This analysis used the Global Burden of Disease (GBD) 2023 database to quantify the global burden of diabetes mellitus among adolescents aged 10–19 years from 1990 to 2023 and to project future trends to 2040. The study evaluated age-standardized incidence (ASIR), prevalence (ASPR), mortality (ASMR), and disability-adjusted life year rates (ASDR) across 204 countries and territories. It also examined disparities by sex, age subgroup, and Socio-demographic Index (SDI) and applied time-series forecasting to estimate future burden trajectories.
The study extracted age-standardized rates for adolescents (10–19 years) from the GBD 2023 database for incidence, prevalence, mortality, and DALYs. Analyses were performed at global, regional, and national levels and stratified by sex, age subgroup (including 10–14 and 15–19 years), and SDI quintile. Temporal trends in ASRs were quantified using estimated annual percentage changes (EAPCs) from semi-log regression models. Health inequality was assessed using recommended WHO metrics. Future trajectories were forecast using ARIMA modeling with validation.
All burden estimates derive from the GBD 2023 integrated meta-regression outputs, which provide standardized estimates for 371 conditions across 204 countries and territories. Countries were categorized into five SDI quintiles (high, high-middle, middle, low-middle, low) based on per capita GDP, average educational attainment (age 15+), and fertility rates for those under 25. The SDI classification used the 2019 reference structure consistent with GBD conventions.
Temporal trends in age-standardized rates were quantified with semi-logarithmic regression to produce EAPCs. The model fits an ln(ASR) against calendar year; the slope is transformed to an annual percentage change with 95% confidence intervals. Trend direction was judged by the sign and confidence bounds of the EAPC: a persistent rise requires the EAPC and its lower 95% CI to be >0, and a persistent decline requires the EAPC and its upper 95% CI to be <0.
Absolute and relative inequality across SDI were assessed using the slope index of inequality (SII) and the concentration index (CI). The SII quantifies the absolute difference in burden between the highest and lowest SDI positions; positive values indicate higher burden in high-SDI settings, negative values indicate concentration in low-SDI settings. The CI, derived from the Lorenz concentration curve, ranges from −1 to 1; a negative CI indicates greater burden concentrated in lower SDI countries.
Sex- and region-stratified ARIMA models were built for each combination of measure (incidence, prevalence, mortality, DALYs), region (global and five SDI regions), and sex (female, male, both). Time series covered age-standardized rates from 1990–2023. The auto.arima() function in R (v4.4.1) selected optimal (p,d,q) orders using the corrected Akaike Information Criterion (AICc). Stationarity and differencing needs were assessed with the augmented Dickey–Fuller (ADF) test. Residual autocorrelation was checked with the Ljung–Box test (acceptable models had p > 0.05). Model fit statistics reported included AIC, BIC, and log-likelihood. Forecasts extended to 2040 with 80% and 95% prediction intervals; full model specifications and goodness-of-fit figures were provided in supplementary materials.
Between 1990 and 2023, global adolescent diabetes burden increased markedly. Incident cases, prevalent cases, and DALYs all showed significant rises over the study period, while deaths increased only modestly. Consequently, ASIR, ASPR, and ASDR rose persistently; ASMR remained comparatively stable with a slight decrease. The study reports that in 2023 an estimated 6.07 million adolescents were living with diabetes, corresponding to an ASPR of 458.07 per 100,000.
By SDI, the highest adolescent ASPR was observed in low-SDI regions (482.25 per 100,000), followed by high-SDI regions (442.81 per 100,000). Low-SDI regions also experienced the highest DALY and mortality burdens. Sex patterns shifted over time: by 2023 males exceeded females in both incidence and prevalence. The authors caution that this sex reversal may partly reflect evolving screening and diagnostic practices rather than purely biological or behavioral changes. Age stratification showed the overall burden increase was primarily driven by adolescents aged 15–19 years.
ARIMA projections indicate that ASIR, ASPR, and ASDR are expected to continue increasing through 2040 at global and SDI-stratified levels. In contrast, ASMR is projected to remain low and relatively stable over the forecast horizon. Forecast uncertainty was expressed with 80% and 95% prediction intervals; model diagnostics (AIC, BIC, Ljung–Box p-values) were used to confirm adequate specification.
The divergent pattern—rising non-fatal burden (incidence, prevalence, DALYs) alongside stable or slightly declining mortality—suggests improvements in acute diabetes care and survival but growing morbidity and long-term disease burden in adolescents. Geographic heterogeneity and concentrated burdens in low-SDI settings highlight persistent inequities. The observed male excess in incidence and prevalence by 2023 and the concentration of increases in older adolescents (15–19 years) point to the need for targeted prevention, early detection, and management strategies tailored by age, sex, and region.
Policy implications emphasized by the authors include prioritizing adolescent-focused diabetes prevention and control, strengthening surveillance and diagnostic capacity (to clarify observed sex patterns), and directing resources to low-SDI regions where DALY and mortality burdens are highest.
The study relies entirely on GBD 2023 modeled estimates; no new primary data were generated. The GBD uses integrative meta-regression to produce standardized estimates across locations and time, but the source acknowledges limitations inherent to secondary data and modelling (specific data gaps or bias sources are not exhaustively itemized in the summary). The ARIMA forecasts are model-based projections subject to uncertainty captured by prediction intervals and dependent on past trends continuing.
In conclusion, adolescent diabetes mellitus burden increased considerably from 1990 to 2023 with notable heterogeneity by age, sex, and geography. Forecasts predict continued rises in incidence, prevalence, and DALYs through 2040 while mortality remains low and stable. The results underscore an urgent need for age-, sex-, and region-specific public health strategies to address growing non-fatal morbidity and persistent inequalities.