Skip to main content
NCBI home page
BMC Public Health logoLink to BMC Public Health
. 2026 Jul 6;26:2680. doi: 10.1186/s12889-026-28343-8

The association between heat waves and pregnancy outcome among women undergoing assisted reproductive technology in Sichuan, China

Tianyi Ma 1,2,3,#, Yuanyuan Hu 1,2,#, Die Tang 4, Yiting Lin 5, Hanyi Liu 5, Ying Long 1,2, Sophia Zuoqiu 3, Tian Tang 1,2,✉
PMCID: PMC13613809  PMID: 42402568

Abstract

Background

With the increasing frequency and intensity of extreme weather events due to climate change, heat waves have emerged as a significant public health threat. To date, the potential effect of extreme heat wave events, particularly when combined with air pollution, remains poorly understood for pregnancy outcomes among women undergoing assisted reproductive technology (ART).

Methods

A retrospective study included 15,198 women receiving ART and 7519 fresh embryo transfer cycles between 2020 to 2022 at the Reproductive Center of West China Second University Hospital in Chengdu, China. Heat wave, a climate change indicator for extreme temperature events, was calculated based on daily temperature during the period of 85 days prior to oocyte retrieval. All environmental exposure variables, including weather and pollution, were matched geospatially to day 0 to day 85 before oocyte retrieval. Generalized linear model (GLMM) were used to assess the association between environmental exposures and ART outcomes, with secondary analysis using interaction terms between heat waves and individual pollutants.

Results

Exposure to one heat wave event was positively correlated with the likelihood of becoming pregnant (+ 34.9% in univariate model and + 34.5% in multivariate model for heat wave events + 1 time) and this association was more pronounced in women under 35 years of age (+ 53.7% heat wave events + 1 time), while no statistical correlation was observed between exposure to two heat wave events and ART outcomes. Additionally, CO exhibited a significant negative association with biochemical pregnancy for women under 35 years old (-66.2% for CO + 1 mg/m3), and SO2 exhibited a significant negative association on biochemical pregnancy rate for women older than 35 years old (-6.5% for SO2 + 1 μg/m3). Results from the interaction model indicated that concurrent exposure to O3 and two heat wave events was statistical associated with clinical pregnancy (OR = 3.77).

Conclusions

Findings from this study suggest that heat waves could be an important climatic indicator that reflects the impact of extreme weather on pregnancy outcomes among women receiving ART treatment. The synergy between exposure to extreme temperatures and air pollution could be further analyzed to provide deeper insight into the environmental impact on reproductive health.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12889-026-28343-8.

Keywords: Heat waves, Assisted Reproductive Technology (ART), Pregnancy outcomes, Ambient air pollution, Environmental exposure

Background

Over the past several decades, the global prevalence of infertility has shown an upward trend, currently affecting 10% to 25% of reproductive age couples [8, 31, 56]. The growing infertile population has driven an increasing demand for assisted reproductive technology (ART). The high cost and potential health risks associated with ART have imposed considerable economic and psychological burdens on infertile couples [31, 56]. Therefore, understanding and improving these treatment outcomes is crucial for alleviating the burden on patients and optimizing the allocation of healthcare resources.

Follicular development is a complex, multi-stage process in humans, with preantral follicles requiring approximately 85 days to mature into preovulatory follicles. Systematic reviews have indicated that external exposure within the 85-day period prior to oocyte retrieval is associated with a negative impact on pregnancy and live birth, particularly external exposure to ambient air pollution, which may impair the success rate of ART by affecting oocyte development and quality [28, 32]. Ambient air pollution is widely recognized as a major global public health concern and has been repeatedly linked to various adverse health outcomes, including reproductive health outcomes [39, 50]. Studies have confirmed that air pollutants exert a significant impact on the conception rate during ART cycles, potentially leading to infertility, miscarriage, stillbirth, and birth defects [9, 41, 54].

In addition to ambient air pollution, there has been growing interest in the effects of climate change on reproductive health in recent years. Prior research has provided scientific evidence for the climate’s effect on public health, including the induction of cardiovascular diseases, respiratory diseases, and other conditions [19, 24, 27, 55]. However, research on the effects of ambient temperature, an indicator of climate, on ART pregnancy outcomes remains limited and contradictory. For example, past studies in the United States and China have shown that frozen embryo transfer cycles performed in summer yield higher clinical pregnancy rates, while research in Brazil has found that fertilization rates are higher in spring [4, 7, 61]. A cohort study conducted in Shanghai, China, suggested that higher temperatures during cold weather were significantly associated with increased, while another study from Hubei, China, suggested that high temperatures after embryo transfer were linked to reduced clinical pregnancy rates [6, 15]. Nevertheless, all these studies focused exclusively on the analysis of meteorological factors without further investigation of the co-exposure of ambient temperature and air pollutants.

Heat waves are typically defined as prolonged episodes of abnormally high temperatures, and this abnormal increase in atmospheric-related thermal stress can adversely affect human health, particularly among vulnerable groups such as pregnant women [3, 34]. As a comprehensive indicator that emphasizes temperature persistence and abnormality, heat waves can more effectively reflect the cumulative effect of heat load and the body’s additional adaptation than the daily average temperature [30]. In previous studies, heat waves have been more commonly used as a typical extreme weather event in the field of disaster risk assessment, including heat-related deaths, acute changes in cardiovascular and respiratory disease risk, and health impact assessment in the context of combined extreme events [16, 21, 46, 51]. However, heat waves are currently rarely incorporated into existing research on reproductive health, which still mainly focuses on general exposure to ambient temperature. Therefore, in this study, heat waves are used as an indicator of extreme heat exposure to assess their association with ART outcomes and further explore potential interaction effects with air pollution exposure.

Biological plausibility supports a potential relationship between heat exposure and ART outcomes [7, 12, 15]. Follicular development and oocyte maturation are sensitive biological processes, and experimental and epidemiological evidence suggests that thermal stress may affect ovarian function, folliculogenesis, granulosa cell activity and oocyte developmental competence [20, 22]. Potential mechanisms include oxidative stress, altered antioxidant capacity, impaired steroidogenesis, and cellular stress response pathways [20]. In addition, short term thermal stress may activate adaptive responses, including heat shock protein expression pathways, suggesting that the reproductive effects of heat exposure may depend on exposure intensity, duration, and timing rather than following a simple linear pattern [23, 29]. Evaluating heat wave frequency, lagged effects, and cumulative heat wave days during the pre-retrieval window is biologically plausible and may help capture different dimensions of heat exposure burden.

This retrospective cohort study aims to investigate the relationship between heat waves, an established climate change indicator, and ART pregnancy outcomes, while also assessing the combined effects of heat waves and air pollutants on ART pregnancy outcomes. This study also explores potential differences in susceptibility to such combined exposure across age groups. The novelty of this study is twofold: 1) We evaluated heat wave events as sustained extreme temperature exposures instead of focusing only on the average ambient temperature. 2) We incorporated co-exposures into the multivariable framework to explore the association between heat waves and ART outcomes. To our knowledge, this is the first study that utilizes climatic indicators, in combination with ambient air pollution, in assessing ART pregnancy outcomes. By integrating these aspects, the present study seeks to guide future research priorities and support the development of effective preventive measures to protect the reproductive health of current and future generations.

