Application of glycemic qualification rate based on fingerstick glucose monitoring in women with gestational diabetes mellitus

Abstract

Objective

To explore the appropriate application of glycemic qualification rate (GQR) calculated by fingerstick blood glucose (BG) monitoring for patients with gestational diabetes mellitus (GDM) by analyzing the relationship between BG control and adverse pregnancy outcomes.

Methods

Fingerstick Blood Glucose data during the second and third trimester of singleton pregnant women diagnosed with GDM were collected. GQR which is defined as the percentage of fingerstick BG values reaching the targets of BG control in a period of time was calculated. Patients were divided into three groups according to tertiles (tertile 1, GQR <56.25%; tertile 2, GQR 56.25–75%; and tertile 3, GQR ≥75%). Pregnant outcomes were compared among the three groups. Univariate analysis and logistic regression were performed to analyze the potential relationship between GQR and pregnancy outcomes. Receiver operating characteristic (ROC) curves were calculated to determine the cutoff values. We also explored that whether twice or three times monitoring per day would be adequate for GQR calculation, so we brought in two or three glucose measuring times per day to explore the relationship between new GQR and adverse outcomes.

Results

A total of 311 patients diagnosed with GDM were analyzed. In univariate analysis, the incidences of cesarean section of tertile 1–3 groups were 61.4%, 58.7%, and 44.9%, respectively (p < .05). The incidences of neonatal hypoglycemia of tertiles 1–3 groups were 19.8%, 18.6%, and 8.7% (p < .05). The difference of composite outcomes was statistically significant (p = .001). After adjustment, the patients with worse BG control (lower GQR) had higher risk of cesarean section (tertile 1 – aOR = 2.029, 1.128–3.648), neonatal hypoglycemia (tertile 1: aOR = 2.498, 1.082–5.766) as well as composite outcomes. The ROC curve of GQR indicated the predictive value for neonatal hypoglycemia (area under the ROC curve (AUC) 0.612 (0.532–0.692)) and neonatal composite outcomes (AUC 0.593 (0.528–0.657)) with optimal cutoff values of 81.1% and 73.5%, respectively. We also explored that whether twice or three times monitoring per day would be adequate for GQR calculation. The result showed that GQR only calculated by FBG + 2hPG after lunch (2h AL) per day also had well relationship with cesarean section (tertile 1: OR = 2.412, 1.322–4.398), neonatal hypoglycemia (tertile 1: aOR = 4.497, 1.607–12.586), and neonatal composite outcomes (tertile 1: aOR = 1.959, 95% confidence interval (CI): 1.114–3.444, p = .020).

Conclusions

The GQR calculated by the easily applicable fingerstick BG is related to occurrence of cesarean section and neonatal hypoglycemia in GDM women. GQR ≥ 80% is recommended for better pregnancy outcomes. As for the number of points monitoring per day, GQR calculated by FBG + 2h AL was an optimal option for better pregnancy outcomes if mothers needed to simplify the process of monitoring.

Keywords:

• Gestational diabetes mellitus
• fingerstick blood glucose monitoring
• neonatal hypoglycemia
• cesarean section
• pregnancy outcome

Background

Gestational diabetes mellitus (GDM) is an abnormal glucose metabolism status diagnosed in the second or third trimester of pregnancy without overt diabetes (ODM) prior to gestation [1], which accounts for 83.4% of gestational hyperglycemia [2]. The average prevalence of GDM in China is 17.5% [3]. As one of the most common complications in pregnancy, GDM has numbers of short- and long-term adverse effects on both mother and child [4]. The most common perinatal adverse outcomes of GDM include pre-eclampsia, cesarean section, neonatal hypoglycemia, macrosomia, fetal growth restriction, and neonatal respiratory complications [5,6].

It has widely demonstrated that maintaining euglycemia is crucial for the better outcomes of GDM pregnancy. Therefore, the monitoring of blood glucose (BG) of women with GDM is particularly important [4,5]. Currently, capillary BG monitoring and glycated hemoglobin (HbA1c) are recognized as the most useful methods [7]. However, the HAPO study [6] has reported that an OGTT or a single BG point at 28 weeks rather than HbA1c is more relevant to pregnancy outcomes. Therefore, at present, the best method to monitor BG during the second and third trimesters would be capillary BG monitoring rather than HbA1c. Recently, a new group of BG indicators derived from continuous glucose monitoring (CGM), such as time in range (TIR) and coefficient of variation, has applied in clinic [8]. Studies have revealed that lower TIR is associated with higher risks of adverse perinatal outcomes among type 1 and type 2 diabetes as well as GDM patients, including neonatal hypoglycemia, large for gestational age (LGA) infants [9–12].

