twenty-four hour blood pressure monitoring in early pregnancy: is it predictive of pregnancy-induced...

Acta Obstet Gynecol Scand 1998; 77: 14–21 Copyright C Acta Obstet Gynecol Scand 1998

Printed in Denmark – all rights reservedActa Obstetricia et

Gynecologica ScandinavicaISSN 0001-6349

ORIGINAL ARTICLE

Twenty-four hour blood pressure monitoringin early pregnancy: is it predictive ofpregnancy-induced hypertension andpreeclampsia?CHIARA BENEDETTO, LUCA MAROZIO, MAURIZIO GIAROLA, LUCIANO CHIAROLINI, VINCENZA MAULA AND

MARCO MASSOBRIO

From the Department of Gynecology and Obstetrics, University of Turin, Turin, Italy

Acta Obstet Gynecol Scand 1998; 77: 14–21. C Acta Obstet Gynecol Scand 1998

Objective. To investigate whether a chronobiological analysis applied to automated 24-hourblood pressure monitoring in early pregnancy provides objective parameters enabling detec-tion of single patients at risk of pregnancy-induced hypertension or preeclampsia.Methods. 24-hour automatic blood pressure monitoring was performed at 8–16 and 20–25gestational weeks in 104 women at risk of pregnancy-induced hypertension or preeclampsia.The subjects were hospitalized to be synchronized to rest-activity and meal-timing schedules.All women were followed longitudinally until post-partum. Chronobiological analysis ofblood pressure values was performed; sensitivity, specificity and predictive values of MESORand hyperbaric index were also calculated. Incidence of pregnancy-induced hypertension orpreeclampsia, gestational week at delivery and birthweight were recorded.Results. Nine thousand nine hundred and eighty-four blood pressure measurements wereanalyzed. In patients who later developed overt hypertension, systolic and diastolic bloodpressure MESOR, hyperbaric index and percent time elevation were already significantlyhigher in early pregnancy than in those who remained normotensive. The best sensitivity andspecificity was obtained between 20–25 weeks of gestation with systolic single cosinor MES-OR and Hyperbaric Index using as cut-off 103 mmHg (sensitivity: 88%; specificity: 75%) and10 mmHg/24 hour (sensitivity: 70%; specificity: 92%), respectively.Conclusions. The chronobiological analysis applied to 24-hour blood pressure monitoringduring pregnancy allows definition of objective cut-off values which can be particularly usefulin the routine clinical practice when the risk of developing pregnancy-induced hypertensionor preeclampsia must be calculated in the individual subject.

Key words: chronobiological analysis; prediction; preeclampsia; 24-hour blood pressure moni-toring

Submitted 18 February, 1997Accepted 19 May, 1997

Pregnancy-induced hypertension (PIH) and preec-lampsia (PE) remain a major cause of maternaland perinatal mortality and morbidity, including

Abbreviations:PIH: pregnancy-induced hypertension; PE: preeclampsia; BP:blood pressure; SBP: systolic arterial blood pressure; DBP: dias-tolic arterial blood pressure; HBI: hyperbaric index; PTE: per-cent time elevation; MESOR: mid-line estimating statistics ofrhythm.

C Acta Obstet Gynecol Scand 77 (1998)

premature delivery, low birth-weight, and intra-uterine growth retardation (1–4).

The ability to prevent PIH and PE is limited bylack of knowledge regarding their cause. However,identification of the high-risk individual, followedby close clinical and laboratory monitoringallowing us to recognize the disease process in itsearly stages, does help to prevent catastrophic ma-ternal and fetal sequelae (5).

Blood pressure monitoring and preeclampsia 15

Certain maternal conditions (i.e. first pregnancy,chronic hypertension, chronic nephropathy, dia-betes, previous obstetric history of PE, family his-tory positive for eclampsia, twin pregnancy) (5–10)are associated with an increased risk for PIH andPE. Nevertheless, although such characteristicsmay help to identify at-risk groups of patients, pre-dicting the development of PIH or PE for individ-ual subjects remains difficult.

