Putting the Artificial Sweetener Analysis into Context by Focusing on the Experimental Data in the Meta-Analysis

If you seriously believe that Classic Coke was healthier than Diet Coke or Coke Zero, you cannot be helped, anyway. If you agree that's bullshit, you may want to read the rest of this article to learn what RCTs say about artificial sweeteners.

If we disregard the "associations" in the notoriously unreliable observational studies and focus solely on the randomized trials that are involved in the latest meta-analysis of the health effects of artificial sweeteners, even the scientists' own conclusion doesn't support what the "press release copy+paste" articles at ScienceDaily & Co seem to suggest to the average reader who will ignore the important difference between association and causation and understand that artificial sweeteners cause "weight gain and increased risk of obesity, diabetes, high blood pressure and heart disease, ..." when he skims (often just parts of) articles like the one ScienceDaily published on July 17, 2017.
It is, however, also true that the average reader of these articles will not understand that there's a practically relevant difference between "being associated" and causing "weight gain and increased risk of obesity, diabetes, high blood pressure and heart disease, ..."

Figure 1: People see through the futility of suspecting a causal link between sales and mustache-length. Confirmation bias, however, makes them fail to see that the same logic must be applied to the link between sweeteners and obesity.

It would be unfair, however, to blame the stupidity of the average reader for the confusion. With all the previous flawed reporting about a causal link between artificial sweetener use, obesity, and metabolic disease, it is, after all, hardly surprising that people are able and willing to understand the futility of the previously illustrated link between mustache-length and sales, but ignore the difference between correlation and causality when it comes to the use of artificial sweetener - and let's be honest: if you read the press coverage, 99.9% of the authors show the same confirmation bias that plagues the objectivity of the average Joes and Janes who unreflectively share their bogus articles on Facebook and Twitter.

How can you be so sure that there's no causal relationship between mustaches and sales?

In all honesty, you can't be 100% certain. Why's that? Well, in contrast to the issue of non-nutritive sweeteners (NNSs), obesity and metabolic disease, there are no RCTs available that refute the claim that long mustaches are driving the increase in sales that's illustrated in Figure 1. It is thus haphazard to abuse the meta-analysis by Azad et al. to 'prove' that "routine intake of nonnutritive sweeteners may be associated with increased BMI and cardiometabolic risk" (Azad 2017).
It is true that Azad et al. conclude, based on the results of a thorough search of MEDLINE, Embase and the Cochrane Library (inception to January 2016) for randomized controlled trials (RCTs) that evaluated interventions for nonnutritive sweeteners and prospective cohort studies that reported on consumption of non-nutritive sweeteners (NNS) among adults and adolescents, that there's a link between the primary outcome of the study, i.e. body mass index (BMI), as well as secondary outcomes including weight, obesity and other cardiometabolic end points and the subjects' NNS consumption.

Table 1: Randomized controlled trials that evaluated nonnutritive sweetener interventions and long-term cardiometabolic health - ASB = artificially sweetened beverage, BMI = body mass index, F = female, HOMA-IR = homeostatic model assessment for insulin resistance, M = male, NNS = nonnutritive sweetener, SD = standard deviation (Azad 2017).

What the average anti-sweetener rant forgets to mention, though, is that none (I repeat "zero") of the 7 randomized controlled trials (1003 participants; median follow-up 6 mo) and thus no experimental evidence supports the notion that this link is a causal one - or, in other words, that artificial sweeteners are to blame for the increased weight and cardiometabolic risk in people who consume aspartame, saccharine, sucralose and co. frequently.

Table 2: If you're on the anti-BMI bandwagon, the sign. reductions in the subject waistlines in the three RCTs that actually assessed the effects of artificial sweeteners may be of interest to you (suppl. material in Azad 2017).

Now, before you start crying "bias", let me point you to the fact that only three of the 7 previously illustrated RCTs were sponsored by the industry (Blackburn => Nutrasweet; Tate => Nestlé; Peters => American Beverage Association). Problems exist, however, with blinding both participants and personal to the treatment. After all, artificial sweeteners don't taste like sugar - especially if you want to investigate the isolated effects of a single sweetener without mixing them in order to improve the taste.

