Study under review: Sugar rush or sugar crash? A meta-analysis of carbohydrate effects on mood
If you want to stay on top of the latest nutrition research, check out the Nutrition Examination Research Digest.
As summer rolls around the corner, you can be sure parents are ready to stop their kids from drinking too much soda at little Billy’s pool party. As the title of the article under review suggests, acute sugar intake (or carbohydrates in isolation) has been commonly thought to lead to a ‘sugar rush’ (burst of energy) followed by a ‘sugar crash’ (fatigue from using up said burst of energy) though there is no strong scientific evidence to support this myth.
The "sugar rush" notion was dispelled in the scientific community when a comprehensive 1995 meta-analysis demonstrated no effect of sugar on behavior or cognitive performance in children. Nonetheless, the myth survives and for almost 25 years after that meta-analysis, researchers continued to study the influence of sugar on mood. This occurred in the context of the U.S population, which uses soda, energy, and sports drinks as one of the top ten sources of energy in the country. More specifically, some studies investigating the interaction between mood and simple sugars have suggested changes in cognitive performance and emotional wellbeing.
One appealing hypothesis regarding carbohydrates (CHOs) and emotional well-being is the serotonin hypothesis. This hypothesis, whose details are depicted in Figure 1, postulates that CHO consumption leads to higher levels of the feel-good neurotransmitter serotonin in the brain as a result of greater tryptophan (a precursor to serotonin) availability. This is supported by reports of individuals suffering from mood disorders (i.e., depression, anxiety, etc.) that tend to self-medicate with CHO-rich meals. However, one study from the 1980s failed to observe increases in tryptophan or serotonin after eating, except in a protein-free meal. This tryptophan-increasing effect appears to be stopped if as little as 5% of the ingested CHO drink or meal is protein.
Other studies have demonstrated conflicting results, reporting no influence or even slightly detrimental effects on the overall relationship between mood and CHO intake. While reviews have recently covered the relationship between CHO intake and mood, studies have not been quantitatively (meta-) analyzed to determine which direction the evidence leans and to quantify how heterogeneous the literature really is. The authors of the study under review conducted a meta-analysis to fill this gap by quantitatively analyzing the short-term effects of isolated CHO intake on mood while considering potential confounding variables such as CHO dose and/or type, fasting interval, and any activity preceding mood assessment.
Carbohydrates are believed to influence behavior and mood. While some studies suggest carbohydrate intake will increase the release of the “feel-good neurotransmitter” serotonin and even boost cognitive performance, some report no influence, if not slightly detrimental effects. The study under review aims to quantify the effect and explain the conflicting results through a review and meta-analysis of results and confounding factors that may influence the putative interaction between CHO intake and mood.
Who and what was studied?
This was the first systematic review and meta-analysis of RCTs that evaluated the effects of acute CHO ingestion on measures of mood in healthy adults. Eligible studies could assess the effects of CHO and another compound (like caffeine) only if the other compound was also present in the control group. The meta-analysis was not preregistered and did not specify whether it followed PRISMA or similar methodological guidelines.
A total of 31 studies including 1,259 participants with an average age of under 30 years old were included in the meta-analysis. Most studies used glucose (n=16), sucrose (n=4), or a combination of the two (n=3), while a few used maltodextrin (n=1) or an unspecified CHO (n=7). The placebo was most commonly an artificial sweetener. Participants were tested either after an overnight fast or were required to fast for two to three hours prior to ingesting test meals or shakes containing 15-128 grams of carbohydrates.
All studies reported composite scores for mood constructs (i.e., alertness, calmness, tension, anxiety, etc.) from two tests. The first was the Bond-Lader visual analog scale, which is a 16-item test with adjective antonym pairs on either end of the scale. Participants choose where they feel between the two antonyms. One example of an antonym pair is “tense” versus “relaxed.” The second construct was the Profile of Mood States (POMS)—a 65-item 5-point scale ranging from 0 (not at all) to 4 (extremely). The composite mood constructs served as the only outcome measures for the study at hand. Overall mood was also analyzed using a combination of the composite scores if data for overall mood was not already assessed in the study.
A random effects meta-analysis was conducted using Hedges g-corrected standardized mean differences (SMDs), which is a way of pooling data from studies that assess the same outcome but measure it in different ways. Check out the sidebar of NERD #48, Volume 1’s “Can alpha-lipoic acid help manage metabolic diseases?” for a refresher on SMDs. The authors evaluated the potential effects of confounding variables (dose and type of CHO, fasting interval, and activity preceding mood assessment) only when heterogeneity was significantly high. Separate meta-analyses were conducted for three time-windows: immediate (0–30 minutes), short term (31–60 minutes), and long term (more than 60 minutes). Publication bias was assessed using funnel plots and statistical tests. Studies were considered as outliers if their results were more than 2.5 standard deviations outside the average and, if that was the case, were removed from the analysis to prevent them from skewing the results.
