This was a cross-sectional analytical study with retrospective and prospective collection of data including women with a history of preeclampsia between January 2015 and June 2021 with a written consent form of participation and a complete medical file. We excluded Women with a cardiovascular comorbidity before pregnancy (chronic hypertension, cardiopathy, kidney disease, diabetes), Women pregnant during our study period or on postpartum less than 6 months.

The study was conducted in three main obstetrical and gynecologic units of Yaounde; The Yaounde Central Hospital, the Yaounde Gyneco-Obstetric and Pediatric Hospital and the Yaounde University Teaching Hospital. Participants were recruited from the 18th of November 2021 to the 1st of June 2022.

The sample size was estimated to 90 using the Fleiss formula, the expected prevalence of family history of hypertension among women with persistent hypertension (80%) and among normotensive women to be (61.3%) [3].

Ethical approvals were obtained from the institutional committee review board of the Faculty of Medicine and Biomedical Sciences of the University of Yaounde I, and administrative authorizations were obtained from all participating hospitals. Participants were identified through their medical records; we identified 714 cases of women with preeclampsia. 378 records were complete, and only 321 had available addresses of these patients, 229 responded favorably to our call, we contacted them by phone and invited them to participate in the study, 163 were included and only 113 were retained. Informed consent was obtained from each participant before inclusion. Data were collected using a data collection sheet.

It was first an interview to collect information on the past medical history, including cardiovascular risk factors (preeclampsia/eclampsia, hypertension, diabetes), obstetrical history (past history of; miscarriage, preeclampsia/Eclampsia, number of years of postpartum, gravidity, parity, new paternity). Afterward, we performed a physical examination to measure weight and height for body mass index (BMI) calculation, blood pressure (measured according to European Society of Hypertension guidelines using a validated automated device (OMRON®MX2 Basic, OMRON HEALTHCARE, INC. Bannockburn, Illinois 60015.CHINA), We considered the average of the last two values of each occasion), and a complete physical exam, Waist circumference, SPO2 (%), pulse rate (bpm), respiratory rate (cpm), headache, visual impairment, chest pain, dyspnea.

ABPM (Ambulatory Measure of Blood Pressure) was performed using a validated automated oscillometric device (OSCAR 2; SunTech Medical; Morrisville NC, USA). Recordings were obtained over a minimum 24-hour period, with measurements taken every 20 minutes during daytime (06:00-22:00) and every 30 minutes during night-time (22:00-06:00). A recording was considered valid if at least 70% of scheduled measurements were successful, including a minimum of 14 valid daytime readings and 7 valid night-time readings, in accordance with the 2018 ESC/ESH Guidelines. Persistent hypertension was defined as a 24-hour mean ≥130/80 mmHg, a daytime mean ≥135/85 mmHg, or a night-time mean ≥120/70 mmHg. The nocturnal dipping pattern was classified as: dipper (10% - 20% nocturnal fall in BP), extreme dipper (>20% fall), non-dipper (<10% fall), and riser (higher night-time than daytime BP).

Echocardiography and electrocardiography were performed and Heart failure was diagnosed according to the 2016 ESC Heart Failure Guidelines, requiring: 1- symptoms and/or signs of heart failure (dyspnoea, fatigue, ankle oedema); 2- structural and/or functional cardiac abnormality on echocardiography; and 3- elevated natriuretic peptides where available. Most cases identified in this study had heart failure with preserved ejection fraction (HFpEF), defined as left ventricular ejection fraction ≥50% with evidence of diastolic dysfunction or elevated filling pressures.

At the end of the interview, blood sampling was collected for lipid profile, serum creatinine, fasting glucose level and uric acid levels.

To address the heterogeneity introduced by the wide range of time elapsed since index delivery (6 months to 7 years), a sensitivity analysis was performed stratifying participants into two groups: those assessed ≤2 years postpartum and those assessed >2 years postpartum. The prevalence of persistent hypertension was compared between these groups using Pearson’s chi-square test.

