This article has Open Peer Review reports available.
Retinoids, anxiety and peripartum depressive symptoms among Chinese women: a prospective cohort study
© The Author(s). 2017
Received: 8 November 2016
Accepted: 27 June 2017
Published: 1 August 2017
The current study aimed to investigate whether serum RBP levels can be a key predictor of peripartum depression (PPD).
This was a prospective cohort study, conducted at a general teaching hospital in South China. Research participants were evaluated at three time points: the third trimester of pregnancy (T1), after delivery at week one (T2), and after delivery week six (T3) using a set of self-reported questionnaires and blood sample assays.
A total of 156 subjects were included for data analysis. The prevalence of anxiety symptoms ranged from 32.69% to 36.53%. The prevalence of PPD was also high and ranged from 27.56% to 35.89%. In the third trimester, significant predictors of depressive symptoms include serum retinol-binding protein (RBP) concentrations and estradiol levels (P = 0.008 and 0.033, respectively). At one week after delivery, serum concentrations of RBP at T2 were still significant predictors of depressive symptoms (P = 0.020, and serum estradiol concentrations at T1 were a significant predictor (P = 0.010). The most stable predictor of depressive symptoms at T3 was anxiety symptoms, especially at T3 time point (P < 0.001). Serum RBP concentrations at T1 and T2 were still significant predictors of depressive symptoms at T3.
A high prevalence of anxiety and depressive symptoms tended to persist in Chinese women during the peripartum period. This study, which found the potential contribution of RBP to the occurrence of PPD, requires that large sample studies be conducted in future with a longer-term follow-up period, in order to confirm its results.
Depression is the leading cause of disease-related disability in women globally . Peripartum anxiety and depression are the most common complications of childbearing, and are associated with substantial adverse effects on pregnancy outcomes, such as limiting a woman’s ability to function effectively in the maternal role [2–4]; as well as with adverse effects on the developing child in terms of preterm delivery, lower birth weight and later cognitive, affective and behavioral disturbances [4, 5]. Consequences of depression and anxiety are often more severe in the peripartum period than during other life periods . A number of studies suggest that prevalence rates of anxiety symptoms range from 25% to 45%, and are increasingly common during the peripartum period [4, 7]. Prevalence of peripartum depression (PPD) is up to 25% around the world . In China, the prevalence of PPD varies between 28.5% and 38.8% [9, 10]. Given the high prevalence and severe consequences of PPD, there is a need for early detection and prevention strategies for women with PPD.
Retinoids, Vitamin A and its derivatives, have been linked to neuropsychiatric symptoms including depression . Excess accumulation of Vitamin A has been reported to induce depression . Mawson and Wang  proposed a theory of retinoid toxicity hypothesis of PPD: Vitamin A accumulates in breast, brain and liver to potentially toxic concentrations in the third trimester. Increased levels of Vitamin A may impair the mobilization and secretion of retinol-binding protein (RBP), which may cause PPD by increasing the levels of circulating retinoic acids and retinyl esters . Serum RBP estimation highly correlates with retinol concentration to predict Vitamin A status . Hence, measuring retinoids by serum RBP concentrations may be a sensitive and specific predictor or biomarker of PPD in women. Therefore, this study aimed to investigate whether serum RBP levels can be a putative biomarker of PPD. This study hypothesized that serum RBP concentrations might be a key predictor of PPD, and could be used as a putative biomarker for the early detection and diagnosis of PPD.
Study design and procedure
This was a prospective cohort study, conducted at a general teaching hospital in South China. Research participants were evaluated at three time points: the third trimester of pregnancy (T1), after delivery at week one (T2), and after delivery week six (T3) using a set of self-reported questionnaires and blood sample assays. This study was approved by the Ethical Review Committee of the Third Affiliated Hospital of Guangzhou Medical University. Prior to subject recruitment and data collection, the researcher provided a full explanation of the study purpose and procedure. All research participants, who voluntarily joined this study, gave written informed consent. This study adhered to STROBE methodology.
Women were recruited at an outpatient obstetric clinic from January to August 2016. Inclusion criteria were pregnant women ages 18 years or older, who were planning to give birth and attend postpartum follow up at the studied hospital, with normal singleton pregnancies. Exclusion criteria included women with severe mental health problems, such as psychiatric disease, or physical illnesses requiring medical treatment or admission to hospital.
