1. Introduction
It has become something of a truism that many medieval women must have died in childbed; the internet is replete with articles like ‘The Historical Horror of Childbirth,’ and grim statements that ‘many [medieval] women suffered greatly and many more died’.Footnote 1 These popular characterisations of the medieval reproductive experience find some support in the historical record. On 5 May 1316, for example, Princess Elizabeth of Rhuddlan, pregnant with her ninth child, went into labour near Quendon, co. Essex where she delivered a daughter, Isabel, and ‘the said noble lady Countess Elizabeth died in childbirth and was buried near Walden’.Footnote 2 The work of professional historians also offers support for the portrayal of medieval childbirth as incredibly dangerous. Ole Benedictow, for example, characterises the life of medieval women as marked by supermortality due in large part to reproductive stress.Footnote 3 Writers who seek to buttress this assertion can also point to the charms or talismans that medieval women sought to alleviate their anxieties about giving birth.Footnote 4 Still others reference bioarchaeological evidence from a variety of medieval English skeletal assemblages in which female mortality is interpreted as peaking between the ages of 25 and 35, ages that might suggest maternal mortality.Footnote 5 Yet the age at death among a sample of 128 female skeletons excavated at three late medieval English nunneries also peaked during their reproductive years, with 68 per cent of the women dying between 26 and 45; unless these were particularly naughty nuns, the pressures of reproduction cannot have been the cause of death.Footnote 6
Maternal mortality not only affects the size and age pyramid of the population but also family structures, the labour market and gender roles. In modern studies of the effects of maternal mortality, scientists have noted that children whose mothers have died are more likely to die than children whose fathers have died, and children in both groups are at higher risk of death than children with both parents.Footnote 7 Today, the risk is particularly high for infants, who rely on their mothers for sustenance; in a community-based study of maternal deaths in India, researchers found that the risk of infant death during delivery increased five times after the death of the mother, and the overall risk of mortality for infants in their first year was 28 times that of an infant whose mother was still alive.Footnote 8 A similar survey of Tanzanian households found a correlation between stunted development and maternal death, even more so than the households that experienced paternal death.Footnote 9 In a study of the relationship between family health and death in New Delhi, researchers concluded that the loss of a mother was ultimately more detrimental to overall household health than the loss of any other adult and had the most negative economic effects on the family, even though the fathers were the primary breadwinners.Footnote 10 Modern consequences that are particularly resonant for our understanding of the medieval past include the higher likelihood of family dissolution and reconstitution, as well as an increased rate of economic participation of children and a breakdown in the female-dependent networks of support for the very young and very old.
Determining maternal mortality in any medieval population carries with it a myriad of problems, particularly when it comes to evidence. Many primary sources come to us via male mediators, and the accurate recounting of the experience of birth, which occurred in a female-dominated space, was of subsidiary concern. Birth scenes in the chivalric literature are generally heavily stylised, and accuracy takes a back seat to artistry. It is highly unlikely, for example, that dying royal women gave elaborate speeches to their newborn children, as Queen Isabella does in the Prose Tristan, lamenting, ‘in sorrow I bore thee, sad was the hour I brought thee forth, and sad is the welcome thy wretched mother bids thee,’ before she bestows an appropriately sorrowful name on her son and expires.Footnote 11 Canonisation inquests and miracle collections have many of the same problems; the ultimate goal is to assert the efficacy of a given holy man or woman, and as such only those cases with positive outcomes were recorded for posterity. In fact, problems related to pregnancy and childbirth constitute only 2 per cent of miracles in English collections from the twelfth and thirteenth centuries, despite the ubiquity of the reproductive experience.Footnote 12
The material culture of childbirth, both secular and religious, offers insight into the lived experiences and concerns of actual women by illuminating how medieval women managed their anxiety regarding the birth process but tells us little about how many pregnancies medieval women experienced and if or when they died giving birth. The coral and jet beads that women willed to other women indicate that they feared haemorrhage and believed in the blood-staunching power of these stones, for example, but not how often they were used.Footnote 13 Wellcome MS 632, a parchment roll with protective prayers and pictorial representations of the instruments of Christ's passion, is perhaps the only extant medieval English birth-girdle which was actually used; rolls like MS 632 were placed over the belly of the parturient woman; the prayers offer protection from death in childbirth, and also from shipwreck or judicial punishment. The manuscript shows signs of wear and spots that could be blood, but who it belonged to, and against which threat to life, remain tantalising mysteries.Footnote 14
Bioarchaeological evidence is equally if not more problematic for quantifying maternal mortality. Despite the supposed frequency of such deaths in pre-modern societies, since 1970 archaeologists have published reports detailing only about 20 skeletons of women who were definitively pregnant or in labour at death, that is, buried with foetal remains in situ under the pubic symphysis.Footnote 15 Both anatomical and social factors account for the rarity of these finds. Foetal remains, while robust, are quite small and prone to misidentification or being missed by inexperienced archaeologists, and while burials of neonates and reproductive-age women in the same grave may suggest a case of maternal mortality, it is only suggestive, not conclusive evidence.Footnote 16 Western cultural taboos about the burial of pregnant or recently delivered women with foetal remains resulted in their burial in marginal spaces, decreasing the likelihood of survival and excavation. In the western world, this practice began with the Lex Caeserea of the Roman Empire and continued through the end of the Middle Ages.Footnote 17