Hypertension on ABPM is ≥130/80 mmHg over 24 hours, ≥135/85 mmHg for daytime average and ≥120/70 mmHg for the night-time average。

Left Ventricular Hypertrophy was considered if Left Ventricular Mass index > 95 g/m².

The criteria for left ventricular hypertrophy (LVH) was any Cornel voltage (Sv3 + Ravl) ≥ 20 mm which is an indirect electrocardiographic sign of chronic hypertension.

All the data collected were analyzed using the software SPSS version 23.0. Quantitative variables are presented with their mean and standard deviation (SD). Qualitative variables were expressed as counts and proportions. The comparison of means was made using the Student’s T-test. Fisher’s test was used to compare proportions. Multivariable logistic regression was performed to identify independent factors associated with persistent hypertension. Variables entered into the model were those with p < 0.20 on univariable analysis and those considered clinically relevant a priori, including: age ≥ 35 years, family history of hypertension, prior preeclampsia, severity of PE at diagnosis, multiparity, BMI, and LVH on ECG. To maintain parsimony given the 46 hypertensive cases (approximately 1 variable per 10 events), a stepwise backward elimination procedure was used, retaining only variables that remained significant at p < 0.05 after adjustment. Adjusted odds ratios (aOR) and 95% confidence intervals (95% CI) are reported.

Serum creatinine and eGFR were not entered simultaneously into the regression model due to collinearity.

Distribution of study population according to hospitals

The distribution of the study population by hospitals in which they were enrolled is displayed in the figure below. Most of the participants were enrolled from the YCH (39.8%) and YGOPH (35.4%) (Figure 2).

Figure 1. Flow chart of the study population.

Figure 2. Distribution of study population according to hospitals.

Sociodemographic characteristic of the study population

Table 1 below shows that, in our study population, ages ranged from 17 to 43 years. The mean age was 29.78 ± 6.19 years. The most represented age group was that of 26 to 35 years. Thirty-four-point five percent (34.5%) of our participant were from the centre region. Sixty-one point one (61.1%) were married; almost all lived in urban areas (86.7%). Majority were schooled (higher: 43.4% and secondary: 38.9%). Forty-seven point eight (47.8%) were workers and 71.7% earned a low monthly revenue.

Table 1. Sociodemographic characteristics of women after preeclampsia in Yaounde.

Others*: north, far north, North West, south, south west, littoral and east.

Proportion of persistent hypertension after preeclampsia

We had 113 women who had pregnancies complicated by preeclampsia in our different hospital giving us a total proportion of 31% for office blood pressure measurement and 40.7% at ABPM (Table 2).

Table 2. Proportion of persistent hypertension after preeclampsia in Yaounde.

BP: blood pressure ABPM: ambulatory blood pressure monitoring

The mean interval between index preeclampsia and study inclusion was 2.25 ± 1.48 years (range: 0.5 - 7.0 years). In the sensitivity analysis stratifying by time since delivery, the prevalence of persistent hypertension was 43.8% in women assessed ≤2 years postpartum and 36.7% in those assessed >2 years postpartum (p = 0.45), suggesting that the pooled prevalence estimate was not substantially driven by time since delivery.

Blood pressure phenotypes at ABPM

The most represented hypertensive phenotypes were sustained hypertension 30 (26.5%) followed by masked hypertension 16 (14.2%) (Table 3).

Table 3. Blood pressure phenotypes of women after preeclampsia in Yaounde.

Frequency of hypertension

Hypertension was most of the time permanent 29 (63%) (Table 4).

Table 4. Frequency of hypertension in women after preeclampsia at Yaounde

Type of hypertension in women after preeclampsia inYaounde

More than half of the hypertensive women had both type of hypertension 38 (82.6%) (Table 5).

Table 5. Type of hypertension in women after preeclampsia at Yaounde.

Dipping pattern,adrenergiccomponentand pulse pressure of our studypopulation

Majority of women (61.9%) in our study population had an adrenergic component involved at ABPM. Almost all had a normal pulse pressure (Table 6).