Demographic and clinical variables included age, marital status, highest level of education, monthly income, parity (first birth or subsequent), delivery week, delivery mode (vaginal birth or Caesarian section), and mode of infant feeding (exclusive breastfeeding rates). Self-report measures were collected to characterize participants’ sleep, anxiety and depressive symptoms. Previous research found that women with symptoms such as anxiety and sleep disturbance were significantly correlated with PPD , so this study also included sleep quality and anxiety symptom measures.
The Pittsburgh Sleep Quality Index (PSQI) was utilized to evaluate subjects’ sleep quality. The PSQI consists of 18 items composed of seven factors; each factor is scored by 0–3 points and the total score ranges from 0 to 21 points; a higher score indicates worse sleep quality . The Chinese version of PSQI has been validated in a Chinese population, and has established good reliability and validity [16, 17]. Anxiety symptoms were assessed using the Self-rating Anxiety Scale (SAS), originally developed by Zung ; SAS is a 20-item Likert-style (4-point) rating scale for anxiety, with a theoretical score range extending from 20 to 80. The original total score plus 1.25 will be the “Anxiety Index” score (<50 as normal; ≥50 as anxiety cases) . The Chinese version of SAS is also widely used among Chinese women during pregnancy . In this study, the internal consistency by Cronbach’s alpha of SAS was 0.89. The Edinburgh Postnatal Depression Scale (EPDS) is a widely used depression assessment scale for peripartum women globally . It has 10 items, with each item divided into 0–3 points; the total score ranges from 0 to 30 points . The Chinese version of the EPDS has been validated in a sample of Chinese women in the perinatal period. The cut-off score of 9/10 can show sensitivity with 82% and specificity with 86% at six weeks postpartum . Hence, subjects in this study with an EPDS score ≥ 10 were taken as “peripartum depression” cases. Studies have shown the scale has good reliability and validity in populations in mainland China [23, 24].
Blood samples for serum RBP assays were obtained in late pregnancy (30–34 weeks of gestation), and at one week and six weeks after delivery. Blood draw was performed in the morning within two hours (8 a.m. to 10 a.m.). The serum levels of RBP were assessed by human ELISA (Enzyme-Linked Immunosorbent Assay) kits for research use only (Abcam Trading Company Ltd., Guangzhou, China). Serum was frozen until assayed. The intra- and inter-assay coefficient of variation averaged 4.5% and 7.2%, respectively. The normal reference value of RBP used was 25–80 mg/L. During the peripartum period, the tremendous changes in gonadal hormones, such as estrogen and progesterone, could be triggers for depression . Hence, laboratory determinations of serum concentrations of estradiol and progesterone were also carried out by commercially available kits using RIA (radioimmunoassay). The coefficients for intra-assay and inter-assay variation were 4.3% and 6.8% for estradiol and 5.1% and 8.8% for progesterone . For both hormones, all samples were analyzed in a single batch.
Statistical analyses were performed using SPSS Version 21.0 and AMOS 21.0. Descriptive analysis of continuous and categorical variables was performed. Continuous variables were expressed as means ± SD, and categorical variables were expressed as proportions (%) by SPSS 21.0. Correlation analysis was used to identify potential predictors for EPDS score at each time assessment. Those variables with p values less than 0.05 were entered into the regression models. In order to control any potential collinear relation of variables, SEM (Structural Equation Modeling) regression analysis was used to identify predictors of depressive symptoms in women in the peripartum period by AMOS 21.0. A p-value of 0.05 was considered to be statistically significant.
Of 200 potential eligible patients approached, 181 agreed to join this study. Among those who declined to join this study, 14 were too busy, and five were planning to give birth in a different hospital, and so would be unable to take part in the follow-up visits. Among the 181 subjects, only 156 completed two follow-up assessments after delivery at week one and six. Hence, a total of 156 subjects were included for data analysis.
Descriptive statistics of variables used in the analysis (N = 156)
26.65 (2.70) (22 to 35)
High school or below
Undergraduate or high
Monthly income (RMB)
Weeks of delivery
38.99 (1.31) (36 to 41)
Exclusive breastfeeding (T2)
Exclusive breastfeeding (T3)
RBP level (mg/L)
Estradiol level (pmol/L)
Progesterone level (nmol/L)
Mean PSQI score
≥ 50 at T1
≥ 50 at T2
≥ 50 at T3
≥ 10 at T1
≥ 10 at T2
≥ 10 at T3
A summary of descriptive statistics of psychosocial variables, RBP and reproductive hormones is also shown in Table 1. Serum RBP concentrations were lowered at T3, serum estradiol and progesterone concentrations dropped sharply from T1 to T3. Sleep quality remained relatively stable. The prevalence of anxiety symptoms was relatively high, ranging from 32.69% to 36.53%. The prevalence of depressive symptoms reached the highest level at T3 (35.89%), followed by the period of late pregnancy (30.76%).
