Hypertensive Disorders of Pregnancy and Future Cardiovascular Health

Hypertensive disorders of pregnancy (HDP) occur in almost 10% of gestations. These women are known to have higher cardiovascular morbidity and mortality later in life in comparison with parous controls who had normotensive pregnancies. Several studies have demonstrated that women with preeclampsia present in a state of segmental impaired myocardial function, biventricular chamber dysfunction, adverse biventricular remodeling, and hypertrophy, a compromised hemodynamic state and indirect echocardiographic signs of localized myocardial ischemia and fibrosis. These cardiac functional and geometric changes are known to have strong predictive value for cardiovascular disease in non-pregnant subjects. A “dose effect” response seems to regulate this relationship with severe HDP, early-onset HDP, coexistence of fetal growth disorders, and recurrence of HDP resulting in poorer cardiovascular measures. The mechanism underlying the relationship between HDP in younger women and cardiovascular disease later in life is unclear but could be explained by sharing of pre-pregnancy cardiovascular risk factors or due to a direct impact of HDP on the maternal cardiovascular system conferring a state of increased susceptibility to future metabolic or hemodynamic insults. If so, the prevention of HDP itself would become all the more urgent. Shortly after delivery, women who experienced HDP express an increased risk of classic cardiovascular risk factors such as essential hypertension, renal disease, abnormal lipid profile, and diabetes with higher frequency than controls. Within one or two decades after delivery, this group of women are more likely to experience premature cardiovascular events, such as symptomatic heart failure, myocardial ischemia, and cerebral vascular disease. Although there is general agreement that women who suffered from HDP should undertake early screening for cardiovascular risk factors in order to allow for appropriate prevention, the exact timing and modality of screening has not been standardized yet. Our findings suggest that prevention should start as early as possible after delivery by making the women aware of their increased cardiovascular risk and encouraging weight control, stop smoking, healthy diet, and daily exercise which are well-established and cost-effective prevention strategies.

Hypertensive disorders of pregnancy (HDP) occur in almost 10% of gestations. These women are known to have higher cardiovascular morbidity and mortality later in life in comparison with parous controls who had normotensive pregnancies. Several studies have demonstrated that women with preeclampsia present in a state of segmental impaired myocardial function, biventricular chamber dysfunction, adverse biventricular remodeling, and hypertrophy, a compromised hemodynamic state and indirect echocardiographic signs of localized myocardial ischemia and fibrosis. These cardiac functional and geometric changes are known to have strong predictive value for cardiovascular disease in non-pregnant subjects. A "dose effect" response seems to regulate this relationship with severe HDP, early-onset HDP, coexistence of fetal growth disorders, and recurrence of HDP resulting in poorer cardiovascular measures. The mechanism underlying the relationship between HDP in younger women and cardiovascular disease later in life is unclear but could be explained by sharing of pre-pregnancy cardiovascular risk factors or due to a direct impact of HDP on the maternal cardiovascular system conferring a state of increased susceptibility to future metabolic or hemodynamic insults. If so, the prevention of HDP itself would become all the more urgent. Shortly after delivery, women who experienced HDP express an increased risk of classic cardiovascular risk factors such as essential hypertension, renal disease, abnormal lipid profile, and diabetes with higher frequency than controls. Within one or two decades after delivery, this group of women are more likely to experience premature cardiovascular events, such as symptomatic heart failure, myocardial ischemia, and cerebral vascular disease. Although there is general agreement that women who suffered from HDP should undertake early screening for cardiovascular risk factors in order to allow for appropriate prevention, the exact timing and modality of screening has not been standardized yet. Our findings suggest that prevention should start as early as possible after delivery by making the women aware of their increased cardiovascular risk and encouraging weight control, stop smoking, healthy diet, and daily exercise which are well-established and cost-effective prevention strategies.

