Journal of Diabetes & Metabolic Disorders

https://doi.org/10.1007/s40200-023-01316-z

RESEARCH ARTICLE

Comparison of BMI and HbA1c changes before and during the

COVID‑19 pandemic in type 1 diabetes: a longitudinal
population‑based study
Marie Auzanneau1,2 · Dorothee M. Kieninger3 · Katharina Laubner4 · Christian Renner5 · Joaquina Mirza6 ·
Gerhard Däublin7 · Kirsten Praedicow8 · Holger Haberland9 · Claudia Steigleder‑Schweiger10 · Bettina Gohlke11 ·
Angela Galler12 · Reinhard W. Holl1,2 · on behalf of the DPV Initiative

Received: 1 June 2023 / Accepted: 17 September 2023

© The Author(s) 2023

Abstract
Purpose To compare the changes in body weight and glycemic control before and during the COVID-19 pandemic in people
with type 1 diabetes (T1D).
Methods In 47,065 individuals with T1D from the German Diabetes Prospective Follow-up Registry (DPV), we compared
the adjusted mean changes in BMI-Z-scores and HbA1c as well as the distribution of individual changes between four periods
from March 2018 to February 2022, by sex and age group (4- < 11, 11- < 16, 16–50 years).
Results At population level, the only significant pandemic effects were a slight increase in BMI Z-score in prepubertal chil-
dren (girls: + 0.03 in the first COVID year vs. before, P < 0.01; boys: + 0.04, P < 0.01) as well as a stabilization of HbA1c
in all subgroups or even improvement in women (− 0.08%, P < 0.01). At individual level, however, heterogeneity increased
significantly (p < 0.01), especially in children. More prepubertal children gained weight (girls: 45% vs. 35% before COVID;
boys: 39% vs. 33%). More pubertal girls lost weight (30% vs. 21%) and fewer gained weight (43% vs. 54%). More children
had a decreasing HbA1c (prepubertal group: 29% vs. 22%; pubertal girls: 33% vs. 28%; pubertal boys: 32% vs. 25%) and
fewer had increasing values. More women had stable HbA1c and fewer had increasing values (30% vs. 37%). In men, no
significant changes were observed.
Conclusion This real-world analysis shows no detrimental consequences of the two first COVID years on weight and HbA1c
in T1D on average, but reveals, beyond the mean trends, a greater variability at the individual level.

Keywords COVID-19 · Lockdown · BMI · HbA1c · Type 1 diabetes

Introduction results are inconsistent. Some analyses report weight gain

associated with the pandemic, either in adults [1, 2] or in
To date, only few studies have examined the impact of the children [3–5], while other report an increase of cases of
COVID-19 pandemic and associated lockdown measures on pediatric anorexia [6, 7], or weight changes in both direc-
people with type 1 diabetes (T1D). In particular, published tions [2]. A problem is that apart from rare exceptions [8, 9],
evidence on the effect of social restrictions and stay-at-home the large majority of these reports are solely based on one
orders on metabolic outcomes, such as BMI or HbA1c, is difference (i.e., the comparison of two values, one before
limited in this population. However, it is important to evalu- and one during the pandemic), and do not consider longer-
ate the potential consequences of the pandemic on diabetes term temporal trends including weight changes before the
management in people with T1D as a vulnerable group. pandemic as control [2, 4]. Another frequent limitation is the
Results of studies investigating the effect of the lockdown report of population means, which hides the heterogeneity
on body weight in the general population are not necessar- of the individual changes [3, 4]. In fact, mean estimates do
ily transferrable to people living with T1D, and published not provide any information on the proportion of individu-
als with weight gain, weight loss, or stable weight, and how
these proportions evolved over time. Most studies assessing
Extended author information available on the last page of the article

