Previous research indicates that high caloric meals [1
] ingested late at night contribute to weight gain and impaired cardiometabolic health [4
]. However, recent data on the health impact of nighttime eating suggests that our understanding of this practice may need modification based on the criteria used to identify late night eating (e.g., majority of daily intake solely at night, consumed during a specified time frame at night, includes dinner, or minutes before bedtime), as well as the content (percent carbohydrate vs. fat vs. protein) and quality (glycemic index) of calories consumed during this period [6
Recent work [7
] suggests that low-calorie (~150 kcals), protein-rich sources may be beneficial or non-detrimental to metabolism, health, and body composition when consumed as a snack ~30 min before going to sleep acutely (one-night) [7
] and when combined with exercise for a longer duration (i.e., 4–12 weeks) [10
]. In general, protein ingestion suppresses appetite, increases energy expenditure and muscle mass, and decreases body fat [14
], encouraging support for ingesting protein-rich foods before going to sleep. Casein protein (CAS) has been designated as an optimal bedtime snack because it is slow to digest and augments muscle protein synthesis [19
] making it ideal for pre-sleep ingestion. Indeed, data show that acute CAS intake (30–40 g per serving), compared to a non-nutritive placebo, is properly digested/absorbed and increases overnight muscle protein synthesis when provided before sleep [11
]. We have previously demonstrated that casein before bed elevates next morning resting energy expenditure (REE) more than a non-nutritive placebo, but does not inhibit fat oxidation, as observed with other proteins and macronutrients relative to placebo, in normal-weight men [9
]. This may suggest that casein before bed may not negatively impair fat metabolism relative to fasting and therefore may be appropriate to consume in close proximity to sleep. Studies in obese women have demonstrated that CAS ingestion before sleep can increase next morning satiety and, augment REE from baseline, however no differences were detected between casein and calorically-matched whey protein or carbohydrate ingestion [8
]. Additionally, regardless of macronutrient type, acute elevations in insulin and insulin resistance have been observed the morning after pre-sleep caloric intake in obese women [8
]. While the latter finding is not ideal for obese populations, it is important to note that a true non-nutritive placebo was absent and food intake throughout the day was not well controlled [8
]. It is possible that the habitual diet during the day could have confounded the effects of nighttime macronutrient ingestion on next morning insulin and, hence, deserves further attention.
The potential mechanisms for improved metabolism and body composition have not been fully explored. To our knowledge, no studies have investigated mechanisms by which pre-sleep protein ingestion alters body fat (i.e., lipolysis) during sleep. Elucidating these mechanisms may be clinically relevant especially for obese individuals as eating the wrong foods at the wrong time (i.e., late at night before sleep) may contribute to increases in body fat and obesity [3
]. Upper body subcutaneous fat depots are the predominate suppliers of circulating free fatty acids during the overnight postabsorptive period [22
] and investigating changes in subcutaneous abdominal adipose tissue (SCAAT) lipolysis with microdialysis during sleep provides a novel opportunity to explore changes in regional fat metabolism that may ensue as a result of bedtime protein ingestion. Therefore, the purpose of this study was to examine the effect of CAS consumption before sleep on SCAAT lipolysis, whole-body substrate utilization, growth hormone, insulin, glucose, and REE compared to consumption of a non-nutritive placebo (PLA) in obese men. Based on our earlier work showing no differences in morning fat oxidation following bedtime CAS as compared to PLA ingestion in normal weight men [9
], we hypothesized that CAS and PLA would yield similar changes in SCAAT lipolysis, whole-body substrate utilization (favoring fat oxidation), growth hormone, insulin and glucose in obese men. We also hypothesized that CAS prior to sleep will increase next morning REE more than PLA, as previously reported [8
]. By measuring SCAAT lipolysis with microdialysis and whole-body fat oxidation, we examined the mechanisms by which CAS consumption before sleep may mediate changes in metabolism and body composition. In addition, we also decided to explore the effect of CAS before sleep on subjective feelings of appetite and explore potential mechanisms for these changes, if any, by measuring ghrelin, an appetite hormone. Others have shown that nutrient ingestion prior to sleep may suppress next morning appetite [8
] but no studies have explored mediating mechanisms. Likewise, whether this holds true in obese men is unknown. It was hypothesized that CAS will decrease appetite to a significantly greater extent than PLA while its effect on ghrelin is exploratory.
