Highlights

• In a large observational study, women who reported drinking more than two diet drinks per day had higher risks of stroke, heart disease and overall mortality. The association between diet drinks and stroke risk was stronger in obese women and black women.
• Consumption of diet drinks was only assessed once, so changes in intake over time could not be assessed. Low- and no-calorie sweetener intake from foods or single-serving packets was not measured.
• Observational studies cannot prove cause and effect, and there are many limitations to this study that may have influenced the results—including incomplete and error-prone dietary intake collection, reverse causality, inability to differentiate between types of low- and no-calorie sweeteners, and residual confounding.
• This study contributes to the body of evidence on the impact of diet drinks on health outcomes, but results should be interpreted with caution. Long-term randomized controlled trials with better collection of food and beverage intake are needed.

A new study published in the journal Stroke is hot off the presses, receiving media attention for its examination of self-reported consumption of beverages sweetened with low- and no-calorie sweeteners (referred to throughout as “diet drinks”) and their impact on cardiovascular health. When studies like these make headlines, it’s easy for the critical details to get lost in the hype. Here’s our take on the latest research.

How Was the Study Done?

The purpose of this study was to investigate the association of diet drinks with risk for cardiovascular disease outcomes like stroke, coronary heart disease and death. This research was conducted as part of the long-running Women’s Health Initiative Observational Study, which started back in the 1990s to study the health of over 90,000 postmenopausal women ages 50 to 79 at baseline. The current study analyzed the diet drink consumption of 81,714 women using a single food frequency questionnaire (FFQ) that was administered after participants had been enrolled in the study for three years. Participants were asked one question on the topic of diet drinks, using a standard serving size of 12 ounces (the amount in one can of soda): “During the past 3 months, how often did you drink these beverages?” The beverage category included low-calorie, artificially sweetened colas, sodas and fruit drinks but did not mention drinks like coffee or tea with low- or no-calorie sweeteners. In addition, no information about the specific type of low- or no-calorie sweetener the drinks contained was collected. Responses were organized into four categories: “never or less than once per week,” “1 to 4 times a week,” “5 to 7 times a week,” and “2 or more times per day.”

After an average follow-up time of nearly 12 years, researchers confirmed the number of several different types of stroke, coronary heart disease, and death from all causes that occurred in the study population.

What Were the Results?

Most women (nearly two-thirds of the 80,000+ participants) were infrequent consumers (never or less than 1 per week) of diet drinks, with only 5.1 percent of women consuming 2 or more per day (equal to 24 or more fluid ounces daily). Women who consumed more diet drinks tended to be younger, have higher education levels, higher incomes, higher energy intake, lower overall diet quality, lower levels of exercise and a past medical history of diabetes, heart attack or stroke. They were more likely to be overweight or obese, report a history of never or past smoking, and report drinking alcohol.

Compared with the least-frequent diet-drink consumers, women who consumed two or more diet drinks per day had:

• 23 percent increased risk of any type of stroke
• 31 percent increased risk of a stroke caused by a blood clot (also known as an ischemic stroke)
• 29 percent increased risk of developing heart disease
• 16 percent increased risk of dying from any cause

These risks were defined after controlling for many—but not all—factors that could influence the results, including things like age, smoking history or having high blood pressure. Risk of stroke was increased further in obese women and in black women who reported higher diet drink consumption as compared with other racial demographics. White women with higher diet drink intake, however, had an increased risk of coronary heart disease as compared with other demographic groups.

How are Food Frequency Questionnaires (FFQs) Used in Nutrition Research?

FFQs like the one used in this particular study play an important role in gleaning more information about dietary patterns over time. Data from FFQs are often used to assess how these dietary patterns may be associated with certain health outcomes. However, this does not mean that FFQs are perfect. In fact, there are many inherent problems with FFQs, including the fact that many of us do a very poor job of accurately reporting what we eat or drink when asked to do so. There could be several reasons for this discrepancy, including underreporting consumption of foods perceived as “unhealthy,” overreporting consumption of “healthy” foods, miscalculating serving sizes or just truly not remembering. Additionally, when studies like this one ask participants about their eating and drinking habits only once, they are unable to measure changes in consumption over time. For example, if a participant said that they consumed an average of one diet drink per day at the start of the study and then later increased or decreased their average daily consumption, there would be no way to know about this change.

