Sunday, September 10, 2017

Meta-Analysis: Dieting Reduces Food, Fat, Sugar, Starch & Junk-Food Cravings | In Whom? Why? Is it Diet Dependent?

Believe it or not - At least after some time, a severely energy restricted diet will reduce your cravings for donuts, pizza, pasta, and even chocolate statistically significantly - the questions from the headline do yet remain: Who benefits? What's the mechanism? And do the effects depend on your diet / its macros?
If you browse blogs and read e-books, you will read highly popular claims like "Dieting is useless, it just makes you hungry. If you want to lose weight, you got to stop eating carbs and reduce insulin." Those claims are popular because they entail the (unwarranted) claim that you could lose weight without cutting back on your energy intake. WRONG!

Fortunately, more and more people seem to understand that. One thing that will still be hard for them to swallow is the conclusion of the latest meta-analysis that addresses the ubiquitous claim that "calorie restriction may increase food cravings" (Kahathuduwa 2017), would thus ruin all your weight loss efforts and should be shunned in favor of macronutrien modulations that would promote satiety and thus allow for effortless weight loss.
Can non-nutritive sweeteners trigger food cravings / make you fat due to other side effects?

Aspartame & Your Microbiome - Not a Problem?

Will Artificial Sweeteners Spike Your Insulin?

Diet Soda Beats Water as Dieting Aid in RCT

Chronic Sweeten-er Intake Won't Effect Microbiome

Sucralose Tricks 'Ur Energy Gauge - Implications?

Sweeteners In- crease Sweet- ness Threshold
The study that addresses this important, if not crucial question was conducted by researchers from the Texas Tech University and the Pennington Biomedical Research Center. For their meta-analysis, C. N. Kahathuduwa, M. Binks, C. K. Martin, and J. A. Dawson dug up studies that were
  • conducted on subjects with obesity, 
  • implemented calorie restriction for at least 12 weeks, and 
  • measured food cravings pre-intervention and post-intervention. 
The eight studies the scientists found mostly used the Food Craving Inventory. "Other comparable methods were," as the authors point out "converted to a similar scale" (Kahathuduwa 2017). For studies w/ multiple follow-ups, the authors used the duration ≥12 weeks, but closest to 16 weeks for studies and performed DerSimonian–Laird random-effects meta-analyses (this is the most widely used method to estimate between studies variance) using the ‘metafor’ package in r software.

Cravings for food, in general, sweets, high-fat foods, starchy foods, and fast food declined significantly over the course of the dieting period in obese subjects.

The above, i.e. the significant reductions in almost all relevant domains of food cravings were observed with a certain heterogeneity (meaning the reductions varied between studies) across all studies; a heterogeneity of which the scientists' analysis of potentially confounding variables revealed that it was explained by baseline body weight, type of intervention, duration, sample size and percentage of female subjects.
For whom are the results of the study at hand relevant? To answer this question we have to take a closer look at both: the subjects and the design of the eight studies that were analyzed for the meta-analysis at hand. Since all subjects were obese and thus equipped with more than ample energy stores to compensate for the 500-1200kcal deficit, it is more than questionable that the results are relevant for a hard-training athlete in his contest prep.
Before we delve further into the intricate differences due to (a) subject characteristics or (b) study design, though, let's briefly appreciate the mean reductions in craving scores I plotted for you in Figure 1. As previously highlighted, the scientists observed observed reductions in pooled effects for overall food cravings (−0.246 [−0.490, −0.001]) as well as cravings for sweet (−0.410 [−0.626, −0.194]), high‐fat (−0.190 [−0.343, −0.037]), starchy (−0.288 [−0.517, −0.058]) and fast food (−0.340 [−0.633, −0.048]) in the meta‐ analysis.

What does it mean if the scientists say that cravings for sweet reduced by "0.410" pts?

