Sugar-sweetened or artificially sweetened beverage consumption, physical activity and risk of type 2 diabetes in US adults

Abstract

Aims/hypothesis

A positive association between sugar-sweetened beverages (SSBs) and diabetes risk has been shown, with inconsistent evidence between artificially sweetened beverages (ASBs) and diabetes. Moreover, it is uncertain if physical activity can mitigate the negative effects of these beverages on diabetes development. Therefore, we aimed to evaluate the independent and joint associations between SSB or ASB consumption and physical activity on the risk of type 2 diabetes.

Methods

We followed 64,029 women in the Nurses’ Health Study (1980–2016), 88,340 women in the Nurses’ Health Study II (1991–2017) and 39,436 men in the Health Professionals Follow-up Study (1986–2016). SSB and ASB consumption was calculated from food-frequency questionnaires administered every 4 years, while physical activity data were collected biennially. A validated supplementary questionnaire on diabetes symptoms, diagnostic tests and treatment confirmed type 2 diabetes cases. Multivariable Cox proportional hazards regression models were used to calculate HRs and 95% CIs for developing type 2 diabetes.

Results

During 5,105,351 person-years of follow-up, we recorded 19,940 new cases of type 2 diabetes. Compared with those who never or rarely consumed SSBs or ASBs, those who consumed ≥2 servings/day had a 41% (HR 1.41 [95% CI 1.33, 1.50]) and 11% (HR 1.11 [95% CI 1.07, 1.16]) higher risk of type 2 diabetes, respectively. For participants meeting physical activity guidelines (≥7.5 metabolic equivalent of task [MET] h/week) and consuming ≥2 servings/week of SSBs or ASBs, the risk was 22% (HR 1.22 [95% CI 1.15, 1.29]) and 7% (HR 1.07 [95% CI 1.02, 1.12]) higher, respectively, compared with those who met physical activity guidelines and never or rarely (<1 serving/month) consumed these beverages. For participants meeting the physical activity guidelines and consuming 1–4 servings/month of SSBs, there was a 9% (HR 1.09 [95% CI 1.02, 1.15]) higher risk of type 2 diabetes. Compared with the reference group (those who met physical activity guidelines and consumed <1 SSB serving/month), adults who did not meet physical activity guidelines (<7.5 MET h/week) and who never or rarely (<1 serving/month) consumed SSBs, had 1–4 SSB servings/month, or had ≥2 SSB servings/week, the HRs (95% CIs) were 1.22 (1.13, 1.31), 1.28 (1.20, 1.37), and 1.51 (1.43, 1.61), respectively. Similarly, for ASB consumption, adults who did not meet physical activity guidelines and who never or rarely (<1 serving/month) consumed ASBs, had 1–4 servings/month, or had ≥2 servings/week, the HRs (95% CIs) were 1.21 (1.14, 1.28), 1.21 (1.13, 1.30), and 1.30 (1.23, 1.37) compared with the reference group (who met physical activity guidelines and consumed <1 ASB serving/month).

Conclusions/interpretation

Even when individuals were physically active, a higher consumption of SSBs or ASBs was associated with a higher risk of type 2 diabetes. Meeting physical activity guidelines reduced the impact of SSB and ASB consumption on diabetes risk, underscoring the need to promote physical activity as part of lifestyle modifications to lower diabetes incidence.

Graphical Abstract

Supplementary Information

The online version contains peer-reviewed but unedited supplementary material available at 10.1007/s00125-024-06351-w.

Introduction

Diabetes is a public health problem, with lifestyle changes and nutritional strategies, such as reducing the consumption of sugar-sweetened beverages (SSBs), being important for its prevention [1]. SSBs are the prime source of added sugars in the American diet and high consumption of these beverages is associated with a higher risk of diabetes, both via body weight gain and other metabolic pathways [2, 3]. Artificially sweetened beverages (ASBs), such as diet soft drinks, contain few to no calories and are often consumed as a substitute for SSBs. ASBs have been suggested as an initial replacement for SSBs to help frequent SSB consumers transition to healthier alternatives, such as water, and may help in diabetes management [4]. However, the evidence on the long-term health effects of ASBs, including the risk of developing diabetes, is inconsistent [2, 5–8].

