key: cord-0732542-477ru6re authors: Dover, A.R.; Ritchie, S.A.; McKnight, J.A.; Strachan, M.W.J.; Zammitt, N.N.; Wake, D.; Forbes, S.; Stimson, R.H.; Gibb, F.W. title: Assessment of the effect of the COVID‐19 lockdown on glycaemic control in people with type 1 diabetes using flash glucose monitoring date: 2020-08-02 journal: Diabet Med DOI: 10.1111/dme.14374 sha: fa41ffb9d25ae8e94c6dd28cc21a5be8690f33ca doc_id: 732542 cord_uid: 477ru6re AIM: To describe the effect of the stringent lockdown measures, introduced in the UK on 23 March 2020 to curtail the transmission of COVID‐19, on glycaemic control in people with type 1 diabetes using flash glucose monitoring. METHODS: We undertook an observational study of 572 individuals with type 1 diabetes for whom paired flash glucose monitoring data were available between early March and May 2020. The primary outcome was change in flash glucose monitoring variables. We also assessed clinical variables associated with change in glycaemic control. RESULTS: Percentage of time in range increased between March and May 2020 [median (interquartile range) 53 (41–64)% vs 56 (45–68)%; P < 0.001], with associated improvements in standard deviation of glucose (P <0.001) and estimated HbA(1c) (P <0.001). There was a small reduction in the number of individuals meeting the hypoglycaemia target of <5% per day (64% vs 58%; P = 0.004). Comparing changes in flash glucose monitoring data from March to May in 2019 with the same period in 2020 confirmed that these differences were confined to 2020. Socio‐economic deprivation was an independent predictor of a ≥5% reduction in time in range during lockdown (odds ratio 0.45 for people in the two most affluent Scottish Index of Multiple Deprivation quintiles; P <0.001). CONCLUSIONS: Lockdown was not associated with a significant deterioration in glycaemic control in people with type 1 diabetes using flash glucose monitoring. However, socio‐economic deprivation appeared to increase the risk of decline in glycaemic control, which has implications for how support is focused in challenging times. A novel coronavirus strain (COVID-19) was identified as the cause of a severe acute respiratory distress syndrome (SARS-CoV-2), which emerged in Wuhan, China in December 2019 [1] . In the following months, the virus rapidly spread across the world, with a marked increase in cases occurring in the UK from March 2020. In response to this, the UK government enacted a lockdown which included the closure of schools and all non-essential business, effective from 23 March 2020 (when just over 6000 cases were confirmed) [2] . The lockdown also included a directive to the population to 'stay at home' with the exception of essential work, food/medicine shopping and a single episode of daily exercise. Across the UK, diabetes services have generally cancelled all routine 'face-to-face' outpatient activity during the COVID-19 crisis and, where possible, have delivered consultations 'remotely' by telephone and/or video calls [3] . It is conceivable that the major upheaval in diet, physical activity, daily routine and stress associated with COVID-19, in addition to interruption in diabetes services, may have had a negative impact on glycaemic control in people with type 1 diabetes. To assess this, we interrogated our centre's database to compare glycaemic variables in those with type 1 diabetes using flash glucose monitoring, both before and approximately 7 weeks after lockdown (whilst lockdown was still in place). In addition, we explored the clinical and demographic factors associated with decline in glycaemic control across this period. This article is protected by copyright. All rights reserved Approximately 3340 individuals with type 1 diabetes attend Royal Infirmary of Edinburgh and Western General Hospital (Edinburgh) diabetes clinics. A total of 1776 individuals (i.e. approximately half of all people with type 1 diabetes attending our clinics) using flash glucose monitoring (FreeStyle Libre, Abbott, Witney, UK) have linked their glucose data to these clinics using the LibreView platform (libreview.com). Of 1776 LibreView users, we identified a cohort with glucose data uploaded within 7 days of 11 March 2020 (baseline) and again within 7 days of 14 May 2020 (follow-up 7 weeks into lockdown). Only individuals with ≥75% flash glucose monitoring data captured were included (n = 572). An additional analysis, comparing change in glycaemic variables between March to May 2019 and March to May 2020, was possible in 271/572 people (i.e. paired data available for both 2019 and 2020). The main study outcomes were assessment of flash glucose monitoring glycaemic variables between March 2020 and May 2020 (using 2-week data capture), including average glucose, standard deviation of glucose, interquartile range of glucose, coefficient of variation glucose, number of low glucose events per 2-week monitoring period, average low glucose event duration, percentage of time spent with glucose below 3.9 mmol/l, percentage of time spent with glucose in range of 3.9-10 mmol/l, percentage of time spent with glucose above 10 mmol/l, daily flash glucose monitoring scans, percentage of flash glucose