Information about insulin detemir in GDM is scarce and originates mainly from studies on pregestational diabetes, T1DM, or T2DM or both pregestational and gestational diabetes [6,7,8,9, 15,16,17,18, 20,21,22,23,24,25,26]. To our knowledge, our study is the first to investigate exclusively insulin NPH versus detemir in a peer cohort of women with real GDM diagnosed from the 24th to 28th week of gestation. Insulin detemir was found to be equally effective compared to NPH with no differences observed with respect to maternal adverse outcomes or neonatal complications. In our retrospective cohort, all women received a single bedtime injection of NPH or detemir per day in order to correct fasting BG, without any additional prandial insulin. Postprandial BG was controlled with an appropriate gestational diabetes diet.

Traditionally, improvements in glycemic control during insulin therapy are associated with weight gain [27]. In our study, there were no statistical differences concerning maternal weight gain (both per week and throughout pregnancy) between the two groups. These results are in agreement with most other observational cohorts or RCTs comparing the use of detemir versus NPH in pregnancies with T1DM, T2DM, or GDM [8, 15, 16]. Only one recent RCT in pregnant women with T2DM by Bartal et al. found that insulin detemir was associated with less weight gain compared to NPH [22].

In our cohort, no cases of hypoglycemia were recorded in either group, this probably due to the fact that there was no need for intensification of insulin treatment. Insulin detemir or NPH was injected once daily, while there was no need for rapid-acting insulin. In two observational studies investigating insulin NPH versus detemir in pregnancies with T2DM or GDM, the rate of hypoglycemia between the two groups was comparable [21, 24]. However, in the RCT by Herrera et al. in 87 pregnant women with GDM and T2DM, NPH was associated with a higher risk of hypoglycemia compared to insulin detemir. In line with this, there are other studies investigating detemir versus NPH in pregnancies with pregestational and gestational diabetes, possibly due to the intensive insulin regimen including short-acting insulins [15, 16, 22].

Furthermore, no hypertensive disorders were recorded in either group. Previous studies comparing insulin detemir versus NPH in pregnancies with T1DM, T2DM, or GDM have shown comparable rates of hypertension [9, 16, 21, 24]. Only the aforementioned RCT in pregnant women with T2DM observed lower adverse maternal outcomes, including hypertensive disorders, in the detemir group compared to the NPH group [22].

No allergic reactions were recorded in our study in the two treatment arms. According to most previous studies, adverse drug reactions were rare and similar between the NPH and detemir groups [16, 21, 22, 24]. Notably, in the study by Mathiesen et al., only eight women of the 152 of the detemir group reported adverse events relating to injection sites [8]. In addition, in the study by Herrera et al., a higher rate of allergic reactions in women treated with insulin detemir was noted, forcing them to switch to an alternate medication [15].

We found no differences between the two groups with respect to mean fasting and postprandial BG levels at the beginning of insulin treatment and at the end of gestation. These findings are similar to those of previous observational and RCTs comparing detemir and NPH in pregnancy [15, 26].

In our study, the detemir group presented a slightly lower HbA1c level at the end of gestation, confirmed by the ANCOVA test after adjusting for age, BMI, weight gain, and initial HbA1c. However, Hba1c is not considered as the gold standard marker for assessing glycemic control during pregnancy due to altered red blood cell kinetics, increased erythropoiesis, and hemodilution. Mean BG levels during pregnancy are more sensitive to changes in glucose control over the short term that these patients are treated for. In addition, the RCT of Ji et al. comparing NPH versus detemir in 240 pregnant women with pregestational and gestational diabetes showed that in the early period of treatment, detemir performed better than NPH in terms of controlling blood glucose, as fasting and postprandial BG levels after 1 week of treatment were lower in the detemir group and the time to reach target was shorter. Furthermore, insulin detemir was at least as effective as NPH in a long-termtreatment as after 3 months, HbA1c between the two groups being shown to be similar [16].

In agreement with the study by Mathiesen et al., we found no differences between the two groups in relation to the duration of insulin treatment or the daily insulin dose [8]. Only in the study by Ji et al. was the total insulin dose in the detemir group higher than in the NPH group [16]. By contrast, in our study insulin doses between the two groups were comparable after adjustment for weight.

Overall, there were no statistically significant differences in neonatal outcomes including birth weight, birth weight adjusted for gestational age, percentage of macrosomia, or SGA and rates of preterm delivery or cesarean section, although the relevant data were unavailable. There were no congenital malformations or cases of severe neonatal hypoglycemia, which suggested that insulin detemir was as good as NPH in terms of safety. We acknowledge the lack of data of our study regarding perinatal parameters, such as induction rates and complications in labor. Although, indeed, one limitation might be that some data are missing, our results are consistent with the findings reported by most previous studies [9, 15,16,17, 21, 24]. Only the recent RCT by Bartal et al. reported lower rates of adverse neonatal outcomes, but with a higher rate of LGA in the detemir group [21].

To the best of our knowledge, this is the largest retrospective study investigating exclusively insulin detemir versus NPH, either of these administered in a single dose at night, in pregnancies with GDM only. In addition, one notable strength of the study is that it was performed in real-life routine conditions among unselected patients in a single center by the same medical team, where the glycemic goals for treatment were constant throughout the study period. The majority of the patients were of Caucasian origin and insulin therapy was initiated taking into consideration the increase of the fetal abdomen circumference on ultrasound.

Limitations of the study are the retrospective nature of the design and the selection of patients, the missing data regarding neonatal outcomes, and the selection of the insulin type. However, the current study provides real-world data, which are very important and are complementary to information derived from RCTs. Moreover, most of the women were of the same ethnicity with a mean prepregnancy BMI under 30 kg/m2. Outcomes from this study should be applied carefully in cases of more serious GDM where an intensive insulin regimen is needed, and especially in pregnancies with T2DM.

In conclusion, this study provides evidence that insulin detemir and NPH are equally effective and safe with respect to glycemic control and total insulin dose needed, maternal outcomes and complications, and neonatal outcomes. The only significant difference, though minor, was observed in relation to final HbA1c. The lower HbA1c in the detemir group seems not to have clinical significance, as evidenced by the absence of differences with regard to maternal adverse outcomes or neonatal complications, between the two GDM groups. Finally, the only differentiating factor appears to be the cost of the two options, where NPH is the more economical one.