Authors’ reply

We appreciate the letter by Talbo et al1 regarding our article.2 We would like to thank the authors for their careful and detailed analysis and kindly reply to their concerns.

Polonsky et al3 indeed studied both patients with type 1 and type 2 diabetes. However, the latter group of participants accounted for only 8.8% of the entire sample, which, in our opinion, made this study suitable for analysis. Also, the small number of studies eligible for quantitative analysis encouraged us to include this particular study. Likewise, its exclusion would adversely affect the comprehensiveness of our review. Nevertheless, we undoubtedly should have referred to the issue of a mixed study sample in the manuscript and regret overlooking this. Still, we believe it did not have any impact on the conclusions or overall results of the study.

The original version of the manuscript included the following statement: Conventional self-monitoring of blood glucose (SMBG) was defined as measuring blood glucose by finger-capillary blood sample at least once a day. The glucose level was measured using a blood glucose meter and other types of continuous glucose monitoring system. As the final version was the result of many comments, suggestions, and discussions with reviewers, it is most likely that last part of the second sentence was somehow overlooked at some stage. Yet, this information was still included in Table 1 in the column referring to the control group (Control group [eg. SMBG or blinded CGM]).

The manuscript contains the following sentence: The only exception was the study by Reddy et al (Diabetes Technol Ther, 2018), in which higher scores on the total HFS-II scale, HFS-W and HFS-B subscales, and the Problem Areas in Diabetes Scale (PAIDS) were found in the CGM group. Although we did not say in the main part of the manuscript that the initial level of fear of hypoglycemia was higher in the intervention group, you may still find this information in Supplementary material, Table S5.

Results regarding the quality of life in the study by Hommel et al (Acta Diabetol, 2014) were not presented very clearly and transparently. They are included in Table 2, which is not transparent and fairly difficult to comprehend because it shows change versus baseline for both groups simultaneously: in the sensor-on arm compared with the sensor-off arm. However, in Table 1 you can find results on the self-reported health-related quality of life in children compared with their parents’ proxy ratings. In the study by Charleer et al (J Clin Endocrinol Metab, 2018), the intervention group comprised adults with type 1 diabetes from the Belgian RT-CGM reimbursement program who were on continuous subcutaneous insulin infusion therapy. The control group consisted of patients who had started the treatment before the program was introduced. It is not clearly stated in the study whether these participants ever used a commercial RT-CGM or not before the program started.

The study by Nørgaard et al (Diabetes Technol Ther, 2013) was a 12-month observational study in patients with type 1 diabetes treated with continuous subcutaneous insulin infusion therapy upon the introduction of continuous glucose monitoring systems. It had no control group. The cross-sectional study3 was a questionnaire survey research which met our inclusion criteria.

We would like to thank Talbo et al1 for their valuable insight. Indeed, some aspects presented in the Discussion section and other elements should have been elaborated on more precisely. However, we remain convinced that it did not affect the results and conclusions of our study. We also would like to emphasize that until the moment of publication, there was no meta-analysis that would comprehensively focus on the quality of life and fear of hypoglycemia in adults with type 1 diabetes. The methodology proved to be challenging, and this is probably one of the reasons why no such work was done earlier. We used our best efforts to make our review reliable and appropriate in academic terms. It must be kept in mind, though, that every study carries a risk of bias and methodological inaccuracies.4