Physical activity plays an important role in the prevention and management of diabetes [9]. A recent literature search on the relationship between physical activity and incident diabetes indicated an inverse association (26% risk reduction [95% CI 20%, 31%]) among those who accumulated ~225 min/week (or 30 min/day) of physical activity vs sedentary individuals. Moreover, individuals who reached twice this amount of physical activity had a greater risk reduction (36% [95% CI 27%, 46%]) [9]. However, it is unknown whether higher levels of physical activity could compensate for the harmful health effects of SSB or ASB consumption. Hence, this study aims to examine the independent and joint associations between the consumption of SSBs or ASBs and physical activity with the risk of type 2 diabetes in women and men from three large prospective cohort studies.

Methods

Details regarding the methodology of this study are comparable to those in our recently published paper that focused on the joint associations between SSB or ASB consumption, physical activity and the risk of cardiovascular disease [10]. Thus, a brief description of the methodology is given below, with further details included in the electronic supplementary material (ESM) Methods.

Study population and design

The Nurses’ Health Study (NHS) and the Nurses’ Health Study II (NHSII) are prospective cohort studies of 121,701 and 116,430 US-registered female nurses, respectively. The NHS was initiated in 1976 and the NHSII in 1989. The Health Professionals Follow-up Study (HPFS) is a prospective cohort study of 51,529 US male health professionals, established in 1986.

We defined baseline as the first year in which SSB intake data were collected via the food-frequency questionnaires (FFQs; https://nurseshealthstudy.org/participants/questionnaires; www.hsph.harvard.edu/hpfs/hpfs-questionnaires/): 1980 for the NHS, 1991 for the NHSII and 1986 for the HPFS. After excluding participants diagnosed with chronic disease at baseline, with missing exposure information, with implausible energy intake data, and those who died before the study baseline or had missing records, our final analysis included 152,369 women (n=64,029 from the NHS and n=88,340 from the NHSII) and 39,436 men (ESM Fig. 1). The protocol was approved by the institutional review board of Brigham and Women’s Hospital and Harvard T.H. Chan School of Public Health (both Boston, MA, USA).

Dietary assessment and SSB and ASB intake

Dietary intake was measured by FFQs, which were administered every 4 years. The FFQs asked participants how often, on average, they consumed a standard portion of specific foods or beverages (i.e. one standard glass, bottle or can), with possible responses including: never or less than once per month, 1–3 times per month, one time per week, 2–4 times per week, 5–6 times per week, one time per day, 2–3 times per day, 4–5 times per day and ≥6 times per day. SSBs were defined as caffeinated colas, caffeine-free colas, other carbonated SSBs and noncarbonated SSBs (fruit punches, lemonades or other fruit drinks). ASBs were defined as caffeinated, caffeine-free and noncarbonated low-energy or diet beverages.

Physical activity assessment

Physical activity was assessed biennially, with participants reporting the mean time per week spent partaking in various moderate- or vigorous-intensity leisure-time activities in the preceding year. For each reported activity, the metabolic equivalent of task (MET) per h was calculated based on intensity and duration, and then it was quantified as the weekly sum of all activities (MET h/week).

Ascertainment of type 2 diabetes

The primary outcome was the incidence of confirmed type 2 diabetes. Cases of type 2 diabetes were identified by biennially mailed questionnaires (https://nurseshealthstudy.org/participants/questionnaires; www.hsph.harvard.edu/hpfs/hpfs-questionnaires/) and confirmed by a validated supplementary questionnaire, which collected data on diabetes symptoms, diagnostic laboratory test results and use of glucose-lowering therapy (please refer to the ESM file for a copy of the 2010 Supplementary Diabetes Questionnaire used for the NHS and the HPFS, and the 2010 Supplementary Diabetes Questionnaire used for the NHSII).

Assessment of covariables

Information on participant characteristics, lifestyle factors and medical history was obtained via self-report in biennial questionnaires (https://nurseshealthstudy.org/participants/questionnaires; www.hsph.harvard.edu/hpfs/hpfs-questionnaires/). BMI was calculated from body weight, which was reported on each follow-up questionnaire, and height, which was reported at study initiation and ascertained at baseline. Demographic and ancestry information was also collected, with the self-reported categories for ancestry including Southern European/Mediterranean, Caucasian/Scandinavian, Other Caucasian, African American, Hispanic, Asian, American Indian and Other. Since these data were collected more than 30 years ago, the categories may not be consistent with the current classifications. A modified Alternative Healthy Eating Index (AHEI) score, with SSBs and alcohol removed, was used as an indicator of overall diet quality, with a higher score (range 0–90) denoting better diet quality [11].