The study under review is a review and random-effects meta-analysis that evaluated the effects of acute isolated carbohydrate ingestion on composite measures of mood at 0-30 minutes, 31-60 minutes, and more than 61 minutes after intake in 31 RCTs on healthy individuals.
What were the findings?
Overall mood improved (p=0.051) more than 60 minutes after consuming CHO compared to placebo, while no statistical differences between the two were noted at less than 30 minutes or 30-60 minutes. For each of the mood subscales, no statistically significant differences were found between CHO and placebo at any time except for a reduction in alertness at 30-60 minutes and an increase in fatigue at less than 30 minutes with CHO.
Heterogeneity was low and non-significant across almost all mood constructs. Only the task preceding the mood assessment was found to exhibit a trend toward significance following analysis for which factors might be causing these differences. Further analysis of this confounding variable revealed that groups ingesting CHO reported significantly less fatigue in comparison to placebo after physically demanding tasks and not after cognitively demanding tasks. A trend (p=0.078) was also found for reductions in fatigue from CHO ingestion following stressful tasks, but this included only one study for the more than 61 minute window.
Publication bias was only found for contentedness in one test (Egger’s test) but this may have been due to the low number of included studies. A trend for publication bias was also noted for overall mood in both Egger’s and Begg’s tests.
The results failed to demonstrate any consistent beneficial effects on mood and suggest some detrimental effects on alertness and fatigue, which peak within an hour of carbohydrate ingestion. On the other hand, carbohydrate intake appears to prevent fatigue an hour or more after completion of physically demanding tasks and might prevent fatigue following stressful tasks. Heterogeneity and publication bias was rare across studies, except for a trend of publication bias, which was observed for overall mood.
What does the study really tell us?
The results from this study suggest that CHOs can have beneficial effects on overall mood an hour after ingestion. However, the trend for publication bias (meaning papers with positive results are more likely to be published) for overall mood suggests potential for selective reporting that would temper the confidence that can be placed in this particular result. For example, one study that was included in the analysis only reported two of the six subscales from the POMS, forcing the overall mood calculation for the study to be based on only the two subscales reported. It is also worth mentioning that meta-analyses do not represent the totality of the available research. After all, 20 studies were ineligible for the study under review because of missing data or non-responsiveness from authors. Moreover, conducting three meta-analyses per mood construct means lots of statistical tests, which may suggest that the few positive results could be noise resulting from multiple testing.
The results also suggest there may be slight reductions in alertness up to one hour following CHO intake and increases in fatigue within 30 minutes of intake, at least at face value. However, when looking at the confidence intervals for the results, a pretty strong pattern emerges: pretty much every single measurement is consistent with small to no effect. You can see this for yourself in Figure 2, where select statistically non-significant results are presented on the top, and all the statistically significant results below. Recall that, as a rule of thumb, “small” standardized effect sizes are around 0.2 and medium ones are about 0.5. Note that the confidence intervals overlap mostly small or no effect. The one exception is the confusion measurement, which overlaps medium effect sizes mainly because its estimate is so imprecise, implying that the data are consistent with a relatively broad range of values. However, this is the exception, rather than the rule: for the most part, the large majority of reported measurements are mostly consistent with either no effect or a small effect, at best. This implies that the effects of CHO on aspects on mood are relatively small, if they exist at all.
The overall low level of heterogeneity suggests that, despite a relatively wide range of study designs and methods across studies included in the meta-analysis, the results are relatively consistent and robust. In other words, even though there were differences between the experimental conditions of the studies that went into this meta-analysis, the outcomes probably wouldn’t change much. The data are generally consistent with the idea that CHO intake makes little to no impact on different aspects of mood, regardless of the particulars of study design.
While these results are robust, there are still some aspects of the meta-analysis at hand that limit the generalizability of the results. Since this study focused on healthy populations, it fails to address whether there might be measurable benefits in people with health issues, especially in individuals with glucose regulation problems. Also, the CHO was provided in isolation from other macronutrients or substances. While this isolates the effects of CHOs for proper science, rarely are CHOs consumed in isolation (especially after fasting) in the real world—most people don’t just eat simple sugars alone unless they’re drinking soda or juice—and combinations with other macronutrients or substances, such as proteins or caffeine, can have varying or synergistic effects on mood.