3.1. Descriptive Characteristics
During the recruitment period, 103 individuals expressed interest in study participation. Of these individuals, 32 returned their screening document for further review and attended scheduled information sessions. Following the information sessions, 12 individuals expressed interest in enrolling in the study. Descriptive data obtained at baseline are presented in Table 1
and dietary characteristics are in Table 2
. One participant did not return his dietary log. Energy intake from the meal-replacement beverages was also included in this table and the difference in caloric intake (habitual (dietary log) vs. controlled (estimated from REE) was 306 kcals with the latter being higher (p
= 0.16). However, carbohydrate intake was significantly higher during the controlled diet (p
= 0.00). More importantly, the protein intake (g/kg/day) was not significantly different between the habitual and controlled diets (p
There were no significant differences in subjective measures of hunger (baseline: 35 ± 6, CAS: 49 ± 6, PLA: 47 ± 6 mm, p = 0.21) or satiety (baseline: 36 ± 6, CAS: 36 ± 4, PLA: 39 ± 5 mm, p = 0.89). A significant group effect for desire to eat was observed (p = 0.03) and post hoc analysis revealed differences between baseline and CAS favoring a greater desire to eat with CAS (baseline: 39 ± 6, CAS: 62 ± 8, PLA: 55 ± 5 mm, p = 0.03).
There were no significant differences in REE (baseline: 2150 ± 119, CAS: 2126 ± 111, PLA: 2145 ± 106 kcals/day, p = 0.99) or RER (baseline: 0.75 ± 0.02, CAS: 0.76 ± 0.01, PLA: 0.76 ± 0.01, p = 0.77). Similarly, neither estimated carbohydrate oxidation (baseline: 0.06 ± 0.03, CAS: 0.08 ± 0.02, PLA: 0.08 ± 0.02 g/min, p = 0.76) nor estimated fat oxidation (baseline: 0.13 ± 0.01, CAS: 0.13 ± 0.01, PLA: 0.13 ± 0.01 g/min, p = 0.79) were different at any collection time point.
3.4. Blood Markers
Blood markers data are presented in Table 3
. Glucose measurements during the PLA trial were not available for one participant, so his data were excluded from all glucose and HOMA-IR analyses. There were no differences in glucose, insulin, HOMA-IR, or ghrelin among baseline, CAS and PLA. Participants met the criteria for hyperinsulinemia (fasting insulin > 180 pmol/L) [43
] and insulin resistance (HOMA-IR value > 2.5) [31
]. Growth hormone data were missing for four participants at various time points due to technical issues (i.e., sample concentrations were unable to be determined for some data points), therefore, only participants with data at all three time points were included in the analysis (n
= 8). There were no differences in growth hormone at any time point.
3.5. Sleep Quantity and Quality
Sleep data at baseline and during the trial periods were only available from a subset of participants (n = 5) due to availability of the actigraph watches. Average minutes resting (lying down but not sleeping) were 408.2 ± 26.9 (~6 h, 48 min), 371.0 ± 19.1 (6 h, 12 min), and 311.2 ± 42.1 (5 h, 12 min) (p = 0.12) at baseline and during the CAS and PLA, respectively. Average minutes sleeping were 288.0 ± 33.1 (~4 h, 48 min), 261.0 ± 27.4 (4 h, 21 min) and 235.0 ± 42.0 (3 h, 55 min) (p = 0.57), for baseline, CAS, and PLA, respectively. Average minutes of sleep latency (the time it takes to fall asleep while resting) were 50.0 ± 22.6, 64.2 ± 25.0, and 17.6 ± 5.2 (p = 0.27), for baseline, CAS, and PLA, respectively. The quality of sleep was assessed by the sleep efficiency scores which were 69.9% ± 4.3%, 71.1% ± 8.1% and 74.7% ± 5.6% (p = 0.85), for baseline, CAS, and PLA, respectively.