Since an ideal predictive test should be simpleand easy to perform, attempts have been made touse blood pressure (BP) recordings obtained beforethe third trimester as early indicators of PE (11).However, a comparison of the larger studies (12–14) reveals a wide disparity in sensitivity and posi-tive predictive values. A possible reason for thesediscrepancies may have been the lack of a rigidlystandardized method of BP measurement. Twenty-four hour BP monitoring allows us to overcomethe bias due to single and random measurementsof BP. In previous studies it has been demonstratedthat 24-hour BP monitoring may be useful in themanagement of hypertensive pregnant women, inorder to define more precisely BP elevations andthe prognosis of the hypertensive disease, and toplan an appropriate follow up of the patients (15,16). Indeed, 24-hour BP monitoring results maybe objectively evaluated using a chronobiologicalanalysis: this allows us to transform a periodicseries of values in single parameters, which giveinformation not only on BP elevations, but also onthe rhythmic characteristics of BP during the 24-hour period. It is possible that in pregnancy, as innon pregnant state (17), early alterations of rhyth-mic patterns of BP precede the development ofovert hypertension.

At present, however, the studies on the utility ofthe 24-hour monitoring for the early detection ofsubjects at risk to develop PIH or PE are still in-sufficient (18, 19); moreover, in none of thesestudies a chronobiological analysis of BP valueshas been used.

Therefore, we decided to investigate whether thechronobiological analysis applied to automated24-hour BP monitoring in early pregnancy pro-vides objective parameters allowing us to detect in-dividual patients at risk of PIH or PE.

Methods

We recruited 104 normotensive pregnant womenwho satisfied one or more of the following criteria:family history of eclampsia; obstetric history ofPE; age over 40 years; current twin pregnancy(Table I). According to the results of our previousstudies (10,20), such risk factors identify a popula-tion at moderate/high risk to develop PIH or PE.


All subjects included in the study were seen con-secutively over a three year period, and they wereselected among patients referred to the Center forthe Prevention, Diagnosis and Treatment of Hy-pertension in Pregnancy of our department.

At the booking visit, all subjects underwentphysical examination, BP measurement, hematol-ogical and urinary examination, renal clearancesand proteinuria, to exclude chronic hypertension,chronic nephropathy and major systemic diseases.Gestational age was assessed by last menstrualphase and confirmed by ultrasound.

In each subject, BP monitoring was firstly per-formed at 8 to 16 and then repeated at 20 to 25gestational weeks, and all women were followedlongitudinally until post-partum from the time ofrecruitment. In our setting, in order to avoid inter-ference due to environmental factors and physicalactivity, and to obtain more standardized chrono-biological results, all subjects were hospitalizedduring monitoring to be synchronized to rest-ac-tivity and meal-timing schedules (breakfast 6.30a.m., lunch 0.30 p.m., dinner 6.30 p.m.). They re-mained in bed for most of the time and the 24-hour monitoring was started between 10.00 a.m.and 1.00 p.m.

Throughout the 24-hour period, systolic (SBP)and diastolic BP (DBP) were measured every 30minutes from the right arm with an automatic os-cillometric monitor (Dinamap 1846 SXP, Critikon,Ethicon, Pratica di Mare, Roma, Italy) that hasbeen validated for use in pregnancy (21, 22); dur-ing hospitalization, each subject underwent routinehematological and urinary examination and ultra-sonography. Renal clearances and proteinuria werealso evaluated in 24-hour urine collection.

The study received ethical approval with the In-stitutional Committee responsible for human ex-perimentation and all women who participated inthe study did so voluntarily, having given their in-formed consent.

The main outcome measure of the study was theappearance of PIH or PE, which was diagnosedusing traditional sphygmomanometric measure-ments and the criteria reported by Davey and MacGillivray (23) (DBP Ø90 mmHg in two consecutivemeasurements four hours or more apart after thetwentieth week of pregnancy in a previously nor-motensive woman for PIH; DBP Ø90 mmHg andproteinuria Ø300 mg/24h after the twentieth weekof pregnancy in a previously normotensive andnonproteinuric woman for PE). The gestationalage at delivery and the newborn weight were care-fully recorded for each patient.

