Birthing people and their children are vulnerable to acute and long-term health effects of poor diet quality and excess sugar consumption during pregnancy.^{1, 2} Sugar-sweetened beverages (SSBs) are the largest source of added sugar for pregnant individuals, who on average consume 50% more calories from added sugar than recommended.¹ One in five pregnant individuals consumes SSBs at least once per day.³ In the general population, SSB intake is associated with weight gain and risk of obesity, type 2 diabetes (T2D), and cardiovascular disease (CVD).^4–6 Growing evidence also supports the hypothesis that greater maternal SSB intake is associated with elevated risk for a range of perinatal complications, including gestational diabetes mellitus (GDM),^{7, 8} gestational and post-partum weight gain,^{9, 10} hypertensive disorders of pregnancy,^{11, 12} preterm birth,^{13, 14} and abnormal birthweight,^{1, 15} conditions which increase the risk of CVD and T2D later in life.^16–18 Not only are pregnant individuals with GDM at elevated risk of excess gestational weight gain (GWG), but excess GWG also increases risk of neonatal complications, notably being small-for-gestational-age (SGA).^{19, 20}

Population-wide interventions to reduce SSB intake and improve diet among pregnant individuals remain critical.²¹ SSB taxation is one promising policy to reduce SSB consumption,^22–24 sugar intake, and chronic disease risk.²⁵ As of December 2019, seven US cities were levying an excise tax of $0.01-$0.02 per fluid ounce of SSBs. While modeling studies suggest that SSB taxes reduce chronic disease risk in the general population,^{26, 27} empirical studies of the health impacts have proven difficult to conduct, because chronic diseases such as T2D develop over a long latency period. It also remains unclear whether SSB taxes improve the health of pregnant individuals and infants by reducing SSB consumption during pregnancy. Pregnancy is not only a critical period for later-life outcomes, but also a time-limited period during which nutrition policies can have observable health effects.^28–32 Thus, pregnancy is a uniquely valuable context for studying SSB taxes.

The present study examined whether SSB taxes were associated with a reduction in maternal and birth complications among pregnant individuals and offspring, using data from >5 million US births. Leveraging a quasi-experimental difference-in-differences design, this study compared before-after SSB tax changes in perinatal outcomes among pregnant individuals in five SSB-taxed US cities, while accounting for secular trends for pregnant individuals in untaxed comparator cities. This study also investigated subgroup effects of SSB taxes: by race and ethnicity motivated by the racial disparities in sugar intake and maternal and infant health outcomes,^33–38 and by pre-pregnancy body mass index (BMI), maternal age, and parity due to their modifying effects on perinatal risk.^{39, 40} As SSB taxes become increasingly common across the US and globally,²⁵ evidence of their health impacts can inform policy making and community-level interventions targeting maternal and child nutrition and chronic disease.⁴¹

Tax-exposed individuals were more likely than those in comparator cities to be non-Hispanic White or Black, older, educated, to have smoked prepregnancy, and have normal BMI, although these differences were modest (Table 1). Perinatal health was similar for individuals living in intervention and comparator cities prior to SSB tax implementation (Table 2), including GDM prevalence (5.4% in intervention cities vs. 5.5% in comparator cities) and GWG z-score (−1.8 vs. −1.9 SD).

SSB taxes were associated with a 41.4% reduction in GDM (−2.2 percentage points [pp]; 95%CI −4.2, −0.2) and 7.9% reduction in GWG z-score (−0.2 SD; 95%CI −0.3, −0.01) (Table 2). In addition, SSB taxes were associated with a 39.1% reduction in SGA risk (−4.4 pp; 95%CI −6.5, −2.1). There was no association between SSB taxes and other secondary outcomes (Table 2).

Event-study estimates revealed that before-tax differences for primary outcomes between tax-exposed and tax-unexposed individuals were small and not significant, suggesting roughly parallel before-tax outcome trends (Figure 1). Before-tax trends were also similar between tax-exposed and comparator individuals for most secondary outcomes, except birthweight, gestational age, and preterm birth (Appendix Figure 2).

Event-study estimates further highlighted that SSB taxes were associated with decreased GDM risk two to four quarters post-tax implementation, attenuated thereafter (Figure 2). For GWG z-score, the intervention-comparator difference particularly grew in later quarters. For SGA risk, there was a sustained decrease for five quarters post-tax implementation (Appendix Figure 2). Event-study plots for other secondary outcomes are provided in Appendix Figure 2 and discussed in the Appendix.

There were subgroup differences in the estimated associations between SSB taxes and outcomes (Figure 2, Appendix Figure 3). Event studies for some subgroups indicated pre-tax imbalance between intervention and control cities (Appendix Figure 8); results for those subgroups should be interpreted cautiously. Reductions in GDM risk and GWG z-score were more pronounced among Asian individuals. Reductions in GWG z-score were largest among those aged 25–29 and 35+, those with a high school diploma, as well as for multiparous and obese individuals. Subgroup findings are further discussed in the Appendix.

Cohort-specific analyses showed larger improvements in Philadelphia versus other cohorts for GWG z-score (Appendix Table 3, Appendix Figure 4). Philadelphia and Seattle/San Francisco also had larger declines in SGA risk (Appendix Table 1, Appendix Figure 5). All cohorts except Berkeley showed declines in GDM risk, yet none attained statistical significance. Results for other cohort-specific effects are provided in the Appendix (Appendix Figure 5).

