Sweetened beverages and risk of frailty among older women in the Nurses' Health Study: A cohort study.

BACKGROUND: Consumption of sugar-sweetened beverages (SSBs) has been consistently associated with a higher risk of obesity, type 2 diabetes, cardiovascular disease, and premature mortality, whereas evidence for artificially sweetened beverages (ASBs) and fruit juices on health is less solid. The aim of this study was to evaluate the consumption of SSBs, ASBs, and fruit juices in association with frailty risk among older women. METHODS AND
FINDINGS: We analyzed data from 71,935 women aged ≥60 (average baseline age was 63) participating in the Nurses' Health Study (NHS), an ongoing cohort study initiated in 1976 among female registered nurses in the United States. Consumption of beverages was derived from 6 repeated food frequency questionnaires (FFQs) administered between 1990 and 2010. Frailty was defined as having at least 3 of the following 5 criteria from the FRAIL scale: fatigue, poor strength, reduced aerobic capacity, having ≥5 chronic illnesses, and weight loss ≥5%. The occurrence of frailty was assessed every 4 years from 1992 to 2014. During 22 years of follow-up, we identified 11,559 incident cases of frailty. Consumption of SSBs was associated with higher risk of frailty after adjustment for diet quality, body mass index (BMI), smoking status, and medication use, specifically, the relative risks (RRs) and 95% confidence interval (95% CI) for ≥2 serving/day versus no SSB consumption was 1.32 (1.10, 1.57); p-value <0.001. ASBs were also associated with frailty [RR ≥2 serving/day versus no consumption: 1.28 (1.17, 1.39); p-value <0.001]. Orange juice was associated with lower risk of frailty [RR ≥1 serving/day versus no consumption: 0.82 (0.76, 0.87); p-value <0.001], whereas other juices were associated with a slightly higher risk [RR ≥1 serving/day versus no consumption: 1.15 (1.03, 1.28); p-value <0.001]. A limitation of this study is that, due to self-reporting of diet and frailty, certain misclassification bias cannot be ruled out; also, some residual confounding may persist.
CONCLUSIONS: In this study, we observed that consumption of SSBs and ASBs was associated with a higher risk of frailty. However, orange juice intake showed an inverse association with frailty. These results need to be confirmed in further studies using other frailty definitions.

Introduction

Frailty is a geriatric syndrome characterized by a progressive decline in physiological systems and functional reserves that leads to a high risk of falls, disability, hospitalization, and death [1,2]. This syndrome involves functional limitation, unintentional loss of weight, malnutrition and, in many cases, partly results from the synergistic effect of several diseases [3]. Due to the aging of the population, an increasing number of people is expected to suffer this condition in the coming decades [4]. Thus, it is important to identify the determinants of frailty to ensure that older adults not only live longer but also maintain healthier lives as they age.

Research on dietary factors associated with frailty is still limited. Some specific components of the human diet, including fruit, proteins, and micronutrients, are thought to decrease frailty risk when consumed in adequate amounts [5-7]. Moreover, dietary patterns with overall good quality have been associated with lower risk of frailty [8,9]. However, the effects of specific food components in low-quality diets are not clear.

Added sugar intake constitutes a significant portion of the US diet, providing an average percentage of daily energy intake of 13.6% among older adults [10]. In addition, 65% of older adults exceed the 10% maximum recommended by the World Health Organization and the 2015 Dietary Guidelines Advisory Committee [10-12]. The largest contribution to added sugar intake in the American diet is from liquid sources including sugar-sweetened beverages (SSBs) (37.1%) and fruit drinks (8.9%) [10]. Consumption of added sugar from these liquid sources may not suppress the intake of solid foods in subsequent meals and thus, energy balance can be altered toward higher total energy intake and weight gain [13]. On the other hand, consumption of foods and beverages high in added sugar could displace nutrient-rich components of the diet, increasing the risk of malnutrition in the older population. These mechanisms, together with the adverse effect of sugar on inflammation, glucose tolerance, and lipid metabolism [14,15], may partially link sugary beverages to adverse health outcomes including diabetes, heart disease, premature mortality [16,17] and, possibly, frailty. SSBs are often replaced by artificially sweetened beverages (ASBs). About 23.1% of women aged 60 and older in the US population consumed ASBs on a given day during 2009 to 2010 [18]. The effects of these beverages on health are not well established.

We hypothesized that higher consumption of sugary beverages is associated with higher risk of frailty in older adults. Therefore, we investigated the association of SSBs, ASBs, and fruit juices with the risk of frailty in a large population of older women from the Nurses’ Health Study (NHS).

Methods

Ethics statement

The Harvard T.H. Chan School of Public Health and the Brigham and Women’s Hospital Human Subjects Committee Review Board approved the protocol for the study, and participants provided written informed consent. There was no formal prospectively written protocol for the current study. All analyses described below were decided a priori, except for the additional adjustment for physical activity and baseline morbidity, the combined analysis of SSBs and ASBs, and the sensitivity analysis defining the weight loss component as 10% weight reduction, which were suggested by the reviewers.

Study design and participants

The NHS was established in 1976 with the enrollment of 121,700 female nurses aged 30 to 55 years at inception [19]. Participants completed biennial mailed questionnaires to update information on medical history and lifestyle. The follow-up rate was approximately 90% at each follow-up cycle.

Dietary assessment

Dietary intake was assessed using a validated food frequency questionnaire (FFQ) administered every 4 years as described in detail elsewhere [20]. For the current study, we used the FFQs prior to frailty assessment in 1990, 1994, 1998, 2000, 2006, and 2010. In each questionnaire, participants were asked how often on average during the previous year they had consumed the foods specified. A standard portion size and 9 possible responses for the frequency of consumption, ranging from “never, or less than once per month” to “6 or more times per day” were given for each food item. The consumptions of the following beverages were summed as SSBs: caffeinated and non-caffeinated colas (e.g., Coke, Pepsi, and other colas with sugar), other carbonated beverages with sugar (e.g., 7 Up), and noncarbonated sweetened beverages (e.g., Hawaiian Punch, lemonade, and other noncarbonated fruit drinks). In addition, ASBs consisted of caffeinated, caffeine-free, and noncarbonated low-calorie or diet beverages. Fruit juices included orange juice, apple juice or cider, grapefruit juice, prune juice, and non-specified fruit juices. To best represent long-term diet during follow-up and to account for changes in food consumption, we used the cumulative average consumption of these beverages from all available dietary questionnaires from baseline through frailty onset or the end of follow-up [21]. We stopped updating diet information when a participant reported a diagnosis of diabetes during follow-up to exclude changes in sugary beverage consumption as a consequence of this endpoint. Correlation coefficients between FFQs and multiple dietary records for SSBs were 0.84 for colas, 0.36 for non-cola carbonated soft drinks, 0.56 for noncarbonated sweetened beverages, and 0.84 for fruit juice [22].

Nutrient intakes were calculated by multiplying the consumption of each food recorded with the FFQs by its nutrient content, using the US Department of Agriculture database and complemented with information from the manufacturers. Total energy and nutrient intakes were calculated by summing the derived intakes from all foods. Previous research showed that, compared with multiple dietary records, 24-h dietary recalls, and biomarkers of diet, the FFQ provides sufficient information to detect important associations with disease [23,24]. A modified Alternate Healthy Eating Index (AHEI) score was used as an indicator of overall diet quality. This score was calculated based on 10 foods and nutrients that are predictive of chronic disease risk, including fruit, vegetables, nuts and legumes, red and processed meat, whole grains, alcohol, sodium, trans fat, long-chain omega-3, and other polyunsaturated fats, and excluding the item for SSBs consumption [25]. A higher score in the AHEI denotes better diet quality (range 0 to 10).