As previously pointed out, Kahathuduwa et al. used the "Food Craving Inventory" (FCI | learn more in White 2002) for their comparisons. For studies which used different methods to quantify the food cravings of their subjects, the authors converted the corresponding scale(s) accordingly. As White et al. point out in their 2002 paper that introduced this tool to the scientific community, "[t]he FCI was designed to measure specific food cravings using two subscales: subjective cravings and consumption of particular foods." White et al. further explain:
"Low Carb(-ing) Reduces Fat & Fast Food (10-20%) Cravings Plus 60% Less Hunger After Meals in Obese Men/Women" - This 2017 study by Heimowitz et al. is not yet part of the meta-analysis at hand.
"We chose to develop these subscales to distinguish between a craving for a food and consumption of that food. Because previous research had often defined craving as consumption, our intention was to determine whether craving and consumption were psychometrically different constructs. The first subscale (subjective) assessed the frequency of subjective cravings for 47 different foods. The foods were chosen according to the classifications provided by Harvey et al. and additional foods were included on the basis of theory and literature review. The second subscale (behavioral) was intended to measure the extent to which participants gave in to craved foods. Participants completed the subjective and behavioral subscales for the same 47 foods" (Kahathuduwa 2017). 
Exemplary foods from the analysis are fried chicken, sausage, milk, tuna, beef, gravy, fried fish, bacon, corn bread, hot dogs, steak, brownies, cookies, cheese, candy, chocolate, donuts etc. As you may already have guessed based on this incomplete list, the FCI includes both, foods people typically crave (cookies, donuts, gravy, etc.), as well as foods you'd not necessarily put on a "food addiction" list, such as steak, veal, milk, tuna, etc.

The FCI assesses both subjective (craving) and objective (eating) factors asking subjects: "Over the past month, how often have you experienced a craving for the food?" and "Of those times in the past month during which you craved a particular food, how often did you 'give in'?". Subjects answer questions with values from 1-5 implying: 1 = never, 2 = rarely (once or twice), 3 = sometimes, 4 = often, and 5 = always/almost every time.
Despite the fact that the meta-analysis at hand refutes the assumption that energy restriction per se will trigger cravings, there's another feature common dieting attempts have in common: dietary monotony (Pelchat 2000). Just like the meta-analysis at hand, the increase in food cravings in response to a monotonous diet in young people on nutrient sufficient diets support the notion that food cravings have an important psychological component.
How can that be true? If you're dieting, you got to crave more... Well, that's what common sense and the "deficit theory" (Weingarten 1990 | cravings occur owing to a deficiency of energy or nutrients) may say. As the authors of the meta-analysis at hand point out, though, their results do not support the "deficit theory". Rather than that, the findings of the scientists from the Texas Tech University support the notion that - at least in the obese - "craving develops as a result of a conditioned association between repeated consumption of a specific type of food with a particular stimulus, environment or occasion (i.e. classical conditioning)" (Kahathuduwa  2017). The scientists speculate that after 12 weeks on an energy restricted diet that doesn't allow for the regular consumption of these usually energy‐dense foods, a dissociation of consumption of food and associated stimuli would have taken place and the hard-wired food cravings would have been erased.

In fact, scientists have observed that extended calorie restriction may suppress food‐cue reactivity of brain regions that regulate food reward and increase activity in regions that have been implicated in inhibitory control over drives towards food ingestion (Rosenbaum 2008) - a phenomenon that would be the "neurophysiological surrogate for ‘craving’" (Kahathuduwa  2017)

