Alternative Eats Part II: Artificial Sweeteners and Weight Management
Exploring the utility of artificial sweeteners for weight-loss
So, last time in the Alternative Eats Series (I) we established that non-nutritive sweeteners provide sweet taste with zero or nearly zero calories attached via one or several of the following mechanisms: supplying high degrees of sweetness with minuscule amounts of sweetener, being excreted un-metabolized via urine, and passing through the GI tract unabsorbed. And, we also alluded to how this lack of calories makes these sugar-substitutes valuable for weight-loss, in theory, if we accept that creating a caloric deficit–where we are consuming less than we burn–is essential for losing weight. Now, we face the question of practicality: though they are EFFICACIOUS in theory, are these compounds EFFECTIVE options for losing weight?
Although we identified a few categories of non-nutritive sweeteners last time, you’ll find that almost all of the research cited below pertains to artificial sweeteners (AS), like aspartame, sucralose, saccharin, and acesulfame-potassium. Furthermore, it’s important to keep in mind that, while there is research considering AS consumption in foods or both drinks and foods, much of the research below pertains to AS use in drinks. Also of note, while there are many other considerations regarding these sweeteners’ impacts on our health from the zoomed-out perspective–risk of cancer, Type 2 Diabetes, cardiovascular disease, and gut dysbiosis, to name a few–we’ll put these to the side for future posts and zero-in on weight-loss alone here.
Friendly Disclaimer: I am not a doctor, and none of the information below is medical advice in any way, shape, or form. You are responsible for conducting your own research, consulting your medical professionals, and making your own decisions.
Bottom-line Up-front
Artificial sweeteners are a prime example of how the picture can change fairly dramatically depending on which type of data we are looking at. Rodent studies (XVII, XVIII) and observational data (X, XIX, XX, XXII, XXV, XLI, XLIII, XLIV, XLVI, XLVIII) generally suggest a negative impact, while randomized-controlled trials (RCTs) generally suggest a benefit (IV, V, XXVI, XXVIII, XXIX, XXX, XXXII, XXXIII, XXXV, XLII, XLIII, XLV).
The rodent data is not necessarily applicable because, simply put, humans aren’t mice and rats.
Though mostly pointing towards a harmful association, the observational data is somewhat inconsistent, with some data associating artificial sweetener consumption with weight-loss. (XXIII) Also, this data is confounded by the potential of reverse-causation, whereby, instead of artificial sweetener consumption causing overweight/obesity, people who become overweight or obese may be more likely to consume artificial sweeteners–the change in weight may cause the artificial sweetener consumption, rather than the other way around. (IV, V, XXVIII, XXIX, XLI, XLIII, XLVI, XLVII) And, the observational data is weakened by the observational studies utilizing food-frequency questionnaires, dietary recalls, and food diaries where subjects self-report their dietary intakes for the past days, weeks, months, years, or even decades. (X, XIX, XX)
Though also imperfect, RCTs probably represent the most reliable data in this realm, in my opinion. In particular, the longer term studies showing that protocols including artificial sweetener consumption lead to significant weight-loss and weight-loss maintenance are convincing. (XXVIII, XXIX, XXX, XXXII) Additionally, several meta-analyses on RCTs suggest benefits for weight management (IV, XXXIII, XXXV, XLII, XLVI), and countless RCTs suggest that AS substitution does not seem to lead to increased calorie consumption in general, consumption of sweets, sugar cravings, or hunger–nor reduced fullness–in comparison to sugar-sweetened options or water. (V, XIV, XV, XXVI, XXVIII, XXIX, XXX, XXXII, XXXIII, XXXVI, XXXVIII, XXXIX)
But, there is nothing special about artificial sweetener consumption in terms of weight-loss, as other options that cut calories work as well. (XXVIII, XXX, XXXII) Artificial sweeteners are just an option to make weight-loss feasible and sustainable over longer periods of time by allowing tasty options to be included in a reduced-calorie diet.
At the end of the day, if we take a zoomed-out approach and look at weight management alone, both the supporters and naysayers will agree that artificial sweetener consumption will lead to weight-loss IF it leads to a caloric deficit. And, whether or not that actually happens likely depends upon how they are used, what they are used for, and who is using them.
