Abstract
Background
Although the influence of temperature on gestational diabetes mellitus (GDM) has been extensively researched, it remains debatable whether this influence stems from a temporary alteration in blood glucose levels due to temperature adaptation or represents a genuine effect on the intrauterine environment. This study aimed to explore the impact of ambient temperature on the diagnosis accuracy of GDM by analyzing pregnancy outcomes among GDM pregnancies diagnosed at different temperatures.
Methods
This cohort study included 65,908 singleton pregnant women who delivered at the Guangdong Women and Children Hospital between January 2015 and June 2022. Participants were categorized based on the 24-hour and morning (8:00–11:00 AM) average ambient temperatures recorded on their oral glucose tolerance test (OGTT) day, using the 5th and 95th percentiles as cut-offs, defining low, moderate, and high temperature groups, respectively. We compared the prevalence of GDM and the subsequent risk of macrosomia or large-for-gestational-age (LGA) infants across these groups. Risk ratios (RRs) for macrosomia and LGA infants in GDM pregnancies diagnosed at low or moderate temperatures, relative to high temperatures were quantified using modified Poisson regression.
Results
Among the 65,908 pregnant women, the prevalence of GDM varied by ambient temperature. When stratified by the 24-hour average temperature, the prevalence was highest in the high-temperature group (20.0%), followed by the moderate- (18.2%) and low-temperature (16.4%) groups. A similar gradient was observed for the morning average temperature, with prevalence rates of 20.2%, 18.3%, and 15.8% in the high-, moderate-, and low-temperature groups, respectively. Compared with GDM pregnancies diagnosed at high 24-hour average temperatures, those diagnosed at low temperatures were more likely to develop an LGA infant (10.2% vs. 6.3%; adjusted RR: 1.48, 95% CI: 1.01, 2.18) and macrosomia (4.6% vs. 2.2%; adjusted RR: 1.90, 95% CI: 1.01, 3.59). A consistent pattern was observed when using the morning average temperature for stratification.
Conclusions
While GDM prevalence rose with increasing ambient temperature on the OGTT day, the rates of subsequent LGA infants and macrosomia among GDM pregnancies fell. The divergent patterns between diagnosis rates and neonatal outcomes suggest that temperature may influence glycemic measurements and diagnostic classification.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12884-026-08968-8.
Keywords: Ambient temperature, Gestational Diabetes Mellitus (GDM), Oral glucose tolerance test (OGTT), Large-for-gestational-age (LGA), Macrosomia
Introduction
As a common complication during pregnancy, gestational diabetes mellitus (GDM) presents significant threats to the health of both pregnant women and their newborns. Women diagnosed with GDM are at an increased risk for suboptimal obstetric outcomes, developing type 2 diabetes and cardiovascular complications later in life [1–4]. Their offspring face higher risks of metabolic disorders and neurocognitive impairments [5–9]. Given the serious consequences of GDM and its rising prevalence, the diagnosis and effective screening of GDM have become critical public health priorities.
Currently, the screening methods and diagnostic thresholds for GDM vary globally [10, 11], largely due to cost-effectiveness considerations [12, 13]. Nonetheless, the current rationale for diagnosing GDM uniformly relies on identifying women at risk of adverse pregnancy outcomes, such as macrosomia or large-for-gestational-age (LGA) newborns [13, 14], largely based on findings from the Hyperglycemia and Adverse Pregnancy Outcomes (HAPO) study [15]. The gold standard for diagnosing GDM remains the oral glucose tolerance test (OGTT), regardless of whether a one-step or two-step method is used.
In recent years, the increasing health risks posed by climate change have brought about a growing focus on the impact of temperature on GDM. Recent studies suggest that higher temperatures correlate with an increased prevalence of GDM [16–18]. However, whether this correlation is a genuine reflection of altered glucose metabolism or merely a transient physiological response of the body, such as elevated peripheral blood glucose resulting from the redistribution of blood flow in high-temperature environments, remains debated [17, 19, 20]. Existing research has focused on examining alterations in glucose-regulating hormones and biomarkers at different temperatures. Since a primary goal of GDM diagnosis is to identify women at risk of adverse pregnancy outcomes, it is crucial to investigate how temperature influences GDM-related pregnancy outcomes. Specifically, the consistency of the effect of GDM diagnosed under different temperatures on pregnancy outcomes needs examination. If GDM diagnosed in high temperatures is associated with fewer adverse outcomes than that in low temperatures, it may imply that the hyperglycemia detected during heat exposure represents a transient physiological response to heat stress, rather than a persistent pathological state. This could further imply a potential for overdiagnosis of GDM during high-temperature periods.
This study aims to address the gap in the literature by investigating how GDM diagnoses made under different temperature conditions affect subsequent pregnancy outcomes. We hypothesize that high-temperature conditions may lead to the possibility of overdiagnosis of GDM, resulting in lower rates of LGA infants or macrosomia, while low-temperature conditions may contribute to the possibility of underdiagnosis and higher rates of LGA infants or macrosomia. By comparing pregnancy outcomes across varying ambient temperatures on the OGTT day, we seek to clarify the true impact of ambient temperature on GDM.