As GDM is a mild hyperglycemia situation and more than 80% GDM women can reach a good glucose control through lifestyle interventions [13], fingerstick BG monitoring is more applicable in GDM management than CGM. However, a single point of glucose level is limited in reflecting the whole glucose control pattern and does not assess glucose fluctuations. Therefore, glucose qualification rate (GQR) based on fingerstick BG monitoring may play a more applicable role in the glucose management of women with GDM during the second and third trimester because of its convenience, economical applicability compared with CGM [14–16]. Glucose qualification rate is defined as the percentage of fingerstick BG values reaching the targets of BG control in a period of time [17]. Similar to TIR, GQR could also reflect the glycemic control during a period of time. The difference is that GQR does not include data on continuous BG monitoring. Although several studies have explored the use of GQR in GDM women [18–20], there were few analyzed the association of GQR with adverse pregnancy outcomes. No consensus is available for the standard application of GQR for GDM women.

In this study, we collected the fingerstick BG values in fasting and 2-h postprandial conditions within one week during the second and third trimester among GDM patients and retrospectively analyzed the relationship between the GQR and adverse pregnancy outcomes in a GDM cohort. It aimed to explore the appropriate target of GQR calculated by fingerstick BG monitoring for GDM patients and increase the convenience of timely adjustment of BG.

Methods

Patients

Data for the singleton pregnant women diagnosed with GDM according to 75-g 2-h OGTT at 24–28 gestational weeks and delivered in the First Affiliated Hospital of Sun Yat-sen University between January 2018 and October 2021 were collected. The patients were evenly divided into three groups, the lowest GQR group was tertile 1 group, GQR < 56.25%, the middle group was tertile 2 group, GQR 56.25–75%, and the highest GQR group was tertile 3 group, GQR ≥75%.

The inclusion criteria were as follows: (1) aged between 18 and 45 years; (2) had a singleton pregnancy; (3) had available data for fingerstick BG of fasting and 2 h postprandial BG that no less than four measures per day during the second and third trimester of pregnancy; (4) had complete basic data of gestation and delivery; (5) diagnosed with GDM according to a 75-g OGTT during 24–28 weeks of pregnancy; and (6) received appropriate intervention or medical treatment after diagnosis. The exclusion criteria were as follows: (1) a diagnosis of pregestational diabetes mellitus or ODM; (2) had endocrine disease, such as hyperthyroidism or hypothyroidism; (3) complicated with severe heart disease, chronic hypertension, active liver disease, or active kidney disease; (4) use of medicine that affects BG while monitoring; and (5) psychological diseases that are likely to interfere with the data collection of the study.

Data collection

The collected baseline characteristics were included as follows: demographic characteristics, gestational weeks, parity, pre-pregnancy weight, pre-pregnancy body mass index (BMI), systolic blood pressure and diastolic blood pressure, OGTT glucose levels as well as glycated hemoglobin (HbA1c) levels at the OGTT test, usage of insulin during pregnancy, family history of diabetes, smoking habits of spouse and history of GDM. Data for fingerstick BG of fasting and 2 h postprandial BG that no less than four measures per day within seven days [21] during the second and third trimester of pregnancy (median gestational week of 32.9 weeks) were collected for the purpose to obtain sufficient information for GQR calculation. The timeframe was chosen according to referent TIR studies [21] that mostly conducted over a period of 3–7 days. Seven days of BG data gave us sufficient information to calculate the GQR and is more practical in clinic as the management of GDM was usually adjusted weekly based on glucose data.

The following maternal outcomes were included: (1) hypertensive disorders, such as pre-eclampsia and gestational hypertension; (2) cesarean section; and (3) polyhydramnios. The following neonatal outcomes were included: (1) neonatal hypoglycemia; (2) abnormal birth weight, such as macrosomia, LGA, small for gestational age (SGA), and low birth weight; (3) preterm birth or fetal growth restriction; (4) fetal distress or abnormal Apgar scores; and (5) transfer treatment to neonate department.

We also explored whether twice or three times monitoring per day would be adequate for GQR calculation, so we brought in two or three glucose measuring times per day to explore the relationship between new GQR and adverse outcomes.