Data on SBP and DBP were downloaded intoan IBM personal computer. An observer editedeach 24-hour record, noting missing values and

16 C. Benedetto et al.

Table I. Characteristics of the patients included in the study

Patients at risk

Primigravidae Multigravidae Total

No. of patients 19 85 104– Age (x∫s.e.m.; range) 26.6∫1.2 32.1∫0.7 31.0∫0.6

(18–36) (22–43) (18–43)

Risk factors– Family history of eclampsia 6 4 10– Previous preeclampsia – 62 62– Age ±40 years 6 11 17– Twin pregnancy 7 8 15

probable artefacts. Rejected or missing values werereplaced by linearly interpolated values, providedthat no more than two consecutive samples re-quired editing. At the end of the study, rejected ormissing values resulted were less than 1% of all BPmeasurements.

We analyzed the data after 3 years because thesample size was large enough to detect a significantdifference for systolic MESOR (13 mmHg) and fordiastolic MESOR (8 mmHg) between subjects whoremained normotensive and those who became hy-pertensive (80% power at aΩ0.05, two-tailed test).

Data series of each BP monitoring session wereanalyzed by the single Cosinor method to evaluatetheir circadian rhythmicity. This procedure consistsof the least squares fit of a 24-hour cosine curve tothe data, thus obtaining the estimates of the fol-lowing parameters: the MESOR (Midline Estimat-ing Statistic Of Rhythm), defined as the averagevalue of the rhythmic function fitted to the BPdata; the amplitude, that is the measure of one half

Fig. 1. This figure gives an example of the chronobiological parameters described in this paper, calculated for diastolic bloodpressure (DBP) values in one pregnant woman become hypertensive. In part A of the figure the parameters evaluated by least-squaresfit of cosine model are represented: MESOR (Midline Estimating Statistic Of Rhythm), Amplitude and Achrophase. Chronogram: isthe individual display of data as a function of time. The continuous sinusoidal curve is the expression of the mathematical functionwhich better approximates the rhythm. In part B the superimposition of the chronogram is shown over the normal range (chrono-desm) of corresponding gestational weeks, which allows for calculation of the hyperbaric index ( HBI ) and percent time elevation(PTE). Chronodesm: is the time-qualified reference interval, calculated with 95% prediction limits.


the extent of rhythmic variation of BP in a cycle;the acrophase, a measure of timing, i.e. the lagfrom a reference timepoint (in this study local mid-night) of the crest-time in the cosine curve appro-priately approximating the BP circadian rhythm(Fig. 1A). The significance value for a rhythm iscalculated by the statistical Zero Amplitude Test;i.e. in the cosine function

f(t)ΩMπA cos (wtπF)

(where MΩMESOR; AΩAmplitude; FΩAcro-phase; wΩangular frequency; tΩtime) is tested thenull hypothesis: AΩ0. Rejection of this null hypo-thesis signifies that the fitted curve (implyingrhythmicity) approximates the data more closelythan does a straight line with zero slope (implyingconstancy). Acceptance of the hypothesis AΩ0 im-plies either that the data are essentially constantwith time or that the single cosine model is other-wise inappropriate (24).

The parametric endpoints obtained for each in-


dividual data series were summarized by popula-tion-mean cosinor method (24), obtaining the cir-cadian BP rhythmometric characteristics for eachgroup of subjects and gestational period.

BP data profiles of subjects were also comparedwith time-qualified reference intervals (chrono-desms), calculated with 95% prediction limits, onthe basis of BP data collected for each period ofpregnancy from 73 healthy normotensive pregnantwomen without risk factors for PIH or PE, of anyparity and aged 18–40 years. They were randomlyrecruited among subjects referred to the antenatalclinics of our department, and gave their informedconsent to undergo hospitalization and 24-hourBP monitoring. Before recruitment, all normoten-sive subjects underwent physical examination, BPmeasurement, hematological and urinary exams,renal clearances and proteinuria to exclude chronichypertension, chronic nephropathy, and major sys-temic diseases. Gestational age was assessed by lastmenstrual phase and ultrasound. Each subject wasre-evaluated at the end of pregnancy to confirmnormotension.

The comparison between a BP circadian profileand the chronodesm, appropriate for gestationalperiod, allowed us to calculate the hyperbaric index(HBI) (25). It corresponds to the total ‘area’ of

Table II. Incidence of PIH/PE in primigravidae and multigravidae in relation with the different risk factors

Primigravidae Multigravidae

Remained Became Remained BecameRisk factors normotensive hypertensive normotensive hypertensive

Family history of eclampsia 5 1 3 1Previous preeclampsia – – 54 8Age ±40 years 4 2 11 –Twin pregnancy 4 3 6 2

Total 13 6 74 11

Table III. MESOR of the population mean cosinor, and mean values of the hyperbaric index (HBI) and percent time elevation (PTE) of systolic (SBP) and diastolicblood pressure (DBP), before 25 weeks of gestation, in patients who remained normotensive until post-partum and in those who became hypertensive

HBI PTEMESOR∫s.e.m. (x∫s.e.m.) (x∫s.e.m.)