Methods
Study population
The study population comprised 86,194 pregnant women who attended antenatal care and planned delivery at the Guangdong Women and Children Hospital in Guangzhou City, Guangdong Province of China, from January 2015 to July 2022. We excluded participants with miscarriages or stillbirths (n = 4,542), multiple births (n = 2,964), gestational age at delivery less than 28 weeks (n = 100), pre-pregnancy diabetes mellitus (n = 362) and incomplete OGTT results or missing test dates (n = 12,318). The final study population comprised 65,908 pregnant women (Supplementary Figure S1). We compared the baseline characteristics between the included and excluded populations and revealed statistically significant differences. However, the magnitude of these differences was small (Supplementary Table S1). The detected statistical significance is probably a consequence of high analytical power afforded by the large sample size, which can detect even minor deviations as significant. The study was approved by the Ethics Committee of Guangdong Women and Children Hospital.
Exposure
Hourly temperature data for Guangzhou City from the official website of the China Meteorological Administration (http://www.weather.com.cn/). To assess the association between temperature exposure and GDM risk, we calculated the 24-hour average ambient temperature on the day of each participant’s OGTT. This exposure window was selected because acute temperature exposure may transiently influence glucose metabolism. Temperature exposures were categorized into low (< 12.3 °C), moderate (12.3–30.3 °C), and high (> 30.3 °C) groups using cutoffs based on the 5th and 95th percentiles of the 24-hour average temperature distribution. To more precisely align the exposure window with the timing of glucose measurement, we performed an additional analysis using the average ambient temperature during the morning of the OGTT (8:00–11:00 AM), with categorization based on 5th and 95th percentiles of this morning-specific average temperature distribution.
GDM and GDM-related pregnancy outcomes
GDM was diagnosed based on the criteria established by the International Association of Diabetes and Pregnancy Study Groups (IADPSG) in 2010 [21], which is currently adopted in China. GDM was confirmed after a one-step 75 g OGTT at 24–28 weeks of gestation, if the blood glucose levels met or exceeded any of the following thresholds: fasting glucose of 5.1 mmol/L, 1-hour postprandial glucose of 10.0 mmol/L, or 2-hour postprandial glucose of 8.5 mmol/L.
To assess fetal overgrowth related to intrauterine hyperglycemia, we analyzed LGA infants and macrosomia as distinct but complementary outcomes. This approach allows for a comprehensive evaluation from both a relative (gestational-age-adjusted) and an absolute perspective. LGA was defined as a birth weight exceeding the 90th percentile for the same gestational age and sex, based on a widely used Chinese population reference standard [22], macrosomia was defined as a birth weight ≥ 4,000 g. An infant could be classified as both an LGA infant and macrosomia, only one, or neither.
Information regarding the OGTT and delivery details was extracted from the hospital’s electronic medical records.
Covariates
The characteristics of participants, including maternal age, education, height, pre-pregnancy weight, weight before delivery, parity, infant sex, gestational age at delivery, and calendar year of delivery were extracted from electronic medical records. Maternal age was dichotomized at the median into < 30 and ≥ 30 years. Education levels were categorized as junior high school or below, high school, and college or above. Pre-pregnancy body mass index (BMI) was calculated as self-reported pre-pregnancy weight (kg) divided by height squared (m²), and categorized as underweight (< 18.5 kg/m²), normal weight (18.5–24.0 kg/m²) and overweight/obese (≥ 24.0 kg/m²) based on Chinese criteria [23]. Gestational weight gain (GWG) was calculated as the difference between the weight before delivery and the pre-pregnancy weight, and categorized as inadequate, adequate, and excessive gain according to the 2021 Chinese GWG recommendations, which account for pre-pregnancy BMI [24]. Parity was classified as primipara and multipara.
Statistical analysis
Continuous variables with normal distribution were expressed as mean ± standard deviations (SD) and compared using the t-test. Categorical variables were presented as numbers (percentages) and were compared using the Chi-square test.
Multivariable logistic regression models were used to estimate the adjusted prevalence of GDM, controlling for maternal age, education, and pre-pregnancy BMI [25]. Modified Poisson regression models with robust error variances were employed to estimate the relative risks (RRs) and 95% confidence intervals (CIs) for the associations of the ambient temperature on the OGTT day (categorized as low, moderate, and high) with adverse neonatal outcomes in the GDM population.
When the outcome was LGA infants, the model was adjusted for maternal age, education level, pre-pregnancy BMI, gestational weight gain, parity, infant sex and calendar year of delivery. For the analysis with macrosomia as the outcome, the model included all the aforementioned covariates and was additionally adjusted for gestational age at delivery, given that the definition of macrosomia is based on an absolute birth weight threshold independent of gestational week. These analyses were also repeated in the non-GDM population to evaluate whether the observed associations were directly driven by ambient temperature on the OGTT day itself or other potential confounding factors, rather than through its effect on the GDM diagnosis.
Stratified analyses were conducted to examine whether the effect was modified by age (< 30 or ≥ 30 years), pre-pregnancy BMI categories (< 18.5, 18.5–24.0, or ≥ 24.0 kg/m2), and gestational weight gain status (inadequate, adequate, and excessive). Multiplicative interaction was evaluated using product interaction terms in adjusted Poisson models via Wald tests. To assess the robustness of our findings, we conducted four sensitivity analyses by: (1) employing alternative temperature cut-offs at the 15th and 85th percentiles (15.8–28.8 °C), with the human thermal comfort threshold (18.0–23.0 °C) included for reference; (2) restricting the analysis to subgroups who underwent OGTT testing in spring (March to May) or autumn (September to November) to minimize potential confounding by seasonal variations (summer and winter were not included due to their narrower temperature range, which is less suitable for stratification); (3) excluding 129 individuals (1.07%) with severe GDM who required insulin therapy to reduce potential confounding effects on pregnancy outcomes; (4) restricting the analysis to the subgroup with available third-trimester hemoglobin A1c (HbA1c) data, with further adjustment for HbA1c levels in adjusted models, to minimize confounding by glycemic control subsequent to the mid-pregnancy diagnosis of GDM. Owing to sample size constraints, stratified analyses and the latter three sensitivity analyses used temperature categories based on the 15th/85th percentiles.