Frailty assessment

We used the FRAIL scale [26] that includes 5 self-reported frailty criteria: fatigue, poor strength (reduced resistance), reduced aerobic capacity, having several chronic illnesses, and significant weight loss during the previous year. In 1992, 1996, 2000, 2004, 2008, and 2012, the participants completed the Medical Outcomes Study Short-Form (SF-36), a 36-item questionnaire with 8 health dimensions, including physical and mental components [27]. From the SF-36, we assessed the first 3 frailty criteria with the following questions: (1) for fatigue: “Did you have a lot of energy?,” with response options “some of the time” or “none of the time” or with the question “I could not get going,” with response options “moderate amount” or “all of the time”; (2) for poor strength: “In a normal day, is your health a limitation to walk up 1 flight of stairs?,” with responses “yes” or “a lot”; and (3) for reduced aerobic capacity: “In a normal day, is your health a limitation to walk several blocks or several miles?,” with response options “yes” or “a lot.” In addition, the illnesses criterion was assessed from the question “In the last 2 years, have you had any of these physician-diagnosed illnesses?.” We considered that this criterion was met when participants reported ≥5 of the following diseases: cancer, hypertension, type 2 diabetes, angina, myocardial infarction, stroke, congestive heart failure, asthma, chronic obstructive lung disease, arthritis, Parkinson disease, kidney disease, and depression. Finally, because weight of the participants was available only biannually, the weight loss criterion was defined as a ≥5% decrease in the weight reported in a 2-year period before the assessment of frailty. At the end of each follow-up cycle, incident frailty was defined as having ≥3 criteria in the FRAIL scale. The recovery rate of frailty was 14%, 6%, and 1% after respectively 4, 8, and 12 years of follow-up, which indicates that frailty is a stable outcome. Despite the absence of performance-based measures, the FRAIL scale has been shown to be correlated with the Fried scale (r = 0.617, p <0.001) [28], the most widely used scale for frailty assessment, which includes both self-reported and performance-based measures, among older adults in care settings.

Ascertainment of mortality

Deaths were reported by the next of kin, the postal system, or ascertained through the National Death Index. Follow-up for mortality was more than 98% complete [29]. We obtained copies of death certificates and medical records to determine causes of death (classified according to the International Classification of Diseases, Ninth Revision). Death records were reviewed and coded by physicians.

Medical history, anthropometric data, and lifestyle factors

From the 1992 questionnaire, we collected information on age, weight, smoking status, and medication use. This information has been updated on each of the subsequent biennial questionnaires. To calculate body mass index (BMI), we used information on height measured in 1976, when the cohort was initiated; BMI was calculated as weight in kilograms divided by the square of height in meters. Discretionary physical activity was reported as the average time spent per week during the preceding year in specific activities (e.g., walking outdoors, jogging, and bicycling). The time spent in each activity was multiplied by its typical energy expenditure, expressed in metabolic equivalent tasks (METs), and then summed overall activities. Detailed information on the validity and reproducibility of self-reported weight and physical activity has been published elsewhere [30,31].

Statistical analysis

For this analysis, we included women aged ≥60 years at baseline with complete information on the exposure and outcome variables. Women younger than 60 years at baseline in 1992 entered the study when they turned 60 during follow-up. Women with an unreasonably high (>3,500 kcal/d) or low (<500 kcal/d) caloric intake were excluded from follow-up, as well as women identified as frail at analytical baseline, leaving a total of 71,935 women for the analysis. The association between sweetened beverages and frailty occurrence was examined up to 2014.

Participants were classified into 6 groups according to sweetened beverage consumption: never or almost never (reference), 1 to 3 servings per month, 1 serving per week, 2 to 6 servings per week, 1 to 2 servings per day, and 2 or more servings per day. Since orange juice consumption represents 65% of total juices reported, a separate analysis for this beverage and a combination of all other juices was performed. We used cause-specific proportional hazards models [32] to calculate relative risks (RRs), approximated by hazard ratios, and their 95% confidence interval (95% CI) for the association between each category of sweetened beverage consumption and frailty, adjusting for potential confounders updated at each 4-year time period. Person-years were calculated from baseline until the occurrence of frailty, death, or the end of the study period (1 June 2014), whichever came first. The Andersen–Gill (counting process) data structure was used to handle time-varying covariates and left truncation [33]. We stratified the analysis jointly by age in years at start of follow-up and calendar year of each questionnaire cycle. Multivariable models were adjusted for BMI at baseline (<25.0, 25.0 to 29.9, and ≥30.0 kg/m2), baseline physical activity (in quintiles of METs-h/wk), smoking status (never, past, and current with 1 to 14, 15 to 24, and ≥25 cigarettes/day), energy intake (quintiles of kcal/d), alcohol intake (0, 1.0 to 4.9, 5.0 to 14.9, or ≥15.0 g/d), and current medication use (yes/no) including postmenopausal hormone therapy, aspirin, diuretics, beta blockers, calcium channel blockers, angiotensin converting enzyme inhibitors, other antihypertensive medication, statins, and other cholesterol-lowering drugs, insulin, and oral hypoglycemic medication. Medication use was included in the model to address the fact that persons with risk factors for chronic disease are possibly at greater risk of developing frailty, although some over adjustment might exist. Similarly, the inclusion of BMI might account for some over adjustment because weight loss is part of the frailty outcome. In addition, we adjusted for diet quality using the AHEI (quartiles of the score). Because it might cause some over adjustment, baseline diseases (heart disease, diabetes mellitus, and cancer) have been added to a separate model. All models were mutually adjusted for the other types of beverages to obtain estimates for a beverage independent of the other beverages consumed. Physical activity is closely related to the outcome; therefore, analyses have been repeated excluding this variable from the model. Linear trends were evaluated using the Wald test on a continuous variable using median intakes of each category of beverage consumption. The risk of frailty associated with 1 serving/d increment in beverage consumption was also calculated. Moreover, the association between sweetened beverage consumption and each criterion of the FRAIL scale was examined separately.

Stratified analyses were done by age (<70 versus ≥70 y), BMI (<25 versus ≥25 kg/m2), physical activity (below versus above the median), and the AHEI (below versus above the median). Interaction was evaluated using the Wald test on cross-product terms based on beverage intake (continuous variable) and the stratification variable.

In sensitivity analysis, only the most recent measurement of beverage consumption was considered in relation to frailty. Also, analyses among women with 0 frailty criteria at baseline were performed to understand whether the effect of beverage consumption on frailty may differ depending on the baseline frailty status. In addition, an analysis including dietary exposure before baseline measured in 1980, 1984, and 1986 in association with the risk of frailty and 6-, 8-, and 12-year lagged analyses were performed. Although the FRAIL scale includes having several diseases as 1 of the frailty criteria, additional analyses were performed excluding women with diabetes, heart disease, or cancer at baseline or those who developed these diseases during the follow-up to assess the independence of the studied associations from main chronic diseases. To evaluate the FRAIL scale including only those with a more severe weight loss, we have performed analysis in which we defined weight loss as a 10% weight reduction in 2 years.

All statistical tests were 2-sided with a p-value <0.05 and performed using SAS software version 9.4 for UNIX (SAS Institute, Cary, North Carolina, US). This manuscript follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) recommendations (S1 STROBE Checklist) [34].

Results

In Table 1, the age-standardized baseline characteristics of the study participants by categories of sweetened beverages are presented. Compared to women in the lowest category of consumption, those with higher consumption of SSBs or ASBs had higher BMI and were less physically active. By contrast, high fruit juice consumption was associated with lower BMI and more physical activity. Medication use was similar across strata, although the use of insulin and oral hypoglycemic drugs was remarkably high among those women in the highest category of ASBs consumption. Total energy intake increased across the categories of SSBs and fruit juices, whereas the AHEI score and alcohol intake were lower only among those with higher SSB consumption.