Method

Study population

This was a retrospective cohort study. The subjects of this study were women who received transfers of their fresh embryos at the Reproductive Center of West China Second University Hospital, Sichuan University, from 2020 to 2022. Only patients living in Sichuan were included and they were assumed not to have relocated during the treatment period. Participants were excluded from our study if they had (1) an incomplete medical history, (2) no qualified embryos available after oocyte retrieval, or (3) undergone preimplantation genetic testing. The study protocol was approved by the ethics committees of West China Second University Hospital, Sichuan University (KS402).

ART procedures

Briefly, the ART process is generally divided into four steps: the beginning of controlled ovarian stimulation (COS), oocyte retrieval, embryo transfer, and pregnancy test. The center followed a standard operating procedure (SOP) to administer COS treatment to all participants. COS treatment included long gonadotropin-releasing hormone (GnRH) agonist (-a) protocols, GnRH antagonist (-ant) regimens, and other regimens, including short gonadotropin agonist protocols, mini-stimulation protocols, or progesterone-prompted ovarian stimulation (PPOS) protocols, which were decided by the physician based on the patient’s ovarian reserve. The dose of gonadotropins was adjusted accordingly based on age, BMI and ovarian response. After two or more follicles reached a diameter of ≥ 18 mm, hCG was injected subcutaneously on the trigger day with or without Triptorelin depending on ovarian response. The oocytes were retrieved after 34–36 h. The selection of subsequent fertilization techniques was contingent upon an evaluation of sperm quality and clinical factors, encompassing the utilization of IVF or ICSI. The fresh cycle protocol involves selecting top-quality embryos for transfer 3 days after oocytes have been successfully retrieved and luteal support has begun on the day of oocyte retrieval.

The primary outcomes were biochemical pregnancy and clinical pregnancy. Biochemical pregnancy is defined as when serum β-HCG is greater than 30 mIU/ml at 14 days after embryo transfer. Clinical pregnancy is defined as the identification of an intrauterine gestational sac by ultrasound after five weeks of embryo transfer.

Environmental pollution and meteorology data

Environmental pollution and meteorological data with 0.01° resolution were matched with participants addresses and based on the date of oocyte retrieval [48, 57–59]. Detailed approaches were explained in earlier publications. In brief, hourly raw concentrations of PM2.5 (μg/m3), PM10 (μg/m3), O3 (μg/m3), NO2 (μg/m3), SO2 (μg/m3), and CO (mg/m3) were obtained from China National Environmental Monitoring Centre (CNEMC, http://www.cnemc.cn/). Hourly raw meteorological data with a spatial resolution of 0.25° were obtained from the National Oceanic and Atmospheric Administration (NOAA, https://nomads.ncep.noaa.gov/). We resampled the temperature and relative humidity into 0.01° grid cells by using bilinear interpolation. Additionally, we performed additional estimation of PM2.5 and PM10 based on aerosol optical depth (AOD) data from the Himawari-8 and the Multi-Angle Implementation of Atmospheric Correction (MAIAC) [10, 45]. We conducted additional estimation of gaseous pollutants (NO2, SO2, O3, and CO) based on tropospheric NO2 columns and the total SO2 (and O3, CO) columns from the Sentinel-5P TROPOspheric Monitoring Instrument (TROPOMI) [10, 44, 45, 47]. Subsequently, six separate machine learning models were used to estimate the hourly gridded concentration for each of the six criteria air pollutants. Then 24-h average concentrations were calculated for PM2.5 (μg/m3), PM10 (μg/m3), NO2 (μg/m3), SO2 (μg/m3) and CO (mg/m3) and 8-h average was calculated for O3 (μg/m3) [47, 48, 57–59].

Definition of heat wave and exposure window

We selected the follicular development stage during the follicular development window, which is the 86-day period starting from the day of oocyte retrieval to 85 days before oocyte retrieval, as the exposure window. This stage covers the key pre-retrieval period of follicular development and oocyte maturation, during which environmental exposures may be biologically relevant to oocyte quality and subsequent ART outcomes.

In this study, we used daily mean temperature to identify heat wave events. Heat wave was defined as ≥ 3 consecutive days with daily mean temperature above 35℃ during the 86-day window from the day of oocyte retrieval to 85 days before retrieval. If there were periods of non-heat wave days between two qualifying heat wave periods, which were counted as separate heat wave events.

For each ART cycle, we counted the number of heat wave events within this window and categorized exposure as no event, one event, and two events, with no event as the reference category in the regression models.

Statistical analysis

In this study, we used generalized linear mixed models (GLMMs) with a logit link to assess the association between environmental data and ART outcomes. To account for repeated cycles within the same woman, we included a subject-specific random intercept to account for multiple cycles per woman and unobserved between woman heterogeneity, which minimizes the effect of individual differences. Models were fit with glmmTMB (Template Model Builder, Laplace approximation) to provide efficient and stable likelihood-based estimation for mixed logistic regression.

We estimated the effects of exposures using two models. Single-exposure adjusted models (univariate models) included one environmental exposure at a time and the prespecified covariates, providing an overall association estimate that is less sensitive to collinearity among exposures. Mutually adjusted multi-exposure models (multivariate models) simultaneously included all environmental exposures to estimate the independent association of each exposure while controlling co-exposures.

For woman i and cycle j, the univariate model for is built as:

graphic file with name d33e598.gif 1a

The multivariate model is:

graphic file with name d33e604.gif 1b

where Inline graphic denotes the ART outcome (biochemical pregnancy and clinical pregnancy), Inline graphic is the main exposure of environmental data, Inline graphic represents the other seven variables (air pollutants and relative humidity) included as covariates in the model simultaneously, and Inline graphic includes prespecified covariates: Age, BMI, smoking status, education level (Above college degree, College degree, Less than college degree), job status (Outdoor, Indoor, Undefined), the number of embryo transferred, and season. The season was mapped based on the oocyte retrieval date, with Spring as the reference. Exposures were interpreted per unit (temperature + 1℃, relative humidity + 1%, PM2.5, PM10, NO2, O3, SO2 + 1 μg/m3, CO + 1 mg/m3).

For heat waves, exposure was the number of events during the window, and no event was the reference. The model replaced Inline graphic by a set of level indicators. The univariate model is:

graphic file with name d33e647.gif 2a

The multivariate model is built as:

graphic file with name d33e653.gif 2b

where Inline graphic is the OR for k events using no event as the reference, Inline graphic represents seven variables included as covariates in the model simultaneously. To avoid collinearity, temperature was not included as a covariate in the multivariate model.

Inline graphic denoted a subject-specific random intercept capturing unobserved individual heterogeneity and accounting for within woman correlation across multiple cycles, assumed Inline graphic. Two models used the intraclass correlation coefficient (ICC) to assess variance partitioning. For logistic mixed models, the residual variance on the latent scale was approximated as Inline graphic:

graphic file with name d33e683.gif 3

which can be interpreted as the proportion of total variability attributable to differences among individuals, with the remaining proportion reflecting within individual (unexplained) variability.

In this study, we estimated parameters by maximum likelihood using the Laplace approximation in glmmTMB. We reported ORs, Wald 95% CIs and two-sided Wald p values for the exposure coefficient (or each heat wave event level). For continuous exposures, Inline graphic is the per unit OR, Inline graphic compares k events with 0 event.

