International Journal of Epidemiology 1999;28:253257

Prolonged pregnancy and the risk of post-maturity are of majorinterests for obstetricians. A prolonged pregnancy is commonlydefined as a gestational age of 42 completed weeks or more (i.e.294 days or more). Prolonged pregnancy is not synonomouswith post-maturity, which is a clinical syndrome thought to bea consequence of a failing placental function.1,2 Several eventsand outcomes of pregnancy have a familial background, e.g.low birthweight-for-gestational-age,3,4 pre-term delivery,5 andpre-eclampsia.6,7 Only a few studies have been performed con-cerning the correlation of gestational age across generations. Ina couple of studies a low correlation for all gestational age across generations was reported.8,9 A recent study indicates atrend of increasing number of post-term births as the gesta-tional age at mothers own birth increases.10 A Norwegian studyfound a relative risk of 2.2 in subsequent birth if the first birthwas post term.11 The main objective in our study was to test thehypothesis of recurrence of prolonged pregnancy across gener-ations which has not, to our knowledge, yet been performed

and also the recurrence of prolonged pregnancy in subsequentbirths.

MethodsSubjects

Data in the Medical Birth Registry at the National Board ofHealth and Welfare in Sweden have been collected since 1973.The register is based on standardized collected data from maternityhealth centres, delivery units and paediatric examinations on all pregnancies ending with a delivery. Approximately 1% of alldeliveries are missing from the register.12 The information islinked to the unique personal identification number of the childand the mother. Data from ledgers at delivery units in the countiesof Vsterbotten and Vsternorrland (northern Sweden), occur-ring in 1955 to 1972, were collected by a unit connected toUme University. This register includes 123 819 births. Informa-tion on all deliveries of inhabitants in the area was extractedfrom the Medical Birth Registry of 1973 to 1990. This data setwas added to the register of 1955 to 1972 yielding a total cohortof 248 907 births.

In the cohort 49 439 mother-daughter pairs (with events of delivery for both generations) were identified. Pairs, whereeither mother or daughter was part of a multiple birth, were

International Epidemiological Association 1999 Printed in Great Britain

Recurrence of prolonged pregnancyIngrid Mogren,a,b Hans Stenlundb and Ulf Hgberga,b

Background We conducted a cohort study in an attempt to determine whether prolongedpregnancy in mother is a risk factor for prolonged pregnancy in daughter, and ifprevious prolonged pregnancy is a risk factor for prolonged pregnancy in sub-sequent pregnancy.

Methods Data from the Swedish Medical Birth Registry were combined with a local registryof births (19551990). Mother-daughter pairs (with events of delivery in eachgeneration) were identified. Relative risk (RR) and its 95% confidence interval(CI) were calculated and population attributable proportion was estimated whenappropriate.

Results If mother had had prolonged pregnancy at delivery of daughter the relative risk(RR) of prolonged pregnancy in daughter was moderately raised (RR = 1.3;CI : 1.01.7) with population attributable proportions ranging between 2.1% and4.6%. If previous pregnancy had been prolonged, the RR of prolonged pregnancyat subsequent birth was increased 23 fold with population attributable propor-tions of 12.5% to 15.8%. Possible confounders such as mothers parity, age andmaternal age did not alter the risks.

Conclusions Although moderate, prolonged pregnancy in mother may be a risk factor forprolonged pregnancy in daughter. A previous prolonged pregnancy increases therisk of prolonged pregnancy in a subsequent birth. However, the familial factorof prolonged pregnancy explains just a minor part of its occurrence in the popula-tion (due to small population attributable proportions).

Keywords Prolonged pregnancy, recurrence, relative risk, cohort, familial

Accepted 7 September 1998

a Department of Obstetrics and Gynecology, Ume University, Ume, Sweden.b Department of Epidemiology and Public Health, Ume University, Ume,

Sweden.

253

excluded thereafter, resulting in a net sample of 48 076 mother-daughter pairs. Secondly, each of two daughters and their firstand sometimes second deliveries (either sex) were identified.Information on last menstrual period was only available for about10% of deliveries in mother-generation, whereas informationon gestational age was nearly complete for deliveries of daughters(Figure 1). Cases with gestational age .45 completed weekswere excluded as probable errors in the mother-generation; all deliveries in the daughter-generation were 42 completed weeks (i.e. >294 days). Term pregnancyincluded gestational age of 3741 completed weeks (259293days), and consequently, pre-term pregnancy was defined asgestational age ,37 completed weeks (or ,259 days). The lastmenstrual period and the date of birth were the basis for assign-ment of gestational age in all subjects of mother-generation.However, the use of ultrasound for gestational length assess-ment was generally introduced during the period of 19841992in different clinics involving the majority (62.7%) of pregnanciesin daughter-generation.

