Association of hyperuricemia with higher miscarriage rates and lower live birth rates in women undergoing IVF/ICSI.

A total of 2189 couples met the inclusion criteria and were included in the final analysis (Fig.  1 ), representing 2189 first transfer cycles and resulting in 1100 clinical pregnancies, 881 live births, and 213 miscarriages. The average age of women was 31.1 ± 4.7 years, and for men, it was 32.7 ± 5.2 years. In the initial transfer cycle, when the serum UA threshold was set at 360 µmol/L, 430 (19.4%) cycles fell into the high UA group, while 1759 (80.6%) cycles were categorized as normal UA. Women in the high UA group tended to have higher values for BMI, blood glucose, lipids, AMH, AFC, TSH, FT4, basal P, E2, and T than those in the normal UA group. Additionally, more patients were diagnosed with PCOS in the high UA group ( p  < 0.05). Table 1 presents detailed results. Fig. 1 Flow chart of participant selection in this study Table 1 Baseline characteristics of the included participants Characteristics Overall Normal group (UA ≤ 360 µmol/L) High UA group (UA > 360 µmol/L) P No. of couples 2189 1759 430 Type of embryo transfer (1st transfer cycle), n (%)  Fresh 1344 1074 (61.1%) 270 (62.8%) 0.543  Frozen 845 685 (38.9%) 160 (37.2%) Female age (y), mean (SD) 31.1 ± 4.7 31.2 ± 4.7 30.0 ± 4.7 0.050* Male age (y), mean (SD) 32.7 ± 5.2 32.6 ± 5.2 32.9 ± 5.2 0.332 Female BMI, kg/m 2 , median (IQR) 22.3 (20.3–24.98) 21.8 (20.0–24.0) 22.5(22.9–27.9) 0.000** No. of patients in different groups based on female weight, n (%) 0.000**  Underweight (BMI < 18.5 kg/m 2 ) 188 175 (9.9%) 13 (3.0%)  Normal-weight(18.5 kg/m 2  ≤ BMI < 24 kg/m 2 ) 1264 1114 (63.3%) 150 (34.9%)  Overweight (24 kg/m 2  ≤ BMI < 28 kg/m 2 ) 555 390 (22.2%) 165 (38.4%)  Obese (BMI ≥ 28 kg/m 2 ) 182 80 (4.5%) 102 (23.7%) Male BMI, kg/m 2 , median (IQR) 24.1 (21.9–26.6) 24.0 (21.93–26.5) 24.2 (21.9–26.9) 0.501 Serum UA, µmol/L, median (IQR) 298 (259–347) 284.0 (249–318) 395 (375–429) 0.000** No. of patients with time interval between UA blood sampling and embryo transfer, n (%) 0.571  Time interval ≤ 1 month 349 295 (16.8%) 54 (12.6%)  1 month < time interval ≤ 3 months 1093 869 (49.4%) 224 (52.1%)  3 months < time interval ≤ 6 months 747 595 (33.8%) 152 (35.3%) AMH, ng/mL, median (IQR) 3.78 (2.09–6.53) 3.70 (2.03–6.40) 3.96 (2.27–6.99) 0.038* AFC, n, median (IQR) 15 (10–24) 15 (10–22) 18 (11–24) 0.000** Basal P level, ng/mL, median (IQR) 0.24 (0.13–0.52) 0.25 (0.16–0.57) 0.20 (0.11–0.37) 0.000** Basal E2 level, pg/mL, median (IQR) 53.32 (33.39–110.10) 54.97 (33.43–121.80) 47.18 (33.27–76.84) 0.002** Basal T level, ng/mL, median (IQR) 0.29 (0.20–0.40) 0.28 (0.20–0.38) 0.33 (0.22–0.45) 0.000** Basal FSH level, mIU/mL, median (IQR) 6.23 (4.68–7.78) 6.32 (4.60–8.02) 6.11 (4.98–7.10) 0.078 Basal LH level, mIU/mL, median (IQR) 5.69 (3.59–9.17) 5.69 (3.61–9.08) 5.65 (3.50–9.61) 0.856 FBG, mmol/L, median (IQR) 5.21 (4.96–5.46) 5.20 (4.94–5.45) 5.27 (5.02–5.56) 0.013* TC, mmol/L, median (IQR) 4.28 (3.77–4.79) 4.23 (3.71–4.76) 4.47 (4.11–4.94) 0.000** TG, mmol/L, median (IQR) 1.10 (0.78–1.58) 1.01 (0.74–1.48) 1.48 (1.16–2.12) 0.000** HDL, mmol/L, median (IQR) 1.37 (1.18–1.58) 1.42 (1.21–1.60) 1.24 (1.08–1.45) 0.000** LDL, mmol/L, median (IQR) 2.53 (2.07–2.97) 2.48 (2.04–2.92) 2.75 (2.45–3.08) 0.000** TSH, µIU/mL, median (IQR) 2.26 (1.59–3.15) 2.23 (1.59–3.11) 2.44 (1.60–3.36) 0.047* FT3, pg/mL, median (IQR) 4.87 (4.49–5.29) 4.86 (4.47–5.27) 4.92 (4.54–5.33) 0.056 FT4, ng/mL, median (IQR) 17.07 (15.49–18.66) 17.12 (15.55–18.69) 16.83 (15.27–18.45) 0.033* Type of infertility, n (%)  Primary 1069 849 (48.3%) 220 (51.2%) 0.282  Secondary 1120 910 (51.7%) 210 (48.8%) Infertility diagnosis, n (%)  Female factor 1781 1427 (81.1%) 354 (82.3%) 0.438  Male factor 102 87 (4.9%) 15 (3.5%)  Both 306 245 (13.9%) 61 (14.2%) Ovarian stimulation protocol, n (%)  Long protocol 1026 803 (45.7%) 223 (51.9%) 0.165  Antagonist protocol 991 817 (46.4%) 174 (40.5%)  Mild stimulation protocol 34 29 (1.6%) 5 (1.7%)  PPOS protocol 113 89 (5.1%) 24 (5.6%)  Other protocols 25 21 (1.2%) 4 (0.9%) Frozen embryo transfer cycle protocols, n (%) 845 685 160 0.375  Natural 18 15 (15/685,2.19%) 3 (3/160, 1.88%)  Stimulated 9 9 (9/685, 1.31%) 0 (0/160, 0%)  Programmed 818 661 (661/685,96.5%) 157 (157/160,98.13%) Fertilization, n (%)  IVF 1806 1443 (82.0%) 363 (84.4%) 0.257  ICSI 383 316 (18.0%) 67(15.6%) No. of oocytes retrieved, n, median (IQR) 12 (8–17) 12 (8–16) 13 (9–17) 0.040* Duration of infertility, years, median (IQR) 3 (2–5) 3 (2–5) 3 (2–5) 0.339 No. of transferred embryos, (1st transfer cycle), n (%)  1 449 366 (20.8%) 83 (19.3%) 0.530  2 1740 1393 (79.2%) 347 (80.7%) Day of embryo transfer, n (%)  Cleavage 1697 1376 (78.2%) 321 (74.7%) 0.122  Blastocyst 492 383 (21.8%) 109 (25.3%) Endometrial thickness on transfer day (cm), median (IQR) 1.0 (0.9–1.2) 1.0 (0.9–1.2) 1.0 (0.9–1.2) 0.326 No. of patients diagnosed with PCOS, n (%) 251 165 (9.4%) 85 (19.8%) 0.000** Data are reported as median (interquartile range) or number (percentage), as applicable UA uric acid, BMI body mass index, AFC antral follicle account, AMH anti-Müllerian hormone, LH luteinizing hormone, FSH follicle-stimulating hormone,  P progesterone, E2 estradiol, T testosterone, FBG fasting blood glucose, TC total cholesterol, TG triglycerides, HDL high-density lipoprotein, LDL low-density lipoprotein, IVF in