Definitions

The diagnosis of GDM [22] was based on the IADPSG guidelines with one or more of the following values from a 75-g OGTT: FPG levels ≥92 mg/dL (5.1 mmol/L), 1 h PG levels ≥180 mg/dL (10.0 mmol/L) and 2 h PG levels ≥153 mg/dL (8.5 mmol/L). If FPG ≥126 mg/dL (7.0 mmol/L) or random plasma glucose ≥200 mg/dL (11.1 mmol/L) (with confirmation of FPG or A1C using a Diabetes Control and Complications Trial (DCCT)-/United Kingdom Prospective Diabetes Study (UKPDS)-standardized assay), it was defined as ODM in pregnancy. The definition of neonatal hypoglycemia was BG < 40 mg/dL (2.2 mmol/L).

Blood glucose control target is based on the recommendations proposed by the American Diabetes Association (ADA) in 2022 [7,23]: fasting BG <95 mg/dL (5.3 mmol/L), and 2-h postprandial BG <120 mg/dL (6.7 mmol/L).

The GQR was calculated by capillary BG values as follows: $$\mathrm{GQR}=\frac{\text{the number of times the BG reached the target }}{\text{the total number of BG monitoring record }}\times \mathrm{ }100\mathrm{\%}.$$

The maternal composite outcome comprised of at least one of the following outcomes: pre-eclampsia, cesarean section, and polyhydramnios. The neonatal composite outcome comprised of at least one of the following outcomes: preterm birth, LGA infant, macrosomia, neonatal hypoglycemia, transfer treatment to neonate department, miscarriage and intrauterine death.

Statistical analysis

SPSS 26.0 (SPSS Inc., Chicago, IL) was used for data analysis. Continuous variables with normal distribution are presented as the $\overline{x}\hspace{0.17em}\pm \hspace{0.17em}s,$ and nonnormally distributed data are expressed as median values with interquartile ranges (IQRs; 25th–75th quartiles). Categorical data are presented as percentages (%). In univariate analysis, we conducted analysis of variance (ANOVA) for continuous variables with normal distribution, and non-parametric tests were used if the variables were not normally distributed. For categorical variables, Pearson’s chi-square test was used, and continuity correction was used when expected counts were <5. In multivariate analysis, the correlations among metrics were evaluated by binary logistic regression analysis, and we used it to estimate the odds of occurrence of adverse outcomes with 95% confidence intervals (CIs) by each group. p < .05 was considered a statistically significant difference.

Results

Baseline characteristics

After the screening by the inclusion and exclusion criteria (Supplementary Figure 1), data of 311 GDM pregnant women were collected. The 5th percentile is BG of our data was 4.3 mmol/L (77 mg/dL), the 95th percentile BG of our data was 8.4 mmol/L (151 mg/dL). The lowest glucose value was 2.5 mmol/L (45 mg/dL), the highest glucose value was 15.5 mmol/L (279 mg/dL). The average GQR was 63.85%, and the median GQR was 66.67%.

The participants were categorized into three groups according to tertiles (tertile 1, GQR <56.25%; tertile 2, GQR 56.25–75%; and tertile 3, GQR ≥75%) in order to analyze the relationships between the outcomes and the levels of GQR.

Among the baseline characteristics, except for the proportion of primiparous women and FPG during OGTT, there were no significant differences in age, GDM history, family history of diabetes, pre-pregnancy BMI, spouse smoking habits, OGTT 1 h, OGTT 2 h, systolic blood pressure, and diastolic blood pressure among the three groups. Patients with lower GQR had a lower proportion of primipara, a higher fasting BG during OGTT and a higher percentage of insulin use (p < .05) (Table 1).

Table 1. Basic characteristics of patients with GDM.