In my opinion, considering the body of evidence as a whole—including the strengths and weaknesses of different study types and designs—artificial sweeteners are likely one effective option for weight management.
There are three main pools of research surrounding the relationship between AS and weight management: rodent models, observational studies, and randomized-controlled trials (RCTs). On top of these, there are reviews and meta-analyses which assess many studies together to make more overarching assessments of the entire body of data. We’ll see below that AS are a prime example of how the picture can change fairly dramatically depending on which type of data we are looking at.
The Rodent Story
The rodent story is short and simple, in that some studies do suggest that AS consumption can lead to increased caloric consumption in rats. (XVII, XVIII) One hypothesis to explain these outcomes is that, “the effectiveness of energy regulation will be reduced under circumstances that degrade the ability of tastes to predict the occurrence of the caloric or nutritive postingestive consequences of eating.” (XVII) In other words, maybe dissociating sweet taste from caloric content–as zero-calorie sweeteners do–messes with the rats’ bodies’ abilities to regulate energy consumption, in that sweet foods no longer cue their nervous and endocrine systems to decrease their hunger after eating/drinking. Others propose that differences in energy expenditure during digestion of caloric and non-caloric sweeteners explains the differences, or that the rats’ insulin responses to some artificial sweeteners are behind the results. (XVIII)
So, case closed, right? Not quite. One reason basing our argument off of rats is insufficient is that, simply put, humans aren’t rats. With regards to energy balance, rats have much faster metabolisms than humans; furthermore, rats, along with other animals used for research, are known to vastly overeat when provided food around the clock. Still, this rodent data is useful as a stepping stone for exploring our hypotheses further through more human-specific research.
Artificial Observation
One flavor of this human-specific research consists of observational cohort and cross-sectional studies, where researchers seek out associations between AS consumption and weight change or risk of being overweight/obese over time. As with all observational data, these studies are vulnerable to confounding variables–or factors that are influencing the results but that we are not accounting for–and they are only capable of showing correlations, not causations. In addition, much of the observational data about AS comes from food-frequency questionnaires and dietary recalls, where subjects self-report their dietary intakes for the past days, weeks, months, years, or even decades. (X, XIX, XX) Take a moment to attempt to recall exactly what and how much you ate for breakfast, lunch, and dinner 3 days ago, and you’ll realize why this is not a perfectly reliable method of data collection.
Credit for this thought experiment goes to Dr. Peter Attia, who I have heard use this example several times to depict the flaws in food-frequency questionnaires and similar protocols.
Other options for tracking food intake include food journals, where subjects record everything they eat in real-time; however, these aren’t perfect either, as they typically entail extrapolating the subjects’ diet for months, years, or decades based on how they ate during the recorded day(s) or week(s). (XXV) This issue is somewhat inherent to this method, as subjects’ entries will likely become less consistent and accurate the longer they have to record their food intake–hence the reason these food journals typically only cover days and weeks worth of time. Even putting the Hawthorne Effect aside, it’s unlikely that looking at what subjects ate over the course of one week will be completely indicative of what they’ve eaten over previous months, years, and decades.
That being said, these observational studies do pose some pro’s, in that their lower maintenance and control in comparison to randomized-controlled trials (RCTs) allows for larger sample sizes. In addition, rather than taking place in a synthetic lab environment, the subjects’ choices and outcomes play out in real-life conditions. So, what do these observational data have to say about AS and weight management?