Regression coefficients of predictors for EPDS score at T1
T1_EPDS score ← Education
T1_EPDS score ← RBP_T1
T1_EPDS score ← Estradiol_T1
T1_EPDS score ← Progesterone_T1
T1_EPDS score ← PSQI_T1
T1_EPDS score ← SAS_T1
Regression coefficients of predictors for EPDS score at T2
T2_EPDS score ← Education
T2_EPDS score ← Income
T2_EPDS score ← RBP_T1
T2_EPDS score ← Estradiol_T1
T2_EPDS score ← Progesterone_T1
T2_EPDS score ← PSQI_T1
T2_EPDS score ← SAS_T1
T2_EPDS score ← RBP_T2
T2_EPDS score ← Estradiol_T2
T2_EPDS score ← Progesterone_T2
T2_EPDS score ← PSQI_T2
T2_EPDS score ← SAS_T2
Regression coefficients of predictors for EPDS score at T3
T3_EPDS score ← Income
T3_EPDS score ← Exclusive breastfeeding
T3_EPDS score ← RBP_T1
T3_EPDS score ← Estradiol_T1
T3_EPDS score ← Progesterone_T1
T3_EPDS score ← PSQI_T1
T3_EPDS score ← SAS_T1
T3_EPDS score ← RBP_T2
T3_EPDS score ← Estradiol_T2
T3_EPDS score ← Progesterone_T2
T3_EPDS score ← PSQI_T2
T3_EPDS score ← SAS_T2
T3_EPDS score ← RBP_T3
T3_EPDS score ← Estradiol_T3
T3_EPDS score ← Progesterone_T3
T3_EPDS score ← PSQI_T3
T3_EPDS score ← SAS_T3
The prevalence of PPD ranged from 27.56% to 35.89% in this study sample. Anxiety symptoms were also very common, with more than one-third of women reporting anxiety symptoms during the peripartum period. Symptoms of anxiety and depression tended to persist during the peripartum period. Compared with one in seven women reporting PPD in the United States , Chinese women in this study experienced a far higher prevalence of PPD, and urgently need more intervention strategies to treat and control depressive symptoms. This study also found a high prevalence of anxiety, so Chinese women in the peripartum period should be screened for these two common symptoms in a routine clinical setting. A routine screening strategy may help childbearing women prevent PPD and achieve better mental health and pregnancy outcomes. Consistent with previous research , there were no socio-demographics or obstetric variables that significantly predicted PPD.
This study partly supported the theory of retinoid toxicity hypothesis of PPD proposed by Mawson and Wang . Serum RBP levels were significant predictors of PPD during periods in late pregnancy and the first week after delivery, but at six weeks after delivery, there was a non-significant predicting relationship between RBP levels and EPDS scores. Mawson and Wang  proposed that vitamin A accumulates in breast, brain and liver to potentially toxic concentrations in the third trimester, such that increased liver concentrations of vitamin A impair the mobilization and secretion of RBP, which may cause PPD by increasing the levels of circulating retinoic acids and retinyl esters. This study found that serum RBP estimation correlates highly with retinol concentration, so that measuring retinoids by serum RBP concentrations could be a sensitive and specific predictor or biomarker of PPD in women. Consistent with Mawson and Wang’s hypothesis, RBP was strongly associated with PPD symptoms. This study did not, however, assess retinyl esters or retinoic acid. Thus, future work should investigate the possible role of other vitamin A metabolites in PPD.
Within this retinoid toxicity theory of PPD, Mawson and Wang  proposed that breastfeeding could reduce body stores of Vitamin A for women and prevent the occurrence of PPD. Although previous research evidence also indicates a strong association between breastfeeding and PPD , the present study did not find a significant predicting relationship between exclusive breastfeeding and EPDS score. One possible explanation may be due to the lower rates of exclusive breastfeeding and short-term follow-up assessment after six weeks of delivery. Therefore, whether breastfeeding could prevent PPD by reducing maternal body stores of Vitamin A needs further investigation.
This study found that the role of gonadal hormones, such as estradiol at T1, were significant predictors of PPD at T1 and T2. As women within one week after delivery, it is a time when their gonadal hormones fluctuate dramatically, and result in a significant predicting relationship between estradiol and the occurrence of PPD . In addition, this study’s findings were consistent with previous empirical research [6, 29], with anxiety levels significantly predicting PPD at three assessment points. Hence, relevant intervention programs should be developed in order to help women improve their rate of breastfeeding, and reduce anxiety levels.
