Impact of COVID-19 lockdowns on glycemic control across age groups based on a retrospective isCGM analysis from Poland 2020–2021

Abstract

Most European countries experienced three COVID-19 waves in 2020–2021. In Poland, Wave 1 (March-July 2020) was associated with a severe lockdown prohibiting external social activities beyond every-day necessities. During Waves 2 and 3 some restrictions were reimplemented, however, impacted everyday life less. The study aimed to analyze the individual as well as long-term impacts of COVID-19 pandemic lockdowns on glycemic control in different age groups of people with diabetes using intermittently scanned continuous glucose monitoring (isCGM), in Poland. We analyzed retrospective data from isCGM users during the COVID-19 pandemic (from January 2020 to September 2021), describing glucometrics according to the International Consensus on time in range (TIR). Longitudinal data were analyzed from 680 Polish patients with diabetes, comprising 470 adults aged 18–64, 66 adults aged 65+, and 144 children and adolescents. The most evident improvement in glycemic indices was observed during the first, most severe lockdown, especially in adults (TIR 68.2 vs. 65.6%). For seniors, most indices were stable during the pandemic, with only small improvement in time below range (TBR) (TBR70 1.9 vs. 2.6%). The comparison of the pre-pandemic and post-lockdown periods showed that most glycemic indices returned to similar levels, however, in children and adolescents, some deterioration was seen. Changes in glycemic indices during the COVID-19 pandemic in 2020–2021 in patients using isCGM differed between the 3 lockdown events and between age groups. Results demonstrate an age dependency in users’ response to the COVID-19 lockdowns in Poland and suggest that behavior changes brought on by dramatic lifestyle changes associated with COVID lockdowns were not sustained following the pandemic.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-20323-z.

Introduction

Coronavirus disease 2019 (COVID-19) is an infectious disease caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)^1–4. On March 11, 2020, the World Health Organization (WHO) declared the outbreak of COVID-19 to be a global pandemic⁵. In many European countries, the number of reported cases indicated a three-wave pattern of the COVID-19 pandemic in 2020–2021, with the first wave occurring from March to July 2020, the second during the last quarter of 2020, and the third from February to May 2021 (Fig. 1)⁶.

Fig. 1.

Number of confirmed COVID-19 cases and deaths related to COVID-19 during the first three waves of the COVID-19 pandemic in Poland (2020–2021). Based on WHO data⁶.

In Poland, the first case of laboratory confirmed COVID-19 was diagnosed on March 4, 2020^7,8. Wave 1 quickly followed and was characterized by particularly severe restrictions prohibiting all external social activities except those involving every-day necessities. Wave 2 of the COVID-19 pandemic began in October 2020, resulting in re-implementation of some of restrictions, however, their impact on everyday life was smaller than it was during Wave 1, moreover, people were more likely not to follow COVID-19 rules. Similarly, in March 2021, some limitations to social life were implemented again as a result of Wave 3, however, restrictions were again much less severe than they were during the first lockdown^9,10. Detailed descriptions on governmental policies concerning COVID-19 in Poland were previously published¹⁰. In the European Union, vaccinations against COVID-19 were introduced in December 2020. In Poland, adult patients with diabetes were among the first groups of people to be vaccinated and starting from 2021, those who were vaccinated were allowed to follow fewer restrictions than the rest of the society^9–11.

The COVID-19 lockdowns in Poland resulted in substantial reduction of in-person medical care, as many specialist out-patient services were switched to telemedicine¹². However, a single-center observation from Poland showed that for patients with type 1 diabetes (T1DM) in outpatient clinic settings, the COVID-19 pandemic was not associated with deterioration of glycemic control as could have been anticipated.

One potential explanation for this phenomenon of sustained glucose control during the pandemic is the availability of glucose monitoring technology, which can help users balance the increased difficulty of accessing stationary health care with more intensive and careful treatment and glucose self-monitoring, either with a glucose meter or with a continuous glucose monitoring system (CGM). Self-monitoring can furthermore be assisted by the recently introduced capability to remotely share data with a healthcare provider via cloud services, some examples of which include LibreView (Abbott, USA), CareLink (Medtronic, USA), Clarity (Dexcom, USA) or DMS (Ascensia, Switzerland).