The paucity of concrete source material means that most demographers of medieval Europe have not taken up maternal mortality, despite the importance of the topic.Footnote 18 Wrigley and Schofield's The Population History of England 1541–1871 and Andrew Hinde's English Population: A History since the Domesday Survey, for example, discuss infant mortality, nuptiality and age at marriage, but not maternal mortality. This is also true of T.H. Hollingsworth's Historical Demography. Peter Laslett's discussion of child marriage in Elizabethan England, in The World We Have Lost, Further Explored, asserts that mothers below age 16 were virtually unknown but makes no mention of the dangers of pregnancy in adolescents. French scholars have been slightly more attentive, though most studies focus on the early modern period and analyse relatively small populations. Bertrand-Yves Marfart's 1994 article is a rare exception, though his evidence is largely bioarchaeological and based on only four skeletal assemblages, totalling 156 women, three of whom were buried with foetal remains. One of the few studies that deal explicitly with quantifying obstetric death in the medieval world is the analysis of David Herlihy and Christine Klapisch-Zuber. Through an analysis of several Florentine Libri dei morti, c. 1424–1430 that recorded cause of death for 2,312 women, the authors observed that only 32 women, less than 2 per cent, were recorded as dying in childbirth (sopra parto or sconciasi).Footnote 19 This is a far lower figure than we find in popular perception, though the methodology employed by Klapisch-Zuber and Herlihy does not match what modern demographers employ to calculate the maternal mortality rate. What are we to believe, therefore, about the birthing experience of the Middle Ages? Was childbirth ‘terribly dangerous,’ as Lawrence Stone wrote in the two sentences devoted to childbirth in his seven-hundred-page work on the English family?Footnote 20 Did ‘everyone … know someone who died in childbirth?’Footnote 21
In the following article, I hope to accomplish two tasks: first, to identify the risk factors which most often contributed to deaths in childbed and, second, to quantify the rate of maternal mortality among late medieval elite Englishwomen. To do so, I employ a quantitative methodology based upon a FileMaker Pro database, ‘Births and Deaths of Englishwomen’ (BADE). The database records the pregnancies of the most well-documented members of medieval English society from the marriage of Eleanor of Provence to King Henry III in 1236 to the death of Elizabeth of York in 1503, offering more than a century of data on either side of the demographic disruption of the Black Death.Footnote 22 By doing so, the database allows for an analysis of the incidence of both maternal death and risk factors identified by modern studies of maternal mortality, namely maternal youth (defined here as maternal age below 15 years at delivery), high parity (for this study, the fifth birth event and beyond), stillbirths and advanced maternal age (maternal age over 35 at delivery).Footnote 23 In populating the database, I began with the queens, then their daughters and daughters-in-law, branching out to include female members of noble families related to the royal family by marriage. The selection of aristocratic women for the study was based on two interrelated trends. First, primary sources about the elites are much more plentiful than for the peasantry. Second, because the propagation of the lineage through marriage and the successful delivery of heirs was of immense concern, the pregnancies of elite women were much more likely to receive comment, either by the immediate family or, for the royal family in particular, by contemporaries. Monastic histories and the chronicles of Matthew Paris, for example, provide primary documentation for the births, pregnancies and deaths of many queens and princesses, organised by year. Secondary scholarship and biographies provide further material for the royal family and select elite women, as well. Finally, in order to account for mistakes in primary sources or transcription and to minimise bias and reduce cherry-picking, I systematically mined the Oxford Dictionary of National Biography, filtering entries by date (c. 1236–1503) in order to capture information on any women previously missed.Footnote 24 For comparison's sake, I also tracked the pregnancies of women in the most well-documented gentry families: the Celys, Pastons, Plumptons and Stonors. In sum, the database tracks 457 pregnancies shared between 102 women.Footnote 25
Though small, the sample size is in many ways constrained by the topic itself. In order to perform this exercise in a responsible and defensible way, scholars need to have several pieces of information. For calculating maternal mortality, the total number of pregnancies must be compared to the number of maternal deaths. To most effectively analyse the reproductive lives of a past population, however, requires more information, including the date of birth of each mother and her children, the total number of pregnancies experienced, the interval between pregnancies, and incidence of both stillbirths and multiple births. The scarcity of this information is, by and large, why quantitative analysis remains lacking for women who lived in societies without consistent and detailed demographic data. For many members of the gentry, for example, maternal date of birth is unknown, allowing only for comparison of the total number of recorded pregnancies. I also recorded the social status of the mother – queen, princess, noblewoman or member of the gentry – allowing for comparison across social strata. The database also keeps track of how long the child survived, if the child married and if married, had issue. It is important to note, too, that the BADE database excludes women who, for a multiplicity of reasons, never reproduced successfully. Some women took the veil (in this study, 8 of the 194 female children who lived to their teens), or never married despite living to adulthood. Nor was marriage a guarantee of successful reproduction. When Anne of Bohemia died after 12 years of marriage to Richard II, a variety of chroniclers including Adam Usk and Jean Froissart remarked that their union had produced no children. Furthermore, no evidence exists for Anne ever conceiving, and contemporary apothecary bills preserved in E 101/402/18 include medicaments that may have been used to regulate her menses in the hopes of becoming pregnant. Ultimately, however, it is impossible to know if the problem lay with her or her husband, who died without issue in 1399.Footnote 26 In the database, 14 of the 160 female children who lived to adulthood, or roughly one in ten (9 per cent), married but never gave birth.Footnote 27 The BADE database, therefore, covers a small but comparatively well-documented section of late medieval English society. The women analysed were elite members of late medieval English society, women with considerable economic and political clout whose births, pregnancies and deaths received contemporary comment, and for whom the ability to reproduce successfully was of widespread concern and therefore received comment.