Wks mmHg mmHg/24 h %

Remained 8–16 SBP 100.4∫1.1 5.1∫2.2 3.3∫1.0normotensive DBP 59.8∫0.8 1.1∫0.8 0.8∫0.5nΩ87 20–25 SBP 98.1∫0.8 2.3∫0.6 1.9∫0.4

DBP 58.7∫0.7 0.4∫0.2 0.4∫0.2

Become 8–16 SBP 108.4∫2.5* 7.9∫2.5 7.8∫3.3*hypertensive DBP 64.7∫2.3 0.5∫0.3 0.9∫0.5nΩ17 20–25 SBP 111.1∫1.9* 37.5∫13.4* 18.2∫5.3*

DBP 66.9∫1.8* 5.6∫2.1* 6.1∫2.2*

* p∞0.05, vs Remained normotensive (in corresponding wks). Statistical differences were calculated as described in the Methods. SBP, systolic blood pressure;DBP, diastolic blood pressure.


BP above the upper limit of the chronodesm (unitmmHg/24h) and it can be considered a measure ofthe total pressure load exerted upon the arterialwall. Another important index that we calculatedto evaluate the load is the percent time elevation(PTE) (25) that indicates the percent of time dur-ing which the BP values of a patient exceed thethreshold constituted by the upper limit of the rel-evant chronodesm (Fig. 1B). These two indices,HBI and PTE, expressed as two simple numbers,give information of the dynamic evaluation of the24-hour monitoring of blood pressure to be usedfor comparison between subjects or in the samesubject in different clinical situations.

The results were analyzed using SPSS, the Stat-istical Package for Social Science.

Statistical differences among the variablesstudied in the two groups of subjects, i.e. womenwho remained normotensive and those who be-came hypertensive, were separately processed foreach gestational period considered. Since severalvariables were not normally distributed, we em-ployed the non-parametric test for two indepen-dent samples of Mann-Whitney.

Sensitivity, specificity, and predictive values wereassessed by the method of Vecchio (26). The bestcut-off values for systolic and diastolic MESOR,


HBI and PTE were calculated using the receiver op-erating characteristics ( ROC ) curves method (27).

Results

Table II illustrates the incidence of PIH or PE inrelation with the different risk factors. None of thesubjects developed clinically overt hypertension,diagnosed following traditional criteria (21), be-fore 30 weeks of gestation. The gestational age atdelivery and the newborn weight in women whoremained normotensive (nΩ87) were 38∫0.2 (wks,mean∫s.e.m.) and 3,123∫63 (g, mean∫s.e.m.), re-spectively, whereas in those who developed hyper-tension (nΩ17) they were 37∫0.8 and 2,918∫227.

We analyzed a total of 9,984 BP measurements.In no instances DBP, recorded automatically,measured zero.

Chronobiological parameters (i.e. the MESORof the population mean cosinor, and the meanvalues of the HBI and PTE), resulting from theanalysis of the automatic measurements in the 24-hour period are reported in Table III. In subjectswho later developed overt hypertension systolicMESOR and PTE, at the very beginning of preg-nancy (8–16 weeks), were already significantlyhigher than in those who remained normotensiveuntil post-partum. Between 20–25 weeks of ges-tation systolic and diastolic MESOR, HBI, PTEwere all significantly higher in subjects who laterdeveloped hypertension than in those who re-mained normotensive.