Statistical analyses

Person-years of follow-up for each participant were calculated from the return of the baseline questionnaire to the date of diagnosis of type 2 diabetes, death or end of follow-up (30 June 2016 for the NHS; 30 June 2017 for the NHSII; and 31 January 2016 for the HPFS), whichever came first. Multivariable Cox proportional hazards regression models were used to estimate HRs and 95% CIs for developing type 2 diabetes according to the independent associations of SSB intake, ASB intake, and physical activity, as well as the joint associations of SSB or ASB intake and physical activity.

To represent the long-term intake of SSBs and ASBs, the primary analysis utilised cumulative averages of data on SSB and ASB consumption, created using repeated measures from the FFQs (calculated by taking the mean intake from all FFQs up to the beginning of the follow-up 4 year questionnaire interval). The same approach was used for the time-varying covariables, such as alcohol intake, total energy intake and dietary factors. Time-varying variables, such as smoking status, physical activity and BMI, were included as repeated measurements in the models. If data were missing at a given time point, we used data from the previous cycle.

For the independent association analysis between SSB or ASB intake or intake of their subtypes (cola, non-cola carbonated beverages and punch for SSBs; diet cola and other diet carbonated beverages for ASBs) and risk of type 2 diabetes, frequency of consumption was collapsed into five categories: (1) never or less than one serving/month (corresponding to the original FFQ answer of ‘never or less than once per month’; used as the reference); (2) 1–4 servings/month (merges ‘1–3 times per month’ and ‘1 time per week’ FFQ answers); (3) 2–6 servings/week (merges ‘2–4 times per week’ and ‘5–6 times per week’ FFQ answers); (4) 1–<2 servings/day (corresponds to ‘1 time per day’ FFQ answer and any intake ≤2 servings/day after the cumulative average calculation); and (5) ≥2 servings/day (merges ‘2–3 times per day’, ‘4–5 times per day’ and ‘≥6 times per day’ FFQ answers, and any intake more than ‘≥6 times per day’ after the cumulative average calculation). The physical activity quintiles for the independent pooled analysis were as follows: 0≤3.0 METs, >3.0≤8.4 METs, >8.4≤16.5 METs, >16.5≤31.7 METs and >31.7 METs. Linear trends were evaluated using the Wald test on a continuous variable representing the median intake of each category. We also evaluated the association for one serving increase of SSB or ASB intake/day in the multivariable models. For the independent association between physical activity and type 2 diabetes, physical activity was categorised in quintiles as well as a continuous variable.

For the joint association analyses between SSB or ASB, physical activity and risk of type 2 diabetes, physical activity was dichotomised to <7.5 MET h/week or ≥7.5 MET h/week. We further collapsed the last three categories of SSB and ASB intake since most participants consumed <2 servings/week, leaving the following categories: (1) never or less than one serving/month (reference); (2) 1–4 servings/month; and (3) ≥2 servings/week (merges the ‘2–6 servings/week’, ‘1–<2 servings/day’ and ‘≥2 servings/day’ categories of the five-level SSB and ASB independent associations analysis). These beverage intake categories were combined with the dichotomised physical activity variable. The reference categories were consuming SSBs or ASBs never or having less than one serving/month and meeting physical activity guidelines (≥7.5 MET h/week).

We also ran a series of sensitivity analyses to assess the robustness of the results, such as focusing on heavy SSB and ASB consumers (top 25% of SSB or ASB consumers vs bottom 75%), exclusion of BMI from models, considering fruit juices in the SSB definition, and joint association analysis with SSB or ASB and tertiles of physical activity, to name a few. Further details can be found in the ESM Methods. Statistical tests were two-sided and p values <0.05 were considered statistically significant. Data were analysed with the SAS package, version 9.4 (SAS Institute, Cary, NC, USA).

