Over the study period between 2008 and 2009 and 2018–2019, a total of 5,922 children were included in the analyses (ranging between 424 and 718 children annually). Participants’ characteristics by the levels of their LCS consumption for the first and last survey years are presented in the Table 1. In 2008–2009, the median age of children was 10 years (IQR, 7.0 to 14.0), 49.2% were boys, 89.0% were of white ethnicity, 64.9% had a BMI in the normal range, and 36.2% belonged to the lowest household income tertile (Table 1). Moreover, children in the High-LCS group were younger and more likely to be of white ethnic group than those in other groups. Across survey years, the characteristics of children in all LCS consumption groups remained largely similar, except for a lower proportion of participants with white ethnic background (89.0% in 2008–2009 vs. 82.3% in 2018–2019; P < 0.01), and a larger proportion with a missing BMI (4.8% vs. 8.3%; P = 0.03).

The proportional distribution of children in each group and their mean intake of LCS products have not substantially changed throughout the 11-year study period (Appendix Fig. 2). In 2008–2009, 70.4% of participants consumed LCS products, with a mean intake of 256.5 g/d among consumers. The prevalence and mean intake, both overall and within each LCS group, did not significantly change by 2018–2019 (Appendix Table 2).

Figure 1 illustrates the mean dietary contribution of each Nova subgroup relative to the total food and beverage intake (%gtotal/day) stratified by LCS consumption groups in 2008–2009 and 2018–2019. In 2008–2009, a gradient of lower water intake was observed with increasing levels of LCS consumption from 18.7%gtotal/day to 6.4%gtotal/day. Compared with the No-LCS group, children in the High-LCS group had a higher intake of non-LCS Nova 4 foods but a lower intake of both Nova 1 beverages and non-LCS Nova 4 beverages. In 2018–2019, the overall patterns remained largely similar to that observed in 2008–2009. However, the proportional intake of water increased in all groups, and the consumption of minimally processed beverages reduced during the study period. There was also a general reduction.

in the proportional intake of non-LCS ultra-processed drinks and a decrease in the proportional intake of non-LCS ultra-processed foods.

Results of multivariable linear regression showed that in 2008–2009, the proportion of total energy intake from free sugars was − 1.9%kcaltotal (95% CI, -2.8, -1.0) lower among children in the High-LCS group compared with those in the No-LCS group and − 3.0%kcaldrinks (95% CI, -3.7, -2.2) lower for free sugars from beverages (Table 2; Fig. 2). Free sugar intake in the No-LCS group showed a downward trend by 0.5%kcaltotal reduction per year (95% CI, -0.6 to -0.4) for intakes from the overall diet and 0.5%kcaldrinks reduction per year (95% CI -0.5 to -0.4) from beverages. The High-LCS group showed a smaller decline in free sugar intake from both overall diet and beverages compared with the No-LCS group. However, after Bonferroni correction, the difference in free sugar intake from overall diet no longer reached statistically significance. By year 2018–2019, there were no statistically significant differences in free sugars intake from neither the overall diet nor beverages between No-LCS and High-LCS groups (Appendix Table 4). When considering the free sugar intake measured in grams consumed, the trends and associations were consistent with the findings of the relative measures (Appendix Table 3, Appendix Fig. 3). Furthermore, findings for total sugar intake were similar to those from free sugar intake.

In year 2008–2009, the mean total energy intake (kcal/day) was significantly higher among the Low-LCS and Mid-LCS groups but similar for the High-LCS group as compared with the No-LCS group (Table 2; Fig. 2). The trend in total energy intake in the No-LCS group did not significantly change over time. However, the total energy intake for the Low-LCS and High-LCS groups showed a decline, with an incremental yearly reduction of 14.8 kcal/day (95% CI, -25.4 to -4.3) and 12.3 kcal/day (95% CI, -22.5, -2.0), respectively compared with the No-LCS group.

Grams of dietary intake by Nova subgroup in 2008–2009 (A) and 2018–2019 (B). A (N = 646). Percentages of dietary intake by Nova subgroups. No-LCS: Nova 2 processed culinary ingredients, 0.4%; Nova 3 processed foods and beverages, 2.3%. Low-LCS: Nova 2 processed culinary ingredients, 0.5%; Nova 3 processed foods and beverages, 1.6%; Nova 4 LCS beverages, 2.6; Nova 4 LCS foods, 0.5%. Mid-LCS: Nova 2 processed culinary ingredients, 0.4%; Nova 3 processed foods and beverages, 1.8%; Nova 4 LCS foods, 0.4%. High-LCS: Nova 2 processed culinary ingredients, 0.3%; Nova 3 processed foods and beverages, 1.7%; Nova 4 LCS foods, 0.3%. B (N = 424). Percentages of dietary intake by Nova subgroups. No-LCS: Nova 2 processed culinary ingredients, 0.6%; Nova 3 processed foods and beverages, 2.0%. Low-LCS: Nova 2 processed culinary ingredients, 0.5%; Nova 3 processed foods and beverages, 2.1%; Nova 4 LCS beverages, 3.0; Nova 4 LCS food, 0.2%. Mid-LCS: Nova 2 processed culinary ingredients, 0.5%; Nova 3 processed foods, 1.7%; Nova 4 LCS foods, 0.1%. High-LCS: Nova 2 processed culinary ingredients, 0.3%; Nova 3 processed foods, 1.3%; Nova 4 LCS foods, 0.1%

The fully adjusted models showed that compared with the No-LCS group in 2008–2009, the mean proportion of water intake among all beverages consumed was significantly lower in Mid-LCS and High-LCS groups (Table 2; Fig. 2). Over the 11-year study period, children in the No-LCS group experienced a significant increase in daily water intake, by a yearly increment of 2.7%gdrinks (95% CI, 2.2 to 3.1). Trends in daily water intake increased similarly in children of higher LCS groups, but the annual increments were smaller in Mid-LCS and High-LCS groups by -1.2%gdrinks (95% CI, -1.9, -0.6) and − 1.7%gdrinks (95% CI, -2.3 to -1.2) compared with children in No-LCS group, respectively.

Over 11 years, children in the No-LCS group showed an annual decrease in the proportion of ultra-processed foods and beverages consumed by -0.8%gtotal/day (95% CI, -1.1, -0.6) (Table 2; Fig. 2). By contrast, there was a smaller decrease in children in the High-LCS compared with the No-LCS group by 0.7%gtotal/day (95% CI, 0.3, 1.1). When considering only the consumption of ultra-processed foods (excluding beverages), Bonferroni-corrected results showed no significant difference in mean daily intake across LCS groups in 2008–2009, nor any statistically significant changes in consumption over 11 years within any of the LCS consumption groups.

The consumption of minimally processed foods and beverages in the No-LCS group increased significantly by 0.9%gtotal/day (95% CI 0.6 to 1.1) annually (Table 2; Fig. 2). This trend was not significantly different among children in the Low-LCS and Mid-LCS groups. However, the annual increase in the High-LCS group was smaller by -0.7%gtotal/day (95% CI -1.1 to -0.3) compared with the No-LCS group.

The trends and associations between the LCS consumption groups and the absolute intakes of total energy, ultra-processed foods and beverages, and minimally processed foods and beverages are presented in the Appendix Table 3. The findings were largely consistent with those presented in the main results.

Trends in daily dietary component intake by levels of LCS product consumption among UK children