Patterns of incidence and characteristics of youth with new-onset diabetes mellitus during the COVID era

INTRODUCTION

The incidence of diabetes mellitus (DM) in children, adolescents, and young adults has been on the rise for many decades.^[1-4] For type 1 diabetes (T1D), the incidence in the USA has been increasing by 2% per year and has followed a seasonal pattern, with the highest rate of new diagnoses in the wintertime and the lowest in the summertime.^[3-5] The incidence of type 2 diabetes (T2D) has been increasing by 5.3% per year, with the peak rate of incidence occurring in the late summer.^[3] The COVID-19 pandemic has been associated with a dramatic increase in new cases of DM^[1,5-23] as well as concern for an increased severity of disease at presentation^{.[6-10,12,14-16,19,21-29]}

During the first wave of the COVID-19 pandemic, it was unclear whether there was a true universal increase in the incidence of DM in pediatric patients, with conflicting reports from different countries^{.[18,19,21,24,26,28,30]} However, more recent publications have reviewed the incidence through the end of 2021 and determined that while the initial shutdown period had a decrease in incidence in most regions where it was reported, the incidence of DM, both T1D and T2D, subsequently increased and has not returned to pre-pandemic rates.^[1,5-20,31] The degree of severity at the time of presentation with new-onset DM has not been universally reported, with some centers seeing an increase in severity and others showing no change compared to the pre-pandemic rates of diabetic ketoacidosis (DKA) and hyperosmolar hyperglycemic syndrome (HHS).^{[6-12,14-16,18-25,28-33]}

Whether the sustained increase in T1D incidence has exceeded the predicted rate based on historical rates of annual increase in DM remains up for debate.^{[1,5,8,12,13,17]} However, the change in the seasonal pattern of incidence was notably consistent throughout Europe and North America in 2020, with the highest rates of new-onset T1D diagnosed in the summer and early fall.^{[1,5,6,8,9,11,13,17]} It was noted that in 2021, the shift in the seasonal pattern of T1D incidence continued in the USA/North America; however, most of the European groups saw a return to the pre-pandemic seasonal pattern of T1D incidence.^[5,13] To the best of our knowledge, these data have not been explored beyond early 2022.

There are fewer studies regarding the incidence of T2D in children and adolescents, both pre-pandemic and during the pandemic, but of those available, there has been a general uptrend in annual incidence, with a more marked increase during 2020 and 2021.^{[2,3,7,9,16,17,19,20,23,24]} The timing of diagnosis shifted from the pre-pandemic peak during August and later summer to a steady increase throughout the pandemic from spring 2020 to early 2021.^[9,16,17,23]

As described above, there has been significant interest and publication regarding the patterns of incidence and characteristics of patients with new-onset DM during the pandemic compared to the years before the pandemic. At this time, the majority of available data has been limited to 2020–2021, and minimal to no data has addressed the pattern of incidence through the end of 2022 in both T1D and T2D. In this retrospective chart review, we aimed to determine the incidence of DM and its subtypes at our center, Wendy Novak Diabetes Institute, during the three years pre-COVID-19 pandemic (2017–2019) and the pandemic (2020–2022). We sought to assess the seasonality and temporal pattern of new-onset DM during the pandemic compared to the 3 years prior. We additionally aimed to describe and compare the clinical characteristics of the patients diagnosed during the pandemic with those diagnosed before the pandemic.

MATERIAL AND METHODS

This retrospective chart review included all patients aged 12 months to 21 years diagnosed with DM at the Wendy Novak Diabetes Institute or Norton Children’s Hospital in Louisville, Kentucky, between January 1, 2017, and December 31, 2022. This facility is a tertiary referral center and provides care to the majority of pediatric patients with DM in this region. The data were dichotomized by year of diagnosis, with 2017–2019 considered pre-pandemic and 2020–2022 considered pandemic. The study was approved by the Institutional Review Board at the University of Louisville.