3.6. SCAAT Interstitial Glycerol and Glucose
During the PLA trial, one participant’s morning dialysate collection vial did not have a dialysate sample (from either probe) so this participant’s interstitial data were not included in data analysis as the calculation for interstitial concentrations required both the overnight and next morning PLA samples. The in vivo probe recoveries for PLA were 88% at 0.3 μL/min and 46% at 2.0 μL/min for glycerol and 86% at 0.3 μL/min and 38% at 2.0 μL/min for glucose. The second probe served as a backup in the event that a sample was not collected in the first probe. Interstitial concentrations for one participant’s overnight samples were obtained from the backup probe due to technical difficulties during sleep. Interstitial concentrations were not statistically different between PLA probes (overnight: glycerol, p
= 0.49; glucose, p
= 0.81; next morning: glycerol, p
= 0.30; glucose, p
= 0.51) and, therefore, the use of data from the backup probe did not affect the results. Overnight (CAS, 177.4 ± 26.7; PLA, 183.8 ± 20.2 μmol/L; p
= 0.83) and next morning (CAS, 171.6 ± 19.1; PLA, 161.5 ± 18.6 μmol/L, p
= 0.44) SCAAT interstitial glycerol concentrations were not significantly different between CAS and PLA (Figure 2
). Similarly, there were no differences in overnight glucose (CAS, 2.9 ± 0.3; PLA, 3.2 ± 0.2 μmol/L, p
= 0.16) or next morning glucose (CAS, 3.0 ± 0.3; PLA, 3.1 ± 0.12 μmol/L, p
= 0.39) between trials (Figure 3
3.7. Blood Flow
Blood flow surrounding the probe is inversely related to the ethanol outflow/inflow ratio. Overnight (CAS, 0.13 ± 0.07; PLA, 0.17 ± 0.07 μmol/L at 0.3μL/min perfusate flow rate; p = 0.72) and next morning (CAS, 0.73 ± 0.07; PLA, 0.72 ± 0.07 μmol/L at 2.0 μL/min perfusate flow rate; p = 0.96) outflow/inflow ratios were not significantly different between CAS and PLA.
The present study is the first to examine the effect of acute pre-sleep CAS consumption on markers of fat metabolism (lipolysis, substrate utilization, and growth hormone), REE, insulin, glucose, and appetite (subjective and ghrelin) compared to PLA in obese men. We used microdialysis which, for the first time, allowed us to investigate the mechanisms that may alter metabolism and body composition with pre-sleep CAS ingestion. The primary findings were that SCAAT lipolysis, fat oxidation, growth hormone, insulin and glucose were not significantly different between CAS and PLA (hypotheses supported). In addition, CAS did not result in greater REE or suppression of appetite and ghrelin compared to PLA (hypothesis rejected).
Previous acute studies in lean men [9
] and obese women [8
] have examined next morning substrate utilization and REE as a result of pre-sleep macronutrient intake the night prior. In lean men, bedtime ingestion of calorically-matched macronutrients (CAS, whey protein, carbohydrates) raised REE to a greater extent than the non-nutritive PLA, and, importantly, fat oxidation was not different between CAS and PLA [9
]. In obese women, isocaloric CAS, whey protein, and carbohydrate elevated REE, although not significantly and a true non-nutritive PLA was not used thereby limiting the interpretation of these data [8
]. Given that REE is the largest component of daily energy expenditure (60%–75%), and that reduced energy expenditure has been linked to weight gain [44
], small increments in metabolism through pre-sleep protein intake may be clinically relevant for obese populations. Likewise, if CAS does not blunt lipolysis and, thus, lipolytic rate is not different from a non-nutritive PLA during sleep, then pre-sleep feeding strategies can be developed without concern for effects on lipolysis.
We have previously compared pre-sleep ingestion of calorically matched macronutrients to PLA in lean men and reported that morning fat oxidation was significantly greater for PLA compared to carbohydrates and whey protein but PLA was not different from CAS [9
]. Potential mechanisms contributing to the protein-related differences in fat oxidation relative to PLA may be due to alterations in postprandial metabolism during sleep. For example, insulin inhibits fat oxidation, and lower insulin responses have been observed following daytime CAS ingestion compared to whey protein [45
]. Insulin is also a potent inhibitor of lipolysis [46
] while growth hormone regulates nocturnal lipolysis [47
]. In the present study, we hypothesized that pre-sleep CAS consumption would not significantly alter insulin or growth hormone to attenuate SCAAT lipolysis during sleep (measured in a single overnight sample) more than PLA. Thus, lipolytic rate would likely not differ with CAS compared to PLA during sleep, which is consistent with our findings. It is important to mention that protein intake was adequate (above the recommended dietary allowance of 0.8 g/kg of body weight/day) and not significantly different between participants’ self-reported habitual diet and the controlled diet consumed prior to bedtime CAS ingestion (Table 2
). Additionally total caloric intake was not different between these diets which allowed us to assess the impact of CAS under well-controlled dietary conditions. No other studies have examined the overnight lipolytic effects of CAS when consumed 30 min prior to sleep. However, our data suggest that postprandial responses following ingestion of low-calorie CAS before sleep may not augment insulin or alter growth hormone concentrations to an appreciable level that would result in lipolytic inhibition compared to the fasted state in obese men. Thus, the additional calories from CAS before sleep did not inhibit lipolysis more than PLA (i.e., fasting) during sleep in obese men and, therefore, likely would not have affected fat oxidation.