Because the population included in the database is confined to the upper echelons of English society, I incorporated a variety of supplementary material on non-elite women, including secondary scholarship on maternal mortality in early modern England. Julia Allison, for example, used parish registers to estimate maternal mortality in six East Anglian parishes between 1539 and 1619, while Alannah Tomkins performed a similar task for north Shropshire in the late eighteenth century through a statistical analysis of midwife books. These sources, which unfortunately do not exist for medievalists, can serve as benchmark comparisons. The medieval evidence written in bone also offers glimpses into the experience of non-elite women. Mary Lewis and her colleagues have performed valuable studies on duration of puberty, age at menarche and age at menopause using a variety of English medieval skeletal assemblages, and these data were also incorporated into the analysis of age at first pregnancy and interval between pregnancies.Footnote 28 For the purposes of this study, I employed the definition of maternal death put forth by the World Health Organization, ‘the death of a woman while pregnant or within 42 days of termination of pregnancy, irrespective of the duration and site of the pregnancy or its management, but not from accidental or incidental causes’.Footnote 29 In modern terms, for example, a woman who died of a heart attack due to increased blood pressure from pregnancy would be counted as a maternal death, while a pregnant woman who died in an automobile accident would not. Other pregnancy-related causes of death include an increased susceptibility to infection, as pregnancy represses the mother's immune system to minimise the chance of foetal rejection, increasing the risk of contracting tuberculosis, influenza and smallpox.Footnote 30 Blanche of Lancaster died of infection while pregnant with her second child in 1409; we cannot know if the illness may have been exacerbated by her pregnancy, but it is possible that her lowered immune response contributed to her death.Footnote 31 Women who died within six weeks of delivering a child were flagged as possible cases of maternal mortality; after consulting primary and secondary source material, these cases were marked as either ‘definite’, ‘questionable’, or excluded as cases of maternal mortality.Footnote 32
1.1 Age at first pregnancy
Age at first pregnancy is a major determinant of the risk of obstetric death. The World Health Organization notes that the risk is highest for adolescent girls under the age of 15.Footnote 33 In an almost 20-year study of 854,377 Latin American women, researchers found that pregnant girls under age 15 were at a much higher risk for puerperal endometritis, preterm delivery, and deliveries requiring episiotomies (a procedure which did not develop until the eighteenth century), as well as small-for-gestational-age infants and low birth weight.Footnote 34 Judith Lewis sampled the raw data produced by T.H. Hollingsworth in the creation of The Demography of the British Peerage to analyse maternal mortality among the English aristocracy between 1558 and 1959 and concluded that age at first pregnancy was one of the most influential factors for obstetric death.Footnote 35 Ole Benedictow, in his study of the demography of medieval Scandinavia, asserts that early marriages, and with them early consummation and pregnancies, would have resulted in an increased risk of maternal mortality among young women, a conclusion he supports through a survey of bioarchaeological evidence including the cemetery at Loddekopinge, and life table analysis. As pointed out above, however, determining maternal mortality in skeletal remains is a difficult endeavour without the presence of foetal remains in situ.Footnote 36 Furthermore, young women could and did die from a variety of ailments, and most of them were unrelated to their reproductive status.
