Authors: Yuyang Liu, Binghai Liu, Jiaxing You, Li Zhang, Wanting Miao, Feiyu Ma, Chaoyi He, Guocheng Liu, Yong Huang
Categories: Original Research, hypertensive disorders of pregnancy, cesarean section, sleep quality, blood pressure, pregnancy
Source: International Journal of Women's Health
Doi: 10.2147/IJWH.S606769
Authors: Yuyang Liu, Binghai Liu, Jiaxing You, Li Zhang, Wanting Miao, Feiyu Ma, Chaoyi He, Guocheng Liu, Yong Huang
Blood pressure instability is common in the early postoperative period after cesarean delivery among women with hypertensive disorders of pregnancy, while postpartum sleep disturbance is highly prevalent. However, evidence regarding the role of postpartum sleep in early blood pressure regulation remains limited.
To investigate the associations of prenatal and postpartum sleep quality with repeated systolic and diastolic blood pressure levels within 72 h after cesarean delivery in women with hypertensive disorders of pregnancy, and secondarily to explore their associations with categorical postpartum blood pressure control status.
This prospective cohort study enrolled women with hypertensive disorders of pregnancy undergoing cesarean delivery. Prenatal sleep quality was assessed before delivery using the Pittsburgh Sleep Quality Index. Postpartum sleep quality was evaluated repeatedly within 72 h after surgery using the Richards–Campbell Sleep Questionnaire (RCSQ), alongside repeated blood pressure measurements. Multivariable logistic regression and mixed-effects models were applied to examine associations with blood pressure control status and with systolic and diastolic blood pressure, adjusting for relevant clinical and perioperative covariates.
A total of 123 women were included. Prenatal sleep quality was not significantly associated with postpartum blood pressure outcomes. In the primary analysis, abnormal postpartum sleep was independently associated with higher systolic (β = 3.63 mmHg, 95% CI 1.82–5.44) and diastolic blood pressure (β = 1.71 mmHg, 95% CI 0.37–3.04), while higher overall RCSQ scores and individual sleep item scores were consistently associated with lower systolic and diastolic blood pressure levels. In the secondary analysis, postpartum sleep quality was not significantly associated with binary blood pressure control status.
Early postpartum sleep quality was independently associated with continuous blood pressure levels but not with categorical blood pressure control in women with hypertensive disorders of pregnancy after cesarean delivery. These findings support the relevance of sleep assessment in relation to postpartum blood pressure burden, whereas associations with categorical blood pressure control require confirmation in larger studies.
Hypertensive disorders of pregnancy are important causes of maternal morbidity and mortality, can lead to organ dysfunction, preterm birth, and adverse perinatal outcomes, and exert sustained effects on both short-term prognosis and long-term cardiovascular risk.1,2 Cesarean section is frequently used in women with hypertensive disorders of pregnancy, and marked blood pressure fluctuations occur in the early postoperative period, especially within the first 72 h after delivery, making blood pressure control during this window particularly critical.3 This early postpartum period is characterized by postoperative stress responses, pain, fluid redistribution, and recovery from anesthesia, all of which may interact with the underlying endothelial and vascular dysfunction of hypertensive disorders of pregnancy and contribute to blood pressure instability. Sleep disturbances are common among postpartum women, influenced by endocrine changes as well as incisional pain, nursing procedures, newborn care, and the ward environment; in this high-risk population with hypertensive disorders of pregnancy, sleep disturbances and loss of blood pressure control may overlap and increase the risk of early postpartum complications.4–6
Regarding the relationship between sleep and blood pressure, studies in the general population and in patients with hypertension have demonstrated associations of shortened sleep duration and decreased sleep quality with higher blood pressure, whereas pregnancy-related research has mostly focused on prenatal sleep and the risks of gestational hypertension and preeclampsia.7 For the early postpartum period, especially for women with hypertensive disorders of pregnancy after cesarean section, current evidence is relatively limited; prior work has mostly used single sleep assessments and single time-point blood pressure measurements, lacking characterization of dynamic changes in sleep and blood pressure within the 72 h after surgery.8 In this context, postpartum sleep may have particular clinical relevance because it occurs during the period of greatest short-term blood pressure lability and may be more directly related to early postpartum blood pressure burden than prenatal sleep. Prenatal and postpartum sleep have different temporal and pathophysiological meanings, and their relative contributions to postpartum blood pressure are unclear; whether the effects of sleep quality on categorical blood pressure control and on continuous blood pressure levels are consistent has also not been systematically studied.9
These gaps limit clinical decision-making that incorporates sleep as a perinatal blood pressure management target. Based on a prospective cohort, the present study enrolled women with hypertensive disorders of pregnancy undergoing cesarean section, assessed sleep using the prenatal Pittsburgh Sleep Quality Index (PSQI) and repeated postpartum Richards-Campbell Sleep Questionnaire (RCSQ) measurements, and combined these with repeated blood pressure monitoring within the 72 h after surgery to analyze the relationships of prenatal and postpartum sleep and their interaction with postpartum blood pressure control status and the levels of systolic blood pressure and diastolic blood pressure, with the aim of elucidating the role of early postpartum subjective sleep quality in blood pressure management for this high-risk population and exploring whether sleep constitutes an interventional window for postpartum blood pressure management.
This study was a prospective cohort study conducted at Guangdong Women and Children Hospital. The consecutive enrollment period was from May 1, 2024, to January 31, 2025. The participants were pregnant women who had completed routine prenatal examinations in the obstetrics outpatient and inpatient departments of this hospital, had a definitive diagnosis of hypertensive disorders of pregnancy, and terminated pregnancy by cesarean section. The diagnosis of hypertensive disorders of pregnancy was based on the clinical diagnosis confirmed by the attending physician in the medical records, specifically including gestational hypertension, preeclampsia, chronic hypertension with superimposed preeclampsia, and chronic hypertension. Because the number of eclampsia cases was very small, they were not included in the study. All participants underwent cesarean section. The study protocol was reviewed and approved by the hospital Ethics Committee (approval number 202401315), and all subjects signed written informed consent after full information was provided.