13
Vol.:(0123456789)
 Journal of Diabetes & Metabolic Disorders

the impact of the pandemic on weight and glycemic control Study of Obesity (Deutsche Adipositas Gesellschaft, DAG,
in T1D present similar limitations [5, 10, 11]. Arbeitsgemeinschaft Adipositas [AGA]) for 0- to 79-year-
Our aim in this analysis was therefore not only to compare old [14]. To adjust for differences between laboratories,
mean trends over the years in the population, but also the HbA1c values were standardized to the reference range
distribution of the individual variations, in order to analyze of the Diabetes Control and Complications Trial (DCCT)
whether the pandemic enlarged the heterogeneity of the indi- (4.05–6.05% [20.7–42.6 mmol/mol]) using the multiple of
vidual changes or not. In addition, we sought to perform the the mean method [15].
analysis in the context of longer temporal trends, comparing We adjusted for migration background and socioeco-
changes during the pandemic with changes occurring before. nomic situation which are both known to influence weight
We therefore compared changes in BMI-Z-score and HbA1c and glycemic control in T1D, also in Germany [16, 17].
means and distributions from March 2018 to February 2022 Migration background is defined as place of birth outside
in a large cohort of children and adults with T1D. Germany for the patient or at least for one parent. To take the
socioeconomic situation of the patients into consideration,
we used the German Index of Socioeconomic Deprivation
Methods of the year 2012 (­ GISD2012) [18], which is open for research
at the data repository of the German GESIS Leibniz-Insti-
Study population tute for the Social Sciences (https://​doi.​org/​10.​7802/​1460).
The ­GISD2012 encompasses aggregated data on education,
For this study, we used data from the multicenter Diabetes occupation, and income at the regional level, following a
Prospective Follow-up Registry (DPV; Diabetes-Patienten- methodology described previously [18]. Individuals were
Verlaufsdokumentation), based at the University of Ulm, assigned to districts and consequently to G ­ ISD2012 quintiles
Germany. At the End of March 2022, 512 diabetes centers using the five-digit postcode of their residence. Districts
mainly located in Germany and Austria have been partici- were categorized into deprivation quintiles, from Q1 (lowest
pating in this registry. All of them prospectively collect and deprivation) to Q5 (highest deprivation). Individuals without
document data on diabetes treatment and outcomes in the postcode were excluded from the analysis (n = 709), since
standardized DPV electronic health record, and transmit they could not be assigned to a district and therefore to a
every six months pseudo-anonymized data to the Univer- deprivation quintile.
sity of Ulm. After plausibility checks, the University of Ulm
reports inconsistent data back to the centers for correction Statistical analysis
and validation. Afterwards, anonymized data is used for
benchmarking and patient-centered analyses. Data analysis We defined four time-periods: time 1 from March 2018 to
is approved by the Ethics Committee of the Medical Faculty February 2019, time 2 from March 2019 to February 2020,
of the University of Ulm (Number 314/21). At the End of time 3 from March 2020 to February 2021, time 4 from
March 2022, 643,759 individuals with any type of diabetes March 2021 to February 2022. For each patient, longitu-
have been documented in the DPV registry, with 156,058 dinal data (age, diabetes duration, BMI-Z-score, HbA1c)
individuals classified as T1D. In this longitudinal study, we was aggregated as median for the respective time-period.
included patients with T1D (age at diagnosis ≥ 6 months; Median age per period of time was categorized into three
diabetes duration ≥ 3 months), and residence in Germany. In groups, roughly related to pubertal status: 4‒ < 11 years
addition, we decided to restrict the analysis to people aged (prepubertal), 11‒ < 16 years (pubertal), and 16‒ ≤ 50 years
between 4 and 50 years to form homogeneous age groups (postpubertal and adult). Over the four periods of time from
(excluding older patients who have more frequently long- March 2018 to February 2022, 76,6% of the individuals
term complications and other comorbidities) and to consider (n = 36,032) remained in the same age category, 9.4% of
only people in school or working age, whom we expect to the individuals (n = 4,443) transited from the prepubertal to
have been more affected by school closures and stay-at-home the pubertal group, and 14.0% of the individuals (n = 6,590)
orders. We then analyzed all visits between March 2018 and transited from the pubertal to the postpubertal group. We
February 2022. took the age group of the first period of time into account to
perform the analyzes stratified by sex and age group.
Variables To investigate the mean changes in each subgroup before
COVID, as well as in the first and second COVID year,
In the DPV database, the definition of T1D is based on a mean BMI-Z-Score and HbA1c were estimated in each
physician’s diagnosis according to the ISPAD or ADA time period (time 1 to time 4) using linear regression mod-
guidelines [12, 13]. The BMI-Z-score is defined according els, adjusted for diabetes duration, migration background,
to the reference values of the German Association for the and quintile of socioeconomic deprivation, with repeated

13
Journal of Diabetes & Metabolic Disorders

measurements within individuals (patient as random effect). built TS1M7 on a Windows server 2019 mainframe (SAS
Then, the mean differences between the estimates of two Institute, Cary, NC).
periods were analyzed using respective models.
To analyze the distribution of the individual changes, we
calculated for each patient the BMI-Z-Score- and HbA1c Results
change between the median values in times 2 and 1 (change
before COVID), times 3 and 2 (change in the first COVID A total of 47,065 children and adults met the inclusion crite-
year), and times 4 and 3 (change in the second COVID year). ria (diagnosis of T1D for at least three months, n = 146,059;
In addition, we calculated in each subgroup and time period age between 4 and 50 years, n = 124.080; visits recorded
the proportion of individuals with weight loss (BMI-Z-score between March 2018 and March 2022, n = 53,443; residence
difference <  − 0.1), stable weight (BMI-Z-score difference in Germany with postcode available, n = 47,065). Character-
between − 0.1 and + 0.1), and weight gain (BMI-Z-score istics of the study population stratified by sex and age group
difference >  + 0.1). Similarly, we calculated the propor- are presented in Table 1. In the prepubertal group, none of
tion of individuals with decreasing, stable, and increasing the observed characteristics differed significantly between
HbA1c before COVID and during the first two COVID girls and boys. From the age of 11 and above, girls had a
years (HbA1c difference <  − 0.2; between − 0.2 and + 0.2, longer diabetes duration compared to boys of the same age
and >  + 0.2 respectively). group, as well as a higher BMI-Z-score (Table 1). Median
Unadjusted patient characteristics are presented as HbA1c was slightly higher in pubertal females compared to
median with lower and upper quartile (Q1-Q3) for continu- males (Table 1).
ous variables, or as proportion for variables with binomial
distribution. Wilcoxon tests and X ­ 2 tests, adjusted for mul- BMI‑Z‑score
tiple comparisons according to the Holm-Bonferroni step-
down procedure, were used to compare these characteristics • Adjusted mean BMI-Z-score trends before and during
between males and females. Results of regression analyses the pandemic
are presented as adjusted estimates with their 95% confi-
dence intervals (95% CI). P-values were adjusted for multi- In all subgroups, the adjusted BMI-Z-score changes
ple comparisons according to the Tukey–Kramer procedure. were similar during the two COVID years compared to
A p-value < 0.01 (two-sided) was considered statistically sig- the year before, except in prepubertal children (Fig. 1).
nificant. Statistical analysis was performed using SAS 9.4, In this age group, we observed a slight but significant

Table 1  Characteristics of the study population

Study population
(n = 47,065)
By age group By sex
Female Male P-values
(n = 22,087) (n = 24,978)