Data summary and analysis were conducted in R and Python [13, 33].

Results

Summary of study population and ART outcomes

From 2020 to 2022, 19,075 patients underwent IVF/ICSI cycles at Reproductive Center of West China Second University Hospital, Sichuan University. Among them, 15,198 patients resided in Sichuan province, with 18,342 ART cycles, were enrolled in our study (Fig. 1), and 7519 cycles received fresh embryo transfer. Figure 1 presents the distribution of ART patients classified by age groups in Sichuan Province. Table 1 presents the demographic characteristics of a total of 15,198 participants living in Sichuan Province. In general, the average age and the average BMI of the female participants in the study were 31.69 years old (SD = 4.58) and 21.93 (SD = 2.98). As people get older, the BMI tends to increase. The majority of the study population were of Han ethnicity (14,151, 93.13%), while among the ethnic minorities, Yi (3.15%), Tibetan (1.68%), and Qiang (0.46%) comprised a significant proportion. Almost all individuals were non-smokers (99.19%), smokers account for a larger proportion in the older age group (1.22%). Over half had a bachelor’s degree or higher (63.14%), indicating a high level of education. Half of the study population worked indoors (49.26%), with a similarly high proportion of individuals in the undefined groups (freelancer, unemployed) (45.31%). Outdoor workers account for a larger proportion in the older age group (12.74%), compared with the younger group (4.42% and 7.35%). The majority of research individuals were due to female factors (70.72%). As age increases, female factors account for a larger proportion, indicating that female fertility is affected by age.

Fig. 1.

Fig. 1

Open in a new tab

Distribution of ART patients classified by age groups in Sichuan Province

Table 1.

Summary statistics of the study population information

Characteristic Full Sample < 35 Years Old 35–39 Years Old ≥ 40 Years Old
(15,198 Women,
100%)		 (11,641 Women,
76.60%)		 (2654 Women,
17.46%)		 (903 Women,
5.94%)
Age (years) 31.69 ± 4.58 29.76 ± 2.93 36.54 ± 1.40 42.32 ± 2.30
BMI (kg/m2) 21.93 ± 2.98 21.78 ± 2.97 22.30 ± 2.88 22.86 ± 3.09
Ethnic groups
 Han 14,151 (93.13%) 10,914 (93.76%) 2421 (91.29%) 816 (90.37%)
 Yi 478 (3.15%) 308 (2.65%) 114 (4.30%) 56 (6.20%)
 Tibetan 255 (1.68%) 190 (1.63%) 49 (1.85%) 16 (1.77%)
 Qiang 70 (0.46%) 49 (0.42%) 13 (0.49%) 8 (0.89%)
 Other ethnic groups 241 (1.59%) 179 (1.54%) 55 (2.07%) 7 (0.78%)
Cigarette smoking status
 Smoker 123 (0.81%) 87 (0.75%) 25 (0.94%) 11 (1.22%)
 Non-smoker 15,074 (99.19%) 11,553 (99.25%) 2629 (99.06%) 892 (98.78%)
Education level
 Above college degree 1033 (6.80%) 738 (6.34%) 230 (8.67%) 65 (9.03%)
 College degree 8560 (56.34%) 6722 (57.76%) 1440 (54.28%) 398 (44.69%)
 Less than college degree 5601 (36.86%) 4178 (35.90%) 983 (37.05%) 440 (48.73%)
Working status
 Outdoor worker 824 (5.42%) 514 (4.42%) 195 (7.35%) 115 (12.74%)
 Indoor worker 7483 (49.26%) 5795 (49.81%) 1314 (49.53%) 374 (41.42%)
 Undefined 6883 (45.31%) 5325 (45.77%) 1144 (43.12%) 414 (45.85%)
Infertility factors
 Female factors 10,744 (70.72%) 8041 (69.10%) 1984 (74.81%) 719 (79.62%)
 Male factors 1791 (11.79%) 1574 (13.53%) 190 (7.16%) 27 (2.99%)
 Mixed factors 2558 (16.84%) 1941 (16.68%) 462 (17.42%) 155 (17.17%)
 Unknown factors 99 (0.65%) 81 (0.70%) 16 (0.60%) 2 (0.227%)
Open in a new tab

Values are mean ± SD or n (%)

Percentages were calculated using the total number of participants in each age group as the denominator

Missing values were not shown separately

In this study, an individual may have multiple treatment records. Therefore, ART cycles were used to represent an independent sample (Table 2), indicating that the number of samples is larger than the study population size (18,342). The average number of retrieved eggs, high quality embryos and embryos transplanted was 9.65 (SD = 6.78), 3.55 (SD = 2.81) and 1.64 (SD = 0.48), respectively, all numbers were decreased as the age increased. Most individuals received only IVF treatment (61.57%), while the rest received other treatment options (such as ICSI, RICSI) (38.16%). The biochemical pregnancy rate of the study population was 54.76%, and the clinical pregnancy rate was 48.41%. The probability of pregnancy was significantly affected by age.

Table 2.

Summary statistics of the study ART cycles information and outcomes

Characteristic Full Sample < 35 Years Old 35–39 Years Old ≥ 40 Years Old
(18,342 Cycles,
100%)		 (13,359 Cycles
72.85%)		 (3476 Cycles,
18.94%)		 (1507 Cycles,
8.21%)
No. of high-quality embryos 3.55 ± 2.81 3.87 ± 2.95 2.78 ± 2.12 1.95 ± 1.40
No. of oocytes retrieved 9.65 ± 6.78 10.99 ± 6.86 7.10 ± 5.35 3.62 ± 3.16
No. of Embryos transferred 1.64 ± 0.48 1.66 ± 0.47 1.62 ± 0.49 1.54 ± 0.50
 1 embryo 2674 (35.56%) 1918 (34.12%) 541 (37.78%) 215 (46.24%)
 2 embryos 4845 (64.44%) 3704 (65.88%) 891 (62.22%) 250 (53.76%)
Final treatment
 ICSI 5836 (32.25%) 4105 (30.97%) 1152 (33.69%) 579 (40.72%)
 IVF 11,141 (61.57%) 8200 (61.87%) 2119 (61.98%) 822 (57.81%)
 IVF + ICSI 481 (2.66%) 419 (3.16%) 57 (1.67%) 5 (0.35%)
 IVF + ICSI + RICSI 2 (0.01%) 2 (0.02%) 0 (0.00%) 0 (0.00%)
 IVF + RICSI 373 (2.06%) 319 (2.41%) 49 (1.43%) 5 (0.35%)
 RICSI 213 (1.18%) 173 (1.31%) 34 (0.99%) 6 (0.42%)
IVF outcomes
 Biochemical pregnancy 4117 (54.76%) 3289 (58.50%) 707 (49.37%) 121 (26.02%)
 Clinical pregnancy 3639 (48.41%) 2932 (52.16%) 616 (43.02%) 91 (19.57%)
 No pregnancy 3402 (45.24%) 2333 (41.50%) 725 (50.63%) 344 (73.98%)
Open in a new tab

Values are mean ± SD or n (%)

Percentages were calculated using the total number of participants in each age group as the denominator

Missing values were not shown separately

Variability of meteorological data and air pollution

Table 3 presents the means and standard deviations of various environmental exposures experienced by the study population within the exposure window (from day 0 to day 85 before oocyte retrieval for different age groups). The exposure levels among the three groups were relatively close: between the groups, the temperature differences are approximately < 1℃, the relative humidity differences are approximately < 1%, both falling within the common (relatively suitable) range of the study area.