During the study period, a total of 158 pregnant women who met the initial conditions were screened. After excluding individual cases according to the prespecified exclusion criteria, 123 cases were finally included in the analysis dataset. The study aimed to evaluate the impact of postpartum sleep disturbance on blood pressure levels and blood pressure control status in women with hypertensive disorders of pregnancy after cesarean section, and to explore the interaction between prenatal and postpartum sleep.
Inclusion 1. age ≥18 years; 2. singleton or multiple pregnancy; 3. clinical diagnosis of any subtype of hypertensive disorders of pregnancy (gestational hypertension, preeclampsia, chronic hypertension with superimposed preeclampsia, chronic hypertension); 4. delivery by cesarean section; 5. able to complete questionnaires and bedside assessments and to sign informed consent. Exclusion 1. lack of repeated measurements of postpartum blood pressure; 2. lack of repeated measurements of postpartum sleep. Meeting any one item resulted in exclusion.
Prenatal sleep was assessed using the Chinese version of the Pittsburgh Sleep Quality Index (PSQI).10 The PSQI covers 7 dimensions and reflects sleep conditions over the past 1 month, with a total score of 0–21, and higher scores indicating worse sleep quality. Research nurses distributed and checked the scale at admission for delivery preparation or on the day of admission to ensure item completeness. According to the commonly used threshold in the literature, a PSQI total score >5 was defined as “abnormal prenatal sleep”, and ≤5 was defined as “normal prenatal sleep”.11 In addition to the binary classification, the PSQI total score was also used as a continuous variable in the model analysis to depict the dose–response relationship.
Postpartum sleep was assessed using the Chinese version of the Richards-Campbell Sleep Questionnaire (RCSQ).12 The questionnaire was administered by trained nurses each morning after the woman was fully awake, based on her sleep experience during the previous night. The RCSQ consists of five core items (items 1–5), each rated on a 0–100 mm visual analog scale, with higher scores indicating better sleep quality. The arithmetic mean of the five items was calculated as the mean RCSQ score. To facilitate clinical interpretation, in accordance with commonly used criteria in previous studies, a mean RCSQ score ≥50 was defined as “normal postpartum sleep quality”, whereas a score <50 was defined as “abnormal postpartum sleep quality”.13 Postpartum sleep quality was analyzed both as a categorical variable (normal vs. abnormal) and as a continuous variable in subsequent statistical analyses. Each morning assessment of postpartum sleep referred to the participant’s sleep experience during the immediately preceding night. Accordingly, the sleep measure was treated as a time-varying exposure representing the prior nocturnal period rather than the concurrent daytime period.
Postpartum blood pressure was measured in the ward according to standard procedures by trained nurses. A hospital-calibrated upper-arm electronic sphygmomanometer was used, with a cuff matched to arm circumference; before measurement, a rest period of ≥5 min was ensured, and the left lateral decubitus or sitting position was adopted, maintaining the same position and the same upper limb; at each measurement, 2 consecutive readings were taken 1–2 min apart, and the average value was taken as the systolic blood pressure and diastolic blood pressure at that time point. The routine monitoring frequency was once every 4 h, continuing from recovery after surgery and returning to the general ward through postoperative day 3, forming a repeated measurement sequence within the 72 h after surgery. For descriptive analysis, the daily means of systolic blood pressure and diastolic blood pressure were calculated.
Blood pressure control status was defined using the clinically accepted threshold of 150/100 mmHg, which is commonly applied in the postpartum management of hypertensive disorders of pregnancy.14 In each 4-hour interval, if systolic blood pressure ≥150 mmHg or diastolic blood pressure ≥100 mmHg, the blood pressure status was classified as “poorly controlled” during that interval; otherwise, it was classified as “well controlled”. For baseline comparisons, overall blood pressure control at the individual level was determined across the 72-hour postoperative period. Participants whose blood pressure remained well controlled during all monitored intervals were classified as having “overall well-controlled” blood pressure, whereas those with at least two poorly controlled interval were classified as “overall poorly controlled”. For temporal alignment with the sleep assessment, blood pressure measurements obtained after the morning RCSQ assessment on a given postoperative day were considered subsequent blood pressure observations for that night’s sleep exposure. In a prespecified sensitivity analysis, an alternative threshold of 140/90 mmHg was also used to define postpartum blood pressure control, to examine the robustness of findings across different clinically relevant cutoffs.
The study outcomes were prespecified according to their clinical and analytical roles. The primary outcomes were repeated postpartum systolic blood pressure and diastolic blood pressure measured within 72 h after cesarean delivery, which were analyzed as continuous longitudinal variables using linear mixed-effects models. These outcomes were selected as primary because blood pressure values were repeatedly assessed during the early postpartum period, allowing evaluation of temporal trajectories and providing greater statistical efficiency for detecting associations with sleep quality.
The secondary outcome was postpartum blood pressure control status defined by the clinical threshold of 150/100 mmHg. Specifically, at each 4-hour interval, blood pressure was classified as poorly controlled if systolic blood pressure was ≥150 mmHg or diastolic blood pressure was ≥100 mmHg; Otherwise, it was classified as well controlled. For descriptive baseline comparisons, participants whose blood pressure remained well controlled during all monitored intervals were classified as having overall well-controlled blood pressure, whereas those with at least two poorly controlled intervals were classified as having overall poorly controlled blood pressure. Because the number of participants with overall poorly controlled blood pressure was limited, analyses of the categorical blood pressure control outcome were considered supportive and exploratory in nature. Sensitivity analyses for the categorical outcome additionally examined an alternative blood pressure control threshold of 140/90 mmHg.
