and γ-tocopherol were assessed using high-performance liquid chromatography (Eurofins Craft Laboratories, Wilson, NC). The reference ranges for serum fat soluble vitamin concentrations for this report were: 25(OH)D ≥32 ng/mL, retinol 20–100 mcg/dL, and α-tocopherol 5–18 mcg/mL.^12,13

The MBT was used to assess dietary fat absorption. After a 12-hour fast and 24 hours without dairy products, the MBT was conducted by providing an oral dose of the test fats pentadecanoic acid (PA), a free fatty acid, and triheptadecaenoin (THA), a triglyceride, in a high-fat/high-calorie liquid test meal (32g fat, 550 kcal). These fats were selected as they are easy to measure in serum and are uncommon in the traditional American diet with minute amounts found in dairy products. Triheptadecaenoin is a triglyceride made up of three heptadecanoic (HA) fatty acids that requires hydrolysis by pancreatic lipase for absorption. The test fats were delivered in equal micromolar amounts (5.0g PA, 5.5 g THA) well above the concentrations that might be consumed with typical dietary intake. Serial serum concentrations of PA and HA were evaluated over time to measure fat absorption.^14,15 Blood was collected before test meal administration and then once every hour over the 8-hour testing period. Concentrations of PA and HA were determined by gas-liquid chromatography (Wake Forest University, Winston-Salem, N.C.).¹⁶

A non-compartmental analysis was used to estimate the area under the curve (AUC) for PA and HA micromolar concentrations as well as for the ratio of AUC HA to AUC PA. For the purposes of this study, the reference range of AUC HA >8 mg^xh/dL was used to reflect a typical pattern of absorption in healthy participants.⁹ The AUC HA was selected as the outcome to demonstrate the difference in lipase-dependent triglyceride digestion and absorption between healthy people and those with fat malabsorption. Subjects were categorized as either within the typical absorption range or below the reference range signifying reduced dietary fat absorption.

Dietary Intake Record

Subjects were provided with a digital food scale, measuring cups and measuring spoons to record their 3-day dietary intake in a prospective manner. All subjects were instructed to eat a moderate-to-high fat diet with a goal of 80 g fat/d or higher. The research dietitian taught subjects how to read food labels and provided them with educational materials to support diet pattern choices for the medium-to-high fat intake.

Dietary intake was analyzed using Nutrition Data System for Research software version 2012 (National Coordinating Center, University of Minnesota, Minneapolis, Minn).¹⁷ Results of that analysis were used to determine daily energy (kcal) and fat (g) intake. Those values were used to determine percentage of daily calories consumed as fat (%) and the percentage of estimated energy requirement (EER; %) for each subject. The EER was calculated based on age, sex, weight, height and level of physical activity.¹⁸ For this report, the low active EER formulas were used.

Stool Samples

Subjects were taught how to collect, store and submit a 3-day stool collection and were provided with collection and shipping containers. Total stool collections were weighed and then homogenized (BagMixer™, Interscience Labratories, Woburn, Mass). An aliquot of the homogenized stool was extracted for BC analysis (Obesity and Diabetes Clinical Research Section, Phoenix, Ariz). The remaining stool sample was assessed for fat content by nuclear magnetic resonance spectroscopy (Mayo Clinic Laboratories, Rochester, Minn).¹⁹

Coefficient of Fat Absorption

Coefficient of fat absorption was used to measure dietary fat absorption. The CFA was determined by using the 3-day prospective, weighed food records and 3-day stool collections.²⁰ The CFA was calculated as:

$$CFA(\%)=\frac{[gramsoffatconsumed-gramsoffatinstool]}{gramsoffatconsumed}\times 100\%$$ Equation 1:

A reference range of ≥93% was used for CFA results and subjects were categorized as within range or below range.²¹

Bomb Calorimetry

Bomb calorimetry was used to assess stool energy loss. Each aliquot was thawed and mixed with 50 g of water. Of the resultant stool slurry, 50 g were desiccated by freeze-drying for approximately 48 hours. Multiple one-gram pellets of desiccated stool were prepared. Individual pellets were placed in a bomb calorimeter (PARR Instrument Co, Moline, Ill) for calorimetric measurement. Two pellets were analyzed for each subject and the mean was reported (cal/g). If the results of the first pellet pair varied by more than 40 cal/g, then two additional pellets were tested and the mean of the second pair was reported if those values were within 40 cal/g of each other. If neither pellet pair resulted in values within 40 cal/g of each other, then all four values were averaged and this value was reported (cal/g). The bomb calorimeter was calibrated prior to the first run and then after every ten runs (personal correspondence, Obesity and Diabetes Clinical Research Section).²²