Table 1

Characteristics of women at study entry, by categories of sweetened beverages consumption, in the NHS.

	SSBs			ASBs			Total fruit juices
	Never or almost never	1/wk	≥2/d	Never or almost never	1/wk	≥2/d	Never or almost never	1/wk	≥2/d
Participants, n	28,981	10,231	1,074	22,312	7,542	5,329	7,290	8,703	3,143
Mean age, y	62.9 (2.4)	62.6 (2.2)	62.4 (2.1)	63.0 (2.5)	62.7 (2.3)	62.2 (2.0)	63.0 (2.4)	62.5 (2.2)	62.9 (2.5)
BMI, kg/m2	26.7 (5.2)	26.4 (5.0)	27.2 (5.9)	25.2 (4.8)	26.5 (4.8)	28.9 (5.9)	26.6 (5.2)	27.0 (5.3)	25.7 (5.0)
Current smoker, %	10	10	21	15	7	11	13	12	11
Discretionary physical activity, METs-h/wk	20.7 (24.2)	19.8 (23.9)	17.0 (22.6)	19.6 (23.2)	20.5 (22.8)	17.9 (22.7)	19.5 (23.8)	18.5 (21.8)	23.2 (27.8)
Medication useb
Aspirin, %	49	49	48	47	50	50	47	49	50
Postmenopausal hormone therapy, %	36	33	26	32	35	31	36	33	33
Diuretics, %	11	11	12	9	11	14	10	11	12
β-Blockers, %	14	14	14	13	14	16	13	14	15
Calcium channel blockers, %	11	10	11	9	10	13	11	10	11
ACE inhibitors, %	10	10	12	9	10	12	9	10	11
Other blood pressure medication, %	9	8	11	8	9	10	9	9	8
Statins, %	18	18	24	15	18	23	17	19	17
Other cholesterol-lowering drugs, %	4	4	6	3	4	5	4	5	3
Insulin, %	3	1	2	1	1	5	3	2	1
Oral hypoglycemic drugs, %	4	2	5	2	2	8	4	3	3
Cancer, %	6	5	5	6	6	6	6	6	6
Heart disease, %	4	3	4	3	3	5	4	3	4
Diabetes, %	6	2	4	2	3	10	6	4	4
Number of frailty criteria, %
0	75	75	66	76	76	66	74	72	76
1	20	20	26	19	20	26	21	23	19
2	5	4	8	4	4	8	5	5	5
Dietary intake
SSBs, s/d	0.00 (0.00–0.00)	0.18 (0.14–0.21)	2.50 (2.36–2.94)	0.14 (0.00–0.48)	0.07 (0.00–0.23)	0.00 (0.00–0.10)	0.00 (0.00–0.07)	0.07 (0.00–0.21)	0.14 (0.00–0.52)
ASBs, s/d	0.43 (0.05–1.00)	0.14 (0.00–0.64)	0.00 (0.00–0.20)	0.00 (0.00–0.00)	0.17 (0.14–0.21)	2.57 (2.42–3.22)	0.24 (0.00–1.00)	0.28 (0.00–0.87)	0.07 (0.00–0.57)
Fruit juice, s/d	0.43 (0.07–1.00)	0.64 (0.25–1.07)	0.64 (0.16–1.14)	0.64 (0.15–1.07)	0.59 (0.20–1.04)	0.39 (0.07–0.97)	0.00 (0.00–0.00)	0.18 (0.14–0.21)	2.50 (2.14–2.93)
Orange juice, s/d	0.14 (0.00–0.49)	0.36 (0.07–0.79)	0.25 (0.07–0.79)	0.29 (0.07–0.79)	0.32 (0.07–0.79)	0.14 (0.02–0.57)	0.00 (0.00–0.00)	0.07 (0.07–0.14)	1.02 (1.00–2.50)
Other fruit juices, s/d	0.07 (0.00–0.21)	0.14 (0.06–0.42)	0.14 (0.02–0.50)	0.14 (0.00–0.43)	0.14 (0.02–0.35)	0.07 (0.00–0.23)	0.00 (0.00–0.00)	0.07 (0.04–0.14)	1.07 (0.33–1.76)
Energy intake, kcal/d	1,576 (1,301–1,886)	1,771 (1,484–2,101)	2,229 (1,870–2,567)	1,714 (1,409–2,069)	1,691 (1,403–2,024)	1,726 (1,412–2,085)	1,479 (1,208–1,790)	1,596 (1,327–1,925)	2,086 (1,760–2,440)
AHEI score	52.4 (45.9–58.9)	50.0 (43.9–56.1)	43.0 (37.4–49.3)	49.8 (43.0–56.8)	51.6 (45.3–57.8)	47.9 (41.8–54.5)	51.0 (43.6–57.9)	49.5 (43.2–56.0)	51.4 (44.7–57.8)
Alcohol intake, g/d	1.8 (0.0–8.7)	1.5 (0.0–6.5)	0.0 (0.0–2.5)	1.1 (0.0–6.7)	1.5 (0.0–6.7)	0.9 (0.0–6.5)	0.9 (0.0–5.8)	1.2 (0.0–6.5)	0.1 (0.0–6.7)

ACE, angiotensin converting enzyme; AHEI, Alternate Healthy Eating Index; ASB, artificially sweetened beverage; BMI, body mass index; IQR, interquartile range; MET, metabolic equivalent task; NHS, Nurses’ Health Study; s, serving; SD, standard deviation; SSB, sugar-sweetened beverage.

Values are means (SD), dietary intake values are medians (IQR), unless otherwise indicated. Data, except age, were directly standardized to the age distribution of the entire cohort.

a Entry was age 60.

b One or more times per week.

During 22 years of follow-up, we identified a total of 11,559 incident frailty cases among the 71,935 women of this study (Table 2). SSB consumption was associated with higher risk of frailty after adjustment for lifestyle factors and medication use. The RRs (95% CI) across categories of increasing consumption were 1.00, 1.00 (0.95, 1.05), 1.09 (1.03, 1.16), 1.11 (1.05, 1.17), 1.33 (1.21, 1.46), and 1.46 [(1.22, 1.74); p-value <0.001]. Additional adjustment for diet quality and baseline morbidity somewhat attenuated the association. By contrast, ASB consumption was also associated with higher risk of frailty in fully adjusted models [RRs across categories of increasing consumption 1.00, 0.99 (0.93, 1.05), 1.00 (0.93, 1.06), 1.05 (1.00, 1.11), 1.11 (1.04, 1.19), and 1.28 [(1.17, 1.39); p-value <0.001]. Joint analyses showed that higher consumptions of both beverages simultaneously had also a direct association with frailty, in comparison with the lowest consumption of both beverages [RR for highest tertile versus lowest tertile: 1.18 (1.08, 1.29); p-value <0.001].

Table 2

RRs (95% CI) of frailty according to categories of sweetened beverages consumption among 71,935 women aged ≥60 y in the NHS.