INTRODUCTION
Numerous epidemiological studies have proven the relationship between hypertensive disorders in pregnancy (HDP) and increased cardiovascular risk factors and diseases (CVD) later in life (1,2). This is not surprising as several studies already demonstrated that women with preeclampsia (PE) present in pregnancy a state of impaired myocardial contractility and relaxation, biventricular systo-diastolic chamber dysfunction, adverse biventricular remodeling and hypertrophy, an impaired hemodynamic state, and indirect echocardiographic signs of localized myocardial ischemia and fibrosis (3)(4)(5). The above described structural and functional cardiovascular changes do not completely reverse at 1 year postpartum notwithstanding the unloading conditions and are known to be highly predictive for future cardiovascular adverse outcome (3,6). Furthermore, clinical data has shown that the cardiopulmonary complications already occur in 6% of severe PE and are associated with increased maternal mortality (7). The American Heart Association has already recognized PE as independent risk factor for CVD and introduced this complication of pregnancy in the algorithms for the evaluation of future cardiovascular risk score (8). The physiopathology for this additional CVD risk is still unknown with the existing literature hypothesizing the possibility of pregnancy-induced CVD risk vs. pre-conceptional susceptibility toward an increased risk of CVD or both. The aim of this review is to elucidate the association between HDP and future cardiovascular risk factors and CVD. HDP include gestational hypertension (GH), PE or eclampsia, chronic hypertension, and PE superimposed on top of chronic hypertension. This review will focus on HDP involving new-onset hypertension after 20 weeks' gestation, but other complications of pregnancy such as normotensive small for gestational age or fetal growth restriction as well as normotensive preterm birth, placental abruption, and stillbirth will not be included. This review will focus on the future cardiovascular health of these women, but other outcomes, such as renal failure, diabetes, or other dysmetabolic conditions will not be examined.

SEARCH STRATEGY AND SELECTION OF ARTICLES
A literature examination was accomplished to identify all reports in the English language literature (Medline, National Library of Medicine) published after 2000 until now as previous meta-analysis showed that there were not suitable papers on this topic from 1946 to 1999 (9). The search terms used were "pre-eclampsia, " "gestational hypertension, " "hypertension during pregnancy, " "hypertensive disorders of pregnancy, " "long term outcomes, " "cardiovascular risk, " "cardiovascular health, " "cardiovascular disease, " "essential hypertension, " "peripheral artery disease, " "asymptomatic atherosclerosis, " "diastolic dysfunction, " "heart failure, " "coronary artery disease, " "cerebrovascular disease." We excluded abstracts of oral communications and posters of congresses even if available on Medline. Furthermore, the bibliography of the selected papers were dissected to further identify pertinent studies. The combined set encompassed 230 articles which were reviewed, and a total of 87 articles were considered appropriate for this review. The relevant data were extracted from the full text of the published papers.

RESULTS
The research yielded 59 prospective (3, 6, 10-66) and 28 retrospective studies (67-93) which were synthetically described in Table 1. The number of women analyzed in each study ranged from 58 to 1,072,330 in prospective studies and from 71 to 1,452,926 in retrospective ones. The year of pregnancy included into the studies ranged from 1952 to 2017 in prospective studies and from 1939 to 2016 in retrospective articles. The following complications of pregnancy were included into the studies: HDP, placental infarction and abruption, preterm delivery, low birth weight offspring, small-for-gestational-age fetuses, fetal growth restriction, stillbirth, gestational diabetes mellitus, and pre-gestational diabetes mellitus. The interval between delivery and cardiovascular risk assessment or between delivery and cardiovascular event occurrence were assessed in the majority of the studies, the median of follow up ranging from at least 7 weeks post-partum to more than 30 years after delivery. The median age of the women at the assessment of CV risk or at the occurrence of CV event ranged from 25 to 71 years in the different studies. The following cardiovascular risks or events were differently assessed in the numerous included studies: essential hypertension, heart failure, coronary, cerebro-vascular and peripheral artery disease, asymptomatic atherosclerosis, dyslipidemia, metabolic syndrome, type II diabetes, and renal dysfunction.
The following outcomes will be analyzed in this review: • Essential hypertension • Peripheral artery disease • Asymptomatic atherosclerosis • Asymptomatic heart failure • Heart failure • Coronary artery disease • Cerebrovascular disease.