monitoring data captured and estimated HbA 1c . The percentage of individuals meeting the international consensus targets for time in range and time below range were also recorded [4] . In addition to this, we compared change across a range of glycaemic variables between March and May 2019 (non-lockdown control period) and March and May 2020 (lockdown period). Clinical and demographic data including age, date of diabetes diagnosis, last clinic HbA 1c value, method of insulin therapy (continuous subcutaneous insulin infusion or multiple daily injections), obesity (BMI > 30 kg/m 2 ), smoking status and socioeconomic deprivation [Scottish Index of Multiple Deprivation (SIMD); simd.scot] were collected from SCI-Diabetes. SIMD data were unavailable in 24 individuals. We also sought to identify independent predictors (clinical and demographic) of decline in glycaemic control following lockdown. This article is protected by copyright. All rights reserved Data were mostly non-normally distributed (as determined by Shapiro-Wilk test) and are presented as median and interquartile range (IQR). Paired data were analysed using the Wilcoxon signed-rank test and unpaired data using the Mann-Whitney U-test. Categorical data were analysed using the chi-squared or McNemar test, when comparing paired repeated measurements. Logistic regression [odds ratios (ORs) and 95% CIs] was performed to identify independent predictors of deterioration in continuous glucose monitoring variables. No sample size calculation was performed. P values < 0.05 were taken to indicate statistical significance and no adjustment was made for multiplicity of statistical tests. All analyses were performed using RSTUDIO version 1.0.153 (https://www.rstudio.com). The study was entirely observational (with no deviation from standard clinical care) and ethics approval was not required. Inclusion in our national clinic database system, SCI-Diabetes (https://www.sci-diabetes.scot.nhs.uk), includes permission to access data for audit purposes and individuals grant specific permission for clinic staff to access LibreView data when linking their devices to the clinic. The median (IQR) age of the cohort was 39 (31-50) years, and the median (IQR) diabetes duration was 18 (9-27) years. Of 572 participants, 301 (53%) were men and 147 (26%) were continuous subcutaneous insulin infusion users. A total of 336 participants out of 548 (61%) belonged to the two least deprived quintiles of SIMD (4 and 5), 143/569 participants (25%) were obese and 41/569 (7.2%) were current smokers. The median (IQR) last clinic-recorded HbA 1c value was 60 (53-67) mmol/mol [7.6 (7.0-8.3)%]. There were statistically significant changes in all reported flash glucose monitoring variables, with the exception of coefficient of variation glucose, low glucose event duration and number of daily flash glucose monitoring scans ( Table 1 ). The direction of all change in glycaemic variables following lockdown was broadly consistent with improved glycaemic control (Table 1 and Fig. 1) , although a small but statistically significant increase in low glucose events and percentage of time with glucose below 3.9 mmol/l was noted. A greater proportion of individuals met the time in This article is protected by copyright. All rights reserved range target and a smaller proportion met the hypoglycaemia target between March and May 2020 ( Table 1) . When comparing changes in flash glucose monitoring variables in the period March to May 2019 with the same period in 2020, there were modest but significant improvements in percentage of time in range, estimated HbA 1c and standard deviation in 2020. There were statistically significant increases in low glucose events and percentage of time below range ( Table 2) . Of 572 participants, 125 (22%) experienced a decrease in percentage of time in range of ≥5% during lockdown. The only statistically significant baseline associations in univariate analysis were lower estimated HbA 1c , greater time in range and greater socio-economic deprivation ( Table 3 ). Logistic regression analysis identified a lower risk in those belonging to the two SIMD Table 4) ]. COVID-19 has resulted in an unprecedented public health emergency and extreme disruption to the lives of people across the world. Whilst there are a number of reasons to suspect that lockdown conditions could have a deleterious effect on glycaemic control in people with type 1 diabetes, this was not observed in a large cohort of individuals using flash glucose monitoring. In fact, these data suggest that, overall, there has been a small but significant improvement in important metrics such as time in range and estimated HbA 1c . Although we have also shown statistically significant increases in low glucose metrics, these were negligible and unlikely to be clinically relevant. Two recently reported smaller series have shown similar results, with no This article is protected by copyright. All rights reserved evident deterioration in glycaemic control during lockdown in Italy and Spain [5, 6] . In fact, the Italian study showed that individuals who were not working during lockdown experienced improvements in key glucose metrics, consistent