	Total (n = 311)	Tertile 1 (GQR < 56.25%, n = 101)	Tertile 2 (56.25 ≤ GQR < 75%, n = 92)	Tertile 3 (GQR ≥ 75%, n = 118)	p
Age (years), median (IQR)	34.0 (31.0–37.0)	34.0 (32.0–37.0)	34.5 (30.0–37.0)	33.0 (30.0–38.0)	.233
Age ≥ 35, n (%)	143/311 (45.8%)	50/101 (49.5%)	46/92 (50.0%)	47/118 (39.8%)	.250
Gestational age at data collection (weeks), median (IQR)	32.86 (30–35.14)	32.86 (30–34.14)	31.93 (29.39–34.29)	33.86 (30.46–36.29)	.077
Primiparous, n (%)	147/311 (47.3%)	37/101 (36.6%)	39/92 (42.4%)	72/118 (60.2%)	.001*
FPG (mmol/L), mean (SD)	4.77 ± 0.62	4.99 ± 0.73	4.76 ± 0.52	4.56 ± 0.50	.000*
OGTT 1 h (mmol/L), mean (SD)	9.95 ± 1.37	9.84 ± 1.52	9.92 ± 1.40	10.09 ± 1.19	.366
OGTT 2 h (mmol/L), mean (SD)	8.82 ± 1.19	8.68 ± 1.26	8.89 ± 1.16	8.86 ± 1.16	.522
HbA1c (%), median (IQR)	5.1 (4.9–5.3)	5.2 (5.0–5.4)	5.1 (4.9–5.4)	5.1 (4.9–5.3)	.207
Pre-pregnancy BMI (kg/m^2), median (IQR)	22.34 (20.04–24.60)	22.49 (20.28–24.65)	22.44 (20.61–24.59)	22.21 (19.66–24.59)	.670
Pre-pregnancy BMI category, n (%)					.580
˂18.5	32/311 (10.3%)	6/101 (5.9%)	10/92 (10.9%)	16 (13.6%)
18.5–24	179/311 (57.6%)	63 (61.8%)	52 (56.5%)	59 (54.2%)
24–28	82/311 (26.4%)	28 (27.5%)	25 (27.2%)	29 (24.6%)
≥28	19/311 (6.1%)	5 (4.9%)	5 (5.4%)	9 (7.6%)
Smoking habits of spouse, n (%)	35/311 (11.3%)	10/101 (9.8%)	12/92 (13.0%)	13/118 (11.0%)	.772
History of GDM	34/311 (10.9%)	13/101 (12.7%)	14/92 (15.2%)	7/118 (5.9%)	.078
Family history of diabetes, n (%)	59/311 (19.0%)	18/101 (17.6%)	18/92 (19.6%)	23/118 (19.5%)	.924
Systolic blood pressure (mmHg), mean (SD)	113.3 ± 11.34	114.7 ± 12.0	112.3 ± 10.6	113.0 ± 11.3	.302
Diastolic blood pressure (mmHg), mean (SD)	70.7 ± 8.28	70.3 ± 8.6	70.4 ± 7.6	71.2 ± 8.6	.671
Insulin use, n (%)	31/311 (10.0%)	21/101 (20.8%)	7/92 (7.6%)	3/115 (2.5%)	.000*

^* p < .05.

Maternal and neonatal outcomes

In total, 311 women delivered 297 living infants, and 14 women experienced miscarriage or intrauterine fetal death. Through univariate analysis, the incidences of cesarean section in the tertile 3 group, with improved glycemic control, were significantly lower than those in the other two groups (p < .05) (Table 2). The incidences for the tertile 1, tertile 2, and tertile 3 groups were 61.4%, 58.7%, and 44.9% for cesarean section, respectively. We observed no among-group difference in pre-eclampsia, polyhydramnios, and gestational hypertension among the three groups (p > .05).

Table 2. Maternal and neonatal outcomes of patients.