Results

Characteristics of the study participants

A total of 19,940 new type 2 diabetes cases were documented (8866 in the NHS, 7505 in the NHSII and 3569 in the HPFS) during a follow-up of 5,105,351 person-years and overall median of 15 years, corresponding to up to 36 years (median 17.2 years), 26 years (median 14.0 years) and 30 years (median 9.3 years), respectively. Adults who met physical activity guidelines and consumed fewer than one SSB serving/month were observed to have the highest diet quality, including a high intake of fruits and vegetables, high intake of whole grains and high AHEI, as compared with adults who met physical activity guidelines but consumed higher amounts of SSBs, or adults who did not meet physical activity guidelines, regardless of frequency of SSB consumption (ESM Table 1). This was also observed in adults who jointly met the physical activity guidelines and consumed 1–4 ASB servings/month or who jointly met the physical activity guidelines and consumed ≥2 ASB servings/week at baseline, as compared with adults who met physical activity guidelines and never consumed ASBs or had less than one serving per month, or adults who did not meet physical activity guidelines, regardless of frequency of ASB consumption (ESM Table 2). ESM Table 3 shows baseline participant characteristics stratified by categories of SSB and ASB intake and quintiles of physical activity.

Independent associations and risk of type 2 diabetes

ESM Table 4 shows the independent associations of SSB or ASB intake and physical activity with type 2 diabetes. In the pooled analysis, after adjusting for known confounders, participants who consumed ≥2 servings/day of SSBs or ASBs had a 41% (pooled HR 1.41 [95% CI 1.33, 1.50]; p_trend <0.0001) and 11% (pooled HR 1.11 [95% CI 1.07, 1.16]; p_trend <0.0001) higher risk of type 2 diabetes, respectively, compared with those who never or rarely consumed these beverages. Comparing the highest with the lowest quintile of physical activity, the HR of type 2 diabetes was 0.71 (95% CI 0.67, 0.74; p_trend <0.0001). In the SSB subtype pooled analysis, per serving increment of non-cola carbonated drinks, punch and cola intake was associated with a 16–22% higher risk of type 2 diabetes (pooled HR 1.16 [95% CI 1.10, 1.23] for non-cola carbonated beverages; pooled HR 1.16 [95% CI 1.11, 1.22] for punch; and pooled HR 1.22 [95% CI 1.18, 1.27] for cola beverages; all p_trend <0.0001; ESM Table 5). The results of the ASB subtype analysis are shown in ESM Table 5.

Subgroup analyses

Significant associations for risk of type 2 diabetes according to one serving increment of SSB intake/day were observed between SSB intake and smoking status, SSB intake and age, ASB intake and age, and ASB intake and BMI in the pooled analysis (ESM Table 6). Cohort-specific results are shown in ESM Table 6.

Joint associations and risk of type 2 diabetes

Table 1 shows the cohort-specific and pooled HRs for type 2 diabetes for the joint associations of SSB or ASB intake and physical activity. Compared with the reference group (those who met physical activity guidelines and consumed <1 serving/month of SSBs), participants who did not meet physical activity guidelines and consumed ≥2 servings of SSBs/week had a pooled HR of 1.51 (95% CI 1.43, 1.61) for type 2 diabetes. Those who met physical activity guidelines but had a higher SSB intake (≥2 servings/week) also showed an increased risk of type 2 diabetes (pooled HR 1.22 [95% CI 1.15, 1.29]). Compared with the reference group (those who met physical activity guidelines and consumed <1 serving/month of ASBs), participants who did not meet physical activity guidelines and did not consume ASBs had a 21% higher risk of type 2 diabetes (pooled HR 1.21 [95% CI 1.14, 1.28]). For those who did not meet physical activity guidelines and consumed ≥2 servings of ASBs/week, the risk was slightly higher (pooled HR 1.30 [95% CI 1.23, 1.37]). There was no evidence of a statistically significant multiplicative interaction between SSB or ASB consumption, physical activity and risk of type 2 diabetes (ESM Table 7). However, the additive interaction between ASB intake, physical activity and risk of type 2 diabetes was significant in the pooled analysis (p=0.002; ESM Table 7). The joint associations between the intake of SSB or ASB subtypes, physical activity and type 2 diabetes risk are reported in ESM Table 8.

Table 1.