There were limitations that may have impacted the interpretations of the study findings. Most of the women in this sample were well-educated, and all were married. Hence, this study’s findings exclude the experiences of single mothers and those with lower levels of education. Additionally, hormone levels were only assayed for estradiol and progesterone; other hormones, such as prolactin, luteinizing, and follicle stimulating hormones should also be assayed in future studies, to examine their possible relationship with PPD. A further limitation of the present study is the relatively short-term follow-up period and relatively small sample size. Thus, larger sample studies with longer-term follow-up periods are required, in order to address these limitations.
This study found that a high prevalence of anxiety and depressive symptoms tended to persist in Chinese women during the peripartum period. This study, which also found the potential contribution of RBP to the occurrence of PPD, requires that large sample studies be conducted in future with a longer-term follow-up period, in order to confirm its results. In addition, there is a need for future work to investigate the possible role of other vitamin A metabolites in PPD, i.e., retinyl esters, the percentage of retinyl esters is higher than 10% of total vitamin A - an indicator of retinoid toxicity - and the concentration of retinoic acid, as per the retinoid toxicity hypothesis of PPD proposed by Mawson and Wang .
Thanks for all Chinese women participated in this study. Thank Miss Yingxue Zhou, a statistician specialized in SEM and path analysis for providing consultation in data analysis.
This study was funded by Guangdong Provincial Department of Science and Technology (2016A0202151171); and Guangzhou Municipal Bureau of Traditional Chinese Medicine (20152A011020).
Availability of data and materials
The data that support the findings of this study are available from Principal Investigator of Yingchun Zeng but restrictions apply to the availability of these data, which were used under obtaining permission of research funding institutes of Guangdong Provincial Department of Science and Technology and Guangzhou Municipal Bureau of Traditional Chinese Medicine. Hence, please first contact the principal investigator by email (email@example.com).
ZYC and LYT designed this study; ZYC acted as principle investigators obtained the below series of funding, ZYC and DC drafted this manuscript. XHA, WSL, and ZJX contributed data collections, and data analysis, all authors approved the content of this manuscript.
Ethics approval and consent to participate
The study was approved by the Ethical Review Committee of the Third Affiliated Hospital of Guangzhou Medical University. All study participants were voluntarily participated in this study and provided written informed consent to participate in this study. The ethical approval letter is available only in Chinese version. If needed, we can send this original letter with English translation for proof.
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- World Health Organization, 2012. The Global Burden of Disease-Depression. http://www.who.int/mediacentre/factsheets/fs369/en/.
- Goodman SH, Rouse MH, Connell AM, Broth MR, Hall CM, Heyward D. Maternal depression and child psychopathology: a meta-analytic review. Clin Child Fam Psychol Rev. 2011;14:1–27.View ArticlePubMedGoogle Scholar
- Sockol LE, Neill Epperson C, Barber JP. Preventing postpartum depression: a meta-analytic reivew. Clin Psychol Rev. 2013;33:1205–17.View ArticlePubMedPubMed CentralGoogle Scholar
- Malak SM, Crowley MJ, Mayes LC, Rutherford HJ. Maternal anxiety and neural responses to infant faces. J Affect Disord. 2015;172:324–30.View ArticlePubMedGoogle Scholar
- Anniverno R, Bramante A, Mencacci C, Durbano F. In: Durbano F, editor. Anxiety disorders in pregnancy and the postpartum period, New insights into anxiety disorders. Rijeka: InTech; 2013.View ArticleGoogle Scholar
- Bjørk MH. Veiby G, a Engelsen B, Gilhus NE. Depression and anxiety during pregnancy and the postpartum period in women with epilepsy: a review of frequency, risks and recommendations for treatment. Seizure. 2015;28:39–45.View ArticleGoogle Scholar