No significant differences in amplitude, acro-

Table IV. Sensitivity, specificity, and positive and negative predictive values of different systolic (SBP) and diastolic (DBP) MESOR and hyperbaric index (HBI) at8–16 and 20–25 weeks of gestation as predictors of pregnancy-induced hypertension/preeclampsia

Positive Negativepredictive predictive

Cut-off Sensitivity Specificity value value

8–16 weeks of gestationMESOR SBP 108 mmHg 60% 91% 43% 95%

110 mmHg 40% 93% 40% 93%MESOR DBP 61 mmHg 80% 59% 18% 96%

62 mmHg 80% 68% 22% 97%HBI SBP 5 mmHg/24 h 80% 77% 29% 97%

10 mmHg/24 h 40% 86% 25% 92%HBI DBP 1 mmHg/24 h 40% 86% 25% 93%

1.5 mmHg/24 h 20% 92% 25% 91%

20–25 weeks of gestationMESOR SBP 103 mmHg 88% 75% 45% 96%

104 mmHg 82% 79% 48% 95%MESOR DBP 60 mmHg 82% 55% 30% 93%

61 mmHg 82% 55% 37% 94%HBI SBP 5 mmHg/24 h 71% 69% 54% 93%

10 mmHg/24 h 70% 92% 67% 93%HBI DBP 2 mmHg/24 h 41% 96% 70% 87%

3 mmHg/24 h 15% 97% 75% 86%


phase and percentage of women with significantBP rhythm were observed between the two groups.

Table IV reports the sensitivity, specificity, andpositive and negative predictive values of differentsystolic and diastolic single cosinor MESOR, andHBI at 8–16 and 20–25 weeks of gestation as predic-tors of PIH/PE. The best sensitivity and specificityseem to be obtained between 20–25 weeks of ges-tation with systolic single cosinor MESOR and HBIusing as cut-off 103 mmHg and 10 mmHg/24h, re-spectively. No difference in the distribution betweensubjects who developed PIH (nΩ10) and those whodeveloped PE (nΩ7) was observed in relation withthe cut-off values employed (Fig. 2).

Discussion

In this study, we assessed the utility of 24-hour BPmonitoring in the early detection of PIH or PE.We decided to recruit subjects at moderate/highrisk for PIH or PE in order to enhance the inci-dence of the main outcome measure (i.e. the ap-pearance of PIH or PE), thus reducing the size ofthe sample required for an adequate power of thestudy. In our group the incidence of PIH or PEwas 16%: this allows us to compare our positivepredictive values with most of those reported in theliterature (14, 18, 28–32).

We have previously observed that in outpatientsSBP and DBP MESORs tend to be significantlyhigher than in hospitalized subjects, suggestingthat ambulatory and hospital 24-hour BP monitor-ing in pregnancy may lead to different results (33).


Fig. 2. Individual values of systolic MESOR and hyperbaricindex (HBI) at 20–25 weeks of gestation in women who re-mained normotensive until post-partum (M), and in those whodeveloped pregnancy-induced hypertension (S) or preeclampsia(P). The interrupted lines represent the cut-off which give thebest sensitivity, specificity, and positive and negative predictivevalues.

Although ambulatory monitoring may be less ex-pensive and more easily acceptable to the patients,we decided to perform the 24-hour BP monitoringin hospitalized subjects to avoid variability and bi-ases due to environmental stressors and physicalactivity.

Our results demonstrate that pregnant womenwho will later develop clinically overt hypertensionshow significantly higher values of SBP and DBPMESOR, HBI and PTE at 20–25 weeks of ges-tation, than subjects who will remain normoten-sive until post-partum. Moreover, in the first groupthe systolic MESOR is already significantly greaterat 8–16 weeks of pregnancy.

These data are in agreement with previousstudies, based on standard measurements of BP(12, 13, 30, 34, 35) or on 24-hour ambulatory BPmonitoring (18), during the first and/or second tri-mester of pregnancy, in which the mean arterialpressure (12, 13, 18, 30, 34) or SBP and DBP (13,18, 35) were significantly higher in women wholater developed PE.

At 20–25 weeks of gestation, a SBP MESOR±103 mmHg has a sensitivity for predicting PIHor PE of 88% and a specificity of 75%, a systolic


HBI ±10 mmHg/24h has a sensitivity of 70% anda specificity of 92%. The respective values for aDBP MESOR ±60 mmHg are 82% and 55%, andfor a diastolic HBI ±2 mmHg/24h are 41% and96% (Table IV).