Table 3.

Summary of individual cycle meteorological and ambient pollutant concentrationa

Age
Group		 Temperature
(Mean ± SD)		 Relative humidity
(Mean ± SD)		 PM2.5
(Mean ± SD)		 PM10
(Mean ± SD)
< 35 18.19 ± 6.64 73.97 ± 6.99 33.63 ± 14.38 53.87 ± 19.69
35–39 18.11 ± 6.96 73.78 ± 6.92 33.88 ± 14.82 54.40 ± 20.62
≥ 40 17.30 ± 6.94 74.33 ± 6.92 35.06 ± 15.48 55.04 ± 20.67
Overall 18.12 ± 6.73 73.96 ± 6.97 33.77 ± 14.54 54.04 ± 19.94

Age

Group

NO2

(Mean ± SD)

SO2

(Mean ± SD)

CO

(Mean ± SD)

O3

(Mean ± SD)
< 35 25.15 ± 10.21 7.80 ± 1.69 0.67 ± 0.11 57.45 ± 18.11
35–39 26.45 ± 11.05 7.79 ± 1.76 0.67 ± 0.12 57.51 ± 18.38
≥ 40 26.77 ± 10.91 7.91 ± 2.30 0.68 ± 0.14 55.13 ± 18.65
Overall 25.50 ± 10.44 7.81 ± 1.75 0.67 ± 0.12 57.32 ± 18.20
Open in a new tab

aIndividual exposure was calculated as the mean concentration across the exposure window

The mean difference of air pollutants was mostly within 1–2 μg/m3 (with CO at the level of 0.01 mg/m3). The standard deviations of the groups were all similar, indicating that the average exposure within the individual window period was relatively stable and age stratification was not the main source of difference in exposure distribution.

We summarized the distribution of heat wave events during the exposure window. Among all 18,342 ART cycles, 17,696 (96.48%) cycles experienced no heat wave event, 522 (2.85%) cycles experienced one heat wave event, and 124 (0.68%) cycles experienced two heat wave events. Among complete case cycles included in the outcome regression analyses, the corresponding numbers were 7247, 218 and 54, respectively. The observed number of heat wave events ranged from 0 to 2, and no cycle experienced more than two heat wave events during the exposure window (Table 4 and Fig. 2).

Table 4.

Distribution of heat wave exposures in the exposure window

Heat wave events All cycles (%) Complete case cycles Biochemical pregnancy Clinical pregnancy
0 17,696 (96.48) 7247 3962 3495
1 522 (2.85) 218 126 116
2 124 (0.68) 54 29 28
Open in a new tab

Fig. 2.

Fig. 2

Open in a new tab

Pregnancy rates by heat wave exposure category

Association between heat waves, environmental exposures and ART outcomes

For environmental exposures, the association between environmental exposure levels and ART outcomes was similar for both biochemical and clinical pregnancy (Table 5). For temperature in the univariate models, both OR of biochemical and clinical pregnancy were above 1, as well as the 95% CI. Especially, two ART outcomes presented statistical associations, which means temperature had a positive effect on pregnancy (separately + 1.7% and + 1.2% for temperature + 1℃). However, after further mutual adjustment for co-exposures, which means in the multivariate models, the temperature effect estimates were attenuated, suggesting that this association may be sensitive to co-exposure confounding and correlations among exposures. We also tested the association between temperature and ART outcomes using a nonlinear model. Results of both biochemical pregnancy and clinical pregnancy showed a shallow U-shaped pattern, with no significant relationship observed (Figure S3).

Table 5.

Regression results of environmental exposures and heat waves on the pregnancy outcomes

Univariate Multivariate
Biochemical Pregnancy Clinical Pregnancy Biochemical Pregnancy Clinical Pregnancy
4117 cycles 3639 cycles 4117 cycles 3639 cycles
Temperature 1.017** (1.005,1.029) 1.012* (1.000,1.025) 1.013 (0.999,1.028) 1.009 (0.995,1.024)
Relative Humidity 1.000 (0.992,1.009) 0.997 (0.989,1.005) 1.009 (0.997,1.020) 1.005 (0.994,1.017)
PM2.5 1.000 (0.995,1.004) 1.001 (0.997,1.006) 1.014 (0.997,1.031) 1.015 (0.998,1.031)
PM10 0.999 (0.996,1.003) 1.001 (0.997,1.004) 0.993 (0.981,1.005) 0.995 (0.983,1.007)
O3 1.004 (1.000,1.008) 1.004 (1.000,1.008) 1.006 (0.999,1.013) 1.006 (0.999,1.013)
SO2 0.972 (0.944,1.000) 0.990 (0.962,1.018) 0.997 (0.964,1.032) 1.008 (0.975,1.043)
NO2 1.001 (0.996,1.007) 1.001 (0.997,1.006) 1.006 (0.997,1.015) 1.003 (0.994,1.012)
CO 0.726 (0.430,1.227) 0.830 (0.493,1.396) 0.511 (0.208,1.256) 0.528 (0.216,1.289)
Heat waves

 One event

(ref = no event)a

 1.204 (0.908,1.597) 1.349* (1.022,1.781) 1.242 (0.921,1.676) 1.345* (1.001,1.807)

 Two events

(ref = no event)b

 1.004 (0.587,1.718) 1.262 (0.741,2.150) 1.000 (0.578,1.730) 1.222 (0.709,2.106)
Open in a new tab

Temperature, Relative humidity, PM2.5, PM10, O3, SO2, NO2 and CO were included in the univariate and multivariate mixed effects logistic regression model (mutually adjusted for co-exposures). Both univariate and multivariate models include control variables: age, BMI, cigarette smoking status, education level, working status, number of embryos transferred, and the season in which the egg retrieval day falls

Heat waves multivariate model excluded temperature as a co-exposure factor

aOne event (ref = no event): Compared to no heat wave event, the odds of pregnancy for individuals experiencing one heat wave event

bTwo events (ref = no event): Compared to no heat wave event, the odds of pregnancy for individuals experiencing two heat wave events

Wald tests significance levels: * < 0.05. 95% confidence interval in parentheses

As for the remaining exposures (relative humidity, PM2.5, PM10, SO2, NO2, O3 and CO), the 95% CI intervals mostly spanned 1, indicating weak overall associations with ART outcomes. However, for CO in the group < 35 years of multivariate models, OR of biochemical pregnancy was less than 1, which presented a negative effect on biochemical pregnancy (−66.2% for CO + 1 mg/m3) (Table S1). Furthermore, biochemical pregnancy presented a negative association for SO2, OR was similarly less than 1, indicating a negative effect on biochemical pregnancy (−6.5% for SO2 + 1 μg/m3). Overall, except for CO and SO2 in certain pregnancy outcomes, for the whole research exposures under mutual adjustment for co-exposures, the odds ratio (OR) for all other environmental exposures were close to 1, indicating a weak overall association between average exposure levels and ART outcomes, with no consistent age gradient observed.

