Postprandial Metabolic Responses Differ by Age Group and Physical Activity Level

Abstract

Objectives

To compare the postprandial metabolic responses to a high-fat meal in healthy adults who differ by age and physical activity level.

Design

Cross-sectional, quasi-experimental design.

Setting

Physical Activity and Nutrition Clinical Research Consortium (PAN-CRC) at Kansas State University (Manhattan, KS, USA).

Participants

Twenty-two healthy adults: 8 younger active (YA) adults (4M/4W; 25 ± 5 yr), 8 older active (OA) adults (4M/4W; 67 ± 5 yr), and 6 older inactive (OI) adults (3M/3W; 68 ± 7 yr).

Intervention

Following an overnight (10-hour) fast and having abstained from exercise for 2 days, participants consumed a high-fat meal (63% fat, 34% CHO; 12 kcal/kg body mass; 927 ± 154 kcal). To assess the metabolic response, blood draws were performed at baseline and each hour following the meal for 6 hours.

Measurements

Fasting and postprandial triglycerides (TG), glucose, Total-C, and HDL-C were measured. Metabolic load index (MLI) and LDL-C were calculated.

Results

There were significant group x time interactions for TG (p < 0.0001) and MLI (p = 0.004). The TG total area-under-the-curve (tAUC) response was significantly lower in YA (407.9 ± 115.1 mg/dL 6 hr) compared to OA (625.6 ± 169.0 mg/dL 6 hr; p = 0.02) and OI (961.2 ± 363.6 mg/dL 6 hr; p = 0.0002), while the OA group TG tAUC was lower than the OI group (p = 0.02). The TG peak was significantly lower in YA (90.5 ± 27.0 mg/dL) than OA (144.0 ± 42.2 mg/dL; p = 0.03) and OI (228.2 ± 96.1 mg/dL; p = 0.0003), and was lower in the OA group compared to the OI group (p = 0.03). Glucose was significantly lower 1 hour after the meal in YA (89.4 ± 10.1 mg/dL; p = 0.01) and OA (87.3 ± 22.3 mg/dL; p = 0.005) versus OI (110.7 ± 26.9 mg/dL). MLI tAUC was significantly lower in YA (936.8 ± 137.7 mg/dL 6 hr; p = 0.0007) and OA (1133.0 ± 207.4 mg/dL; p = 0.01) versus OI (1553.8 ± 394.3 mg/dL), with no difference (p = 0.14) between YA and OA groups. Total-C and LDL-C were generally lower in younger compared to older participants at baseline and throughout the postprandial period, while no group or time effects were evident in HDL-C.

Conclusion

Both physical activity status and aging appear to affect the postprandial metabolic, namely TG, response to a high-fat meal. These findings point to an inherently diminished metabolic capacity with aging, but suggest that physical activity may help minimize this decrement.

Introduction

Cardiovascular disease (CVD) represents a serious health risk for many individuals in Western society. There are several lifestyle factors that are broadly accepted as modifiers of a person's CVD risk, but a particular lifestyle feature that has received consistent attention over the last several decades is dietary intake (1). Specifically, there is evidence that consumption of a single high-fat meal (HFM) can increase a person's CVD risk through a variety of mechanisms (2, 3, 4). There is a strong association between the postprandial triglyceride (TG) response and CVD risk, with higher non-fasting TG being related to increased CVD risk (5, 6). The impact of the postprandial TG response (“postprandial lipemia) is important, as individuals spend the majority of their day in a postprandial state (7). Other deleterious postprandial phenomena include evidence of a strong relationship between postprandial hyperglycemia with type 2 diabetes mellitus and CVD development (8, 9, 10), as well as a decrease in high-density lipoprotein cholesterol (HDL-C) following consumption of a HFM (11).

There is a large body of evidence that suggests that exercise can be an effective means for reducing postprandial lipemia (12, 13). Acute exercise performed 10-12 hours before, immediately before, or immediately after consumption of a HFM have all been found to lessen the postprandial TG response (14, 15). However, there are indications that the capacity of physical activity to reduce postprandial lipemia is primarily transitory in nature (16), as active individuals who avoid exercise for several days prior to a HFM may exhibit a post-meal TG response similar to matched inactive individuals (17, 18).

Interestingly, nearly all postprandial lipemia studies, including those investigating the effects of exercise, involve young or middle-aged individuals. While the evidence is scarce, it appears that older individuals tend to exhibit a greater postprandial TG response compared to younger individuals (19). Nevertheless, the physiological explanations for this difference have not been fully elucidated, but could partially be due to decreasing physical activity with increasing age (20). To date, no study of which we are aware has investigated both physical activity and aging as potential factors in altering the postprandial HFM response.