The caloric density (cal/g) of the dry stool sample for each subject was converted to kcal/g and then to calculate the caloric density of the original sample of wet stool (kcal/g), the total (kcal) and daily (kcal/d) stool energy loss, and the coefficient of energy absorption (CEA; %) for each subject. The caloric density of the wet stool was calculated (Equation 2) and then used to calculate the total stool energy loss (Equation 3).

$$Caloricdensityofwetstool(kcal/g)=caloricdensityofdrystool\times \frac{massofdrystool}{massofwetstool}$$ Equation 2:

$$Totalstoolenergyloss(kcal)=totalmassofwetstool\times caloricdensityofwetstool$$ Equation 3:

The total stool energy loss was averaged over the 3-day collection period to determine daily stool energy loss (Equation 4) and the CEA (Equation 5).⁷

$$Dailystoolenergyloss(kcal/d)=\frac{totalenergylossinstool}{3days}$$ Equation 4:

$$CEA(\%)=\frac{[kcalofenergyconsumed-kcalofstoolenergyloss]}{kcalofenergyconsumed}\times 100\%$$ Equation 5:

From the above values, the daily stool energy loss, dry stool caloric density, and CEA were used as primary outcomes for BC testing.

Multiple investigators have reported values for daily stool energy loss, dry stool caloric density, and CEA for healthy subjects using different approaches.^7,8,22–24 For our analysis, we weighted the means and standard deviations of the published results from those studies by sample size to create an overall weighted reference range for each of the three specified measures of stool energy content.^7,8,22–24 The resulting weighted reference ranges were then used to categorize our subjects as within range, above range, or below range for each of the three measures of stool energy loss calculated from the BC analysis (daily stool energy loss, dry stool caloric density, and CEA). In addition, we revised the weighted reference ranges from the literature by adding our weighted results from this study to the pooled reference range analysis.

Statistical Analysis

Categorical variables were analyzed for frequency counts and reported as percentages. Continuous variables were analyzed for means and reported with standard deviations and ranges with minimums and maximums. Associations between variables were evaluated by Pearson’s correlation coefficient (r) with P < 0.05 considered statistically significant.

Results for individual measures were categorized as within, above or below reference range as previously described.

Literature Review

A literature review was conducted to identify foundational articles describing the methodology for CFA, BC, and MBT analysis as well as to identify key articles to serve as the basis for generally accepted reference values for each test. An online search in PubMed was carried out to identify articles indexed in that database through December 31, 2018. The primary search queries were ‘Coefficient of Fat Absorption,’ ‘Stool Bomb Calorimetry,’ and ‘Malabsorption Blood Test.’ Additional, more specific search terms were used as needed to refine searches. Cited reference searches were done for selected articles to identify additional sources.

Foundational articles for each test were summarized in an annotated table (see Supplemental Table 1, annotated table of foundational articles for each measure of fat and energy absorption). Key articles that provided background for the reference ranges for healthy populations for each measure of fat and energy absorption were annotated and summarized as well (see Supplemental Table 2, annotated table of key articles defining the reference ranges for each measure of fat and energy absorption) and used for reference range determination as described.

RESULTS

A total of 19 healthy subjects fulfilled the inclusion criteria for secondary analysis. Of the 24 healthy subjects enrolled in the original study, two were excluded from that study and this secondary analysis due to undisclosed illnesses at initial enrollment: one had a history of chronic pancreatitis and the other had clinically significant liver disease requiring medication that affected dietary fat absorption.⁹ An additional three subjects were excluded from this secondary analysis due to opiate medication use for non-gastrointestinal conditions that were likely to affect gastrointestinal transit and thus may impact the results of stool analyses.

Demographics, anthropometrics, and micronutrient concentrations are presented in Table 1. The subjects included for secondary analysis ranged in age from 30–61 years and were 63% female and 58% white. The BMI range was 16–35 kg/m² with one underweight subject (BMI <18.5 kg/m²), eight overweight subjects (BMI 25–29.9 kg/m²), and three obese subjects (BMI ≥30 kg/m²).

TABLE 1.

Subject Characteristics

Characteristic	n = 19
Demographics
Age, mean (SD), y	41.6 (9.8)
Sex, female, n (%)	12 (63)
Race, n (%)
White	11 (58)
African American	7 (37)
East Asian	1 (5)
Anthropometrics, mean (SD)
Weight, kg	75.0 (16.4)
Height, cm	170.6 (6.4)
Body mass index, kg/m2	25.6 (4.7)
Fat-soluble vitamin concentrations, mean (SD)
25(OH)D, ng/ml	30.9 (16.2)
Retinol, μg/dl	55.9 (15.9)
α-Tocopherol, μg/ml	11.9 (1.9)
λ-Tocopherol, μg/ml	1.8 (0.7)

25(OH)D indicates 25-hydroxy Vitamin D.