Based on clinical relevance and prior literature, the following covariates were prespecified and included as fixed effects in the multivariable maternal age (years), gestational age at delivery (weeks), breastfeeding status (yes/no), ward residence type (single room/shared room), medication use (yes/no, including antihypertensive medications during pregnancy and postpartum use of magnesium sulfate and antihypertensive agents). All covariates were entered into the multivariable models as fixed effects. Time-varying covariates were modeled using repeated-measures structures where appropriate. In the longitudinal mixed-effects models, postoperative measurement time (scheduled 4-hour interval since surgery) was additionally included as a fixed effect to capture temporal variation in blood pressure during the early postpartum period.
After uniform training, the research staff carried out recruitment, questionnaire distribution, and bedside recording, using standardized paper forms and electronic data captured with dual-track entry. The sphygmomanometers were calibrated monthly by the medical engineering department, and calibration certificates were retained. All scales were checked on site for completeness and logical consistency; if missing items occurred, they were supplemented immediately on site. Data entry adopted double entry by two people, and discrepancies were adjudicated by a third party by checking the original medical records. For abnormal values, source verification and tracing to the original records were performed, and group discussion was organized when necessary to confirm retention or disposal. Variable coding was predefined in the data dictionary and version-locked to ensure consistency before and after analysis.
Statistical analyses followed a prespecified analysis plan. Descriptive statistics were first performed, with continuous variables expressed as mean ± standard deviation and categorical variables expressed as frequency and percentage. Baseline sociodemographic and clinical characteristics were compared between groups according to overall postoperative blood pressure control status (well controlled vs. poorly controlled). For continuous variables, the two independent-samples t-test was used, and for categorical variables, the Pearson chi-square test or Fisher’s exact test was used as appropriate.
The primary analysis evaluated the associations between sleep quality and repeated postpartum blood pressure levels within 72 hours after cesarean delivery. Linear mixed-effects models were constructed using repeated postoperative systolic blood pressure and diastolic blood pressure measurements as dependent variables. An individual-level random intercept was specified to account for within-subject correlation arising from repeated blood pressure assessments over time. Postpartum sleep quality was modeled in parallel forms, including (1) a binary indicator of postpartum sleep quality (normal vs. abnormal), (2) continuous scores of the Richards–Campbell Sleep Questionnaire (RCSQ) core items (items 1–5), and (3) the mean RCSQ score representing postpartum sleep quality. These sleep indicators were entered into the models separately rather than simultaneously to avoid collinearity and to facilitate clinical interpretation. Prenatal sleep quality, assessed by the Pittsburgh Sleep Quality Index (PSQI), was included as both a binary variable and a continuous score, and interaction terms between prenatal and postpartum sleep were examined accordingly.
The secondary analyses evaluated the associations between sleep and postpartum blood pressure control status defined by the threshold of 150/100 mmHg. For prenatal sleep (PSQI), which was measured once before delivery, multivariable logistic regression was used. For postpartum sleep indicators measured repeatedly within 72 h after surgery, generalized linear mixed models with a logit link and an individual-level random intercept were applied to account for within-subject correlation. Both binary and continuous sleep indicators were examined, and the interaction between prenatal and postpartum sleep was also tested. Given the limited number of participants with overall poorly controlled blood pressure, results for the categorical blood pressure control outcome were interpreted as secondary and supportive.
Given the limited number of participants with overall poorly controlled blood pressure, analyses of the categorical blood pressure control outcome were considered secondary and exploratory. To reduce the risk of overfitting, a sensitivity analysis was performed using a clinically prioritized reduced adjustment set including maternal age, gestational age at delivery, breastfeeding status, number of patients in the ward, and postpartum medication use. In addition, to assess the robustness of the findings to outcome definition, an alternative threshold of 140/90 mmHg was used in sensitivity analyses for postpartum blood pressure control. Estimates from these categorical models were interpreted cautiously because of the limited number of events and the possibility of imprecise effect estimation.