Dating as far back as 1986, countless observational studies show positive correlations between AS consumption and negative health outcomes, such as metabolic syndrome, glucose intolerance, Type 2 Diabetes, cardiovascular disease, cardiovascular disease mortality, all-cause mortality, and obesity. (X, XIX, XX, XXII, XXV, XLI, XLIII, XLIV, XLVI, XLVIII)
But, the data isn’t crystal clear. First, while some observational data suggests that AS consumption is associated with an increased risk of negative health outcomes across long periods of time (i.e. up to 10+ years follow-up), other observational data either shows no effect, that adjusting for confounding variables attenuates or expunges the once statistically significant associations, or that increased consumption of artificially-sweetened beverages is actually associated with weight loss. (XXIII, XLIII, XLVI)
Second, the positive associations between AS and overweight/obesity may be due to reverse causation, whereby, instead of AS consumption causing overweight/obesity, people who become overweight or obese are more likely to consume AS–the change in weight causes the AS consumption, rather than the other way around. (IV, V, XXVIII, XXIX, XLI, XLIII, XLVI, XLVII)
In this regard, it’s interesting to note that some data suggests AS consumption increases with age. (XLVII) Anecdotally, I find it is common for older individuals to start making adjustments to their diet and lifestyle in their later years when the impacts and limitations of their health conditions are becoming more tangible and problematic. So, older individuals may be more likely to consume AS because they are attempting to mitigate chronic health conditions (i.e. obesity, Type 2 Diabetes, CVD, metabolic syndrome, etc.)–their health issues may be causing them to consume AS, rather than the other way around.(XLVII)
On the other hand, some data suggests that overweight/obese individuals are more likely to consume both sugar-sweetened AND diet soft drinks, suggesting that diet soft drinks are not being used as complete substitutes for sugar-sweetened soft drinks. (XXIV) For example, some data from 1980 to the 2000’s shows that sugar consumption has only slightly increased by ~15% while AS consumption has more than doubled; consequently, some individuals may be consuming AS in addition to their sugar-sweetened beverages, rather than as a substitute, and are confounding the observational data as a result. (XLVII)
To this point, additional data suggests that AS consumption is associated with greater weight-loss amongst those who lose weight over follow-up periods but greater weight-gain amongst those who gain weight across follow-up periods. (XLVII) In other words, AS consumption is associated with outcomes in both directions–weight-loss and weight-gain–depending on the context. So, AS may be associated with overweight/obesity in the context of individuals who are adding them to their diet, rather than using them as a sugar-substitute. But, they may also be associated with weight-loss if we are looking at individuals who actually do use AS as a substitute for caloric sweeteners with the intent of reducing their caloric intake. (XLVII)
So, the observational data generally point in the direction that AS are associated with weight-gain and increased risk of overweight/obesity. But, the data are not uniform, in that some observational studies point in the opposite direction; furthermore, there are several lines of logic suggesting that confounding variables are behind the positive associations. To clear up these inconsistencies and confounders, let’s look at what the RCTs show regarding AS and weight management.
Artificial Intervention
Where observational studies offer large sample sizes and real-life conditions, RCTs offer greater control of variables and a stronger link to causation. Specifically, RCTs are valuable because they involve actually implementing an intervention–in our case consuming or avoiding AS–and recording the outcomes that follow. For this reason–their specificity to actual application in humans and controlling of confounding variables–RCTs are commonly viewed as the golden standard of research.
With that said they do have some limitations, in that the lab environment is often not perfectly representative of real-life conditions. And, the further RCTs go from the lab environment–for example, instead of keeping subjects in the lab and giving them a specific diet for a week, sometimes researchers simply give the subjects food to eat at home or instructions on how to eat–the less controlled they become. But, moving away from the lab becomes more and more inevitable the longer the trial is, as you can only house subjects in a facility for so long. And, it’s important to eventually utilize longer trial designs, as we ultimately care about long-term changes in health outcomes. Taking these considerations into account, let’s check out the RCT data surrounding AS and weight management.
First, there are many studies exploring how AS affect weight management indirectly through hunger, cravings, and calorie consumption. In this realm, some researchers hypothesized that, since AS do not contain any calories, individuals may not be satiated when consuming AS and may compensate by eating more in other meals or snacks as a result. (V, XXXIII, XXXVII, XLV, XLVII) For example, though you can cut your caloric intake by drinking a diet soda instead of a full-sugar soda, you may end up eating back the calories you cut later on.
RCTs that explore this question of caloric compensation generally find that AS consumption results in incomplete caloric compensation in the short and long term. (V, XIV, XXXIII, XXXV, XXXVII, XLVII) For example, one meta-analysis shows that aspartame substitution for sugar leads to an average ~15-32% compensation of calories–meaning that subjects ate back ~15-32% of the calories they cut with the aspartame substitution, achieving an ~85-68% reduction in the calories after compensating. (XXXIII) Interestingly, the compensation was greater (~32%) when the aspartame was substituted in solid food than when it was substituted in drinks (~15%). (XXXIII) In other words, subjects ate back less of the calories they cut from artificially-sweetened drinks than artificially-sweetened foods in terms of relative percentage of cut calories. This doesn’t necessarily mean they ended up cutting more calories in total, as foods typically contain greater calories than drinks–a point we’ll come back to later. Check out Figure I below for a visual of what the caloric compensation and net intake might look like if an individual replaced a 1,000 calorie pint of ice cream with a 400 calorie artificially-sweetened option or 2 cans of soda with a zero calorie diet soda.