Dietary intakes and measures of dietary fat absorption are presented in Table 2. Subjects consumed a range of 1068–4454 kcal/d. Four subjects consumed <1600 kcal/d and four consumed >3000 kcal/d.² Individually, subjects had a range of 40–192% of their estimated energy requirement (EER) with a mean of 103%. Daily fat intake ranged from 45–196 g fat with a mean of 109 g fat. Six subjects consumed <80 g fat/d and four subjects consumed >150 g fat/d. Fat contributed 27–62% of daily calories in our subjects’ diets. Total stool weight ranged from 168–765 g with a mean of 430 g and daily stool weight ranged from 56–255 g with a mean of 143 g.

TABLE 2.

Outcome Measures

Outcome Measure	n = 19
Dietary intake per day*
Energy intake, kcal	2420 (894)
Estimated energy requirement, %	103 (39)
Fat intake, g	109 (45)
Energy intake from fat, %	41 (8)
Stool weight†
Total, g	430 (142)
Average daily, g/d	143 (47)
Coefficient of fat absorption
Stool fat, g/d	4.1 (2.6)
Coefficient of fat absorption, %	96 (2.0)
Bomb calorimetry‡
Daily stool energy loss, kcal/d	178 (69)
Dry stool caloric density, kcal/g	5.11 (0.29)
Coefficient of energy absorption, %	93 (3.1)
Malabsorption blood test‡
AUC PA, mgxh/dL	22.6 (10.5)
AUC HA, mgxh/dL	18.5 (10.5)
AUC HA/PA ratio	6.7 (1.2)

Data presented as mean (SD)

^*Calculated from 3-day prospective, weighed food records

^†From 3-day stool collection

^‡n = 18

AUC indicates area under the curve; PA, pentadecaenoic acid; HA, heptadecaenoic acid.

Outcome Measures

The results of the outcome measures are shown in Table 2 with subject categorization delineated in Table 3 and individual subject results reported in Table 4. Stool fat loss ranged from 1–11 g/d with a mean of 4.1 g/d. A reference range of 2–7 g/d was used (Mayo Clinic Laboratories). Three subjects had daily fat output below range and one subject had a stool fat output above range. The CFA for the group ranged from 92 to 99% with a mean of 96%. Only one subject had a CFA below the range of 93% at 92%.

TABLE 3.

Subject Outcomes by Method and Reference Range

	Reference Range	Subjects Within Range, n (%)	Subjects Below Range, n (%)	Subjects Above Range, n (%)
Fat-soluble vitamins (n = 19)
25(OH)D, ng/mL	≥32*	8 (42)	11 (58)	−
Retinol, mcg/dL	20–100†	19 (100)	0	−
α-Tocopherol, mcg/mL	5–18†	19 (100)	0	−
Outcome measures
Coefficient of fat absorption (n = 19)
Stool fat, g/d	2–7‡	15 (79)	3 (16)	1 (5)
Coefficient of fat absorption, %	≥93§	18 (95)	1 (5)	−
Bomb calorimetry (n = 18)
Daily stool energy loss, kcal/d	120–200||	10 (55)	3 (17)	5 (28)
Dry stool caloric density kcal/g	4.9–6.0¶	13 (72)	5 (28)	0
Coefficient of energy absorption, %	≥90#	15 (83)	3 (17)	−
Malabsorption blood test (n = 18)
AUC HA, mgxh/dL	≥8**	16 (89)	2 (11)	−

^*Reference range per Hollis et al 2005¹²

^†Reference range per Kratz et al 2004¹³

^‡Reference range per Mayo Clinic Laboratories, Rochester, Minn.

^§Reference range per Borowitz et al 2007²¹

^||Reference range extrapolated from multiple sources^7,8,22–24

^¶Reference range extrapolated from multiple sources^7,24

^#Reference range extrapolated form multiple sources^7,8

^**Reference range per Brownell et al 2019⁹

AUC indicates area under the curve; PA, pentadecaenoic acid; HA, heptadecaenoic acid; 25(OH)D, 25-hydroxy vitamin D; −, not applicable

TABLE 4.