	Never or almost never	1/mo to 3/mo	1/wk	2 to 6/wk	1-2/d	≥2/d	P for trend	Per 1 serving/d increase
SSBs
Participants, n	28,981	14,715	10,231	13,492	3,442	1,074
Person-year	372,001	244,222	160,434	210,786	42,745	10,435
Frailty cases, n	3,926	2,604	1,890	2,461	545	133
Age adjusted	1.00	0.97 (0.93, 1.02)	1.09 (1.03, 1.16)	1.17 (1.11, 1.23)	1.52 (1.39, 1.67)	1.98 (1.66, 2.36)	<0.001	1.32 (1.27, 1.38)
Multivariable modela	1.00	1.00 (0.95, 1.05)	1.09 (1.03, 1.16)	1.11 (1.05, 1.17)	1.33 (1.21, 1.46)	1.46 (1.22, 1.74)	<0.001	1.18 (1.13, 1.23)
Multivariable modelb	1.00	0.97 (0.93, 1.03)	1.05 (0.99, 1.11)	1.04 (0.99, 1.10)	1.22 (1.11, 1.34)	1.32 (1.10, 1.57)	<0.001	1.12 (1.07, 1.17)
Multivariable modelc	1.00	0.98 (0.94, 1.04)	1.06 (1.00, 1.12)	1.05 (1.00, 1.11)	1.23 (1.12, 1.35)	1.32 (1.10, 1.57)	<0.001	1.12 (1.07, 1.18)
ASBs
Participants, n	22,312	7,885	7,542	19,542	9,325	5,329
Person-year	324,602	145,224	117,297	292,308	109,672	51,520
Frailty cases, n	3,292	1,614	1,297	3,334	1,294	728
Age adjusted	1.00	1.04 (0.98, 1.11)	1.12 (1.05, 1.19)	1.28 (1.22, 1.35)	1.63 (1.52, 1.74)	2.32 (2.14, 2.52)	<0.001	1.29 (1.26, 1.31)
Multivariable modela	1.00	0.98 (0.93, 1.04)	1.00 (0.93, 1.06)	1.07 (1.01, 1.12)	1.15 (1.07, 1.22)	1.36 (1.25, 1.48)	<0.001	1.11 (1.09, 1.14)
Multivariable modelb	1.00	0.99 (0.93, 1.05)	1.00 (0.94, 1.07)	1.06 (1.01, 1.11)	1.12 (1.05, 1.20)	1.31 (1.20, 1.42)	<0.001	1.10 (1.07, 1.12)
Multivariable modelc	1.00	0.99 (0.93, 1.05)	1.00 (0.93, 1.06)	1.05 (1.00, 1.11)	1.11 (1.04, 1.19)	1.28 (1.17, 1.39)	<0.001	1.09 (1.06, 1.12)
Total fruit juices
Participants, n	7,290	7,534	8,703	25,587	19,678	3,143
Person-year	78,221	103,854	116,549	424,287	280,634	37,077
Frailty cases, n	840	1,216	1,356	4,995	2,840	312
Age adjusted	1.00	0.96 (0.87, 1.04)	0.92 (0.84, 1.00)	0.84 (0.78, 0.90)	0.78 (0.72, 0.85)	0.78 (0.68, 0.89)	<0.001	0.89 (0.86, 0.92)
Multivariable modela	1.00	0.97 (0.89, 1.06)	0.95 (0.87, 1.03)	0.92 (0.85, 0.99)	0.88 (0.81, 0.95)	0.90 (0.79, 1.03)	<0.001	0.96 (0.92, 0.99)
Multivariable modelb	1.00	0.96 (0.88, 1.05)	0.94 (0.86, 1.03)	0.92 (0.85, 0.99)	0.88 (0.81, 0.95)	0.91 (0.80, 1.04)	0.001	0.96 (0.93, 0.99)
Multivariable modelc	1.00	0.97 (0.89, 1.06)	0.94 (0.87, 1.03)	0.92 (0.86, 1.00)	0.88 (0.81, 0.95)	0.91 (0.79, 1.04)	0.01	0.96 (0.93, 0.99)
	Never or almost never	1/mo to 3/mo	1/wk	2 to 6/wk	≥1/d			Per 1 serving/d increase
Orange juice
Participants, n	14,390	13,028	8,696	22,275	13,546
Person-year	174,406	169,201	132,442	391,756	172,817
Frailty cases, n	2,040	1,844	1,572	4,517	1,586
Age adjusted	1.00	0.93 (0.87, 0.99)	0.88 (0.82, 0.94)	0.81 (0.76, 0.85)	0.78 (0.73, 0.83)		<0.001	0.83 (0.80, 0.87)
Multivariable modela	1.00	0.95 (0.89, 1.01)	0.92 (0.86, 0.98)	0.87 (0.82, 0.92)	0.83 (0.77, 0.89)		<0.001	0.89 (0.86, 0.93)
Multivariable modelb	1.00	0.94 (0.88, 1.00)	0.91 (0.85, 0.98)	0.87 (0.82, 0.91)	0.82 (0.76, 0.88)		<0.001	0.89 (0.85, 0.93)
Multivariable modelc	1.00	0.94 (0.89, 1.01)	0.92 (0.86, 0.98)	0.87 (0.82, 0.92)	0.82 (0.76, 0.87)		<0.001	0.89 (0.85, 0.93)
Other juicesd
Participants, n	21,345	15,762	12,954	17,393	4,481
Person-year	268,995	256,046	197,037	272,042	46,502
Frailty cases, n	2,904	2,960	2,218	3,058	419
Age-adjusted	1.00	1.01 (0.96, 1.06)	0.98 (0.92, 1.03)	1.01 (0.96, 1.07)	1.07 (0.96, 1.19)		0.22	1.02 (0.96, 1.08)
Multivariable modela	1.00	1.05 (0.99, 1.10)	1.02 (0.97, 1.08)	1.09 (1.03, 1.15)	1.12 (1.00, 1.24)		0.004	1.07 (1.01, 1.14)
Multivariable modelb	1.00	1.05 (1.00, 1.11)	1.04 (0.98, 1.10)	1.13 (1.07, 1.19)	1.16 (1.05, 1.29)		<0.001	1.11 (1.05, 1.17)
Multivariable modelc	1.00	1.06 (1.00, 1.11)	1.04 (0.98, 1.10)	1.13 (1.07, 1.19)	1.15 (1.03, 1.28)		<0.001	1.10 (1.04, 1.17)

a Adjusted for age (years), calendar time (4-y intervals), BMI (<25.0, 25.0–29.9, ≥30.0 kg/m2), smoking status (never, past, and current 1–14, 15–24, and ≥25 cigarettes/day), alcohol intake (0, 1.0–4.9, 5.0–14.9, or ≥15.0 g/d), energy intake (quintiles of kcal/d), physical activity (quintiles), and medication use (aspirin, postmenopausal hormone therapy, diuretics, β-blockers, calcium channel blockers, ACE inhibitors, other blood pressure medication, statins and other cholesterol-lowering drugs, insulin, and oral hypoglycemic medication).

b Adjusted for variables in model a and additionally adjusted for the AHEI (quartiles).

c Adjusted for variables in model b and additionally adjusted for cancer, heart disease, and diabetes (yes/no). All beverages were mutually adjusted for each other.

d This group includes apple juice or cider, grapefruit juice, prune juice, and non-specified fruit juices

95% CI, 95% confidence interval; ACE, angiotensin converting enzyme; AHEI, Alternate Healthy Eating Index; ASB, artificially sweetened beverage; BMI, body mass index; NHS, Nurses’ Health Study; RR, relative risk; SSB, sugar-sweetened beverage.

Fruit juices were associated with lower risk of frailty [RRs: 1.00, 0.97 (0.89, 1.06), 0.94 (0.87, 1.03), 0.92 (0.86, 1.00), 0.88 (0.81, 0.95), and 0.91 (0.79, 1.04); p-value 0.01]. This inverse association was entirely due to orange juice consumption [≥1 s/d versus never or almost never: 0.82 (0.76, 0.87); p-value <0.001], whereas other types of juices showed a slight positive association [≥1/d versus never or almost never: 1.15 (1.03, 1.28); p-value <0.001]. Excluding physical activity from the models did not change the results.