All analyses were performed using R software version 4.2.3 and SAS version 9.4, with P < 0.05 indicating a statistically significant difference.
Results
Descriptive characteristics of the study population
The baseline characteristics of the 65,908 participants are summarized in Table 1. Among them, 12,015 (18.2%) were diagnosed with GDM. Compared to the non-GDM group, pregnant women with GDM were older, had a higher proportion of multiparity, and a greater prevalence of overweight or obesity before pregnancy. However, they exhibited a lower rate of excessive weight gain, which may reflect dietary and lifestyle interventions initiated after GDM diagnosis under medical supervision.
Table 1.
Characteristics of the study participants stratified by GDM status
| Variables | Total | GDM | Non-GDM | P value |
|---|---|---|---|---|
| Numbers | 65908 | 12015 | 53893 | |
| Maternal age (n, %) | <0.001 | |||
| < 30 years | 35880 (54.4) | 4629 (38.5) | 31251 (58.0) | |
| ≥30 years | 30028 (45.6) | 7356 (61.5) | 22642 (42.0) | |
| Maternal education (n, %) | 0.217 | |||
| Junior high school or below | 8234 (12.5) | 1555 (12.9) | 6679 (12.4) | |
| High school | 13903 (21.1) | 2545 (21.2) | 11358 (21.1) | |
| College or above | 43771 (66.4) | 7915 (65.9) | 35856 (66.5) | |
| Pre-pregnancy BMI (n, %) | <0.001 | |||
| <18.5 kg/m2 | 12869 (19.5) | 1476 (12.3) | 11393 (21.1) | |
| 18.5–24.0 kg/m2 | 44229 (67.1) | 7972 (66.4) | 36257 (67.3) | |
| ≥24.0 kg/m2 | 8810 (13.4) | 2567 (21.4) | 6243 (11.6) | |
| Gestational weight gain (n, %) | <0.001 | |||
| Inadequate | 7003 (10.6) | 2510 (20.9) | 4493 (8.3) | |
| Adequate | 33657 (51.1) | 6430 (53.5) | 27227 (50.5) | |
| Excessive | 25248 (38.3) | 3075 (25.6) | 22173 (41.1) | |
| Parity (n, %) | <0.001 | |||
| Primiparous | 33046 (50.1) | 5290 (44.0) | 27756 (51.5) | |
| Multiparous | 32862 (49.9) | 6725 (56.0) | 26137 (48.5) | |
| Infant sex (n, %) | 0.036 | |||
| Female | 30731 (46.6) | 5498 (45.8) | 25233 (46.8) | |
| Male | 35177 (53.4) | 6517 (54.2) | 28660 (53.2) | |
| Gestational age at delivery (week) | 38.8 ± 1.5 | 38.5 ± 1.6 | 38.8 ± 1.5 | <0.001 |
| Gestational age group | <0.001 | |||
| Preterm (<37 weeks) | 3506 (5.3) | 885 (7.4) | 2621 (4.9) | |
| Early Term (37–38 weeks) | 19390 (29.4) | 3945 (32.8) | 15445 (28.7) | |
| Full Term (≥39 weeks) | 43012 (65.3) | 7185 (59.8) | 35827 (66.5) | |
| Calendar year of delivery (n, %) | <0.001 | |||
| 2017 | 12469 (18.9) | 10147 (18.8) | 2322 (19.3) | |
| 2018 | 12073 (18.3) | 9892 (18.4) | 2181 (18.2) | |
| 2019 | 12994 (19.7) | 10895 (20.2) | 2099 (17.5) | |
| 2020 | 11101 (16.8) | 9111 (16.9) | 1990 (16.6) | |
| 2021 | 11361 (17.2) | 9021 (16.7) | 2340 (19.5) | |
| 2022 (collected through July only) | 5910 (9.0) | 4827 (9.0) | 1083 (9.0) | |
| OGTT by season (n, %) | <0.001 | |||
| Spring | 17992 (27.3) | 3146 (26.2) | 14846 (27.5) | |
| Summer | 16687 (25.3) | 3191 (26.6) | 13496 (25.0) | |
| Autumn | 15770 (23.9) | 3054 (25.4) | 12716 (23.6) | |
| Winter | 15459 (23.5) | 2624 (21.8) | 12835 (23.8) |
Data are mean ± standard deviation or number (%)
Abbreviations: BMI body mass index
GDM prevalence stratified by ambient temperature on the OGTT day
Table 2 presents the prevalence of GDM and its adjusted values stratified by ambient temperatures on the OGGT day. A consistent upward trend in the prevalence of GDM with increasing temperature is observed. When stratified by the 24-hour average temperature, the crude prevalence of GDM was 16.4% in the low-temperature group (< 12.3 °C), 18.2% in the moderate-temperature group (12.3–30.3 °C), and 20.0% in the high-temperature group (> 30.3 °C). After adjustment, the corresponding prevalence was 15.1% (95% CI: 13.9, 16.4), 16.9% (95% CI: 16.6, 17.4), and 19.1% (95% CI: 17.7, 20.5), respectively (P < 0.001). Similarly, for the morning average temperature, the crude prevalence was 15.8% (low), 18.3% (moderate), and 20.2% (high). The corresponding adjusted rates were 14.6% (95% CI: 13.4, 15.9), 17.0% (95% CI: 16.7, 17.4), and 18.9% (95% CI: 17.5, 20.4) (P < 0.001).