Prepubertal: 4- < 11 years Age, years (median, Q1-Q3) 8.0 (5.9–9.6) 7.9 (5.8–9.6) 0.39
(n = 13,607) Diabetes duration, years (median, Q1-Q3) 1.3 (0.6–3.3) 1.4 (0.6–3.4) 0.04
Migration background, (%) 30.1 30.5 0.65
BMI-Z-score, AGA, (median, Q1-Q3) 0.25 (-0.32–0.85) 0.27 (-0.31–0.89) 0.53
HbA1c, % (median, Q1-Q3) 6.97 (6.42–7.52) 6.90 (6.37–7.50) 0.05
Pubertal: 11- < 16 years Age, years (median, Q1-Q3) 13.5 (12.3–14.7) 13.7 (12.4–14.8) < 0.001
(n = 15,652) Diabetes duration, years (median, Q1-Q3) 3.9 (0.9–7.4) 3.0 (0.7–6.9) < 0.001
Migration background, n (%) 28.3 27.5 0.65
BMI-Z-score, AGA, (median, Q1-Q3) 0.47 (-0.20–1.16) 0.20 (-0.48–0.93) < 0.001
HbA1c, % (median, Q1-Q3) 7.20 (6.54–7.92) 7.12 (6.43–7.88) < 0.001
Postpubertal and adults: 16–50 years Age, years (median, Q1-Q3) 20.6 (17.5–34.3) 20.2 (17.4–35.1) 0.40
(n = 17,806) Diabetes duration, years (median, Q1-Q3) 9.6 (4.8–15.9) 8.4 (3.6–14.7) < 0.001
Migration background, n (%) 26.4 24.5 0.25
BMI-Z-score, AGA, (median, Q1-Q3) 0.46 (-0.33–1.20) 0.03 (-0.80–0.87) < 0.001
HbA1c, % (median, Q1-Q3) 7.43 (6.64–8.48) 7.47 (6.64–8.67) 0.12

Unadjusted data. Q1-Q3: lower–upper quartile

13
 Journal of Diabetes & Metabolic Disorders

Fig. 1  Distribution of BMI-Z-scores changes before and during socioeconomic deprivation, and repeated measurements within
the COVID-19 pandemic in patients with type-1-diabetes, strati- patients (random effect). *significant (P < 0.01), adjusted for multi-
fied by sex and age group. Δ2-1, Δ3-2 or Δ4-3: mean change between ple comparisons according to the Tukey–Kramer procedure). IQR:
time 2 (March 2019-February 2020) and time 1 (March 2018-Feb- interquartile range. Distribution of individual BMI Z-score changes:
ruary 2019), time 3 (March 2020-February 2021) and time 2, time Before COVID (time 2 – time 1) in blue; With COVID first year
4 (March 2021-February 2022) and time 3. adj.: Estimated mean (time 3 – time 2) in red, With COVID second year (time 4-time 3) in
change adjusted for diabetes duration, migration background, and green

BMI-Z-score increase in the first COVID year com- COVID continue to evolve similarly during the two years
pared to the year before, without significant difference by of the pandemic: the BMI-Z-score continue to increase
sex (adjusted mean change in the first COVID year vs. in pubertal girls, and to a lesser extent in boys (Fig. 1).
before: + 0.03 [P < 0.01] vs. 0.00 [P = 0.99] in girls; + 0.04 The BMI-Z-score continued to decrease slightly in females
[P < 0.01] vs. 0.00 [P = 0.49] in boys; comparison by sex: aged 16–50 years and remained stable during the whole
p = 0.07). In the other age groups, the BMI trends before observation period in males of the same age group (Fig. 1).

13
Journal of Diabetes & Metabolic Disorders

• Distribution of the individual BMI-Z-score changes vs. 21%), whereas those with weight gain decreased in a
before and during the pandemic similar proportion (43% vs. 54%, both differences: P < 0.001,
Table 2). Observation of Fig. 1C conforms to these results:
Regarding the distribution of the individual BMI-Z- although the adjusted mean changes remained similar over
scores changes, more heterogeneity was observed dur- the years, the unadjusted mean changes decreased in the
ing the first COVID year (Interquartile range [IQR] wider second COVID year (green curve shifted to the left). In
resulting from a lower proportion of individuals with sta- the group aged 16–50 years, changes were not significant
ble weight) compared to the year before (Fig. 1). This was (Table 2).
especially the case in the prepubertal group, in both girls
and boys, as well as in pubertal boys (Fig. 1). In the second HbA1c
COVID year, the heterogeneity declined slightly in all chil-
dren aged < 16 years, and to a greater extent in postpubertal • Adjusted mean HbA1c trends before and during the
individuals and adults. In the prepubertal group, the greater pandemic
heterogeneity during the two COVID years was character-
ized by a higher proportion of children with weight gain The adjusted mean HbA1c trend changed with the sec-
(45% in the second COVID year vs. 35% before COVID ond COVID year compared to the years before in all chil-
in girls and 39% vs. 33% in boys, both P < 0.001, Table 2). dren < 16 years (Fig. 2): mean HbA1c values increased
In the pubertal group, the proportions of boys with stable before and during the first COVID year, and then stabi-
weight, weight loss, and weight gain in the three time peri- lized. In older subjects, the HbA1c trend changed simi-
ods remained similar (Table 2). By contrast, the proportion larly but already in the first COVID year (Fig. 2): after
of pubertal girls with weight loss increased considerably in an increase before COVID (+ 0.07% [+ 0.02; + 0.12]
the second COVID year compared to the year before (30% in females and + 0.07% [+ 0.02; + 0.11] in males, both

Table 2  Proportion of individuals with weight loss, stable weight, and weight gain before COVID and during the first two COVID years, strati-
fied by age group and sex
Age group Sex Time period* With P-value*** With stable P-value*** With P-value***
weight weight** weight
loss** (%) (%) gain** (%)