vitro fertilization, ICSI intracytoplasmic sperm injection, SD standard deviation, PPOS progestin primed ovarian stimulation, PCOS polycystic ovary syndrome, IQR interquartile range, SD standard deviation A P value ≤ 0.05 was considered statistically significant (* for P  ≤ 0.05; ** for P  ≤ 0.01) Flow chart of participant selection in this study Baseline characteristics of the included participants Data are reported as median (interquartile range) or number (percentage), as applicable UA uric acid, BMI body mass index, AFC antral follicle account, AMH anti-Müllerian hormone, LH luteinizing hormone, FSH follicle-stimulating hormone,  P progesterone, E2 estradiol, T testosterone, FBG fasting blood glucose, TC total cholesterol, TG triglycerides, HDL high-density lipoprotein, LDL low-density lipoprotein, IVF in vitro fertilization, ICSI intracytoplasmic sperm injection, SD standard deviation, PPOS progestin primed ovarian stimulation, PCOS polycystic ovary syndrome, IQR interquartile range, SD standard deviation A P value ≤ 0.05 was considered statistically significant (* for P  ≤ 0.05; ** for P  ≤ 0.01) The LASSO logistic regression analysis was conducted to comprehensively analyze the role of all the variables in prediction. We achieved a risk model based on the optimal lambda (λ) (Fig.  2 ), specifically λ-min (λ corresponding to the minimum mean error), and selected variables with non-zero coefficients through λ-min (Supplementary Table 1). Fig. 2 a - c LASSO coefficient profiles of variables. a LASSO regression analysis of factors related to live birth. b LASSO regression analysis of factors related to clinical pregnancy. c LASSO regression analysis of factors related to miscarriage. Each continuous variable was shown as a colored line; ( d - f ) The lambda (λ) selection process in the LASSO regression is depicted. The value of λ that gave the minimum average binomial deviance was applied to select the features. The left vertical line represents the λ-min, which screened out 11 valuable features for live birth ( d ), 8 for clinical pregnancy ( e ) and 3 for miscarriage ( f ) a - c LASSO coefficient profiles of variables. a LASSO regression analysis of factors related to live birth. b LASSO regression analysis of factors related to clinical pregnancy. c LASSO regression analysis of factors related to miscarriage. Each continuous variable was shown as a colored line; ( d - f ) The lambda (λ) selection process in the LASSO regression is depicted. The value of λ that gave the minimum average binomial deviance was applied to select the features. The left vertical line represents the λ-min, which screened out 11 valuable features for live birth ( d ), 8 for clinical pregnancy ( e ) and 3 for miscarriage ( f ) Analysis revealed similar rates of oocyte retrieval, embryo utilization, high-quality cleavage embryos, blastocyst formation between groups, and the normal group had a higher fertilization rate. During the first embryo transfer cycles, CPR between groups was approximately 50%. However, women with higher UA values exhibited lower LBR (41.7% vs 34.2%, p  < 0.05), indicating a higher risk of miscarriage in the HUA group (16.4% vs 31.6%, p  < 0.05). Table 2 presents detailed findings. Table 2 Reproductive outcomes in the high and normal UA groups Outcome measures Normal group High UA group P a Oocytes retrieval rate (95% CI) 67.2 (66.4–68.0) 67.4 (65.7–69.2) 0.946 a Fertilization rate (95% CI) 84.7 (83.9–85.6) 82.7 (81.1–84.3) 0.004** a Embryo utilization rate (95% CI) 92.8 (91.8–93.8) 94.0 (92.4–95.6) 0.126 a High-quality cleavage embryo rate (95% CI) 50.2 (48.7–51.7) 53.1 (50.1–56.1) 0.469 a Blastocyst formation rate (95% CI) 52.4 (51.0–53.8) 52.1 (49.4–54.7) 0.823 a High-quality blastocyst rate (95% CI) 65.6 (64.0–67.2) 64.7 (61.7–67.7) 0.799 b Clinical pregnancy rate (n/N) 50.3 (885/1759) 50.0 (215/430) 0.914 b Miscarriage rate (n/N) 16.4 (145/885) 31.6 (68/215) 0.000** b Live birth rate (n/N) 41.7 (734/1759) 34.2 (147/430) 0.000** Detailed definitions of reproductive outcomes are provided in the Materials and Methods a rates were calculated individually for each cycle and non-parametric test (Mann–Whitney test) was carried out b pregnancies or miscarriages or live births were pooled in each group for calculation and the chi-square test was used A P value ≤ 0.05 was considered statistically significant (* for P  ≤ 0.05; ** for P  ≤ 0.01). UA  uric acid; High UA group, UA > 360 µmol/L; Normal group, UA ≤ 360 µmol/L Reproductive outcomes in the high and normal UA groups Detailed definitions of reproductive outcomes are provided in the Materials and Methods a rates were calculated individually for each cycle and non-parametric test (Mann–Whitney test) was carried out b pregnancies or miscarriages or live births were pooled in each group for calculation and the chi-square test was used A P value ≤ 0.05 was considered statistically significant (* for P  ≤ 0.05; ** for P  ≤ 0.01). UA  uric acid; High UA group, UA > 360 µmol/L; Normal group, UA ≤ 360 µmol/L Various confounding factors were considered to explore the association between elevated serum UA and adverse reproductive outcomes. Model 1 represents an unadjusted model, while Model 2 was adjusted for confounding factors selected by LASSO regression based on Model 1. Table 3 presents the ORs for relevant reproductive outcomes through logistic regression. In