Nevertheless, analyses on glucose control across multiple individual lockdowns are lacking. The aim of the study was to further analyze the individual and long-term impacts of the COVID-19 pandemic lockdowns on glycemic control in different age groups of people with diabetes using isCGM in Poland, 2020–2021, in order to assess users’ long-term ability to sustain glucose control over multiple lockdown events.

Materials and methods

Subjects

The FreeStyle Libre system is an isCGMS with a sensor that measures interstitial glucose levels for up to 14 days¹³. Data collected by the FreeStyle Libre Reader or LibreLink App can be uploaded to the LibreView cloud to generate the Ambulatory Glucose Profile and other personal reports. To use LibreView, patients provide their nationality and age (but no clinical and personal information) during account creation, at which time they may also consent to sharing their device data for research purposes. Once the patient’s consent is obtained, their data are altered such that personal identifying information is removed, enabling data to be collected safely and anonymously, and then stored in a cloud database. Thus, the dataset used in this study was built from the anonymous data that were stored in LibreView after patients’ consent for data uploading and storing was obtained. While users’ nationalities and age groups are included in the de-identified LibreView data, no other information, including sex, type of diabetes, diabetes duration, type of hypoglycemic therapy, education, socioeconomic status, or reimbursement status were available to the authors.

Time points

For this study, the de-identified longitudinal CGM data was collected and analyzed from LibreView users in Poland over a period of 21 months, from January 2020 to September 2021. Users’ sensor data were collected within the 45-day periods at the start of each quarter in 2020 and first three quarters in 2021, for a total of 7 timepoints: 20Q1 (1 Jan 2020–15 Feb 2020), 20Q2 (1 Apr 2020–15 May 2020), 20Q3 (1 Jul 2020–15 Aug 2020), 20Q4 (1 Oct 2020–15 Nov 2020), 21Q1 (1 Jan 2021–15 Feb 2021), 21Q2 (1 Apr 2021–15 May 2021), and 21Q3 (1 Jul 2021–15 Aug 2021). The selection criteria also required that analyzed users have sufficient data (> 120 h of readings) within every 45-day period, for an apples-to-apples comparison between each time point. The analyzed time periods were selected at these specific points in time and at regularly spaced intervals to capture the glucose control of Polish users between important COVID-related public health events, with sufficient time between periods for users to adjust their diabetes management behavior accordingly. The generous size of these time periods (45 days) also (1) permits more users into the analysis who might not have continuous sensor utilization (and who therefore have time gaps in their sensor data) and (2) allows the analysis to gather more data per user in each time period, increasing confidence that the data selected for each user is indeed representative of their overall behavior and glucose control.

In this study, 20Q1 therefore represents the pre-pandemic period, 20Q2 corresponds to the severe lockdown of Wave 1, 20Q3 to the period during which the lockdown restrictions were lifted, and 20Q4 to the more liberal lockdown of Wave 2. 21Q1 corresponds to a period between lockdowns with restrictions lifted, 21Q2 to the hard lockdown of Wave 3, and 21Q3 to the period following the end of Wave 3¹⁰. The relation between time points, pandemic waves and public health events is summarised in Supplementary Table 1¹⁰.

CGM-delivered data

Within each time point, data were aggregated at the user-level to measure each user’s average glucose, glucose standard deviation (SD), CV, TIR 70–180 mg/dL, TAR > 180 mg/dL (TAR180), TAR > 250 mg/dL (TAR250), TBR < 70 mg/dL (TBR70), and TBR < 54 mg/dL (TBR54). Mean average glucose, SD, CV, TIR, TAR180, and TAR250, as well as median TBR70 and TBR54 were then aggregated according to the age ranges 6–17 (pediatrics), 18–64 (adults), and 65+ (seniors) at each timepoint. Analyzed glucose ranges were defined in accordance with the International Consensus on Time in Range¹⁴.

Data analyses

To check the immediate impacts of the three COVID-19 lockdowns on glycemic control, users’ mean change in these metrics were evaluated from Q1 to Q2 2020, from Q3 to Q4 2020, and from Q1 to Q2 2021. Next, differences in glycemic metrics between the pre-pandemic period (20Q1) and the post-lockdown period (21Q3) were assessed to evaluate users’ overall change in glucose control during the pandemic. The changes over these selected evaluation periods were compared within each age group to assess the relative impact of each event in time on that age group. Within each evaluation period, the changes were also compared between age groups to assess how similar or different the age groups’ responses were to the corresponding event.