Individual Subject Detail

	Demographics				3 Day Diet Record				3 Day Stool Collection			Bomb Calorimetry			MBT
Subject No.	Age, y	Race	Sex	BMI, kg/m2	Energy Intake, kcal/d	EER, %	Fat Intake, g/d	Energy Intake as Fat, %	Stool Weight, g/d	Stool Fat, g/d	CFA, %	DSEL, kcal/d	DSCD, kcal/g	CEA, %	AUC HA, gxh/dL
1	30	W	F	26.2	4454	192	182	37	110	3	–	–	–	–	n/a
2	31	W	F	21.2	1736	76	91	47	122	5	–	n/a	n/a	n/a	–
3	31	W	F	26.3	2256	96	104	41	126	5	–	–	–	–	–
4	32	AA	M	29.2	1748	65	77	39	113	1	–	↓	↓	–	–
5	33	W	M	31.3	3536	122	167	43	180	11	–	↑	–	–	–
6	33	AA	M	23.3	2874	121	97	30	113	2	–	–	–	–	–
7	38	W	F	19.1	2703	125	135	45	181	7	–	↑	–	–	–
8	38	AA	F	29.3	1202	50	45	34	166	2	–	–	↓	↓	–
9	38	AA	F	19.1	2161	100	69	29	109	4	–	–	–	–	–
10	39	AA	F	27.5	3331	137	196	53	176	4	–	↑	↓	–	–
11	39	AA	F	16.2	1585	77	64	36	56	1	–	↓	↓	–	↓
12	39	W	F	25.4	2362	102	96	37	191	2	–	–	–	–	–
13	44	As	F	24.2	2582	119	110	38	189	5	–	–	–	–	–
14	47	W	M	29.5	1068	40	74	62	155	6	↓	–	–	↓	–
15	53	AA	M	31.4	1502	58	45	27	111	1	–	↓	↓	–	↓
16	54	W	M	34.7	2811	101	127	41	109	3	–	–	–	–	–
17	54	W	F	25.7	2091	92	98	42	255	7	–	↑	–	↓	–
18	56	W	F	24.2	2247	120	122	49	87	3	–	–	–	–	–
19	61	W	M	23.4	3726	168	172	42	176	6	–	↑	–	–	–

W, White; AA, African-American; As, Asian; M, male; F, female; EER, estimated energy requirement; CFA, coefficient of fat absorption; DSEL, daily stool energy loss; DSCD, dry stool caloric density; CEA, coefficient of energy absorption; MBT, malabsorption blood test; AUC, area under the curve; HA, heptadecaenoic acid; –, within reference range for test; n/a, sample not available for analysis; ↓ below reference range; ↑ above reference range

Bomb calorimetry was completed in 18 of 19 subjects. The BC sample for one subject was contaminated during analysis so that subject was excluded from all of the BC outcomes. Subjects lost a range of 58–331 kcal/d in the stool with a mean of 178 kcal/d. Using a reference range of 120–200 kcal/day, three subjects had stool energy losses below range and five had losses above range. The dry stool caloric density ranged from 4.7–5.5 kcal/g with a mean of 5.11 kcal/g. A reference range of 4.9–6.0 kcal/g was used for dry stool caloric density. Of our subjects, 13 were within range, five were below range and no subject was above range. The CEA ranged from 86–96% with a mean of 93%. With a reference range of ≥ 90%, three subjects had a CEA below range at 86%, 86% and 87%.

Daily stool energy loss was compared to daily stool weight, dry stool caloric density, daily stool fat loss, and CFA to assess for correlation between measures. Daily stool energy loss correlated positively with daily stool fat loss (Pearson’s r = 0.90, P < 0.00001; Fig. 2A), daily stool weight (Pearson’s r = 0.75, P < 0.001; Fig. 2B), and dry stool caloric density (Pearson’s r = 0.59, P < 0.05; Fig. 2C) and negatively with CFA (Pearson’s r = −0.59, P < 0.01; Fig. 2D).

FIGURE 2.

Scatter plots demonstrating daily stool energy loss plotted by measure of stool fat or energy absorption. A, Relationship between stool energy loss (kcal/d) and daily stool fat loss (g). Daily stool fat loss was positively correlated with daily stool energy loss (Pearson’s r = 0.90, P < 0.00001). B, Relationship between stool energy loss (kcal/d) and daily stool wet weight (g). Daily stool production was positively correlated with daily stool energy loss (Pearson’s r = 0.75, P < 0.001). C, Relationship between stool energy loss (kcal/d) and dry stool caloric density (kcal/g). Dry stool caloric density was positively correlated with daily stool energy loss (Pearson’s r = 0.59, P < 0.05). D, Relationship between stool energy loss (kcal/d) and coefficient of fat absorption (%). The CFA was negatively correlated with daily stool energy loss (Pearson’s r = −0.59, P < 0.01).

