Glycemic Control of Patients with Type 1 Diabetes Using an Insulin Pump Before and During the COVID-19–Associated Quarantine

Dear Editor,

The unprecedented experience of the COVID-19 lockdown altered the normality and daily routine of a great proportion of the global population, including patients with type 1 diabetes (DM1). Currently, very few and conflicting data exist regarding the influence of the COVID-19–related quarantine on the glycemic control of these people. The aim of this analysis is to compare glycemic control as obtained by continuous glucose monitoring (CGM) in adults with DM1 using a CGM-augmented insulin pump with or without a closed-loop system, before and during the lockdown period.

Consecutive adult patients with DM1 were recruited at their first postlockdown appointment at the insulin pump outpatient clinic of the participating hospitals (see first three affiliations). CGM recordings up to three months before that visit were downloaded and analyzed. CGM-derived glycemic metrics, defined according to the international consensus on time in range (TIR) recommendations,¹ from three phases were compared: phase 1, the 15-day period immediately before the lockdown date; phase 2, the 15-day period immediately after the lockdown (early lockdown); and phase 3, the 15-day period after phase 2 (late lockdown). TIR_70–180, times below (TBR_<70/TBR_<54) and above range (TAR_>180/TAR_>250), average glucose (AvgGlu), and glucose management index (GMI) were compared between the two postlockdown periods and the prelockdown period, using a paired testing statistical approach. All patients signed an informed consent to allow for CGM data download. The study was performed according to the declaration of Helsinki principles and was approved by the ethics committees of all participating hospitals.

A total of 46 patients (16 men), using a CGM device for >80% of the requested time periods, (mean age 38.2 [standard deviation: 12.9] years, median DM1 duration 19.5 [interquartile range: 12–28] years), were included. Thirty-six patients had been using a closed loop system designed to “suspend before low” (Medtronic MiniMed 640G/Enlite^a), whereas 10 had been using an open loop system with various types of insulin pump and the same flush CGM device (Freestyle libre, Abbott^b).

The comparisons between CGM-obtained glycemic metrics are given in Table 1. TIR_70–180 was significantly longer in late lockdown, as compared with prelockdown phase. Overall, glucose values tended to be lower during both quarantine phases, as indicated by the lower AvgGlu, the lower GMI, the longer TBR_<70, the shorter TAR_>180 and TAR_>250, and the longer times spent in automatic suspension (in closed-loop users) (Table 1). After applying Bonferroni adjustment for multiple comparisons, only the GMI and the time in suspension mode remained statistically significantly different during the two quarantine periods. No significant difference was observed among the three periods regarding the dose of insulin (either basal or bolus) and the amount of consumed carbohydrates. None of the patients was ever diagnosed with COVID-19.

Our retrospective analysis shows that adults with DM1 treated by advanced technology systems did not experience a deterioration of glycemic control during the quarantine period. On the contrary, CGM-obtained indices during both the early and the late quarantine phases pointed toward lower glucose values. Importantly, in patients using a “sustain before low” system, the time spent in automatic suspension increased by ∼30%–40% during the total lockdown period, indicating a higher tendency toward hypoglycemia and lower insulin needs. In this sense, it seems reasonable to hypothesize that closed-loop systems offered protection against hypoglycemia to their users. Hitherto, in two reports from Italy^2,3 (a country heavily affected by COVID-19), glucose control improved slightly during the quarantine to a very similar extent as to our report. In another study from India, however, in DM1 patients not using advanced insulin delivery and/or glucose monitoring systems, glucose control deteriorated.⁴ It seems plausible that the direction of glycemic control during such a disastrous situation may differ depending on several parameters related to the catastrophe itself and its consequences as well as on the individual characteristics of the patient.

The population of our study is characterized by good diabetes education, use of advanced technology, good glycemic control, and a feeling of security due to the very low incidence of COVID-19 in Greece. This combination of events might have been responsible for tilting the balance of glycemia toward lower values. Important limitations of this analysis are its retrospective design, the use of different pumps/CGM devices, and its reflection of a single health care system.

In conclusion, in a country with low prevalence of COVID-19 but early implementation of quarantine measures, patients with DM1 did not experience a deterioration of glycemic control. Lower glycemic levels were observed during quarantine, as well as a tendency toward hypoglycemia that was successfully anticipated by an increase of the automatic insulin suspension time of the closed loop systems.