Joint associations between frequency of SSB or ASB intake, physical activity^a and risk of type 2 diabetes in three large US cohorts

Beverage type/cohort	Meets US physical activity guidelines			Does not meet US physical activity guidelines
	Never or <1 serving/month	1–4 servings/month	≥2 servings/week	Never or <1 serving/month	1–4 servings/month	≥2 servings/week
SSB intake
T2D cases/person-years
NHS	581/227,480	1521/424,906	2260/517,395	618/145,247	1347/258,900	2539/395,673
NHSII	595/296,950	1136/462,074	2013/705,294	530/127,779	1028/195,738	2203/391,821
HPFS	451/150,121	737/222,686	1299/370,093	209/46,887	278/59,390	594/106,900
HR (95% CI)b
NHS	1.00	1.15 (1.04, 1.26)	1.36 (1.23, 1.49)	1.27 (1.13, 1.42)	1.29 (1.17, 1.43)	1.60 (1.46, 1.76)
NHSII	1.00	1.05 (0.95, 1.16)	1.19 (1.09, 1.31)	1.15 (1.02, 1.30)	1.27 (1.15, 1.41)	1.43 (1.30, 1.57)
HPFS	1.00	1.05 (0.93, 1.18)	1.09 (0.97, 1.22)	1.23 (1.04, 1.45)	1.28 (1.10, 1.49)	1.53 (1.34, 1.73)
Pooled	1.00	1.09 (1.02, 1.15)	1.22 (1.15, 1.29)	1.22 (1.13, 1.31)	1.28 (1.20, 1.37)	1.51 (1.43, 1.61)
ASB intake
T2D cases/person-years
NHS	1229/423,774	703/197,000	2430/549,007	1349/316,427	674/128,516	2481/354,877
NHSII	541/314,917	406/190,736	2797/958,664	608/175,611	373/83,299	2780/456,428
HPFS	500/208,270	512/159,179	1475/375,453	256/73,344	188/41,099	637/98,734
HR (95% CI)b
NHS	1.00	1.00 (0.92, 1.10)	1.02 (0.95, 1.10)	1.18 (1.09, 1.28)	1.15 (1.04, 1.26)	1.22 (1.14, 1.31)
NHSII	1.00	1.02 (0.89, 1.16)	1.07 (0.97, 1.17)	1.22 (1.09, 1.37)	1.20 (1.05, 1.37)	1.28 (1.16, 1.40)
HPFS	1.00	1.17 (1.03, 1.33)	1.19 (1.07, 1.32)	1.30 (1.11, 1.51)	1.45 (1.22, 1.71)	1.61 (1.43, 1.82)
Pooled	1.00	1.05 (0.98, 1.12)	1.07 (1.02, 1.12)	1.21 (1.14, 1.28)	1.21 (1.13, 1.30)	1.30 (1.23, 1.37)

Total sample size: N=191,805 (NHS, n=64,029; NHSII, n=88,340; HPFS, n=39,436)

^aPhysical activity was dichotomised to meets US Physical activity guidelines or does not meet US Physical activity guidelines of ≥150 min/week of moderate–vigorous-intensity physical activity (corresponds to ≥7.5 MET h/week or <7.5 MET h/week, respectively)

^bHRs and 95% CIs were derived from a multivariable model adjusted for age, race, ancestry, smoking status, alcohol intake, multivitamin use, postmenopausal status and menopausal hormone use (in women), aspirin use, family history of diabetes, energy intake, BMI and AHEI. For AHEI, SSB and alcohol intake were removed from the index calculation

T2D, type 2 diabetes

Sensitivity analyses

The pooled HRs for risk of type 2 diabetes in ‘heavy SSB consumers’ (individuals in the top 25th percentile) who met and did not meet the physical activity guidelines were 1.17 (95% CI 1.12, 1.22), and 1.44 (95% CI 1.38, 1.50), respectively, compared with those who were not in the top 25th percentile for SSB consumption who met physical activity guidelines (Table 2). Similar findings were observed for ‘heavy SSB or ASB consumers’ who were in the top 25th percentile for physical activity level (‘high physical activity level’) vs those who were not in the top 25th percentile for SSB/ASB consumption and jointly not in the top 25th percentile for physical activity level (ESM Table 9). Findings relating to the impact of diet, BMI and fruit juice intake on independent and/or joint associations of SSB or ASB consumption and physical activity on type 2 diabetes are discussed in the ESM Results and ESM Tables 10–12. In addition, the impact of adjusting for socioeconomic status, and for dietary factors instead of AHEI are also discussed in the ESM Results. The joint association analysis between SSB or ASB consumption and physical activity in tertiles is shown in ESM Table 13 and discussed in the ESM Results.