Kahathuduwa et al do also point out that "it should be noted that a food craving is likely a complex bio‐psycho‐social phenomenon that cannot be fully explained using a simple psychological model alone" (ibid.). Two of the missing physiological phenomena I would suspect to have an impact - especially on sweet, starch and fast food cravings - are (a) the ongoing reductions in inflammation and insulin resistance that will help restore a functional energy sensing system over the course of the diet and the restoration of normal sweet and fat taste perception. Obviously, further research is warranted to confirm or falsify this assumption. If we do yet assume that they're true, they point to a previously highlighted problem with the study at hand: as all studies were conducted with obese individuals, the chances to see similar effects in lean athletes are slim.
The -0.41 point reduction Kahathuduwa et al. report calculated for the sweet cravings of their subjects, is yet not a -0.41 reduction from say 4.0 to 3.59 pts on the Likert scale of the FCI, but an effect size the scientists calculated based on their analysis of all 8 studies. Effect sizes tell you something about the efficacy of an intervention. In this particular case, a negative effect size means that the intervention, i.e. the reduction in energy intake due to the diet, lead to reduced cravings. A positive effect size, on the other hand, would indicate that the cravings increased.
Figure 1: If one pools the data from all studies in the meta-analysis, both, the overall food cravings, as well as all individual food cravings, the scientists assessed (figure shows only fast food) were reduced significantly (Kahathuduwa 2017).
With effect sizes of -0.246 for general food cravings, -0.410 for sweet foods, -0.190 for high-fat foods, -0.288 for starchy foods and -0.340 for fast-food, the effect of dieting was generally positive, all cravings that were assessed in the study were reduced. Furthermore, the scientists observed a...
  • small effect (d<0.30) on general, high-fat, and starchy food cravings, and 
  • medium effect (0.30 < d < 0.50) on cravings for sweet and fast-food.
Accordingly, you cannot expect that category 5 cravings (FCI: almost craving) will be reduced to category 1 (FCI: never craving) if you start dieting. What's also worth knowing is that the scientists' meta-regression analysis shows that for all food cravings...
  • a  higher baseline weight was associated with a decreased effect size - this means the heavier participants saw greater reductions in cravings 
  • a higher percentage of women in the study population was likewise associated with increased effect sizes - in other words, women are more prone to develop cravings, accordingly, studies with more women found slightly lower (d-0.016) reductions in cravings
  • a more extensive list of items the subjects were not allowed to eat at all was associated with an increased effect size (d+0.707) - this means that the more restrictive (in terms of food choices) the diet was, the smaller the beneficial effect on food cravings
In other words: (1) The more body fat you have to draw on while you're dieting, the less prone you're going to be to suffer from cravings. (2) Being a woman makes it a bit more difficult to diet. And (3) restricting certain food items, groups and/or macronutrients can, in line with both, the deficiency theory and hedonistic explanations of food cravings, trigger increased cravings for the prohibited (macro)nutrients [deficiency theory] or food groups [hedonistic explanation].
Figure 2: The link between low-carb and increased cravings for sweet/starch is mainly due to a single study (red), which isn't even a true low-carb vs. high-carb comparison - another study (green) even shows low-carb advantages.
The last-mentioned conclusion, however, seems to be driven by the inter-group differences in the low-carb vs. high-carb group of a 2012 study by Jakubowicz et al., a study that used a diet the defining feature of which was that it was low calories and comparatively high in protein: With a significantly reduced energy intake (-1400-1600kcal total) and whopping ~180g protein, the diet left room for only a marginal inter-group difference of carbohydrate intake of 10% (20% vs. 10% of total energy came from carbs in the high vs. low carb group). This is obviously hardly what you would expect from a "low-carb vs. high-carb" comparison and would warrant only one conclusion:

A low-calorie high protein (52%) medium-fat (38%), low-carb diet (10%) will increase your cravings for sweet, starch and all the other sources of readily available energy.