Prepubertal: 4- < 11 years Female Before COVID 33.01 < 0.001 31.98 0.004 35.02 < 0.001
With COVID first year 31.31 26.93 41.76
With COVID second year 26.98 28.20 44.82
Male Before COVID 34.63 0.329 32.28 0.050 33.09 < 0.001
With COVID first year 34.36 27.19 38.45
With COVID second year 32.17 29.25 38.58
Pubertal: 11- < 16 years Female Before COVID 20.73 < 0.001 25.31 1.000 53.96 < 0.001
With COVID first year 25.76 25.22 49.01
With COVID second year 30.17 26.71 43.12
Male Before COVID 27.00 1.000 28.11 0.477 44.89 1.000
With COVID first year 30.94 24.70 44.36
With COVID second year 28.16 26.26 45.58
Postpubertal and adults: Female Before COVID 31.19 1.000 37.61 1.000 31.19 1.000
16–50 years With COVID first year 31.77 35.24 32.99
With COVID second year 30.56 39.31 30.12
Male Before COVID 32.12 1.000 32.33 1.000 35.55 1.000
With COVID first year 33.41 30.92 35.68
With COVID second year 31.17 33.73 35.09
*
Before COVID: changes between time 2 (March 2019-February 2020) and time 1 (March 2018-February 2019); With COVID first year:
changes between time 3 (March 2020-February 2021) and time 2; With COVID second year: changes between time 4 (March 2021-February
2022) and time 3
**
Weight loss: BMI-Z-score difference < -0.1; stable weight BMI-Z-score difference between -0.1 and + 0.1; Weight gain: BMI-Z-score differ-
ence >  + 0.1
***
­Chi2-Test for comparison before COVID vs. COVID second year. P-values adjusted for multiple tests according to the Bonferroni procedure

13
 Journal of Diabetes & Metabolic Disorders

Fig. 2  Distribution of HbA1c changes before and during the COVID- vation, and repeated measurements within patients (random effect).
19 pandemic in patients with type-1-diabetes, stratified by sex and *significant (P < 0.01), adjusted for multiple comparisons according
age group. Δ2-1, Δ3-2 or Δ4-3: mean change between time 2 (March to the Tukey–Kramer procedure). IQR: interquartile range. Distribu-
2019-February 2020) and time 1 (March 2018-February 2019), time tion of individual HbA1c changes: Before COVID (time 2 – time 1)
3 (March 2020-February 2021) and time 2, time 4 (March 2021-Feb- in blue; With COVID first year (time 3 – time 2) in red, With COVID
ruary 2022) and time 3. adj.: Estimated mean change adjusted for second year (time 4-time 3) in green
diabetes duration, migration background, and socioeconomic depri-

P < 0.01), mean HbA1c then stabilized (males) or even Except in women > 16 years, more individual HbA1c
decreased (females: − 0.08% [− 0.13; − 0.03], P < 0.01). changes were observed in all subgroups during the first
COVID year (IQR wider) compared to the year before
• Distribution of the individual HbA1c-changes before (Fig. 2). This was especially the case in pubertal boys.
and during the pandemic However, this greater heterogeneity decreased in the

13
Journal of Diabetes & Metabolic Disorders

second COVID year (Fig. 2). After two COVID years, the Discussion
lower proportion of children with increasing HbA1c was
replaced by a higher number with decreasing HbA1c (29% This longitudinal analysis of registry-based data documented
vs. 22% in both girls and boys in the prepubertal group, between 2018 and 2022 reveals that the two pandemic years
33% vs. 28% in pubertal girls, and 32% vs. 25% in pubertal were not associated with systematic detrimental conse-
boys, all P < 0.001, Table 3). In women, we observed more quences on weight and glycemic control in our population
homogeneity: the proportion with increasing HbA1c fell with T1D. However, the individual responses of people
from 37 to 30% and the proportion with stable HbA1c with T1D to the COVID-19 pandemic, in terms of BMI or
grew from 28 to 33%, both P < 0.001. In men, the distribu- HbA1c-changes, differed widely.
tion of the changes remained similar before and during the Considering only the adjusted mean changes, BMI
pandemic (Table 3). trends were not affected by the pandemic, except a slight

Table 3  Proportion of individuals with decreasing, stable, and increasing HbA1c before COVID and during the first two COVID years, stratified
by age group and sex
Age group Sex Time period* With decreas- P-value*** With stable P-value*** With increas- P-value***
ing HbA1c** HbA1c** ing HbA1c**
(%) (%) (%)

Prepubertal: Female Before COVID 21.68 < 0.001 34.23 1.000 44.09 < 0.001
4- < 11 years With COVID first 21.72 32.83 45.45
year
With COVID second 29.10 34.37 36.53
year
Male Before COVID 21.80 < 0.001 35.52 1.000 42.68 < 0.001
With COVID first 22.26 33.11 44.62
year
With COVID second 29.13 36.92 33.95
year
Pubertal: Female Before COVID 27.60 < 0.001 25.59 1.000 46.81 < 0.001
11- < 16 years With COVID first 32.62 25.26 42.13
year
With COVID second 33.42 26.62 39.96
year
Male Before COVID 25.35 < 0.001 27.11 1.000 47.55 < 0.001
With COVID first 30.97 24.85 44.18
year
With COVID second 32.08 27.93 39.99
year
Postpubertal and Female Before COVID 35.17 1.000 27.76 0.001 37.07 < 0.001
adults: 16–50 years With COVID first 36.30 28.22 35.47
year
With COVID second 36.47 33.19 30.33
year
Male Before COVID 32.56 1.000 31.50 1.000 35.94 1.000
With COVID first 32.89 31.84 35.27
year
With COVID second 32.55 31.89 35.56
year
*
Before COVID: changes between time 2 (March 2019-February 2020) and time 1 (March 2018-February 2019),
With COVID first year: changes between time 3 (March 2020-February 2021) and time 2,
With COVID second year: changes between time 4 (March 2021-February 2022) and time 3
**
Decreasing HbA1c: HbA1c difference < -0.2; stable HbA1c: HbA1c difference between -0.2 and + 0.2; Increasing HbA1c: HbA1c differ-
ence >  + 0.2
***
­Chi2-Test for comparison before COVID vs. COVID second year. P-values adjusted for multiple tests according to the Bonferroni procedure