The MBT was conducted in all 19 subjects; one subject, however, was excluded from analysis due to poor compliance with the test protocol (failed to consume the MBT test meal within the specified timeframe). The AUC PA, representing the absorption of the free fatty acid, ranged from 10.9–48.1 mg^xh/dL with a mean of 22.6 mg^xh/dL. The AUC HA, representing the absorption of the fatty acid liberated from the administered THA by lipase digestion, ranged from 7.1–47.8 mg^xh/dL with a mean of 18.5 mg^xh/dL. The ratio of AUC HA to AUC PA ranged from 3.7–8.9 with a mean of 6.7. Using the cut point of AUC HA >8 mg^xh/dL, two subjects were below the cut point (7.1, 7.4 mg^xh/dL) and all other subjects were above the cut point.

Fat Soluble Vitamins

The means and standard deviations for fat-soluble vitamin concentrations are outlined in Table 1 with the reference ranges and subject categorizations shown in Table 3. The 25(OH)D concentration ranged from 4.8–74.3 ng/mL with a mean of 30.9 ng/mL. Eleven subjects were below the reference range of 32 ng/dL and no subject was at risk for vitamin D toxicity (>100 ng/mL).¹² Vitamin A (retinol) concentrations ranged from 28.3–85.7 μg/dl and no subject was out of reference range.¹³ Vitamin E concentrations ranged from 8.4–15.4 μg/ml and no subject was out of reference range.¹³

DISCUSSION

This study is a comprehensive inquiry of dietary fat and energy absorption in a healthy, community-dwelling American population, and provides a review of the history of methods. Dietary absorption was investigated using CFA, generally considered the gold standard for measuring stool fat loss, and the MBT, a serum-based test for measuring intestinal fat digestion and absorption. Total stool energy loss was investigated by BC, a long-standing method for measuring total energy content (fat, carbohydrate and protein). The results confirm previous published and utilized CFA reference ranges in our contemporary sample of subjects with these collection and analytical methods. In addition, our results contribute to the growing body of evidence for BC assessment of stool energy loss in healthy subjects. Lastly, our results provide additional data for the MBT, a method for detecting fat malabsorption patterns at the gut mucosal level in research and clinical care.

Although development of the CFA began at the turn of the 20^th century, the process was refined during the late 1940s and early 1950s with the aim to determine the best laboratory method for extracting fat from stool (See Supplemental Table 1, annotated table of foundational articles for each measure of fat and energy absorption).^1,25,26 The van de Kamer titration method became the preferred method to extract fat from stool; now newer analytical methods, such as nuclear magnetic resonance spectroscopy (NMR), are commonly used and correlate well with the titration method.¹⁹

Early investigators established reference ranges for the CFA in healthy individuals, reporting cut points of 90 to >95%.^27–29 More recent CFA studies have recommended a cut point of ≥93% to indicate typical fat absorption in healthy people (see Supplemental Table 2, annotated table of key articles defining the reference ranges for each measure of fat and energy absorption).²¹ Coefficient of fat absorption has been measured in various disease states where fat malabsorption is common such as cystic fibrosis, chronic pancreatitis and acute infectious diarrhea.^8,21,30 These studies confirm the utility of 93% as a cut-point to identify fat malabsorption.

In our study of healthy adult participants, only one subject had a CFA slightly lower than the 93% cut point at 92% (Table 4). This subject had a total fat intake of 74 g/d; however, due to his low energy intake (EER 40%), he had the highest percent energy intake from fat in our sample (62%). Interestingly, this subject’s stool fat content was within the normal range (6g). Of note, the investigators correlating NMR spectroscopy with gravimetric methods cautioned that irrespective of methodology, fecal fat results close to the cut-off point for malabsorption (7 [standard deviation, 1] g) should be considered in the context of the patient’s clinical symptoms suggesting the cut-point is more of a transition zone than a clear division between normal absorption and malabsorption.¹⁹ Conversely, one subject had an elevated stool fat (11 g) with a CFA in the normal range. The results from these two subjects suggest that both stool fat content and CFA may be considered when determining the presence of fat malabsorption as each of these healthy subjects had one abnormal and one normal value. Overall, the CFA values in our contemporary sample confirmed the reference range of ≥93% for healthy people eating a moderate-to-high fat diet. Our study is the first study of community-dwelling, healthy American adults to confirm the reference range for CFA assessed by NMR spectroscopy.