Of note, many of these compensation studies were blinded, in that the subjects did not know whether they were consuming artificially-sweetened or sucrose-sweetened products. Interestingly, one study found that substituting aspartame for sucrose while equating calories for the substituted food led subjects aware of the substitution to tend to consume more calories on days when they consumed the aspartame meal than subjects on days they consumed the sucrose meal, a plain meal (unsweetened), and the aspartame meal while unaware of the substitution. (XXXVIII) Check out Figure II below for a visual of these data.
None of the differences in total daily caloric intake reached statistical significance, but the authors proposed that the trend towards greater calorie consumption on the aspartame days for those aware of the substitution is because those subjects used the calorie-free substitution to justify eating more throughout the day. In support of this point, subjects aware of the substitution ate the most calories on average during lunch and dinner on the days they consumed the aspartame-sweetened breakfast compared to all the averages for all other conditions (i.e. aspartame unaware, sucrose aware, sucrose unaware, etc.). (XXXVIII) The authors did not mention whether the subjects were aware that the breakfasts all contained the same amount of calories, despite the sweetener substitution. However, based on their explanation of the trend towards a difference in caloric intake, it seems that the subjects aware of the substitution likely believed the aspartame meal contained less calories than the others.
This is important, as, though it seems that individuals do not completely compensate for their calorie consumption following AS substitution, it’s key to remember that there is no difference in overall calorie consumption if calories are equated when you substitute an AS for a caloric sweetener. (XXXVIII) In other words, there is no difference in overall calorie consumption if the artificially-sweetened option still has the same calories as the sugar-sweetened option. (XXXVIII)
In addition to caloric compensation, some researchers propose that consuming AS may ultimately lead to weight-gain due to increased sugar cravings from the extreme sweetness of AS. (V, XV, XXVI, XXXVIII) They also propose that the presence of sweetness without calories attached may lead to overall increased calorie consumption. (XXXIX) In this regard, according to most RCTs, AS substitution does not seem to lead to increased calorie consumption in general, consumption of sweets, sugar cravings, or hunger–nor reduced fullness–in comparison to sugar-sweetened options or water. (V, XIV, XV, XXVI, XXVIII, XXIX, XXX, XXXII, XXXIII, XXXVI, XXXVIII, XXXIX)
To be fair, there are a select few studies that show opposite results. (XXXIX, XLIX) However, the differences in outcomes may be ascribed to very short study duration (only 2 days long) (XXXIX) and administration of artificially-sweetened or sugar-sweetened beverages across 12 weeks without instructions or intent to reduce calories and/or lose weight. (XLIX)
Some data actually points to improvements in hunger/craving metrics, where AS-use leads to less hunger as compared to drinking water on a reduced calorie diet–though probably not enough to be clinically significant. (XXVIII, XXIX) Also, compared to substituting with water, substituting artificially-sweetened beverages for sugar-sweetened beverages led to similar reductions in calories and a greater reduction in dessert and caloric sweetener consumption in one study–though the water intervention also led to a reduction in dessert and caloric sweetener consumption. (XV) These results are depicted in Figure III below.
Furthermore, some meta-analyses show that AS consumption leads to decreased energy intake in comparison to sucrose consumption. (IV, XXXIII, XXXV, XLII, XLVI)
One particularly interesting point is that AS consumption could potentially lead to decreased sweet cravings if it induced weight-loss, as some research suggests that weight-loss is correlated with increased sensitivity to sweet taste. (V) If true, this change in sensitivity would be beneficial, in that it means less sweetness is required to reach the same level of taste satisfaction. Still, though interesting, this is only a hypothesis, and the data connecting weight-change and sensitivity to sweet taste are inconclusive. (V)
The RCT data above are important with regards to challenging the hypotheses that explain the rodent and observational data showing an association between AS consumption and weight-gain or risk of overweight/obesity. Nonetheless, more so than impacts on sugar cravings, reported calorie consumption, hunger, or other indirect measures, the data that directly measure how AS consumption impacts actual weight is most important for our purposes here. So, what do the RCTs assessing AS-use and weight management show?