For hypothesis testing, p < 0.05 was considered significant. For each analyzed metric in each age group, a difference for any time point was first tested with a repeated-measures ANOVA, and if significant, was followed by pairwise t-tests between individual time points (to determine which timepoint had a difference). For each metric tracked across any two timepoints, age groups’ changes in that metric were compared by first conducting a one-way ANOVA, followed by post-hoc independent samples t-tests.

Bioethics

The data analyzed in this retrospective, observational study was de‑identified and originally collected in routine clinical care and under EULA terms. No deviations from standard care occurred and no identifiable patient data were accessed. The study was exempt from requiring ethics approval (the Bioethics Committee at the Regional Medical Chamber in Krakow, Poland). All research was conducted in compliance with the Declaration of Helsinki¹⁵.

Results

CGM data across 2020–2021 by age group

Longitudinal data from 680 Polish patients with diabetes were available for analysis, comprising 470 adults aged 18–64, 66 adults aged 65 or more, and 144 children and adolescents (or pediatric users). Figure 2 tracks their selected glycemic markers, aggregated by age group, across the 7 timepoints of the study period (2020–2021). These results, including their hypothesis testing results, are compiled in greater detail in Table 1. Number of sensor wear days across the 7 timepoints by age group is shown in Supplementary Table 2.

Fig. 2.

Glycemic indices in isCGM users across seven quarters of the COVID-19 pandemic in Poland: (A) mean glucose, (B) TIR, (C) TBR70, (D) TBR54, (E) TAR180, (F) TAR250, (G) SD, (H) CV, (I) daily scans.

Table 1.

Glycemic indices in IsCGM users across the seven quarters of the COVID-19 pandemic in poland, by age group.