with the present findings [5] . This suggests that standardization of daily routine may have mediated some of the benefits observed. It is perhaps unsurprising that individuals with higher levels of time in range and lower estimated HbA 1c are at greater risk of subsequent deterioration, as this may simply reflect regression to the mean. Of potentially greater interest is the observation that higher levels of socio-economic deprivation were a strong, independent predictor of deteriorating glycaemic control during lockdown. The negative impact of socio-economic deprivation on glycaemic control is well recognized [7] and there are several plausible explanations as to why the COVID-19 crisis may have exacerbated this effect. For example, socio-economic deprivation may be disproportionately associated with poorer diet and reduced exercise during lockdown, as well as increased stress from childcare commitments and uncertainty around employment and income. Even greater challenges may exist for those not currently using modern glucose monitoring technologies and remote datasharing. These individuals tend to have higher levels of socio-economic deprivation [8] and may be at greater risk of deteriorating glycaemic control in the midst of this crisis. Emerging evidence also suggests that socio-economic deprivation and elevated HbA 1c values are associated with adverse outcomes in individuals with COVID-19 [9] . A key strength of the present study is the cohort size, with the study representing the largest assessment of the effect of COVID-19 lockdown on glycaemic control in type 1 diabetes. An observational study is the only possible method of addressing this research question but raises the possibility that changes observed are not directly attributable to lockdown. We sought to address this potential confounder by comparing changes across the same months (March to May) in both 2019 and 2020, with the former acting as a control. In doing so, we showed that our results are unlikely simply to represent seasonal variation in glycaemic control [10] or incremental improvement in glycaemic control with longer duration of flash glucose monitoring. Another potential criticism of the study is that it is not generalizable to the wider type 1 diabetes population. Flash glucose monitoring, even in the era of National Health Service funding, is skewed towards younger and more affluent individuals [8] and continuous subcutaneous insulin infusion users were also over-represented in the present cohort. Our data were also limited to This article is protected by copyright. All rights reserved people with at least 75% data capture and therefore skewed towards more proactive flash glucose monitoring users. The reassuring findings from this study are likely to be generalizable to populations proactively using flash glucose monitoring, but not beyond this. This study was also limited to interstitial glucose monitoring metrics and we cannot comment on the psychological impact and other person-centred outcomes, which may be equally important to people with diabetes during this crisis. The large sample size identified statistically significant differences in several flash glucose monitoring variables, the magnitude of which would not be considered to be clinically meaningful. In conclusion, these data provide reassurance that people with type 1 diabetes using flash glucose monitoring have not, in general, experienced significant deterioration in glycaemic control during the COVID-19 lockdown; however, the association between deteriorating control and socioeconomic deprivation, even in those using flash glucose monitoring, is concerning. This has important implications for service design to provide more effective support to those who require it most, particularly in challenging times. None. A Novel Coronavirus from Patients with Pneumonia in China Guidance on social distancing for everyone in the UK Accepted Article This article is protected by copyright. All rights reserved Remodelling Diabetes Services and Promoting Innovation in the Time of Covid-19 Clinical Targets for Continuous Glucose Monitoring Data Interpretation: Recommendations From the International Consensus on Time in Range Glycaemic Control Among People with Type 1 Diabetes During Lockdown for the SARS-CoV-2 Outbreak in Italy No deleterious effect of lockdown due to COVID-19 pandemic on glycaemic control, measured by glucose monitoring, in adults with type 1 diabetes Glycaemic control trends in people with type 1 diabetes in Scotland Marked improvement in HbA1c following commencement of flash glucose monitoring in people with type 1 diabetes OpenSAFELY: factors associated with COVID-19-related hospital death in the linked electronic health records of 17 million adult NHS patients. medRxiv Seasonal variation of HbA1c in intensive treatment of children with type 1 diabetes FIGURE 1 Change in (a) estimated HbA 1c , (b) percentage glucose below range Index of Multiple Deprivation Data are median (interquartile range). P compares those in whom TIR fell by ≥5% between March and May 2020 and those in whom this did not occur F.W.G. and A.R.D. have received speaker fees from Abbott Diabetes Care.