There were certainly medieval girls who became mothers at an age we would consider shocking today. In 1455, Edmund Tudor, the 25-year-old half-brother of King Henry VI, married his 12-year old ward, the heiress Margaret Beaufort. When Edmund died of plague a year later, he left behind a 13-year-old and heavily pregnant widow. According to Margaret's confessor, John Fisher, her delivery of the future Henry VII was a dangerous affair; Fisher noted that Margaret was neither ‘a woman of great stature’ nor ‘yet fourteen’ when she gave birth.Footnote 37 Margaret married again the following year, to Henry Stafford. Though their 14-year marriage seems to have been ‘genuinely affectionate’, they had no children; the same was true of her third marriage to Thomas Stanley, which lasted until the latter's death in 1504.Footnote 38 Margaret may have been physically damaged by the early consummation of her marriage or the resulting pregnancy, to such a degree that she never conceived again.Footnote 39 Margaret's experience, however, was not the norm for the women surveyed.
Table 1 displays the age at first pregnancy for all mothers for whom date of birth and date of pregnancy could be determined. Only six birth events occurred to mothers who were likely below the age of 15.Footnote 40 Among the women studied, the average age at first pregnancy was 20 years. This is lower than the mean of 22 years found by Judith Lewis in her study of maternal mortality among aristocratic women between 1558 and 1959, which she considered ‘remarkably steady over four centuries’, but still later than might be expected considering how early the royals and aristocracy married off their daughters.Footnote 41 To what can we attribute this lag between age at marriage and age at first pregnancy? While most girls in the modern western world begin puberty around age 10, considerable evidence exists that medieval girls entered into this process later. The Knowing of Woman's Kind in Childing, a late medieval gynaecological treatise, notes that girls below the age of 15 do not menstruate, ‘for they are joyefull and so yong her mete defyeth as the resseyue it. And so the blood and other humoures pass a-way’.Footnote 42 Bioarchaeological evidence supports this assertion: an analysis of a skeletal assemblage of 314 late medieval English girls who died between 14 and 25 years of age found not only that puberty began later but also that it lasted longer.Footnote 43 Shapland, Lewis and Watts observed ossification of the iliac crest, a developmental event associated with the onset of menstruation, occurring at an average age of 15, with complete fusion of the crest, marking the end of pubertal development, rarely occurring before age 20.Footnote 44 The women included in the database, as elites with diets higher in animal protein and iron, probably achieved full fertility earlier than the young women analysed by Shapland et al., but still well after modern British girls who, on average, begin menstruating at age 13.Footnote 45
Table 1. Age at first pregnancy of royal, noble, and gentry women in England, c. 1236–1503
Notes: Percentages have been rounded to produce integral numbers. The dataset includes all royal and noble women, as well as three gentry women for whom age at first pregnancy could be determined. Eighteen gentry and noble women (18 per cent) for whom the date of birth could not be determined are excluded.
Source: BADE Database.
Medieval parents also seem to have experienced anxiety when marrying off their daughters at a tender age, and some made provisions to delay consummation. Margaret Beaufort's difficult delivery was attributed to both her short stature and her young age, and she objected to the early marriage of her granddaughter to James IV of Scotland because she feared that the bridegroom would not wait until the bride was of age to consummate the marriage.Footnote 46 Her biographers, Michael Jones and Malcolm Underwood, posit that her reluctance was due in part to her own experience.Footnote 47 The gentry seems to have shared similar anxieties; when the 13-year old Elizabeth Clifford married William Plumpton in 1453, the marriage contract included a clause that the marriage would not be consummated for another three years.Footnote 48 Their contract was not remarkable in that regard: many included clauses stipulating that, should either bride or groom die before 16, the dowry should be returned. Barbara Harris takes these clauses as indicative of a delay between age at marriage and age at consummation.Footnote 49
Given these biological and cultural factors, the delay between age at marriage and age at first pregnancy seems entirely reasonable. Even when they married early, as many medieval elites did, some would delay consummation and others, who did not, were simply not physiologically able to conceive. Very early pregnancies were the exception rather than the rule and not, as Benedictow asserts, a main driver of adolescent female mortality.Footnote 50 Confined to the upper classes, these early marriages and consummations were the result of calculated political and economic decisions rather than a widespread practice crossing social classes.Footnote 51
1.2 High parity
Women who have experienced a high number of previous pregnancies are at risk not only for conditions beginning during pregnancy, such as preeclampsia, but also issues during delivery, such as cardiac or respiratory arrest.Footnote 52 As Allison notes in an article surveying maternal mortality in early modern East Anglia, ‘grand multiparous women’, defined as women who gave birth to five or more infants, were also older, compounding their risk of dying in childbirth.Footnote 53 The BADE database allows us to evaluate the experience of late medieval Englishwomen as marked by many pregnancies. Some women spent most of their adult lives pregnant. By the time of her husband, Thomas Grey's, death in 1501, Cecily Bonville had given birth to seven sons and eight daughters – 15 pregnancies in 27 years.Footnote 54 Is it wise, however, to take Cecily as representative of elite Englishwomen writ large?