In this study, we defined a heat wave event as one record of a research individual’s exposure to a heat wave event. In both modeling strategies, experiencing one heat wave event was associated with increased odds of clinical pregnancy (+ 34.9% per event for the univariate model), and this association remained positive after mutual adjustment (+ 34.5% per event for the multivariate model, Table 5). This result indicated that the impact of one heat wave event on clinical pregnancy was not influenced by other exposures. This effect was more obvious in females under 35 years old in the multivariate model (+ 53.7% for heat wave events + 1 time), which has the same level of significance (Table S1). For two heat wave events, the estimated odds ratios (OR) for both ART outcomes were with 95% CI spanning 1, which was primarily reflected in the fact that no association was observed between two heat wave events and pregnancy outcomes. In a sensitivity analysis additionally adjusting for male age, the multivariate associations between environmental exposures, heat waves, and ART outcomes remained highly consistent with the main results. Specifically, there was still a positive association between one heat wave event and clinical pregnancy after male age adjustment, with OR = 1.347 (95%CI: 1.002—1.809, Table S2).

We further supplemented the distributed lag sensitivity analysis to study the impact of heat wave events on the ART pregnancy outcomes (Figure S1). The result showed that cumulative heat wave exposure in the short period before oocyte retrieval was positively associated with pregnancy outcomes, with stronger evidence for clinical pregnancy than for biochemical pregnancy. The cumulative association was most apparent within the lag 0–14 days window (OR = 1.88, 95%CI:1.03–3.43), whereas estimates for longer cumulative windows were imprecise due to the wide confidence intervals. The lag response curves suggested a short-term window of elevated odds for clinical pregnancy approximately 6–17 days before oocyte retrieval, while no lag association was observed for biochemical pregnancy. This finding indicated a modest elevation in odds during the first two weeks before oocyte retrieval.

To evaluate potential effect modification of heat waves by ambient co-exposures, we extended the main multivariate model by adding an interaction term between heat wave events (one heat wave event compared with no heat wave event, and two heat wave events compared with no heat wave event) and exposures (relative humidity, PM2.5, PM10, NO2, SO2, CO, O3) at a time (Table S3). In this part, we used Wald tests for assessing the coefficients and likelihood ratio test (LRT) for comparing interaction models and the corresponding main model to evaluate the evidence for effect modification.

Across environmental exposures, the LRT revealed no improvement in model fit when adding interaction terms between heat waves and environmental exposures (p for LRT > 0.05, Table S3), and the Wald tests similarly showed no significant interaction coefficients among most environmental exposures. These results indicated that the association between heat waves and ART outcome is independent of most environmental exposures. However, Wald test results showed a significant interaction signal between O3 and two heat wave events in relation to clinical pregnancy, with an interaction OR of 3.77. This indicates that, for each 1 IQR increase in O3 concentration, the odds of clinical pregnancy outcomes associated with two heat wave events were 3.77 times higher compared to the reference group with no heat wave event.

Additional sensitivity analyses using alternative exposure metrics and definitions. 1) We evaluated cumulative hot days as an alternative exposure metric to capture the duration and cumulative effect of extreme high temperature (Figure S2 and Table S4). Cumulative hot days were defined as the total number of days with daily mean temperature above 35℃ during the 0–85 days before oocyte retrieval. The results suggested that each additional hot day was associated with higher odds of clinical pregnancy (+ 2.7% for each additional hot day). 2) Categorized the 0–85 days mean temperature before oocyte retrieval into quintiles and using the lowest quintile as the reference. Compared with the lowest quintile, the highest quintile was associated with higher odds of both biochemical and clinical pregnancy (separately OR = 1.408 and OR = 1.383), while intermediate quintiles showed positive but no significant associations (Table S5). 3) Conducted the sensitivity analysis using alternative heat wave definitions based on the 95th and 99th percentiles, results showed broadly similar but no significant associations (Table S6).

Discussion

Main findings

To our knowledge, this study was the first investigation to examine the impact of exposure to heat waves and ambient air before oocyte retrieval on females receiving ART treatment. Exposure to one heat wave event was positively correlated with clinical pregnancy, and this association was more pronounced in females younger than 35 years. At the same time, exposure to two heat wave events was non-significantly associated with clinical pregnancy.

Effects of heat waves on pregnancy outcomes

Our study results suggest that exposure to extreme climate events, as indicated by heat waves, is associated with the probability of ART pregnancy. Heat waves represent extreme temperature conditions and are more informative than averaging temperature over a period of time, an exposure variable used by many previous studies. Previous studies demonstrated the influence of ambient temperature on ART outcomes exhibited variations across different stages, temperature regimes, and transfer modalities [4, 6, 15, 18, 61]. Recently, studies in Shanghai and Hubei, China, have provided support for the relationship between temperature anomalies and pregnancy outcomes by showing that higher temperatures in cold weather were associated with elevated clinical pregnancy and live birth rates, whereas elevated temperatures in hot weather showed opposite effects [6, 15, 18]. Further subgroup analysis revealed that women aged 35 years are more susceptible to this effect, suggesting a possible age-related biological response to thermal stimulation [6]. In the current study, the exposure period covered up to 85 days prior to oocyte retrieval. Clinically, the follicular development stage is a critical window highly sensitive to temperature fluctuations, and the impact of temperature on folliculogenesis presents contradictory patterns across studies. During this period, higher ambient temperatures were shown to reduce antral follicle count (AFC) and the number of retrieved oocytes in oocyte donation recipients and autologous oocyte patients, indicating potential adverse effects on ovarian reserve [14, 25, 43]. Yet, higher temperatures during ovarian stimulation were associated with increased mature oocyte count, fertilization rate, and post-thaw oocyte survival rate, suggesting that moderate heat exposure may positively regulate oocyte maturation and select heat-resistant oocytes with high developmental potential [6, 7, 26, 35]. Our study results confirm that the period prior to oocyte retrieval is a sensitive window to meteorological conditions.

The association between heat wave exposure and ART pregnancy outcome also differed between exposure to one heat wave event and exposure to two heat wave events prior to oocyte retrieval. Optimal temperature may be beneficial to follicular development, while extreme temperatures may affect the ovarian microenvironment by causing vitamin D deficiency and hormone disorders, influencing vitamin D levels, estrous cycle, vasoconstriction and altering ovarian blood supply, thereby affecting the oocyte meiosis process, hormone regulation, and ART outcomes [11, 17, 43, 49, 60]. Yet, our results suggest that exposure to excessive heat, as represented by two heat wave events, prior to oocyte retrieval would hinder the success of ART pregnancy outcome, suggesting that there could be a possible optimal exposure to heat.