Therefore, the purpose of this investigation was to study the independent effects of aging and physical activity status on the postprandial metabolic response. We assessed postprandial responses to a HFM in three groups: younger active (YA), older active (OA), and older inactive (OI) adults. To our knowledge, no previous study has assessed the postprandial metabolic response in clearly defined cohorts of men and women that differ by both age and physical activity level. We predicted that both age and physical activity level would independently alter the postprandial metabolic response. Specifically, we hypothesized that: 1) Participants in the YA group would exhibit a smaller postprandial TG response as compared to both OA and OI groups, but the OA group would show a smaller response as compared to the OI group; 2) Given the macronutrient composition of the test meal, there would be no differences over time or between groups with regard to postprandial glucose; and 3) Metabolic Load Index (MLI), representative of the total metabolic challenge and calculated by summing TG and glucose (21), in the postprandial period would be smaller in YA compared to the OA and OI groups, but OA would exhibit a lesser postprandial MLI response than the OI group.

The hypotheses of the present study were based on TG and glucose (and subsequently MLI), as these are substrates widely recognized to change acutely following meal consumption (22). Total-C, LDL-C, and HDL-C were secondary metabolic markers in the current study – informative of overall metabolic status, but not viewed as primary outcomes in the postprandial period.

Methods

Participants and physical activity level

Twenty-two participants participated in the present study: eight YA adults (age 18-35 years; 4M/4W), eight OA adults (age 60+ years; 4M/4W), and six OI adults (3M/3W). Active participants were regularly meeting physical activity guidelines (≥150 minutes/week of moderate- to vigorous-intensity physical activity; MVPA) (23). Inactive participants were not regularly engaging in planned exercise (<30 minutes/week) and reported engaging in a generally inactive lifestyle (i.e. not meeting physical activity guidelines). Participants had not changed physical activity habits dramatically in the past five years. OA and OI adults reported to having been generally active or insufficiently active, respectfully, for most of their lives. Because there is not a validated questionnaire to assess lifetime physical activity, whether participants fit the physical activity inclusion criteria was determined via extensive interviewing with an investigator, as has been done previously (24). Current physical activity status was objectively measured using accelerometry (Actical; Respironics; Bend, OR, USA). Accelerometers were worn on the non-dominant wrist for 5-7 continuous days, including at least 1 weekend day, and were initialized to record data in 30-second epochs. Participants were free of any ongoing chronic disease, as confirmed via medical history questionnaire. No participants were taking lipid-lowering medications. This study was approved by the Institutional Review Board at Kansas State University.

Initial assessment

Participants reported to the laboratory on two occasions: an initial assessment and a meal assessment. The initial assessment entailed paperwork and anthropometric testing. Height was measured via portable stadiometer (Invictus Plastics, Leicaster, England) and weight was assessed using a digital scale (Pelsar LLC, Alsip, IL, USA). Height and weight were each measured twice, and a third measurement was performed if the values differed by more than 0.5 cm or 0.5 kg, respectively. The values were then averaged together. Body composition was assessed via a dual-energy X-ray absorptiometry (DEXA) scan (GE Lunar Prodigy, Madison, WI, USA).

Meal Test Protocol

The HFM used in the present study was chocolate pie (Marie Callender's Chocolate Satin Pie; Conagra Brands; Omaha, NE, USA). The primary ingredients of the pie were sugar, water, eggs, enriched wheat flour, soybean oil, palm oil, milk, butter, margarine, high fructose corn syrup, cocoa powder, and milk chocolate. The macronutrient distribution was 63% fat, 34% carbohydrate, and 3% protein. The amount of test meal that each participant consumed was relative to their body mass (12 kcal/kg body mass; 0.84 g/kg fat, 1.02 g/kg carbohydrate, 0.09 g/kg protein). When accounting for participant body mass, the HFM contained 927 ± 154 kcal across all of the participants. Table 1 displays kcal consumed in the test meal by group. The amount of pie consumed was generally similar to a typical serving at a restaurant or social gathering (1-2 servings).

Table 1.