Table 6. Adrenergic component and pulse pressure of our study population.

Table 7 below shows that, more than half of our study population were non-dipper (68.1%), and followed by dipper (31%).

Table 7. Dipping pattern of women after preeclampsia in Yaounde.

Comparison between ages

In our study population, ages ranged from 17 to 43 years. The mean age was 29.78 ± 6.19 years (32.37 ± 5.63 for hypertensive women and 28 ± 5.95 for normotensive women). The most represented age group was that of 26 to 35 years (Table 8).

Table 8. Age comparison between hypertensive and normotensive women afterpreeclampsia.

Comparison between region of origin,residenceand marital status amonghypertensive and normotensive women

The region of origin and the marital status were significant with persistent hypertension (Table 9).

Table 9. Comparison between region of origin, residence and marital status among hypertensive and normotensive women.

Comparison between educational level,monthlyrevenueand religion amonghypertensive and normotensive women

Other characteristics like educational level, monthly revenue and religion were distributed in a homogenic manner, hence not significant.

Hypertensive and normotensive women(Table 10)

Table 10. Comparison between educational level, monthly revenue and religion among hypertensive and normotensive women.

Comparison of the reproductive characteristics among hypertensive andnormotensive women

Table 11. Comparison of the reproductive characteristics among hypertensive and normotensive women.

Table 11 above shows that gestation (p = 0.025), parity (p < 0.001) and prior preeclampsia (p < 0.001), were significantly associated with persistent hypertension. The mean parity was 1.81 ± 1.87 weeks (2.52 ± 1.918 for hypertensive women and 1.31 ± 1.68 for normotensive women). Parity ranged from zero to eight.

Comparison of personalpast historyamong hypertensive and normotensivewomen

Table 12 below shows that medical and environmental past history variables are evenly distributed among the population. Hence not significantly associated to persistent hypertension.

Table 12. Comparison of personal past history among hypertensive and normotensive women. Hypertensive and normotensive women.

HTN: hypertension.

Comparison of familypast historyamong hypertensive and normotensive women

Table 13 below shows that having a family of hypertension (p = 0.001) in the first degree was significantly associated with persistent hypertension.

Table 13. Comparison of frequencies of clinical symptoms between women with and without persistent hypertension following preeclampsia

Table 14 below shows that GA at delivery (P = 0.014), severity of preeclampsia (p ≤ 0.001) and mode of delivery (p = 0.026) were significantly associated with persistent hypertension.

Table 14. Comparison of the information on labor and delivery among hypertensive and normotensive women.

GA: Gestational age PE: Preeclampsia.

Comparison of the clinical characteristics of hypertensive and normotensive women

The mean interval period between preeclampsia and the diagnosis of hypertension was 2.25 ± 1.48 years ranging from 6 months to 7 years postpartum.

Comparison of the clinical symptoms of hypertensive and normotensive women

Clinical symptoms are significantly associated to persistent hypertension (Table 15).

Table 15. Comparison of symptoms between hypertensive and normotensive women.

Table16 below shows that, waist circumference (p = 0.038) and BMI (p = 0.003) were significantly associated with persistent hypertension. The mean waist circumference was 90.04 ± 14.69 cm (96.08 ± 6.96 for hypertensive and 82.76 ± 12.50 for normotensive women). The mean for BMI was 27.50 ± 6.86 Kg/m² (31.86 ± 8.67 for hypertensive and 25.80 ± 4.92 for normotensive women).

Table 16. comparison of the waist circumference and BMI between hypertensive and normotensive women of the waist circumference and BMI between hypertensive and normotensive women.

BMI: Body mass index.