We found a significant interaction for SSB and orange juice with age; however, the stratified results do not show large differences in estimates for women aged <70 compared to women aged ≥70. Results did not vary strongly across other subgroups in the stratified analyses (Table 3). Additionally, the associations between sweetened beverages and each frailty criterion are shown in Fig 1. Both SSBs and ASBs were associated with a higher risk of all the individual frailty criteria, whereas orange juice was associated with lower risk of the fatigue, poor strength, and reduced aerobic capacity criteria.

Table 3

RRs (95% CI) of frailty according to sweetened beverages consumption (serving/d), stratified by lifestyle factors, among 71,935 women aged ≥60 y in the NHS.

	Person years	Frailty cases	SSBs	P for interaction	ASBs	P for interaction	Total fruit juices	P for interaction	Orange juice	P for interaction	Other juicesa	P for interaction
Age <70	598,149	2,815	1.15 (1.07, 1.23)	0.01	1.07 (1.03, 1.11)	0.27	0.93 (0.87, 0.99)	0.35	0.82 (0.75, 0.90)	0.03	1.10 (0.99, 1.21)	0.53
Age ≥70	442,472	8,744	1.09 (1.03, 1.16)		1.10 (1.06, 1.13)		0.97 (0.93, 1.01)		0.91 (0.86, 0.95)		1.11 (1.04, 1.19)
BMI <25 kg/m2	456,022	4,256b	1.17 (1.08, 1.28)	0.44	1.13 (1.07, 1.19)	0.46	0.99 (0.93; 1.05)	0.10	0.88 (0.82; 0.96)	0.37	1.20 (1.09, 1.32)	0.11
BMI ≥25 kg/m2	507,392	6,560	1.11 (1.05, 1.17)		1.09 (1.06, 1.12)		0.92 (0.88, 0.97)		0.86 (0.82, 0.91)		1.05 (0.97, 1.13)
Low physical activity (<median)	510,848	9,268	1.10 (1.04, 1.17)	0.37	1.08 (1.05, 1.11)	0.59	0.95 (0.91, 1.00)	0.63	0.88 (0.83, 0.93)	0.97	1.10 (1.02, 1.18)	0.42
High physical activity (≥median)	527,805	2,291	1.16 (1.08, 1.25)		1.11 (1.06, 1.15)		0.95 (0.90, 1.01)		0.89 (0.83, 0.95)		1.10 (1.00, 1.20)
Low AHEI level (<median)	517,268	6,846	1.15 (1.09, 1.21)	0.64	1.09 (1.06, 1.13)	0.97	0.9 (0.93, 1.01)	0.34	0.90 (0.85, 0.95)	0.46	1.11 (1.03, 1.20)	0.34
High AHEI level (≥median)	523,354	4,713	1.12 (1.02, 1.24)		1.09 (1.04, 1.13)		0.94 (0.89, 0.99)		0.87 (0.81, 0.94)		1.06 (0.97, 1.16)

Models were adjusted for age (years), calendar time (4-y intervals), BMI (<25.0, 25.0–29.9, ≥30.0 kg/m2), smoking status (never, past, and current 1–14, 15–24, and ≥25 cigarettes/day), alcohol intake (0, 1.0–4.9, 5.0–14.9, or ≥15.0 g/d), energy intake (quintiles of kcal/d), physical activity (quintiles), medication use (aspirin, postmenopausal hormone therapy, diuretics, β-blockers, calcium channel blockers, ACE inhibitors, other blood pressure medication, statins and other cholesterol-lowering drugs, insulin, and oral hypoglycemic medication), AHEI (quartiles), cancer, heart disease, and diabetes, except for the stratification variable. All beverages were mutually adjusted for each other.

a This group includes apple juice or cider, grapefruit juice, prune juice, and non-specified fruit juices.

b The number of events is different because of missing values for BMI.

95% CI, 95% confidence interval; ACE, angiotensin converting enzyme; AHEI, Alternate Healthy Eating Index; ASB, artificially sweetened beverage; BMI, body mass index; NHS, Nurses’ Health Study; RR, relative risk; SSB, sugar-sweetened beverage.

Fig 1

RRs (95% CI) of frailty components according to sweetened beverages consumption (serving/d) among women aged ≥60 y in the NHS.

Adjusted for age (years), calendar time (4-y intervals), BMI (<25.0, 25.0–29.9, ≥30.0 kg/m2), smoking status (never, past, and current 1–14, 15–24, and ≥25 cigarettes/day), alcohol intake (0, 1.0–4.9, 5.0–14.9, or ≥15.0 g/d), energy intake (quintiles of kcal/d), physical activity (quintiles), medication use (aspirin, postmenopausal hormone therapy, diuretics, β-blockers, calcium channel blockers, ACE inhibitors, other blood pressure medication, statins and other cholesterol-lowering drugs, insulin, oral hypoglycemic medication), AHEI (quartiles), cancer, heart disease, and diabetes. All beverages were mutually adjusted for each other. The group “other juices” includes apple juice or cider, grapefruit juice, prune juice, and non-specified fruit juices. 95% CI, 95% confidence interval; ACE, angiotensin converting enzyme; AHEI, Alternate Healthy Eating Index; ASB, artificially sweetened beverage; BMI, body mass index; NHS, Nurses’ Health Study; RR, relative risk; SSB, sugar-sweetened beverage.

When only the most recent information on beverage consumption before the development of frailty was used, we still observed an increased risk of frailty for higher SSBs and ASBs consumption. Orange juice remained inversely associated with frailty (S1 Table). In addition, analysis among the women without frailty criteria at baseline showed similar associations (S2 Table), as well as analysis including cumulative diet information including the time period before baseline, and also latency analysis (S3 and S4 Tables). Finally, when excluding women with heart disease, cancer, or diabetes or when using a weight loss criterion defining weight loss as a 10% weight reduction in 2 years, the association between the intake of sweetened beverages and frailty remained similar (S5 and S6 Tables).

Discussion

In this analysis of a large prospective cohort in the US, we found that habitual consumption of SSBs and ASBs was associated with higher risk of frailty, whereas orange juice was associated with lower risk. The relationships were independent of lifestyle, medication use, and diet quality and remained similar across different subgroups of women.

The association between SSBs and frailty showed a positive association across increasing categories of consumption, especially in the 2 highest categories (above 1 serving a day). Of note is that participants in the NHS had an average SSBs intake of 0.23 (SD 0.41) servings a day, which is lower than the average intake from the nationally representative population of the National Health and Nutrition Examination Survey (NHANES) in the same age category in 2009 to 2010 (0.61 servings a day) [35]. Thus, the excess risk of frailty observed in the NHS may be of particular concern for a large fraction of the older US population with higher levels of SSBs intake.

So far, only 1 previous study has investigated the association between sweetened beverages and frailty. In a cohort of community-dwelling older people from Spain, participants consuming SSBs did not have an increased risk of frailty after 3 years of follow-up when compared with those who never consumed those beverages [36]. Besides its smaller sample size and a short duration of follow-up, another plausible explanation for their results was the very low intake of SSBs observed. Moreover, in the Spanish study, frailty was defined using the Fried criteria, so their results might not be directly comparable with ours.