Co-effects of heat wave and air pollution on pregnancy outcomes

We found a significant positive interactive effect between O3 and two heat wave events on clinical pregnancy. Heat waves and O3 share common underlying meteorological drivers or conditions, including abundant solar insolation, high air temperatures, low wind speeds, and precipitation, resulting in a synergistic enhancing effect on the mortality rate of cardiovascular-related accidents [38, 52, 53]. However, the co-effects of O3 and temperature exposure on pregnancy are limited and may not be additive. A study in Australia on preterm birth and low birthweight reported that the effect of exposure to O3 during the entire pregnancy was enhanced at low temperatures [5]. However, a case-crossover study in California found that the effect of O3 exposure on fetal death was attenuated from significant to non-significant after adjustment for apparent temperature [36]. Another study in the United States suggested that the day of pregnancy loss occurrence and the preceding 3 days of O3 exposure and high temperature during the warm season increased the risk for pregnancy loss, but temperature could modify the relationship between O3 and pregnancy loss [42]. The effects of O3 and temperature exposures on pregnancy may share biological pathways. They could induce oxidative stress, systemic inflammation and change endometrial blood flow [36, 37, 40]. Therefore, the joint effect of co-exposure could reach a saturation point, resulting in a synergistic enhancement or antagonism. Besides, O3 has been shown to be effective in treating endometritis and decreasing blood pressure and preventing the progression of hypertensive disease, thus reducing both inflammation and uterine infection and improving endometrial receptivity [1, 2].

Strengths and limitations

The current study has several strengths. First, this study employs a relatively large sample size compared to previous studies. A larger sample size could provide better statistical power in assessing associations and making inferences. Second, this study utilized a new exposure indicator that builds on previous understanding of temperature’s relationship with pregnancy outcomes. This new indicator is more relevant in the context of climate change and could provide better evidence for making public health inferences.

However, this study is limited by the following: 1) data from a single Chinese province restricts generalizability to regions with distinct climates, such as warmer areas may have higher heat tolerance, potentially underestimating effects; 2) exposure assessment based on residential addresses ignored nonresidential activities, which could have resulted in attenuation in the relationship between exposure and health outcome; 3) the exposure period was 85 days prior to oocyte retrieval, which could not provide distinction between shorter time frames for analysis within that time period. 4) Although we additionally adjusted for male age in sensitivity analyses, other detailed paternal factors, such as semen quality parameters and male factor infertility diagnosis, were not consistently available in the current datasets. Residual confounding from unmeasured paternal factors therefore cannot be fully excluded.

Conclusions

This study examines the relationship between heat exposure prior to oocyte retrieval and pregnancy outcome among women receiving ART treatment in southwest China. The results suggest that mild heat exposure (as represented by one heat event) is positively associated with clinical pregnancy. Additionally, concurrent exposure to heat and air pollution also exhibited significant association with pregnancy outcome, though the direction of association differed depending on the type of pollutants (i.e. O3, SO2, and CO). Findings from this study could provide valuable insight for understanding the potential reproductive health risks among women seeking ART in light of rapid global climate change.

Supplementary Information

Supplementary Material 1. (305.5KB, docx)

Acknowledgements

The authors would like to thank all the participants who contributed their time and effort to this study.

Abbreviations

ART

Assisted Reproductive Technology

COS

Controlled Ovarian Stimulation

SOP

Standard Operating Procedure

GnRH

Gonadotropin-releasing Hormone

PPOS

Progesterone-prompted Ovarian Stimulation

AFC

Antral Follicle Count

Authors’ contributions

T.Y.M.: Data curation, Formal analysis, Methodology, Software, Validation, Visualization, Writing-original draft, Writing-review&editing Y.Y.H.: Data curation, Methodology, Writing-original draft, Writing -review & editing D.T.: Data curation, Methodology Y.T.L.: Data curation, Methodology H.Y.L.: Data curation, Investigation, Project administration Y.L.: Data curation, Methodology, Project administration S.Z.: Conceptualization, Funding acquisition, Investigation, Project administration, Resources, Supervision, Writing—original draft, Writing—review & editing T.T.: Conceptualization, Funding acquisition, Investigation, Project administration, Resources, Supervision, Writing—original draft, Writing—review & editing.

Funding

This work was supported by the Science and Technology Department of Sichuan Province Project [Grant No.2024YFFK0365].

Data availability

The datasets and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the Helsinki Declaration and approved by the Medical Ethics Committee (Institutional Review Board) of West China Second University Hospital, Sichuan University (Approval No.2023278).

Written informed consent to participate was obtained from all participants included in the study.

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.

Tianyi Ma and Yuanyuan Hu contributed equally to this work.