Participant characteristics

	Younger Activen = 8	Older Activen = 8	Older Inactiven = 6	P-value
Age (years)	25.1 ± 4.8a	66.5 ± 5.2b	68.2 ± 7.4b	<0.0001
Height (cm)	174.2 ± 10.8	175.0 ± 8.8	168.3 ± 8.5	0.40
Body mass (kg)	71.9 ± 10.4	81.6 ± 15.4	78.6 ± 11.3	0.32
Body mass index (kg/m2)	23.6 ± 2.0	26.6 ± 4.1	27.9 ± 4.7	0.11
Body fat (%)	21.5 ± 8.9	30.0 ± 11.7	33.4 ± 8.9	0.09
Trunk fat (%)	22.5 ± 9.2a	32.6 ± 11.5ab	37.7 ± 5.9b	0.02
Steps (x103)/day	16.6 ± 6.5a	14.9 ± 5.5ab	8.8 ± 2.8b	0.03
MVPA (minutes/day)	159.4 ± 48.6a	182.4 ± 77.1a	62.3 ± 14.1b	0.003
Test meal energy (kcal)	863 ± 125	980 ± 185	943 ± 136	0.32

Data are Mean ± SD. The P-value column denotes main effects between groups assessed via one-way ANOVA. Within main effects (by row), column values with shared superscript letters are not significantly different, determined by post hoc pairwise comparisons. Rows with no superscript letters present contain no significant differences. See Results section for post hoc pairwise comparison p-values; n, number of participants; MVPA, moderate- to vigorous-intensity physical activity

Participants were instructed to avoid planned exercise for two full days before their main assessment. Participants were given a 270-kcal snack (Little Debbie Swiss Cake Roll; McKee Foods; Collegedale, TN, USA) that they were instructed to consume in the evening, ten hours prior to their appointment.

On the morning of the meal assessment, participants arrived to the laboratory after a 10-hour overnight fast. An indwelling safelet catheter was inserted into a forearm vein via 24-gauge needle (Exelint International, Redondo Beach, CA, USA). The catheter was kept clear with a consistent infusion of 0.9% NaCl solution (~1 drip/second) and maintained stationary via placement of tegaderm film (3M Healthcare, Neuss, Germany). When the catheter was set in place, a fasting blood draw was conducted. For each blood draw, a 3 mL syringe (BD, Franklin Lakes, NJ, USA) was used to remove saline from the line, after which the actual blood sample was drawn into a 5 mL syringe (BD, Franklin Lakes, NJ, USA). The 5 mL syringe was emptied into a 6 mL Vacutainer test tube (BD, Franklin Lakes, NJ, USA) coated with EDTA (anticoagulant) and inverted three times to ensure adequate mixing with the EDTA. Whole blood from the blood draws was utilized to measure TG, glucose, total cholesterol (Total-C), and high-density lipoprotein cholesterol (HDL-C) using a Cholestech LDX analyzer (Alere Cholestech, San Diego, CA, USA). Low-density lipoprotein cholesterol (LDL-C) was calculated by the LDX analyzer using the Friedewald equation (25). For each blood sample, a few drops of whole blood were drawn into a capillary tube and plunged into a Cholestech LDX Lipid+Glu cassette (Alere Cholestech, San Diego, CA, USA). The cassette was then placed in the Cholestech LDX analyzer for measurement. After the baseline blood draw, participants consumed the test meal within 20 minutes. Water was available ad libitum with the meal and during the post-meal period. Participants stayed in the laboratory for six hours after consumption of the test meal (the six-hour time period began after the last bite of the test meal). Blood draws were conducted serially each hour for six hours post-HFM.

Statistical Analyses

An a priori sample size calculation based on the findings of previous studies (26, 27) revealed that 3-5 participants would need to be recruited to each group to detect statistically significant differences in the postprandial TG response (power = 0.80; α = 0.05). We aimed to recruit 8 participants to each group (24 participants total) in order to increase power to detect differences in other metabolic markers. However, due to the considerable challenge of recruiting inactive individuals over age 60 who were not taking lipid-lowering medications, only 6 participants recruited to the OA group met our a priori inclusion criteria.

MLI is calculated by adding TG and glucose values. As postprandial TG and glucose responses have been shown to be interrelated (28, 29), MLI is intended to represent the cumulative metabolic challenge faced by the body, either fasting or following a meal (21).

Total area under the curve (tAUC), incremental area under the curve (iAUC), peak value, and time to peak value were determined for each of the metabolic markers. tAUC and iAUC were calculated using the trapezoid method. All data were assessed for normality via Shapiro-Wilk formal normality test and analysis of frequency distribution. If data were not normally distributed, a square root transformation was performed. Differences between groups with regard to participant characteristics and postprandial values were tested via one-way Analysis of Variance (ANOVA) with Holm-Sidak adjustment for multiple comparisons. However, since iAUC analyses can potentially produce negative values, a square root transformation was not performed. Instead, a non-parametric Friedman test was utilized to test for group differences if iAUC data were not normally distributed.