Generally, RCTs show that AS consumption is effective for decreasing body weight, BMI, fat mass, energy intake, and waist circumference. (IV, V, XXVI, XXVIII, XXIX, XXX, XXXII, XXXIII, XXXV, XLII, XLIII, XLV) But, there are some nuances to the data depending on the study design you look at. For example, AS consumption is advantageous for weight management/loss in comparison to sugar consumption (IV, V, XXVI, XXXIII, XXXV, XLII, XLIII, XLV); however, in some cases, the differences between outcomes in AS and sugar groups is the result of less weight-gain rather than actual weight-loss.(XLV) In addition, some studies show that AS consumption shows no difference in weight-loss in comparison to water, nothing, or attention-control (i.e. receiving the same dietary advice but without AS consumption). (IV, V, XXX, XXXII, XLV)
Taking a step back, this data further supports the idea that there is nothing special about AS for weight management, in that they simply aid in reducing caloric intake. And, when subjects do so using water to replace sugar-sweetened beverages or simply by eating less, for example, they can lose similar amounts of weight as those utilizing AS to reduce their calorie intake. (XXX, XXXII)
With that said, there is some evidence to suggest that AS consumption makes reducing calories easier, especially in the longer-term. For example, subjects in this study achieved statistically significantly greater weight-loss following a weight-loss protocol including daily artificially-sweetened beverage consumption for 12 weeks than those on the same protocol but consuming water (~13.1 lbs vs. ~9.0 lbs)–though, the subjects in the water group did achieve significant weight-loss as well. (XXVIII) And, the subjects in the AS group also outperformed those in the water group in terms of weight maintenance, with those completing the study achieving ~18.5 lbs of weight-loss on average at the 1-year mark compared to the water groups’ completers’ ~7.5 lbs of weight-loss on average. (XXIX)
Another study found similar results, where female subjects consuming aspartame as a part of their weight-loss program lost similar amounts of weight during an initial weight-loss period as those on the same program but excluding aspartame consumption. (XXXII) Then, the subjects in the aspartame group maintained their weight-loss significantly more than the subjects in the no-aspartame group across 3 years after the active weight-loss period. (XXXII) Of note, the aspartame group did regain some of the weight, but they ultimately lost 11.2 lbs on average compared to baseline at the 3-year follow-up while the no-aspartame group’s average weight returned to baseline at this time. (XXXII)
In one more study, replacing caloric beverages with artificially-sweetened beverages for 6 months resulted in similar average amounts of weight-loss as replacing caloric beverages with water or solely following weight-loss advice (i.e. attention-control group). (XXX) But, subjects in the AS group were 2.29 times more likely to achieve at least 5% weight-loss than those in the attention-control group. (XXX) Though, again, the average weight-loss in all groups–AS, water, and attention-control–was significant and similar at the 6-month mark. (XXX)
Together, these studies suggest that artificial sweeteners are most useful in making weight-loss feasible and sustainable over longer periods of time, provided that you actually use them to replace sugar and reduce your caloric intake. And, this makes sense, in theory, as AS allow you to cut calories while still enjoying sweet drinks and foods. These data also agree with the point that there is nothing special about AS consumption in terms of weight-loss, as other options that cut calories work as well.
Now, though the RCT data as a whole generally points towards AS-use as being beneficial for weight-loss, these studies aren’t flawless either. For example, many of them utilize similarly faulty collection methods as the observational studies we mentioned above. (XXVI, XXVIII, XXIX, XXX, XXXII) Also, the outcomes of each RCT are specific to the population, protocols, and conditions of the specific RCT, so it is not certain that the results are generalizable to everybody. Furthermore, as I mentioned in a recent post, inter-individual variability means that our personal results will vary from the averages taken from these studies. For example, though AS consumption may have decreased hunger, lowered caloric intake, and led to weight-loss on average in these RCTs, some individuals may find the opposite results when they try out the protocols themselves.
Conclusion
So, with all of this taken into account–the rodent, observational, RCT, and review/meta-analyses data, along with all of their strengths and limitations–are artificial sweeteners helpful, harmful, or neutral with regards to weight management?