All multivariable models were adjusted for the same prespecified set of covariates, including maternal age, gestational age at delivery, breastfeeding status, ward residence type, medication use. Regression results were reported as regression coefficients or odds ratios, with 95% confidence intervals and two-sided P values. The significance level was set at α = 0.05. Missing data were handled using multiple imputations prior to model fitting to minimize potential bias due to incomplete observations. Statistical analyses were performed using R software (version 4.3.2).
During the study period, a total of 158 pregnant women with hypertensive disorders of pregnancy who met the initial screening criteria were enrolled. Twenty-four participants were excluded due to fewer than two assessments of postpartum sleep quality, leaving 134 eligible participants. Subsequently, 11 participants were excluded because fewer than two repeated postpartum blood pressure measurements were available. Ultimately, 123 women were included in the final analysis. Based on postoperative blood pressure control status, participants were categorized into two 99 with well-controlled postpartum blood pressure and 24 with poorly controlled postpartum blood pressure (Figure 1). No significant differences were observed between the two groups in terms of age, education level, parity, prepregnancy body mass index, gestational diabetes mellitus, previous gestational hypertension, multiple pregnancy, type of pregnancy, surgical type, anesthesia method, admission blood pressure, neonatal interdepartmental transfer, gestational weight gain, or prenatal PSQI score (all P > 0.05). In contrast, gestational age at delivery, breastfeeding status, postpartum RCSQ item scores (items 1–5), and mean RCSQ score, and mean systolic and diastolic blood pressure differed significantly between the two groups (all P < 0.05). Specifically, women with poorly controlled postpartum blood pressure exhibited shorter gestational age at delivery, lower breastfeeding rates, poorer postpartum sleep quality, and higher blood pressure levels compared with those with well-controlled postpartum blood pressure (Table 1).Table 1Baseline CharacteristicsVariableOverall (n = 123)Well Controlled N = 99Poorly Controlled N = 24P valueAge, year31.8 ± 4.831.7 ± 4.632.0 ± 5.40.766Education level, n (%)0.632Junior high school and below22.0 (17.9%)17.0 (17.2%)5.0 (20.8%)Senior high school/technical secondary school11.0 (8.9%)10.0 (10.1%)1.0 (4.2%)University and above90.0 (73.2%)72.0 (72.7%)18.0 (75.0%)Parity, n (%)0.730050.0 (40.7%)39.0 (39.4%)11.0 (45.8%)≥173.0 (59.3%)60.0 (60.6%)13.0 (54.2%)Medicine used, n (%)0.052No18.0 (14.6%)18.0 (18.2%)0.0 (0.0%)Yes105.0 (85.4%)81.0 (81.8%)24.0 (100.0%)PreBMI, kg/m224.6 ± 5.924.1 ± 4.126.6 ± 10.40.059Gestational weeks, week37.3 ± 1.937.6 ± 1.536.0 ± 2.7<0.001Anesthesia method, n (%)0.871CSEA108.0 (87.8%)87.0 (87.9%)21.0 (87.5%)GA with ETI1.0 (0.8%)1.0 (1.0%)0.0 (0.0%)Else14.0 (11.4%)11.0 (11.1%)3.0 (12.5%)Admission blood pressure, n (%)0.732Well controlled93.0 (75.6%)76.0 (76.8%)17.0 (70.8%)Poorly controlled30.0 (24.4%)23.0 (23.2%)7.0 (29.2%)Gestational weight gain, kg12.0 ± 10.912.5 ± 7.510.3 ± 19.70.376Previous gestational hypertension, n (%)0.582No109.0 (88.6%)89.0 (89.9%)20.0 (83.3%)Yes14.0 (11.4%)10.0 (10.1%)4.0 (16.7%)Surgical type, n (%)0.796Emergency cesarean section31.0 (25.2%)24.0 (24.2%)7.0 (29.2%)Elective cesarean section69.0 (56.1%)57.0 (57.6%)12.0 (50.0%)VD to CS23.0 (18.7%)18.0 (18.2%)5.0 (20.8%)Neonatal interdepartmental transfer, n (%)0.116No90.0 (73.2%)76.0 (76.8%)14.0 (58.3%)Yes33.0 (26.8%)23.0 (23.2%)10.0 (41.7%)Gestational diabetes mellitus, n (%)0.363No79.0 (64.2%)66.0 (66.7%)13.0 (54.2%)Yes44.0 (35.8%)33.0 (33.3%)11.0 (45.8%)Breastfeeding, n (%)0.023No40.0 (32.5%)27.0 (27.3%)13.0 (54.2%)Yes83.0 (67.5%)72.0 (72.7%)11.0 (45.8%)Multiple pregnancy, n (%)>0.999No120.0 (97.6%)97.0 (98.0%)23.0 (95.8%)Yes3.0 (2.4%)2.0 (2.0%)1.0 (4.2%)Number of patients in the ward, n (%)0.889Single room37.0 (30.1%)29.0 (29.3%)8.0 (33.3%)Shared room86.0 (69.9%)70.0 (70.7%)16.0 (66.7%)Type of pregnancy, n (%)0.713Natural pregnancy103.0 (83.7%)84.0 (84.8%)19.0 (79.2%)Assisted reproduction20.0 (16.3%)15.0 (15.2%)5.0 (20.8%)Prenatal PSQI Score7.9 ± 3.47.9 ± 3.38.3 ± 4.00.580Postpartum sleep quality score 158.2 ± 15.260.2 ± 14.450.1 ± 15.80.003Postpartum sleep quality score 261.2 ± 18.764.4 ± 17.248.3 ± 19.5<0.001Postpartum sleep quality score 358.6 ± 15.060.7 ± 14.249.6 ± 15.3<0.001Postpartum sleep quality score 462.6 ± 16.465.4 ± 15.151.0 ± 16.7<0.001Postpartum sleep quality score 558.3 ± 16.260.9 ± 15.247.8 ± 16.5<0.001Mean score59.8 ± 14.162.3 ± 13.049.3 ± 14.0<0.001Mean SBP129.4 ± 8.8127.2 ± 7.7138.9 ± 6.6<0.001Mean DBP82.1 ± 6.280.6 ± 5.688.4 ± 4.6<0.001Notes: Continuous variables are expressed as mean ± standard deviation and compared using the two independent-samples t-test; categorical variables are expressed as frequency (percentage) and compared using the Pearson chi-square test. Mean the mean of items 1–5.Abbreviations: CSEA, combined spinal–epidural anesthesia; GA with ETI, general anesthesia with endotracheal intubation; VD to CS, vaginal delivery to cesarean section; SBP, systolic blood pressure; DBP, diastolic blood pressure. Figure 1Flowchart for the selection of study participants.Flowchart of participant selection for gestational hypertension study.