Table 2.

Joint associations between SSB or ASB intake, physical activity^a and risk of type 2 diabetes in non-heavy and heavy SSB/ASB consumers in three large US cohorts

Beverage type/cohort	Meets US physical activity guidelines		Does not meet US physical activity guidelines
	Non-heavy consumers	Heavy consumersb	Non-heavy consumers	Heavy consumersb
SSB intake
T2D cases/person-years
NHS	3119/902,619	1243/267,162	2980/571,337	1524/228,483
NHSII	1731/759,024	2013/705,294	1558/323,517	2203/391,821
HPFS	1157/358,273	1330/384,627	473/101,335	608/111,841
HR (95% CI)c
NHS	1.00	1.29 (1.21, 1.38)	1.18 (1.12, 1.24)	1.50 (1.40, 1.60)
NHSII	1.00	1.15 (1.08, 1.23)	1.19 (1.11, 1.27)	1.38 (1.29, 1.47)
HPFS	1.00	1.03 (0.95, 1.12)	1.22 (1.10, 1.37)	1.44 (1.30, 1.60)
Pooled	1.00	1.17 (1.12, 1.22)	1.19 (1.14, 1.23)	1.44 (1.38, 1.50)
ASB intake
T2D cases/person-years
NHS	2964/874,253	1398/295,528	2996/603,014	1508/196,806
NHSII	1493/735,538	2251/728,780	1485/354,834	2276/360,503
HPFS	983/358,402	1504/384,499	434/111,266	647/101,911
HR (95% CI)c
NHS	1.00	1.03 (0.97, 1.10)	1.17 (1.11, 1.23)	1.25 (1.17, 1.33)
NHSII	1.00	1.08 (1.01, 1.15)	1.21 (1.12, 1.30)	1.28 (1.20, 1.37)
HPFS	1.00	1.10 (1.01, 1.19)	1.27 (1.13, 1.43)	1.48 (1.34, 1.64)
Pooled	1.00	1.06 (1.02, 1.11)	1.19 (1.14, 1.24)	1.30 (1.25, 1.36)

Total sample size: N=191,805 (NHS, n=64,029; NHSII, n=88,340; HPFS, n=39,436)

^aPhysical activity was dichotomised to meets US Physical activity guidelines or does not meet US Physical activity guidelines of ≥150 min/week of moderate–vigorous-intensity physical activity (corresponds to ≥7.5 MET h/week or <7.5 MET h/week, respectively)

^bHeavy SSB/ASB consumers were classified as individuals with intake of these beverages in the top 25th percentile

^cHRs and 95% CIs derived from a multivariable model adjusted for age, race, ancestry, smoking status, alcohol intake, multivitamin use, postmenopausal status and menopausal hormone use (in women), aspirin use, family history of diabetes, energy intake, BMI and AHEI. For AHEI, SSB and alcohol intake were removed from the index calculation

T2D, type 2 diabetes

Discussion

In three large prospective cohorts of men or women, we observed that adults who did not meet US physical activity guidelines and more regularly consumed either SSBs or ASBs had a higher risk of type 2 diabetes as compared with those who met physical activity guidelines and rarely consumed SSBs or ASBs. We also identified positive linear associations between SSB and ASB (to a lesser extent) consumption and the risk of type 2 diabetes, while physical activity showed an inverse association. This is the first prospective study to examine the joint associations between SSB or ASB consumption and physical activity on type 2 diabetes risk.

