key: cord-0991716-hgg0kc4c
authors: D'Onofrio, Luca; Pieralice, Silvia; Maddaloni, Ernesto; Mignogna, Carmen; Sterpetti, Sara; Coraggio, Lucia; Luordi, Cecilia; Guarisco, Gloria; Leto, Gaetano; Leonetti, Frida; Manfrini, Silvia; Buzzetti, Raffaella
title: Effects of the COVID‐19 lockdown on glycaemic control in subjects with type 2 diabetes: the glycalock study
date: 2021-04-06
journal: Diabetes Obes Metab
DOI: 10.1111/dom.14380
sha: 8e061a651637794b0639508b3c97479b62dee5af
doc_id: 991716
cord_uid: hgg0kc4c

AIM: To assess the effect of the coronavirus disease 2019 (COVID‐19) lockdown on glycaemic control in subjects with type 2 diabetes (T2D). MATERIALS AND METHODS: In this observational, multicentre, retrospective study conducted in the Lazio region, Italy, we compared the differences in the HbA1c levels of 141 subjects with T2D exposed to lockdown with 123 matched controls with T2D who attended the study centres 1 year before. Basal data were collected from 9 December to 9 March and follow‐up data from 3 June to 10 July in 2020 for the lockdown group, and during the same timeframes in 2019 for the control groups. Changes in HbA1c (ΔHbA1c) and body mass index (ΔBMI) during lockdown were compared among patients with different psychological well‐being, as evaluated by tertiles of the Psychological General Well‐Being Index (PGWBS). RESULTS: No difference in ΔHbA1c was found between the lockdown and control groups (lockdown group −0.1% [−0.5%−0.3%] vs. control group −0.1% [−0.4%−0.2%]; p = .482). Also, no difference was found in ΔBMI (p = .316) or ΔGlucose (p = .538). In the lockdown group, subjects with worse PGWBS showed a worsening of HbA1c (p = .041 for the trend among PGWBS tertiles) and BMI (p = .022). CONCLUSIONS: The COVID‐19 lockdown did not significantly impact glycaemic control in people with T2D. People with poor psychological well‐being may experience a worsening a glycaemic control because of restrictions resulting from lockdown. These findings may aid healthcare providers in diabetes management once the second wave of COVID‐19 has ended.

group according to the same exclusion criteria used for the lockdown group.

Routine clinical and biochemical data collected during the diabetes visits and used in this study were age, sex, drug therapy, HbA1c, fasting blood glucose, body mass index (BMI), total cholesterol, HDL cholesterol and triglycerides. LDL cholesterol concentrations were estimated with the Friedewald formula: total cholesterol -[HDL + (triglycerides/5)]. Self-reported data about daily physical activity and adherence to the prescribed diet were also collected. Employment status during lockdown was categorized as follows: people who were retired; people who continued their normal work; people who continued their work from home (smart working); and, finally, people who were either unemployed or were housewives/househusbands. Finally, to assess self-perceived psychological health and wellbeing during lockdown, we used the short version of the Psychological General Well-Being Index (PGWBS), validated for the Italian population. 7 The short version of the PGWBS is based on only six questions that explain 90% of the variance of the original questionnaire. 7 The six questions evaluated different psychological domains of anxiety, vitality, depression, self-control and positive well-being 7 ; a higher score represents better psychological health.

Descriptive statistics are presented for categorical variables as numbers with proportions, and for continuous variables as appropriate measures of central tendency and dispersion. Distributions of variables were tested for normality with the Shapiro-Wilk test. ANOVA and Kruskall-Wallis tests were used to analyse differences between groups for parametric and non-parametric continuous variables, respectively. Differences between basal and follow-up (Δ) between the lockdown and control groups, namely, ΔHbA1c, ΔBMI, ΔGlucose, ΔTotal cholesterol, ΔLDL, ΔHDL and ΔTriglycerides, were tested using the t-test for continuous variables with parametric distribution, while the Mann-Whitney test was used for non-parametric variables.

Categorical variables were compared with a χ 2 or Fisher's exact test as appropriate.