Logistic regression and generalized linear mixed models were used to evaluate the associations of prenatal sleep quality, postpartum sleep quality, and their interaction with postpartum blood pressure control. Prenatal sleep was modeled in two (1) a binary indicator (normal vs. abnormal prenatal sleep quality defined by PSQI > 5) and (2) the continuous PSQI total score. Postpartum sleep was also modeled in two forms based on repeated RCSQ assessments within 72 h after (1) a binary indicator (normal vs. abnormal postpartum sleep quality defined by mean RCSQ score < 50) and (2) continuous sleep scores, including the mean RCSQ score and the five item scores (items 1–5). Interaction terms between prenatal and postpartum sleep were examined correspondingly for both binary indicators and continuous scores (Table 2). The results showed that neither prenatal sleep quality nor postpartum sleep quality was significantly associated with postpartum blood pressure control when assessed as a binary outcome, regardless of whether sleep was modeled as categorical variables or continuous scores. Moreover, no significant interaction was observed between prenatal and postpartum sleep quality in either the binary or continuous formulations. These findings remained consistent before and after adjustment for potential confounders (all P > 0.05; Table 2). Overall, the results indicate that prenatal and postpartum sleep quality, either as binary classifications or continuous scores—and their interaction were not independently associated with the likelihood of achieving blood pressure control in the early postpartum period. In these longitudinal analyses, postpartum sleep referred to the previous night and was linked to blood pressure measurements obtained subsequently after the morning sleep assessment, thereby preserving the intended temporal ordering between exposure and outcome.Table 2Association Between Prenatal and Postpartum Sleep Quality and Postpartum Blood Pressure ControlVariablesModel 1Model 2OR (95% CI)POR (95% CI)PNormal prenatal sleep qualityRef.Ref.Abnormal prenatal sleep quality0.865(0.652, 1.150)0.3191.030(0.747, 1.430)0.086Normal postpartum sleep qualityRef.Ref.Abnormal postpartum sleep quality2.273(0.028, 182.383)0.7132.145(0.194, 23.371)0.533The interaction of quality0.889Prenatal sleep quality score1.040(0.999, 1.080)0.0531.040(0.982, 1.080)0.060Mean of postpartum sleep quality score0.980(0.887, 1.083)0.6940.986(0.880, 1.104)0.803The interaction of score0.868Postpartum sleep quality score 10.986(0.894, 1.088)0.7780.995(0.892, 1.110)0.925Postpartum sleep quality score 20.985(0.911, 1.066)0.7110.985(0.945, 1.026)0.467Postpartum sleep quality score 30.986(0.898, 1.082)0.7610.983(0.937, 1.031)0.482Postpartum sleep quality score 40.984(0.903, 1.072)0.7050.982(0.937, 1.029)0.457Postpartum sleep quality score 50.986(0.904, 1.074)0.7410.984(0.940, 1.030)0.484Notes: Analyzed using logistic regression models. Analyzed using generalized linear mixed models. Model 2 was adjusted for the following age, gestational age at delivery, breastfeeding status, number of patients in the ward, medication use.
Sensitivity analyses using a reduced clinically prioritized covariate set and an alternative blood pressure control threshold of 140/90 mmHg yielded broadly similar conclusions. Although some point estimates changed in magnitude, the associations between sleep indicators and the categorical blood pressure control outcome remained statistically non-significant, and confidence intervals remained wide (Supplementary Table 1).
Figures 2 and 3 illustrate the temporal distribution of blood pressure control status and mean blood pressure levels across different time points during the early postoperative period. For both systolic and diastolic blood pressure, the proportion of patients with uncontrolled values progressively decreased from postoperative day 1 to day 3. In contrast, mean systolic and diastolic blood pressure levels demonstrated a gradual upward trend across daily time points, with higher values consistently observed during the evening and nighttime hours. These findings indicate a dissociation between categorical blood pressure control and continuous blood pressure although the proportion of patients exceeding clinical thresholds declined over time, absolute blood pressure values remained relatively elevated later in the day. This pattern suggests possible time-of-day–related variation in postoperative blood pressure regulation and highlights the importance of considering temporal dynamics when evaluating blood pressure control during the early postpartum period. However, because time-of-day was not modeled as a separate covariate in the primary mixed-effects analyses, these findings should be interpreted as descriptive rather than confirmatory evidence of a circadian effect. Figure 2Diastolic blood pressure control status and temporal trends after cesarean delivery.A mixed stacked bar and line graph. Legend: BP Control with Uncontrolled and Controlled; Time of Day with Morning, Noon, Evening, Night. Horizontal axis Postoperative Days (unit not shown). Categories shown in three groups, each group listing Morning, Noon, Evening, Night. Left vertical axis Number of Patients (unit: patients), range 0 to 200. Right vertical axis Average Diastolic BP (unit: millimeters of mercury), range 40 to 100. Stacked bars (Controlled with Uncontrolled on top), approximate totals by group 1 Morning about 110, Noon about 115, Evening about 112, Night about 80; group 2 Morning about 120, Noon about 118, Evening about 114, Night about 86; group 3 Morning about 108, Noon about 60, Evening about 45, Night about 43. Uncontrolled segments are small caps on several bars. Line values (Average Diastolic BP): Morning about 82, 81, 84; Noon about 83, 84, 85; Evening about 81, 82, 86; Night about 80, 84, 86. The detailed data points are as - For postoperative Day 1, Morning, the average diastolic BP (mmHg) was 82.0, patients (controlled) were 106, patients (uncontrolled) were 3 and patients (total) were 109. - For postoperative Day 1, Noon, the average diastolic BP (mmHg) was 83.5, patients (controlled) were 109, patients (uncontrolled) were 6 and patients (total) were 115. - For postoperative Day 1, Evening, the average diastolic BP (mmHg) was 82.5, patients (controlled) were 108, patients (uncontrolled) were 4 and patients (total) were 112. - For postoperative Day 1, Night, the average diastolic BP (mmHg) was 82.0, patients (controlled) were 80, patients (uncontrolled) were 0 and patients (total) were 80. - For postoperative Day 2, Morning, the average diastolic BP (mmHg) was 82.5, patients (controlled) were 116, patients (uncontrolled) were 3 and patients (total) were 119. - For postoperative Day 2, Noon, the average diastolic BP (mmHg) was 82.0, patients (controlled) were 112, patients (uncontrolled) were 5 and patients (total) were 117. - For postoperative Day 2, Evening, the average diastolic BP (mmHg) was 83.5, patients (controlled) were 113, patients (uncontrolled) were 0 and patients (total) were 113. - For postoperative Day 2, Night, the average diastolic BP (mmHg) was 82.5, patients (controlled) were 82, patients (uncontrolled) were 4 and patients (total) were 86. - For postoperative Day 3, Morning, the average diastolic BP (mmHg) was 84.0, patients (controlled) were 105, patients (uncontrolled) were 0 and patients (total) were 105. - For