The results on the joint associations of SSB or ASB consumption, physical activity and type 2 diabetes risk are novel, offering evidence contrary to the idea that physical activity could completely counterbalance the potential health risk of consuming SSBs or ASBs. Our study demonstrated that participants not meeting the physical activity recommendations had a higher risk of type 2 diabetes across all levels of SSB and ASB intakes, compared with those who never or rarely consumed these beverages and met physical activity guidelines; yet, those that did meet the physical activity still had a higher risk of type 2 diabetes if they consumed ≥2 servings of SSBs or ASBs per week. The trend persisted in the sensitivity analyses of individuals in the top 25th percentile of SSB and ASB consumers (‘heavy consumers’), with the impact of SSBs and ASBs remaining significant among physically active participants who were classified as heavy consumers of these beverages. This trend was also observed in participants who did not meet the physical activity guidelines. Consistent results were observed in our previous article that focused on cardiovascular disease outcomes [10].

Prospective cohort studies [2, 12] consistently link higher SSB intake in adults to an increased risk of type 2 diabetes, with physical activity considered as a covariate. A recent meta-analysis of 17 studies and 38,253 incident cases of type 2 diabetes found that each additional serving of SSB per day was associated with an RR of 1.18 (95% CI 1.09, 1.28) before BMI adjustment and an RR of 1.13 (95% CI 1.06, 1.21) after adjustment for BMI [2]. In our pooled analysis, we noted an attenuation in the association between SSB intake and type 2 diabetes risk once we removed BMI from the model. Regarding ASBs, the same meta-analysis indicated that one serving/day increase was associated with a 25% higher risk of type 2 diabetes before adjustment for adiposity, while after adjustment for adiposity, the incidence estimate was attenuated [2]. In our study, the results for ASB consumption did not follow a similar pattern to SSB consumption, showing a stronger association between a one serving/day increment in ASB intake and type 2 diabetes risk after removing BMI from the model.

We found a 29% lower risk of type 2 diabetes incidence among those in the highest quintile of physical activity compared with the reference group (those in the lowest quintile of physical activity), after adjusting for sociodemographic characteristics and lifestyle factors. Comparing our findings with other studies is challenging due to the heterogeneity in measurements, type and thresholds used for physical activity. However, a systematic review and meta-analysis focusing on moderate-intensity physical activity and type 2 diabetes (including ten prospective studies and 9367 incident cases of type 2 diabetes) reported an RR of 0.69 (95% CI 0.58, 0.83) for regular moderate-intensity physical activity compared with a sedentary lifestyle [13].

It is noteworthy to highlight that findings from two of our prior studies [5, 6] have shown similar results to this study; we have previously reported that SSB consumption was associated with a significantly higher risk of type 2 diabetes (HR 1.24 [95% CI 1.09, 1.40]; p_trend <0.01) independent of lifestyle factors, yet the association between ASBs and type 2 diabetes was principally explained by health status, pre-enrolment weight change, dieting and BMI [5]. Additionally, the substitution of SSBs with coffee (~237 ml) was associated with a 17% risk reduction of type 2 diabetes [5]. Similarly, we have also reported that increasing total sugary beverage intake (including both intake of SSBs and 100% fruit juices) by >0.50 serving/day over a 4 year period was associated with a 16% (95% CI 1%, 34%) higher diabetes risk in the subsequent 4 years, after adjustment for BMI and initial diet and lifestyle covariates and changes in diet and lifestyle covariates over the study period [6]. Furthermore, we showed that increasing ASB consumption by >0.50 serving/day was associated with 18% (95% CI 2%, 36%) higher diabetes risk.

The strengths of the current study include the large sample size, long follow-up duration, and the repeated measures of diet and lifestyle variables. Potential shortcomings include limited causality and residual confounding, since this is an epidemiological observational study. Also, the exposures were self-reported, which may lead to some degree of measurement error. However, the use of repeated measurements minimises random measurement errors caused by within-person variation. Our study population consisted of primarily middle-aged and well-educated, white health professionals, limiting generalisability. Nonetheless, we believe our findings can be generalised to adult men and women since the suggested biological mechanisms in which SSBs or ASBs impact health, including an inadequate compensatory reduction in solid food consumption in response to the liquid energy provided by SSBs, hyperinsulinaemia due to rapid absorption of glucose, and potential stimulation of the dopaminergic reward system in the brain [14], are unlikely to be sex dependent. In the case of ASBs, the intense sweetness of artificial sweeteners has been hypothesised to stimulate appetite and lead to increased energy consumption, as well as the predisposition of ASB consumers to gain weight and develop insulin resistance by altering the gut microbiota [15, 16]. Additionally, when we adjusted our models for socioeconomic status, results were consistent, further supporting that the mechanism by which SSB and ASB consumption affect health is not dependent on sociodemographic characteristics. Finally, we must highlight that our results relating to ASB consumption should be interpreted with caution since reverse causation is a plausible limitation.

