By Jeffrey T. Jensen, MD, MPH, Editor

SYNOPSIS: Participants from the Women’s Health Initiative (WHI) study who reported consumption of more than two soft drinks per day showed a higher incidence of hip fracture compared to those consuming no soft drinks.

SOURCE: Kremer PA, Laughlin GA, Shadyab AH, et al. Association between soft drink consumption and osteoporotic fractures among postmenopausal women: The Women's Health Initiative. Menopause 2019;26:1234-1241.

Epidemiologic studies have demonstrated an association between the consumption of carbonated soft drinks and low bone density and fracture in children and young adults, but the relationship is less clear for postmenopausal women.

In this secondary analysis of data from the Women’s Health Initiative (WHI) study, researchers evaluated the association between soft drink consumption and osteoporotic fractures in postmenopausal women. The authors noted that more than 8.9 million osteoporotic fractures occur annually worldwide, and the incidence of hip fracture is expected to reach 2.6 million by 2025, nearly doubling to 4.5 million by 2050. Given the high morbidity (40% of people with hip fractures are unable to walk independently one year after the fracture) and mortality (21.2% at one year) of hip fracture in elderly populations, modifiable factors that affect this risk deserve study. In this study, the term soft drink refers to carbonated, nonalcoholic beverages.

The authors used data collected from women enrolled in the observational study of the WHI. They sent annual questionnaires to a group of 72,342 eligible women, who provided information on soda intake (total, caffeinated, and caffeine-free) during the sixth follow-up year. These questionnaires included questions about various health behaviors and outcomes, including hospitalizations and clinical fractures. Bone mineral density (BMD) measurements were available only for a subgroup of these women.

When possible, the investigators used hospital records to validate the self-report of hip fracture, a primary outcome. To evaluate the main exposure of total soft drink intake, a validated dietary questionnaire specifically asked about consumption of common carbonated, nonalcoholic beverages by their brand names, with caffeinated and caffeine-free soft drinks categorized separately. For the analysis, the investigators categorized intake of soft drinks as none; up to 2; 2.1-5; 5.1-14; and more than 14 servings per week.

The investigators used a variety of statistical models to adjust minimally and fully for baseline characteristics, including age, race/ethnicity, body mass index (BMI), hormone therapy, smoking status, alcohol use, oral contraceptive use, self-rated general health status, diabetes mellitus, osteoporosis therapy, thiazides, glucocorticoids, thyroid hormone, calcium and vitamin D supplements, coffee intake, total energy expenditure, income, and maternal hip fracture. They then constructed hazard ratios (HRs) and 95% confidence intervals (CI).

A total of 72,342 women contributed 700,388 woman/years of follow-up, with a median duration of 11.9 years. Hip fracture was reported by 2,578 women (3.5% of the sample). Both the minimally and fully adjusted survival models showed a 26% increased risk of hip fracture among women who drank on average 14 or more soft drink servings per week compared with no servings (HR, 1.26; 95% CI, 1.01-1.56). Of interest, the HRs associated with consumption of 14 or more servings per week of non-caffeinated soft drinks remained significantly elevated in most of the models, while those for caffeinated beverages did not. However, the strength of the association for both types of beverages increased in a subgroup analysis that only considered fully adjudicated hip fractures (overall: HR, 1.45; 95% CI, 1.07-1.95; caffeine-free: HR, 1.43; 95% CI, 0.97-2.11; and caffeinated: HR, 1.48; 95% CI, 1.00-2.26).

The authors concluded that high soda consumption has a modest impact on increasing the risk of fracture in postmenopausal women.


We know that low BMD and hip fractures are predictable consequences of menopause, and that hormone therapy prevents postmenopausal bone loss, maintains or improves BMD, and prevents fractures.1,2 So why focus on this study of soft drink exposure?