Time-course changes in metabolic markers in the postprandial period were determined via two-way (group x time) repeated measures ANOVA with a Tukey's adjustment for multiple comparisons. A type 1 error rate of 0.05 was used in all analyses for the determination of statistically significant differences. Statistical analyses were conducted using GraphPad Prism statistical software (Version 6.05; GraphPad Software, Inc; La Jolla, CA).

Results

Participant characteristics

Participant characteristics are displayed in Table 1. The YA group was younger than the OA (p < 0.0001) and the OI (p < 0.0001) groups. The OA and OI groups did not differ in age (p = 0.60). With regard to anthropometric variables, there were no differences between groups (p > 0.05) with respect to height, body mass, body mass index (BMI), or percentage body fat. However, the OI group had significantly more trunk fat compared to the YA group (p = 0.02), although there were no differences between YA and OA (p = 0.09) or OA and OI (p = 0.33).

Table 2.

Metabolic values measured in fasting participants

Fasting Values (mg/dL)	Optimal Values (mg/dL)	Younger Active n = 8	Older Active n = 8	Older Inactive n = 6	P-value
TG	< 150	47.4 ± 4.6a	52.3 ± 9.0ab	75.8 ± 35.3b	0.03
Glucose	< 100	89.6 ± 11.2	90.6 ± 8.3	101.3 ± 9.7	0.08
MLI	N/A	137.0 ± 11.5a	142.9 ± 15.0a	177.2 ± 39.5b	0.01
Total-C	< 200	137.5 ± 24.5	165.4 ± 18.1	165.8 ± 32.1	0.06
LDL-C	100-129	76.3 ± 18.9a	103.9 ± 17.1b	101.3 ± 19.9ab	0.03
HDL-C	40-60	52.0 ± 10.9	50.8 ± 6.0	48.3 ± 17.4	0.84

Data are Mean ± SD. The P-value column denotes main effects between groups assessed via one-way ANOVA. Within main effects (by row), column values with shared superscript letters are not significantly different, determined by post hoc pairwise comparisons. Rows with no superscript letters present contain no significant differences. Optimal values are based on references 48 and 50. n, number of participants; TG, triglycerides; MLI, metabolic load index; Total-C, total cholesterol; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol

Physical activity level

There was no difference in steps/day in the YA group versus the OA group (p = 0.56) or the OA group compared to the OI group (p = 0.06). However, the YA group did obtain significantly more steps than the OI group (p = 0.03). With regard to minutes per day spent in MVPA, the YA and OA groups both engaged in more MVPA compared to the OI group (p = 0.01; p = 0.003, respectively), while YA and OA did not differ (p = 0.44).

Fasting values

There was no difference in fasting TG in the YA group compared to the OA group (p = 0.55) or the OA group compared to the OI group (p = 0.07). Fasting TG was significantly lower in the YA group versus the OI group (p = 0.03). No participants presented with fasting TG >150 mg/dL and only one participant (belonging to the OI group) had fasting TG >100 mg/dL (136 mg/dL). Thus, all participants presented with optimal fasting TG (Goodman et al. 1988). There were no group differences (p > 0.05) with regard to fasting glucose. However, fasting MLI was higher in the OI adults compared to both the YA (p = 0.01) and OA adults (p = 0.03), with no difference between the YA and OA (p = 0.62). Fasting Total-C was not different between groups (p > 0.05). Only one participant (an OI adult) presented with fasting total cholesterol >200 mg/dL (209 mg/dL). Fasting Total-C was in the optimal range for all other participants (Goodman et al. 1988). There were group differences with regard to fasting LDL-C, as OA had higher LDL-C levels compared to YA (p = 0.04). However, the OI group was not different from the YA group (p = 0.06) or the OA group (p = 0.80). No participants presented with fasting LDL-C greater than 129 mg/dL, the upper limit for optimal fasting LDL-C (Goodman et al. 1988). There were no group differences with regard to fasting HDL-C (p > 0.05). Three participants (1 YA, 2 OI) had below-optimal HDL-C (40-60 mg/dL) (Goodman et al. 1988).

Postprandial metabolic responses

Postprandial metabolic values and time-course responses are presented in Table 3 and Figures 1 and 2, respectively. In a two-way repeated measures ANOVA, a significant group x time interaction was detected in the postprandial TG response (p < 0.0001). TG peaked at a significantly lower level in the YA group compared to the OA (p = 0.03) and OI (p = 0.0003) groups, while the OA group peaked at a significantly lower level than the OI group (p = 0.03). TG tAUC (representative of the total magnitude and duration of the postprandial response for the 6-hour period) was lower in YA adults versus OA adults (p = 0.02) and OI adults (p = 0.0002), whereas TG tAUC was lower in OA adults compared to OI adults (p = 0.02). With regard to TG iAUC (representative of the magnitude and duration of the postprandial response above fasting/baseline), YA had a lower response than OA (p = 0.006) and OI (p = 0.0002), with no difference (p = 0.06) in OA and OI groups.