Essential Hypertension
All authors who addressed the issue of the relationship between HDP and future development of essential hypertension unanimously found that women who developed HDP have an increased risk of having high blood pressure later in life (Table 1). There seems to be a "dose-dependent" effect of HDP and future risk of developing chronic hypertension depending on the severity of the hypertension in pregnancy, the onset of the complication in pregnancy, the need of iatrogenic preterm delivery, the association to fetal growth disorders and the numbers of pregnancies complicated by HDP.      (94). The relation between HDP, future occurrence of essential hypertension and other pre-pregnancy CV risk has also been assessed to try and isolate the effect of pregnancy on this outcome. On this subject, Cho et al. found that the development of high blood pressure in pre-eclamptic women was related to pre-pregnancy factors such as family history, obesity and high blood pressure (86). This latter result is only partially confirmed by Mito et al. who found that women who had HDP presented an higher risk of developing high blood pressure 5 years after the index pregnancy "even after adjusting for confounding factors such as age, body mass index, family history of hypertension and salt intake (odds ratio 7.1, 95% CI, 2.0-25.6, P < 0.003)" (44). Bergen et al. also found that although adjustment for BMI attenuate the relationship between HDP in childbirth age and future chronic hypertension by 65%, however it remained significant (46). In contrast, Wang et al. (83) found that "history of HDP had a four-fold multivariable-adjusted risk (95% CI 2.29-6.24) of hypertension, " while no influence was established by preconceptional anthropometric indices on the incidence of high blood pressure after delivery (83). A stratification of the risk of hypertension depending on the year post-partum after a gestation affected by HDP has also been addressed in several studies. Interestingly, Behrens et al. found that the risk of hypertension associated with HDP was high closely after a complicated gestation and continued for more than two decades (40). About 30% of subjects who suffered from a HDP may develop high blood pressure within 10 years after a complicated gestation; "this risk was 12-to 25-fold higher in the first year, up to 10-fold in 10 year, 2-fold after 20 year" (40). In term of number of women to be screened to detect one case of essential hypertension, Groenhof et al. found that at the age of 35, 9 women with HDP needed to be screened to detect 1 clinically relevant hypertension (64). The risk of developing essential high blood pressure after a gestation affected by HDP has also been correlated to the presence of specific echocardiographic findings (6). Specifically, Melchiorre et al. found that formerly pre-eclamptic normotensive women with moderate-severe echocardiographic left ventricle (LV) anomalies identified at 1 year after delivery were more likely to develop high blood pressure at 2 years after delivery (50%) in comparison to those with normal LV function/geometry or mild LV dysfunction/remodeling (3.5%) with a relative risk of chronic high blood pressure in women with LV moderate-severe abnormalities of 14.5 (95% CI 5.14 to 40.89, P < 0.001) (6). In the above mentioned study the severity of LV dysfunction and hypertrophy was graded according to the European Association and American Society of Echocardiography guidelines (EAE/ASE) (95,96). Specifically: LV remodeling-hypertrophy was defined mild, moderate or severe if LV mass/body surface area (g/m 2 ) was between 96 and 108, 109 and 121 or ≥122, respectively; LV diastolic dysfunction was defined as mild in the case of impaired myocardial relaxation pattern (grade I), moderate in the case of pseudo-normal filling pattern (grade II) or severe if restrictive filling was seen (grade III) accordingly to the EAA/ASE diagnostic algorithms for the diagnosis and grading of diastolic dysfunction (96). LV systolic function was graded based on the ejection fraction value (EF) as mildly, moderately or severely abnormal if EF (%) was between 45 and 54, 30 and 44 or <30. This finding has been confirmed by the subsequent echocardiographic studies (97).

Peripheral Artery Disease
HDP have also been associated with increased risk of developing peripheral artery disease (98). "Peripheral artery disease" (PAD) refers to an abnormal narrowing of arteries other than those that supply the heart or brain, commonly caused by atherosclerosis. PAD most commonly affects the lower extremities vessels (99). "Depending on the degree of narrowing at each vascular site, a range of severity of symptoms may occur, while many patients will remain asymptomatic throughout their life. Occasionally acute events occur, often associated with thrombosis and/or embolism and/or occlusion of a major artery" (ESC task force PAD) (99).
Ray et al. in their population-based retrospective cohort study on more than one million women, assessed the association between HDP, placental abruption and infarction (defined as maternal placental syndromes) and the occurrence of hospital admission or revascularization for PAD at least 90 days after the delivery discharge date (67). They found that the risk of PAD in women who suffered from maternal placental syndromes was 3.8 (2.4-5.9) higher than that of women who had an uneventful pregnancy (67). The future risk of PAD remained significant after adjusting for traditional risk factors for CVD, including maternal smoking and metabolic syndrome (adjusted risk 3: 1.9-4.8) and was highest in women who had HDP in combination with fetal compromise, compared to women who had not (67). Lykke et al. also found that severe PE increased the risk of subsequent thromboembolism of 1.9-fold (range 1.35-2.70) and that the relationship was "dose-dependent" (16). Subsequent studies confirmed the increased risk of PAD in women who had HDP.