	Age group	Number of cases	20Q1	20Q2	20Q3	20Q4	21Q1	21Q2	21Q3	ANOVA, p	Mean Change 20Q2-20Q1	p	Mean Change 20Q4-20Q3	p	Mean Change 21Q2-21Q1	p	Mean Change 21Q3-20Q1	p
Average glucose [mg/dl]	pediatrics	144	153.2	152.5	154.7	158.6	158.5	156.5	157.2	< 0.001	−0.7	0.640	+ 3.9	0.002	−2.0	0.084	+ 4.0	0.054
	adults	470	155.6	153.2	154.4	155.6	154.2	154.7	157.5	< 0.001	−2.4	0.004	+ 1.2	0.091	+ 0.4	0.538	+ 1.9	0.078
	seniors	66	149.1	151.8	151.9	152.2	151.9	149.4	150.2	0.560	+ 2.7	N/A	+ 0.3	N/A	−2.5	N/A	+ 1.1	N/A
Glucose Standard Deviation [mg/dL]	pediatrics	144	59.1	58.1	61.1	60.6	61.3	62.3	64.4	< 0.001	−1.0	0.188	−0.5	0.459	+ 0.9	0.144	+ 5.3*,’	< 0.001
	adults	470	55.4	52.8	54.4	53.1	53.2	54.1	56.3	< 0.001	−2.6	< 0.001	−1.4	< 0.001	+ 0.9	0.002	+ 0.8*	0.108
	seniors	66	44.8	44.1	44.1	44.3	43.4	43.0	44.0	0.603	−0.7	N/A	+ 0.2	N/A	−0.4	N/A	−0.9’	N/A
CV [%]	pediatrics	144	38.1	37.7	39.1	37.9	38.3	39.5	40.6	< 0.001	−0.4	0.162	−1.2	< 0.001	+ 1.2*,’	< 0.001	+ 2.4*,’	< 0.001
	adults	470	35.2	34.1	34.9	33.8	34.2	34.7	35.5	< 0.001	−1.1	< 0.001	−1.0	< 0.001	+ 0.5*	< 0.001	+ 0.3*	0.245
	seniors	66	29.6	28.6	28.5	28.5	28.2	28.3	28.8	0.054	−1.1	N/A	+ 0.0	N/A	+ 0.1’	N/A	−0.8’	N/A
TBR54 [%]	pediatrics	144	1.41	1.25	1.25	1.10	1.05	1.56	1.68	< 0.001	−0.16*	0.270	−0.15	0.131	+ 0.52*,’	< 0.001	+ 0.26*,’	0.181
	adults	470	1.39	0.66	0.72	0.54	0.60	0.76	0.84	< 0.001	−0.74*,^	< 0.001	−0.18	< 0.001	+ 0.16*	0.012	−0.55*	< 0.001
	seniors	66	0.52	0.23	0.13	0.16	0.13	0.15	0.17	< 0.001	−0.29^	0.042	+ 0.03	0.449	+ 0.02’	0.614	−0.35’	0.006
TBR70 [%]	pediatrics	144	5.2	5.1	5.4	4.8	4.7	6.0	6.3	< 0.001	−0.1	0.729	−0.5	0.033	+ 1.3*,’	< 0.001	+ 1.1*,’	0.006
	adults	470	4.9	4.2	4.6	4.0	4.2	4.5	4.8	< 0.001	−0.7	< 0.001	−0.5	< 0.001	+ 0.4*	0.008	−0.18*	0.365
	seniors	66	2.6	1.9	1.5	1.6	1.8	2.0	2.1	0.010	−0.7	0.018	+ 0.1	0.482	+ 0.2’	0.512	−0.5’	0.262
TIR [%]	pediatrics	144	66.3	66.9	65.4	63.7	63.9	63.4	62.6	< 0.001	+ 0.6*	0.368	−1.8	0.019	−0.5	0.342	−3.7*,’	< 0.001
	adults	470	65.6	68.2	67.0	66.9	67.6	66.7	64.7	< 0.001	+ 2.5*	< 0.001	−0.2	0.625	−0.9^	0.006	−0.9*	0.107
	seniors	66	73.8	73.9	74.1	73.2	73.4	75.4	74.2	0.788	+ 0.1	N/A	−0.9	N/A	+ 2.1^	N/A	+ 0.4’	N/A
TAR180 [%]	pediatrics	144	28.5	28.0	29.2	31.5	31.4	30.6	31.1	< 0.001	−0.5	0.519	+ 2.3	0.003	−0.8	0.212	+ 2.6	0.018
	adults	470	29.4	27.6	28.4	29.1	28.2	28.8	30.5	< 0.001	−1.8	< 0.001	+ 0.7	0.072	+ 0.5	0.136	+ 1.1	0.069
	seniors	66	23.7	24.2	24.4	25.2	24.9	22.6	23.7	0.630	+ 0.6	N/A	+ 0.8	N/A	−2.3	N/A	+ 0.0	N/A
TAR250 [%]	pediatrics	144	9.1	8.3	9.5	10.5	10.4	10.0	10.6	< 0.001	−0.8	0.146	+ 1.0	0.041	−0.4	0.288	+ 1.5	0.031
	adults	470	9.5	8.4	8.9	8.7	8.6	8.8	10.1	< 0.001	−1.1	< 0.001	−0.2	0.483	+ 0.2	0.534	+ 0.6	0.128
	seniors	66	5.4	5.9	5.7	5.9	6.1	4.9	4.8	0.333	+ 0.5	N/A	+ 0.2	N/A	−1.2	N/A	−0.6	N/A
Daily scans [n/day]	pediatrics	144	15.4	13.8	14.0	12.6	12.6	12.4	12.8	0.004	−1.6	0.450	−1.5	0.614	−0.2	1.000	−2.6	0.041
	adults	470	15.7	14.5	14.9	13.6	13.7	13.7	14.2	0.054	−1.2	N/A	−1.3	N/A	0.0	N/A	−1.5	N/A
	seniors	66	12.9	11.3	11.4	10.5	10.5	10.0	10.3	0.137	−1.5	N/A	−0.9	N/A	−0.5	N/A	−2.5	N/A

Data shown as means. Significant differences between age groups within the time periods are marked by *,^,’ for differences between pediatrics and adults, adults and seniors, and pediatrics and seniors, respectively.

