Diet quality and mortality: Epidemiological lessons for prevention.

Abstract

In the current issue of the Journal of Internal Medicine, Zhao et al. provided new evidence supporting the importance of diet quality in preventing premature deaths, as shown in a large US population of ∼371,000 men and women aged 50–71 years, followed for ∼23 years [6]. Carbohydrate and fat qualities are known to be a key determinant of health outcomes, but in this prospective study, it is shown in the context of both low and high carbohydrate intake. The effects of a low-carbohydrate diet (LCD) versus a low-fat diet (LFD) for the management of obesity and type 2 diabetes have been extensively studied in short-term trials during the last two decades, but the role of diet quality in these diets has not been sufficiently addressed in trials [1]. There is, however, now clear evidence that diet quality (e.g., fiber-rich carbohydrates) is a key determinant of health outcomes [2]. Nevertheless, moderate lowering of either total fat or carbohydrates has been shown to be more or less equally effective in reducing energy intake, promoting weight loss, and reducing cardiometabolic risk factors in obese individuals with or without diabetes at a 1–2-year basis [1]. Both LFD and LCD are also recommended for people with diabetes, but it should be emphasized that diet quality should be satisfactory for both LCD and LFD and without extreme (either high or low) intake of carbohydrates [3]. With increasing popularity, however, parts of the healthy population may also want to adopt LCD patterns, and questions have been raised regarding the long-term safety and efficacy of these diets [4]. Moreover, there has been concern that not all LFDs may be beneficial, especially if they increase the amount of refined carbohydrates and sugars instead of the preferred fiber-rich, high-quality carbohydrates [2]. In the short-term perspective of obesity management, a negative energy balance is the key, and dietary quality is of minor importance. However, it is unknown from the available randomized controlled trials whether LCD results in reduced mortality also after weight loss has been achieved and during weight stability or weight gain. Thus, both the long-term efficacy and safety of LCD and LFD are of high interest, and more research is needed into the types of foods used in these diets. Further, evidence from long-term observational studies in the general population indicates that low (<40% total energy) and high (>70% total energy) intake of carbohydrates are associated with greater premature mortality—but, importantly, the carbohydrate and fat qualities were not taken into account in those studies [5]. Therefore, the study by Zhao et al. [6] is of high interest and supports a favorable role of diet quality in both LCD and LFD. The LCD and LFD dietary patterns were defined as “healthy” or “unhealthy” based on a dietary quality score (e.g., taking into account different types of carbohydrate rich-foods, saturated and unsaturated fats, and plant or animal based protein). Overall, the results suggest more favorable associations between LFD and LCD and mortality at healthier versions of both these dietary patterns [6]. Thus, higher quality of both diets was a consistent determinant of better survival. If the dietary patterns of the LCD and LFD contained more healthy food sources of carbohydrates, fats, and protein—typically from plant-based foods—the mortality risk was reduced compared with dietary patterns of LCD and LFD that contained more saturated fat, more refined, low-quality carbohydrates, and more animal-based proteins [6]. These results are of clear interest, but not completely novel. More than a decade ago, Fung et al. demonstrated that plant-based LCDs were associated with lower mortality, whereas an animal-based LCD instead was associated with increased mortality [7]. Similar relationships have been observed more recently, albeit in smaller cohorts in comparison—for example, in a population-based cohort of ∼37,000 US adults [8] showing associations that were in line with the results by Zhao et al. That study suggested that the associations of LCD and LFD with mortality seemed to depend on the quality and food sources of macronutrients. Overall LCD and LFD scores were not associated with total mortality, whereas both unhealthy LCD and LFD were associated with higher total mortality [8]. Conversely, healthy LCD and LFD were associated with lower total mortality [8]. In the considerably larger study by Zhao et al. [6], overall LCD and unhealthy LCD were both associated with higher mortality, whereas all LFD patterns were associated with lower mortality. However, it is worth noting that both diets were associated with lower mortality if the dietary pattern was healthy with high diet quality, although the healthy LFD showed a greater mortality reduction than the healthy LCD (18% vs. 5% reduction, respectively) [6]. Thus, this new finding suggests that it may be even more important to have high diet quality if adhering to an LCD rather than an LFD—that is, a more plant-based LCD low in saturated fat [9]. Apart from being the largest study so far on this topic, the study by Zhao et al. included middle-aged and elderly individuals, a highly relevant target group at high risk of chronic disease [6]. In addition, they performed substitution analyses showing that isocaloric replacement of 3% energy from saturated fat with other macronutrient subtypes was associated with lower all-cause mortality, and that replacing low-quality carbohydrates with unsaturated fat and plant protein was associated with reduced mortality [6]. This highlights the importance of dietary fat quality, and that partial replacement of saturated fat with unsaturated fat (e.g., polyunsaturated fats from vegetable oils, seeds, and nuts) could contribute to the reduced risk. If diet quality is absent in any dietary pattern, it is not likely to be effective in prevention. For example, this was partly learned from the randomized Women's Health Initiative (WHI) study, a long-term intervention that reduced all types of fat (including unsaturated) and increased not only high-quality carbohydrates. However, when linking carbohydrate quality (e.g., increasing fiber-rich carbohydrates) using new biomarker-calibrated analyses in the WHI cohorts, the benefits on cardiovascular disease risk were more apparent, as this risk tended to increase with higher animal-protein diets but was reduced with plant-based, higher fiber diet [10]. Overall, plant-based, fiber-rich diets were associated with relatively low chronic disease risk in postmenopausal US women using such analyses [10], thus in accordance with the data from Zhao et al. [6]. Note that the current results from Zhao et al. do not inform about the adherence to the LCD and LFD dietary patterns over time, as diet quality was assessed only once at baseline (self-reported dietary intake during the preceding year). However, as this population was middle-aged and elderly, it is likely that the participants were fairly stable with regard to their dietary habits and lifestyle, at least when compared with younger populations that might be more prone to testing different diets and changing their lifestyle habits. Finally, the results may not be representative for very-LCDs, as the overall carbohydrate content at baseline in this population was not in that low range [6]. Dietary guidelines worldwide recommend high-quality diets (e.g., the Mediterranean diet) in which the foods appear more important than the relative amounts of the total macronutrients. Overall, healthy dietary patterns contain more whole foods and fewer refined foods, and they commonly include a significant amount and variety of plant-based foods rich in unsaturated fat, dietary fiber from whole grains, fruits, and vegetables but are limited in saturated fat, red and processed meat, salt, added sugars, and refined carbohydrates including added sugars from soft drinks. Thus, the current study by Zhao et al. supports international dietary guidelines and adds to the accumulating evidence suggesting a key role of diet quality, which indirectly suggests that successful prevention of premature deaths in the general population is partly dependent on the consumption of high-quality foods with adequate quality of fat, protein, and carbohydrates. In summary, although the current observational data cannot replace findings from randomized controlled trials and should not be interpreted as causal, they do provide useful information that overall accords with the previous literature. The results also suggest that individuals and patients may need clear guidance when choosing to adhere to an LCD or LFD pattern. Clearly, it is time to communicate diets in terms of diet and food quality rather than merely of quantity or composition of total macronutrients. For LCD, it may be particularly important to choose plant-based foods—including vegetable fats and proteins—to ensure that the diet is favorable also in the long-term perspective. The author has no conflict of interests to declare.