Model l (unadjusted model), the high UA group had a significantly lower risk of live birth (OR, 0.744; 95% CI: 0.598–0.925) and a higher risk of miscarriage (OR, 2.329 95% CI: 1.662–3.264) than the normal UA group, with no significant change in the CPR. After adjusting for confounding factors (Model 2), a similar risk of live birth (OR, 0.682; 95% CI: 0.538–0.866) and miscarriage (OR, 1.919; 95% CI: 1.384–2.660) was still noted in the high UA group. This trend is further illustrated in Fig.  3 through GAMs. Table 3 Associations between HUA and reproductive outcomes Model Response Dependent variables Live birth Clinical pregnancy Miscarriage Model 1 B −0.296 0.005 0.845 S.E 0.111 0.106 0.172 Wald 7.058 0.002 24.090 P 0.008** 0.961 0.000** OR (95% CI) 0.744 (0.598–0.925) 1.005 (0.817–1.237) 2.329 (1.662–3.264) Model 2 B −0.382 −0.099 0.652 S.E 0.122 0.114 0.167 Wald 9.850 0.753 15.306 P 0.002** 0.386 0.000** OR (95% CI) 0.682 (0.538–0.866) 0.906 (0.724–1.133) 1.919 (1.384–2.660) HUA hyperuricemia. Model 1 is an unadjusted model, and Model 2 is adjusted for confounding factors selected by LASSO regression based on Model 1. Model 2 for live birth is adjusted for age, AFC, AMH, number of embryos transferred, basal FSH, basal LH, TG, FT3, endometrial thickness on transfer day, and type of embryos transferred (cleavage or blastocyst). Model 2 for clinical pregnancy is adjusted for age, AFC, AMH, TG, FT3, number of embryos transferred, endometrial thickness on transfer day, and type of embryos transferred (cleavage or blastocyst). Model 2 for miscarriage is adjusted for age, basal LH and FT3 A P value ≤ 0.05 was considered statistically significant (* for P  ≤ 0.05; ** for P  ≤ 0.01) Fig. 3 Serum UA level associated with live birth ( a ), and miscarriage rates ( b ) using GAMs by Model 1(unadjusted model) and Model 2(model after adjusting for confounding factors selected by LASSO regression) Associations between HUA and reproductive outcomes HUA hyperuricemia. Model 1 is an unadjusted model, and Model 2 is adjusted for confounding factors selected by LASSO regression based on Model 1. Model 2 for live birth is adjusted for age, AFC, AMH, number of embryos transferred, basal FSH, basal LH, TG, FT3, endometrial thickness on transfer day, and type of embryos transferred (cleavage or blastocyst). Model 2 for clinical pregnancy is adjusted for age, AFC, AMH, TG, FT3, number of embryos transferred, endometrial thickness on transfer day, and type of embryos transferred (cleavage or blastocyst). Model 2 for miscarriage is adjusted for age, basal LH and FT3 A P value ≤ 0.05 was considered statistically significant (* for P  ≤ 0.05; ** for P  ≤ 0.01) Serum UA level associated with live birth ( a ), and miscarriage rates ( b ) using GAMs by Model 1(unadjusted model) and Model 2(model after adjusting for confounding factors selected by LASSO regression) Tables 4 and 5 demonstrate that elevated serum UA levels had a more pronounced impact on live birth in patients undergoing FET cycles than in those undergoing fresh-embryo transfer cycles. After adjusting for confounding factors, a lower risk of live birth was still noted in the high UA group (OR, 0.566; 95% CI: 0.380–0.845). Table 4 Reproductive outcomes in patients with high and normal UA levels in different groups Outcome measures Group P Group P Fresh embryo transfer Frozen-thawed embryo transfer Normal group High UA group Normal group High UA group b Clinical pregnancy rate (n/N) 48.5 (521/1074) 50.0 (135/270) 0.683 53.1 (364/685) 50 (80/160) 0.483 b Miscarriage rate (n/N) 15.4 (80/521) 31.1 (42/135) 0.000** 17.9 (65/364) 32.5 (26/80) 0.005** b Live birth rate (n/N) 40.5 (435/1074) 34.4 (93/270) 0.070 43.6 (299/685) 33.8 (54/160) 0.026* Age < 35 years Age ≥ 35 years Normal group High UA group Normal group High UA group a Oocytes retrieval rate (95% CI) 65.6 (64.5–66.7) 66.6 (64.0–69.1) 0.471 71.7 (69.4–73.9) 66.9 (61.9–71.8) 0.077 a Fertilization rate (95% CI) 84.8 (83.2–85.5) 81.8 (79.5–84.2) 0.057 84.2 (81.7–86.7) 84.2 (80.1–88.4) 0.987 a Embryo utilization rate (95% CI) 93.3 (91.6–94.9) 92.7 (90.2–95.1) 0.747 93.3 (91.0–95.7) 95.3 (91.5–99.1) 0.473 a High-quality cleavage embryo rate (95% CI) 48.6 (46.5–50.6) 48.2 (44.0–52.5) 0.799 54.1 (49.7–658.5) 61.2 (51.3–71.1) 0.238 a Blastocyst formation rate (95% CI) 53.5 (51.5–55.5) 52.2 (48.4–56.0) 0.566 48.0 (43.5–52.4) 48.3 (39.2–57.4) 0.948 a High-quality blastocyst rate (95% CI) 67.0 (64.7–69.2) 62.9 (59.0–66.9) 0.081 68.1 (62.5–73.7) 67.6 (56.0–79.2) 0.934 b Clinical pregnancy rate (n/N) 54.4 (849/1562) 52.3 (56/107) 0.690 37.0 (157/424) 39.6 (38/96) 0.642 b Live birth rate (n/N) 47.0 (627/1335) 36.8 (123/334) 0.001** 25.2 (107/424) 25.0 (24/96) 1.000 b Miscarriage rate (n/N) 13.2 (96/728) 30.5 (54/177) 0.000** 31.2 (49/157) 36.8 (14/38) 0.563 BMI < 24 kg/m 2 BMI ≥ 24 kg/m 2 Normal group High UA group Normal group High UA group a Oocytes retrieval rate (95% CI) 66.8 (65.8–67.7) 68.4 (65.5–71.4) 0.686 68.5 (67.0–69.9) 66.8 (64.7–69.0) 0.359 a Fertilization rate (95% CI) 85.0 (84.1–86.0) 83.4 (80.9–85.9) 0.094 83.9 (82.2–85.5) 82.3 (80.2–84.3) 0.081 a Embryo utilization rate (95% CI) 92.1 (91.1–93.1) 93.2 (90.8–95.6) 0.420 94.7 (92.1–97.4) 94.5 (92.4–96.7) 0.584 a High-quality cleavage embryo rate (95% CI) 49.6 (47.8–51.3) 53.1 (48.1–58.2) 0.756 52.0 (48.9–55.1) 53.1 (49.3–56.8) 0.905 a Blastocyst formation rate (95% CI) 52.7 (51.0–54.4) 53.6 (49.7–57.5) 0.669 51.8 (49.3–54.4) 51.2 (47.7–54.7) 0.665 a High-quality blastocyst rate (95% CI) 65.5 (63.6–67.4) 63.6 (58.8–68.4) 0.696 65.9 (62.9–68.8) 65.4 (61.6–69.2) 0.887 b Clinical pregnancy rate (n/N) 49.7 (640/1289) 51.5 (84/163) 0.678 52.1 (245/470) 49.1 (131/267) 0.444 b Live birth rate (n/N) 41.9 (540/1289) 33.1 (54/163) 0.034* 41.3 (194/470) 34.8 (93/267) 0.099 b Miscarriage rate (n/N) 15.2 (97/640) 35.7 (30/84) 0.000** 19.6 (48/245) 29.0 (38/131) 0.040* Detailed definitions of reproductive outcomes are provided in the Materials and Methods a rates were calculated individually for each cycle and non-parametric test (Mann–Whitney test) was carried out b pregnancies or miscarriages or live births were pooled in each group for calculation and the chi-square test was used A P value ≤ 0.05 was considered statistically significant (* for P  ≤ 0.05; ** for P  ≤ 0.01). UA uric acid; High UA group, UA > 360 µmol/L; Normal group, UA ≤ 360 µmol/L Table 5 Associations between HUA and reproductive outcomes in different subgroups Group Dependent variables Model 1 Model 2 P ORs 95% CI P ORs 95% CI Fresh embryo transfer Live birth 0.069 0.772 (0.584–1.020) 0.089 0.771 0.571–1.041 Clinical pregnancy 0.662 1.061 (0.813–1.386) 0.961 0.993 0.751–1.313 Miscarriage 0.000** 2.490 (1.611–3.847) 0.001** 2.034 1.336–3.097 Frozen-thawed embryo transfer Live birth 0.023* 0.658 (0.458–0.943) 0.005** 0.566 0.380–0.845 Clinical pregnancy 0.474 0.882 (0.625–1.244) 0.190 0.773 0.527–1.135 Miscarriage 0.004** 2.215 (1.292–3.798) 0.035* 1.748 1.039–2.943 Young (Age < 35 years old) Live birth 0.001** 0.658 (0.514–0.843) 0.002** 0.658 0.506–0.857 Clinical pregnancy 0.614 0.940 (0.739–1.195) 0.403 0.898 0.697–1.156 Miscarriage 0.000** 2.890 (1.966–4.249) 0.000** 2.215 1.519–3.231 Older (Age ≥ 35 years old) Live birth 0.962 0.988 (0.592–1.647) 0.615 0.865 0.491–1.522 Clinical pregnancy 0.641 1.114 (0.708–1.755) 0.832 0.949 0.583–1.543 Miscarriage 0.506 1.286 (0.613–2.696) 0.523 1.244 0.636–2.431 Normal-weight (BMI < 24 kg/m 2 ) Live birth 0.033** 0.687 (0.487–0.970) 0.033* 1.034 1.002–1.067 Clinical pregnancy 0.651 1.078 (0.778–1.494) 0.935 0.986 0.699–1.391 Miscarriage 0.000** 3.110 (1.894–5.106) 0.000** 2.369 1.465-3.832 Over-weight (BMI ≥ 24 kg/m 2 ) Live birth 0.085 0.760 (0.557–1.038) 0.067 0.721 0.508–1.023 Clinical pregnancy 0.424 0.885 (0.655–1.195) 0.192 0.805 0.581–1.115 Miscarriage 0.039* 1.677 (1.025–2.742) 0.145 1.428 0.884–2.305 HUA hyperuricemia. Model 1 is an unadjusted model, and Model 2 is adjusted for confounding factors selected by LASSO regression based on Model 1. Model 2 for live birth is adjusted for age, AFC, AMH, number of embryos transferred, basal FSH, basal LH, TG, FT3, endometrial thickness on transfer day, and type of embryos transferred (cleavage or blastocyst). Model 2 for clinical pregnancy is adjusted for age, AFC, AMH, TG, FT3, number of embryos transferred, endometrial thickness on transfer day, and type of embryos transferred (cleavage or blastocyst). Model 2 for miscarriage is adjusted for age, basal LH and FT3 A P value ≤ 0.05 was considered statistically significant (* for P  ≤ 0.05; ** for P  ≤ 0.01) Reproductive outcomes in patients with high and normal UA levels in different groups Detailed definitions of reproductive outcomes are provided in the Materials and Methods a rates were calculated individually for each cycle and non-parametric test (Mann–Whitney test) was carried out b pregnancies or miscarriages or live births were pooled in each group for calculation and the chi-square test was used A P value ≤ 0.05 was considered statistically significant (* for P  ≤ 0.05; ** for P  ≤ 0.01). UA uric acid; High UA group, UA > 360 µmol/L; Normal group, UA ≤ 360 µmol/L Associations between HUA and reproductive outcomes in different subgroups HUA hyperuricemia. Model 1 is an unadjusted model, and Model 2 is adjusted for confounding factors selected by LASSO regression based on Model 1. Model 2 for live birth is adjusted for age, AFC, AMH, number of embryos transferred, basal FSH, basal LH, TG, FT3, endometrial thickness on transfer day, and type of embryos transferred (cleavage or blastocyst). Model 2 for clinical pregnancy is adjusted for age, AFC, AMH, TG, FT3, number of embryos transferred, endometrial thickness on transfer day, and type of embryos transferred (cleavage or blastocyst). Model 2 for miscarriage is adjusted for age, basal LH and FT3 A P value ≤ 0.05 was considered statistically significant (* for P  ≤ 0.05; ** for P  ≤ 0.01) GAMs revealed a nonlinear relationship between UA and reproductive outcomes in different subgroups. As UA increased, LBR decreased and the miscarriage rate increased in both the fresh and FET subgroups. However, the probability of LBR decreased more in the FET subgroup (Fig.  4 a). Fig. 4 GAMs revealed a nonlinear relationship between serum UA level and reproductive outcomes in different subgroups by using Model 2 GAMs revealed a nonlinear relationship between serum UA level and reproductive outcomes in different subgroups by using Model 2 The study population was further divided into subgroups based on female age and BMI. Patients aged < 35 years were categorized as the young group, and those aged ≥ 35 years were designated as the older group. There were 1669 patients in the young group and 520 in the older group. According to the recommendations