Findings from additional sensitivity analyses were similar to the primary analysis but smaller in magnitude for GWG z-score and SGA, yet differed somewhat for GDM (Appendix Tables 1–2, Appendix Figures 6–7). Sensitivity estimates for GDM risk were also more muted and did not attain statistical significance for several models. Placebo analyses showed no significant association between prepregnancy smoking nor marital status and SSB taxes, suggesting time-varying confounding of the estimates is less likely (Appendix Table 2).

This study provides the first estimates, to the authors’ knowledge, of the association between city-level SSB taxes and perinatal outcomes. Using quasi-experimental difference-in-differences methods, SSB taxes were associated with a 41.4% reduction in GDM risk and 7.9% decrease in GWG z-score. GDM and excess GWG are well-studied predictors of birthweight, preeclampsia, and preterm delivery, and even risk of CVD and T2D later in life for the birthing person and child,^{18, 68, 69} highlighting these conditions as important mechanisms through which SSB taxes may affect longer-term health.^{7–9, 53} SSB taxes were also associated with a 39.1% decrease in SGA risk, in line with the improvements in GDM and GWG z-score, which are predictors of SGA.^{19, 20} Sensitivity and placebo analyses supported the robustness of our findings for GWG z-score and SGA, but conflicted somewhat with the GDM findings. Studies that leverage additional datasets with more information on mechanisms and less measurement error of maternal outcomes⁷⁰ may address this knowledge gap. Overall, our results suggest that SSB taxes induce clinically relevant changes that may translate into population-wide health improvements, although the findings should be interpreted with caution.

There were no consistent overall improvements in other maternal and neonatal health outcomes. Further research is needed to investigate the long-term health effects of SSB taxes as more post-period data become available. One plausible explanation for the null findings is that individuals undercut the health benefits of SSB taxes through compensatory caloric intake, although recent literature has tended to find such substitution effects to be limited.²⁴

Given recent evidence that Philadelphia experienced among the largest behavioral effects on SSB purchasing and consumption due in part to its slightly higher SSB tax rate, it was hypothesized that Philadelphia would also see larger improvements in perinatal health, compared to other cities. ^{24, 60} Our cohort-specific findings support this hypothesis, strengthening the case for a causal link between SSB taxes and improvements in perinatal health. Nevertheless, further research leveraging individual-level SSB consumption data would help further test the causal relationship.

There were notable differences in the association of SSB taxes and perinatal health across subgroups. Growing evidence has similarly underscored the role of disparities in perinatal outcomes among these groups.^{39, 71, 72} Differential exposure to SSBs has been driven partly by targeted marketing of SSBs to racial/ethnic minority and low-income communities,^{33–38, 73} contributing to notably higher SSB consumption among Black and Hispanic vs. White women during pregnancy (28% and 33% vs. 15%).³ This highlights the need to disaggregate results by subgroup. Our findings show larger declines in GDM and GWG z-score among non-Hispanic Asian pregnant people exposed to SSB taxes; violations of parallel trends for other racial subgroups precluded the study from inferring the impact of SSB taxes on racial disparities and perinatal health. There were large improvements in GWG z-score across other subgroups, including young (25–29), less educated (high school GED), and overweight individuals, indicating that SSB taxes could produce downstream health benefits, even among structurally disadvantaged subgroups. Additional years of follow-up data, along with larger sample sizes for subgroups of interest, might clarify the persistence of the perinatal effects, including among subgroups.

The present study provides the first quasi-experimental evidence suggesting that SSBs contribute to GDM, excess GWG risk, and SGA, and that policies to reduce excess sugar consumption among pregnant individuals may be an effective strategy to improve perinatal health. These findings are consistent with a recent federal commission suggesting that SSB taxes can reduce T2D risk,⁷⁴ and come at a critical time, as pregnancy complications in the US are highly prevalent and have grown in recent years. In population-based studies, the reported US prevalence for GDM is 6.2% (~230,000 births)⁷⁵ and 47.5% for excess GWG (~1,780,000 births),⁷⁶ GDM prevalence has increased by ≥4% over the last two decades, with a marked rise among non-White, overweight, and low-income groups.⁷⁷ Women with GDM also have a 7-fold increased risk of developing T2D and 43% increased risk of myocardial infarction or stroke post-pregnancy, with even higher rates among Black and Hispanic people.^{16, 17, 78, 79} Our study also builds on previous findings that singular nutrition policies—e.g., healthy revisions to the food package offered through WIC—can improve perinatal health.^28–32 Tobacco taxation also provides a model of an effective policy intervention that has translated into population-wide improvements in perinatal health.⁸⁰

A 2021 federal commission recommended a federal excise tax on SSBs as a strategy to control and prevent diabetes, yet empirical data regarding this relationship have been scant.⁷⁴ This study’s results suggest that SSB taxes in five US cities were associated with declines in GDM, GWG z-score, and SGA at the population level, with cohort differences by city, and subgroup differences. These results provide critical evidence that SSB taxes may improve maternal outcomes, as well as some improvements to neonatal outcomes, strengthening the evidence base for improving maternal nutrition during pregnancy.