Diet Sodas Contain More Than 5 Non-Sweetening Ingredients that May Mess W/ Your Glucose Management | learn more

Next to the type of sweetener that was used, the high heterogeneity of the 7 RCTs may be ascribed to (a) the weight status of the participants (mostly obese, some only overweight), and (b) the study duration which may explain why the 12-month trial showed a significant reduction in waist circumference.

The two 6-month interventions, on the other hand, did not find a significant effect on waist circumference and another 6-month trial failed to identify effects on the subjects' percentage of body fat.

Waist reductions only in industry-funded studies? Yes, but...

While it must not be forgotten, that there was also (c) an effect of the previously discussed industry funding on the significance of the results, with a statistically significant reductions in waist circumference being observed only in the industry-funded studies, you should be aware of the fact that the corresponding non-funded study by Madjid et al. (2015) did not reflect regular sweetener use as it can be observed in the general population. Instead of making non-nutritively sweetened beverages one of the main sources of hydration in the subjects, the scientists had their obese participants consume only a single serving of a wantonly undefined "diet beverage" with only one meal (lunch) on only 5 days of the week. For the rest of the day, both groups were asked not to drink DBs or water during the meal and also not add low-calorie sweeteners such as aspartame or sucralose to beverages such as tea or coffee. You don't really think that's an appropriate study design, do you?

Figure 2: Changes in body weight, waist circumference & hunger in the course of the 12-week study (Peters. 2014)

The previously discussed study by Peters et al. (2014), on the other hand, may be industry funded, with an NNS consumption of at least 24 ounces (>710ml/d) per day (vs. water only), it is yet unquestionable that it mirrors the real-world consumption patterns of artificially sweetened beverages much more accurately than the study by Madjid et al.

Figure 3: Overview of the differences in the outcomes and study design (wrt to how efficiently the studies simulated a realistic intake scenario) of the RCT that focussed on weight loss (Madjid 2016, Peters 2014, Tate 2012, Blackburn 1998).

That neither Azad, not the various "science news" outlets address important details like this is typical for a debate of which I don't say that it was settled, yet. In fact, I would be happy if someone finally conducted a series of blockbuster studies in both healthy, lean and sick, obese individuals which show that artificial a reasonable and realistic use of artificial sweeteners does, in fact, increase your risk of increased BMI and cardiometabolic risk. As of now, this evidence is not available - in fact, the majority of the experimental data suggests that the epidemiologically observed association between NNS intake, obesity & co is a corollary, not a causative one.
References:

• Azad, et al. "Nonnutritive sweeteners and cardiometabolic health: a systematic review and meta-analysis of randomized controlled trials and prospective cohort studies." Canadian Medical Association Journal 189.28 (2017): E929 DOI: 10.1503/cmaj.161390
• Blackburn, George L., et al. "The effect of aspartame as part of a multidisciplinary weight-control program on short-and long-term control of body weight." The American journal of clinical nutrition 65.2 (1997): 409-418.
• Ferri, Letícia AF, et al. "Investigation of the antihypertensive effect of oral crude stevioside in patients with mild essential hypertension." Phytotherapy Research 20.9 (2006): 732-736.
• Hsieh, Ming-Hsiung, et al. "Efficacy and tolerability of oral stevioside in patients with mild essential hypertension: a two-year, randomized, placebo-controlled study." Clinical therapeutics 25.11 (2003): 2797-2808.
• Madjd, Ameneh, et al. "Effects on weight loss in adults of replacing diet beverages with water during a hypoenergetic diet: a randomized, 24-wk clinical trial." The American journal of clinical nutrition 102.6 (2015): 1305-1312.
• Peters, John C., et al. "The effects of water and non‐nutritive sweetened beverages on weight loss during a 12‐week weight loss treatment program." Obesity 22.6 (2014): 1415-1421.
• ScienceDaily. Canadian Medical Association Journal. "Artificial sweeteners linked to risk of weight gain, heart disease and other health issues." ScienceDaily, 17 July 2017. <www.sciencedaily.com/releases/2017/07/170717091043.htm>.
• Tate, Deborah F., et al. "Replacing caloric beverages with water or diet beverages for weight loss in adults: main results of the Choose Healthy Options Consciously Everyday (CHOICE) randomized clinical trial." The American journal of clinical nutrition 95.3 (2012): 555-563.