Table 17 below shows that left ventricular hypertrophy, left atrial volume and left ventricular geometry were significantly associated with persistent hypertension. LVH was found in both groups (8%) which can be considered a possible outcome. Other parameters such as repolarization, rhythm, left atrial volume, atrioventricular and intraventricular conductions were normal. The mean cornel voltage was 13.78 ± 5.74 bpm and 12.5 ± 4.40 bpm respectively in hypertensive and normotensive women. The mean heart rate was 72.79 ± 10.03 bpm and 71.43 ± 11.49 bpm respectively in hypertensive and normotensive women. Concerning echographic parameters, geometry and LAVi are significantly associated with persistent hypertension and ejection fraction was conserved for all participants. Of the 10 with an abnormal geometry, 2 (1.8%) had heart failure with conserved left ventricular ejection fraction.

Table 17. Comparison of the electrocardiographic and echocardiographic profile among the study population.

LAV: Left atrial volume, LVH: Left ventricular hypertrophy, ECG: electrocardiogram.

Biological profile of the study population

ϖ Comparisonof biological parameters among hypertensive and normotensivewomen.

Table 18 below shows us that total cholesterol was significantly associated with persistent hypertension. Other biological factors were evenly distributed in the two groups.

Table 18. Comparison of biological parameters among hypertensive and normotensive women.

eGFR: estimated Glomerular filtration rate, HDL: High density lipoprotein, LDL: Low density lipoprotein, FGL: Fast Glucose Level.

Factors associated to persistent hypertension after preeclampsia at Univariateanalysis

Table19 below shows the factors associated to persistent hypertension after preeclampsia at Univariate analysis.

Table 19. Univariate analysis of factors associated with persistent hypertension.

PE: preeclampsia LVH: left ventricular hypertrophy ECG: electrocardiogram.

Factors associated to persistent hypertension after preeclampsia at multivariateanalysis

Table 20. Factors associated with persistent hypertension after multiple logistic regression.

HTN: Hypertension PE: preeclampsia.

Table20 above shows that, after adjustment for confounders using multiple logistic regression, the factors found to be associated to persistent hypertension were: Family history of hypertension, prior pregnancy with preeclampsia, and Severe PE at diagnosis.

Using office blood pressure measurements, we found a proportion of 31%. Some authors worked on the same subject as us [3][8][10][11]: among them, there are those who had results comparable to ours, this is the case of the work carried out by Nganou-Gnindjio et al. and Amougou et al. in Cameroon, who found a prevalence of 32.6% [3] and 23.53% respectively. Similarly, Nakimuli et al. in Uganda and Fajardo Tornes et al. in Cuba found a prevalence of 34% and 27.8% respectively. Other authors reported the prevalence of persistent hypertension lower than that reported by our study. This is the case of Manten et al. in 2007 who found a prevalence of 22% [12]. A study carried out by Kaze et al. in 2016 found a prevalence of 14.8%. This low prevalence can be explained by the difference in study follow-up time period. Other authors have had superior results to ours ranging from 38.7% [13] to 57.4% [14]. The high prevalence reported by these studies could be due to contextual differences that exist among the different study populations and settings. Using 24-hour ABPM, we found a proportion of 40.7% this is similar to results obtained by Benschop et al. [14] who had 41.5%. But this result is inferior to that found by Ditisheim et al. [15] who had 50%. This difference in prevalence could be explained by the fact that we excluded women with superimposed preeclampsia which was part of their study population. Lower prevalence of 29.5% was reported by Ditisheim et al. [15]. This difference may be explained by the fact that our study population was all black which would be more likely to develop hypertension.

The wide range of time since delivery (6 months to 7 years) introduces heterogeneity into the study population. Although our sensitivity analysis did not reveal a significant difference in hypertension prevalence between women assessed early versus late postpartum, this remains a methodological limitation inherent to the cross-sectional design. A prospective cohort study with standardized follow-up intervals would provide more precise estimates of the temporal trajectory of post-PE hypertension.