Excluded from this study were individuals who did not meet the American Diabetes Association diagnostic criteria,^[34] those with DM secondary to another medical condition, medication-induced DM, a diagnosis of maturity-onset diabetes in the young, and those who did not receive their new-onset DM care at our facility. Data for patients who were diagnosed with DM, T1D, or T2D during the time period listed above were collected from their records at diagnosis and first follow-up visit.

Demographic and clinical information were collected from patient medical records in the electronic medical record system. DM type was determined by the presence or absence of autoantibodies and insulin resistance at diagnosis. Patients who were positive for at least one autoantibody (GAD65, Islet antigen 2, and insulin) were classified as T1D. Those without positive autoantibodies who did not have evidence of insulin resistance were also classified as T1D. Those who had evidence of insulin resistance on physical examination without positive autoantibodies were classified as T2D. DKA, HHS, and mixed DKA-HHS were defined per the most recent International Society for Pediatric and Adolescent Diabetes clinical consensus guidelines published in 2022.^[35]

RESULTS

Of the 1139 charts reviewed, 969 patients met the inclusion criteria and had sufficient data for analysis (763 T1D and 206 T2D). Patient demographics and characteristics are shown in Tables 1 and 2.

The incidence of DM was higher during the pandemic period, with 561 new diagnoses, compared to the pre-pandemic period, with 408 new diagnoses, for an overall increase of 38% (Incidence rate ratio [IRR] 1.38, 95% confidence interval [CI] 1.20–1.58). Of those patients diagnosed with DM during the pandemic period, 97 were tested for COVID, and eight were positive.

Pattern of incidence

The average incidence rate by season for type 1 and type 2 diabetes are shown in Figure 1. The lowest rates of T1D incidence occurred during the pre-pandemic period summer, with the highest rate in the fall (IRR 1.68, 95% CI 1.23–2.30), followed by spring (IRR 1.43, 95% CI 1.03–1.97) and winter (IRR 1.40, 95% CI 1.03–1.91). During the pandemic, the highest rates of T1D incidence occurred in the summer (IRR 1.82, 95% CI 1.22–2.72), followed by fall (IRR 1.41, 95% CI 0.92–2.16), winter (IRR 1.39, 95% CI 0.90–2.16), and spring (IRR 1.38, 95% CI 0.93–2.07) [Figure 2]. Following the pandemic onset, the incidence of T2D increased with the highest rates in the summer (IRR 8.45, 95% CI 3.41–20.96), followed by winter (IRR 7.49, 95% CI 2.88–19.47), spring (IRR 6.61, 95% CI 2.7–16.17), and fall (IRR 6.22, 95% CI 2.40–16.13).

Close

Figure 1:

The average incidence rate by season for all DM, T1D, and T2D from January 2017– December 2022. DM: Diabetes mellitus, T1D: Type 1 diabetes, T2D: Type 2 diabetes.

Export to PPT

Close

Figure 2:

Comparison of the average seasonal incidence rates of type 1 diabetes between January 2017–February 2020 and March 2020–December 2022. *Incidence rate ratio 1.82 (95% confidence interval 1.22–2.72).

Export to PPT

The average incidence trend, adjusted for seasonal patterns and random error, showed an increase during the pandemic to a peak rate for all new diagnoses of DM in March 2021 (4.72/100K), followed by an overall decline until summer 2022 and then stabilized at 3.72/100K [Figure 3]. The graphs for T1D and T2D showed a very similar trendline, with the rate of new diagnosis highest in spring 2021 and then overall decline until spring 2022, after which the trend of new diagnosis stabilized at (2.84/100K, 0.88/100K, respectively).

DISCUSSION

This study identified an increase in the incidence of new-onset T1D and T2D during the pandemic, with a shift in seasonal patterns that persisted through the end of 2022. In concert with what has been reported,^[1,5-20,31] the rate of new-onset DM increased in our population during 2020–2021. The monthly incidence trend for T1D and T2D peaked in spring 2021, gradually declined until summer 2022, and then remained stable through December 2022 [Figure 3]. For T2D, the trend stabilized at an incidence rate lower than in spring 2021 but higher than in the pre-pandemic period. The trend rate of new-onset T1D at the end of 2022 is higher than the majority of the pre-pandemic period, with the exception of winter 2018–2019.