- Clout D, Brown R. Sociodemographic, pregnancy, obstetric, and postnatal predictors of postpartum stress, anxiety and depression in new mothers. J Affect Disord. 2015;188:60–7.View ArticlePubMedGoogle Scholar
- Silveira ML, Ertel KA, Dole N, Chasan-Taber L. The role of body image in prenatal and postpartum depression: a critical review of the literature. Arch Womens Ment Health. 2015;18:409–21.View ArticlePubMedPubMed CentralGoogle Scholar
- Zeng Y, Cui Y, Li J. Prevalence and predictors of antenatal depression among Chinese women in their third trimester: a cross-sectional survey. BMC Psychiatry. 2015;15:66.View ArticlePubMedPubMed CentralGoogle Scholar
- Huang M, Wang H, Cao S, Lin F, Kong B, Huang J, Chen Y. A survey of prevalence and influence factors of peripartum depression in Guangzhou, China. Modern Clin Nurs. 2014;11:14–7. [in Chinese]Google Scholar
- Bremner JD, McCaffery P. The neurobiology of retinoic acid in affective disorders. Prog Neuro-Psychopharmacol Biol Psychiatry. 2008;32:315–31.View ArticleGoogle Scholar
- O'Reilly K, Bailey SJ, Lane MA. Retinoid-mediated regulation of mood: possible cellular mechanisms. Exp Biol Med. 2008;233:251–8.View ArticleGoogle Scholar
- Mawson AR, Wang X. Breastfeeding, retinoids, and postpartum depression: a new theory. J Affect Disord. 2013;150:1129–35.View ArticlePubMedGoogle Scholar
- Mahmood K, Samo AH, Jairamani KL, Ali G, Talib A, Qazmi W. Serum retinol binding protein as an indicator of vitamin a status in cirrhotic patients with night blindness. Saudi J Gastroenterol. 2008;14:7–11.View ArticlePubMedPubMed CentralGoogle Scholar
- Langan R, Goodbred AJ. Identification and Management of Peripartum Depression. Am Fam Physician. 2016;93:852–8.PubMedGoogle Scholar
- Tsai PS, Wang SY, Wang MY, Su CT, Yang TT, Huang CJ, Fang SC. Psychometric evaluation of the Chinese version of the Pittsburgh sleep quality Index (CPSQI) in primary insomnia and control subjects. Qual Life Res. 2005;14:1943–52.View ArticlePubMedGoogle Scholar
- Li G, Kong L, Zhou H, Kang X, Fang Y, Li P. Relationship between prenatal maternal stress and sleep quality in Chinese pregnant women: the mediation effect of resilience. Sleep Med. 2016; In Press.Google Scholar
- Zung WWK. 1971. A rating instrument for anxiety disorders. Psychosomatics. 1971;12:371–9.View ArticlePubMedGoogle Scholar
- Wang XD. Measurement manual of psychological health. Chin J Psychol Health. 1993;160:205. [in Chinese]Google Scholar
- Li Y, Zeng Y, Zhu W, Cui Y, Li J. Path model of antenatal stress and depressive symptoms among Chinese primipara in late pregnancy. BMC Pregnancy Childbirth. 2016;16:180.View ArticlePubMedPubMed CentralGoogle Scholar
- Cox JL, Holden JM. Perinatal mental health: a guide to Edinburgh postnatal depression scale. London: Gaskell; 2003.Google Scholar
- Ngai FW, Chan SWC, Ip WY. The effects of a childbirth psychoeducation program on learned resourcefulness, maternal role competence and perinatal depression: a quasi-experiment. Int J Nurs Studies. 2009;46:1298–306.View ArticleGoogle Scholar
- Liu Y, Guo C, Li L, Zhang L, Di Y. Effects of multi angle cognitive - behavioural nursing intervention on elimination process of postpartum depression and anxiety. J Nurs Admin. 2015;15:899–900.Google Scholar
- He Q, Hu J, Zhou B, Li Y. Changes of sex hormone level in perinatal depression in different perinatal periods. Chin J Perinat Med. 2016;19:336–40. [in Chinese]Google Scholar
- Stahl SM. Natural estrogen as an antidepressant for women. J Clin Psychiatry. 2001;62:404–5.View ArticlePubMedGoogle Scholar
- Klier CM, Muzik M, Dervic K, Mossaheb N, Benesch T, Ulm B, Zeller M. The role of estrogen and progesterone in depression after birth. J Psychiatr Res. 2007;41:273–9.View ArticlePubMedGoogle Scholar
- Dias CC, Figueiredo B. Breastfeeding and depression: a systematic review of the literature. J Affect Disord. 2015;171:142–54.View ArticlePubMedGoogle Scholar
- Brummelte S, Galea LAM. Depression during pregnancy and postpartum: contribution of stress and ovarian hormones. Prog Neuro-Psychopharmacol Biol Psychiatry. 2010;34:766–76.View ArticleGoogle Scholar
- Biaggi A, Conroy S, Pawlby S, Pariante CM. Identifying the women at risk of antenatal anxiety and depression: a systematic review. J Affect Disord. 2016;191:62–77.View ArticlePubMedPubMed CentralGoogle Scholar