The positive predictive values of systolic and di-astolic MESOR and HBI are not significantlyhigher than those reported in the literature and ob-tained with traditional sphygmomanometric meas-urements (12, 13, 28–32, 34, 35). However, the cal-culation of the parameters based on 24-hour BPmonitorings could supply the clinician with moreobjective cut-offs, which are not influenced by thewell known bias of the traditional technique of BPmeasurement due to subjective evaluation. Thismay be particularly important when it must be de-cided if the single patient is at real risk and there-fore has to be included in a clinical protocol ofantenatal care. Indeed, our data suggest that ad-ditional biochemical or biophysical predictivemarkers of PIH or PE are still needed to achievebetter sensitivity and positive predictive value.

We did not observe a different distribution be-tween subjects who developed PIH and those whodeveloped PE in relation with the above mentionedcut-off values (Fig. 2). This finding does not sup-port the statement made, with regard to mean ar-terial pressure, by Chesley and Sibai (36) that if anincreased second-trimester level predicts anything,it is transient hypertension rather than PE.

In our series, only two patients developed severePE (BP ±160/110 mmHg; proteinuria ±3 g/24hour): for this reason we could not evaluate theutility of chronobiological parameters for the pre-diction of the severe forms of the disease.

In normal pregnancy, as in the non pregnantstate, both SBP and DBP have a significant circa-dian rhythm, characterized by a decrease duringthe night. It has been previously observed that inPIH or PE the rhythm of BP may be lost, due toa less pronounced nocturnal fall of BP values(15,37). In this study, we did not observe a bluntingof the nocturnal BP decrease before the onset ofclinically overt hypertension as already pointed outby Kyle et al. (18). This suggests that the circadianrhythm of BP is not grossly altered before the ap-pearance of clinical disease.

Our data stress the importance of SBP measure-ments in the prediction of PE: indeed, systolicMESOR is significantly higher, as early as 8–16weeks of gestation, in patients who will later be-come hypertensive, and at 20–25 weeks systolicMESOR, and HBI have a greater sensitivity thanthe corresponding diastolic parameters. This couldbe due to higher plasma concentrations of nor-adrenaline early in pregnancy in women who willdevelop PE. However, Tunbridge and Donnai ob-


served that patients who developed hypertensionin late pregnancy have plasma noradrenaline levelssignificantly lower than controls at the same stageof gestation (38). It has been suggested that araised sympathetic activity increases cardiac out-put, which in turn could augment peripheral vas-cular resistance. Then, after the onset of clinicallyovert hypertension, cardiac output and noradren-aline concentration could subsequently decrease(39). At this regard, Schobel et al. (40) have re-cently found with a microneurographic techniquethat sympathetic vasoconstrictor discharge to so-matic muscle is markedly elevated in patients withPE but is normal in normotensive pregnant sub-jects. These findings suggest that the increase inperipheral vascular resistance seen in PE may bemediated, at least partly, by a substantial increasein sympathetic vasoconstrictor activity.

In any case, DBP MESOR, HBI, and PTE arealso significantly increased in women who will laterdevelop PIH or PE. Therefore, it can be hypo-thesized that these patients do not undergo thephysiologic vasodilatation typical of normal preg-nancy and that their peripheral-vascular tone is in-appropriately elevated throughout gestation eitherbecause of abnormal sensitivity of their vascula-ture to circulating substances or because of a de-creased production of vasodilating stimuli.

In conclusion, based upon our experience, wethink that the chronobiological analysis applied to24-hour BP monitoring during pregnancy allowsus to define more objective cut-off values thanthose obtained with traditional sphygmomano-metric measurements. This may be particularly im-portant in the routine clinical practice, when therisk of developing PIH or PE must be calculatedfor each individual subject.

Acknowledgments

Supported by the Italian Ministry for University and Scientificand Technological Research (MURST 40%) and the Italian Na-tional Research Council (CNR) – targeted project ‘Preventionand Control Disease Factor’, subproject ‘Maternal-Infant Dis-ease’, contract no. 9300691.PF41.

We thank Professor Alberto Angeli for his encouragement.We feel indebted to Dr. Marco Acutis for his invaluable statisti-cal advice. We are also grateful to Ethicon S.p.A for technicalsupport.

References

1. Department of Health. Report on confidential enquiriesinto maternal deaths in the U.K. 1988–1990. London,HMSO, 1994.

2. Brans YW, Escobedo MB, Hayashi RH, Huff RW, Kagan-Hallet KS, Ramamurthy RS. Perinatal mortality in a largeperinatal center: Five-year review of 31,000 births. Am JObstet Gynecol 1984; 148: 284–9.