Observational studies using FFQs can be important first steps in building the greater body of scientific understanding, but they are just that—first steps. These types of studies cannot establish cause and effect; rather, they can help inform experimental studies such as randomized control trials—the gold standard of research design. Having both observational and experimental data leads to scientific advancements in nutrition and food science.

Study Strengths and Limitations

Strengths of this study included its large sample size and relatively lengthy follow-up time of more than one decade. Researchers also attempted to reduce the impact of reverse causality by conducting an analysis that excluded women who had been diagnosed with diabetes or cardiovascular disease before the FFQ was administered. Since it’s possible that these women switched from consuming sugar-sweetened beverages (SSB) to diet drinks because of their diagnosis, removing them reduced the potential for blaming diet drinks as the reason for their health condition—when really it was the health condition that helped instigate their consumption of diet drinks. However, authors of an editorial that accompanied the article emphasized that this analysis “cannot entirely exclude the possibility of reverse causality, however, because they did not exclude all participants with prediabetes, overweight and obesity, or the metabolic syndrome.” Each of these conditions are linked to the development of cardiovascular disease over time, and these participants may have also switched to diet drinks in an attempt to lose weight or control blood sugar.

A major limitation of this study is its observational design. We’ve talked about the limitations of observational studies time and time again, with the primary problem being that they can’t prove cause and effect (that is, that diet drinks cause strokes or heart disease). Causal effects can only be tested through randomized controlled trials (RCTs), which assign some people to receive an intervention (in this case, it would be diet beverages) and others to receive either a placebo or other replacement (such as water or a beverage without low-calorie sweeteners) and measure the differences between groups. RCTs of diet drinks and health outcomes have been conducted, but they’re usually too short in duration to determine a link between outcomes that develop over time, such as cardiovascular disease. More long-term RCTs are needed to gather high-quality information.

It’s also critical to highlight the poor methods by which researchers assessed low- and no-calorie sweetener intake. Information on diet beverage consumption was only collected once, from a single question, so there is no way of knowing whether women changed their consumption over time. In addition, as previously mentioned, self-reported dietary intake is prone to errors and inaccuracy from both known and unknown causes. Low- and no-calorie sweeteners are also found in several foods, including yogurts, nutrition bars, sauces and candies, and are commonly added to foods and drinks through use of single-serving packets. No reporting of overall sweetener intake is given, so the impact of sweeteners from both foods and beverages, or their combined effect, cannot be assessed by this study. Lastly, the way that information was collected prohibited differentiation between different types of low- and no-calorie sweeteners, all of which are biologically distinct and may have very different effects on the body. Lumping all sweeteners together doesn’t provide useful information on exactly which sweeteners may have impacted the results and which had no—or even beneficial—effects. Future studies need to do a better job of measuring the duration and patterns of diet beverage use as well as the reasons for consuming them.

Finally, issues related to confounding and generalizability to other populations are important to consider when interpreting the study results. Since this study was conducted in postmenopausal women, the conclusions are not directly applicable to other populations, including men and women younger than 50 years old. And even though the researchers took into account many factors that could influence the relationship between diet drinks and cardiovascular disease, it’s likely that other important factors weren’t brought into the analysis, leaving the door open for residual confounding. For example, weight change over time and dietary habits before the study began were not considered. This is not just our perspective—authors of the editorial also commented that these are “likely important explanatory factors for the observed associations.”

Closing Thoughts

The role of beverages in our diet is a hotly debated topic in nutrition. The American Heart Association (AHA) recently published a science advisory stating that there was not enough scientific research to conclude that diet drinks do—or do not—alter risk factors for heart disease and stroke. The AHA also stated that diet drinks may help replace high-calorie, sweetened beverages, but recommends water as the best choice for a no calorie drink. The 2015 U.S. Dietary Guidelines make a similar recommendation.

While low- and no-calorie sweeteners continue to be studied by independent researchers, they have been thoroughly reviewed and deemed safe by regulatory bodies around the world, including the U.S. Food and Drug Administration and the European Food Safety Authority.

As the editorial authors state, “Continued public health campaigns to reduce consumption of SSB are prudent. Replacing SSB with water is challenging for heavy consumers, however.” Choosing diet drinks instead of full-calorie options is one of many tools consumers can use to help control added sugar intake and keep calories in check—both important components in reducing risk for diet-related disease and maintaining good health.

This article includes contributions by Kris Sollid, RD.