Differences between age groups

Wave 1

The Q1/2020 – Q2/2020 evaluation period resulted in significant one-way ANOVAs across age groups, for their changes in TBR54 (p < 0.001) and TIR (p = 0.02). A difference in the change in TBR54 was detected for Adult users (p < 0.001 and p = 0.006 for the t-tests against the changes measured for Peds and Seniors, respectively). All 3 age groups showed an average decrease in TBR54, with the Adult users showing the greatest decrease of −0.74%. A difference in the change in TIR was also detected between Pediatric and Adult users (p = 0.02), with Pediatric users gaining + 0.6% TIR on average, in contrast with the + 2.5% mean increase in TIR for Adult users. The age groups’ mean differences over Q1/2020-Q2/2020 are again found in Table 1.

Wave 2

The one-way ANOVA conducted across age groups for the change in the Q3/2020-Q4/2020 evaluation period was not significant.

Wave 3

The Q1/2021 – Q2/2021 evaluation period results in significant one-way ANOVAs across age groups, for their changes in glucose CV (p = 0.03), TBR54 (p = 0.02), TBR70 (p = 0.006), TIR (p = 0.01), and TAR180 (p = 0.01). A difference in the change in Glucose CV was detected for Pediatric users (p = 0.03 for both t-tests against the changes measured for Adults and Seniors), who increased their CV by + 1.2%, in contrast with the CV increases of + 0.5% and + 0.1% of adults and seniors, respectively. For TBR54, a difference in its change was again detected for Pediatric users (p = 0.03 and p = 0.001 for the t-tests against the changes measured for Adults and Seniors, respectively), who again increased their TBR54 by + 0.52%, whereas adults and seniors measured small average increases of + 0.16% and + 0.02%, respectively. Pediatric users similarly differed from the other age groups regarding their change in TBR70 over this period, with p = 0.007 and p = 0.01 for the t-tests against the changes measured for Adults and Seniors, respectively. While Adults and Seniors showed small average increases in TBR70 by + 0.4% and + 0.2%, respectively, Pediatric users demonstrated a slightly greater increase in TBR70 by + 1.3%. A difference in the change in TIR over this period was detected between Seniors and Adults (p = 0.04)—whereas Adults experienced an average decrease in TIR by −0.9%, Seniors increased mean TIR by + 2.1%. For TAR180, a difference was again only detected between Seniors and Adults (p = 0.05). Adults increased TAR180 by + 0.5%, in contrast with the − 2.3% decrease experienced by Seniors.

Pre-pandemic vs. post-lockdowns

The Q1/2020-Q3/2021 evaluation period resulted in significant one-way ANOVAs across age groups, for their changes in Glucose SD (p < 0.001), Glucose CV (p < 0.001), TBR54 (p < 0.001), TBR70 (p = 0.005), and TIR (p = 0.02). A difference was detected for the change in Glucose SD of Pediatric users (p < 0.001 for both t-tests against Adults and Seniors), who showed an overall increase of + 5.3 mg/dL, contrasting with the smaller changes of + 0.8 mg/dL and − 0.9 mg/dL in the Adult and Senior groups. Similarly, a difference was also detected for the change in Glucose CV of Pediatric users (p < 0.001 again for both t-tests against Adults and Seniors), where Pediatric users increased their CV by + 2.4%, again in contrast with the smaller changes of + 0.5% and − 0.8% of Adults and Seniors, respectively. A difference from the other age groups was detected for the change in TBR54 of Pediatric users (p < 0.001 and p = 0.008 for the t-tests against Adults and Seniors, respectively) and a difference was similarly detected for the change in TBR70 of Pediatric users (p = 0.004 and p = 0.006 for the t-tests against Adults and Seniors, respectively). Pediatric users increased their TBR54 and TBR70 by + 0.26% and + 1.1%, respectively, whereas Adults had overall decreases of −0.55% and − 0.18% (in TBR54 and TBR70, respectively) and Seniors had overall decreases of −0.35% and − 0.5% (again in TBR54 and TBR70, respectively). We also detect a difference from the other age groups in the change in TIR of Pediatric users (p = 0.01 and p = 0.03 for the t-tests against Adults and Seniors, respectively)—whereas Adults and Seniors showed small overall changes of −0.9% and + 0.4%, respectively, Pediatric users experienced an overall decrease of −3.7% TIR.

