Examination of the association between male gender and preterm delivery☆

1. Introduction 

Preterm birth remains one of the major problems of pregnancy. Among children registered on the Western Australian cerebral palsy register (1986–1992) 34.7% were born before 37 weeks [1]. There are multiple risk factors but fetal illness and infection appear to be most important mechanisms. In addition, there is a well-established small male excess in preterm births [2]. This is not an artefactual consequence of the accepted increased proportion of males conceived when fertilisation takes place early in the cycle [3]. Indeed, a male ‘disadvantage’ is present for multiple other outcomes too: male fetuses are more likely to miscarry, develop congenital malformations, to be delivered abdominally [4] be born in need of resuscitation and intensive support and to die [5], and to subsequently develop cerebral palsy [6]. Further, different patterns of adult illness and longevity between the sexes are well established.

Several possible reasons have been suggested. Pre-pregnancy risk factors for preterm birth might alter the sex ratio. Sex-linked hormonal differences may be important: Tan et al. [7] demonstrated that male–male pairs of twins were at greater risk of preterm delivery than female–female, or even female–male pairs. Or the larger birth weight of males could be responsible. Infection, which has a major role in preterm birth, might be more common among males. Perhaps most importantly, there is an increased risk of many antenatal obstetric complications with males [8]: this has not been examined as a possible mechanism.

The aim of this study is to investigate possible reasons why males are more likely to be born preterm by examining the putative causal pathways.

2. Materials and methods 

This is a retrospective database review of all pregnancies from 1st January 1992 to 31st December 2002, where delivery occurred in the John Radcliffe Hospital, Oxford. This is a major teaching hospital serving the local population and also pregnancies referred from the surrounding region. Pregnancy and delivery details are recorded prospectively in a database (OXMAT). This information was transferred for analysis to SPSS 14 (Chicago, IL): some additional details on babies admitted to the neonatal unit were gathered from unit records and entered retrospectively. Exclusions to this analysis were pregnancies ending before 24 completed weeks’ gestation, those affected by major congenital abnormalities, pregnancy terminations, multiple pregnancies and those of indeterminate sex. Ethical approval for the use of OXMAT for this study was granted (COREC) in September 2005.

Preterm delivery was defined as that occurring before 37 completed weeks’ gestation and further divided into extreme (24–28 weeks), severe (29–32 weeks) and moderate (33–36 weeks). Initially, demographic, antenatal, intrapartum and outcome variables were compared for preterm male and female neonates. For binary variables the chi-squared test was used (or Fisher's Exact test where appropriate); for continuous variables that were normally distributed Student's t-test was used. Odds ratios (OR) for preterm delivery at different gestation bands were calculated, with 95% confidence intervals (CIs), using the term cohort (37–42 completed weeks) for comparison. Putative simplified causal pathways that might explain the significant associations of fetal sex with preterm delivery (Fig. 1) were drawn up. Binary logistic regression was used to test these pathways by calculating adjusted odds ratios (AORs) with 95% confidence intervals (CIs), for preterm delivery at different gestation bands. Adjustments were performed for potentially causal factors that were associated with male sex and might reasonably account for the sex difference.

View Large Image Download to PowerPoint

Fig. 1. Simplified putative causal pathways underlying the male excess in preterm deliveries.

3. Results 

Total pregnancies analysed numbered 75,725. Of these, 4003 (5.3%) delivered before 37 weeks, and 71,722 (94.7%) after. Males accounted for 38,601 (51%) of total deliveries: 36,447 (50.8%) at term and 21,54 (53.9%) preterm. Males were more likely to deliver preterm, both overall (OR=1.13; 95% CI 1.06–1.20) and in all gestation groups except prior to 28 weeks (Table 2). Table 1 shows the demographic and pregnancy details for preterm deliveries in the cohort according to fetal sex. Table 2, Table 3, Table 4 explore the relationship between fetal sex and preterm delivery using binary logistic regression testing the putative causal pathways as shown in Fig. 1 and detailed below.

Table 1. Demographic and obstetric differences in preterm deliveries according to fetal sex