References

  • 1.Akcilar R, Akcer S, Simsek H, Akcilar A, Bayat Z, Genc O. The effect of ozone on blood pressure in DOCA-salt-induced hypertensive rats. Int J Clin Exp Med. 2015;8(8):12783–91. [PMC free article] [PubMed] [Google Scholar]
  • 2.Avila ACA, Diniz NC, Serpa RT, Chaves M, Viu MAO, de Oliveira RA. Effectiveness of ozone therapy in the treatment of endometritis in mares. J Equine Vet Sci. 2022;112:103900. 10.1016/j.jevs.2022.103900. [DOI] [PubMed] [Google Scholar]
  • 3.Boni Z, Bieńkowska Z, Chwałczyk F, Jancewicz B, Marginean I, Serrano PY. What is a heat (wave)? An interdisciplinary perspective. Clim Change. 2023;176(9):129. [Google Scholar]
  • 4.Braga DP, Setti A, Figueira Rde C, Iaconelli A Jr, Borges E Jr. Seasonal variability in the fertilization rate of women undergoing assisted reproduction treatments. Gynecol Endocrinol. 2012;28(7):549–52. 10.3109/09513590.2011.649812. [DOI] [PubMed] [Google Scholar]
  • 5.Chen G, Guo Y, Abramson MJ, Williams G, Li S. Exposure to low concentrations of air pollutants and adverse birth outcomes in Brisbane, Australia, 2003–2013. Sci Total Environ. 2018;622–623:721–6. 10.1016/j.scitotenv.2017.12.050. [DOI] [PubMed] [Google Scholar]
  • 6.Cheng G, Qu R, Song G, Li X, Zhang F, Wang J, et al. Association of ambient temperature and diurnal temperature range with the outcome of in vitro fertilization in women from Hubei, China: a retrospective cohort study. Environ Res. 2024;263(Pt 1):120072. 10.1016/j.envres.2024.120072. [DOI] [PubMed] [Google Scholar]
  • 7.Correia KFB, Farland LV, Missmer SA, Racowsky C. The association between season, day length, and temperature on clinical outcomes after cryopreserved embryo transfer. Fertil Steril. 2022;117(3):539–47. 10.1016/j.fertnstert.2021.11.014. [DOI] [PubMed] [Google Scholar]
  • 8.Cox CM, Thoma ME, Tchangalova N, Mburu G, Bornstein MJ, Johnson CL, et al. Infertility prevalence and the methods of estimation from 1990 to 2021: a systematic review and meta-analysis. Hum Reprod Open. 2022;2022(4):hoac051. 10.1093/hropen/hoac051. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Deng L, Chen G, Duan T, Xie J, Huang G, Li X, et al. Mixed effects of ambient air pollutants on oocyte-related outcomes: a novel insight from women undergoing assisted reproductive technology. Ecotoxicol Environ Saf. 2024;280:116525. 10.1016/j.ecoenv.2024.116525. [DOI] [PubMed] [Google Scholar]
  • 10.Di Q, Kloog I, Koutrakis P, Lyapustin A, Wang Y, Schwartz J. Assessing PM2.5 exposures with high spatiotemporal resolution across the continental United States. Environ Sci Technol. 2016;50(9):4712–21. 10.1021/acs.est.5b06121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Ding M, Lu Y, Wen Q, Xing C, Huang X, Zhang Y, et al. Ovarian PERK/NRF2/CX43/StAR/progesterone pathway activation mediates female reproductive dysfunction induced by cold exposure. Sci Rep. 2024;14(1):10248. 10.1038/s41598-024-60907-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Farland LV, Correia KFB, Missmer SA, Racowsky C. Seasonal variation, temperature, day length, and IVF outcomes from fresh cycles. J Assist Reprod Genet. 2020;37(10):2427–33. 10.1007/s10815-020-01915-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Foundation-Python-Software. Python: a dynamic, open-source programming language. 2024.
  • 14.Gaskins AJ, Minguez-Alarcon L, VoPham T, Hart JE, Chavarro JE, Schwartz J, et al. Impact of ambient temperature on ovarian reserve. Fertil Steril. 2021;116(4):1052–60. 10.1016/j.fertnstert.2021.05.091. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Geng L, Yang Y, Chen Y, Ye T, Qiu A, Bukulmez O, et al. Association between ambient temperature exposure and pregnancy outcomes in patients undergoing in vitro fertilization in Shanghai, China: a retrospective cohort study. Hum Reprod. 2023;38(12):2489–98. 10.1093/humrep/dead192. [DOI] [PubMed] [Google Scholar]
  • 16.Guo Y, Gasparrini A, Armstrong BG, Tawatsupa B, Tobias A, Lavigne E, et al. Heat wave and mortality: a multicountry, multicommunity study. Environ Health Perspect. 2017;125(8):087006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Hoshino Y. Updating the markers for oocyte quality evaluation: intracellular temperature as a new index. Reprod Med Biol. 2018;17(4):434–41. 10.1002/rmb2.12245. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Jiang D, Cai X, Fang H, Li Y, Zhang Z, Chen H, et al. Coexposure to ambient air pollution and temperature and its associations with birth outcomes in women undergoing assisted reproductive technology in Fujian, China: a retrospective cohort study. J Hazard Mater. 2024;481:136539. 10.1016/j.jhazmat.2024.136539. [DOI] [PubMed] [Google Scholar]
  • 19.Jiao A, Sun Y, Avila C, Chiu V, Slezak J, Sacks DA, et al. Analysis of heat exposure during pregnancy and severe maternal morbidity. JAMA Netw Open. 2023;6(9):e2332780. 10.1001/jamanetworkopen.2023.32780. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Kawano K, Sakaguchi K, Madalitso C, Ninpetch N, Kobayashi S, Furukawa E, et al. Effect of heat exposure on the growth and developmental competence of bovine oocytes derived from early antral follicles. Sci Rep. 2022;12(1):8857. 10.1038/s41598-022-12785-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Kenney WL, Craighead DH, Alexander LM. Heat waves, aging, and human cardiovascular health. Med Sci Sports Exerc. 2014;46(10):1891. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Khan A, Khan MZ, Umer S, Khan IM, Xu H, Zhu H, et al. Cellular and molecular adaptation of bovine granulosa cells and oocytes under heat stress. Animals (Basel). 2020;10(1):110. 10.3390/ani10010110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Kregel KC. Invited review: heat shock proteins: modifying factors in physiological stress responses and acquired thermotolerance. J Appl Physiol. 2002;92(5):2177–86. 10.1152/japplphysiol.01267.200. [DOI] [PubMed] [Google Scholar]
  • 24.Lakhoo DP, Brink N, Radebe L, Craig MH, Pham MD, Haghighi MM, et al. A systematic review and meta-analysis of heat exposure impacts on maternal, fetal and neonatal health. Nat Med. 2025;31(2):684–94. 10.1038/s41591-024-03395-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.LaPointe S, McAloon M, Lee JC, Thornburgh S, Nagy ZP, Shapiro DB, et al. Ambient temperature in relation to ovarian reserve and early outcomes following ovarian stimulation and in vitro fertilization. Environ Res. 2025;271:121117. 10.1016/j.envres.2025.121117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Leathersich SJ, Roche CS, Walls M, Nathan E, Hart RJ. Season at the time of oocyte collection and frozen embryo transfer outcomes. Hum Reprod. 2023;38(9):1714–22. 10.1093/humrep/dead137. [DOI] [PubMed] [Google Scholar]
  • 27.Li-Maloney C, Wagar KE, Tetzlaff EJ, Kenny GP. Pregnancy and extreme heat events: a rapid review of evidence related to health outcomes, risk factors and interventions. Women Birth J Aust Coll Midwives. 2025;38(4):101931. 10.1016/j.wombi.2025.101931. [DOI] [PubMed] [Google Scholar]
  • 28.Liu J, Dai Y, Yuan J, Li R, Hu Y, Su Y. Does exposure to air pollution during different time windows affect pregnancy outcomes of in vitro fertilization treatment? A systematic review and meta-analysis. Chemosphere. 2023;335:139076. 10.1016/j.chemosphere.2023.139076. [DOI] [PubMed] [Google Scholar]
  • 29.Oksala NKJ, Ekmekçi FG, Özsoy E, Kirankaya Ş, Kokkola T, Emecen G, et al. Natural thermal adaptation increases heat shock protein levels and decreases oxidative stress. Redox Biol. 2014;3:25–8. 10.1016/j.redox.2014.10.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Perkins SE, Alexander LV. On the measurement of heat waves. J Clim. 2013;26(13):4500–17. [Google Scholar]