	Tertile 1 (GQR < 56.25%, n = 101)	Tertile 2 (56.25 ≤ GQR < 75%, n = 92)	Tertile 3 (GQR ≥ 75%, n = 118)	p
Maternal outcomes
Cesarean section, n (%)	62/101 (61.4%)	54/92 (58.7%)	53/118 (44.9%)	.031*
Polyhydramnios, n (%)	5/101 (5.0%)	6/92 (6.5%)	2/118 (1.7%)	.221
Gestational hypertension, n (%)	7/101 (6.9%)	4/92 (4.3%)	8/118 (6.8%)	.702
Pre-eclampsia, n (%)	7/101 (6.9%)	4/92 (4.3%)	1/118 (0.8%)	.055
Maternal composite outcomes	66/101(65.3%)	56/92(60.9%)	55/118 (46.6%)	.013*
Neonatal outcomes
Miscarriage and intrauterine fetal death, n (%)	5/101 (5.0%)	6/92 (6.5%)	3/118 (2.5%)	.404
Gestational age at delivery (weeks), median (IQR)	38.43 (37.29–39.14)	38.43 (36.86–39.00)	38.64 (37.96–39.43)	.363
Preterm birth, n (%)	16/96 (16.7%)	18/86 (20.9%)	11/115 (9.6%)	.074
Birthweight (kg)	3.08 ± 0.55	3.01 ± 0.48	3.02 ± 0.42	.567
Low birth weight, n (%)	12/96 (12.5%)	11/86 (12.8%)	11/115 (9.6%)	.719
SGA, n (%)	8/96 (8.3%)	9/86 (10.5%)	21/115 (18.3%)	.074
LGA, n (%)	4/96 (4.2%)	3/86 (3.5%)	1/115 (0.9%)	.278
Macrosomia, n (%)	4/96 (4.2%)	2/86 (2.3%)	1/115 (0.9%)	.343
Neonatal hypoglycemia, n (%)	19/96 (19.8%)	16/86 (18.6%)	10/115 (8.7%)	.047*
Fetal distress, n (%)	18/96 (18.8%)	12/86 (14.0%)	23/115 (20.0%)	.520
Fetal growth restriction, n (%)	10/101 (9.9%)	5/92 (5.4%)	6/118 (5.1%)	.322
ARDS, n (%)	8/96 (8.2%)	6/86 (7.0%)	4/115 (3.5%)	.325
Abnormal Apgar 1 min (<8), n (%)	6/96 (6.3%)	5/86 (5.8%)	5/115 (4.3%)	.813
Apgar 1 min (<10), n (%)	21/96 (21.9%)	11/86 (12.8%)	13/115 (11.3%)	.079
Transfer treatment to neonate department, n (%)	34/96 (35.4%)	21/86 (24.4%)	28/115 (24.3%)	.140
Neonatal composite outcomes	54/101 (53.5%)	45/92 (48.9%)	39/118 (33.1%)	.006*

ARDS: acute respiratory distress syndrome.

The data are presented as the mean (SD) or median (IQR) as appropriate. The maternal composite outcome comprised of at least one of the following outcomes: pre-eclampsia, cesarean section, and polyhydramnios. The neonatal composite outcome comprised of at least one of the following outcomes: preterm birth, LGA infant, macrosomia, neonatal hypoglycemia, transfer treatment to neonate department, miscarriage and intrauterine death.

* p < .05.

Regarding neonatal outcomes, the incidence of neonatal hypoglycemia was statistically different (p < .05), those with improved glycemic control, tertile 3, had lower rates of neonatal hypoglycemia. The incidences for the tertile 1, tertile 2, and tertile 3 groups were 19.8%, 18.6%, and 8.7%, respectively. There were no differences among the incidences of preterm birth, neonatal birth weight, macrosomia, LGA infants, SGA infants, low birth weight infants, abnormal Apgar 1 min, fetal distress, fetal growth restriction, and transfer treatment to neonate department (p > .05).

For the maternal and neonatal composite outcomes, the difference between the patients in tertile 3 group with improved glycemic control and the remaining two groups was statistically significant (p < .05) (Table 2). Those with better BG control in tertile 3 group experienced better composite outcomes.

Correlation of GQR and adverse pregnancy outcomes

Taking the confounding factors into account, we performed binary logistic regression analysis on the potential influencing factors of adverse maternal and infant outcomes (Table 3). After adjusting for age, BMI before pregnancy and parity, it was found that mothers of the tertile 1 group experienced more cesarean sections (model 1: adjusted odds ratio (aOR) = 2.029, 95% CI: 1.128–3.648, p = .018) and more of their babies suffered from neonatal hypoglycemia (model 1: aOR = 2.498, 95% CI: 1.082–5.766, p = .034) compared to the reference tertile 3 group (Table 3). Moreover, a lower percentage of full scores of Apgar 1 min appeared in the tertile 1 group than the reference group (model 1: aOR = 2.657, 95% CI: 1.201–5.878, p = .016). As for maternal and neonatal composite outcomes, patients in tertile 1 group were worse than tertile 3 group (maternal – model 1: aOR = 2.145, 95% CI: 1.180–3.897, p = .012; neonatal – model 1: aOR = 2.389, 96% CI: 1.363–4.186, p = .002). Similarly, patients in the tertile 2 group had a higher percentage of neonatal hypoglycemia, cesarean section, and neonatal composite outcomes. After further adjusting for OGTT results, insulin use, smoking habits of spouse, family history of diabetes and the number of points recorded in model 2 and model 3, the results indicated that there were also significant differences in these outcomes among the groups. In general, it is known that GQR based on fingerstick BG monitoring may be an independent risk factor associated with these pregnancy outcomes mentioned above. After adjusting for confounders, no statistically significant associations of GQR with gestational hypertension, polyhydramnios, pre-eclampsia, birth weight, preterm birth, fetal distress, fetal growth restriction, ARDS, transfer treatment to neonate department were found (p > .05) (Table 3).