Bomb calorimetry was delineated in the 1950s and further developed in the late 1960s and early 1970s (See Supplemental Table 1, annotated table of foundational articles for each measure of fat and energy absorption).^31,32 One of the earliest articles describing BC for human stool analysis was published in 1969.²³ Despite development shortly after CFA, the utilization and reference ranges for BC are less well defined than those for CFA. Original studies found variance in stool energy output related to dietary intake of unavailable (non-digestible) carbohydrate and minimal difference among subjects of different demographic groups (male/female, younger/older adults) consuming the same diet.²³ More recent BC studies have examined stool energy loss patterns in: American women; lean and obese individuals consuming diets with varying energy intakes; healthy Dutch individuals; and, healthy Norwegian individuals (see Supplemental Table 2, annotated table of key articles defining the reference ranges for each measure of fat and energy absorption).^7,8,22,24 All of these studies reported stool energy loss (kcal/d or kJ/d) while only some studies provided the dry stool caloric density (kcal/g stool or kJ/g stool) or CEA for stool (%). As previously described, for this study we used the results from these previously published studies to calculate weighted reference ranges for three measures of stool energy loss derived from BC assessment as part of our comparisons.

For daily stool energy loss, we calculated a range of 120 to 200 kcal/d using the results of the healthy participants in the aforementioned studies.^7,8,22–24 Ten of our subjects fell within this reference range with three below and five above range (Table 3). The details of individual subject results are provided in Table 4. The three subjects who fell below the range (with daily stool energy losses of 58, 105, and 113 kcal/d) also had low stool fat losses of 1g/d (reference range of 2 to 7 g/d). Of the five subjects who had daily stool energy losses above range, one had a stool fat loss of 11 g (above range), two had stool fat losses of 7 g/d (at the cut-point), and one had a stool fat loss of 6 g/d (within range). This pattern demonstrates that in these healthy adults, stool fat loss was a major source of total stool energy losses even in a healthy population without fat malabsorption.

Stool energy loss can also be described by measuring the dry stool caloric density. Our range of 4.9 to 6.0 kcal/g was derived from the two previous studies that reported dry stool caloric density.^7,24 Thirteen of our subjects were within range, five subjects were below range, and no subject was above range (Table 3). Of the five subjects below the reference range, three had low stool calories losses, one had stool calorie losses within the reference range and one had high stool calorie losses (Table 4). This pattern suggests that stool energy loss (kcal/g) was less related to stool caloric density than it is to stool fat loss. Three of the five subjects had both low energy and fat intakes while one subject had high energy and fat intakes. The dry stool caloric density was more variable and might be less useful in differentiating healthy people from those with malabsorption. Additional evidence is needed to further evaluate the utility of dry stool caloric density as a measure of stool energy loss.

Coefficient of energy absorption is the final measure calculated from the BC results. Similar to the CFA for fat, CEA estimates the amount of energy absorbed in the gastrointestinal tract relative to the amount of dietary energy consumed. The reference value for CEA of ≥90% was suggested by two recent European studies in healthy, ambulatory people.^7,8 Of the 18 subjects in the current study whose stool was evaluated by BC, only three were below the CEA range (86%, 86%, 87%, Table 3). There were no identifiable patterns among the markers of fat and energy absorption for these three subjects with low CEA. One of the subjects had an elevated daily stool energy loss (301 kcal/d) while the other two had low energy intakes, both when evaluated by mean daily energy intake and by EER. Of the two subjects with low energy intake, one had a low percentage of dietary fat (34%) and one had a high percentage of dietary fat (62%). Of the three measures of stool energy loss determined from BC (daily stool energy loss, dry stool caloric density, and CEA), CEA had the fewest subjects outside of the reference range. This pattern suggests that CEA might be a better BC outcome for determining stool energy loss than either the daily stool energy loss or dry stool caloric density.

The CEA cut-point of ≥90% has been confirmed in a number of studies evaluating patients with disease states known to cause malabsorption such as diarrheal illness in intensive care unit patients, patients with pancreatic cancer, and patients with pancreatic insufficiency due to chronic pancreatitis.^8,33,34 The mean CEA for subjects with conditions associated with malabsorption was <90% in each of these studies . These findings support using a reference range of ≥90% for CEA.

It is important to understand how our results compare to the two recent European stool energy studies as both included healthy, community-dwelling people following protocols for dietary record reporting and stool collection and evaluation similar to our study. The first study by Wierdsma et al described intestinal absorptive capacity in 23 healthy Dutch adults ages 22–60 y who were consuming their usual Western European diet in an ambulatory setting.⁷ Participants recorded a 4-day prospective, weighed dietary record with stool collection occurring over the last three days of that period. Demographic information of the subjects (separated by male and female participants) and selected results are shown in Table 5.

TABLE 5.