Table 2.
GDM prevalence by ambient temperature on the OGTT day
| Temperature Group | n/N | Crude rate, % | Adjusted rate (95%CI), % | P value |
|---|---|---|---|---|
| 24-hour average ambient temperature | ||||
| Low (< 5th, 12.3 °C) | 538/3278 | 16.4 | 15.1 (13.9, 16.4) | < 0.001 |
| Moderate (5–95th, 12.3–30.3 °C) | 10,806/59,281 | 18.2 | 16.9 (16.6, 17.4) | |
| High (> 95th, 30.3 °C) | 671/3349 | 20.0 | 19.1 (17.7, 20.5) | |
| morning average ambient temperature (8:00 AM to 11:00 AM) | ||||
| Low (< 5th, 11.9 °C) | 505/3205 | 15.8 | 14.6 (13.4, 15.9) | < 0.001 |
| Moderate (5–95th, 11.9–30.8 °C) | 10,906/59,707 | 18.3 | 17.0 (16.7, 17.4) | |
| High (> 95th, 30.8 °C) | 604/2996 | 20.2 | 18.9 (17.5, 20.4) | |
n/N denotes the number of GDM cases over the total number of OGTTs performed in each temperature group. Adjusted prevalence rates were estimated using logistic regression model after adjusting for maternal age, education, and pre-pregnancy body mass index
Associations between ambient temperature on the OGTT day and neonatal outcomes
The rates of LGA infants (8.4% vs. 7.0%) and macrosomia (3.6% vs. 2.9%) were significantly higher in the GDM group than in the non-GDM group, with adjusted RRs of 1.21 (95% CI: 1.13, 1.30) and 1.46 (95% CI: 1.32, 1.63), respectively.
Contrary to the trend for GDM prevalence, the risks of LGA and macrosomia among GDM pregnancies increased as the ambient temperature on the OGTT day decreased (Table 3). Using the 5th and 95th percentiles of the 24-hour average temperature as cut-offs, the rates of LGA infants and macrosomia rose from 6.3% and 2.2% in the high-temperature group to 10.2% and 4.6% in the low-temperature group, respectively. After multivariable adjustment, the relative risks (RRs) in the low-temperature group were significantly elevated compared to the high-temperature group for both LGA infants (RR: 1.48, 95% CI: 1.01, 2.18) and macrosomia (RR: 1.90, 95% CI: 1.01, 3.59). The moderate-temperature group showed a non-significant increase in risk. A comparable inverse association was also observed when using the morning average temperature for stratification. In contrast, no significant associations between temperature and neonatal outcomes were observed in the non-GDM population.
Table 3.
Adjusted RRs (95% CIs) of LGA infants and Macrosomia according to ambient temperature on OGTT day in women with and without GDM
| Temperature Group | LGA infants | Macrosomia | ||||
|---|---|---|---|---|---|---|
| n (%) | Model 1 | Model 2 | n (%) | Model 1 | Model 2* | |
| 24-hour average ambient temperature | ||||||
| GDM | ||||||
| High (> 95th, 30.3 °C) | 42 (6.3) | Ref. | Ref. | 15 (2.2) | Ref. | Ref. |
| Moderate (5–95th, 12.3–30.3 °C) | 913 (8.4) | 1.35 (1.00, 1.82) | 1.26 (0.93, 1.71) | 398 (3.7) | 1.65 (0.99, 2.74) | 1.58 (0.95, 2.64) |
| Low (< 5th, 12.3 °C) | 55 (10.2) | 1.63 (1.11, 2.40) | 1.48 (1.01, 2.18) | 25 (4.6) | 2.08 (1.11, 3.90) | 1.90 (1.01, 3.59) |
| Non-GDM | ||||||
| High (> 95th, 30.3 °C) | 172 (6.4) | Ref. | Ref. | 66 (2.5) | Ref. | Ref. |
| Moderate (5–95th, 12.3–30.3 °C) | 3426 (7.1) | 1.10 (0.95, 1.28) | 1.09 (0.94, 1.26) | 1435 (3.0) | 1.20 (0.94, 1.53) | 1.23 (0.97, 1.56) |
| Low (< 5th, 12.3 °C) | 190 (6.9) | 1.08 (0.88, 1.34) | 1.07 (0.88, 1.32) | 83 (3.0) | 1.23 (0.89, 1.69) | 1.27 (0.92, 1.74) |
| morning average ambient temperature (8:00 AM to 11:00 AM) | ||||||
| GDM | ||||||
| High (> 95th, 30.8 °C) | 40 (6.6) | Ref. | Ref. | 13 (2.2) | Ref. | Ref. |
| Moderate (5–95th, 11.9–30.8 °C) | 915 (8.4) | 1.27 (0.93, 1.72) | 1.19 (0.96, 1.47) | 401 (3.7) | 1.71 (0.99, 2.95) | 1.63 (0.95, 2.77) |
| Low (< 5th, 11.9 °C) | 55 (10.9) | 1.64 (1.11, 2.43) | 1.34 (1.02, 1.75) | 24 (4.8) | 2.21 (1.14, 4.29) | 1.95 (1.01, 3.74) |
| Non-GDM | ||||||
| High (> 95th, 30.8 °C) | 156 (6.5) | Ref. | Ref. | 68 (2.8) | Ref. | Ref. |
| Moderate (5–95th, 11.9–30.8 °C) | 3445 (7.1) | 1.08 (0.93, 1.26) | 1.00 (0.92, 1.11) | 1438 (2.9) | 1.04 (0.82, 1.32) | 1.07 (0.84, 1.35) |
| Low (< 5th, 11.9 °C) | 187 (6.9) | 1.06 (0.86, 1.30) | 1.02 (0.89, 1.17) | 78 (2.9) | 1.02 (0.74, 1.40) | 1.07 (0.78, 1.47) |
Model 1: Unadjusted;
Model 2: Adjusted for maternal age, education level, pre-pregnancy body mass index, gestational weight gain, parity, infant sex and calendar year of delivery;
Model 2*: Additionally adjusted for gestational week at delivery in addition to the covariates in Model 2
Stratified and sensitivity analysis
