孕晚期糖脂代谢紊乱是大于胎龄儿的独立危险因素：基于LASSO回归的风险预测模型

Third-trimester glycolipid metabolic disorders as independent risk factors for large for gestational age infants: a LASSO-based risk prediction model

摘要

目的 构建基于LASSO回归的大于胎龄儿(large for gestational age, LGA)风险因素预测模型,尤其关注孕晚期糖脂代谢指标的影响。方法 基于前瞻性母婴队列研究西安纵向母子队列,收集西北妇女儿童医院2016—2019年孕产妇资料,基于LASSO回归优化模型的特征选择,通过多因素Logistic回归构建预测模型并绘制列线图,借助受试者工作特征(ROC)曲线、校准曲线、决策分析曲线判定预测模型区分度、校准度和临床应用价值,并使用Bootstrap法进行内部验证。结果 共纳入15 304对母子。多因素Logistic回归分析显示,LGA的独立危险因素为经产妇(OR=1.36,95% CI:1.24～1.50)、孕前超重(OR=1.31,95% CI:1.19～1.45)、孕前肥胖(OR=1.83,95% CI:1.53～2.19)、有既往不良孕产史(OR=1.19,95% CI:1.09～1.29)、孕期增重过多(OR=1.92,95% CI:1.76～2.10)、妊娠期贫血(OR=1.18,95% CI:1.08～1.29)、孕中期空腹血糖≥5.1 mmol/L(OR=1.29,95% CI:1.14～1.47)、孕晚期糖化血红蛋白(HbA1c)＞5.3%(5.3%～5.7%:OR=1.21,95% CI:1.10～1.34;＞5.7%:OR=1.78,95% CI:1.58～2.00)、孕晚期总胆固醇＞5.74 mmol/L(5.74～6.16 mmol/L:OR=1.40,95% CI:1.17～1.67;＞6.16 mmol/L:OR=1.39,95% CI:1.13～1.71)、孕晚期甘油三酯＞2.47 mmol/L(2.47～3.14 mmol/L:OR=1.23,95% CI:1.09～1.38;3.14～3.43 mmol/L:OR=1.41,95% CI:1.22～1.63;＞3.43 mmol/L:OR=1.58,95% CI:1.39～1.80)。模型ROC曲线下面积为0.697(95% CI:0.687～0.706),内部验证模型显示优异性能。结论 所构建的LGA风险预测模型可靠且具临床适用性;孕晚期糖脂代谢指标(HbA1c、总胆固醇及甘油三酯)是LGA的独立预测因子,提示需全孕期关注糖脂代谢以降低LGA风险。

Abstract

Objective To develop a prenatal risk prediction model for large for gestational age (LGA) infants using least absolute shrinkage and selection operator (LASSO) regression, with a focus on third-trimester glycometabolic indicators. Methods Based on the prospective mother-infant cohort study, the Xi'an Longitudinal Mother-child Cohort (XAMC), data of mother-infant dyads were collected from Northwest Women's and Children's Hospital between 2016 and 2019. Feature selection for the model was optimized using the LASSO regression. A predictive model was then constructed using multivariate Logistic regression based on the selected features, and a nomogram graph was plotted. The discriminative ability, calibration, and clinical utility of the predictive model were evaluated using the receiver operating characteristic (ROC) curve, calibration curve, and decision curve analysis, respectively. Internal validation was made using the Bootstrap method. Results A total of 15 304 mother-infant pairs were included in this study. The results of multivariate Logistic regression showed that the independent risk factors for LGA were as follows: multiparity (OR=1.36, 95% CI:1.24-1.50), pre-pregnancy overweight (OR=1.31, 95% CI:1.19-1.45), pre-pregnancy obesity (OR=1.83, 95% CI:1.53-2.19), a history of previous adverse maternal outcomes (OR=1.19, 95% CI:1.09-1.29), excessive gestational weight gain (OR=1.92, 95% CI:1.76-2.10), gestational anemia (OR=1.18, 95% CI:1.08-1.29), second-trimester fasting plasma glucose (FPG) ≥5.1 mmol/L (OR=1.29, 95% CI:1.14-1.47), third-trimester glycated hemoglobin A1c (HbA1c) ＞5.3% (5.3%-5.7%: OR=1.21, 95% CI:1.10-1.34; ＞5.7%: OR=1.78, 95% CI:1.58-2.00), third-trimester total cholesterol ＞5.74 mmol/L (5.74-6.16 mmol/L: OR=1.40, 95% CI:1.17-1.67; ＞6.16 mmol/L: OR=1.39, 95% CI:1.13-1.71), and third-trimester triglycerides ＞2.47 mmol/L (2.47-3.14 mmol/L: OR=1.23, 95% CI:1.09-1.38; 3.14-3.43 mmol/L: OR=1.41, 95% CI:1.22-1.63; ＞3.43 mmol/L: OR=1.58, 95% CI:1.39-1.80). The area under the ROC curve of the model was 0.697 (95% CI: 0.687-0.706), with excellent performance in internal validation. Conclusion The model developed in this study for prenatal risk prediction model of LGA exhibited reliability and clinical applicability. The third-trimester glycemic and lipid indicators (HbA1c, total cholesterol, triglycerides) were identified as predictors of LGA. These findings emphasize the importance of monitoring glycemic and lipid metabolism throughout the entire pregnancy to reduce the risk of LGA.