		Males n/mean (%/S.D.)	Females n/mean (%/S.D.)	p
	Total deliveries	38,601 (51)	37,124 (49)
	Total preterm	2,154 (5.6)	1,849 (5.0)
	Population characteristics
	Maternal age	29.07 (5.89)	29.09 (5.99)	0.921
	Race
	Caucasian	1,952 (90.7)	1,651 (89.3)	0.057
	Afro-Carribean	12 (0.6)	12 (0.6)	0.366
	Asian	106 (4.9)	108 (5.8)	0.069
	Oriental	23 (0.6)	11 (0.6)	0.091
	Mixed	13 (0.6)	23 (1.2)	0.07
	Other	47 (2.2)	44 (2.4)	0.276
	Pre-pregnancy factors
	Nulliparous	1,111 (51.6)	898 (48.6)	0.031
	Diabetes	41 (1.9)	28 (1.5)	0.21
	Endocrine disease	49 (2.3)	48 (2.6)	0.289
	Psychiatric illness	175 (8.1)	140 (7.6)	0.278
	Kidney disease	162 (7.5)	142 (7.7)	0.448
	Heart disease	61 (2.8)	51 (2.8)	0.483
	Thromboembolic disease	28 (1.3)	24 (1.3)	0.555
	Pregnancy factors
	Placental abruption	123 (5.7)	83 (4.5)	0.047
	Pre eclampsia	354 (16.4)	397 (21.5)	<0.001
	Prolonged SROM	440 (26.5)	358 (26.0)	0.405
	Renal tract infection	497 (23.1)	374 (20.2)	0.016
	Birth weight (g)	2,214 (748)	2,084 (731)	<0.001
	Low birth weight (<2500g)a	(59.8%)	1,240 (67.1)	0.462
	High birth weight (>4500g)a	2 (0.1)	0	0.29
	Gestation (weeks)	33.63 (2.99)	33.55 (3.10)	0.296
	Prelabor caesarean	503 (23.4)	546 (29.5)	<0.001
	Induction	414 (19.2)	381 (20.6)	0.146
	Labored spontaneously	1,237 (57.4)	922 (49.9)	<0.001
	Intrapartum factors
	Intrapartum caesarean	245 (6.1)	169 (4.2)	0.012
	Instrumental delivery	191 (8.9%)	127 (6.9)	0.011
	Duration of labor (min)	317 (220)	296 (216)	0.433
	EFM	1,605 (74.5)	1,358 (73.8)	0.309
	Epidural analgesia	367 (17.0)	305 (16.5)	0.339
	Oxytocin used	504 (23.4)	421 (22.8)	0.333
	Mechanisms of damage
	Birth trauma	13 (0.6)	9 (0.5)	0.391
	Fever in labor (>37.5°C)	81 (6.4)	80 (7.7)	0.115
	Arterial pH	7.231 (0.096)	7.231 (0.101)	0.202
	Arterial pH <7.10	166 (10.7)	148 (11.2)	0.374
	Arterial pH <7.00	44 (2.7)	43 (3.0)	0.304

SROM, spontaneous rupture of membranes; EFM, electronic fetal monitoring; Birth trauma, shoulder dystocia or failed instrumental delivery.

a Birth weight centiles are not used because these are fetal sex specific.

Table 2. The association between fetal sex and preterm delivery (spontaneous+iatrogenic)

		Male n (%)	Female n (%)	OR (95% CI)
	Overall preterm	2154 (53.8)	1849 (46.2)	1.13 (1.06–1.20)
	Extreme (24–28)	186 (49.7)	188 (50.3)	0.96 (0.78–1.17)
	Severe (29–32)	362 (55.9)	286 (44.1)	1.23 (1.05–1.43)
	Moderate (33–36)	1606 (53.9)	1375 (46.1)	1.13 (1.05–1.23)
	All term	36447 (50.8)	35725 (49.2)	1.00

Table 3. The association between fetal sex and iatrogenic preterm delivery

		Male n (%)	Female n (%)	OR (95% CI)
	Overall preterm	917 (49.7)	927 (50.3)	0.96 (0.87–1.05)
	Extreme (24-28)	96 (44.2)	121 (55.8)	0.77 (0.59–0.99)
	Severe (29-32)	187 (50.0)	187 (50.0)	0.97 (0.79–1.19)
	Moderate (33-36)	634 (50.6)	619 (49.4)	0.99 (0.89–1.11)
	All term	36447 (50.8)	35725 (49.2)	1.00

Table 4. The association between fetal sex and spontaneous preterm delivery

		Male n (%)	Female n (%)	OR (95% CI)	AOR1 (95% CI)	AOR2 (95% CI)	AOR3 (95% CI)
	Overall	1237 (57.3)	922 (42.7)	1.30 (1.19–1.42)	1.30 (1.19–1.42)	1.30 (1.19–1.42)	1.87 (1.68–2.08)
	Extreme (24–28)	90 (57.3)	67 (42.7)	1.30 (0.95–1.78)	1.30 (0.95–1.78)	1.31 (0.95–1.80)	3.87 (1.11–13.51)
	Severe (29–32)	175 (63.9)	99 (36.1)	1.71 (1.34–2.19)	1.71 (1.34–2.19)	1.73 (1.45–2.21)	3.39 (2.20–5.23)
	Moderate (33–36)	972 (56.3)	756 (43.8)	1.24 (1.13–1.37)	1.24 (1.13–1.37)	1.25 (1.13–1.37)	1.83 (1.64–2.04)
	All term	36447 (50.8)	35725 (49.2)	1.00

AOR1, adjusted for nulliparity; AOR2, adjusted for UTI and retroplacental clot; AOR3, adjusted for birth weight.