Discussion

In this retrospective, anonymous database study, we conduct the first examination of the individual impact of Wave 1, Wave 2, and Wave 3 of the COVID-19 pandemic on glycemic control in isCGM users of different ages. We show that changes in glycemic indices could have been dependent on the severity of lockdown and age group. Our results indicate that the most improvement in glycemic indices was observed during the first, most severe lockdown, especially in adults. From Q1/2020 to Q2/2020, the adult age group appeared to make favorable changes in every metric examined, whereas for seniors, changes were detected in only their TBR54 and TBR70. This could probably be explained by subjects’ pre-pandemic lifestyles, as working adults (and therefore their diabetes management behaviors) were impacted more heavily by the first COVID-19 wave. A study from Italy compared isCGM data from patients who stopped working with that of patients who continued working during the lockdown, showing that quitting work during the lockdown (suggesting a slowing of routine daily activities) could have been, at least in the short-term, beneficial in terms of diabetes management¹⁶.

Additionally, there are published studies on glycaemic control in patients using intermittently scanned CGM (isCGM) from different populations (France, Italy, Greece, Spain, UK), where various strategies against the COVID-19 pandemic were implemented. Most studies showed no deterioration in glycaemic control, and some even indicated short-term improvement^17–22. A meta-analysis of 70 observational studies on glycaemic control in adults with T1DM diabetes using isCGM/real-time CGM (rtCGM) during the lockdown due to COVID-19 furthermore demonstrated improvements in mean glucose, glucose management index (GMI), coefficient of variation (CV), time in range (TIR), and time above range (TAR), but not time below range (TBR)²³.

Despite published studies showing improvement in glycemic control during lockdowns; analyses that further track and compare glucose control across multiple individual lockdown events are lacking²³. In our study, we show CGM-delivered data for each individual lockdown, as well as for the periods following those lockdowns. After 1.5 years of the COVID-19 pandemic, glycemic control for seniors was similar to its measurement before the outbreak, but a contrasting result for children and adults was observed that suggested deterioration of their glucose control, a result that appears at first glance to be contrary to the previous publications demonstrating sustained glucose control during the pandemic. The differing responses of age groups to subsequent COVID-19 events over the long-term reveal greater complexity in the relationship between lockdown restrictions and real-world diabetes management (beyond the first and most severe lockdown event), and warrant further study to better understand the limitations of lifestyle-change practices with regards to long-term diabetes management. Our study found in general that seniors had better glycemic control than did adults and pediatric users, the latter of which seemed to have the worst glycemic control. However, among the many possible explanations for this observation, the contexts of mode of therapy and sensor reimbursement policy in Poland are particularly important - from 2020 to 2021 in Poland, only T1DM patients aged 18 or less were reimbursed for their sensors While we could therefore expect that most pediatric users were T1DM patients, we must consider that the adult and senior groups likely included some patients with T2DM, most of which are likely using intensive insulin therapy, but also some of which might be using less intensive hypoglycemic intervention, and all of whom were required to pay the full cost of sensors out-of-pocket. For such users without reimbursement, this decision to pay for sensors out-of-pocket suggests a pro-active and motivated attitude towards diabetes management that likely also drives improved glucose control, and we may therefore also see some impact of this selection bias in the earlier evaluation and comparison of age groups.

The potential therapy differences between age groups may be further suggested by their initial difference in time in hypoglycemia—pediatric users in Q1/2020 had a median TBR54 of 0.72%, in contrast with the 0.11% median TBR54 of senior users in that same period. Figure 2 illustrates how TBR54 decreases for children and adults beginning with Wave 1 (−0.16% and − 0.74% for children and adults, respectively), and that this decrease is generally sustained until Wave 3. While the 3rd COVID wave also included a lockdown, its critical difference was that the COVID vaccine had become available in Poland several months prior (December 28, 2020), effectively at the start of Q1/2021, and so it is possible that some of the analyzed users were subjected to fewer COVID restrictions as a result of their vaccination status - by September of 2021 the majority of people (~ 60%) in Poland were vaccinated, amounting to 23 million people²⁴. Pediatric users’ mean change in TBR54 from this time to Q2/2020 was + 0.52%, a reversal of the preceding trend of decreasing TBR54.

Although there again are many potential explanations for this observation, it does raise the possibility that the first 2 lockdowns caused a lifestyle change that increased the quality or efficiency of insulin dosing, and also that the vaccination (which permitted people to resume normal day-to-day activities) then enabled people to revert to old habits correlated with lower quality insulin dosing. Investigating this possibility could yield further insights for the standard of diabetes care in Poland. For future pandemics it would be beneficial to have standardized guidelines on diabetes management not only for lockdowns but for time when restrictions will be lifted and deterioration in glucose control could be expected.