13
 Journal of Diabetes & Metabolic Disorders

increase in prepubertal children. Nevertheless, beyond the cases of eating disorders, in particular in young women, in
mean, it appears clearly that the first year of the pandemic conjunction with the pandemic [6, 7, 22].
has been accompanied by more individuals with BMI Our findings indicate a stabilization or even an improve-
increases or decreases, and that this greater heterogeneity ment of glycemic control in the second COVID-year in both
particularly affected children and adolescents < 16 years. children and adults with T1D. During the pandemic, the pro-
With the second COVID year, the proportion of individu- portion of individuals with increasing HbA1c fell in nearly
als with stable BMI increased again, but we observed a all subgroups. Few studies found no significant changes in
larger proportion of prepubertal children with weight gain, children [11] or only very slight changes in adults [10], but a
as well as a larger proportion of pubertal girls with weight greater number of analyses [23, 24], including three system-
loss. While increasing before COVID, mean HbA1c trends atic reviews [5, 25, 26], found better glycemic control, in line
stabilized during the pandemic in all subgroups or even with our results. To explain the possible positive effect of the
improved in women. At individual level, no significant lockdown on glycemic control in T1D, several hypotheses
change was observed in adult men, but in all other sub- have been formulated [5]. In particular, the stay-at-home
groups, the proportion of individuals with increasing orders may have been associated with more time, not only
HbA1c decreased. for diabetes management, but also for self-care, healthy
Whereas some studies performed in the general popula- meals and exercise. Moreover, the lockdown could have
tion have found a greater BMI increase in all children dur- enabled a more predictable daily routine, with more regular
ing the pandemic compared to previous years [3, 5], other meal and sleep times [27]. Last but not least, the increasing
found that this was more pronounced or, like in our results, use of telemedicine during the pandemic [28] combined with
restricted to prepubertal children [8, 19]. In general, the a widespread use of diabetes technology (in particular, con-
accelerated increase in BMI in children due to the pandemic tinuous glucose monitoring [CGM] and automated insulin
has often been explained as a consequence of reduced physi- delivery [AID]) in T1D seems to have contribute to an effec-
cal activity and altered eating habits due to the widespread tive diabetes management despite lockdown [11].
closure of schools and leisure or sport facilities during the A limitation of this study is that we did not take the use
lockdown [3]. In line with these studies, we found a higher of diabetes technology into account, although diabetes tech-
proportion of children with weight gain, but the average nology and CGM in particular is associated with improved
excess BMI-Z-score gain related to the pandemic in this glycemic control [29]. For example, a more frequent use of
age group was very small. It is possible that the impact of CGM during the pandemic could have improved glycemic
the COVID-lockdown on children’s weight development has outcomes. Another bias could have resulted from potential
been attenuated in the presence of diabetes. This disease differences in frequency of weight and HbA1c measure-
being part of well-documented risk factors for worse COVID ments during the pandemic compared to the period before.
outcomes, parents of children with this chronic condition However, a recent study from Germany indicates that access
could have paid particularly attention to alimentation and/ to healthcare did not change considerably during the pan-
or to physical activities during the pandemic [20]. In addi- demic for children and adolescents with T1D [30]. For our
tion, working from home, which was widespread during the study population, BMI and HbA1c were documented only
lockdown, may have made it easier to prepare healthy meals, slightly less frequently during the pandemic compared to the
and improved parental supervision. Moreover, children with pre-pandemic period (in mean: 3.6 instead of 4.2 times per
diabetes and their parents may have benefited from more fre- year for BMI, 3.1 instead of 3.4 times per year for HbA1c).
quent medical consultations and health advices than people Nevertheless, a possible selection bias could have affected
without chronic disorder. Thus, T1D could have constituted the results, if individuals with worse glycemic control had
a moderating factor against weight gain in this time period fewer medical visits or less health care utilization during the
when compared to the general population. pandemic than those with better glycemic control. Inversely,
Regarding older children and adults, longitudinal studies one could consider that especially individuals with worse
in the general population which have taken a pre-pandemic glycemic control were more prone to visit their diabetes
control period into consideration indicate that BMI trends center during the pandemic than other individuals with bet-
were not modified by the pandemic [9, 10, 21]. Our findings ter health outcomes. A strength of this longitudinal analysis
in adolescents and adults with T1D confirm these results. In is the use of repeated measurements over four years, which
pubertal girls however, even if the proportion with weight allows the comparison of trends between the COVID and
gain was still greater that the proportion with weight loss the pre-COVID periods. In addition, real world evidence
after two years pandemic, about one third more have lost could have been provided by the use of a nationwide registry
weight compared to the year before. This aspect is not appar- comprising about 90% of children and adolescents and about
ent when only the adjusted mean values are evaluated. These 30% of adults with T1D in Germany. Moreover, the report
results are in line with many other studies indicating more of the distribution of individual weight and HbA1c changes