Table 3.

Postprandial metabolic outcomes

	Younger Active n=8	Older Active n=8	Older Inactive n=6	P-value
Triglycerides
Peak (mg/dL)	82.0 (75.5-97.5)a	133.0 (120.3-160.0)b	182.0 (155.3-341.5)c	0.0004
Time to peak (hours)	3.5 (2.0-4.0)	4.0 (3.0-4.8)	4.5 (3.0-5.0)	0.18
tAUC (mg/dL 6 hr)	372.8 (334.0-436.1)a	570.0 (486.5-712.3)b	794.8 (677.6-1367.0)c	0.0003
iAUC (mg/dL 6 hr)	78.3 (59.3-154.6)a	288.0 (216.5-358.3)b	453.0 (307.3-687.8)b	0.0002
Glucose
Peak (mg/dL)	99.9 ± 7.8	99.0 ± 16.4	118.2 ± 17.2	0.04
Time to peak (hours)	1.5 (0.0-3.5)	1.0 (0.0-3.8)	1.0 (0.8-1.3)	0.96
tAUC (mg/dL 6 hr)	528.9 ± 43.2	507.3 ± 82.2	592.5 ± 59.9	0.07
iAUC (mg/dL 6 hr)	-8.9 ± 58.5	-36.4 ± 49.5	-15.5 ± 44.9	0.56
Metabolic Load Index
Peak (mg/dL)	174.0 (165.3-193.5)a	222.0 (207.0-245.8)a	275.0 (256.3-445.3)b	0.0009
Time to peak (hours)	3.4 ± 1.1	3.6 ± 0.9	4.2 ± 1.0	0.35
tAUC (mg/dL 6 hr)	936.8 ± 137.7a	1133.0 ± 207.4a	1553.8 ± 394.3b	0.0009
iAUC (mg/dL 6 hr)	114.8 ± 131.2a	275.6 ± 135.1ab	490.7 ± 255.7b	0.003
Total Cholesterol
Peak (mg/dL)	144.1 ± 24.2a	180.9 ± 13.8b	176.8 ± 26.2b	0.006
Time to peak (hours)	2.1 ± 2.0	3.1 ± 1.4	2.8 ± 2.3	0.56
tAUC (mg/dL 6 hr)	803.8 ± 123.8a	1023.5 ± 84.0b	993.7 ± 149.9b	0.003
iAUC (mg/dL 6 hr)	-21.3 ± 57.6	31.1 ± 53.9	-1.4 ± 54.1	0.19
LDL-Cholesterol
Peak (mg/dL)	77.5 (65.5-90.5)a	104.5 (95.8-123.5)b	106.0 (87.0-122.0)b	0.01
Time to peak (hours)	1.0 (0.0-6.0)	1.0 (0.0-3.5)	0.0 (0.0-1.5)	0.52
tAUC (mg/dL 6 hr)	427.9 ± 88.4a	579.4 ± 80.2b	523.8 ± 121.9ab	0.03
iAUC (mg/dL 6 hr)	-30.1 ± 39.5	-43.8 ± 41.1	-84.2 ± 64.7	0.16
HDL-Cholesterol
Peak (mg/dL)	57.5 ± 11.9	56.4 ± 9.0	51.0 ± 17.7	0.62
Time to peak (hours)	3.4 ± 2.1	2.6 ± 2.3	1.8 ± 2.3	0.45
tAUC (mg/dL 6 hr)	305.8 ± 67.4	317.2 ± 47.1	276.9 ± 106.1	0.60
iAUC (mg/dL 6 hr)	-6.2 ± 17.4	12.7 ± 27.0	-13.1 ± 19.9	0.10

Normally distributed data are Mean ± SD, and non-normally distributed data are Median (Interquartile Range). Non-normally distributed data were transformed before analysis. A one-way ANOVA was conducted to test for differences between groups. However, due to negative numbers, a non-parametric Friedman test was conducted for TG iAUC. The P-value column denotes main effects between groups assessed via one-way ANOVA. Within main effects (by row), column values with shared superscript letters are not significantly different, determined by post hoc pairwise comparisons. Rows with no superscript letters present contain no significant differences. See Results section for post hoc pairwise comparison p-values. tAUC, total area under the curve; iAUC, incremental area under the curve; LDL, low-density lipoprotein, HDL, high-density lipoprotein; ANOVA, analysis of variance

Figure 1.