The present study is a cohort study of two birth-giving gener-ations. The relative risk (RR) and its 95% confidence interval (CI)were calculated for the following three comparisons and thepopulation attributable proportion (= p(RR 1)/[1 + p(RR 1)],p = the proportion exposed in the population)which is theproportion of cases in the population that should not haveoccurred, had the exposed had the incidence of the unexposedwas calculated when appropriate:

(1) The risk for the daughters to have prolonged pregnancy ifmother had had prolonged pregnancy at the delivery of thedaughter.

(2) The risk of a prolonged subsequent pregnancy ifprolonged pregnancy had occurred previously.

(3) The risk of prolonged pregnancy if sister had hadprolonged pregnancy.

The reference group in calculation of relative risk of pro-longed pregnancy was subjects with term pregnancy. The datawere also analysed with logistic regression.

Statistical analysis was done using the Statistical Package forSocial Sciences, Version 7.0 (SPSS, 1995) and Epi Info, Version6.03 (1996).

ResultsThe prevalence of prolonged pregnancy was higher in themother-generation compared to the daughter-generation (Table 1). Range of gestational age and maternal age is shownin Tables 1 and 2. The relative risk for the daughter to haveprolonged pregnancy if mother had had prolonged pregnancy at the delivery of the daughter was moderately raised in themajority of comparisons and the population attributable pro-portions were small (Table 3). The relative risk of prolongedpregnancy at subsequent pregnancy if prolonged pregnancyoccurred at previous pregnancy was increased 23 fold indaughters and the population attributable proportions of pro-longed pregnancy were 12.515.8% (Table 4). The risk of pro-longed pregnancy in the younger daughter was elevated if theelder daughter had had prolonged pregnancy in either of herdeliveries (Table 5).

To study the influence of potential confounders, such asmothers age (below or above 35 years of age), parity of motherand age of daughter, we used logistic regression. However, therisks were unaltered using such analyses. The relative risk ofprolonged pregnancy among daughters if mother had hadprolonged pregnancy at both of her deliveries could not becalculated because there were too few subjects.

254 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

Figure 1 Deliveries of two generations

RECURRENCE OF PROLONGED PREGNANCY 255

Table 1 Distribution of gestational age and maternal age (median, 10th and 90th percentiles) in specified pregnancy (comparison betweenmother and daughter)

Maternal age (years)

Variable (number) Term number (%) Post-term number (%) 10th Median 90th

Mother-elder daughter pair:

Mothers delivery of elder daughter (n = 2339) 1837 (78.5) 348 (14.9) 20 25 37

Elder daughters first delivery (n = 2339) 1934 (82.7) 264 (11.3) 19 24 28

Elder daughters second delivery (n = 1349) 1176 (87.2) 116 (8.6) 22 26 30

Mother-younger daughter pair:

Mothers delivery of younger daughter (n = 274) 223 (81.4) 33 (12.0) 21 28 37

Younger daughters first delivery (n = 274) 224 (81.8) 30 (10.9) 19 23 28

Younger daughters second delivery (n = 144) 126 (87.5) 12 (8.3) 22 25 29

Table 2 Distribution of gestational age and maternal age (median, 10th and 90th percentiles) in specified pregnancy (comparison between firstand subsequent pregnancy in the individual)

Maternal age (years)

Variable (number) Term number (%) Post-term number (%) 10th Median 90th

Elder daughters

First delivery (n = 22 617) 18 474 (81.7) 2678 (11.8) 19 23 29

Second delivery (n = 14 249) 12 334 (86.6) 1290 (9.1) 22 26 31

Younger daughters

First delivery (n = 2949) 2385 (80.9) 341 (11.6) 19 23 25

Second delivery (n = 1642) 1426 (86.8) 136 (8.3) 22 25 29

Table 3 Comparison between mother and daughter. Relative risk (RR), its 95% confidence interval (CI) and population attributable proportion ofprolonged pregnancy at elder and younger daughters deliveries in relation to prolonged pregnancy at mothers delivery of the daughter

Prevalence of prolonged pregnancy Population attributable

Variable (number) RR 95% CI Exposeda Unexposedb proportion

Elder daughters

First delivery (n = 2339) 1.3 1.01.7 14.1% 10.8% 4.6%

Second delivery (n = 1349) 1.1 0.71.8 9.8% 8.7% 2.1%

Younger daughters

First delivery (n = 274) 0.8 0.32.6 9.1% 10.8%

Second delivery (n = 144) 1.5 0.46.4 11.8% 7.8% 4.5%

a Exposed = daughter whose mother had prolonged pregnancy with daughter.b Unexposed = daughter whose mother did not have prolonged pregnancy with daughter.