Stratified analyses based on the 15th/85th percentile temperature categories are presented in Fig. 1. Although there were no statistically significant interactions between ambient temperature on the OGTT day and maternal age, pre-pregnancy BMI or gestational weight gain for either LGA infants or macrosomia, the inverse association between temperature and neonatal outcomes appeared more pronounced in subgroups with pre-pregnancy BMI ≥ 24.0 kg/m2.
Fig. 1.
Stratified analysis of the association between 24-hour ambient temperature on the OGTT day and risks of LGA infants (A) and macrosomia (B) in GDM pregnancies. LT-GDM: GDM diagnosed under low temperatures (< 15th, 15.8℃); HT-GDM: GDM diagnosed under high temperatures (> 85th, 28.8℃); MT-GDM: GDM diagnosed under moderate temperatures (15.8–28.8℃, 15th & 85th Percentiles of ambient temperature). RRs (95% CIs) were estimated using modified Poisson regression adjusting for maternal age, education level, pre-pregnancy BMI, gestational weight gain, parity, and infant sex, excluding the stratification variables
Sensitivity analyses confirmed the robustness of the primary findings. The trends in GDM prevalence remained consistent across alternative temperature cut-offs (15th/85th percentiles) and the human thermal comfort range (18.0–23.0 °C) (Supplementary Table S2). Similarly, the association between ambient temperature on the day of the OGTT and neonatal outcomes persisted when different temperature categorizations were implemented (Supplementary Table S3). These associations remained consistent when the analysis was restricted to seasons with wider temperature variations (spring and autumn) (Supplementary Table S4) and when excluding GDM cases that required insulin therapy (Supplementary Table S5). Furthermore, a restricted analysis in the subgroup with available third-trimester HbA1c data continued to show consistent associations after adjusting for HbA1c levels (Supplementary Table S8). Notably, this subgroup showed comparable baseline characteristics to those without HbA1c data (Supplementary Table S6) and exhibited no significant differences in HbA1c levels across temperature groups (Supplementary Table S7), which supports the robustness of the findings.
Discussion
In this study, we found that ambient temperature on the OGTT day was associated not only with the likelihood of a GDM diagnosis, but also with the risk of related adverse outcomes, specifically LGA infants and macrosomia. Intriguingly, while the frequency of GDM diagnosis was higher in warmer conditions, women diagnosed with GDM under these conditions exhibited a lower incidence of LGA infants and macrosomia compared to those diagnosed in colder conditions. This pattern suggests a potential diagnostic discrepancy, elevated temperatures might contribute to the possibility of overdiagnosis of GDM, whereas lower temperatures could be linked to the possibility of underdiagnosis, as reflected by the subsequent occurrence of typical GDM complications.
The observed positive association between higher temperatures and increased GDM prevalence aligns with previous studies [26–32]. The underlying mechanisms remain inconclusive. Competing mechanistic explanations have led to varying views on whether this association may be a potential diagnostic artifact. Some researchers propose that heat-induced changes in peripheral blood flow may transiently elevate blood glucose levels without altering long-term metabolic risk. For example, Akanji and Oputa [33] found that elevated plasma glucose levels at higher temperatures led to the misclassification of non-diabetic subjects as having impaired glucose tolerance, likely due to altered blood flow responses. Similarly, Moses et al. [20] suggested that blood flow redistribution, rather than direct metabolic effects, accounts for the nonlinear relationship between ambient temperature and glucose levels. In contrast, other research proposes that temperature-related changes in GDM prevalence reflect genuine metabolic alterations. Retnakaran et al. [30] reported that elevated temperatures shortly before the OGTT were independently associated with GDM and pancreatic β-cell dysfunction. Shen et al. [16] further reported an inverse relationship between temperature at birth and umbilical cord C-peptide and glucose levels, which was the first study to consider neonatal glucose metabolism. Nevertheless, these studies primarily addressed short-term glycemic responses and could not determine whether temperature exerts lasting effects on the intrauterine environment or fetal growth.