Postprandial responses for triglycerides, glucose, and metabolic load index

Figure 2.

Postprandial cholesterol responses

For the postprandial glucose response, there was no significant group x time interaction following the HFM (p = 0.24), although there was a significant time effect (p = 0.002). There was a significant group effect (p = 0.04) for peak glucose values. However, no group pairwise comparisons were statistically significant (p > 0.05). With the exception of a few time-point specific differences (Figure 1), there were no other significant differences between groups in the postprandial glucose response.

The repeated measures two-way ANOVA revealed a significant group x time interaction in the MLI response to the HFM (p = 0.004). The OI group exhibited a greater MLI peak compared to the YA (p = 0.0007) and OA (p = 0.02) groups, while there was no difference between YA and OA (p = 0.09). With regard to tAUC, there was no difference between the YA and OA groups (Mean diff: -196.2 mg/dL x 6 hr; 95% CI: (-516.9, 124.5); p = 0.14), but the OI group exhibited a greater MLI response compared to YA (Mean diff: 617.0 mg/dL x 6 hr; 95% CI: (270.6, 963.4); p = 0.0007) and OA (Mean diff: 420.8 mg/dL x 6 hr; 95% CI: (74.4, 767.2); p = 0.01). The YA group also displayed a significantly lower (Mean diff: -375.9 mg/dL x 6 hr; 95% CI: (-614.7, -137.2); p = 0.002) MLI iAUC response compared to the OI group, with no difference between YA and OA (Mean diff: -160.9 mg/dL x 6 hr; 95% CI: (-381.9, 60.2); p = 0.08) or OA and OI (Mean diff: -215.0 mg/dL x 6 hr; 95% CI: (-453.8, 23.7); p = 0.07).

Postprandial Total-C levels in each group are presented in Figure 2. No significant group x time interaction was observed in Total-C (p = 0.49), although there was a significant group effect (p = 0.004). Peak Total-C levels were lower in the YA group compared to the OA (Mean diff: -36.8 mg/dL; 95% CI: (-64.2, -9.3); p = 0.009) and OI (Mean diff: -32.7 mg/dL; 95% CI: (-62.4, -3.1); p = 0.02) groups, with no difference between OA and OI (Mean diff: 4.04 mg/dL; 95% CI: (-25.6, 33.7); p = 0.73). Total-C tAUC was significantly lower in YA compared to OA (Mean diff: -219.8 mg/dL x 6 hr; 95% CI: (-370.9, -68.6); p = 0.005) and OI (Mean diff: -189.9 mg/dL x 6 hr; 95% CI: (-353.2, -26.7); p = 0.02), but there was no difference between OA and OI (Mean diff: 29.8 mg/dL x 6 hr; 95% CI: (-133.4, 193.1); p = 0.65).

With regard to LDL-C, in a two-way repeated measures ANOVA, although there was no significant group x time interaction (p = 0.46), significant time (p < 0.0001) and group (p = 0.03) effects were detected. YA adults exhibited a significantly lower LDL-C peak compared to OA (p = 0.02) and OI (p = 0.04) adults, with no difference between OA and OI adults (p = 0.68). The LDL-C tAUC response was lower (Mean diff: -151.5 mg/dL x 6 hr; 95% CI: (-285.3, -17.8); p = 0.03) in the YA group versus the OA group, but there was no difference in YA versus OI (Mean diff: -95.9 mg/dL x 6 hr; 95% CI: (-238.9, 47.1); p = 0.20) or OA versus OI (Mean diff: 55.6 mg/dL x 6 hr; 95% CI: (-78.1, 189.4); p = 0.30).

There was no significant group x time interaction (p = 0.47) with regard to HDL-C. There were also no statistical group differences (p > 0.05) in peak, time to peak, tAUC, or iAUC.

Discussion

Main Findings

The main finding of the present study was the distinct difference in the postprandial TG response between the three groups. Supporting our first hypothesis, the YA group exhibited an attenuated postprandial lipemic response relative to OA and OI groups, and OA displayed a tempered response compared to OI. These findings suggest that both age and physical activity status independently impact the postprandial TG response. With regard to our second hypothesis, we predicted no group- or time-based differences in the postprandial glycemic response. Interestingly, we found that glucose was significantly higher in the OI group compared to the YA and OA groups one hour after the HFM, although there was not an overall group x time interaction. Finally, in partial agreement with our third hypothesis, the YA and OA groups displayed smaller postprandial MLI responses compared to the OI group, with no differences between YA and OA groups.