4.2.1. Hypertensive Phenotypes at ABPM

In our study, the most represented hypertensive phenotype was sustained hypertension 30 (26.5%). Some authors have worked on the same subject as us [14]-[16]. Among them there are those who had results similar to ours, this was the case of the work of Ditisheim et al. [15], who reported a prevalence of 20.5%. Studies carried out by Benschop et al. in 2018 and Ditisheim et al. in 2018 had a lower prevalence of 14.5% [14] and 14.2% [15] respectively. This can be explained by the fact that most of their study population was under anti-hypertensive medications. In our context only 4 women were initially under antihypertensive medications but the level of observance was very low (majority stopped taking medication 4 - 8 months prior our study) knowing that around three months after stopping medications, the effect of the drug is not more present. Masked Hypertension represented 14.2% of hypertension at ABPM, this is similar to others studies carried out by Pechere et al. in 2013, Benschop et al. in 2018 and Ditisheim et al. in 2018 who had prevalence of 11.6% [16], 17.5% [14] and 17.9% [15] respectively. Masked hypertension has been associated with an increased risk of developing sustained hypertension and cardiovascular events, independent of office BP. A known risk factor for masked hypertension is prehypertension which affected 28.6% of women in our study. Considering the potential long-term consequences for the mother, masked hypertension in the postpartum needs closer attention. Our study found that 2.7% of women presented white-coat hypertension. In some studies carried out by Benschop et al. [1] in 2018 in the Netherlands and Ditisheim al in 2018 in Switzerland. WCH was associated with modest increased risk of stroke and target organ damage. If the management of WCH is uncertain, especially in the postpartum period, its identification may prevent unnecessary use of antihypertensive medication. However, WCH should not be treated with antihypertensive medication because this can lead to hypotension.

4.2.2. Dipping Pattern (Mean Circadian Rhythm)

Non-dipping profile is also associated with increased cardiovascular risk and target organ damage. Early studies on circadian variations of BP during preeclampsia showed the loss of normal nocturnal fall and this was correlated to the severity of the disease [17]. In the present study, we found that 68.1% of women had abnormal nocturnal dipping. Sleep disruption of some breast-feeding mother may explain the absence of nocturnal BP drop in both groups. This result was similar to those obtained by Benschop et al. (50.5%) [14] and Ditisheim et al. (59.8%) [15] with similar reasons. Several pathological mechanisms have been suggested to explain the insufficient reduction of night-time BP, including deficient decrease of night-time sympathetic activity. Sympathetic activity is inversely associated with the difference in day-to-night BP, suggesting that it may influence 24- hour BP pattern in hypertensive individuals [14]. In women with preeclampsia, sympathetic over activity is described both during and after pregnancy. This might explain the high prevalence of adrenergic component and a disadvantageous dipping pattern in our study population.

Age

The mean age of women was 29.78 ± 6.19 years. The average age found could be explained by the fact that Cameroonian population is young, especially women, as affirmed by the Cameroon Demographic and Health Survey (CDHS) 2018 which reports an age ranging from 15 to 49 years of 62.32% [18]. Some authors have worked on the same subject as us. Among them, there are some who had results similar to ours [7][8][10][12][18] The results similar to ours can be explained by the fact that they also carried out their studies on the same target population as us. However, some results are superior to ours. This is the case of Benschop et al.[14] and Ditisheim et al. [15] both in 2018. These can be explained by the fact that their study population consisted of adults with an age varying from 21 to 45 years and an average age of 33.7 ± 5.7 years.

Family history of hypertension was distributed in a heterogeneous manner in our study population, hence significantly associated with persistent hypertension. This results corroborates to those found by Dohou et al. [16], Amougou et al. [10]. We found an association between parity and the risk of chronic hypertension. This result was similar to that found by Nkwabong et al. in 2016 in Cameroon [6] and Nakimuli et al. in Uganda in 2011 [11]. We also found that, severity of preeclampsia at diagnosis was significantly associated with persistent hypertension. This result corroborates to that of Nakimuli et al. in Uganda in 2011 [11] and Smith et al. in 2009 in Canada [19]. Personal history of preeclampsia in previous pregnancies was found to be associated with the persistent hypertension in our study. Smith et al. in 2009 in Canada [19], Benschop et al. in 2018 in the Netherlands [14] and Eldow et al. in 2009 in Phildelphia [13], reached the same conclusion. Gestational age at delivery was significantly associated with persistent hypertension in our study. Previous studies such as that by Nkwabong et al. [6] and Nakimuli et al. [11] have all shown similar results.