The women included in the database averaged 4.4 pregnancies; this is double the number of Englishwomen today, who average two, but consistent with the global average per woman c. 1970–1975 (4.5 births).Footnote 55 Barbara Harris performed a similar task for elite women between 1450 and 1550, using correspondence, wills, cases in the Chancery and Star Chamber, marriage contracts and more. Her own, much larger assemblage, also averaged 4.4 children per woman.Footnote 56 The proportions of aristocratic women with five or more children in my sample were 36 per cent (see Figure 1), while Harris’ sample was 40 per cent. Twenty-five per cent of the women analysed in BADE had six or more children, compared to Harris's 29 per cent.Footnote 57
Figure 1. Total number of pregnancies of royal, noble and gentry women in England, c. 1236–1503. Notes: Percentages have been rounded to produce integral numbers. The dataset includes all royal and noble women, as well as three gentry women for whom age at first pregnancy could be determined. Eighteen gentry and noble women 18 per cent for whom the date of birth could not be determined are excluded. Source: BADE Database.
It is possible that this low number of recorded pregnancies may be indicative of the low health status of medieval people more generally. While the relatively short stature of medieval people has been remarked upon in a variety of English skeletal assemblages, the study of a known individual is relatively rare. One exception, however, is the bioarchaeological analysis performed on the remains of Anne Mowbray, Duchess of York, who died shortly before her ninth birthday in 1481. In 1965, excavators rediscovered and analysed the remains of the child duchess, including bones, hair and nails.Footnote 58 Anne was as tall as a modern six year old, and her hair had atrophied roots, indicating ill-health.Footnote 59 Neutron activation analysis of her hair also found high levels of arsenic and antimony, possibly administered as medicine.Footnote 60 The consumption of these materia medica is known to have deleterious effects on reproduction, causing male and female infertility and provoking spontaneous abortion in pregnant women.Footnote 61 As a wealthy heiress and then a member of the royal household, it is unlikely that Anne ever wanted for food, and yet her stature still lagged well behind that of a modern child of equivalent age, illustrating the comparatively low level of nutrition in the medieval English diet, even among the upper echelons of society.Footnote 62 Poor nutrition can retard fertility, with negative effects including delayed, anovulatory, or the complete cessation of menses for women, and decreased sperm number, decreased sperm mobility and a stop in sperm production for males. Among well-nourished modern communities which do not practice contraception, such as the Hutterites, researchers have documented families with 11 to 12 live births; among poorer historical communities, however, couples who lived to completed fertility, that is, the onset of menopause, averaged only six to seven live births.Footnote 63
1.3 Interval
A short interval between pregnancies means that the mother is not able to build up sufficient stores of nutrients between births, a condition which is dangerous for both mother and infant. Scholars call the decrease in maternal health due to high numbers of pregnancies and short intervals ‘maternal depletion syndrome’ (MDS), with symptoms including anaemia, goiter, oedema and osteomalacia (a softening of the bones due to vitamin D or calcium deficiency).Footnote 64 In her survey of maternal mortality among early modern English aristocratic women, Lewis found that age at first pregnancy and birth intervals had the highest impact on risk of death; she notes ‘for most women, it was the long years of short birth intervals that killed’.Footnote 65 Perhaps no English queen illustrates this better than Philippa of Hainault. She spent much of her twenties with child, giving birth at ages 23, 24, 26, 27 and 28, and half of her pregnancies occurred within two years of the previous delivery.Footnote 66 Philippa gave birth to her seventh child, Edmund of Langley, for example, only 13 months after she birthed her sixth, John of Gaunt.Footnote 67
Of the 382 birth events with definitive dates in the BADE database, 74, about 19 per cent, were within two years of the previous pregnancy.Footnote 68 Short intervals occurred in all social groups including the gentry – Margery Punt, the wife of George Cely, gave birth to five children in five years, for example. In fact, 21 women shared all the pregnancies which occurred less than two years since the previous birth event, meaning that one in five of the women analysed experienced a birth interval of less than 24 months at least once during her reproductive life.
Birth intervals were often shorter for the elite than the majority of the population, as elite mothers farmed their babies out to wet-nurses rather than nursing them themselves, allowing for the resumption of ovulation, menstruation and conception.Footnote 69 The demographer R.G. Potter, looking at the interval between first and second pregnancies in modern populations, found an average birth interval of 18 months for non-lactating mothers versus 27 months for lactating mothers.Footnote 70 A study of nitrogen isotopes in the skeletal assemblage of Wharram Percy found that the medieval women buried there breastfed for about 18 months, placing them squarely in the lactating mothers category.Footnote 71 Whether or not the employment of wet-nurses was a conscious choice made by the elites in part to maximise the number of heirs, it had the effect of increasing the fertile period of women, the number of pregnancies experienced, and their risk of obstetric death.Footnote 72 While the danger of short intervals remains true, the frequency with which these women experienced short birth intervals is less than might be expected, as 80 per cent of all recorded pregnancies occurred more than two years after the previous birth, allowing for the considerable restoration of maternal nutrient stores, particularly among elites with better access to nutrient-dense food.