The results suggest that carbohydrates have beneficial effects on overall mood an hour after ingestion, but a trend for publication bias suggests there is an influence of selective reporting. Within an hour of carbohydrate ingestion, alertness was reduced and fatigue was increased. These specifics should be taken lightly, though, due to the multiple comparisons done in this study, combined with the overall trend of most measures being consistent with low to no effect. Overall, the data tend to suggest that if there’s any impact on mood, it’s small. The low heterogeneity across studies included suggests these results are relatively robust.
The big picture
Recent reviews investigating CHO-mood interactions have observed similar findings as the meta-analysis under review. One systematic review identified small or partial effects on mood for less than half of the studies investigating glucose, while no beneficial effects were observed from sucrose ingestion. A review of the effects of glucose in isolation and in combination with caffeine was also unable to demonstrate significant effects from CHO intake on mood. However, it did show that CHO with caffeine had a synergistic effect on cognition, but not mood—this might explain the appeal of many beverages, ranging from the classic Coke to various energy drinks.
In support of the serotonin hypothesis, previous studies have demonstrated small increases in tryptophan availability and serotonin synthesis an hour or more following isolated CHO ingestion. The details of the relative timing of tryptophan and serotonin levels derived from rats who are fed a carbohydrate diet are provided in Figure 3. The meta-analysis under review did suggest a benefit in overall mood at 60 minutes and longer, which would support the serotonin hypothesis. However, as we mentioned before, confidence in this finding should be tempered due to possible publication bias. Even ignoring the publication bias issue, though, this possible benefit is not very practical as it would only surface after an hour of increased fatigue and decreased alertness (according to the meta-analysis under review) and any small amounts of protein might interfere with the serotonin boost. Future studies should investigate how common snacks might influence mood to reach more practical findings.
One review argues that the benefits in mood and serotonin levels from CHO are primarily observed in stress-prone individuals or people with mood disorders. Given that the study under review only included studies of healthy individuals, further research on individuals with mood disorders could potentially reveal an effect.
Another particular circumstance where beneficial effects have been observed are in times of physical activity or psychological stress. In the study under review, moderator analysis identified physical activity preceding mood assessment as a potential explanation for differences in results for the mood construct of fatigue one hour after CHO ingestion. In principle, the finding that CHO administration reduces fatigue following physical activity (or even stressful activity) appears relatively sound, as the ingested CHO acts to quickly recover what was recently used. This would fit along the lines of greater energy availability and align nicely with the fact that our brain requires 100-150 grams of glucose per day and it’s been argued that CHOs are the brain’s preferred fuel source. However, this is all speculation as there are aspects of fatigue that are still not completely understood.
Carbohydrates alone might have a slight beneficial effect on overall mood in healthy populations. The potential benefits appear to surface when participants are subjected to physical or cognitive stress. Further research needs to clarify whether CHOs have a positive impact on mood in individuals with mood disorders.
Frequently asked questions
Q. Are there cognitive benefits to CHO intake?
It would make sense that CHO intake provides cognitive facilitation when the brain requires 100-150 grams of glucose per day to function, but starvation studies have shown that ketones (metabolites of fat) can fuel 60-70% of the brain’s energy needs, and the liver can make enough glucose for the rest. Given that the average person consumes about half of their energy intake as CHO, why might an extra soda with 50 grams of sugar make a difference in cognition? It might be as simple as the availability of those carbohydrates, but this is just speculation.
Nevertheless, a review suggests that intake of 25 grams of glucose can facilitate memory, but CHOs efficacy for other cognitive tasks (i.e., speed of processing, problem solving) need more research (especially testing different doses) as the results are not very robust. Even if there are cognitive benefits associated with carbohydrate intake beside memory facilitation, it appears the differences might be small and future reviews will need to identify potential confounders like as study design, sample population, and others.
What should I know?
This review and random effects meta-analysis evaluated data from 31 RCTs on healthy individuals that investigated the effects of acute isolated carbohydrate ingestion on composite measures of mood at 0-30 minutes, 31-60 minutes, and more than 61 minutes after intake. The review fails to demonstrate consistently beneficial effects on mood from CHO intake. While researchers did find detrimental effects on alertness and fatigue within an hour of ingestion and a putative and a beneficial effect on total mood scores 60 minutes after CHO ingestion, the number of tests done suggest taking these specifics lightly. Instead, the overall pattern seems to indicate that the data are consistent with little to no effect on most aspects of mood.
Future studies need to explore potential benefits in other populations (e.g. individuals with mood disorders), investigate the potential for weight gain, and determine whether observed effects from intake of CHO in isolation can actually be observed for the typical convenience store snacks (e.g. chocolate bars and chips) that people gravitate toward when they want to battle fatigue and improve their mood.