of the Working Group on Obesity in China [ 39 , 40 ], the normal weight group was defined as having BMI < 24 kg/m 2 and included 1452 women, and the overweight group had BMI ≥ 24 kg/m 2 , comprising 737 women. In the young group, HUA emerged as an independent factor affecting the occurrence of live births and miscarriages. However, in the older group, reproductive outcomes became less sensitive to UA levels (Tables 4 and 5 ). In the older group, UA had a limited impact on reproductive outcomes, with wide confidence intervals (Fig.  4 b). The increase in UA levels negatively impacted live birth and promoted miscarriage, particularly in the normal weight group (Tables 4 and 5 , Fig.  4 c). After adjusting for confounding factors using Model 2, in the normal weight group, the probability of live birth rapidly decreased when UA exceeded 360 µmol/L, whereas this phenomenon occurred in the overweight group when UA was at smaller values (Fig.  4 c).

This retrospective cohort study was conducted at the First Affiliated Hospital of Kunming Medical University from March 2016 to April 2023. The study protocol was reviewed and approved by the Ethics Committee of the First Affiliated Hospital of Kunming Medical University (Approval No. [2022] Ethical Review No. 60). Given the retrospective design and the use of de-identified data extracted from medical records, the requirement for informed consent was waived by the ethics committee. Infertile couples who received IVF/ICSI treatment during the study period were potentially eligible for participation in this study. Couples were included in this study if: 1) at least one oocyte was retrieved in each cycle; 2) at least one embryo transferred in each cycle; 3) records of UA levels within 6 months of embryo transfer exist.4) the females were aged 20–42 years old; Cycles were excluded if:1) the female or male partner had a chromosomal abnormality; 2) cycles were performed with frozen sperm or sperm from testicular/epididymal aspiration; 3) the female was diagnosed with cancer, endometriosis, gout, pseudogout or immune disorder (such as systemic lupus erythematosus, rheumatoid arthritis and ulcerative colitis, etc.); 4) had known uterine anomalies, including intrauterine adhesion, septal uterine cavity, adenomyosis and fibroids with a diameter larger than 4 cm; 5) pregnancy loss had occurred ≥ 2 times; 6) DNA fragmentation index (DFI) of the male was greater than 30% [ 32 ]. Only the first oocyte retrieval cycle was included in this study. According to clinical guidelines and previous studies, group analyzes were performed by level of UA threshold in serum of 360 µmol/L [ 33 , 34 ], which was defined as HUA. No specific intervention or follow-up protocol was applied to patients with elevated serum UA levels during the study period, as this was a retrospective observational study based on routine clinical data. Blood samples were acquired after overnight fasting for at least 8 h. The levels of basal sex hormones were measured by chemiluminescence immunoassay (ECLIA) (Cobas e601, Roche, Switzerland), including luteinizing hormone (LH) level, follicle-stimulating hormone (FSH) level, progesterone (P) level, estradiol (E2) level, testosterone (T) level. Serum UA, fasting blood glucose (FBG), total cholesterol (TC), triglyceride (TG), low-density lipoprotein (LDL), and high-density lipoprotein (HDL) concentrations were quantified using an autoanalyzer (Cobas 8000, Roche, Switzerland). The levels of serum free triiodothyronine (FT3), free thyroxine (FT4) and thyroid stimulating hormone (TSH) were measured by ECLIA (i2000 SR, Abbott, United States). Serum UA, blood lipids, blood sugar and thyroid hormone are matched according to the embryo transfer date through Python. If there are multiple results, the most recent one is matched. If the time interval between the serum UA record and the first transfer data exceeds 6 months, this patient will be excluded. If the time interval between record of blood lipid/sugar sampling and the first transfer date exceeds 6 months, this blood test will be marked as a missing value. All patients received a standardized ovarian stimulation protocol, oocyte retrieval, fertilization, embryo transfer, and luteal support. Patients received ovarian stimulation protocols according to their age and ovarian reserve (evaluated based on anti-Müllerian hormone concentration and antral follicle count). Follicle development was monitored using transvaginal ultrasound and reproductive hormone levels. When two to three follicles reached 18 mm in diameter, final oocyte maturation was triggered by administering human chorionic gonadotropin. Transvaginal ultrasound-guided oocyte retrieval was performed 34–36 h after triggering maturation. All follicles with diameters of ≥ 10 mm were aspirated. Fertilization was achieved using either conventional IVF or ICSI, depending on the patient’s condition. For a detailed description of these protocols, please refer to our previous publication [ 35 ]. The primary outcome was the live birth rate (LBR) of the first transfer cycle. Secondary outcome measures include clinical pregnancy rate (CPR), miscarriage rate, oocyte retrieval, fertilization, embryo utilization, blastocyst formation, high-quality cleavage embryo rate and high-quality blastocyst formation rate. A clinical pregnancy was defined by the presence of a gestational sac