Summary of Recent Similar Studies

	Reference Range	Wierdsma et al, 20147	Erchinger et al, 20188	Current Study 2019
Study location		Netherlands	Norway	United States
Sample size, n		23	12	19
Demographics
Age, y		−	52	41.6 (9.8)
Male		43.2 (11.8)	−	44.7 (12.0)
Female		42.9 (14.2)	−	39.8 (8.2)
Sex, female, n (%)		61	75	63
Body mass index, kg/m2		−	24	25.6 (4.7)
Male		25.3 (2.7)	−	29.0 (4.2)
Female		23.1 (2.2)	−	23.7 (3.9)
Outcome Measures
Daily fat intake, g		−	95	109 (45)
Male		89 (16)	−	108 (49)
Female		68 (19)	−	109 (45)
Coefficient of fat absorption
Stool fat, g/d	2–7	5.2 (2.2)	6.7 (3.1)	4.1 (2.6)
Coefficient of fat absorption, %	≥93	92.5 (3.7)	92.1 (2.9)	96 (2.0)
Bomb calorimetry
Daily stool energy loss, kcal/d	120–200	213 (67)*	201.3 (73.9)	178 (69)†
Stool caloric density kcal/g	4.9–6.0	5.40 (0.6)*	−	5.11 (0.29)†
Coefficient of energy absorption, %	≥90	89.4 (3.8)	90.0 (3.1)	93 (3.1)†

Data are presented as mean or mean (SD) unless otherwise indicated.

^*Data reported as in kJ in original report; converted to kcal for this report

^†n = 18

− indicates not available.

In comparison to our subjects, the Dutch subjects were of a similar age, sex predominance, and BMI but had lower mean daily fat intake than our subjects. Overall, the Dutch subjects had a higher mean stool energy loss, a higher mean dry stool caloric density, and a lower mean CEA than our subjects. These differences suggest that the Dutch subjects had greater stool energy losses than our subjects did. The Dutch group also had a lower mean CFA than our group (92.5% vs 96%) despite lower daily fat intake suggesting that a significant amount of stool energy loss was likely due to fat loss in this healthy population. Of note, the Dutch group assessed CFA by the van de Kamer method. Despite a low mean CFA, the mean stool fat lost per day was within the reference range, again suggesting that these values must be evaluated together to determine the presence of fat malabsorption. It is possible that these differences are due to the different methods used to assess CFA or to the difference in fat composition between a Northern European and an American diet. One early study found that CFA could be decreased by as much as 5% from the expected value for healthy people when large amounts of animal fats (e.g. sheep, cow, pig) were consumed.²⁹ Specific details of the food composition of the Dutch subjects’ fat intake were not reported.

The other recent European study, published by Erchinger et al, was a prospective interventional study of Norwegian patients with chronic pancreatitis and a healthy comparison group evaluating the response of increasing doses of pancreatic enzyme replacement therapy (PERT) on stool fat and energy losses.⁸ All subjects were community dwelling and consumed their regular diet. Dietary intake was assessed by analysis of 3-day prospective, weighed food records and stools were collected the subsequent three days of each week. All subjects were evaluated on increasing doses of PERT. Demographic information and selected results are shown in Table 5.

The 12 Norwegian healthy subjects were older, included more females, and had a similar BMI to both our subjects and to the Dutch subjects. Their mean daily fat intake was closer to that of our participants and higher than that of the Dutch subjects. The Norwegian healthy subjects had a higher mean daily stool energy loss than our subjects but lower daily stool energy loss than the Dutch subjects. The Norwegian subjects also had a lower mean CEA than our subjects, more consistent with the CEA of the Dutch subjects. Their mean CFA was also lower than our findings and again closer to the Dutch result. Of note, the CFA was assessed by a modified van de Kamer method developed by their group.³⁵ As with the Dutch study, in spite of a low mean CFA, the mean stool fat lost per day was within the reference range, again suggesting that these values should be evaluated together when determining the presence of fat malabsorption for a community-dwelling population. The Norwegian study did not report the dry stool caloric density.

Both of the European studies compared daily stool energy loss to other markers of stool fat and energy absorption to identify correlations among different measures. For healthy subjects, both studies found a positive correlation between stool weight and stool energy loss. The Norwegian study also found a positive correlation between stool fat loss and stool energy loss. We found a similar, strong positive correlation between stool energy loss and stool fat loss in our subjects (Fig. 2A) and a less strong positive correlation between stool energy loss and stool weight (Fig. 2B). There was a moderate positive correlation between dry stool caloric density and stool energy loss (Fig. 2C) and a moderate negative correlation between CFA and stool energy loss (Fig. 2D). These findings suggest that daily stool fat content and daily stool weight may be good proxy measures for total stool energy loss in healthy adults. Dry stool caloric density may not be as useful for estimated overall stool energy losses. Interestingly, CFA was less strongly correlated with stool energy loss than stool fat loss, potentially due to its reliance on reported dietary intake data, which was variable as evidenced by the ranges of energy and fat intake for all three studies. Overall, the approach of combining data from multiple similar studies should be considered in the development of generally accepted reference ranges for BC-derived stool measures of energy absorption in a healthy population.