13
Journal of Diabetes & Metabolic Disorders

over the years provides useful information to understand the provided by the German Center for Diabetes Research (DZD, grant
real impact of two years COVID in people living with T1D. number: 82DZD14E03), the German Diabetes Association (DDG),
the German Robert Koch Institute (RKI, grant number 1368–1711).
To summarize, this longitudinal analysis contributes to
reduce the concerns over potential detrimental impact of the Data Availability Aggregated data might be made available upon rea-
COVID pandemic on health outcomes of individuals with sonable request via email to the senior author reinh​ard.​holl@​uni-​ulm.​
T1D. First, the pandemic seems to have been associated with de.
perturbations or instability in terms of weight and glycemic
Declarations
control at the individual level, but eventually, in the sec-
ond COVID year, these variations have been significantly Competing Interests The authors declare no conflicts of interest.
reduced in all subgroups. It is possible that most individuals
progressively adapted their lifestyle to the pandemic situa- Open Access This article is licensed under a Creative Commons Attri-
tion with its lockdown restrictions, and that the psychologi- bution 4.0 International License, which permits use, sharing, adapta-
cal impact of the pandemic consequently reduced over time. tion, distribution and reproduction in any medium or format, as long
as you give appropriate credit to the original author(s) and the source,
In addition, in our large population with T1D, the pandemic provide a link to the Creative Commons licence, and indicate if changes
was not associated with clinically relevant changes in BMI were made. The images or other third party material in this article are
trends, but with a slight improvement in the HbA1c trend. included in the article’s Creative Commons licence, unless indicated
Stay-at-home orders combined with the fear of adverse otherwise in a credit line to the material. If material is not included in
the article’s Creative Commons licence and your intended use is not
COVID outcomes in individuals with diabetes may have permitted by statutory regulation or exceeds the permitted use, you will
encouraged a heathier lifestyle [5]. In addition, the develop- need to obtain permission directly from the copyright holder. To view a
ment of telemedicine [28] associated with the increasing copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
use of diabetes technology may have played a positive role
during the pandemic.
In conclusion, even if the COVID pandemic did not affect References
weight and glycemic control negatively in individuals with
T1D on average, our data indicates greater variation in some 1. Ramirez Manent JI, Altisench Jane B, Sanchis Cortes P, Busquets-
subgroups, in particular children and adolescents. Since Cortes C, Arroyo Bote S, Masmiquel Comas L, et al. Impact of
COVID-19 Lockdown on Anthropometric Variables, Blood Pres-
weight and glycemic variability are associated with adverse
sure, and Glucose and Lipid Profile in Healthy Adults: A before
diabetes-related outcomes [31, 32], it is important to pay and after Pandemic Lockdown Longitudinal Study. Nutrients.
attention to these changes to improve the care of vulnerable 2022;14(6):1237. https://​doi.​org/​10.​3390/​nu140​61237.
groups in the future. 2. Bakaloudi DR, Barazzoni R, Bischoff SC, Breda J, Wickramasin-
ghe K, Chourdakis M. Impact of the first COVID-19 lockdown on
Acknowledgements Special thanks to A. Hungele and R. Ranz for sup- body weight: A combined systematic review and a meta-analysis.
port and the development of the DPV documentation software (Institute Clin Nutr. 2021. 20:S0261–5614(21)00207–7. https://​doi.​org/​10.​
of Epidemiology and Medical Biometry, ZIBMT, University of Ulm). 1016/j.​clnu.​2021.​04.​015
Furthermore, we thank all participating centers in the DPV initiative, 3. Knapp EA, Dong Y, Dunlop AL, Aschner JL, Stanford JB, Hartert
especially the centers contributing data to this investigation and their T, et al. Changes in BMI During the COVID-19 Pandemic. Pedi-
patients. A full list of all participating centers is available at https://​ atrics. 2022;1;150(3):2022056552. https://​doi.​org/​10.​1542/​peds.​
www.d-​p-v.​eu. 2022-​056552.
4. Chang TH, Chen YC, Chen WY, Chen CY, Hsu WY, Chou Y,
Author Contributions M.A. and R.W.H. designed the study. D.M.K., et al. Weight Gain Associated with COVID-19 Lockdown in Chil-
K.L., C.R., J.M., G.D., K.P., H.H., C.S.-S., B.G., A.G. contributed to dren and Adolescents: A Systematic Review and Meta-Analysis.
data collection. M.A. performed the statistical analysis, created the Nutrients. 2021;13(10):3668. https://d​ oi.o​ rg/1​ 0.3​ 390/n​ u1310​ 366.
figures, performed literature research, and wrote the manuscript. All 5. Eberle C, Stichling S. Impact of COVID-19 lockdown on glyce-
authors reviewed and edited the manuscript. R.W.H. is the coordinator mic control in patients with type 1 and type 2 diabetes mellitus: a
for the DPV Initiative. M.A. is the guarantor of this work and, as such, systematic review. Diabetol Metab Syndr. 2021;13:95–9. https://​
had full access to all the data in the study and takes responsibility for doi.​org/​10.​1186/​s13098-​021-​00705-9.
the integrity of the data and the accuracy of the data analysis. 6. Agostino H, Burstein B, Moubayed D, Taddeo D, Grady R, Vyver
E, et al. Trends in the Incidence of New-Onset Anorexia Ner-
Funding Open Access funding enabled and organized by Projekt vosa and Atypical Anorexia Nervosa Among Youth During the
DEAL. This article is part of a project (www.imisophia.eu) that has COVID-19 Pandemic in Canada. JAMA Netw Open. 2021;4:
received funding from the Innovative Medicines Initiative 2 Joint e2137395. https://d​ oi.o​ rg/1​ 0.1​ 001/j​ amane​ twork​ open.2​ 021.3​ 7395.
Undertaking under grant agreement No 875534. This Joint Undertak- 7. Springall G, Cheung M, Sawyer SM, Yeo M. Impact of the coro-
ing support from the European Union’s Horizon 2020 research and navirus pandemic on anorexia nervosa and atypical anorexia ner-
innovation programme and EFPIA and T1D Exchange, JDRF, and vosa presentations to an Australian tertiary paediatric hospital. J
Obesity Action Coalition. In addition, this analysis is supported by the Paediatr Child Health. 2022;58:491–6. https://​doi.​org/​10.​1111/​
German Diabetes Foundation (grant number FP-0446–2022), and by jpc.​15755.
a financial support program for female researchers, Office for Gender 8. Wu AJ, Aris IM, Hivert MF, Rocchio C, Cocoros NM, Klompas
Equality, Ulm University. Additional funding for the DPV registry is M, et al. Association of Changes in Obesity Prevalence With the
COVID-19 Pandemic in Youth in Massachusetts. JAMA Pediatr.