postoperative Day 3, Noon, the average diastolic BP (mmHg) was 85.0, patients (controlled) were 60, patients (uncontrolled) were 0 and patients (total) were 60. - For postoperative Day 3, Evening, the average diastolic BP (mmHg) was 86.0, patients (controlled) were 42, patients (uncontrolled) were 3 and patients (total) were 45. - For postoperative Day 3, Night, the average diastolic BP (mmHg) was 85.5, patients (controlled) were 43, patients (uncontrolled) were 0 and patients (total) were 43.A mixed bar and line graph showing diastolic blood pressure control and averages by time of day. Figure 3Systolic blood pressure control status and temporal trends after cesarean delivery.A mixed stacked bar and line graph. Left vertical axis Number of Patients, range 0 to 200. Right vertical axis Average Systolic BP (mmHg), range 80 to 160. Horizontal axis Postoperative Days, with grouped categories Morning, Noon, Evening, Night repeated across three postoperative days separated by vertical dashed lines. Legend BP Control: Uncontrolled and Controlled. Legend Time of Day: Morning, Noon, Evening, Night. Bars (patients): Day 1 Morning about 110 total with about 10 uncontrolled; Day 1 Noon about 115 total with about 10 uncontrolled; Day 1 Evening about 112 total with about 5 uncontrolled; Day 1 Night about 80 total with about 3 uncontrolled. Day 2 Morning about 120 total with about 5 uncontrolled; Day 2 Noon about 118 total with about 7 uncontrolled; Day 2 Evening about 112 total with about 2 uncontrolled; Day 2 Night about 85 total with about 7 uncontrolled. Day 3 Morning about 110 total with about 5 uncontrolled; Day 3 Noon about 60 total with about 5 uncontrolled; Day 3 Evening about 45 total with about 5 uncontrolled; Day 3 Night about 42 total with about 5 uncontrolled. Lines (average systolic blood pressure): Day 1 Morning about 129, Noon about 130, Evening about 129, Night about 131. Day 2 Morning about 131, Noon about 132, Evening about 130, Night about 134. Day 3 Morning about 132, Noon about 133, Evening about 136, Night about 137. The detailed data points are as - For postoperative Day 1 in the Morning, the average systolic BP (mmHg) was 130, controlled (patients) was 100, uncontrolled (patients) was 10 and total patients was 110. - For postoperative Day 1 at Noon, the average systolic BP (mmHg) was 130, controlled (patients) was 105, uncontrolled (patients) was 10 and total patients was 115. - For postoperative Day 1 in the Evening, the average systolic BP (mmHg) was 129, controlled (patients) was 105, uncontrolled (patients) was 7 and total patients was 112. - For postoperative Day 1 at Night, the average systolic BP (mmHg) was 129, controlled (patients) was 75, uncontrolled (patients) was 5 and total patients was 80. - For postoperative Day 2 in the Morning, the average systolic BP (mmHg) was 131, controlled (patients) was 115, uncontrolled (patients) was 5 and total patients was 120. - For postoperative Day 2 at Noon, the average systolic BP (mmHg) was 132, controlled (patients) was 110, uncontrolled (patients) was 8 and total patients was 118. - For postoperative Day 2 in the Evening, the average systolic BP (mmHg) was 130, controlled (patients) was 110, uncontrolled (patients) was 3 and total patients was 113. - For postoperative Day 2 at Night, the average systolic BP (mmHg) was 133, controlled (patients) was 78, uncontrolled (patients) was 7 and total patients was 85. - For postoperative Day 3 in the Morning, the average systolic BP (mmHg) was 132, controlled (patients) was 100, uncontrolled (patients) was 7 and total patients was 107. - For postoperative Day 3 at Noon, the average systolic BP (mmHg) was 133, controlled (patients) was 52, uncontrolled (patients) was 6 and total patients was 58. - For postoperative Day 3 in the Evening, the average systolic BP (mmHg) was 136, controlled (patients) was 40, uncontrolled (patients) was 5 and total patients was 45. - For postoperative Day 3 at Night, the average systolic BP (mmHg) was 135, controlled (patients) was 38, uncontrolled (patients) was 4 and total patients was 42.A mixed bar and line graph showing blood pressure control and average systolic blood pressure by time of day.
Linear mixed-effects models showed that prenatal sleep quality, either as a binary variable or a continuous PSQI score, was not significantly associated with postpartum systolic or diastolic blood pressure, and no significant interaction between prenatal and postpartum sleep quality was observed (all P > 0.05). In contrast, abnormal postpartum sleep was significantly associated with higher systolic (Model β = 3.630 mmHg, 95% CI 1.820–5.440, P < 0.001) and diastolic blood pressure (Model β = 1.710 mmHg, 95% CI 0.370–3.040, P = 0.012). Moreover, higher mean RCSQ scores and individual item scores were consistently associated with lower systolic and diastolic blood pressure, and these associations remained significant after adjustment for covariates (Table 3 and Table 4).Table 3The Association Between Sleep Quality and Postpartum Systolic Blood PressureVariablesModel 1Model 2β (95% CI)Pβ (95% CI)PNormal prenatal sleep qualityRef.Ref.Abnormal prenatal sleep quality0.131(−3.430, 3.690)0.942−0.139(−3.600, 3.320)0.937Normal postpartum sleep qualityRef.Ref.Abnormal postpartum sleep quality3.410(1.730, 5.080)<0.0013.630(1.820, 5.440)<0.001The interaction of quality0.796Prenatal sleep quality score0.164(−0.295, 0.622)0.481−0.096(−0.367, 0.560)0.680Mean of postpartum sleep quality score−0.092(−0.131, −0.053)<0.001−0.090(−0.129, −0.051)<0.001The interaction of score0.251Postpartum sleep quality score 1−0.064(−0.101, −0.028)<0.001−0.063(−0.100, −0.027)<0.001Postpartum sleep quality score 2−0.080(−0.111, −0.049)<0.001−0.078(−0.108, −0.047)<0.001Postpartum sleep quality score 3−0.054(−0.089, −0.020)0.002−0.054(−0.089, −0.019)<0.001Postpartum sleep quality score 4−0.088(−0.122, −0.053)<0.001−0.085(−0.120, −0.050)<0.001Postpartum sleep quality score 5−0.053(−0.086, −0.020)0.002−0.052(−0.085, −0.019)<0.001Notes: Analyses were performed using linear mixed-effects models. Model 2 was adjusted for age, gestational age delivery,breastfeeding status, number of patients in the ward, medication use. Table 4The Association Between Sleep Quality and Postpartum Diastolic Blood PressureVariablesModel 1Model 2β (95% CI)Pβ (95% CI)PNormal prenatal sleep qualityRef.Ref.Abnormal prenatal sleep quality−0.521(−3.000, 1.960)0.678−0.717(−3.050, 1.620)0.544Normal postpartum sleep qualityRef.Ref.Abnormal postpartum sleep quality1.720(0.486, 2.960)0.0061.710(0.374, 3.040)0.012The interaction of quality0.844Prenatal sleep quality score0.141(−0.178, 0.459)0.383−0.064(−0.248, 0.377)0.684Mean of postpartum sleep quality score−0.051(−0.080, 0.022)<0.001−0.046(−0.075, −0.017)0.002The interaction of score0.684Postpartum sleep quality score 1−0.038(−0.065, −0.011)0.006−0.036(−0.062, −0.009)0.009Postpartum sleep quality score 2−0.045(−0.067, −0.022)<0.001−0.040(−0.063, −0.018)<0.001Postpartum sleep quality score 3−0.034(−0.060, −0.008)0.010−0.030(−0.056, 0.005)0.021Postpartum sleep quality score 4−0.043(−0.069, −0.017)0.001−0.037(−0.060, −0.012)<0.001Postpartum sleep quality score 5−0.028(−0.052, −0.003)0.026−0.024(−0.048, 0.000)0.051Notes: Analyses were performed using linear mixed-effects models. Model 2 was adjusted for age, gestational age delivery, breastfeeding status, number of patients in the ward,medication use.