Our study results contribute to the development of reference ranges for BC assessments of stool energy losses in healthy, community-dwelling adults. When the data from our subjects were weighted and added to the previous data to create the ranges utilized in this report, all of the previous ranges were affected.^7,8,22–24 After the new calculations, the reference range for stool energy loss narrows from 120 to 200 kcal/d to 130 to 200 kcal/d, the range for dry stool caloric density adjusts downward from 4.9 to 6.0 kcal/g to 4.8 to 5.7 kcal/g, and the CEA range adjusts upwards from ≥90.0% to ≥90.5%.

Studying a community-dwelling population decreases the burden and cost of using a hospital-based research metabolic unit for nutrient balance studies and might be more acceptable for clinical care and research. Limitations, however, exist for both CFA and BC when conducted in a less controlled setting. Both CFA and BC rely on the meticulous reporting of dietary intake consumed and the diligent collection of all stool passed during the study period. In the community setting, even when participants weigh and record their dietary intake as instructed, there is room for error in intake calculations due to inaccurate food and beverage measurement, omission of intake, and estimations required to report consumption of non-labeled foods (i.e. restaurant prepared meals where each component cannot be individually weighed). In addition, subjects must collect all stool during the study period, which can be challenging when participants leave the home. In our study, CFA results were not affected by transitioning to community-based evaluation, but may still be affected by changing from gravimetric to NMR spectroscopy-based methods. The evidence for BC was too limited to make a determination on historical reference range application by changing to the community setting. With three similar contemporary studies of healthy community-dwelling subjects, however, reference ranges for BC in ambulatory settings are evolving. Ultimately, the development of an absorption test that does not require admission to a metabolic unit or reliance on subject adherence and diligence will be beneficial.

An alternative to dietary intake and stool collection for absorption testing has been recommended for some time. Various methods including spot stool evaluations, breath tests, and blood-based tests have been developed and explored without widespread clinical adoption.^36–39 The MBT was developed to address this clinical care and research gap, and results have demonstrated the ability to detect fat malabsorption.³⁶ Since it is performed in a single day, the MBT is less time consuming than either the CFA or BC. By using precise amounts of selected fats and serum sampling, the MBT emancipates the investigator and subject from relying on the dual burdens of dietary reporting and stool collection. The strength of the MBT lies in its ability to measure fat absorption at the gut mucosal site independent of these factors, as well as any others that may affect stool output, such as intestinal motility.

The MBT has been evaluated in both healthy participants and those with known malabsorption with cystic fibrosis and chronic pancreatitis (See Table, Supplemental Digital Content 1, annotated table of foundational articles for each measure of fat and energy absorption).¹⁵ For the purposes of this report, a cut-point for AUC HA of ≥8 was selected. In this healthy population, only two subjects had values that were below the cut point of 8.0 (7.1, 7.4). Both of those subjects also had daily stool energy losses and stool caloric densities below range. Interestingly, each subject only lost one gram of fat in the stool, consistent with normal fat absorption by CFA. The results of this study will add to the evidence and experience necessary to continue to refine the development of the MBT for use in clinical care and research.

In summary, our results confirm the previous CFA reference range in a contemporary sample of healthy community-dwelling subjects and contribute to the growing understanding of BC assessment of stool energy loss in healthy subjects in a community setting. While CFA is considered the standard for identifying fat malabsorption, it is infrequently utilized due to the burden of dietary intake recording and 3-day stool collection. In evaluating CFA in a community-dwelling population, it might be informative to examine both total stool fat loss (g) and CFA (%) to consider the presence of fat malabsorption in an individual. Although it has been utilized less often than CFA, BC was performed in recent stool fat and energy absorption studies. Like CFA, BC relies on detailed dietary intake records and precise stool collections. When BC is used, different outcome measures are available to assess the results. Our results suggest that CEA might be more useful than daily stool energy loss or dry stool caloric density for delineating healthy people from those people with elevated stool energy loss. In addition, stool fat content and stool wet weight might be sufficient proxies for stool energy loss in healthy people. Lastly, these results provide additional experience and data for the utility of MBT, a potential method for detecting fat malabsorption independent of dietary intake or stool collection.

Supplementary Material

Supplemental Data File (doc, pdf, etc.)

Supplemental Digital Content 1: Annotated table of foundational articles for the measures of fat and energy absorption used in this study .pdf

Supplemental Digital Content 2: Annotated table of key articles defining the reference ranges for each measure of fat and energy absorption.pdf