with fetal heart activity observed during ultrasound examination 35 days after embryo transfer. The CPR was calculated as the number of cycles with the presence of at least one gestational sac on ultrasound divided by the number of transfer cycles. The termination of a pregnancy before the 28 weeks of gestation was considered a miscarriage. The miscarriage rate was calculated as the number of miscarriages divided by the number of patients diagnosed with clinical pregnancy. A live birth was defined as the delivery of any viable infant at 28 weeks’ gestation or later. The LBR was calculated as the number of live births after 28 weeks’ gestation divided by the number of patients receiving embryo transfer. The oocyte retrieval rate was defined as the ratio of the total number of oocytes retrieved to the number of follicles punctured (≥ 10 mm in diameter) on the day of oocyte retrieval. The fertilization rate was calculated as the number of normally fertilized oocytes (two pronuclei) divided by the number of oocytes used for insemination (IVF cycle) or injection (ICSI cycle). The embryo utilization rate was calculated as the number of grades I-III embryos divided by the total number of embryos evaluated on day 3 after fertilization. The high-quality cleavage embryo rate refers to the percentage of grade I-II embryos of all of the embryos evaluated on day 3. The blastocyst formation rate was calculated as the number of embryos that reached the blastocyst stage at day 5/6 divided by the number of two-pronuclei embryos subjected to blastocyst culture. The high-quality blastocyst rate was calculated as the number of blastocysts rated as grade 4BB or higher divided by the total number of embryos that underwent blastocyst culture. Statistics, computation, and plotting were performed using Python programming language (version 3.12.1) and SPSS software (version 23.0, Inc., Chicago, IL, USA). All tests were two-tailed, and P value < 0.05 was considered statistically significant. The distribution of data was assessed by Kolmogorov–Smirnov test at first. If data followed a Gaussian distribution, Student’s t-test were carried out and data were presented as mean ± SD. If data were not normally distributed, non-parametric tests (Mann–Whitney test for two groups or Kruskal–Wallis test with Dunn’s post-test for three or more groups) and data were presented as median (inter quartile range, IQR). The categorical variables are described in terms of n (%), the difference test was the chi-square test, and Fisher’s exact probability test was used when the predicted number was small. Considering that reproductive outcomes (live birth, miscarriage and clinical pregnancy) are binary variables, logistic regression was performed. Potential confounding variables included HUA, age, body mass index (BMI), anti-Müllerian hormone (AMH), antral follicle count (AFC), basal LH level, basal FSH level, basal P level, basal E2 level, basal T level, number of embryos transferred, endometrial thickness on transfer day, ovarian stimulation protocol, cleavage embryo or blastocyst transferred, FBG, TC, TG, LDL HDL, FT3, FT4, TSH levels and PCOS (diagnosed according to"Rotterdam criteria" [ 36 ]). In order to remove the collinearity between confounding factors, a Least Absolute Shrinkage and Selection Operator (LASSO) logistic regression were used to preliminarily select the most valuable features. Based on the selected features, multivariable logistic regression analysis was used to establish the model to obtain the odds ratios (ORs) for ART outcomes. All ORs were reported as OR with 95% confidence interval (CI). Unlike HUA, which is a binary variable differentiated by thresholds, serum UA levels were further treated as original continuous attributes using generalized additive models (GAMs) to visually examine their nonlinear relationships with reproductive outcomes. The red dotted lines are the 95% CI. The implementation of LASSO regression and GAMs is based on open-source Python packages, namely “glmnet_py” and “pygam.” In this study, the sample size was determined using the"events per variable"(EPV) principle, which is the ratio of outcome events to candidate predictors. Insufficient outcome events relative to predictors can lead to biased regression coefficients, overfitting, and poor model performance. The literature suggests an EPV range of 5 to 20 for reliable results, with 10 often cited as the standard for multivariable logistic and Cox regression analyses [ 37 , 38 ]. Therefore, we adopted an EPV of at least 10 to ensure a robust and generalizable multivariable model.

Our findings indicate that elevated serum UA levels are independently associated with decreased LBR and increased risk of miscarriage in the general population undergoing IVF/ICSI, particularly in younger, normal-weight patients and those undergoing programmed FET cycles. While these findings are derived from a retrospective, single-center study and should be interpreted cautiously, they underscore the potential value of incorporating UA monitoring into routine ART assessments. Tailored intervention strategies, especially in patients with poorly controlled UA or undergoing programmed FET cycles, may help mitigate pregnancy loss and improve reproductive outcomes. Future multicenter prospective studies are needed to confirm these observations and clarify the underlying mechanisms.