Postprandial TG Responses

While there is a well-established connection between postprandial lipemia and CVD risk (30), most studies investigating postprandial lipemia have utilized young and middle-aged individuals, as opposed to older adults (31, 32). This omission is problematic, considering that: 1) older individuals are at a higher risk for CVD (33), and 2) older adults, as a segment of the population, are becoming more numerous in Western society (34). The current study found an age-related increase in postprandial lipemia: YA adults showed a comprehensively lower postprandial TG response compared to OA and OI adults. The finding of an increase in the postprandial lipemic response with age is in agreement with previous studies (26, 27, 35, 36, 37, 38). However, the reason(s) for the greater postprandial TG response in older individuals remains unclear, with several potential mechanisms in consideration (19). First, there is evidence that lipoprotein lipase (LPL) activity decreases with age (39). As LPL is the rate-limiting step for the clearance of TG in circulation (19), the increase in postprandial lipemia with age could be partly due to a decrease in LPL activity. Next, age-related changes in liver physiology result in impairment of uptake and metabolism of chylomicron remnants, including the TG portion (40). Related to this, liver fat content is typically higher in older individuals (41), and there is evidence that increased liver fat content is accompanied by increased circulating lipid concentrations, including TG (42). Therefore, there are several potential physiological explanations for the greater postprandial TG response displayed by older adults.

To our knowledge, only one previous study has assessed postprandial lipemia in older adults, while also considering chronic physical activity level (27). Miyashita and colleagues (27) tested the postprandial TG response in 26 older adults (mean age: ~70 years), divided into active and inactive groups based on whether they obtained less or more than 150 minutes per week of MVPA. The authors found a significantly lower postprandial TG response in the active older adults, despite requiring that all participants avoid physical activity for 48 hours prior to their assessment. There are several noteworthy differences in study design between the present investigation and that of Miyashita et al. (27), including that the present study matched groups by sex, obtained metabolic values every hour post-meal (instead of every other hour), and utilized a fattier test meal (63% vs 35% kcal from fat). Despite these differences, both investigations found a significantly lower postprandial lipemic response in OA than OI groups. These data disagree with the previously accepted notion that the lipid-lowering effects of physical activity were limited to the acute time-frame, as OA adults appear to experience a lesser post-meal TG response than their inactive counterparts, even when required to abstain from exercise for two days prior to the meal. Therefore, there are potentially chronic physical activity effects on the postprandial TG response in older adults that may not be present in younger individuals.

Finally, the most notable difference between the present study and that of Miyashita and colleagues (27) is that our study included a YA group. In doing so, the present study was designed to assess whether both aging and physical activity impacted the postprandial TG response. Results from the current investigation indicated that there was a lower lipemic response in OA compared to OI, while YA exhibited the lowest lipemic response of the three groups. As there were no differences between the YA and OA groups with regard to physical activity level, it appears that both physical activity status and aging may have independent effects on the postprandial lipemic response.

Other Metabolic Markers

For glucose, there were no group differences with regard to the primary postprandial indices. This is not surprising, given the quicker postprandial response of glucose relative to TG, so that group differences would be difficult to detect over a six-hour period. Notably, we did find that glucose was significantly higher in the OI group compared to the YA and OA groups one hour after the meal. This finding was unexpected, since the HFM only included 34% of kcals as carbohydrate. While the present study was not designed or intended to robustly assess glucose tolerance, our findings point to potentially greater capacity for glucose clearance in OA adults compared to OI adults.

In accordance with this, we also utilized the MLI response to the HFM in assessing postprandial metabolism in the three groups. Noteworthy group differences were exhibited with regard to both TG and glucose, but these differences generally occurred at different phases of the postprandial period. Specifically, there were group differences in glucose (but not TG) early in the postprandial period, while group differences were evident in TG (but not glucose) later in the postprandial period. However, consideration of the postprandial MLI response reveals significant group differences throughout the postprandial period. Thus, the metabolic differences between active and inactive adults in the present study were not merely an issue of lipid clearance or glucose uptake, but an overall difference in metabolic capacity to clear the mixed meal, as represented by the group differences in MLI throughout the postprandial period. (However, it should be noted that the mixed meal in the present study was 63% fat, 34% carbohydrate. Thus, the MLI response primarily represents the postprandial lipemic response.) As both elevated postprandial glycemia and lipemia have been independently linked to CVD risk (21, 22), this finding of an overall diminished metabolic capacity in OI adults is noteworthy.