3. Kaunitz AM, Hughes JM, Grimes DA, Smith JC, Rochat


RW, Kafrissen ME. Causes of maternal mortality in UnitedStates. Obstet Gynecol 1985; 65: 605–12.

4. Redman CWG. Pathophysiology of pre-eclampsia and ec-lampsia – why is it dangerous. In: Maternal Mortality –The way forward. N. Patel Ed. Royal College of Obste-tricians and Gynaecologists, 1992; 14–20.

5. National high blood pressure education program workinggroup report on high blood pressure in pregnancy. Am JObstet Gynecol 1990; 163: 1689–712.

6. McMullan PF, Norman RJ, Marivate M. Pregnancy-in-duced hypertension in twin pregnancy. Br J Obstet Gynae-col 1984; 91: 240–3.

7. Campbell DM, MacGillivray I, Carr-Hill R. Pre-eclampsiain second pregnancy. Br J Obstet Gynaecol 1985; 92: 131–40.

8. Chesley LC, Cooper DW. Genetics of hypertension in preg-nancy: Possible single gene control of pre-eclampsia andeclampsia in the descendants of eclamptic women. Br J Ob-stet Gynaecol 1986; 93: 898–908.

9. Cunningham FG, Cox SM, Harstad TW, Mason RA, Prit-chard JA. Chronic renal disease and pregnancy outcome.Am J Obstet Gynecol 1990; 163: 453–9.

10. Parazzini F, Bortolus R, Chatenould L, Restelli S, Ricci E,Marozio L et al. Risk factors for pregnancy-induced hyper-tension in women at high risk for the condition. Epidemi-ology 1996; 7: 306–8.

11. Dekker GA, Sibai BM. Early detection of preeclampsia.Am J Obstet Gynecol 1991; 165: 160–72.

12. Page EW, Christianson R. The impact of mean arterialpressure in the second trimester upon the outcome of preg-nancy. Am J Obstet Gynecol 1976; 125: 740–6.

13. Moutquin JM, Rainville C, Giroux L, Raynauld P, AmyotG, Bilodeau R et al. Prospective study of blood pressure inpregnancy: prediction of preeclampsia. Am J Obstet Gyn-ecol 1985; 151: 191–6.

14. Villar MA, Sibai BM. Clinical significance of elevated meanarterial blood pressure in second trimester and thresholdincrease in systolic or diastolic blood pressure during thirdtrimester. Am J Obstet Gynecol 1989; 160: 419–23.

15. Benedetto C, Zonca M, Marozio L, Dolci C, CarandenteF, Massobrio M. 24-hour blood pressure patterns in physi-ologic pregnancy and in pregnancy-induced hypertension,preeclampsia and chronic hypertension. Obstet Gynecol1996; 88: 503–10.

16. Benedetto C, Zonca M, Giarola M, Maula V, Chiarolini L,Carandente F. 24-hour blood pressure monitoring for theevaluation of nifedipine effects in preeclampsia and inchronic hypertension in pregnancy. Br J Obstet Gynaecol1997; 104: 682–8.

17. Carandente F, Ahlgren A, Halberg F. Mesor-hypertension:hints by chronobiologists. Chronobiologia 1984; 11: 189–94.

18. Kyle PM, Clark SJ, Buckley D, Kissane J, Coats AJ, DeSwiet M et al. Second trimester ambulatory blood pressurein nulliparous pregnancy: a useful screening test for preec-lampsia? Br J Obstet Gynaecol 1993; 100: 914–19.

19. Halligan A, O’Brien E, O’Malley K, Mee F, Atkins N, Con-riy R et al. Twenty-four hour ambulatory blood pressuremeasurement in a primigravid population. J Hypertens1993; 11: 869–73.

20. Italian study of aspirin in pregnancy. Low-dose aspirin inprevention and treatment of intrauterine growth retar-dation and pregnancy-induced hypertension. Lancet 1993;341: 396–400.

21. Milsom I, Svahn S-O, Forssman L, Sivertsson R. An evalu-ation of automated indirect blood pressure measurementsduring pregnancy. Acta Obstet Gynecol Scand 1986; 65:721–5.

22. Franx A, Van der Post JAM, Elfering IM, Veerman DP,


Merkns HM, Boer K et al. Validation of automated bloodpressure recording in pregnancy. Br J Obstet Gynaecol1994; 101: 66–9.