In my opinion, the specificity, control, and consistency of the RCT data out-weighs the rodent and observational data that links AS consumption with weight-gain and overweight/obesity. I find this to be especially true when considering the confounding variables and potential reverse-causation in the observational data.
And, the meta-analyses data for RCTs is strengthened when considering that several meta-analyses produced similar effect sizes and quantities of weight-loss as a result of AS consumption. (IV, XXXIII, XLII, XLV) In particular, De la Hunty’s meta-analysis of RCTs of aspartame consumption showed that the expected caloric reduction based on total displaced calories with aspartame substitution, as well as the typical amount of caloric compensation, predicts a weight-loss nearly identical to what their meta-analysis found in terms of weight-loss seen in RCTs. (XXXIII) In other words, not only were the calorie intake, caloric compensation, and weight-loss data in agreement directionally, but the amount of calories consumed and compensated for matched the amount of weight lost. (XXXIII)
With that said, I think there are some interesting nuances to account for. If, in addition to weight-loss, the goal is to reduce cravings for sweet foods altogether, then reducing the sweetness of the diet and acclimating to less sweet foods, rather than substituting with artificial sweeteners, may be optimal. (XLVIII) Interestingly, some data suggests that this process–reducing the magnitude of a sweet, fatty, salty, etc. sensation–does not reduce overall satisfaction; instead, doing so decreases the magnitude of the taste stimulus needed to elicit the same degree of satisfaction. (XLVIII) For this reason, cutting down on all sweet foods may be ideal if you’re attempting to curb sweet-cravings.
Also, some authors suggest that AS could inadvertently lead to negative health impacts if, for whatever reason, individuals consuming them decided to stop and required an increase in sucrose consumption to satisfy their habituation to and consequent craving of highly sweet foods. (XLVI) To me, this sounds similar to the idea that individuals on GLP-1 agonists–medications that reduce the desire to eat and lead to weight-loss as a result–will ultimately regain the weight they lost if they discontinue their drug-use. In both cases–temporarily using AS or GLP-1 agonists–the intervention seems to be a band-aid fix, rather than a sustainable solution.
At the end of the day, if we take a zoomed-out approach and look at weight management alone, both the supporters and naysayers will agree that AS consumption will lead to weight-loss IF it leads to a caloric deficit. And, whether or not that actually happens is easily testable at the N=1 level: you simply replace caloric drinks/foods with lower-calorie, artificially-sweetened options and see if it helps you lose weight.
In this regard, though the question of artificial sweeteners’ utility is generally presented in a binary picture–where they are either awesome or terrible–these compounds’ usefulness may fall along a spectrum, ultimately depending upon how they are used, what they are used for, and who is using them. (XLVIII) In other words, they have pro’s–like reducing energy intake and improving adherence in the short and long terms by allowing palatable foods while dieting–and they have con’s–like potentially allowing a strong desire for sweet foods to persist, rather than decreasing this desire by reducing sweetness in the diet for a prolonged period of time. And, these pro’s and con’s are likely of different significance and magnitude depending on the individual. (XLVIII)
In my personal experience, though most of the talk around AS is based on sugar-free beverages, the greatest potential for cutting calories lies with artificially-sweetened foods. For example, a Coke Zero saves you 140 calories per 12 oz can of Coca-Cola Original. But, one pint of Halo-Top can save you 600+ calories in many cases, depending on what brand of original ice cream you eat.
In this way, I suppose it comes back to personal preference and identifying which parts of your current diet you can substitute out for lower-calorie, artificially-sweetened options. And, this could simplify weight-loss dieting for a lot of people, as making a single swap of a low-calorie option for a high-calorie option while keeping the rest of the diet the same could lead to significant weight-loss.
If you’re interested, I described the concept of high-impact foods–those that are difficult to remove from your diet due to cravings and are also high in calories–and how finding low-calorie substitutes for these foods may be an effective strategy for weight-loss in another post.
You can always just cut calories without using AS as well. And, there are many people today that suggest you should, claiming that artificial sweeteners cause several other health issues, from cardiovascular disease to Type 2 Diabetes, cancer, and gut dysbiosis. Stay tuned, and we’ll explore the research around artificial sweeteners and these topics as well in this Alternative Eats Series.