1.4 Stillbirths
Stillbirths are, even today, much more dangerous than live births for the mother. If the stillborn foetus begins to decompose in utero, prior to the advent of drugs like Pitocin and other pharmaceuticals that induce uterine contractions, infection, sepsis and death could quickly follow. If the child died during labour, even in the birth canal, this could also increase the risk, as tools which allowed for the extraction of a foetus were not available until the early modern period.Footnote 73 Tomkins found in her analysis of Shropshire midwife books that the risk of maternal death was four to five times higher when delivering a stillbirth, and R.S. Schofield's estimates are roughly similar.Footnote 74 Allison's study of parish registers found that the MMR for mothers giving birth to stillborn children was 125 per 1,000 births, and therefore, the highest predictor of maternal death.Footnote 75
Medieval people understood how dangerous this situation could be, and both religious and medical thought reflected this knowledge. The Trotula notes that ‘there are some women who are so afflicted in the function of birth that hardly ever or never do they deliver themselves … sometimes because the foetus is dead and cannot aid Nature in its movement’.Footnote 76 It further recommends rue, mugwort, wormwood and pepper in wine to prompt the expulsion of the dead child.Footnote 77 A Middle English translation of the Macer Floridus, itself derived in part from the Regimen Sanitatis Salernitanum, also declares that imbibing rue ‘putteth out the child’.Footnote 78The Knowing of Woman's Kind in Childing directs that the parturient woman drink hyssop in hot water ‘yif the child be ded in the modris wombe’, indicating a clear understanding on the part of medical practitioners that allowing a dead foetus to remain was dangerous to the life of the mother.Footnote 79 This text was not only about women, but intended for their use: the author notes that he has translated the work from French and Latin into English because that is the language literate women can read, and he hopes that the literate women might pass their knowledge on to their illiterate counterparts.Footnote 80 If we use miracle stories to illuminate popular understandings, we find that knowledge of the dangers of stillbirth to maternal health was not restricted to the upper tiers of medieval society. William of Canterbury records in his twelfth-century collection of miracles relating to Thomas Becket that the pregnant wife of Ansfrid, a Sussex knight, knew that her life was in danger after the time for the birth came and went and the foetus did not move. Indeed, ‘the living one thought she was being made rotten by the dead one, [and] she began to lose her senses’.Footnote 81 Only a vow to St. Thomas saved her life; she delivered the dead child and, together with her husband, made the promised pilgrimage to Canterbury.
Determining the stillbirth rate among past populations is a difficult task. Stillborn infants in medieval Europe were often buried outside or on the margins of the churchyard and produced less documentation than infants who were born alive and survived to baptism.Footnote 82 Even parish registers, one of the most detailed demographic sources for early modern England, likely under-recorded children born dead, while midwife books noting deliveries and their outcomes vary in completeness regarding stillbirths.Footnote 83 Some midwives counted them in their ledgers and journals, and others did not.Footnote 84 In the BADE database, births resulting in unnamed children were considered as possible stillbirths and cross-referenced with material relating to the mother to evaluate the likelihood of a stillbirth. For example, in January 1278, at age 37, Eleanor of Castile gave birth to her twelfth child, probably a daughter, at Westminster.Footnote 85 The child was either stillborn or died very soon after birth, as no name was recorded, or preparations made for baptism. This methodology resulted in seven possible stillbirths, or a ratio of one stillbirth for every 65 live births. This result echoes the 48 aristocratic women analysed by Lewis in her article, ‘Maternal Health in the English Aristocracy, Myths and Realities, 1790–1840’; these women experienced 381 conceptions, resulting in 345 live births, 31 miscarriages and 5 stillbirths, a ratio of one stillbirth per 69 live births.Footnote 86 Both rates are favourable when compared to that observed in two early modern English midwife books, where the ratio of stillbirths to live births was one stillbirth per 25–33 live births.Footnote 87 The women of the upper classes, it seems, experienced fewer stillbirths than their less wealthy counterparts.
1.5 Age at last pregnancy
Over age 35, the risk of pregnancy complications including diabetes and high blood pressure rises dramatically; the risk of prematurity or low birth weight increases concomitantly. A large-scale study of American women, for example, found that the risk of severe complications was 20 per cent higher for women aged 35 to 39 than for women aged 25 to 29, and this risk more than quintupled for mothers over age 50.Footnote 88 The authors concluded that the age of the cerebrovascular and respiratory systems contributed to this increased risk, as the heart, lungs and brain of the mother were unable to cope effectively with the stresses of pregnancy, including a rise in blood volume and blood pressure.