13
 Journal of Diabetes & Metabolic Disorders

2022;176:198–201. https://​doi.​org/​10.​1001/​jamap​ediat​rics.​2021.​ 23. Garofolo M, Aragona M, Rodia C, Falcetta P, Bertolotto A, Campi

5095. F, et al. Glycaemic control during the lockdown for COVID-19
9. Wing RR, Venkatakrishnan K, Panza E, Marroquin OC, Kip KE. in adults with type 1 diabetes: A meta-analysis of observational
Association of COVID-19 Stay-at-Home Orders With 1-Year studies. Diab Res Clin Pract. 2021;180: 109066. https://​doi.​org/​
Weight Changes. JAMA Netw Open. 2022;5: e2217313. https://​ 10.​1016/j.​diabr​es.​2021.​109066.
doi.​org/​10.​1001/​jaman​etwor​kopen.​2022.​17313. 24. Kaushal T, Tinsley L, Volkening LK, Ambler-Osborn L, Laffel L.
10. Hartmann B, Tittel SR, Femerling M, Pfeifer M, Meyhofer S, Improvement in Mean CGM Glucose in Young People with Type
Lange K, et al. COVID-19 Lockdown Periods in 2020: Good 1 Diabetes During 1 Year of the COVID-19 Pandemic. Diabetes
Maintenance of Metabolic Control in Adults with Type 1 and Technol Ther. 2022;24:136–9. https://​doi.​org/​10.​1089/​dia.​2021.​
Type 2 Diabetes. Exp Clin Endocrinol Diab. 2022. https://d​ oi.o​ rg/​ 0258.
10.​1055/a-​1743-​2537 25. O'Mahoney LL, Highton PJ, Kudlek L, Morgan J, Lynch R, Schof-
11. Hammersen J, Reschke F, Tittel SR, Pappa A, Dost A, Kostner ield E, et al. The impact of the COVID-19 pandemic on glycaemic
K, et al. Metabolic control during the SARS-CoV-2 lockdown in control in people with diabetes: A systematic review and meta-
a large German cohort of pediatric patients with type 1 diabetes: analysis. Diabetes Obes Metab. 2022. https://​doi.​org/​10.​1111/​
Results from the DPV initiative. Pediatr Diab. 2022;23:351–61. dom.​14771
https://​doi.​org/​10.​1111/​pedi.​13319. 26. Karatzi K, Poulia KA, Papakonstantinou E, Zampelas A. The
12. Mayer-Davis EJ, Kahkoska AR, Jefferies C, Dabelea D, Balde N, Impact of Nutritional and Lifestyle Changes on Body Weight,
Gong CX, et al. ISPAD Clinical Practice Consensus Guidelines Body Composition and Cardiometabolic Risk Factors in Children
2018: Definition, epidemiology, and classification of diabetes in and Adolescents during the Pandemic of COVID-19: A System-
children and adolescents. Pediatr Diab. 2018;19(Suppl 27):7–19. atic Review. Children (Basel) 2021;8:https://d​ oi.o​ rg/1​ 0.3​ 390/c​ hild​
https://​doi.​org/​10.​1111/​pedi.​12773. ren81​21130. https://​doi.​org/​10.​3390/​child​ren81​21130
13 American Diabetes Association Professional Practice Committee. 27. Capaldo B, Annuzzi G, Creanza A, Giglio C, De Angelis R,
2 Classification and Diagnosis of Diabetes: Standards of Medical Lupoli R, et al. Blood Glucose Control During Lockdown for
Care in Diabetes-2022. Diabetes Care. 2022;45:S17–38. https://​ COVID-19: CGM Metrics in Italian Adults With Type 1 Dia-
doi.​org/​10.​2337/​dc22-​S002. betes. Diabetes Care. 2020;43:e88–9. https://​doi.​org/​10.​2337/​
14. Kromeyer-Hauschild K, Moss A, Wabitsch M. Body mass index dc20-​1127.
reference values for German children, adolescents and adults. Adi- 28 Giani E, Dovc K, Dos Santos TJ, Chobot A, Braune K, Cardona-
positas - Ursachen, Folgeerkrankungen, Therapie. 2015;09:123–7. Hernandez R, et al. Telemedicine and COVID-19 pandemic: The
15. Rosenbauer J, Dost A, Karges B, Hungele A, Stahl A, Bachle C, perfect storm to mark a change in diabetes care. Results from
et al. Improved metabolic control in children and adolescents with a world-wide cross-sectional web-based survey. Pediatr Diab.
type 1 diabetes: a trend analysis using prospective multicenter data 2021;22:1115–9. https://​doi.​org/​10.​1111/​pedi.​13272.
from Germany and Austria. Diab Care. 2012;35:80–6. https://d​ oi.​ 29. Šoupal J, Petruželková L, Grunberger G, Hásková A, Flekač M,
org/​10.​2337/​dc11-​0993. Matoulek M, et al. Glycemic Outcomes in Adults With T1D Are
16. Addala A, Auzanneau M, Miller K, Maier W, Foster N, Kapellen Impacted More by Continuous Glucose Monitoring Than by Insu-
T, et al. A Decade of Disparities in Diabetes Technology Use lin Delivery Method: 3 Years of Follow-Up From the COMISAIR
and ­HbA1c in Pediatric Type 1 Diabetes: A Transatlantic Com- Study. Diab Care; 43(1):37–43. https://d​ oi.o​ rg/1​ 0.2​ 337/d​ c19-0​ 888
parison. Diab Care. 2021;44(1):133–40. https://​doi.​org/​10.​2337/​ 30. Göldel JM, Kamrath C, Minden K, Wiegand S, Lanzinger S, Sen-
dc20-​0257. gler C, et al. Access to Healthcare for Children and Adolescents
17. Auzanneau M, Eckert AJ, Meyhöfer SM, Heni M, Gillessen A, with a Chronic Health Condition during the COVID-19 Pandemic:
Schwettmann L, et al. Area deprivation and demographic factors First Results from the KICK-COVID Study in Germany. Children.
associated with diabetes technology use in adults with type 1 dia- 2022;10(1):10. https://​doi.​org/​10.​3390/​child​ren10​010010.
betes in Germany. Front Endocrinol. 2023;14:1191138. https://​ 31. Chen J, Yi Q, Wang Y, Wang J, Yu H, Zhang J, et al. Long-
doi.​org/​10.​3389/​fendo.​2023.​11911​38. term glycemic variability and risk of adverse health outcomes in
18. Kroll LE, Schumann M, Hoebel J. Regional health differences patients with diabetes: A systematic review and meta-analysis of
developing a socioeconomic deprivation index for Germany. J cohort studies. Diab Res Clin Pract. 2022;192: 110085. https://​
Health Monitoring. 2017;2(98):114. https://​doi.​org/​10.​17886/​ doi.​org/​10.​1016/j.​diabr​es.​2022.​110085.
RKI-​GBE-​2017-​048.2. 32 Massey RJ, Siddiqui MK, Pearson ER, Dawed AY. Weight vari-
19. Woolford SJ, Sidell M, Li X, Else V, Young DR, Resnicow K, ability and cardiovascular outcomes: a systematic review and
et al. Changes in Body Mass Index Among Children and Adoles- meta-analysis. Cardiovasc Diabetol. 2023;9;22(1):5. https://​doi.​
cents During the COVID-19 Pandemic. JAMA. 2021;326:1434–6. org/​10.​1186/​s12933-​022-​01735-x.
https://​doi.​org/​10.​1001/​jama.​2021.​15036.
20. Pla B, Arranz A, Knott C, Sampedro M, Jimenez S, Hernando Publisher's Note Springer Nature remains neutral with regard to
I, et al. Impact of COVID-19 Lockdown on Glycemic Con- jurisdictional claims in published maps and institutional affiliations.
trol in Adults with Type 1 Diabetes Mellitus. J Endocr Soc.
2020;4:bvaa149. https://​doi.​org/​10.​1210/​jendso/​bvaa1​49.
21. Kim MJ, Lee KH, Lee JS, Kim N, Song JY, Shin YH, et al. Trends
in body mass index changes among Korean adolescents between
2005–2020, including the COVID-19 pandemic period: a national
representative survey of one million adolescents. Eur Rev Med
Pharmacol Sci. 2022;26:4082–91. https://​doi.​org/​10.​26355/​eur-
rev_​202206_​28978.
22. Linardon J, Messer M, Rodgers RF, Fuller-Tyszkiewicz M. A
systematic scoping review of research on COVID-19 impacts on
eating disorders: A critical appraisal of the evidence and recom-
mendations for the field. Int J Eat Disord. 2022;55:3–38. https://​
doi.​org/​10.​1002/​eat.​23640.