Close

Figure 3:

Graphical representation of the trend based on monthly incidence. The average incidence trend by season for all patients diagnosed with DM, those diagnosed with T1D, and those diagnosed with T2D. DM: Diabetes mellitus, T1D: Type 1 diabetes, T2D: Type 2 diabetes.

Export to PPT

To the best of our knowledge, this is the first report of the stabilized incidence for T1D and T2D in 2022. The decrease and stabilization of the incidence of DM in pediatric patients during the second half of 2021 and throughout 2022 was likely multifactorial. The return to in-person school and structured environments with increased opportunities for physical activity may have had positive effects on weight status and metabolic changes that occurred during the sedentary period. In addition, the decrease in rates of COVID infection and social isolation likely led to some improvement in psychological stress. It remains to be seen how this trend will evolve as components of lifestyle and social environments continue to change.

In an effort to understand the increase in incidence of DM since the SARS-CoV2 virus reached pandemic status, prior work has evaluated for COVID infection in patients with new-onset DM. These studies showed very low rates of coinfection with COVID-19 at the time of diagnosis or before diagnosis.^{[13,16-18,36]} In our group, less than 20% of patients diagnosed during the pandemic had COVID test results, and a very small proportion of those tested were positive for COVID-19 infection at the time of DM diagnosis.

Theories not related to the direct effects of SARS-CoV-2 infection have been proposed that include the immunologic and downstream metabolic effects of stress, both psychologic and physiologic, related to the drastic environmental changes put in place to prevent the spread of COVID.^{[2,5,13,37,38]} The lockdown measures led to a significant increase in social isolation, a decrease in physical activity, an increase in sedentary lifestyles, the start of virtual school, and the loss of structured environments. Subsequently, the rates of obesity rose at a more rapid rate than in previous years.^[39] Woolford et al. found the greatest increase in obesity during this period (2020–2021) occurred in the younger groups, 5–11 years and 12–15 years, in contrast to the pre-pandemic years when the older adolescents had the highest rates of obesity.^[39] In addition to this increase in obesity, there is concern that the increase in stress and sedentary lifestyle, along with the decrease in physical activity, further increased insulin resistance in those already at risk. Consistent with the increase in obesity in the younger groups, the age of T2D diagnosis during the pandemic decreased in our population, with the highest rate of new onsets in the 12–15-year-old group. In addition, our study showed that the patients diagnosed with T1D during the pandemic had a higher BMI percentile compared to those diagnosed before the pandemic. Combined, this supports the idea that stress and environmental changes have an impact on immunologic and metabolic function, making them likely contributors to the development of DM.

Due to the retrospective design, no ZnT8 antibody levels were available. In addition, evaluation of prior COVID-19 infection was limited due to the paucity of data available. As this study includes those diagnosed at a single center, and it will be important to evaluate this in diverse populations. While self-reported race was used, we acknowledge this is a social construct and that differences identified are most likely resultant of social determinants of health. This highlights the importance of further study to better characterize the effect of social determinants of health on the incidence of diabetes in youth.

CONCLUSION

This study supported the reported global increase in the incidence of DM in the pediatric population during the pandemic period and identified a persistent shift in the historical seasonal pattern of incidence. We noted a more pronounced change in the rate of T2D diagnosis that occurred at a younger age and an increased BMI percentile of new-onset T1D, parallel to the increased rate of obesity in younger age groups during the pandemic. Interestingly, we demonstrated a peak incidence of both T1D and T2D in spring 2021, followed by a gradual decline in the rate of new diagnoses until it stabilized in summer 2022 for the remainder of the year. Further studies are needed to evaluate the underlying mechanisms of the persistent increase in incidence.