For the primary aim of the study, we evaluated differences in ΔHbA1c between the groups. Differences in follow-up HbA1c were also reported for descriptive purposes. Differences in Δ between the groups in all other collected variables were also tested as exploratory analyses. If a statistical difference was observed, differences in the follow-up of that specific variable were also tested.

For subgroup analyses, tertiles of PGWBS, categories of employment status (as specified above), educational level (based on the presence or absence of any degree), type of antidiabetes treatment (treatment with or without multiple daily doses of insulin [MDI] ) and adherence to the prescribed diet and to physical activity, were used.

Based on a clinically relevant change in HbA1c of 0.5% between the lockdown and control groups, we estimated that 123 patients were needed to detect a medium effect size (0.5) in the change of HbA1c from baseline with 80% power. IBM SPSS Statistics v. 21 software was used for data analysis and Prism 7.0 software was used for graphical representations.

The study was performed in accordance with the Declaration of Helsinki, and the study procedures were approved by the institutions' which was slightly higher in the lockdown group, clinical characteristics did not differ between the two study groups (Table 1) . Further, in the lockdown group, 54% of subjects were retired and 21% were unemployed. Among the workers, 8% continued their usual job during the lockdown period, whereas 10% of participants worked from home (smart working). Finally, the proportions of patients who reported continuing a regular physical activity and following a prescribed diet during the lockdown were 41% and 43%, respectively.

No significant difference in terms of ΔHbA1c was found between the lockdown group and the control group (lockdown group −0.1% Table 2 . Accordingly, no absolute difference in follow-up HbA1c was found (lockdown group 7.3% [6.6%-8.0%] vs. control group 7.4% [6.8%-7.9%]; p = .482).

As for the secondary aim of the study, no differences were found in ΔBMI (p = .316), ΔGlucose (p = .538), ΔHDL (p = .142) or ΔTriglycerides (p = .887) between the lockdown and control groups ( 

After subdividing subjects enrolled in the lockdown group by employment categories, only retired patients showed an improvement in HbA1c during the lockdown (prelockdown HbA1c 7.4% [6.8%-8.1%] vs. postlockdown HbA1c 7.3% [6.4%-7.9%]; p = .006), while no differences were observed among workers (p = .146), subjects who were smart working (p = .462), or people who were unemployed (p = .517).

Also, no differences were found in prelockdown and postlockdown HbA1c levels after subdividing patients by education level (patients with a degree, p = .729; patients without a degree, p = .154), treatment with MDI (patients with MDI, p = .684; patients without MDI, p = .071), self-reported physical activity (patients who reported physical activity, p = .182; patients who did not report physical activity, p = .930) or self-reported prescribed diet (patients who reported a correct diet, p = .317; patients who did not report a correct diet, p = .775) during lockdown (Table S1 ).

Most of the studies published during 2020 were focused on the acute effects of COVID-19; few studies explored the indirect effect of the pandemic on specific populations, especially in subjects with T2D. This is one of the first studies investigating the effect of lockdown on glycaemic control in patients affected by T2D. In particular, data observed in people experiencing lockdown measures were compared with data obtained from age-and gender-matched subjects enrolled during the same time period 1 year before the lockdown.

No significant difference in ΔHbA1c was found comparing subjects with stable therapy exposed to lockdown measures compared with controls.

These observations follow the same trend of previous findings in patients with insulin-treated diabetes. 2, 3 Recently, similar results were also obtained among Indian people with T2D, 6 despite the possible differential impact of ethnicity and geography in diabetes management. 8 healthcare providers could reduce the effort in following up stable patients, for example, by extending follow-up appointments by a few months. The spared resources could then be used in the management of patients with T2D and relevant co-morbidities who are at an increased risk of developing a poor prognosis of COVID-19. [10] [11] [12] Interestingly, among subjects exposed to lockdown measures, only retired patients showed an improvement in HbA1c. This finding is consistent with results observed in people with insulin-treated diabetes, 3 especially patients with type 1 diabetes, 13 suggesting that a more stable rhythm of life increases the time available to cope with daily management of diabetes. This speculation could be applicable not only to young people, but also to adults and old people.