There is consistent evidence that liquid sources of carbohydrates are associated with less satiety than solid sources and that their intake is not compensated by reduced intake of other foods, so total daily energy intake is increased [13]. This is 1 of the mechanisms that may partially explain the association between SSBs and obesity, diabetes, and cardiovascular disease [16,37]. Frailty may also result from other biological pathways that contribute to the association between SSBs and those diseases. For example, added sugar in SSBs may lead to inflammation, impaired glucose, and lipid metabolism [14,15], leading to the occurrence of several chronic diseases and possibly increasing the risk of frailty. These mechanisms also impair muscle glucose handling and intracellular energy production and reduce protein synthesis, which leads to sarcopenia and less efficient muscle contraction [38-39]. Furthermore, high-fructose corn syrup used in SSBs produces a significant dose–response increase in uric acid concentrations [40], which has been associated with frailty incidence [41]. Our results also suggest that the association between sweetened beverages and frailty was not entirely mediated or due to obesity and other diseases since main analyses were adjusted for BMI, and the sensitivity analyses excluding participants with cardiovascular disease, diabetes, cancer, or overweight still showed a significant direct association.

While other types of fruit juices were not related to a lower risk of frailty, orange juice was inversely associated with risk. Many antioxidant nutrients and bioactive substances (including vitamins, carotenoids, flavonoids, and polyphenols) are found in juices and especially in orange juice. These compounds may limit oxidative stress and inflammation, which are core mechanisms of the decline in muscle function and strength in old people as well as of frailty [42-44]. Flavonoids and the carotenoid beta-cryptoxanthin may also lower the risk of cognitive decline [45]. Our results showed that orange juice was inversely associated with all the individual criteria of frailty, except for the illnesses criterion. Although the criteria mostly reflect physical frailty, cognitive impairment is closely related to physical frailty and might also play a role in the development of the individual criteria used to define frailty in this study [2]. Therefore, the potential beneficial effect of orange juice observed might be partly attributed to an improvement in cognitive status. However, our results need to be confirmed in further studies before public health recommendations can be made.

Similar to our results, other studies have found positive associations between ASBs and several outcomes including mortality, diabetes, and cardiovascular diseases [17,46,47]. Little is known about possible biological mechanisms that could explain these associations. It has been suggested that the potential adverse effects of ASB may be caused by a detrimental effect on gut microbiota that, in turn, may have a negative effect on glucose tolerance [48]. On the other hand, the authors suggested that misclassification and reverse causation could account for the results found [17,46]. In our study, the results for ASBs consumption and frailty held among different subgroups of participants, with healthy and unhealthy lifestyle behaviors.

Frailty is an important outcome because it is the consequence of alterations in many physiological systems. Frailty and pre-frail status as defined by the FRAIL scale has shown to be a significant predictor of disability among older adults [49]. Although reverting frailty development is challenging for many patients, at an early stage, frailty might be reversible and is therefore a valuable tool to identify those at risk for further adverse health effects [50]. Some previous research has shown that several diet-related factors (e.g., Mediterranean diet, fruits, and vegetables) that lower the risk of frailty [9,51-52] also lower the risk of disability [53,54]. Thus, we could speculate that that sweetened beverages might also have a detrimental effect on disability.

Strengths of this study are the large sample size, the repeated diet measurements that allowed calculating cumulative average consumptions, and the use of updated information on covariates in a cohort with high rates of follow-up. However, several limitations need to be acknowledged. First, since dietary information was self-reported, measurement error and misclassification could occur; however, the FFQ used here has been extensively validated against diet records and biomarkers, showing good correlations [20], and the repeated measures reduced random error. Second, although we were able to adjust for many potential confounders, and sensitivity analysis among subgroups of healthy participants showed robustness of the results, some unmeasured and residual confounding cannot be ruled out. Reverse causation cannot be totally discarded; however, latency analyses showed similar associations to main analyses. Third, although studying the risk of frailty among only female nurses helped to increase internal validity, the observed associations might not apply to other populations. Fourth, frailty is a dynamic condition and therefore, potentially reversible. However, as well as other chronic conditions, once it occurs, it is unlikely to reverse. Finally, performance-based measures were not available in this large cohort of older women. Due to the use of self-reported information, misclassification of frailty might have occurred. Our results should be confirmed in studies using other frailty definitions that include more objective measurements [2, 4].

In conclusion, we found that habitual consumption of ≥1 serving/d of SSBs, as well as ASBs, was associated with a higher risk of frailty. By contrast, consumption of orange juice was associated with lower risk of frailty. Whether ASBs consumption has a detrimental effect on frailty or is a spurious finding is unclear as this was not a prior hypothesis, and plausible mechanisms have not been established. Further studies of both SSBs and ASBs in relation to frailty would be valuable. Considering the high intake of SSBs in the US population and its many adverse health effects, possibly including frailty, older adults should be advised to limit their SSBs intake.

Relative risks (95% confidence interval) of frailty according to categories of the most recent information of sweetened beverages consumption before the onset of frailty among 67,739 women.

(DOCX)

Click here for additional data file.

Relative risks (95% confidence interval) of frailty according to categories of sweetened beverages consumption among 57,760 women with 0 frailty criteria at baseline.

(DOCX)

Click here for additional data file.

Relative risks (95% confidence interval) of frailty according to categories of sweetened beverages consumption among 82,430 women aged ≥60 y in the Nurses’ Health Study, including diet before baseline in 1980, 1984, and 1986.

(DOCX)

Click here for additional data file.

Relative risks (95% confidence interval) for latency analysis of frailty according to sweetened beverage intake (serving/d) among women aged ≥60 y in the Nurses’ Health Study.

(DOCX)

Click here for additional data file.

Relative risks (95% confidence interval) of frailty according to categories of sweetened beverages consumption among 60,402 women aged ≥60 y without heart disease, diabetes, or cancer in the Nurses’ Health Study.

(DOCX)

Click here for additional data file.

Relative risks (95% confidence interval) of frailty with the weight loss criteria defined as a 10% weight reduction in 2 years according to categories of sweetened beverages consumption among 72,180 women aged ≥60 y in the Nurses’ Health Study.

(DOCX)

Click here for additional data file.

(DOC)

Click here for additional data file.

19 May 2020

Dear Dr Struijk,

Thank you for submitting your manuscript entitled "Sweetened beverages and risk of frailty among older women in the Nurses´ Health Study" for consideration by PLOS Medicine.

Your manuscript has now been evaluated by the PLOS Medicine editorial staff as well as by an academic editor with relevant expertise and I am writing to let you know that we would like to send your submission out for external peer review.

However, before we can send your manuscript to reviewers, we need you to complete your submission by providing the metadata that is required for full assessment. To this end, please login to Editorial Manager where you will find the paper in the 'Submissions Needing Revisions' folder on your homepage. Please click 'Revise Submission' from the Action Links and complete all additional questions in the submission questionnaire.

Please re-submit your manuscript within two working days, i.e. by .

Login to Editorial Manager here: https://www.editorialmanager.com/pmedicine

Once your full submission is complete, your paper will undergo a series of checks in preparation for peer review. Once your manuscript has passed all checks it will be sent out for review.

Feel free to email us at plosmedicine@plos.org if you have any queries relating to your submission.

Kind regards,

Thomas J McBride, PhD,

PLOS Medicine

6 Jul 2020

Dear Dr. Struijk,

Thank you very much for submitting your manuscript "Sweetened beverages and risk of frailty among older women in the Nurses´ Health Study" (PMEDICINE-D-20-02037R1) for consideration at PLOS Medicine.

Your paper was evaluated by a senior editor and discussed among all the editors here; it was also sent to independent reviewers, including a statistical reviewer. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of these reviews, I am afraid that we will not be able to accept the manuscript for publication in the journal in its current form, but we would like to consider a revised version that addresses the reviewers' and editors' comments. Obviously we cannot make any decision about publication until we have seen the revised manuscript and your response, and we plan to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We expect to receive your revised manuscript by Jul 27 2020 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see http://journals.plos.org/plosmedicine/s/submission-guidelines#loc-methods.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

We look forward to receiving your revised manuscript.