Asymptomatic Atherosclerosis
Several studies have addressed the issue of the relationship between HDP and asymptomatic/pre-clinical atherosclerosis ( Table 1). "Asymptomatic atherosclerosis" refers to the coronary/carotid artery inflammatory disease while the condition is still in a subclinical stage but the presence of atherosclerosis can be well-identified and quantified by several invasive and non-invasive techniques, including coronary angiography, ultrasonography, computed tomography, and magnetic resonance imaging (100). In a prospective cohort study, "a history of PE was associated with an increased risk of coronary artery calcifications (CAC) >30 years after the index pregnancy", even after controlling individually for traditional risk factors although this association was not more significant when corrected for current hypertension (39). Specifically, the authors found that the odds of presenting a higher CAC score was 3.54 (CI: 1.39-9.02) times greater in formerly pre-eclamptic women compared to women who did not develop PE without adjustments and it was 2.61 (CI: 0.9-7.14) times greater after correction for current high blood pressure (39). On the contrary, the association between CAC score and history of PE remained significant after adjusting for body mass index alone (odds: 3.20; CI: 1.21-8.49) (39). It has been also shown that one third of subjects who had PE express features of coronary atherosclerosis on vascular computed tomography imaging as compared to one fifth of women from the reference group imaging and this result is manifest as early as at age 45-55 years when women are on average 16 ± 6 years postpartum (59). In contrast, a prospective case-control study showed that the "average maximum carotid intima media thickness (CIMT) was similar among women with vs. without PE (0.831 mm vs. 0.817, p = 0.38), and PE was not a significant predictor of CIMT in a multiple linear regression model (p = 0.63), despite more electrocardiograms compatible with coronary disease" (74). Specifically, the authors of the above mentioned prospective study defined "abnormal electrocardiograms" the ones with at least one of the following: the presence of Q waves/isolated infarct, new left bundle branch block, ST elevation, ST depression, T wave inversion (74). The conventional measurement of common carotid artery intima-media thickness (CCA-IMT) does not represent this (60) as opposed to measurement of the individual CCA intima and media thicknesses which visibly reflect augmented vascular risk (60).

Asymptomatic Heart Failure
Several studies assessed the relationship between HDP and the subsequent development of asymptomatic (stage B) or symptomatic (stage C) heart failure (97,101). The American Heart Association and American College of Cardiologists define asymptomatic stage B heart failure as any subject being affected by LV hypertrophy or dysfunction (102). At this regard, it should be taken into account that several echocardiographic studies in pregnancy already showed a state biventricular dysfunction and hypertrophy, low cardiac output and high total vascular resistancein pre-eclamptic women in pregnancy vs. normotensive matched pregnant controls (3-5) A subsequent prospective longitudinal follow-up study of PE vs. controls assessed at 1 and 2 years post-partum showed, integrating conventional echocardiography, color and pulsed wave tissue Doppler, strain and strain rate techniques, that the prevalence of asymptomatic stage B heart failure (HF-B) at one year postpartum follow-up was significantly higher in preterm vs. term preeclampsia and controls (70 vs. 25 vs. 10%, respectively; p < 0.001), but not in term preeclampsia vs. controls (6). Similarly, moderate-severe left ventricular (LV) dysfunction and remodeling (according to the EAE/ASE diagnosis and grading of abnormal LV structure and function) were significantly more prevalent in formerly preterm PE women compared with those who had term preeclampsia and controls (3,6). Similarly, Soma-Pillay et al. showed that women with early onset PE had an increased risk of diastolic dysfunction at 1 year post-partum (RR 3.41, 95% CI: 1.11-10.5, p = 0.04) and regardless of the presence of chronic hypertension (56). Breetveld et al. in their prospective longitudinal cohort study confirmed that the prevalence of HF-B was consistently high (1 in 4) amongst formerly PE women at 1 and 4 years postpartum (41). Moreover, in a subsequent observational case-control study, he confirmed that the prevalence of HF-B in formerly PE women was threefold higher than that detected for healthy parous controls (25 vs. 8%, P < 0.05) at more than 4 years post-partum (47). The increased risk of asymptomatic HF in formerly PE women was further confirmed by Ghossein-Doha et al. who also found that prehypertension increased this risk significantly, while metabolic syndrome elements did not (42). Orabona et al. conducted a prospective study on hemolysis, elevated liver enzymes, and a low platelet count syndrome (HELLP) and PE women and also documented in both HELLP and PE groups a higher prevalence of LV concentric remodeling, diastolic dysfunction and reduced LV ejection fraction at 6 months to 4 years postpartum vs. matched healthy controls (81). Bokslag et.al. also found that history of PE predisposes in middle age, 9-16 years after pregnancy, to worse LV diastolic function, which could increase the likelihood of later heart failure with preserved ejection fraction (84).