In this study, we investigated the impact of HUA on reproductive outcomes in patients undergoing IVF/ICSI treatment. Additionally, we stratified patients based on normal weight and overweight categories according to female BMI for further analysis. We focused solely on the first transfer cycle to calculate LBR, CPR, and miscarriage rate, ensuring the independence of each dataset and the objectivity of our results. Our findings indicated that patients with HUA experienced lower LBR and higher miscarriage rates, highlighting HUA as an independent factor affecting reproductive outcomes during ART. The HUA group had a higher incidence of PCOS; however, PCOS was excluded as a confounding variable in our LASSO regression analysis due to its minimal impact on reproductive outcomes. This finding aligns with previous studies, including a recent one in 2023 [ 41 ], which found that PCOS does not significantly affect early and overall late abortion rates after IVF/ICSI fresh cycle pregnancy. These results support the view that HUA, rather than PCOS, is a more relevant predictor of miscarriage in ART patients. We also observed a lower fertilization rate in the HUA group, though embryo quality parameters—including cleavage, blastulation, and high-quality blastocyst rates—were comparable between groups. These results are consistent with previous reports [ 42 ]indicating that elevated serum UA levels do not impact the potency of oocytes and embryo quality. While UA levels had minimal impact on clinical pregnancy rates [ 42 ], HUA significantly increased miscarriage risk, thereby reducing LBR. Mechanistically, HUA as a high-risk factor for endometrial hyperplasia, may disrupt endometrial homeostasis, cause vascular damage [ 43 ], interfere with decidualization [ 26 ] and promote sterile inflammation [ 44 ]. In parallel, UA’s pro-oxidative properties can exacerbate oxidative stress [ 45 ], which is particularly harmful during early pregnancy due to the placenta’s limited antioxidant capacity. Additionally, UA may induce stress indirectly via complement activation [ 46 ] or insulin resistance [ 47 ], both of which can lead to miscarriage [ 48 , 49 ]. However, the specific mechanism by which HUA affects ART outcomes require further study. Although HUA has been linked to spontaneous pregnancy complications — such as preeclampsia [ 50 ], GDM [ 51 ], premature rupture of membranes (PROM) [ 52 ], PB [ 53 ] and LBW [ 54 ] — data on its association with core ART outcomes (pregnancy, live birth and miscarriage) remain limited. One population-based study linked higher serum UA levels with increased infertility risk in U.S. women, even after adjusting for confounders [ 55 ]. Additionally, a strong association between insulin resistance and elevated UA levels has been noted in patients with recurrent pregnancy loss (RPL) [ 56 ]. These findings, together with our data, suggest that HUA is broadly detrimental to reproductive success, in both natural and ART conceptions. Programmed FET cycles were associated with a higher occurrence of placental disorders and obstetric and perinatal complications [ 57 – 59 ]. While the mechanisms remain unclear, recent studies suggest that the absence of a corpus luteum (CL) may play a critical role [ 60 – 62 ]. In contrast to fresh cycles where a CL is always present, programmed FET cycles are CL-deficient and depend on exogenous hormone replacement for endometrial preparation and luteal support. This absence of a CL can disrupt early pregnancy vascular adaptation, potentially causing delayed or attenuated cardiovascular responses in programmed FET cycles compared with fresh-embryo transfer cycles [ 62 , 63 ]. The CL plays a crucial role in endometrial decidualization and implantation preparation, with studies showing that programmed FET cycles alter key signaling pathways (estrogen receptor, VEGF, integrins), negatively affecting endometrial receptivity [ 64 ]. In addition to the absence of a CL, suboptimal steroid hormone administration in programmed FET cycles may impair trophoblast invasion and placental function, while prematurely elevated estradiol levels may inhibit spiral artery remodeling, increasing the risk of hypertensive disorders [ 65 ]. In our study, programmed FET cycles accounted for 96.8% (818/845) of all FET cycles, and we suspect that the effect of HUA on live birth in the frozen subgroup may be amplified by the negative impact of programmed FET cycles on the intrauterine environment, resulting in a lower probability of live birth in patients exposed to HUA. Our logistic regression analysis revealed that reproductive outcomes among women aged ≥ 35 years were not significantly influenced by HUA. This finding may be due to the small sample size or the overriding effect of aging on reproductive outcomes, even after controlling for known confounders. Even among women without HUA, those aged ≥ 35 had a substantially higher miscarriage rate than younger counterparts (31.2% vs 13.2%; Table  4 ). Aging affects fertility through various mechanisms and further exploration of reproductive outcomes in older women with HUA is warranted on a larger scale. Interestingly, HUA had a greater negative impact on live birth and miscarriage in the normal-weight group than in the overweight group. This may reflect both the difficulty of isolating UA’s effect in obese women—where baseline reproductive risk is already elevated—and the smaller sample size in the overweight subgroup, which may have limited statistical power. Obesity impairs fertility through multiple mechanisms, including altered adipokine secretion [ 66 ], oxidative stress and inflammation [ 67 , 68 ], hormonal imbalances, and impaired ovarian steroidogenesis [ 69 ]. These factors, together with mitochondrial dysfunction and altered endometrial receptivity [ 70 – 72 ], can compromise embryo quality and implantation, contributing to higher miscarriage rates [ 73 , 74 ]. Obese patients often require higher gonadotropin doses and tend to have lower cumulative LBRs [ 75 , 76 ]. While HUA is one factor in this complex landscape, its independent contribution is more detectable in leaner patients, where confounding risks are fewer. In this investigation, we found that serum UA levels in women undergoing assisted ART within 6 months prior to embryo transfer significantly impact reproductive outcomes. Given the rising prevalence of HUA, monitoring serum UA levels will become an essential part of ART protocols. Serum UA levels should be evaluated at the initial fertility consultation to establish a baseline. For those with normal UA levels, it is recommended to continue UA monitoring every 2–3 months during the 6-month period before embryo transfer to ensure levels remain stable [ 77 ]. For patients with elevated UA levels or risk factors for HUA, such as obesity or metabolic syndrome, monthly monitoring is advisable [ 77 ]. Given the broader systemic effects of HUA, additional related tests should be incorporated into ART protocols to optimize reproductive outcomes including kidney function tests, insulin and glucose metabolism, lipid profile, thyroid function tests and inflammation markers. For patients already diagnosed with HUA, lifestyle modifications such as weight loss and dietary changes are essential. It is advisable to reduce the consumption of red meat, seafood, shellfish, fructose, sugar-sweetened