Sincerely,

Emma Veitch, PhD

PLOS Medicine

On behalf of Clare Stone, PhD, Acting Chief Editor,

PLOS Medicine

plosmedicine.org

-----------------------------------------------------------

Requests from the editors:

*Please revise your title according to PLOS Medicine's style - as well as setting out the study question this should also summarise the study design/framework used, after a colon (eg, ": prospective cohort).

*At this stage, we ask that you include a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract. Please see our author guidelines for more information: https://journals.plos.org/plosmedicine/s/revising-your-manuscript#loc-author-summary

*The abstract should have some minor changes to structure, please use the headings Background, Methods and Findings, Conclusions (nb, "Methods and Findings" is a single subsection).

*In the last sentence of the Abstract Methods and Findings section, please include a brief note about any key limitation(s) of the study's methodology.

*Did your study have a prospective protocol or analysis plan? Please state this (either way) early in the Methods section.

a) If a prospective analysis plan (from your funding proposal, IRB or other ethics committee submission, study protocol, or other planning document written before analyzing the data) was used in designing the study, please include the relevant prospectively written document with your revised manuscript as a Supporting Information file to be published alongside your study, and cite it in the Methods section. A legend for this file should be included at the end of your manuscript.

b) If no such document exists, please make sure that the Methods section transparently describes when analyses were planned, and when/why any data-driven changes to analyses took place.

c) In either case, changes in the analysis-- including those made in response to peer review comments-- should be identified as such in the Methods section of the paper, with rationale.

*It's encouraging the paper has been reported using the STROBE guideline - however we'd suggest also appending the completed STROBE checklist as a supporting information file alongside the submitted revised paper - checklist can be downloaded at https://www.equator-network.org/reporting-guidelines/strobe/. When completing the checklist, please use section and paragraph numbers, rather than page numbers.

-----------------------------------------------------------

Comments from the reviewers:

Reviewer #1: In this manuscript, authors investigate the association of SSB, ASB and fruit juice intake and risk of frailty in women 60 years and older, using data from the Nurses' Health Study. They report a statistically significant increased risk of frailty with high consumption of SSB, ASB and fruit juice other than orange juice, and a lower risk of frailty with high consumption of orange juice. The manuscript is well written, and easy to read. There are number of strengths to this analysis: authors use a well-established sizable cohort with good quality data; prospective study design; long follow up; and repeated exposure measurement during follow up. They also conducted a comprehensive set of statistical analyses, including a number of sensitivity analyses. Nonetheless, physical activity variable notably misses from their multivariable models, despite being an important confounder (major limitation). They report from a stratified analysis by PA level (low/high), though this does not seem enough to rule out the confounding effect of PA. I suggest repeating all the models including METh/wk.

Table 3: it would be helpful to see p for interaction. Perhaps a footnote indicating a statistically significant interaction.

-----------------------------------------------------------

Reviewer #2: Struijk et al. present a large prospective epidemiological study on sweetened beverages and risk for frailty. The study has indisputable strengths, including a very large sample size (n=almost 76,000 women), and a long follow-up (up to 22 years) between exposures (comprehensive exposure of most sweetened beverages) and ascertainment of frailty in older age. Very few studies on this topic have been conducted so far, therefore this report is important for the research field.

It is generally acknowledged that frailty has many risk factors with small to moderate effect sizes, and given the aging of the population worldwide, the observed decreased risk for frailty with higher orange juice intake could lead to a strong reduction in the number of cases.

In sum, the paper is well written overall although the way nutritional aspects are discussed should be improved. The authors would benefit from deciphering underlying mechanisms, instead of just mentioning obesity, diabetes, cardiovascular diseases or inflammation, and anti-oxidant properties of fruits to explain their results. A more complicated approach of the food matrix effect of fruit intakes compared with fruit juices intake would have been interesting. Moreover, a discussion about the possible reversibility of frailty would have reinforce the discussion section.

More specific remarks:

Abstract:

The "dose-response manner" mentioned in the conclusion is not ascertained by the retained results of the abstract. Please remove.

Introduction:

Are the artificially sweetened beverages (ASBs) largely consume among USA residents? The prevalence of consumption is only described for sugar-sweetened beverages SSBs and fruit juices.

A mention of the association between higher fruits intake or between carotenoids levels and the lower risk of frailty among European older people should be considered (Garcia-Esquinas et al 2016 and Pilleron et al 2019).

Methods:

Regarding the exposure, could the cumulative consumption of SSBs, ASBs and fruit juices be a relevant analysis? Indeed, some people may be considered both as SSBs consumers, OR ASBs consumers, while they consumed both beverages. I wondered how this cumulative exposure could influence the results.

Regarding the identification of incident frail participants, one out of five criteria is weight loss. I'm surprised to observe that the threshold retained was a reduction of 5% of weight during the last 2 years. How participants lost to follow-up for a visit were classified? -the delay between 2 visits being higher than 2 years? Moreover, why the authors have not considered the GLIM criteria (Cederholm et al 2018) to approach undernutrition (the consensus of experts proposed a reduction of weight of 5% for less than 6 months or 10% for 6 months and over).

Regarding frailty as a whole, it's acknowledged that subjective criteria are used, but it's surprising not to considered physical activity (MET are available) and a more objective scale than the FRAIL one (which also need a clinical judgment form the practitioner which seems not available here). The authors should acknowledge that the scale is a proxy of a more subjective scale and a misclassification could occur. Finally, the concordance between the FRAIL scale and the Fried criteria has been described in a care setting, but not yet among community-dwellers. This is another limitation.

More importantly, how were considered participants with frailty and disability (in instrumental activities of daily living IADL): it may occur an overlap between both frailty and IADL disability, while some authors argued that frailty is not disability (and the scale used to identify frail people should also not identify IADL disabled participants, see Zamudio-Rodríguez A et al, Age Ageing 2020). This important point should be discussed in depths.

Regarding the statistical approach, I identified 2 major limits: the first one is the potential reversibility of the frailty status. It's acknowledged that frailty might be transitory (even more when the duration of the follow-up is high), and then, the identification of frail people who are then censured at the first time they are identified frail could induce a bias. The second limit is the lack of consideration of the risk for death: however, when the duration of the study is so long, the proportion of older people who died during the follow-u is high, and in the present study, the exposure can also be associated with this specific risk of death. Therefore, the analysis is mainly appropriate among the survivors, and the competitive risk for death compared with frailty cannnot be dismissed. To limit this bias, an illness-death analysis should be performed. It would reinforce the results.

Results.

Table 1 provides the description of low, moderate or high consumers of specific beverages, while the statistical part provided data on 6 levels of consumption: how were these 6 different levels rearranged on 3 different levels of intake?

Table 1: is alcohol expressed as g/d? please specify.

The text relative to the description of Table 2 (and others) would benefit from details (for instance, 3 different RR are provided, but it's not easy to understand which exposure is linked to which RR).

The test of interaction (p value) between exposure, BMI … should be provided to ensure the relevance of such stratified analysis.

Discussion.

In addition to the limits already discussed above, the relevance of such results among women only should be addressed.

The discussion about the differences observed between orange juice or other fruits juices intake appears limited and only in the field of cognition. This should be discussed further, because of the public health relevance of such results. I suggest revising with deeper discussion on nutritional aspects.

Overall, it remains minor issues that should be addressed to improve this paper which is already of high quality.

-----------------------------------------------------------

Reviewer #3: The authors present an investigation of associations between SSBs, ASBs and fruit juices, and frailty in the NHS. This is a well written, well argued, coherent manuscript which generates useful results with potential value to public health policy makers. I recommend publication of his paper with only a couple of very minor suggestions for improvement below.