In women with hypertensive disorders of pregnancy undergoing cesarean section, there were no substantial differences between the well-controlled and poorly controlled groups in baseline characteristics such as age, underlying diseases, prepregnancy weight, type of pregnancy, and surgical type. The differences were mainly reflected in perinatal and postpartum those with poorly controlled blood pressure had shorter gestational age at delivery, a lower proportion of breastfeeding, higher mean postoperative blood pressure, and lower subjective sleep quality as indicated by RCSQ scores and the mean sleep score, with sleep impairment and increased blood pressure showing a synchronous trend. Hypertensive disorders of pregnancy are accompanied by endothelial dysfunction and increased sympathetic activity, and blood loss after cesarean section, fluid redistribution, and incisional pain increase the hemodynamic load.15 Nighttime monitoring care and newborn caregiving lead to repeated interruptions and fragmentation of sleep, with sustained sympathetic activation and activation of the hypothalamic–pituitary–adrenal axis, impeding adequate decline of blood pressure in the early postoperative period.16 Previous studies have mostly focused on the risks of hypertension and preeclampsia associated with sleep disturbances during pregnancy or in the general population, and data on the relationship between early postpartum sleep and blood pressure levels are limited; evidence for this high-risk subgroup of women with hypertensive disorders of pregnancy undergoing cesarean section is even more scarce.17,18 In this study, sleep quality was repeatedly quantified in the early postoperative period and combined with repeated blood pressure measurements, suggesting that postpartum sleep disturbances are relatively common in this population and closely related to early postpartum blood pressure, providing new evidence for incorporating sleep management into postpartum follow-up and care for women with hypertensive disorders of pregnancy.
Analysis using linear mixed-effects models indicated that postpartum subjective sleep quality was persistently associated with systolic and diastolic blood pressure within the 72 h after surgery in women with hypertensive disorders of pregnancy undergoing cesarean section, whereas the relationship between prenatal sleep level and postpartum blood pressure was not obvious, suggesting that the key time window affecting early postpartum blood pressure lies mainly after delivery. Women with abnormal postpartum sleep quality generally had higher blood pressure, and the RCSQ core items and the mean score were consistently negatively associated with blood pressure; every small change in the sleep score corresponded to continuous differences in blood pressure. Such a dose–response relationship remained after adjustment for multiple covariates, including gestational age, metabolic status, medication use, and the ward environment.19,20 These findings support that sleep is not merely a passive indicator, but an independent factor involved in postpartum hemodynamic regulation. Sleep insufficiency and sleep fragmentation readily cause sympathetic activation and a decrease in vagal tone, attenuating the nocturnal decline in blood pressure and even leading to a non-dipping pattern,21 while also activating the renin–angiotensin–aldosterone system, increasing sodium and water retention and small-artery resistance; in women with hypertensive disorders of pregnancy whose endothelial function is already impaired, this is more likely to manifest as sustained elevations in blood pressure.22 There is abundant evidence in the general population and among patients with hypertension demonstrating that shortened sleep duration, sleep-disordered breathing, and decreased sleep quality are associated with higher daytime blood pressure and an increased risk of incident hypertension,23 whereas pregnancy-related research has mostly focused on the effects of late-pregnancy sleep on the occurrence of preeclampsia or gestational hypertension.24 The results of this study suggest that, in this high-risk population of women with hypertensive disorders of pregnancy undergoing cesarean section, early postpartum sleep patterns are likewise reflected in blood pressure levels over a short time frame, providing new support for advancing sleep management to the postoperative inpatient stage and using it as a potential intervention entry point for blood pressure management.
The magnitude of association also deserves clinical interpretation. In the adjusted models, abnormal postpartum sleep quality was associated with a 3.63 mmHg higher systolic blood pressure. Although this difference is smaller than the blood pressure reductions typically sought with anti-hypertensive therapy, it is unlikely to be clinically trivial in a population with hypertensive disorders of pregnancy, in whom short-term blood pressure instability may precipitate severe hypertensive episodes and postpartum complications. In the broader blood pressure literature, a 5-mmHg reduction in systolic blood pressure has been associated with an approximately 10% lower risk of major cardiovascular events, suggesting that even modest shifts in systolic blood pressure may be relevant at the population level. Extrapolating cautiously, a difference of about 3 mmHg may therefore still represent a meaningful contribution to blood pressure burden, particularly in this high-risk postpartum setting. In addition, available postpartum hypertension trials suggest that some active pharmacologic interventions produce blood pressure differences of roughly 4–6 mmHg over short follow-up periods; Thus, the association observed for postpartum sleep in our study is smaller than, but not negligible relative to, the magnitude of early treatment effects reported for selected postpartum interventions. Taken together, these considerations support the clinical relevance of postpartum sleep quality as a potentially modifiable contributor to blood pressure burden, even if its effect size is more modest than that of antihypertensive medication.