13
Journal of Diabetes & Metabolic Disorders

Authors and Affiliations

Marie Auzanneau1,2 · Dorothee M. Kieninger3 · Katharina Laubner4 · Christian Renner5 · Joaquina Mirza6 ·
Gerhard Däublin7 · Kirsten Praedicow8 · Holger Haberland9 · Claudia Steigleder‑Schweiger10 · Bettina Gohlke11 ·
Angela Galler12 · Reinhard W. Holl1,2 · on behalf of the DPV Initiative

8
* Marie Auzanneau Clinic for Children and Adolescent Medicine, Diabetology
[email protected] and Endocrinology, Helios Clinical Centre Aue,
Aue‑Bad Schlema, Germany
1
Institute of Epidemiology and Medical Biometry, CAQM, 9
Paediatric Diabetology, Klinik Für Kinder- Und
Ulm University, Albert‑Einstein‑Allee 41, 89081 Ulm,
Jugendmedizin, Sana Klinikum Lichtenberg, Berlin,
Germany
Germany
2
German Center for Diabetes Research (DZD), Munich, 10
Department of Paediatrics, Paracelsus Medical University,
Neuherberg, Germany
Salzburg, Austria
3
Diabetes Division, Department of Paediatrics, 11
Paediatric Endocrinology and Diabetology, University
Universitätsmedizin Johannes Gutenberg Universität Mainz,
of Bonn, Bonn, Germany
Mainz, Germany
12
4 Charité, Universitätsmedizin Berlin, corporate member
Division of Endocrinology and Diabetology, Department
of Freie Universität Berlin Und Humboldt-Universität
of Medicine II, Medical Center, University of Freiburg,
Zu Berlin, Sozialpädiatrisches Zentrum, Paediatric
Faculty of Medicine, University of Freiburg, Freiburg,
Endocrinology and Diabetology, Berlin, Germany
Germany
5
Pediatric Practice Deggendorf, Deggendorf, Germany
6
Kinderkrankenhaus Amsterdamer Straße, Paediatric
Diabetology, Klinik Für Kinder- Und Jugendmedizin,
Kliniken Köln, Cologne, Germany
7
Children’s Hospital Aurich, Aurich, Germany

13