Heart Failure
Several large registry-based studies addressed the issue of the relationship between HDP and symptomatic heart failure. Specifically, in a retrospective registry-cohort study on short term cardiovascular outcome in women who had HDP, Jarvie et al. scrutinized all hospital-based deliveries in Florida from 2004 to 2010 and following cardiovascular (CV), non-CV and any second hospitalization to any Florida hospital within 36 months of index delivery excluding subsequent deliveries (89). They found that "women with HDP had twice the risk of CV readmission within 3 years of delivery (6.4 vs. 2.5/1,000 deliveries; P < 0.001), with higher rates among African American women. Heart failure was the most common reason for CV readmission accounting for 78.6% of all CV readmissions and 84.4% of CV readmissions in women with HDP" (89). Furthermore, in a retrospective longitudinal registry study on long term outcome in women who had HDP, Kuo et. al. using National Health Insurance Research Database, found that formerly PE and eclamptic women were at increased risks of congestive heart failure (hazard ratio 9.1, and 7.4, respectively) with a drastic increase of congestive heart failure occurrence at 3 and 10 years since the index pregnancy (90). Similarly, Chen et al. in a retrospective cohort study found that heart failure incidence was greater in the HDP group than controls (9.83 vs. 1.67 per 10,000 person-years), was more likely to develop within 5 years and that severe or recurrent HDP had a greater risk (85).

Coronary Heart Disease
Numerous studies have assessed the relationship between HDP and subsequent coronary heart disease ( Table 1).  "Coronary artery disease" refers to the clinical manifestations of atherosclerosis and comprehends a range of diseases that result from atheromatous change in coronary vessels. Specifically, it includes the following: (stable and unstable) angina pectoris, myocardial infarction, sudden cardiac death (100).
Interestingly, a previous study on cardiac function and structure in pre-eclamptic women showed the presence of echocardiographic findings suggestive of segmental myocardial ischemia and fibrosis (4-6). Specifically, significantly higher prevalence of basal septal post-systolic shortening, identified by Color Tissue Doppler-derived strain analysis, associated to septal bugling and segmental abnormal myocardial strain and strain rate indices were found in a minority of pre-eclamptic women with severe disease vs. normotensive pregnant controls and these echocardiographic findings have been associated in human autopsy studies and in animal in vitro studies to regional myocardial ischemia (4-6). Not surprising the prevalence of ischemic heart disease in formerly PE women is significantly higher than in parous controls with an uneventful pregnancy (1, 101) ( Figure 1). As well as it has been demonstrated in chronic hypertension, this higher risk of developing coronary artery disease in women who had HDP show a "dose dependent effect, ", being higher in the presence of a more severe disease in pregnancy, in the case of associated fetal growth abnormalities and iatrogenic preterm birth and if PE recurred in subsequent pregnancies (55,94) (Figures 2, 3). Moreover, the association was not explained by adjustment for confounding variables, although it was attenuated by the presence of other CV risk factors (55,94). In particular, Tooher et. al. directed a retrospective cohort study on all subjects who delivered at a tertiary hospital in Sydney between the years 1980 and 1989 (n = 31,656) of whom 4,387 had HDP (82). The whole cohort were studied for linkage analysis to future CVDs and he found that the formerly HDP women were at increased risk of admission for future CVD vs. normotensive parous controls (OR 2.1; 95% CI, 1.7-2.6) (82). The median time from the index gestation to the occurrence of CVD was 20 years with a range of 3-29 years (82). Another study demonstrated that among a cohort of subjects with acute coronary syndrome, positive history of pregnancy adverse outcome was related with more severe disease and worse outcome (103). In particular, at presentation with acute coronary syndrome, women who had PE were younger and had more conventional CV risk factors such as essential high blood pressure and an increased soluble fms-like tyrosine kinase:placental growth factor ratio compared to women who had uncomplicated gestation (103). There was also an increased risk of recurrence of acute coronary syndrome at 1 year in women with previous PE (hazard ratio, 6.8) (103). Moreover, women with ≥2 complicated gestation had an increased cardiac mortality risk as shown by Theilen et al. (aHR = 3.3, 95% CI 2-5.4) (91). A large registry-study in Norway also revealed that GH was associated with increased risk of subsequent cardiovascular disease and the highest risk was noticed when GH was combined with fetal growth abnormalities infants and/or preterm delivery (65). Subsequent studies confirmed this association between HDP and increased risk of myocardial infarction at 40-70y with an HR of 2.08 for formerly PE women and 1.56 or formerly GH women (62). Furthermore, an interesting study on the relationship between coronary flow velocity reserve (CFVR) by Doppler echocardiography and previous HDP found that the mean coronary flow velocity reserve was significantly impaired in the early-onset than in the late-onset PE and in the control group with a positive relation between gestational age at PE diagnosis and coronary flow velocity reserve which notable persisted significant after adjustment for conventional cardiac risk factors such as body mass index, blood pressure, and glycated hemoglobin (61).