Table 3. Odds ratios from different levels of glycemic qualification rate for adverse pregnancy outcomes.

		Tertile 1 (GQR < 56.25%)		Tertile 2 (56.25 ≤ GQR < 75%)		Tertile 3 (GQR ≥ 75%)
		OR (95% CI)	p Value	OR (95% CI)	p Value	OR (reference)
Cesarean section	Crude	2.050 (1.195–3.518)	.009*	1.804 (1.039–3.131)	.036*	1.00
	Model 1	2.029 (1.128–3.648)	.018*	1.858 (1.024–3.372)	.041*	1.00
	Model 2	1.882 (0.999–3.544)	.049*	1.631 (0.897–2.966)	.055	1.00
	Model 3	2.141 (1.223–3.748)	.008*	1.850 (1.049–3.262)	.034*	1.00
Neonatal hypoglycemia	Crude	2.558 (1.127–5.806)	.025*	2.400 (1.030–5.593)	.043*	1.00
	Model 1	2.498 (1.082–5.766)	.032*	2.382 (1.012–5.604)	.047*	1.00
	Model 2	2.919 (1.249–6.826)	.013*	2.487 (1.058–5.848)	.037*	1.00
	Model 3	2.556 (1.104–5.920)	.029*	2.425 (1.028–5.720)	.043*	1.00
APGAR 1 min < 10	Crude	2.372 (1.100–5.116)	.028*	1.392 (0.592–3.270)	.448	1.00
	Model 1	2.657 (1.201–5.878)	.016*	1.500 (0.632–3.557)	.358	1.00
	Model 2	2.412 (1.072–5.424)	.033*	1.427 (0.601–3.386)	.420	1.00
	Model 3	2.286 (1.035–5.050)	.041*	1.396 (0.565–3.204)	.502	1.00
Maternal composite outcomes	Crude	2.160 (1.250–3.731)	.006*	1.782 (1.025–3.099)	.041*	1.00
	Model 1	2.145 (1.180–3.897)	.012*	1.836 (1.009–3.340)	.046*	1.00
	Model 2	1.819 (1.018–3.250)	.043*	1.710 (0.973–3.004)	.062	1.00
	Model 3	2.179 (1.238–3.835)	.007*	1.795 (1.019–3.163)	.043*	1.00
Neonatal composite outcomes	Crude	2.327 (1.346–4.025)	.003*	1.939 (1.107–3.397)	.021*	1.00
	Model 1	2.389 (1.363–4.186)	.002*	1.975 (1.120–3.481)	.019*	1.00
	Model 2	3.035 (1.676–5.496)	.000*	2.100 (1.187–3.714)	.011*	1.00
	Model 3	2.225 (1.271–3.897)	.005*	1.920 (1.088–3.390)	.024*	1.00

The OR was calculated by logistic regression analysis. Model 1 adjusted for age, BMI before pregnancy and parity (primipara or multipara). Model 2 adjusted for OGTT 0 min, 1 h and 2 h (normal or abnormal) and insulin use. Model 3 adjusted for history of GDM, family history of diabetes and smoking habits of spouse.

* p < .05.

To determine the predictive value of the metric for neonatal hypoglycemia, we used the receiver operating characteristic (ROC) curve of GQR compared to HbA1c (area under the curve (AUC) = 0.523 (0.429–0.616), p = .640) and OGTT fasting BG (AUC = 0.500 (0.398–0.603), p = .998). According to the curve (Figure 1), the GQR had a certain predictive value for neonatal hypoglycemia with an AUC value of 0.612 (0.532–0.692), achieving a sensitivity of 98%, a specificity of 28%, and an optimal cutoff value of 81.1%. Regarding the composite outcomes, GQR served better as a metric for predicting the neonatal composite outcomes with an AUC value of 0.593 (0.528–0.657), achieving an optimal cutoff value of 73.5%, a sensitivity of 71% and a specificity of 47%.