Adverse neonatal outcomes such as LGA infants and macrosomia are well-established consequences of maternal hyperglycemia, which promotes fetal hyperinsulinemia and accelerates growth [34, 35]. Our findings, however, indicate that among women diagnosed with GDM, the risks of these outcomes were actually lower when diagnosis occurred under warmer conditions. Moreover, in the non-GDM population, temperature variations did not significantly affect the incidence of LGA infants or macrosomia. These findings further suggest that the temperature on the OGTT day influences the risk of LGA infants and macrosomia primarily through its effect on the GDM diagnosis itself, rather than acting as a direct, independent risk factor for these outcomes. The persistence of this pattern in seasonal subgroup analyses (spring and autumn) further strengthens the argument by minimizing confounding from other seasonally variable factors such as variation in diet, physical activity, or exposure to sunlight which could affect vitamin D and other light-sensitive hormones. Stratified analyses suggested that the inverse relationship between temperature and neonatal outcomes among GDM women was more pronounced in those with pre-pregnancy overweight/obesity, although the formal test for interaction was not statistically significant. This phenomenon may arise from the increased volume of adipose tissue in individuals with obesity, which restricts heat dissipation and prompts compensatory enhancements in skin blood flow [36]. The subsequent redistribution of blood flow may diminish glucose uptake by visceral organs and skeletal muscle, consequently leading to elevated peripheral blood glucose levels.
To our knowledge, this is the first study to investigate whether temperature influences the potential for overdiagnosis or underdiagnosis of GDM from the perspective of GDM-related adverse outcomes. Overdiagnosis of GDM can lead to unnecessary antenatal monitoring and increased healthcare costs. Furthermore, excessive treatment following diagnosis may lead to adverse outcomes from medical interventions, such as earlier gestational age at birth and cesarean sections [37]. It may also induce psychological distress, which in pregnant women can exceed that associated with type 1 diabetes [38–40]. Conversely, underdiagnosis deprives at-risk women of beneficial interventions, increasing avoidable perinatal complications. Therefore, recognizing and addressing temperature-related diagnostic variability is essential for optimizing maternal care and resource allocation.
Several limitations should be considered. First, temperature exposure was assessed using regional monitoring data rather than individual-level measurements, which may not account for time spent indoors, air conditioning use, or personal behavioral adaptations. While this likely resulted in non-differential misclassification, it may have attenuated the observed associations. Second, our observational study design is its inability to conclusively distinguish between two potential mechanisms, a direct effect of ambient temperature at diagnosis versus an effect mediated through differences in post-diagnosis management. Although we adjusted for third-trimester HbA1c as a marker of average glycemia, we lacked detailed data on other critical aspects of clinical management (e.g., timing and intensity of dietary intervention, frequency of self-monitoring, specific pharmacological treatment protocols, and individual patient adherence) that influence fetal growth. While standardized institutional protocols were in place to minimize variation, unmeasured differences in management execution or response remain a plausible alternative explanation for the observed associations between temperature at diagnosis and neonatal outcomes. Third, Guangzhou’s subtropical climate limits exposure to extreme cold, which may restrict the generalizability of findings to colder regions. Fourth, neonatal outcomes were limited to birth weight metrics; incorporating additional biomarkers such as cord-blood C-peptide or clinical outcomes like NICU admission could provide deeper insight into true diabetic fetopathy.
Despite these limitations, our findings highlight the need for clinical awareness of temperature as a potential source of diagnostic variability. While revising diagnostic criteria based solely on temperature is premature, practical steps, such as scheduling OGTTs in thermally stable settings or interpreting results with ambient conditions in mind could help minimize bias. Future prospective studies incorporating individual-level exposure assessment and longer-term metabolic follow-up are warranted to validate these observations and inform evidence-based adjustments.
Conclusion
This study suggests that ambient temperature on the OGTT day may be associated with variations in GDM diagnosis and subsequent risks of LGA infants and macrosomia. The divergent patterns between diagnosis rates and neonatal outcomes suggest that temperature may influence glycemic measurements and diagnostic classification. Clinicians should be aware of ambient conditions as a potential, though not definitive, technical factor when interpreting OGTT results. Given the study’s inherent limitations and the need to distinguish between diagnostic and post-diagnostic effects, further prospective research is required before any formal recommendations regarding temperature adjustments in clinical practice can be made.
Supplementary Information
Acknowledgements
We gratefully acknowledge the valuable support provided by the staff of the Guangdong Women and Children Hospital.
Authors’ contributions
Q.F. researched data, wrote the initial draft of the manuscript, and critically reviewed and revised the manuscript. J.P., L.W., L. Zhang, and L. Zheng contributed to data acquisition and reviewed and edited the manuscript. L.S., Q.L., and G.F. contributed to discussion and review for significant intellectual content. Y.W. and G.L. conceptualized and designed the study, and reviewed and approved the final manuscript. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
Funding
This study was supported by the National Natural Science Foundation of China (82073660).
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
This study was conducted in accordance with the Declaration of Helsinki, and ethical approval was obtained from the ethics committee of Guangdong Women and Children Hospital (approval number: 202201203). In this study, due to its retrospective nature, informed consent was waived with the approval of the ethics committee of Guangdong Women and Children Hospital.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Guocheng Liu and Youjie Wang contributed equally to this work.