The average age at last pregnancy in the sample was 28 (see Table 2 for the age distribution at last pregnancy). Though queens experienced a slightly longer fertile period, with an average age at last pregnancy of 31, this may simply reflect the high level of documentation related to the royal family. The relative youth of the women at their last pregnancy may be due in part to an under-recording of pregnancies resulting in miscarriages or stillbirths, particularly those ending before the mother could feel foetal movement, called ‘the quickening’. This event also signified that the foetus was endowed with a soul; it is possible, therefore, that pregnancies lost prior to this point received little to no contemporary comment.Footnote 89 The risk of miscarriage also increases with age, which would decrease the average age at last pregnancy. A large-scale study of over 420,000 pregnancies between 2009 and 2013 found that risk of miscarriage is lowest for women aged 25–29 (10 per cent), increasing after age 30 and eventually reaching 53 per cent for women over age 45.Footnote 90 Elizabeth Woodville's fertile period of 23 years was particularly long, though the longest observed was that of Eleanor of Castile, who gave birth for the first time at age 13, and for the last time at 43. Six of the 80 women, or one in 13, were fertile even after age 40, including four queens, one noblewoman and a member of the gentry. In contrast to Benedictow's assertion that ‘women's reproductive role ceased’ after age 40, some women conceived and gave birth even after this point.Footnote 91 Most of the pregnancies occurred between the ages of 20 and 29, however, when the mothers were fully grown but before the risk factors associated with age developed.Footnote 92
Table 2. Maternal age at last pregnancy for royal, noble and gentry women, c. 1236–1503
Notes: Percentages have been rounded to the first decimal. The dataset excludes 22 women for whom age at last pregnancy could not be determined.
Source: BADE Database. See note to Table 1.
2. Discussion
Between 1840 and 1936, when maternal mortality began to drop precipitously, the maternal mortality rate was 5 maternal deaths per 1,000 live births.Footnote 93 The introduction of antibiotics, sulphide drugs and increased prenatal care further contributed to this decline and, between 1990 and 2015, maternal mortality dropped by 44 per cent worldwide.Footnote 94 This steep decline makes it difficult to compare the MMR of past societies to modern countries today. We do, however, have some material for comparison, particularly the early modern parish registers, some of which include maternal deaths related to stillbirths, which are left out of the WHO data, and secondary scholarship on the aristocracy and midwife books.
Out of the 457 birth events recorded in the database, 44.6 per cent were births which had one or more risk factors.Footnote 95 The most frequent risk factor experienced by elite women in late medieval England was high parity (see Table 3); this is unsurprising given the desire to produce multiple heirs, including males to continue the family line and females to link to other powerful families. The number is slightly skewed by the extremely high number of pregnancies experienced by Eleanor of Castile, Joan Beaufort, Cecily Bonville, Philippa of Hainault and Elizabeth Woodville; though they comprise only one per cent of the mothers, their pregnancies constitute 28 per cent of the high parity births. Short intervals between births were also common, occurring in more than one in ten pregnancies, which was in large part a result of wet nursing and reflective of the high social status of the mothers. Advanced maternal age was relatively common as well, with 10 per cent of the pregnancies for which maternal age could be determined occurring to mothers over age 35.
Table 3. Risk factors experienced by elite Englishwomen, c. 1236–1503
Notes: A pregnancy with more than one risk factor is included in all applicable categories, such as ‘maternal age’ as well as ‘multiples’. Percentages have been rounded to the first decimal.
Source: BADE Database.
a For maternal age and maternal youth, the number of pregnancies analysed was confined to those with a known maternal age, totalling 358 birth events.
In the database, six women died while giving birth or soon after.Footnote 96 Four of the six women shared the risk factor of high parity (experiencing their sixth, eighth and ninth pregnancies), and the other two women died giving birth for the first time. Elizabeth of York's final pregnancy involved two risk factors; she was pregnant with her eighth child at 37 years of age. In total, 6 per cent of the women analysed in the database died during or because of childbirth. Ninety-four per cent, however, did not. This pairs remarkably well with Lewis's cohort of 65 aristocratic women born between 1575 and 1599, of whom 4, 6.1 per cent, died in childbirth; in her total sample of 1,251 women who lived between 1558 and 1899, 61 women (4.8 per cent) were classified as cases of maternal mortality.Footnote 97
To determine the maternal mortality rate, the number of maternal deaths must be compared to the number of birth events. If the likely stillbirths are excluded, as they are in modern calculations, the MMR of late medieval England was 13 deaths per 1,000 live births, roughly similar to Lewis's 10.34 deaths per 1,000 births among the aristocracy between 1558 and 1700.Footnote 98 Put another way, excluding stillbirths, the elite woman ran about a one per cent chance of death with every birth. As the average number of pregnancies experienced was roughly 4 (457 birth events among 102 women), we would expect to see, and indeed do, that around 5 per cent of elite women died in childbirth. It is important to note, too, that the pregnancies analysed are only those which entered the documentary record in some way: it is quite likely that medieval women experienced miscarriages or birthed sons and daughters who were never remarked upon. These birth events would increase the overall total used in calculating the maternal mortality rate, perhaps dropping it even lower. This problem is allayed only somewhat by focusing on royalty and the upper echelons of the aristocracy, for whom the propagation of the bloodline was of primary importance and who created comparatively ample documentation.