Figure 1. ROC curves. (a) ROC curve of GQR, HbA1c, and fasting BG during the OGTT for predicting the diagnosis of neonatal hypoglycemia. (b) ROC curve of GQR, HbA1c, and fasting BG during the OGTT for predicting the diagnosis of neonatal composite outcomes. AUC: area under the ROC curve.

Influence of twice or three times monitoring points on outcomes

Most patients in this study calculated GQR by fasting glucose, 2 h postprandial glucose (2hPG) after breakfast, after lunch, and after dinner. We wanted to explore whether the absence of BG values at a certain time point would affect the relationship between GQR and outcomes, and which time point had the strongest correlation with them. Therefore, we brought in two or three glucose measuring times per day including fasting + 2 h after breakfast (2h AB), fasting + 2 h after lunch (2h AL), fasting + 2 h after dinner (2h AD), fasting + 2 h after breakfast and lunch (2h AB + AL), fasting + 2 h after breakfast and dinner (2h AB + AD), fasting + 2 h after lunch and dinner (2h AL + AD) respectively to calculate GQR. In this model, the major confounders including age, parity, and pre-pregnancy BMI were adjusted. As the result, we found that GQR calculated by FBG + 2 h after lunch (2h AL) had relationship with cesarean section (tertile 1: aOR = 2.412, 95% CI: 1.322–4.398, p = .018), neonatal hypoglycemia (tertile 1: aOR = 4.497, 95% CI: 1.607–12.586, p = .004), and neonatal composite outcomes (tertile 1: aOR = 1.959, 95% CI: 1.114–3.444, p = .020) (Supplemental Table 1).

Discussion

With the economic improvement and the increase of elderly mothers, the incidence of GDM has remarkably increased in recent years in China [3]. Even it is a mild hyperglycemia situation, numbers of studies have demonstrated that GDM increases the risk of both neonatal and pregnancy complications as well as long-term maternal and offspring abnormal metabolism disorders [4].

Therefore, a strict glucose control based on proper glucose monitoring is essential in GDM management. Recently, many studies have focused on the qualified rate of BG and the frequency of BG monitoring. Compared to average BG, these indicators emphasize the stability of BG level and reflect the fluctuation of BG. However, there has been no consensus regarding the standard TIR for GDM patients due to the unnecessary or unavailability of CGM in the management of GDM during the second and the third trimesters. In 2020, Tian et al. [17] used a new definition, “glycemic qualification rate” (GQR), to explore the level of BG control at different gestational weeks. GQR is defined as the percentage of fingerstick BG values reaching the targets of BG control in a period of time. Although studies on the qualified rate of BG have emerged, there are different opinions on the standard of poor/good glycemic control, and there is no consensus on the standard of GQR available for the clinical applications. Li et al. [18] defined the standard of poor BG control as the daily self-monitoring of BG below 60% on the correlation among BG control and glycated albumin in Chinese women with GDM. Ghesquière et al. [20] defined good (<20% not within target range), acceptable (20–40% not within target range), or poor control (>40% not within target range) when assessing the effect of isolation on glycemia in GDM patients. They used their standards according to clinical experience but few data could prove the reliability.

Our data for the first time has demonstrated that GQR based on fingerstick BG was related to adverse pregnancy outcomes. Compared to GDM patients with GQR ≥75% during second and third trimester, GDM patients with GQR <75% have significantly more cesarean sections, higher incidence of hypoglycemia. Further ROC curve analysis showed that GQR has a certain predictive value for neonatal hypoglycemia and neonatal composite outcomes, of which the optimal cutoff values were 81.1% and 73.5%, respectively. Therefore, we suggested that a GQR equal to or greater than 80% based on fingerstick BG values may be an optimal cutoff value to reduce the occurrence of neonatal hypoglycemia and improve pregnancy outcomes.

Meanwhile, in consideration of the more economy, convenience, and less painful fingersticks, we tried to explore whether the difference in the daily measurement frequency can affect the relationship between GQR and pregnancy outcomes. We found that FBG + 2hPG after lunch (2h AL) were enough to reflect the correction with cesarean section, neonatal hypoglycemia and composite outcomes. Therefore, fasting glucose and 2h AL can provide the most valuable glucose control message among GDM women. Patients can skip some points of monitoring except the fasting and the 2h AL points.