23. Davey DA, MacGillivray I. The classification and defi-nition of the hypertensive disorders of pregnancy. Clin andExper Hyper – Hyper in Pregnancy 1986; B5 (1): 97–133.

24. Nelson W, Tong YL, Lee JK, Halberg F. Method for cosin-or-rhythmometry. Chronobiologia 1979; 6: 305–23.

25. Carandente F, Halberg F. Chronobiology of blood pressurein 1985. Chronobiologia 1984; 11: 189–321.

26. Vecchio TJ. Predictive value of a single diagnostic test inunselected populations. N Engl J Med 1966; 274: 1171–3.

27. Richardson DK, Schwartz JS, Weinbaum PJ, Gabbe SG.Diagnostic tests in obstetrics: a method for improved evalu-ation. Am J Obstet Gynecol 1985; 152: 613–19.

28. Phelan JP. Enhanced prediction of pregnancy-induced hy-pertension by combining supine pressor test with mean ar-terial pressure of middle trimester. Am J Obstet Gynecol1977; 129: 397–402.

29. Quaas L, Robrecht D, Kaltenbach FJ. The mean arterialpressure versus roll-over test as predictors of hypertensionin pregnancy In: Bonnar J, MacGillivray I, Symonds EM.Pregnancy Hypertension. Lancaster, MTP Press. 1980;113–18.

30. Oney T, Kaulhausen H. The value of the mean arterialblood pressure in the second trimester (MAP-2 value) as apredictor of pregnancy-induced hypertension and pre-ec-lampsia. A preliminary report. Clin Exp Hypertens 1983[B]; 2: 211.

31. Marya RK, Rathee S, Mittal R. Evaluation of three clinicaltests for predicting pregnancy-induced hypertension. Am JObstet Gynecol 1988; 158: 683–7.

32. Conde-Agundelo A, Belizan JM, Lede R, Bergel EF. Whatdoes an elevated mean arterial pressure in the second halfof pregnancy predict: gestational hypertension or preec-lampsia? Am J Obstet Gynecol 1993; 169: 509–14.

33. Benedetto C, Valensise H, Marozio L, Tranquilli A, Chiaro-lini L, Paesano PG. Ambulatory and hospital 24-hour


blood pressure monitoring in pregnancy: a comparison be-tween these two procedures. Ital J Gynaecol Obstet 1996;4: 129–34.

34. Fallis NE, Langford HG. Relation of second trimesterblood pressure to toxaemia of pregnancy in the primigravidpatient. Am J Obstet Gynecol 1963; 87: 123–5.

35. Gallery EDM, Hunyor SN, Ross M, Gyory AZ. Predictingthe development of pregnancy-associated hypertension.The place of standardized blood-pressure measurement.The Lancet 1977; i : 1273–5.

36. Chesley LC, Sibai BM. Clinical significance of elevatedmean arterial pressure in the second trimester. Am J ObstetGynecol 1988; 159: 275–9.

37. Miyamoto S, Shimokawa H, Sumioki H, Touno A, NakanoH. Circadian rhythm of plasma atrial natriuretic peptide,aldosterone and blood pressure during the third trimesterin normal and preeclamptic pregnancies. Am J Obstet Gyn-ecol 1988; 158: 393–9.

38. Tunbridge RDG, Donnai P. Plasma noradrenaline in nor-mal pregnancy and in hypertension of late pregnancy. Br JObstet Gynaecol 1981; 88: 105–8.

39. Tunbridge RDG, Donnai P. Pregnancy-associated hyper-tension, a comparison of its prediction by ‘roll-over test’and plasma noradrenaline measurement in 100 primigra-vidae. Br J Obstet Gynaecol 1983; 90: 1027–32.

40. Schobel HP, Fischer T, Heuszer K, Geiger H, SchmiederRE. Preeclampsia: a state of sympathetic overactivity. NEngl J Med 1996; 335: 1480–5.

Address for correspondence:

Chiara Benedetto, M.D., Ph.D.Dipartimento di Discipline Ginecologiche ed OstetricheUniversita di TorinoVia Ventimiglia 310126 TorinoItaly

twenty-four hour blood pressure monitoring in early pregnancy: is it predictive of pregnancy-induced...

Documents