We also noticed a higher percentage of metformin users and a lower percentage of sulphonylurea users among people in follow-up during the lockdown period. While this may be the result of the progressive reduction in sulphonylurea prescriptions in Italy, 14 this observation might also highlight the relevance of using drugs with a low hypoglycaemic risk and with possible benefits for better COVID-19 outcomes 15 during the lockdown.

As suggested by our subgroup analysis using a validated questionnaire (i.e. the PGWBS), an ability to cope with the pandemic is strictly related to the psychological stress experienced during lockdown. In this respect, Khare et al. reported that psychological stress because of the pandemic, as assessed by a self-designed questionnaire, was responsible for the worsening of glycaemic control. 9 Psychological distress is commonly observed in countries facing the pandemic and adopting lockdown measures. 16 The presence of chronic illness was one of the risk factors associated with mental distress. 16 In particular, a report from South America observed that patients with diabetes T A B L E 2 Differences in characteristics between the patient and control groups We also chose to stop study recruitment in mid-July to reduce any possible bias as a result of the distance in time from the end of the lockdown. Further, data on employment status, education and adherence to prescribed diet and physical activity were only available for a few patients in the control group, therefore not allowing sufficient controls for subgroup analyses.

The study strengths include the multicentre design, thus reducing the bias of the single-centre study, and the availability of various data about key factors possibly influencing glycaemic control during lockdown. Further, another major strength was that data collected were compared with data from age-and gender-matched subjects enrolled during the same timeframe 1 year before, in order to obtain a control group that was not experiencing the effects of lockdown measures. We also collected information on telemedicine during the lockdown period, excluding possible bias resulting from an intervention by medical staff during the pandemic. Finally, we took advantage of this unique opportunity to study the effect of lockdown on glycaemic control as soon as the lockdown ended in Italy.

In conclusion, this study shows that there are no differences in diabetes control before and after lockdown in patients not requiring intensification of their usual hypoglycaemic treatment. Further, psychological stress could have a detrimental effect on glycaemic control.

These results might be useful to help healthcare providers plan and organize diabetes management in the future once the second wave of COVID-19 has ended.

A tale of two pandemics: how will COVID-19 and global trends in physical inactivity and sedentary behavior affect one another?

Blood glucose control during lockdown for COVID-19: CGM metrics in Italian adults with type 1 diabetes

Effects of COVID-19 lockdown on glucose control: continuous glucose monitoring data from people with diabetes on intensive insulin therapy

Short-term impact of COVID-19 lockdown on metabolic control of patients with wellcontrolled type 2 diabetes: a single-centre observational study

The effect of COVID-19 lockdown on glycemic control in patients with type 2 diabetes mellitus in Turkey

Improved glycemic control amongst people with long-standing diabetes during COVID-19 lockdown: a prospective, observational, nested cohort study

Development and validation of the short version of the psychological general well-being index (PGWB-S)

Frailty and geography: should these two factors be added to the ABCDE contemporary guide to diabetes therapy?

Observational study on effect of lock down due to COVID 19 on glycemic control in patients with diabetes: experience from Central India

Cardiometabolic multimorbidity is associated with a worse Covid-19 prognosis than individual cardiometabolic risk factors: a multicentre retrospective study (CoViDiab II)

COVID-19 in people with diabetes: understanding the reasons for worse outcomes

Clinical features of patients with type 2 diabetes with and without Covid-19: a case control study (CoViDiab I)

Stay-at-home orders during the COVID-19 pandemic, an opportunity to improve glucose control through behavioral changes in type 1 diabetes

Metformin use is associated with reduced mortality rate from coronavirus disease 2019 (COVID-19) infection

Impact of COVID-19 pandemic on mental health in the general population: a systematic review

Mental health in the era of COVID-19: prevalence of psychiatric disorders in a cohort of patients with type 1 and type 2 diabetes during the social distancing

Effects of COVID-19 lockdown on type 2 diabetes, lifestyle and psychosocial health: a hospital-based cross-sectional survey from South India

Effects of the COVID-19 lockdown on glycaemic control in subjects with type 2 diabetes: the glycalock study

We thank Chiara Moretti and Antonio Siena for their help provided during the enrolment of patients. No funding supported this study.

All the authors declare no conflicts of interests related to this manuscript.

LDO: design, conducted data collection, analysis and manuscript writ-