beverages, and alcoholic drinks (especially beer), as these can elevate the risk of HUA [ 78 ]. Conversely, low-fat dairy products and vitamin C have protective effects [ 79 ]. Caution should be exercised with medications, as the teratogenic potential of uric acid-lowering drugs like allopurinol remains unclear [ 80 ]. Patients with a history of hypertension, diabetes, chronic kidney disease, or a family history of HUA should be closely monitored and consult specialists to determine whether serum UA management is necessary prior to ART. In summary, a comprehensive, multi-faceted approach to managing HUA, including regular UA monitoring and broader metabolic and renal assessments, is essential for optimizing ART outcomes. This is the first study to examine the association between HUA and core reproductive outcomes—specifically miscarriage and live birth—in a general ART population. Prior studies focused on non-PCOS groups and only assessed clinical pregnancy, overlooking pregnancy maintenance. By using UA levels measured within six months before embryo transfer, we reduced temporal bias. Subgroup analyses by age, BMI, and transfer type further clarified the differential impact of HUA. We found that HUA did not impair embryo quality or the establishment of pregnancy, but significantly increased the risk of miscarriage, leading to lower live birth rates. This effect remained significant after adjusting for confounders, suggesting that elevated UA levels—or the metabolic state they reflect—compromise pregnancy stability. Clinically, this implies that even when biochemical or ultrasound-confirmed pregnancies are achieved, HUA patients are at higher risk of pregnancy loss. Importantly, HUA emerged as an independent risk factor particularly in younger, non-obese women—populations typically expected to have favorable ART outcomes. These findings challenge the assumption that younger or lean patients are metabolically low-risk, and indicate that HUA may serve as a warning sign even in the absence of overt risk factors like obesity. In overweight women, miscarriage risk is already elevated, making the effect of HUA more difficult to isolate. In contrast, in normal-BMI women with few confounders, elevated UA appears to signal underlying metabolic or inflammatory dysregulation that undermines pregnancy maintenance. This highlights the clinical importance of monitoring UA levels beyond traditional demographic indicators such as age and BMI. Patients with high UA may benefit from closer surveillance in early pregnancy, including earlier and more frequent ultrasounds, hormone monitoring, and optimized luteal support. In cases of poorly controlled UA, programmed FET cycles—already associated with reduced physiological endometrial preparation—should be considered cautiously. Incorporating UA screening into ART protocols and tailoring management for HUA patients may help reduce miscarriage risk and improve live birth outcomes. Our study has some limitations. First, this was a retrospective study spanning approximately 7 years during which time clinical practice, biochemical measurements, and IVF or ICSI treatments might have changed, and various potential confounding factors could have affected the results. Although we tried to eliminate the influence of confounding factors as much as possible, some known or unknown confounding factors may still affect the statistical results. Second, the exclusion of some diseases was only based on clinical database records, which may lead to inconsistency with their respective diagnostic criteria. Third, being a single-center study, it requires validation through larger multicenter studies. Fourth, the small sample size in the overweight and older groups may limit the generalizability of our conclusions. Finally, serum UA levels were collected within 6 months prior to embryo transfer. While this time window allowed for sufficient sample size and broad patient inclusion, it may not fully reflect short-term metabolic fluctuations close to transfer. In real-world ART practice, UA is not routinely assessed at standardized time points, and narrower windows (e.g., within 1–3 months) would have resulted in substantial data loss. Nonetheless, variability in testing intervals and the inherent fluctuations in UA may have introduced exposure misclassification. Future prospective studies with standardized UA assessments are needed to address this limitation.

Uric acid (UA), a by-product of purine metabolism, is primarily synthesized endogenously by the liver, intestines, and other tissues. Elevated production or reduced excretion of UA can result in higher-than-normal serum UA levels, leading to hyperuricemia (HUA) [ 1 ]. The prevalence of HUA has been increasing in the recent years, including asymptomatic HUA [ 2 ]. In China, the prevalence of HUA was reported to be 14% from 2018 to 2019, with men and women experiencing rates of 24.4% and 3.6%, respectively [ 3 ]. Similarly, 20% of adults in the United States are affected by HUA [ 4 ]. HUA triggers inflammation and oxidative stress [ 5 – 9 ], which are closely linked to various diseases [ 10 – 13 ], such as chronic kidney disease (CKD) [ 14 ], cardiovascular disease [ 15 ], gout [ 16 ], hypertension [ 17 ], metabolic syndrome (MetS) [ 18 ], and diabetes [ 19 ]. Emerging research has expanded the understanding of HUA beyond being solely a metabolic disorder, recognizing its involvement in inflammatory, immune disorders, and hemodynamic disturbance [ 20 ]. Moreover, it has been implicated in diseases of the female reproductive system. The prevalence of HUA in women with polycystic ovary syndrome (PCOS) (25.48%) is significantly higher than that in women without PCOS (8.74%) [ 21 ], and serum UA levels have been correlated with pregnancy complications and adverse fetal outcomes [ 22 ]. HUA can increase the risk of gestational hypertension [ 23 ] and gestational diabetes mellitus (GDM) [ 24 ]. An inverse correlation between UA levels and birth weight was found, with HUA being associated with an increased risk of low birth weight (LBW) and small for gestational age (SGA), particularly in the third trimester [ 25 ]. Studies have demonstrated a close association between UA and endometrial decidualization [ 26 ], with HUA identified as a high-risk factor for endometrial hyperplasia [ 27 ]. Additionally, it serves as a predictor for maternal and fetal complications [ 28 , 29 ], with monitoring of amniotic fluid UA considered valuable in predicting fetal growth [ 30 ]. There are various indications of the relationship between serum UA and maternal and fetal aspects. While studies have shown that HUA in women with PCOS is associated with significantly lower probabilities of live birth and clinical pregnancy and a higher risk for low birthweight [ 31 ] in their first fresh-embryo transfer cycles, evidence regarding the relationship between serum UA and reproductive outcomes in the general population of women undergoing assisted reproductive technology (ART) remains scarce. Hence, the objective of our current study was to investigate the association between serum UA levels and reproductive outcomes among women undergoing in vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI) treatments, providing a novel perspective on serum UA levels in predicting reproductive risk.

Supplementary Material 1. Supplementary Material 1.