Our study should be interpreted in the context of its strengths and limitations. Firstly, this is a retrospective and observational database analysis prone to many biases typical for this type of research. Additionally, while we analyzed 21-months of longitudinal data for several hundred isCGM users, limited clinical data were available. For example, no data on type of diabetes, mode of therapy, daily activities, and reimbursement status could be assessed. Furthermore, neither the extent of in-person or remote care (e.g. via data-sharing in LibreView) nor incidence of COVID-19 in analyzed subjects within the study period were available. Next, we analyzed data only from patients whose CGM-delivered data were available for all analyzed periods. Thus, patients who started or quit using isCGM during the pandemic, or patients whose sensor utilization was highly irregular were not included in analyses. This could have influenced the results of the study, as we cannot exclude that our studied group represents a specific group of patients, reducing the generalizability of our results towards those users who did not demonstrate the same commitment to or need for glucose monitoring. Moreover, as the number of subjects in the Senior group was relatively small as compared to groups of younger subjects, their measurements of glucose control could be underpowered. Finally, from a clinical point of view, even improvement in glycemic control should be analyzed with caution, as some of this effect could have been accompanied by less physical activity and weight gain. Moreover, the effects of the lockdowns on mental health were not examined. The deterioration in glycemic control observed in children and adolescents in 2021 is also an observation that appeared typical of that age group even prior to COVID-19²⁵; however, in light of the complex lifestyle impact of events like the COVID-19 pandemic, the age group’s actual behavior under such conditions still requires further examination and careful assessment.

Conclusions

In summary, the successive lockdowns due to the COVID-19 pandemic waves in Poland had different impacts on glycemic control in patients with diabetes using is-CGM, which appeared to be dependent on age group. The severe lockdown of Wave 1 resulted in an initial improvement in glycemic indices for adults with diabetes, consistent with previous research regarding the onset of the COVID-19 pandemic, but this improvement ultimately was not sustained after subsequent lockdown events. In contrast, seniors showed very little change in glucose control over the course of the pandemic. These dependencies between age group and severity of lockdown raise important questions about the ideal implementation and rate of behavior change that, when studied in more depth, can inform the standard of diabetes care. Recently expanded data-sharing features of CGM (i.e. telehealth), users’ pre-pandemic lifestyles, their diabetes and therapy type, and their reimbursement status were all likely to be important factors and should also be studied further to assist the development of diabetes care in this regard. Long-term effects of glycemic control should be analyzed with caution - while most indices were similar when comparing the post-lockdown to the pre-pandemic period, in some cases clear deterioration was observed, revealing wide complexity in the lifestyle and behavior impacts of the COVID-19 lockdowns in Poland on real-world diabetes management.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

COVID

19-Coronavirus disease 2019

CV

coefficient of variation

GMI

glucose management index

isCGM

intermittently scanned continuous glucose monitoring

rtCGM

real time continuous glucose monitoring

TAR180

time above range > 180 mg/dL

TAR250

time above range > 250 mg/dL

TBR54

time below range < 54 mg/dL

TBR70

time below range < 70 mg/dL

TIR

time in range

Author contributions

JH: Design, Conduct/data collection, Analysis, Writing manuscript; KK: Design, Conduct/data collection, Analysis, Writing manuscript; LB: Design, Supervision; MTM: Design, Supervision.

Funding

This study was funded by Abbott Diabetes Care, Alameda, CA, USA.

Data availability

The datasets used during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

KK and LB are employees of Abbott Diabetes Care, Alameda, CA, USA. JH and MTM have received fees from Abbott Diabetes Care, Dexcom, Ascensia for lecturing and participating in the advisory panels.

Bioethics committee approval

The data analyzed in this retrospective, observational study was de‑identified and originally collected in routine clinical care and under EULA terms. No deviations from standard care occurred and no identifiable patient data were accessed. The study was exempt from requiring ethics approval (the Bioethics Committee at the Regional Medical Chamber in Krakow, Poland). All research was conducted in compliance with the Declaration of Helsinki.