With regard to change over time, great care must be taken when drawing conclusions, as the sample analysed in BADE represents only a small, elite, but well-documented section of medieval English society. Even absent a larger sample size, however, what is striking is the level of continuity in the reproductive experience of elite Englishwomen between the end of the medieval period and through the early modern; among the women analysed in BADE and Lewis's two samples spanning 1558–1899 CE, the percentage of women who died in childbirth remained stable at around 5 per cent. This relative continuity points to the absolutely integral role of social class in characterising the childbearing experiences of pre-modern women. Early modern medical writers believed that elite women were at a higher risk of dying in childbirth, and certainly, they experienced a higher number of pregnancies than poorer women, due to the practice of wet nursing and the desire to maximise potential heirs.Footnote 99 The upper classes, however, were also more likely to consume diets rich in iron and B12 and experience a higher level of sanitation during the birthing process.Footnote 100 Early modern parish registers confirm the centrality of social standing in studying MMR. Julia Allison's analysis of parish registers for six parishes in East Anglia observed an MMR across social classes of 7.5 maternal deaths per 1,000 birth events (including stillbirths), but MMR shifted depending on the economic standing of the parish's population.Footnote 101 The comparatively wealthy coastal parish of Wells experienced the lowest MMR, 4 deaths per 1,000 birth events, while Newport, which was undergoing an economic downturn, had the highest, 14.9 maternal deaths per 1,000 births.Footnote 102 Modern studies of maternal mortality confirm the important role of economic and social status in risk of death; 99 per cent of all maternal deaths in the world today occur in developing nations.Footnote 103 Social class ultimately played a larger role in the risk of obstetric calamity than the number of pregnancies experienced. Poorer women, who began with lower nutritional reservoirs, experienced more danger in the childbed than their elite, earlier-married, frequently pregnant counterparts.Footnote 104
While a statistical analysis of this population remains impossible due to a lack of systematic documentation, other methods offer a means of ingress into nutrition and health among lower-status medieval women; these factors, in turn, increased their risk of maternal mortality. Stable-isotope analysis offers one way forward, allowing for analysis of how social status determined diet. Upper-class individuals in medieval urban centres had access to foodstuffs rich in both iron and B12, such as beef, pork and marine proteins, while the middling class relied more on grains.Footnote 105 Prior to the fall in food prices which followed the Black Death of the mid-fourteenth century, many of the poor were forced to subsist on pottage, bread and other iron- and B12-poor foods, increasing their risk of anaemia and haemorrhage.Footnote 106 Skeletal indicators of poor health, including porous bone lesions such as porotic hyperostosis and cribra orbitalia, which manifest as pitting in the bone's surface, are observed more frequently in the remains of low status, urban individuals, who lived in a location with a higher parasite load and poorer hygiene than their counterparts in the countryside.Footnote 107
3. Conclusions
The ability to propagate one's lineage was of supreme importance to medieval people. Doing so was not only the fulfilment of the charge to ‘be fruitful and multiply’ but also served to cement political, economic and social ties. As such, the reproductive lives of late medieval aristocrats received considerable commentary; queens, princesses and countesses could be noted for their lack of children, as Anne of Bohemia was, or for their large broods, like Philippa of Hainault, and there was little confusion over which state was preferable. Conversely, there was also widespread awareness of how dangerous giving birth could be, and in the intervening centuries, the characterisation of the Middle Ages as a time of remarkably high maternal mortality has become commonplace. An evaluation of this portrayal through quantitative analysis shows, however, that the reproductive experiences of elite Englishwomen were not as foreign or as dangerous as originally assumed. On average, these women experienced their first pregnancy at about age 20 and their last at 28, and three or four pregnancies in the interim. Each birth carried a risk of about one per cent, and therefore, approximately 1 out of every 20 women would die in childbirth.
While the rate of maternal mortality observed in the BADE database would be unacceptable today, we have only recently been able to consider it so. The British Registrar General began to record maternal deaths in 1837, but it was not until 1870, when recording cause of death became mandatory, that these records became reliable. It was not until 1930 and the advent of antibiotics that the rate of maternal mortality began to drop, falling to below 1 death per 1,000 births in only 40 years.Footnote 108 In 2015, the MMR of the United Kingdom was only 9 maternal deaths per 100,000 live births.Footnote 109 Medieval women did die in childbirth, but the process of labour and delivery was hardly the main driver of female mortality during their reproductive years. Most women would experience pregnancy, and most would deliver successfully. In the end, though childbirth may have defined the lives of many medieval elite Englishwomen, it infrequently ended them.