关键词

大于胎龄儿(LGA) / LASSO回归 / 糖脂代谢 / 孕晚期 / 预测模型

Key words

large for gestational age (LGA) / LASSO regression / glycolipid metabolism / third trimester / prediction model

参考文献

[1] SIMEONI U, BARKER D J. Offspring of diabetic pregnancy: long—term outcomes[J]. Semin Fetal Neonatal Med, 2009, 14(2): 119-124.

[2] BROWN J, ALWAN N A, WEST J, et al. Lifestyle interventions for the treatment of women with gestational diabetes[J]. Cochrane Database Syst Rev, 2017, 5(5): Cd011970.

[3] DE BRUN M, MAGNUSON A, MONTGOMERY S, et al. Changing diagnostic criteria for gestational diabetes (CDC4G) in Sweden: a stepped wedge cluster randomised trial[J]. PLoS Med, 2024, 21(7): e1004420.

[4] 苏莉, 王琪 . 我国妇女孕期体重增加的变化及其对母儿的影响[J]. 中国妇幼保健, 2002, 17(8): 503-504.

[5] SU L, WANG Q. Changes in weight gain among pregnant women in China and their impacts on maternal and neonatal health[J]. Matern Child Health Care China, 2002, 17(8): 503-504.

[6] CATALANO P M. Obesity, insulin resistance, and pregnancy outcome[J]. Reproduction, 2010, 140(3): 365-371.

[7] MCLACHLAN K A, O'NEAL D, JENKINS A, et al. Do adiponectin, TNFalpha, leptin and CRP relate to insulin resistance in pregnancy? Studies in women with and without gestational diabetes, during and after pregnancy[J]. Diabetes Metab Res Rev, 2006, 22(2): 131-138.

[8] SONAGRA A D, BIRADAR S M, DATTATREYA K, et al. Normal pregnancy—a state of insulin resistance[J]. J Clin Diagn Res, 2014, 8(11): CC01-CC03.

[9] Obstetrics Subgroup, Chinese Society of Obstetrics and Gynecology, Chinese Medical Association, et al. Guideline of diagnosis and treatment of hyperglycemia in pregnancy (2022) (Part one)[J]. Chin J Obstet Gynecol, 2022, 57(1): 3-12.

[10] CHANDRA M, PARAY A A. Natural physiological changes during pregnancy[J]. Yale J Biol Med, 2024, 97(1): 85-92.

[11] LIBERIS A, PETOUSIS S, TSIKOURAS P. Lipid disorders in pregnancy[J]. Curr Pharm Des, 2021, 27(36): 3804-3807.

[12] JI J, HE Z, QU P, et al. The Xi'an longitudinal mother—child cohort study: design, study population and methods[J]. Eur J Epidemiol, 2021, 36(2): 223-232.

[13] HE L R, YU L, GUO Y. Birth weight and large for gestational age trends in offspring of pregnant women with gestational diabetes mellitus in southern China, 2012—2021[J]. Front Endocrinol (Lausanne), 2023, 14: 1166533.