  • 31.Qian J, Si Y, Chen Y, Zhu Y, Zhu J, Mo C, et al. Ambient temperature and female infertility prevalence: an ecological study based on the 2019 global burden of disease study. Reprod Biol Endocrinol. 2025;23(1):27. 10.1186/s12958-025-01365-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Qiao JC, Sun LJ, Xie PP, Li ZY, Zhang MY, Gui SY, et al. Association between ambient air pollution exposure and pregnancy outcomes in women treated with assisted reproductive technology: an updated systematic review and meta-analysis. BMC Public Health. 2025;25(1):1639. 10.1186/s12889-024-19301-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.R-Core-Team. R: a language and environment for statistical computing. Vienna: R Foundation for Statistical Computing; 2025. [Google Scholar]
  • 34.Robinson PJ. On the definition of a heat wave. J Appl Meteorol. 2001;40(4):762–75. [Google Scholar]
  • 35.Roth Z. Effect of heat stress on reproduction in dairy cows: insights into the cellular and molecular responses of the oocyte. Annu Rev Anim Biosci. 2017;5:151–70. 10.1146/annurev-animal-022516-022849. [DOI] [PubMed] [Google Scholar]
  • 36.Sarovar V, Malig BJ, Basu R. A case-crossover study of short-term air pollution exposure and the risk of stillbirth in California, 1999–2009. Environ Res. 2020;191:110103. 10.1016/j.envres.2020.110103. [DOI] [PubMed] [Google Scholar]
  • 37.Schifano P, Asta F, Dadvand P, Davoli M, Basagana X, Michelozzi P. Heat and air pollution exposure as triggers of delivery: a survival analysis of population-based pregnancy cohorts in Rome and Barcelona. Environ Int. 2016;88:153–9. 10.1016/j.envint.2015.12.013. [DOI] [PubMed] [Google Scholar]
  • 38.Schnell JL, Prather MJ. Co-occurrence of extremes in surface ozone, particulate matter, and temperature over eastern North America. Proc Natl Acad Sci U S A. 2017;114(11):2854–9. 10.1073/pnas.1614453114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Segal TR, Giudice LC. Systematic review of climate change effects on reproductive health. Fertil Steril. 2022;118(2):215–23. 10.1016/j.fertnstert.2022.06.005. [DOI] [PubMed] [Google Scholar]
  • 40.Sharkhuu T, Doerfler DL, Copeland C, Luebke RW, Gilmour MI. Effect of maternal exposure to ozone on reproductive outcome and immune, inflammatory, and allergic responses in the offspring. J Immunotoxicol. 2011;8(2):183–94. 10.3109/1547691X.2011.568978. [DOI] [PubMed] [Google Scholar]
  • 41.Song H, Clemens T, Doherty RM, Stocker J, Bhattacharya S. Assessing ambient air pollution’s effects on birth outcomes: a Scottish IVF cohort study (2010–2018). Environ Health. 2025;24(1):54. 10.1186/s12940-025-01204-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Tong M, Wang M, Li P, Gong J, Zhu T, Xue T. The short-term effect of ozone on pregnancy loss modified by temperature: findings from a nationwide epidemiological study in the contiguous United States. Sci Total Environ. 2023;902:166088. 10.1016/j.scitotenv.2023.166088. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Wang C, Chen J, Lin Z, Shi L, Ruan Q, Long J, et al. The impact of season, temperature, and direct normal irradiance on IVF pregnancy outcomes: a retrospective cohort study. Int J Biometeorol. 2025;69(8):2053–65. 10.1007/s00484-025-02951-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Wang W, Liu X, Bi J, Liu Y. A machine learning model to estimate ground-level ozone concentrations in California using TROPOMI data and high-resolution meteorology. Environ Int. 2022;158:106917. 10.1016/j.envint.2021.106917. [DOI] [PubMed] [Google Scholar]
  • 45.Wang W, Mao F, Du L, Pan Z, Gong W, Fang S. Deriving hourly PM2.5 concentrations from Himawari-8 AODs over Beijing–Tianjin–Hebei in China. Remote Sens. 2017;9(8):858. [Google Scholar]
  • 46.Wang XY, Guo Y, FitzGerald G, Aitken P, Tippett V, Chen D, et al. The impacts of heatwaves on mortality differ with different study periods: a multi-city time series investigation. PLoS One. 2015;10(7):e0134233. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Wu S, Huang B, Wang J, He L, Wang Z, Yan Z, et al. Spatiotemporal mapping and assessment of daily ground NO(2) concentrations in China using high-resolution TROPOMI retrievals. Environ Pollut. 2021;273:116456. 10.1016/j.envpol.2021.116456. [DOI] [PubMed] [Google Scholar]
  • 48.Wu Y, Di B, Luo Y, Grieneisen ML, Zeng W, Zhang S, et al. A robust approach to deriving long-term daily surface NO(2) levels across China: correction to substantial estimation bias in back-extrapolation. Environ Int. 2021;154:106576. 10.1016/j.envint.2021.106576. [DOI] [PubMed] [Google Scholar]
  • 49.Xu F, Wolf S, Green O, Xu J. Vitamin D in follicular development and oocyte maturation. Reproduction. 2021;161(6):R129–37. 10.1530/REP-20-0608. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Xu Q, Guan Q, Lu Y, Xu J, Deng S, Dong C, et al. Effect of short-term ambient air pollution exposure on early miscarriage and pregnancy hormones with critical window identification. J Hazard Mater. 2023;460:132328. 10.1016/j.jhazmat.2023.132328. [DOI] [PubMed] [Google Scholar]
  • 51.Xu R, Sun H, Zhong Z, Zheng Y, Liu T, Li Y, et al. Ozone, heat wave, and cardiovascular disease mortality: a population-based case-crossover study. Environ Sci Technol. 2023;58(1):171–81. [DOI] [PubMed] [Google Scholar]
  • 52.Xu R, Sun H, Zhong Z, Zheng Y, Liu T, Li Y, et al. Ozone, heat wave, and cardiovascular disease mortality: a population-based case-crossover study. Environ Sci Technol. 2024;58(1):171–81. 10.1021/acs.est.3c06889. [DOI] [PubMed] [Google Scholar]
  • 53.Xu R, Zheng Y, Wei J, Wang S, Chen Q, Luo L, et al. Widowhood disparities in nonaccidental mortality attributable to heat waves and ozone. Eco-Environ Health. 2025;4(3):100169. 10.1016/j.eehl.2025.100169. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Xue T, Tong M, Li J, Wang R, Guan T, Li J, et al. Estimation of stillbirths attributable to ambient fine particles in 137 countries. Nat Commun. 2022;13(1):6950. 10.1038/s41467-022-34250-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Yang X, Xu X, Wang Y, Yang J, Wu X. Heat exposure impacts on urban health: a meta-analysis. Sci Total Environ. 2024;947:174650. 10.1016/j.scitotenv.2024.174650. [DOI] [PubMed] [Google Scholar]
  • 56.Yu J, Fu Y, Zeng L, Xie P, Li L, Zheng Y. Burden of female infertility in China from 1990 to 2019: a temporal trend analysis and forecasting, and comparison with the global level. Sex Health. 2023;20(6):577–84. 10.1071/SH23029. [DOI] [PubMed] [Google Scholar]
  • 57.Zhan Y, Luo Y, Deng X, Chen H, Grieneisen ML, Shen X, et al. Spatiotemporal prediction of continuous daily PM2.5 concentrations across China using a spatially explicit machine learning algorithm. Atmos Environ. 2017;155:129–39. 10.1016/j.atmosenv.2017.02.023. [DOI] [Google Scholar]
  • 58.Zhan Y, Luo Y, Deng X, Grieneisen ML, Zhang M, Di B. Spatiotemporal prediction of daily ambient ozone levels across China using random forest for human exposure assessment. Environ Pollut. 2018;233:464–73. 10.1016/j.envpol.2017.10.029. [DOI] [PubMed] [Google Scholar]
  • 59.Zhan Y, Luo Y, Deng X, Zhang K, Zhang M, Grieneisen ML, et al. Satellite-based estimates of daily NO(2) exposure in China using hybrid random forest and spatiotemporal kriging model. Environ Sci Technol. 2018;52(7):4180–9. 10.1021/acs.est.7b05669. [DOI] [PubMed] [Google Scholar]
  • 60.Zhang L, An G, Wu S, Wang J, Yang D, Zhang Y, et al. Long-term intermittent cold exposure affects peri-ovarian adipose tissue and ovarian microenvironment in rats. J Ovarian Res. 2021;14(1):107. 10.1186/s13048-021-00851-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61.Zhao M, Zhang H, Waters THB, Chung JPW, Li TC, Chan DYL. The effects of daily meteorological perturbation on pregnancy outcome: follow-up of a cohort of young women undergoing IVF treatment. Environ Health. 2019;18(1):103. 10.1186/s12940-019-0538-7. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material 1. (305.5KB, docx)

Data Availability Statement

The datasets and/or analyzed during the current study are available from the corresponding author on reasonable request.

Articles from BMC Public Health are provided here courtesy of BMC

RESOURCES