Page 6, para 2: the FFQ is 'reasonably valid' - this is vague; what does reasonably valid mean?

Page 9, line 1: the number of participants included in the study belongs in the results, not the methods.

-----------------------------------------------------------

Reviewer #4: This is a well-conducted study on the association between sweetened beverages and risk of frailty among older women using the Nurses´ Health Study (NHS) cohort. However, there are a few issues needing attention.

1) Competing risk. All the relative risks (RR) in table 2 were approximated by hazard ratios from the Cox proportional hazards models. However, as the outcome is frailty rather than all cause mortality, there is an issue of competing risk (from death). During the 22 years follow-up, potentially many participants (n=?) died. Therefore competing risk analyses need to be performed to derive the true HR/RR.

2) Many baseline variables were adjusted in table 2 as shown in the Multivariable model a and b. However, models were not adjusted for baseline co-morbidities which is not adequate. Although authors argued that the frailty outcome consists of some of the conditions, the baseline co-morbidities are different and need to be adjusted. In sensitivity analyses (supplementary table 5) the authors made some effort to address the issue but it is a bit patchy not systematic. It would be good to have a multivariable model c to additionally adjust these co-morbidities at baseline.

3) Not clear whether SSBs, ASBs, orange juice and etc are mutually exclusive in the questionnaires. What if a person takes two more different types of drinks during the follow-ups? How did the authors adjust this potential overlap in the analyses?

4) Table 1, in the Dietary intake section, most variables appeared to be skewed with non-normal distributions therefore should be summarised as median and IQR rather than mean and SD.

-----------------------------------------------------------

Any attachments provided with reviews can be seen via the following link:

[LINK]

26 Jul 2020

Submitted filename: Response to reviewers_PLOSMed.docx

Click here for additional data file.

13 Oct 2020

Dear Dr. Struijk,

Thank you very much for re-submitting your manuscript "Sweetened beverages and risk of frailty among older women in the Nurses´ Health Study: a prospective cohort study" (PMEDICINE-D-20-02037R2) for review by PLOS Medicine.

I have discussed the paper with my colleagues and the academic editor and it was also seen again by the statistical reviewer. I am pleased to say that provided the remaining editorial and production issues are dealt with we are planning to accept the paper for publication in the journal.

The remaining issues that need to be addressed are listed at the end of this email. Any accompanying reviewer attachments can be seen via the link below. Please take these into account before resubmitting your manuscript:

[LINK]

Our publications team (plosmedicine@plos.org) will be in touch shortly about the production requirements for your paper, and the link and deadline for resubmission. DO NOT RESUBMIT BEFORE YOU'VE RECEIVED THE PRODUCTION REQUIREMENTS.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

In revising the manuscript for further consideration here, please ensure you address the specific points made by each reviewer and the editors. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments and the changes you have made in the manuscript. Please submit a clean version of the paper as the main article file. A version with changes marked must also be uploaded as a marked up manuscript file.

Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. If you haven't already, we ask that you provide a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract.

We expect to receive your revised manuscript within 1 week. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

We ask every co-author listed on the manuscript to fill in a contributing author statement. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

If you have any questions in the meantime, please contact me or the journal staff on plosmedicine@plos.org.

We look forward to receiving the revised manuscript by Oct 20 2020 11:59PM.

Sincerely,

Thomas McBride, PhD

Senior Editor

PLOS Medicine

plosmedicine.org

------------------------------------------------------------

Requests from Editors:

1- With apologies for recommending this in the previous revisions, please remove “prospective” from the subtitle.

2- Thank you for providing your STROBE checklist. Please replace the page numbers with paragraph numbers per section (e.g. "Methods, paragraph 1"), since the page numbers of the final published paper may be different from the page numbers in the current manuscript. Please also refer to the checklist (S1 Checklist) when you mention it in the Methods section, and add the checklist to the list of supplemental files at the end of the main text.

3- Thank you for noting that there was no formal prospectively written protocol and mentioning analyses that were added at the request of reviewers. Should you also note the combined analysis of SSB and ASB intakes, added in response to reviewer 2?

4- Please include the results of the sensitivity analysis for weight loss defined as 10% weight reduction (in response to reviewer 2) in the supplementary information (referenced from the main text).

5- Similarly, please note in the main text and include in supplementary information the assessment of frailty reversal that was also included in response to reviewer 2.

6- Abstract Methods, please note the setting of the Nurses Health Study, and provide some demographic information (eg, average age).

7- In the Abstract Conclusions, please address the study implications without overreaching what can be concluded from the data; the phrase "In this study, we observed ..." may be useful. The last sentence of the Abstract Conclusions could specify what types of further studies are necessary.

8- Author summary, point 2, perhaps, “... an increasing number of people are at risk of developing frailty.”

9- Author summary, point 5, “... while higher orange juice consumption was associated with a lower risk.”

10- Author summary, point 6, no need to limit this to the US population, SSB intake is high in many countries. (Similarly for the Discussion conclusions).

11- Methods could note that participants provided consent as part of enrollment.

12- Methods, Dietary Assessment section. The bracketed sections at the bottom of page 7 (“ [e.g., Coke, Pepsi, and other colas with sugar],”; “ [e.g., 7-Up],” ; “ [e.g., Hawaiian Punch, lemonade, and other noncarbonated fruit drinks] ”)and beginning of page 8 should be parentheticals.

13- Results, page 17, please include the p value for the comparison of highest tertile vs. lowest tertile of ASB and SSBs

14- Please start of the Discussion: "In this analysis of a large prospective cohort in the United States, we found ..."

Comments from Reviewers:

Reviewer #4: Many thanks authors for their great effort to improve the manuscript. I am satisfied with the response and the revision. All my concerns were comprehensively addressed. No further issues needing attention.

Any attachments provided with reviews can be seen via the following link:

[LINK]

29 Oct 2020

Dear Dr. Struijk,

On behalf of my colleagues and the academic editor, Dr. Carol Brayne, I am delighted to inform you that your manuscript entitled "Sweetened beverages and risk of frailty among older women in the Nurses´ Health Study: a cohort study" (PMEDICINE-D-20-02037R3) has been accepted for publication in PLOS Medicine.

PRODUCTION PROCESS

Before publication you will see the copyedited word document (within 5 business days) and a PDF proof shortly after that. The copyeditor will be in touch shortly before sending you the copyedited Word document. We will make some revisions at copyediting stage to conform to our general style, and for clarification. When you receive this version you should check and revise it very carefully, including figures, tables, references, and supporting information, because corrections at the next stage (proofs) will be strictly limited to (1) errors in author names or affiliations, (2) errors of scientific fact that would cause misunderstandings to readers, and (3) printer's (introduced) errors. Please return the copyedited file within 2 business days in order to ensure timely delivery of the PDF proof.

If you are likely to be away when either this document or the proof is sent, please ensure we have contact information of a second person, as we will need you to respond quickly at each point. Given the disruptions resulting from the ongoing COVID-19 pandemic, there may be delays in the production process. We apologise in advance for any inconvenience caused and will do our best to minimize impact as far as possible.

PRESS

A selection of our articles each week are press released by the journal. You will be contacted nearer the time if we are press releasing your article in order to approve the content and check the contact information for journalists is correct. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact.

PROFILE INFORMATION

Now that your manuscript has been accepted, please log into EM and update your profile. Go to https://www.editorialmanager.com/pmedicine, log in, and click on the "Update My Information" link at the top of the page. Please update your user information to ensure an efficient production and billing process.

Thank you again for submitting the manuscript to PLOS Medicine. We look forward to publishing it.

Best wishes,

Thomas McBride, PhD

Senior Editor

PLOS Medicine

plosmedicine.org