[14] JUAN J, YANG H X. Prevention, and lifestyle intervention of gestational diabetes mellitus in China[J]. Int J Environ Res Public Health, 2020, 17(24): 9517.

[15] 高永祥, 张晋昕 . Logistic回归分析的样本量确定[J]. 循证医学, 2018, 18(2): 122-124.

[16] GAO Y X, ZHANG J X. Determination of sample size in Logistic regression analysis[J]. The Journal of Evidence—Based Medicine, 2018, 18(2): 122-124.

[17] ZONG X L, LI H, ZHANG Y Q, et al. Reference standards and growth curves for weight/length ratio, body mass index, and ponderal index of Chinese newborns with different gestational ages[J]. Chin J Pediatr, 2021, 59(3): 181-188.

[18] 中国营养学会 . 中国妇女妊娠期体重监测与评价: T/CNSS 009—2021[S].2021.

[19] Chinese Nutrition Society. Weight monitoring and evaluation during pregnancy period of Chinese women: T/CNSS 009—2021[S].2021.

[20] 中国高血压防治指南起草委员会, 高血压联盟, 中国医疗保健国际交流促进会高血压病学分会, 等 . 中国高血压防治指南(2024年修订版)[J]. 中国卒中杂志(中英文), 2006(8): 575-582.

[21] Chinese Hypertension Prevention and Treatment Guidelines Drafting Committee, Hypertension Alliance, Hypertension Branch of China Association for the Promotion of International Exchange in Healthcare. Chinese guidelines for the management of hypertension[J]. Chin J Stroke, 2006(8): 575-582.

[22] METZGER B E, GABBE S G, PERSSON B, et al. International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy[J]. Diabetes Care, 2010, 33(3): 676-682.

[23] LI N, LIU E, GUO J, et al. Maternal prepregnancy body mass index and gestational weight gain on pregnancy outcomes[J]. PLoS One, 2013, 8(12): e82310.

[24] JENSEN D M, DAMM P, SØRENSEN B, et al. Pregnancy outcome and prepregnancy body mass index in 2459 glucose—tolerant Danish women[J]. Am J Obstet Gynecol, 2003, 189(1): 239-244.

[25] RASMUSSEN K M, YAKTINE A L. Weight gain during pregnancy: reexamining the guidelines[M]. Washington(DC):National Academies Press (US), 2009: 2009.

[26] CATALANO P M, SHANKAR K. Obesity and pregnancy: mechanisms of short term and long term adverse consequences for mother and child[J]. BMJ, 2017: 356.

[27] FONSECA A, LOPES J, CLODE N. Glucose intolerance in the third trimester is not predictive of adverse outcomes[J]. Int J Gynaecol Obstet, 2019, 147(1): 108-114.

[28] SACKS D A, HADDEN D R, MARESH M, et al. Frequency of gestational diabetes mellitus at collaborating centers based on IADPSG consensus panel—recommended criteria: the hyperglycemia and adverse pregnancy outcome (HAPO) study[J]. Diabetes Care, 2012, 35(3): 526-528.

[29] RYAN E A, ENNS L. Role of gestational hormones in the induction of insulin resistance[J]. J Clin Endocrinol Metab, 1988, 67(2): 341-347.

[30] YAMAMOTO J M, KELLETT J E, BALSELLS M, et al. Gestational diabetes mellitus and diet: a systematic review and meta—analysis of randomized controlled trials examining the impact of modified dietary interventions on maternal glucose control and neonatal birth weight[J]. Diabetes Care, 2018, 41(7): 1346-1361.

[31] METZGER B E, LOWE L P, DYER A R, et al. Hyperglycemia and adverse pregnancy outcomes[J]. N Engl J Med, 2008, 358(19): 1991-2002.

[32] DE B M, MAGNUSON A, MONTGOMERY S, et al. Changing diagnostic criteria for gestational diabetes (CDC4G) in Sweden: a stepped wedge cluster randomised trial[J]. PLoS Med, 2024, 21(7): e1004420.

基金资助

陕西省重点研发计划一般项目(2021SF-009)