Waist circumference (p = 0.038), and BMI (p = 0.003) were significantly associated with persistent hypertension in our study population. Amougou et al. [10], Fajardo et al. [8] and Dohou et al. [16] have all shown similar results.

We assessed electrocardiographic LVH with the development of persistent hypertension in women with history of preeclampsia. We found a prevalence of 8%, this is similar to results found by Drost et al. in 2012 in the Netherlands who had a prevalence of 5.7% [20] but these results are different from those of Hoogsteder et al. who found a prevalence of 0.3% [21]. This difference in prevalence could be explained by the small size of our study population also, their study included all HDP. Other echographic parameters were normal.

Concerning echographic parameters;

In our study population, 8.9% of women had an abnormal geometry. Ejection fraction was preserved for all participants with slightly enlarged atrial dimensions (18.6% 22/113). This can be explained by the fact that cardiac changes may persist following hypertensive pregnancies. This result corroborates to the findings of Melchiorre et al. in the United Kingdom who had 10.9% of women with abnormal geometry with conserved ejection fraction and slightly enlarged atrial dimensions [22]. Scantlebury etal. had higher value of 15.8% [23]. This difference may be due to the fact that they included all hypertensive disorders of pregnancy.

In our study population 1.8% of women had heart failure with conserved ejection fraction. This result is similar with that of Scantlebury et al. in the United States who had 0.8%. However, this result is different from that of Breetveld et al. who had a higher proportion of 19%. This difference may be due to small size of their study population.

In our study, high total cholesterol levels were significantly associated with persistent hypertension. This result is similar to those found by Manten et al. in 2007 in the Netherlands [12]; Smith et al. in Canada [19] and Eldow et al. in Philadelphia [13]. It can be justified by the fact that dyslipidemia is associated to the risk of chronic hypertension.

In our study, we have identified four independent associated risk factors:

Family history of hypertension: We assessed the link between family history of hypertension and the risk of future chronic hypertension (OR = 5.18 (1.32 - 20.37); p = 0.001) after PE and found that the presence in the siblings of the affected woman was independently associated to progression to chronic hypertension. The presence of a family history of hypertension especially in first degree relative, is a well-known risk factor of chronic hypertension and once more points the fact that a genetic component is implicated in the pathogenesis of hypertensive diseases in pregnancy and primary hypertension. This result corroborates to those found by Dohou et al. [16], Amougou et al. [10] and Ditisheim et al. [15].Prior pregnancy with preeclampsia: We also found that, severe preeclampsia at diagnosis represents 74.3% of our study population. This result corroborates to that of Nakimuli et al. in Uganda in 2011 [11], who found that 47 out of 64 cases (73.4%) had severe preeclampsia. Smith et al. in 2009 in Canada reported that 51.4% of cases had severe preeclampsia [19]. This could be due to the fact that severe preeclampsia has been noted to be more strongly associated with a wide range of internal placental factors with which their subsequent release into maternal circulation leads to endothelial dysfunction, generalized vasospasms and ultimately impaired multiple organ dysfunctions [21].Severe at preeclampsia diagnosis: A history of preeclampsia in previous pregnancies was found to be associated with the persistent hypertension in our study. This can be explained by the fact that, with each pregnancy women have an increased vascular and metabolic risk and after pregnancy the residual risk is greater in women with preeclampsia than those with a normal pregnancy. With each recurrence of preeclampsia, this risk is slightly higher-this is known as “additional memory” of risk [9]. Thus, a woman with her second episode of pregnancy is more likely to develop chronic hypertension than a woman with her first episode. Smith et al. in 2009 in Canada [19], Benschop et al. in 2018 in the Netherlands [14]and Eldow et al. in 2009 in Philadelphia [13], reached the same conclusion.