If you have followed my commentaries, you know my concern for studies that show statistically significant but clinically unimportant relationships. At first glance, this paper illustrates this type of study. Given the large number of participants, this secondary analysis yielded statistically significant CIs for some outcomes. But the magnitude of the risk estimates (less than 2.0) reflects a weak association that in most cases should not influence practice patterns. Well-designed double-blind, randomized studies do provide greater evidence for validity of a weak association. The randomized estrogen-only and combined estrogen/progestin study arms of the WHI provided this degree of resolution. In contrast, the observational arm of the WHI did not involve randomization. Women enrolled in this study comprised a cohort based on providing answers to a questionnaire on soft drink exposure during year six of the observation period.

Since patterns of soft drink consumption were evaluated as an exposure, and not an experimental condition, the opportunity for confounding exists. To account for this, the investigators used several models of statistical adjustment for the analysis and considered several baseline characteristics, such as BMI, race, and smoking as potential confounders. In general, the result of an increase in the risk of fracture associated with consumption of two or more soft drinks per day was durable across the several models of adjustment. As expected, increasing the number of adjustments widened the CIs, with many comparisons losing statistical significance (e.g., CI includes 1.0).

Although the potential for unmeasured confounding variables always remains, let’s assume that the point estimates supporting the main conclusion that soft drink consumption increases the risk of fracture by 26% (HR, 1.26; 95% CI, 1.01-1.56) reflects truth. In the main WHI randomized studies, women assigned to conjugated equine estrogens (CEE) plus medroxyprogesterone acetate (MPA) had increased HRs of 1.24 (95% CI, 1.01-1.53) for breast cancer and 1.18 (95% CI, 0.95-1.45) for coronary heart disease, and a decrease fracture risk of 33% (HR, 0.67; 95% CI, 0.47-0.95) compared to placebo. Considering the importance attributed to risk estimates of this magnitude in the main WHI findings, the magnitude of the impact of soda consumption (a modifiable risk) on fracture risk (26%) seems more compelling.

A contrarian would argue that none of these relationships necessarily has clinical importance. One woman’s modifiable risk is another’s elixir of life. For example, premenopausal exposure to alcohol does not influence bone density or fracture risk,3 but it does increase breast cancer risk.4 Also, good news for caffeine in this study (no increase in risk).

I believe this paper provides additional useful information when counseling women on the risks and benefits of menopausal hormone therapy. The clinician should guide the discussion to consider lifestyle choices important to individual women. Putting risk into perspective requires a translation into real numbers of events.5 Among other outcomes from the WHI main findings, the attributable risk associated with CEE/MPA was nine additional cases of breast cancer and six additional cases of coronary heart disease in 10,000 exposed women. At the same time, CEE/MPA prevented six cases of fracture and six cases of colorectal cancer per 10,000 women.6 So, raise your glass of wine, mug of beer, or can of soda. Life is full of risks and benefits. Lucky are those who get to choose.


  1. Rossouw JE, Anderson GL, Prentice RL, et al. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: Principal results from the Women's Health Initiative randomized controlled trial. JAMA 2002;288:321-333.
  2. Anderson GL, Limacher M, Assaf AR, et al. Effects of conjugated equine estrogen in postmenopausal women with hysterectomy: The Women's Health Initiative randomized controlled trial. JAMA 2004;291:1701-1712.
  3. Tudor-Locke C, McColl RS. Factors related to variation in premenopausal bone mineral status: A health promotion approach. Osteoporos Int 2000;11:1-24.
  4. Hvidtfeldt UA, Tjonneland A, Keiding N, et al. Risk of breast cancer in relation to combined effects of hormone therapy, body mass index, and alcohol use, by hormone-receptor status. Epidemiology 2015;26:353-361.
  5. Jensen JT, Trussell J. Communicating risk: Does scientific debate compromise safety? Contraception 2012;86:327-329.
  6. Manson JE, Chlebowski RT, Stefanick ML, et al. Menopausal hormone therapy and health outcomes during the intervention and extended poststopping phases of the Women's Health Initiative randomized trials. JAMA 2013;310:1353-1368.