Cerebro-Vascular Disease
There is general agreement on the link between HDP and subsequent risk of stroke ( Table 1). It has been demonstrated that not only PE/Eclampsia increase the risk of cerebrovascular events with an HR of 10.7 (p < 0.0001) (90), but Haug et al. (62) and Tooher et al. (82) is also associated with increased risk of stroke. A large population-based study found that formerly GH women had a 1.3-fold (95% CI 0.9-1.7) higher risk of suffering from cerebrovascular disease compared with parous controls with uncomplicated gestation (55). In general, HDP women seem to have an increased risk of cerebrovascular events (HR 1.47, 1.15-1.87) at 40-70 years (62). Moreover, no significant difference was found in relation to the use of antihypertensive drugs or the duration of HDP and subsequent hospitalization for stroke (82). Again, there seems to be a "dose dependent" effect of HPD on cerebrovascular mortality, with women who had more than two complicated gestation showing an increased mortality from stroke (aHR = 5.10), as well as the other causes of cardiovascular mortality, compared to parous women who had only 1 or 0 pregnancies complicated by HDP (1,91,94) (Figures 1, 3).

CONCLUSIONS
HDP is associated with increased risk of cardiovascular diseases later in life. This is not surprising as several studies have demonstrated that women with PE present in a state of cardiac dysfunction, ventricular hypertrophy and indirect echocardiographic signs of localized myocardial ischemia and fibrosis (103,104). Moreover, the structural and functional cardiovascular changes do not completely reverse at 1 year post-partum and are known to have strong prognostic value for future cardiovascular morbidity and mortality in nonpregnant subjects (104). The relationship is stronger in the case of severe or early-onset HDP, concomitant fetal growth disorders, need for iatrogenic preterm delivery and recurrent HDP. Adjustments for confounders (such as family history of cardiovascular diseases, high BMI, hypertension, diabetes, and dyslipidemia) do not eliminate, but attenuate this relationship. The underlying mechanisms have not been fully elucidated, but a concomitance of pre-pregnancy predisposition to increased risk of cardiovascular disease and a direct effect of pregnancy on the cardiovascular system may play a role in determining this excess of cardiovascular morbidity and mortality in women who experienced PE and GH in pregnancy. Guidelines regarding timing and extent of cardiovascular follow-up as well as strategies of prevention after HDP are lacking, but it is reasonable to recommend that screening should be started as early as 1 year after delivery and should primarily include awareness of the women of their future increased CV risk and lifestyle modifications with weight control, smoking cessation, healthy diet, and daily exercise. Future studies should address the issue of a structured screening for cardiovascular disease and the impact of timely preventive intervention in improving cardiovascular health in this group of young women.

AUTHOR CONTRIBUTIONS
AK, BT, and KM conceived the study. Articles were examined by AM, VG, AR, and KM. KM drafted the paper. All authors contributed in writing and revising the paper.