Previous research suggests that Total-C is elevated in older individuals compared to younger individuals (43), and that it is not very responsive in the acute postprandial period (44). These notions were supported by the present findings, as Total-C was lower at baseline and throughout the postprandial period in the younger adults compared to the older adults, and there was no main effect of time for Total-C. Additionally, despite previous evidence suggesting lower Total-C in regularly active individuals (45), the lack of difference between the OA and OI adults is not surprising, since we excluded individuals who were taking lipid-lowering medication.

In the present study, although it was generally lower in the younger adults, LDL-C did not dramatically change in the postprandial period in any of the groups. On the other hand, it has previously been demonstrated that HDL-C has a tendency to decrease following consumption of a HFM, another potentially deleterious feature of the postprandial metabolic response (11). Accordingly, we hypothesized that we would see a decrease in HDL-C following the HFM, with the OI adults displaying the greatest decrease and that YA adults showing the most negligible decrease. Contrary to expectation, we found no changes in the postprandial period, or differences among groups, with regard to HDL-C. The explanation for this null finding remains unclear.

Strengths and Experimental Considerations

There are several strengths of the present investigation that deserve to be highlighted. First, a strong point of the present study is its realistic test meal. Prior postprandial metabolic studies have tended to utilize test meals that are not very “true-to-life (e.g. 1500 kcal) (46). In the present investigation, the average kcal value for the test meal was ~930 kcal. Thus, while the meal was not small, it was nevertheless considerably smaller than many previous studies (46), and yet, in our view, still representative of an unhealthy Western meal. Finally, an asset of this study was the blood sample frequency and the duration of postprandial period. It is ideal to collect frequent blood samples during the postprandial period to optimally characterize the response curve. With regard to duration of assessment, measuring the postprandial TG response for four hours is typically sufficient (47). However, as older adults tend to have a delayed TG response relative to younger adults, it is imperative that postprandial metabolic tests in older adults extend beyond four hours (19). In utilizing a six-hour postprandial period, with hourly blood samples, we view the present study to be well-suited to characterize the postprandial metabolic response.

However, the present study is not without limitations, and considerations need to be made in interpreting the findings. First, the present study did not include a younger inactive group. The primary research purpose was to determine whether there were independent effects of aging and chronic physical activity on the postprandial response. We did not view including a younger inactive group as pertinent to accomplishing this purpose. Specifically, we were most interested in comparing YA adults to OA adults, and OA adults to OI adults. Further, the comparison between younger active and inactive individuals has been tested numerous times in the past (16). Consequently, we directed our resources toward obtaining sufficient data to adequately answer our research question with the essential YA, OA, and OI groups. However, relevant to our research question, had we included a younger inactive group, we could have compared the effects of physical activity in younger participants to the activity effects in older participants. Thus, we acknowledge not including a younger inactive group as a limitation. Another consideration is the limited degree to which we can connect the postprandial responses in the present study to CVD risk. While there is a fasting TG recommendation of <150 mg/dL (48), and on average the OI group exceeded this cut point for most of the postprandial period, there is no established post-meal TG cut point for increased CVD risk. However, an expert panel statement has suggested non-fasting TG ≥180 mg/dL is “undesirable (49). More research is needed to better determine TG reference ranges that represent increased risk in the acute postprandial period. Similarly, MLI is a novel, recently proposed marker of metabolic status. As such, the strength of MLI as an independent CVD risk factor is yet to be determined, and no cut points exist for fasting and postprandial MLI. Therefore, at this point, the higher MLI response in OI participants cannot be directly linked to increased CVD risk. Next, it was crucial that individuals on lipid-lowering medications be excluded from the present study. However, not only does this requirement make recruiting more difficult, but it potentially limits external applicability for the OI group. There were simply not very many OI individuals who were not on lipid-lowering medications in our target population. However, it could be argued that OI individuals on lipid-lowering medications could, theoretically, display a similar response as the OI adults in the present study who were not on lipid-lowering medications. This is an interesting issue, and should be investigated further. In addition, given the important roles of insulin in regulating substrate utilization, it would have been valuable to assess fasting and postprandial insulin.

Conclusions and Future Directions

We found that both aging and physical activity level likely impact the postprandial TG response, as YA exhibited a lesser response than OA and OI groups, but OA adults showed a reduced response compared to OI. OI adults also displayed higher glucose in the 1-2 hours post-meal, pointing to an overall diminished capacity for metabolic clearance, as supported by the MLI findings. Overall, these findings point to the possibility of an age-related decline in metabolic clearance capacity, but this deterioration can be partially alleviated with chronic physical activity – an important public health message. Future research should continue to study the relationship between physical activity, postprandial metabolism, and CVD risk in aging individuals, as older adults represent a growing segment of the population that is at increased risk for CVD development.

Conflicts of Interest

We declare that there are no conflicts of interest.