The only 'real" low vs. high carb comparison in the meta-analysis, confirms the notion that low-carb or, I should say, high-fat, low-carb diets will not make you crave carbs. After all, Martin et al. who compared a diet with only 20 g carbs per day (in the form of low-glycemic index vegetables) to an isocaloric diet that contained 55% of the energy intake in form of carbs in 2011, found no detrimental effects of a significantly reduced carbohydrate intake in terms of either starch and sweet cravings (see Figure 2, green) - and that despite the fact that their diet had a similar energy deficit as it was used by Jakubowicz et al (see Figure 3, red).
Other "craving news": Even though it seems unlikely that the results of studies in obese, sedentary individuals translate to avid gymrats, it is worth noting that Rocha et al. have recently demonstrated that moderate intensity activity will decrease total food cravings, specific cravings of high-fat foods, fast-food fats, and carbs (moreover, there was a trend with a large effect size for cravings of sweets (p = 0.052, d = −0.86) to be lower after the exercise intervention) within the same 12-week time frame that was investigated in the meta-analysis at hand - albeit not in the obese, but in healthy, normal weight, untrained young men (Rocha 2016).
Another study that's worth mentioning comes from School of Medicine, The University of Leeds. It found that resisting food cravings is rewarded with significantly increased weight loss - in view of the fact that even the meta-analysis at hand shows that dieting will only reduce the cravings, this finding should remind us that it's not necessarily the severity of the craving, but the ability to resist that's relevant; and the effects of dieting on the latter has not been studied, yet.
Bottom line: The overall message of Kahathuduwa is that a mere reduction in food intake will not increase your general and/or special food cravings for sweet, high-fat, high starch and/or fast foods after 12 weeks of dieting. While this sounds awesome, there are some caveats we must not forget: the effects are (a) only small to medium, you cannot expect to no longer have to think about chocolate or a burger if you're dieting; the meta-analysis (b) didn't investigate cravings in the first weeks of dieting - the latter may be significantly elevated or at least not reduced - this is also supported by the negative correlation of study duration and food cravings observed in the study at hand (i.e. longer diet = reduced cravings);  the benefits (c) have been observed exclusively in obese individuals, and can be expected to differ significantly in dieting athletes, who (i) don't have the (fat) resources of an obese individual to draw on and (ii) don't see the same improvements in glucose management with every pound they lose; lastly, there's (d) the potential effect of limiting both, the energy intake and the intake of certain foods/macronutrients that could nullify if not reverse the beneficial effects of dieting.

With respect to (d), i.e. the effects of restricting certain food groups and macronutrients, it's, however, worth pointing out that the conclusion that this study would prove that low-carbing will increase your hunger for sweets and starchy foods is unwarranted. This assumption is supported only by one study in the meta-analysis and this study did not compare classic low- vs. high-carb diets (cf. previous elaborations) - both, the study by Martin et al that was part of the meta-analysis and a more recent study by Heimowitz et al., I discussed in March 2017, clearly refute the assumption that carbohydrate reductions would trigger cravings for sweets/starches... what we shouldn't forget, however, is that this study was also conducted in obese, inactive individuals and is thus of questionable relevance for athletes | Comment
  • Jakubowicz, Daniela, et al. "Meal timing and composition influence ghrelin levels, appetite scores and weight loss maintenance in overweight and obese adults." Steroids 77.4 (2012): 323-331.
  • Joyner, Michelle A., Ashley N. Gearhardt, and Marney A. White. "Food craving as a mediator between addictive-like eating and problematic eating outcomes." Eating behaviors 19 (2015): 98-101.
  • Kahathuduwa, C. N., et al. "Extended calorie restriction suppresses overall and specific food cravings: a systematic review and a meta‐analysis." Obesity Reviews (2017).
  • Martin, Corby K., et al. "Change in Food Cravings, Food Preferences, and Appetite During a Low‐Carbohydrate and Low‐Fat Diet." Obesity 19.10 (2011): 1963-1970.
  • Pelchat, Marcia Levin, and Susan Schaefer. "Dietary monotony and food cravings in young and elderly adults." Physiology & behavior 68.3 (2000): 353-359.
  • Rocha, Joel, et al. "Effects of a 12-week aerobic exercise intervention on eating behaviour, food cravings, and 7-day energy intake and energy expenditure in inactive men." Applied Physiology, Nutrition, and Metabolism 41.11 (2016): 1129-1136.
  • Rosenbaum, Michael, et al. "Leptin reverses weight loss–induced changes in regional neural activity responses to visual food stimuli." The Journal of clinical investigation 118.7 (2008): 2583.
  • Weingarten, Harvey P., and Dawn Elston. "The phenomenology of food cravings." Appetite 15.3 (1990): 231-246.
  • White, Marney A., et al. "Development and validation of the food‐craving inventory." Obesity 10.2 (2002): 107-114.