In conclusion, consumption of SSBs and ASBs and physical activity are independently associated with the risk of type 2 diabetes. Not meeting physical activity guidelines alongside a greater intake of SSBs or ASBs was associated with a higher risk of type 2 diabetes among three large prospective cohort studies. Since physical activity partially mitigated the risk associated with SSB and ASB intake, this study emphasises the detrimental impact of SSB consumption and low physical activity on type 2 diabetes risk. These findings support public health initiatives aimed at reducing SSB consumption and promoting increased physical activity.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AHEI

Alternative Healthy Eating Index

ASB

Artificially sweetened beverage

FFQ

Food-frequency questionnaire

HPFS

Health Professionals Follow-up Study

MET

Metabolic equivalent of task

NHS

Nurses’ Health Study

NHSII

Nurses’ Health Study II

SSB

Sugar-sweetened beverage

Acknowledgements

The authors thank the participants and staff of the NHS, NHSII and HPFS for their valuable contributions. Some of the data were presented as an abstract at the American Heart Association's Epidemiology and Prevention/Lifestyle and Cardiometabolic Health 2023 Scientific Sessions.

Data availability

The datasets analysed during the current study are available from the corresponding author on request.

Funding

The research reported in this manuscript was supported by National Institutes of Health grants UM1 CA186107, U01 CA167552 and U01 CA176726. Additionally, this work was supported by grant T32 DK007703-26 (LSP), K01HL169414 (LSP) from the National Heart, Lung, and Blood Institute and K01 DK136968 (DEH) from the National Institute of Diabetes and Digestive and Kidney Diseases. Other funding sources include R01DK125803 (DKT), R01DK131753 (SNB), R01DK120560 (SNB) and R01DK125273 (CBE) from the National Institutes of Health, the Thrasher Research Fund (CBE), the New Balance Foundation (CBE), the Arnold Ventures (CBE) and Dairy Farmers of Canada (J-PD-C), and the Novo Nordisk Foundation grant NNF18CC0034900 (MG-F).

Authors’ relationships and activities

LSP has been supported by grants T32 DK007703-26 from the NIH; DKT has been supported by grant R01DK125803 from the NIH and is currently a member of the 2025 US Dietary Guidelines Scientific Advisory Committee, reviewing beverages and cardiovascular disease; SNB has been supported by grants R01DK131753 and R01DK120560 from the NIH and discloses a pending patent entitled ‘system and methods for conducting nutritional health risk assessments and adapting risk analyses based on actual health outcomes’ (provisional application filed on 20 May 2022; US Application No. 63/344,293); WCW is part of the EAT Foundation Advisory Board (unpaid position) and receives royalties for the textbook Nutritional Epidemiology, Oxford University Press; DSL discloses royalties for a book that recommends a carbohydrate-modified diet; CBE has been supported by grant R01DK125273 from the NIH, as well as grants from the Thrasher Research Fund, New Balance Foundation, and Arnold Ventures, and honoraria from the Columbia Cornell Obesity course and the University of Arkansas for Medical Sciences (speaker); J-PD-C has been supported by grants and received payments (research grant review committee) from the Dairy Farmers of Canada. MG-F is supported by a Novo Nordisk Foundation grant NNF18CC0034900. MG-F is the PI of a grant from the International Nut Council (INC). All other authors declare that there are no relationships or activities that might bias, or be perceived to bias, their work.

Contribution statement

LSP formulated the study question and design, performed the statistical analyses, interpreted the results and wrote the complete first draft of the manuscript. DKT, YL, SNB, DEH, J-PD-C, DSL, CBE and WCW revised and edited the manuscript and made substantial contributions to the interpretation of data. FBH and MG-F formulated the study question and design, interpreted the results and edited the manuscript. FBH and WCW obtained funding. YL assisted in statistical analyses and interpretation of data. All authors participated in critical revision of the manuscript and approved the final version of the manuscript. The corresponding author (LSP) and senior author (MG-F) take full responsibility for the work and/or the conduct of the study, had access to the data and controlled the decision to publish.