References
- 1.Ye W, Luo C, Huang J, Li C, Liu Z, Liu F. Gestational diabetes mellitus and adverse pregnancy outcomes: systematic review and meta-analysis. BMJ. 2022;377:e067946. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Greco E, Calanducci M, Nicolaides KH, Barry EVH, Huda MSB, Iliodromiti S. Gestational diabetes mellitus and adverse maternal and perinatal outcomes in twin and singleton pregnancies: a systematic review and meta-analysis. Am J Obstet Gynecol. 2024;230(2):213–25. [DOI] [PubMed] [Google Scholar]
- 3.Kramer CK, Campbell S, Retnakaran R. Gestational diabetes and the risk of cardiovascular disease in women: a systematic review and meta-analysis. Diabetologia. 2019;62(6):905–14. [DOI] [PubMed] [Google Scholar]
- 4.Vounzoulaki E, Khunti K, Abner SC, Tan BK, Davies MJ, Gillies CL. Progression to type 2 diabetes in women with a known history of gestational diabetes: systematic review and meta-analysis. BMJ. 2020;369:m1361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lowe WL Jr., Scholtens DM, Kuang A, Linder B, Lawrence JM, Lebenthal Y, McCance D, Hamilton J, Nodzenski M, Talbot O, et al. Hyperglycemia and Adverse Pregnancy Outcome Follow-up Study (HAPO FUS): Maternal Gestational Diabetes Mellitus and Childhood Glucose Metabolism. Diabetes Care. 2019;42(3):372–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Lowe WL Jr., Lowe LP, Kuang A, Catalano PM, Nodzenski M, Talbot O, Tam WH, Sacks DA, McCance D, Linder B, et al. Maternal glucose levels during pregnancy and childhood adiposity in the Hyperglycemia and Adverse Pregnancy Outcome Follow-up Study. Diabetologia. 2019;62(4):598–610. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Bianco ME, Josefson JL. Hyperglycemia During Pregnancy and Long-Term Offspring Outcomes. Curr Diab Rep. 2019;19(12):143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Chen S, Persson M, Wang R, Dalman C, Lee BK, Karlsson H, Gardner RM. Random capillary glucose levels throughout pregnancy, obstetric and neonatal outcomes, and long-term neurodevelopmental conditions in children: a group-based trajectory analysis. BMC Med. 2023;21(1):260. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Xiang AH. Association of Maternal Diabetes With Autism in Offspring. JAMA. 2017;317(5):537–8. [DOI] [PubMed] [Google Scholar]
- 10.Farrar D, Duley L, Dowswell T, Lawlor DA. Different strategies for diagnosing gestational diabetes to improve maternal and infant health. Cochrane Database Syst Rev. 2017;8(8):Cd007122. [DOI] [PMC free article] [PubMed]
- 11.Tieu J, McPhee AJ, Crowther CA, Middleton P, Shepherd E. Screening for gestational diabetes mellitus based on different risk profiles and settings for improving maternal and infant health. Cochrane Database Syst Rev. 2017;8(8):Cd007222. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.ACOG Practice Bulletin No. 190: Gestational Diabetes Mellitus. Obstet Gynecol. 2018;131(2):e49–64. [DOI] [PubMed] [Google Scholar]
- 13.Sweeting A, Wong J, Murphy HR, Ross GP. A Clinical Update on Gestational Diabetes Mellitus. Endocr Rev. 2022;43(5):763–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Vandorsten JP, Dodson WC, Espeland MA, Grobman WA, Guise JM, Mercer BM, Minkoff HL, Poindexter B, Prosser LA, Sawaya GF, et al. NIH consensus development conference: diagnosing gestational diabetes mellitus. NIH Consens State Sci Statements. 2013;29(1):1–31. [PubMed] [Google Scholar]
- 15.Hadar E, Hod M. Establishing consensus criteria for the diagnosis of diabetes in pregnancy following the HAPO study. Ann N Y Acad Sci. 2010;1205:88–93. [DOI] [PubMed] [Google Scholar]
- 16.Shen EX, Moses RG, Oats JJN, Lowe J, McIntyre HD. Seasonality, temperature and pregnancy oral glucose tolerance test results in Australia. BMC Pregnancy Childbirth. 2019;19(1):263. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Molina-Vega M, Gutiérrez-Repiso C, Muñoz-Garach A, Lima-Rubio F, Morcillo S, Tinahones FJ, Picón-César MJ. Relationship between environmental temperature and the diagnosis and treatment of gestational diabetes mellitus: An observational retrospective study. Sci Total Environ. 2020;744:140994. [DOI] [PubMed] [Google Scholar]
- 18.Vasileiou V, Kyratzoglou E, Paschou SA, Kyprianou M, Anastasiou E. The impact of environmental temperature on the diagnosis of gestational diabetes mellitus. Eur J Endocrinol. 2018;178(3):209–14. [DOI] [PubMed] [Google Scholar]
- 19.Antoine-Jonville S, El Khoury D, Faure C, Charlot K, Hue O, Hardy-Dessources MD. Metabolic response to oral glucose tolerance test performed in neutral and warm environmental temperature. Int J Hyperth. 2019;36(1):625–31. [PubMed] [Google Scholar]