Table 4 Comparison in the individual. Relative risk (RR), its 95% confidence interval (CI) and population attributable proportion of prolongedpregnancy at subsequent pregnancy if prolonged pregnancy occurred in previous pregnancy

Prevalence of prolonged pregnancy Population attributable

Variable (number) RR 95% CI Exposeda Unexposedb proportion

Elder daughters

Second delivery (n = 14 147) 2.6 2.32.9 19.9% 7.7% 15.8%

Younger daughters

Second delivery (n = 1635) 2.2 1.53.1 16.4% 7.6% 12.5%

a Exposed = previous pregnancy prolonged.b Unexposed = previous pregnancy not prolonged.

DiscussionAlthough moderate prolonged pregnancy in mother may be a risk factor for prolonged pregnancy in daughter. A previousprolonged pregnancy increases the risk of prolonged pregnancyoccurring at a subsequent birth.

The question of how to determine accurately the expecteddate of delivery is of great importance in the clinical manage-ment of fetal well-being, fetal growth and events at delivery and during the neonatal period. The issue of the normal lengthof singleton pregnancy is not yet settled.1319 The last men-strual period has historically been the basis for assigning gesta-tional age. There is a great variation in cycle length, not onlybetween women, but also, within the individual. Furthermore,the variability in cycle length is more pronounced for youngerwomen.16 The curve of conception in relation to last menstrualperiod shows a skewness towards late conception.20 Mothersaged >35 tend to give birth 2 days earlier than those ,35 whichmight reflect a shortening in cycle length with increasing age.21

Ultrasonic fetal biometry can be used to estimate gestationallength if performed before mid-pregnancy.19

In the present study gestational age was assigned by lastmenstrual period in all subjects in the mother-generation butonly in part in the daughter-generation. Gestational length wasultrasound corrected in 62.7% of pregnancies in the daughter-generation. The prevalence of prolonged pregnancy differedbetween mother-generation and daughter-generation (Table 1),which might be due to the introduction of ultrasound techniqueand so-called term-correction.16,22

In a literature review, the incidence rates for post-term preg-nancy ranged from 4% to 14%, with an average of about 10%.11

The level of prolonged pregnancy in our study complies withthese figures and was also similar to the findings in a study ofapproximately 427 000 Swedish singleton births, where 12.6%of women in the total data set or 10% in a subset (with gesta-tional age assigned by certain menstrual dates) gave birth postterm.21

The final sample of comparisons across generations consistedof mothers delivered in the Department of Obstetrics and Gyn-ecology in Umethe residential town in Vsterbotten county,which receives admissions from the rest of the region. The

daughters were delivered in all 19 delivery units in Vsterbottenand Vsternorrland counties. The rate of induction of labour orcaesarean section (in the mother-generation) might be elevatedbecause of admissions of complicated pregnancies and theseconditions could therefore influence inter-generational correla-tions of prolonged pregnancy. The majority of daughters in theanalysis of recurrence of prolonged pregnancy in subsequentbirths had mothers who delivered somewhere other than Ume.

To evaluate possible selection bias in the mother-generationwe used the data set prepared for comparisons across genera-tions and calculated the relative risk of prolonged pregnancy atsubsequent delivery for elder daughter if previous delivery wasprolonged (RR = 2.2; CI : 2.43.2). This obviously did not differto any major degree from the larger data set. Thus, we do notconsider any major selection bias in the mother-generation.

The accuracy of records in the Swedish Birth Registry hasbeen investigated earlier.23,24 A quality control of data collected1974 and 1986, was performed, where the quality of each itemin the register was independently assessed by the authors witha scoring scale of 13 (1 = poor, 2 = acceptable, 3 = good). Dataon personal identification number, infants personal identi-fication number, date of birth, single/multiple birth and order inbirth were scored as good, whereas the number of previouspregnancies and diagnoses during pregnancy were scored as ac-ceptable.23 A quality control of the contents in the local registryof births (Vsterbotten and Vsternorrland counties) has beenperformed.25