Although our findings suggest that CAS prior to sleep does not influence fat metabolism and REE in obese men, we cannot ignore the influence of circulating hormones. We have previously reported increased insulin and insulin resistance in obese women the morning following acute nighttime protein ingestion [8
]. However, prior food intake was not well controlled and it is possible that the habitual diet during the day may have confounded the negative effects on insulin. In the present study, morning insulin concentrations did not differ among CAS and PLA trials when food intake was standardized by providing meal-replacement beverages to match the participants total daily caloric needs, however hyperinsulinemia (fasting insulin > 180 pmol/L) [43
] and insulin resistance [31
] were apparent at all collection periods (Table 3
). Under conditions of normal insulin signaling, high levels of insulin would favor fat storage by increasing lipoprotein lipase activity and decreasing hormone-sensitive lipase activity, thereby promoting triglycerides synthesis [46
]. Plasma fatty acids and glycerol were not measured in the present study (markers of whole-body lipolysis). However, we observed low RER values (~0.76 ± 0.01) and a corresponding predominance of estimated fat oxidation (fat oxidation: ~0.13 ± 0.01 g/min vs. carbohydrate oxidation: 0.08 ± 0.02 g/min) following both CAS and PLA. While the substrate oxidation rates were estimated from RER it must be noted that protein was not accounted for in these calculations and it is possible that these data may be an over-estimation of true values. Nevertheless, this suggests that plasma free fatty acids and glycerol were elevated, as RER has been shown to be inversely related to plasma fatty acid concentrations [48
]. However, it is noteworthy that the addition of a CAS beverage prior to sleep did not further augment insulin concentrations.
It is also possible that the hyperinsulinemia and low RER may be a result of sleep restriction [49
]. Rao et al. demonstrated that sleep restriction (~4 h) for five nights leads to lower RER (increase fat utilization) with no change in RMR but increases insulin resistance, cortisol and catecholamines [49
]. Despite poor sleep efficiency scores (i.e., below 85%) [42
] at baseline and during CAS and PLA trials, no differences in the quality or quantity of sleep between trials were present in the subset of participants (n
= 5). These data demonstrate that neither microdialysis nor bedtime nutrient ingestion affect sleep quality patterns. Although participants were instructed to consume CAS or PLA within 30 min of going to bed, the time spent lying down but not sleeping was ~6 h, while the average time spent sleeping was ~4.5 h (baseline: 4 h and 48 min ± 33 min; CAS: 4 h and 21 min ± 27 min; PLA: 3 h and 55 min ± 42 min). Thus, it is evident that the participants may have been sleep-restricted at baseline and during the CAS and PLA trials resulting in the low RER values and insulin resistance observed. The present study did not screen for irregular sleep patterns. Hence, future nighttime feeding studies in obese individuals should assess sleep quality and/or screen for sleep disturbances (i.e., sleep apnea).
While other studies have shown that nutrient intake prior to sleep may influence next morning appetite [8
], whether this holds true in obese men is unknown. We have previously reported that acute pre-sleep ingestion of calorically-matched protein or carbohydrate supplements increased feelings of satiety while concomitantly reducing the desire to eat the following morning, regardless of macronutrient type in obese women [8
]. Likewise, less hunger was reported the following morning in response to CAS compared to PLA administration during sleep in elderly men [12
]. Surprisingly, our hypothesis that CAS would result in greater appetite suppression compared to PLA was not supported. In fact, our participants reported greater desire to eat during the CAS trial relative to baseline but ghrelin, an orexigenic hormone, did not mediate these changes. Despite previous work showing blunted ghrelin concentrations in obesity [50
], obese individuals are highly sensitive to ghrelin’s orexigenic effects as low dose infusion (1 pmol/kg/min) reportedly increases food intake [51
]. Changes in subjective appetite, however, were only apparent with a high dose infusion (5 pmol/kg/min) [51
]. In the present study, we did not standardize dietary intake prior to the baseline visit, however participants were asked to fast for at least 8 h before their baseline visit. In contrast, liquid meal replacement beverages (Ensure Plus®
) were used to standardized dietary intake on the days participants consumed CAS and PLA at bedtime (i.e., the day prior to appetite measurements). It is conceivable that the increased feelings for a “desire to eat” may have been mistaken for a “desire to chew” given that their food intake on those days were in liquid form (different from food intake the day prior to baseline) and their subjective feelings of hunger and ghrelin levels were not altered by either trial. While standardizing meal intake with solid meals as opposed to liquid meal replacements would have been ideal, economically were unable to do this in the present study. Regardless, since food intake was standardized and no differences between CAS and PLA trials were present we feel that the same effect would have been observed. Contrary to the present study, data from our laboratory [8
] and others [12
] have demonstrated a satiating effect of proteins.