In practical application, GQR is suitable for the BG management of GDM patients in the second and third trimester of pregnancy. In the study, we included women who underwent fingerstick glucose monitoring after the diagnosis of GDM, the vast majority of whom were in the second or third trimester of pregnancy, with a median gestational week of 32.9 weeks at data collection. After calculating GQR with 4 points per day for a week, we could adjust the following BG management plan according to whether it reaches the standard, and may give some advice on whether or not to use medicine in the future. For daily use, further studies are needed to confirm whether this standard is appropriate for pregnant women who has less BG data during pregnancy and what the best timing to use medicine. Our study showed preliminary evidence that fasting + 2h AL glucose was more noteworthy.

There were several limitations in the present study. First, it was a retrospective study with a relatively small sample size. Some patients were excluded due to the difficulty in saving fingerstick BG data of pregnant women after delivery. To some extent, it affected the general applicability and persuasiveness. The relationship between GQR and pregnancy outcomes in patients lack of BG data during pregnancy was not explained. However, in order to ensure the scientific rationality and good comparability among patients of our research with a relatively small number of patients, we used the present inclusion and exclusion criteria, excluding patients in situations that affect BG levels, such as thyroid diseases, hypertension, and glucocorticoids use. In addition, we compared the baseline characteristics among 311 patients in our study with GDM patients in the hospital those were excluded, including age, percentage of primipara, OGTT, pre-pregnancy BMI, history of GDM, etc., and we found no significant differences. Second, the percentage of cesarean section seemed to be higher compared with other countries [24]. The reason was reported that Chinese parents tended to choose cesarean section because they thought it to be the safest delivery option. Hospitals in urban areas have better equipment and more qualified staff than in more remote rural areas, which makes c-sections more accessible [24,25]. In addition, the percentages of miscarriage/IUFD in tertiles 1–3 groups were 5.0%, 6.5%, and 2.5%, respectively. The percentages of IUFD in tertiles 1–3 groups were 2.9%, 2.2%, and 0.9%. It seemed to decrease as the GQR improved, but there was no statistical difference of miscarriage/IUFD rate among groups (p > .05). For the reliability of the data, we referred to other large sample size studies [26,27] in China and found that the miscarriage and IUFD rate in these studies is not significantly different from ours. However, we noted that some of these studies, like ours, were retrospective, there might be some bias. The reason for the relatively higher proportion of mothers with miscarriage or IUFD outcomes compared to prospective studies may be that our inclusion criteria required good compliance and willingness to seek medical treatment, which made us have more chance to get in touch with them. At last, the GQR calculated from the fingerstick BG value did not cover the entire glucose pattern during pregnancy. Therefore, for a better use of GQR based on fingerstick BG among GDM women, a well-organized prospective study with a large sample size and increased frequency of fingerstick BG should be performed.

Conclusions

In conclusion, GQR calculated by fingerstick BG monitoring in GDM is practical and effective. The findings of the present study for the first time indicated that higher GQR is related to lower occurrence of cesarean section and neonatal hypoglycemia. Our study suggested that GQR in one week equal to or greater than 80% is the cutoff point for better maternal and neonatal outcomes. As for the number of points monitoring per day, GQR calculated by FBG + 2h AL were also an optimal option to reflect the relation with cesarean section, neonatal hypoglycemia, and composite outcomes.

Ethical approval

All methods were carried out in accordance with the relevant institutional guidelines and regulations. The study was approved by the ethics committees for clinical research and animal trials of the First Affiliated Hospital, Sun Yat-sen University (ethical approval numbers: [2020] 048). The study conforms to Declaration of Helsinki.

Author contributions

R.Z. performed the analysis and wrote the manuscript. X.C. and Y.L. conceived of the idea and designed the study. R.Z., N.C., and L.P. designed the research. C.X., C.W., H.L., and W.D. contributed to data collating, suggestion on statistical analysis, and revision of the manuscript. All authors read and approved the final manuscript. X.C. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Consent form

Informed consent for study participation was obtained from all participants (if subjects are under 16, from a parent and/or legal guardian).

Acknowledgements

The authors expressed appreciation to Jiabin Xie, for helpful advice for data analysis during the development of the study. Xie received no financial support for his participation.

Disclosure statement

The authors report there are no competing interests to declare.

Data availability statement

All data generated and analyzed in this study are included in this published article. The datasets are available from the corresponding author on reasonable request.

Additional information

Funding

The study was supported by grants from the 5010 Project Foundation of Sun Yat-sen University (No. 2017001) and The Science and Technology Foundation of Guangzhou City (No. 201803010101).