Several events and outcomes of pregnancy have a familialcomponent. Pre-eclampsia in pregnancy of mother is a riskfactor for development of pre-eclampsia in pregnancy ofdaughter.6,7,26 Newborns with low birthweight-for-gestational-age are more likely to have mothers who themselves had lowbirthweight-for-gestational-age.3,4 There is also a tendency ofrepeating reproductive outcome in subsequent deliveries, e.g.pre-term delivery, gestational age5 and low birthweight.27

Different factors such as twins, previous history of perinataldeath, rhesus haemolytic disease, pre-eclampsia and recurrenthaemorrhage may all act as confounding variables and reducethe likelihood of being delivered post term.2 In our study onlysingleton deliveries in both generations were included. Post-term pregnancy is associated with an excess incidence of con-genital malformation compared with delivery at term.2 Parityhas also been shown to influence the proportion of post-termbirth with a considerable drop between primiparity and multi-parity,11 but our study indicated just a moderate drop of theprevalence of prolonged pregnancy at subsequent birth (Table 2).

Earlier research has reported low correlation of all gestationalage across generations (,0.1).9 However, a recent study haspresented results suggesting that there is a weak genetic effectof parental gestation on gestational age of 37 weeks or more.10

In our study we found a moderate excess risk of prolongedpregnancy across generations in almost all comparisons andsmall attributable proportions overall (Table 3). Exceptionally, ayounger daughter presented a reduced risk of prolonged preg-nancy at her first delivery if her mother had had prolongedpregnancy at delivery of her younger daughter (Table 3). How-ever, the risk for younger daughter at first delivery was elevatedif mother had had a prolonged delivery with her elder daughter(RR = 2.4; CI : 1.06.0, not included in Tables). The relative riskof recurrence of prolonged pregnancy in subsequent birth was

256 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

Table 5 Comparison between sisters. Relative risk (RR), and its 95%confidence interval (CI) of prolonged pregnancy in younger daughter if elder daughter had prolonged pregnancy

Variable RR 95% CI

If prolonged pregnancy occurred at elder daughters firstdelivery:

Younger daughters

First delivery 1.4 (1.4a) 1.11.9 (1.02.1a)

Second delivery 1.2 (1.1a) 0.71.9 (0.62.0a)

If prolonged pregnancy occurred at elder daughters seconddelivery:

Younger daughters

First delivery 1.4 (1.7a) 1.01.9 (1.12.5a)

Second delivery 0.7 (0.7a) 0.31.5 (0.31.7a)

a Ultrasound corrected pregnancies in daughter-generation.

considerable and a 23 fold risk is apparently in accordancewith a relative risk of 2.2 presented by Bakketeig and Bergsj.11

The authors also found a lower perinatal mortality rate amongbabies of mothers who repeated post-term births compared to mothers whose previous birth was not post term. We alsoinvestigated the perinatal and neonatal mortality among oursubjects and found no obvious differences between the adverseoutcome of term and post-term pregnancies. Furthermore, the small mortal number prevented further investigation ofprolonged pregnancy as a risk factor for adverse outcome insubsequent delivery.

ConclusionsAn accurate gestational age is most important in clinical practicein directing attention to the perceived risks to the fetus of thepost-mature syndrome or in avoiding unnecessary intervention.Earlier research has not investigated whether a risk for thedaughter of having prolonged pregnancy was related to themother having had prolonged pregnancy at the delivery of the daughter. Our findings, however, indicated a moderateexcess of prolonged pregnancy across generations and alsoassociations between deliveries in a family, although further re-search is required to settle this issue. We confirm earlier resultsthat the relative risk of prolonged pregnancy in a subsequentdelivery is approximately doubled. The population attributableproportions are small concerning recurrence of prolongedpregnancy across generations, but it is elevated in the individualwhen comparing previous to subsequent birth. Thus, it seems asif familial predisposition explains just a minor part of the occur-rence of prolonged pregnancy in the population, but the indi-vidual expresses a disposition towards a similar outcome in asubsequent pregnancy which is the case for several other repro-ductive outcomes.

AcknowledgementsThanks to Gran Lindahl, UMDAC, Ume University, EvaJohansson, and the late Jan Gunnarskog, National Board ofHealth and Welfare, Sweden, for data management. Specialthanks to Bengt Haglund, National Board of Health and Welfare,Sweden, for valuable co-operation and alertness. This researchwas supported by grants from the Swedish Council for SocialScience Research (No 95-0043:IB), National Institute of PublicHealth and Medical Faculty, Ume University.

References1 Ahmed AI, Versi E. Prolonged pregnancy. Curr Opin Obstet Gynecol

1993;5:66974.2 Crowley P. Post-term pregnancy: induction or surveillance? In: Enkin

M, Keirse MJ, Chalmers I (eds). Effective Care in Pregnancy and Child-birth. Oxford: Oxford University Press, 1989, pp.77691.