Males are not over-represented among iatrogenic deliveries (OR 0.96; 95% CI 0.87–1.05) (Table 2). Indeed, among the extreme iatrogenic preterm deliveries, males were significantly under-represented (OR 0.77; 95% CI 0.59–0.99), and when spontaneous labour only was included, there was a non-significant excess of males in the extreme preterm gestation band (OR 1.30, 95% CI 0.95–1.87). The male excess is therefore the result of an increased incidence of spontaneous preterm labour in mothers carrying boys.

Of pre-pregnancy factors that might lead to spontaneous labour (nulliparity, demographic factors such as age, or pre-pregnancy disease, Table 1), only nulliparity was associated with male sex, but the association with preterm spontaneous labour remains significant after adjustment for this (AOR1=1.30; 95% CI 1.19–1.42) (Table 3). Of recorded pre-labour pregnancy complications (Table 1) that could lead to spontaneous labour, only maternal urinary tract infection (UTI) and placental abruption were more common among males. Nevertheless, after adjusting for them, the association remains significant (AOR2 1.30; 95% CI 1.19–1.42) (Table 3). Furthermore, pre-eclampsia, a major association of spontaneous (and iatrogenic) preterm delivery, was actually more common among preterm deliveries in females (p<0.001) (Table 1).Although preterm males were, on average, 130g heavier than females (Table 1), adjusting for birth weight did not remove the association (AOR3 1.87; 95% CI 1.68–2.08) (Table 3). Of recorded markers of an inflammatory response/infection, there is no increase in the incidence of fever in labour (Table 1) and although UTIs were more common among pregnancies with males, the effect remained after adjustment for them (Table 3).

4. Discussion 

The excess of males among preterm deliveries is consistent with previous reports [2], [8]. With the exception of theories regarding hormonal differences between the sexes [7], our analysis of the putative causal pathways goes further to examine why.

The excess is restricted to spontaneous preterm delivery. In the extreme preterm gestation band, being male is not associated with increased risk overall, yet there is an increased risk of spontaneous labour. This is probably because pre eclampsia, the most common indication for iatrogenic extreme preterm birth, is more common among preterm females, an observation that has been made before [9].

We found no evidence that infection could account for the male excess. Others have also reported no increase in preterm pre-labour rupture of the membranes, clinical chorioamnionitis or endometritis in pregnancies with male babies [10]. Our data cannot exclude infection, because it is frequently subclinical and we had only maternal fever and urinary tract infection as variables. Nevertheless, preterm males do not more frequently have histological evidence of acute inflammation [11].

We found no evidence to support the hypothesis that women at higher risk of pregnancy complications implicated in preterm birth (e.g. maternal disease) are more likely to carry males, although we do not have data on some risk factors for preterm delivery, such as socio-economic class. Despite the observed increased incidence of pregnancy complications with males, a fact that has been widely described before [8], this does not appear to be responsible for the male excess. This is important and was largely unexpected. It is supported by the fact that although pregnancy disease is a cause of iatrogenic preterm delivery as well as spontaneous delivery, we show that iatrogenic preterm delivery is not more common in males.

The effect of birth weight poses a more complicated problem. Clearly, the male excess is not, as has been suggested, simply due to increased size (Table 4) or uterine stretch. Indeed, polyhydramnios may be rarer among males [12]. Yet optimally defined intrauterine growth restriction (IUGR) cannot be excluded by this analysis. It is now accepted that suboptimal fetal conditions, e.g. IUGR and pre eclampsia are associated with spontaneous preterm birth as a ‘fetal survival response’ as well as with iatrogenic delivery as medical intervention [1]. IUGR is not simply small for dates, but a state where adverse outcome is more likely because fetal growth is less than would be expected when taking account of constitutional determinants [13]. Studies of IUGR and birth weight according to gender have been dogged by this and other problems. For instance, although boys are, on average, larger, and centiles of birth weight take account of this, males could more frequently be small compared to their optimal weight. A further problem is that centiles of birth weight at preterm gestations probably underestimate the normal in utero size of a baby at that gestation, because being small predisposes to preterm birth [14]. Adjustment for these problems is beyond the scope of this analysis. In an important study, it has been suggested that males at all gestations are more vulnerable than females to growth deviations from expected norms (up as well as down) [15]. Our data allow us only to conclude that although low birth weight is an indication for iatrogenic preterm delivery and is strongly associated with pre eclampsia, there is no excess of males in iatrogenic preterm births and pre eclampsia is more common among preterm females.

Preterm delivery is an important risk factor for cerebral palsy and death. Many theories for the male excess exist. We demonstrate that the risk is (1) due only to spontaneous delivery and cannot be attributed to (2) the larger birth weight of males or (3) their acknowledged higher risk of antenatal complications, and probably not because of an increased risk of infection. Whilst this is a negative finding, these have previously thought to be important. It may indeed be that a ‘male factor’, such as hormonal differences, are responsible.