- 20.Moses RG, Patterson MJ, Regan JM, Chaunchaiyakul R, Taylor NAS, Jenkins AB. A non-linear effect of ambient temperature on apparent glucose tolerance. Diabetes Res Clin Pr. 1997;36(1):35–40. [DOI] [PubMed] [Google Scholar]
- 21.Metzger BE, Gabbe SG, Persson B, Buchanan TA, Catalano PA, Damm P, Dyer AR, Leiva A, Hod M, Kitzmiler JL, et al. International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy. Diabetes Care. 2010;33(3):676–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Dai L, Deng C, Li Y, Zhu J, Mu Y, Deng Y, Mao M, Wang Y, Li Q, Ma S, et al. Birth weight reference percentiles for Chinese. PLoS ONE. 2014;9(8):e104779. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zhou BF. Predictive values of body mass index and waist circumference for risk factors of certain related diseases in Chinese adults–study on optimal cut-off points of body mass index and waist circumference in Chinese adults. Biomed Environ Sci. 2002;15(1):83–96. [PubMed] [Google Scholar]
- 24.Weight monitoring and evaluation during pregnancy period of Chinese women. [2025 Mar 20]. http://www.nhc.gov.cn/fzs/s7848/202208/2c1c388fcd0c47c58630e5f971ebb468/files/3e00de62d92749fbbe749427a5ab75ef.pdf
- 25.Bieler GS, Brown GG, Williams RL, Brogan DJ. Estimating model-adjusted risks, risk differences, and risk ratios from complex survey data. Am J Epidemiol. 2010;171(5):618–23. [DOI] [PubMed] [Google Scholar]
- 26.Janghorbani M, Stenhouse E, Jones RB, Millward A. Gestational diabetes mellitus in Plymouth, U.K.: prevalence, seasonal variation and associated factors. J Reprod Med. 2006;51(2):128–34. [PubMed] [Google Scholar]
- 27.Katsarou A, Claesson R, Ignell C, Shaat N, Berntorp K. Seasonal Pattern in the Diagnosis of Gestational Diabetes Mellitus in Southern Sweden. J Diabetes Res. 2016;2016:8905474. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Verburg PE, Tucker G, Scheil W, Erwich JJ, Dekker GA, Roberts CT. Seasonality of gestational diabetes mellitus: a South Australian population study. BMJ Open Diabetes Res Care. 2016;4(1):e000286. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Chiefari E, Pastore I, Puccio L, Caroleo P, Oliverio R, Vero A, Foti DP, Vero R, Brunetti A. Impact of Seasonality on Gestational Diabetes Mellitus. Endocr Metab Immune Disord Drug Targets. 2017;17(3):246–52. [DOI] [PubMed] [Google Scholar]
- 30.Retnakaran R, Ye C, Kramer CK, Hanley AJ, Connelly PW, Sermer M, Zinman B. Impact of daily incremental change in environmental temperature on beta cell function and the risk of gestational diabetes in pregnant women. Diabetologia. 2018;61(12):2633–42. [DOI] [PubMed] [Google Scholar]
- 31.Su WL, Lu CL, Martini S, Hsu YH, Li CY. A population-based study on the prevalence of gestational diabetes mellitus in association with temperature in Taiwan. Sci Total Environ. 2020;714:136747. [DOI] [PubMed] [Google Scholar]
- 32.Zhang H, Wang Q, Benmarhnia T, Jalaludin B, Shen X, Yu Z, Ren M, Liang Q, Wang J, Ma W, et al. Assessing the effects of non-optimal temperature on risk of gestational diabetes mellitus in a cohort of pregnant women in Guangzhou, China. Environ Int. 2021;152:106457. [DOI] [PubMed] [Google Scholar]
- 33.Akanji AO, Oputa RN. The effect of ambient temperature on glucose tolerance. Diabet Med. 1991;8(10):946–8. [DOI] [PubMed] [Google Scholar]
- 34.Van Assche FA, Holemans K, Aerts L. Long-term consequences for offspring of diabetes during pregnancy. Br Med Bull. 2001;60:173–82. [DOI] [PubMed] [Google Scholar]
- 35.Herrera E, Ortega-Senovilla H. Disturbances in lipid metabolism in diabetic pregnancy - Are these the cause of the problem? Best Pract Res Clin Endocrinol Metab. 2010;24(4):515–25. [DOI] [PubMed] [Google Scholar]
- 36.de Graaf AA, Freidig A, van Ommen B. Human thermoregulation and obesity. FASEB J. 2009;23(S1):596.592–596.592. [Google Scholar]
- 37.Wei YM, Yang HX. Comparison of the diagnostic criteria for gestational diabetes mellitus in China. Chin J Obst Gynecol. 2011;46(8). [PubMed]
- 38.Schmidt CB, Voorhorst I, van de Gaar VHW, Keukens A, van Potter BJ, Snoek FJ, Honig A. Diabetes distress is associated with adverse pregnancy outcomes in women with gestational diabetes: a prospective cohort study. BMC Pregnancy Childbirth. 2019;19(1):223. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Lydon K, Dunne FP, Owens L, Avalos G, Sarma KM, O’Connor C, Nestor L, McGuire BE. Psychological stress associated with diabetes during pregnancy: a pilot study. Ir Med J. 2012;105(5 Suppl):26–8. [PubMed] [Google Scholar]
- 40.Egan AM, Dunne FP, Lydon K, Conneely S, Sarma K, McGuire BE. Diabetes in pregnancy: worse medical outcomes in type 1 diabetes but worse psychological outcomes in gestational diabetes. QJM. 2017;110(11):721–7. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.