3 Ounsted M, Ounsted C. Rate of intra-uterine growth. Nature 1968;220:599600.

4 Klebanoff MA, Graubard BI, Kessel SS, Berendes HW. Low birthweight across generations. JAMA 1984;252:242327.

5 Bakketeig LS, Hoffman HJ, Harley EE. The tendency to repeatgestational age and birth weight in successive births. Am J ObstetGynecol 1979;135:1086103.

6 Arngrimsson R, Bjornsson S, Geirsson RT, Bjornsson H, Walker JJ,Snaedal G. Genetic and familial predisposition to eclampsia and pre-eclampsia in a defined population. Br J Obstet Gynaecol 1990;97:76269.

7 Sutherland A, Cooper DW, Howie PW, Liston WA, MacGillivray I. The incidence of severe pre-eclampsia amongst mothers and mothers-in-law of pre-eclamptics and controls. Br J Obstet Gynaecol 1981;88:78591.

8 Carr-Hill R, Campbell DM, Hall MH, Meredith A. Is birth weightdetermined genetically? Br Med J 1987;295:68789.

9 Magnus P, Bakketeig LS, Skjaerven R. Correlations of birth weightand gestational age across generations. Ann Hum Biol 1993;20:23138.

10 Hennessy, E, Alberman E. Intergenerational influences affecting birthoutcome. II. Preterm delivery and gestational age in the children of the 1958 British birth cohort. Paediatr Perinat Epidemiol 1998;12:6175.

11 Bakketeig LS, Bergsj P. Post-term pregnancy: magnitude of theproblem. In: Enkin M, Keirse MJ, Chalmers I (eds). Effective Care inPregnancy and Childbirth. Oxford: Oxford University Press, 1989,pp.76575.

12 National Board of Health and Welfare, Centre of Epidemiology,Stockholm, Sweden. The Activities at the Centre of Epidemiology, 1996 (In Swedish).

13 Mittendorf R, Williams MA, Berkey CS, Cotter PF. The length ofuncomplicated human gestation. Obstet Gynecol 1990;75:92932.

14 Saunders N, Paterson C. Can we abandon Naegeles rule? Lancet 1991;337:60001.

15 Saunders N, Paterson C. Effect of gestational age on obstetric perform-ance: when is term over? Lancet 1991;338:119092.

16 Geirsson RT. Ultrasound instead of last menstrual period as the basisof gestational age assignment. Ultrasound Obstet Gynecol 1991;1:21219.

17 Todros T, Roncho G, Lombardo D, Gagliardi L. The length of preg-nancy: an echographic reappraisal. J Clin Ultrasound 1991;19:1114.

18 Kieler H, Axelsson O, Nilsson S, Waldenstrm U. The length of humanpregnancy as calculated by ultrasonographic measurement of thefetalbiparietal diameter. Ultrasound Obstet Gynecol 1995;6:35357.

19 Mongelli M, Opatola, B. Duration and variability of normal preg-nancy. Implications for clinical practice. J Reprod Med 1995;40:64548.

20 Pryll W. Kohabitationstermin und kindsgeschlecht. MuenchenerMedizinische Wochenschrift 1916;45:157982.

21 Bergsj P, Denman III DW, Hoffman HJ, Meirik O. Duration of humansingleton pregnancy. Acta Obstet Gynecol Scand 1990;69:197207.

22 Hgberg U, Larsson N. Early dating by ultrasound and perinataloutcomea cohort study. Acta Obstet Gynecol Scand 1997;76:90712.

23 Cnattingius S, Ericson A, Gunnarskog, J, Klln B. A quality study ofa medical birth registry. Scand J Soc Med 1990;18:14348.

24 National Board of Health and Welfare, Bureau 3. Validation of theInformation in the Swedish Medical Birth Register, 1974. Stockholm, 1977.(In Swedish).

25 Sandstrm A, Nystrm, L. Building up a medical birth registry for the period of 195572. Socialmedicinsk tidskrift 1985;23:9597 (InSwedish).

26 Chesley LC, Annitto JE, Cosgrove RA. The familial factor in toxemiaof pregnancy. Obstet Gynecol 1968;32:30311.

27 Khoury MJ, Calle EE, Joesoef RM. Recurrence of low birth weight insiblings. J Clin Epidemiol 1989;42:117178.

RECURRENCE OF PROLONGED PREGNANCY 257