Based on analyses using logistic regression and generalized linear mixed models, prenatal and postpartum sleep indicators, whether characterized in binary or continuous form, were not statistically significantly associated with the binary outcome of well or poorly controlled blood pressure; in contrast with their stable associations with continuous systolic and diastolic blood pressure, this suggests that the effect of sleep quality is insufficient in the short term to move more patients across the clinical threshold of 150/100 mmHg. The binary outcome relies heavily on an artificially defined cutoff, and small but persistent changes in blood pressure are more readily reflected in continuous indicators yet may not necessarily drive individuals to cross the threshold in either direction; in the early postpartum stage, women with hypertensive disorders of pregnancy generally receive close monitoring and individualized adjustments of antihypertensive therapy, and drug effects may partially mask the subtle differences attributable to sleep,25 coupled with the relatively limited sample size of those with poorly controlled blood pressure, rendering the statistical power for the categorical outcome insufficient to detect small-to-moderate effects.26 In previous studies, when “at target or not” or “occurrence of hypertensive events” were used as outcomes, sleep-related factors often showed directionally consistent but unstable statistical significance, whereas using continuous blood pressure or dynamic blood pressure load made it easier to observe robust associations.27 In the present study, sleep still showed an independent association with continuous blood pressure levels, suggesting that in postpartum management for women with hypertensive disorders of pregnancy, even if it is difficult to markedly change the categorization of blood pressure control in the short term, nursing interventions focusing on sleep quality may still yield a reduction of several mmHg in blood pressure; such seemingly small improvements remain clinically meaningful for a cardiovascular high-risk population, and they also suggest that future studies evaluating the effects of sleep interventions should focus more on continuous blood pressure and overall blood pressure load, rather than relying solely on binary target attainment. Even when an alternative threshold of 140/90 mmHg was examined in sensitivity analyses, categorical models remained less precise than analyses of repeated continuous blood pressure, underscoring the limitations of dichotomized outcomes in a relatively small cohort.
This study was a single-center study with a relatively small sample size, including only women with hypertensive disorders of pregnancy undergoing cesarean section at a single provincial women and children hospital; the case mix and diagnostic and treatment patterns may differ from those in other regions, limiting the generalizability of the results. The observation window covered only 72 hours after surgery, which may not fully reflect longer-term postpartum sleep patterns and blood pressure changes, and prevents assessment of the impact of sleep improvement on long-term cardiovascular outcomes. Sleep assessment relied on subjective scales such as the PSQI and RCSQ without simultaneous use of polysomnography or actigraphy, introducing a risk of information bias. Although the models adjusted for covariates such as pregnancy outcomes, medication use, and ward environment, factors such as emotional status, history of chronic pain, analgesia strategies, and nighttime breastfeeding frequency were difficult to fully quantify, and residual confounding remained. In addition, although sleep quality was not significantly associated with the binary outcome of blood pressure control, this finding should be interpreted cautiously. In the present cohort, only a limited number of participants met the definition of overall poorly controlled blood pressure, which reduced the statistical power for detecting small-to-moderate associations in categorical analyses. By contrast, repeated continuous systolic blood pressure and diastolic blood pressure measurements provided greater sensitivity to detect temporal blood pressure variation during the early postpartum period. For this reason, the continuous blood pressure analysis was treated as the primary outcome, whereas categorical blood pressure control was considered a secondary and supportive outcome. In addition, although blood pressure trajectories showed variation across daily time points, we did not fit a dedicated circadian model or include a separate time-of-day term in the primary analyses; thus, potential circadian influences on postpartum blood pressure require confirmation in future studies using denser temporal sampling and explicit rhythm modeling. Future studies with multicenter, large-sample, and extended follow-up designs are needed, combining objective sleep monitoring and 24-hour ambulatory blood pressure measurement, to validate these findings and further evaluate the effects of sleep interventions on blood pressure control and cardiovascular prognosis in high-risk women.
From a clinical perspective, the observed postoperative blood pressure trajectory highlights the early postpartum period, particularly the first 72 hours as a high-risk window for hypertensive complications. Enhanced blood pressure surveillance, judicious fluid management, adequate multimodal analgesia, and timely optimization of antihypertensive therapy are essential during this period. For patients exhibiting persistent or progressive blood pressure elevation, prolonged hospitalization and closer monitoring may be warranted to prevent severe complications such as postpartum preeclampsia, eclampsia, heart failure, or cerebrovascular events. These findings emphasize that postpartum blood pressure management should be proactive, individualized, and integrated into routine postoperative care pathways for women with hypertensive disorders of pregnancy.
In women with hypertensive disorders of pregnancy undergoing cesarean section, early postpartum subjective sleep quality was consistently associated with systolic and diastolic blood pressure within the 72 h after surgery, independent of factors such as gestational age,medication use, and the ward environment; prenatal sleep and its interaction with postpartum sleep were not associated, and the effect of sleep on continuous blood pressure was insufficient to change the binary determination of blood pressure control. Although the observed blood pressure differences were modest in magnitude, they may still be clinically relevant in this high-risk postpartum population, in whom even small increases in blood pressure burden may contribute to short-term hemodynamic instability and the risk of hypertensive complications. The inpatient stage in the early postpartum period is therefore a key window for blood pressure management, and sleep may represent an intervenable target with practical relevance for postpartum care. Clinical assessment should emphasize continuous blood pressure and overall blood pressure load and consider incorporating sleep screening and sleep-supportive management into routine care and follow-up for high-risk women, with the aim of reducing early postpartum blood pressure burden.