Fetal And Maternal Factors Associated With Infant Mortality In Vervet Monkeys

Abstract

Background

Causes of infant death remain unknown in significant proportions of human and nonhuman primate pregnancies.

Methods

A closed breeding colony with high rates of infant mortality had pregnancies assessed (n=153) by fetal measurements and maternal characteristics. Infant outcome was classified as neonatal death (stillborn or died <48 hours from birth), postnatal death (died 2 – 30 days) or surviving (alive after 30 days).

Results

Fetal size did not predict outcome. Poor maternal glycemic control and low social ranking increased odds for adverse outcome (OR=3.72, p=0.01 and 2.27, p=0.04 respectively). Male sex was over-represented in stillbirths (p=0.04) and many were macrosomic, but size did not associate with maternal glycemic control measured as glycated hemoglobin A1c. Postnatally dead infants were smaller (p<0.01), which associated with behavioral factors and glycemic control.

Conclusions

Fetal growth estimates predicted gestational age but not fetal outcome. Maternal social status and metabolic health, particularly glycemic control, increased risks of adverse pregnancy outcome.

Introduction

The African green monkey, or vervet (Chlorocebus aethiops), is widely utilized in biomedical research and has been well-characterized as a model of maternal-infant interactions [1, 2]. Here we investigate an unusually high and sustained infant mortality rate of greater than 30% [3] that had been refractory to efforts at reduction, in a closed breeding colony of captive vervets.

Stillbirth and infant mortality continue to have significant prevalence rates in women of industrialized nations, accounting for nearly 1% of all pregnancies in the United States [4]. Stillbirth is the greater proportion of this percentage, and the majority of cases have unknown causative factors [5, 6]. Infant mortality in captive nonhuman primate breeding colonies is typically low, constituting < 20% of the detected pregnancies [7-10]. High reproductive failure is a problem from a production and animal welfare standpoint. However an opportunity exists to learn about risk of neonatal death from situations where environmental, nutritional, and genetic conditions are relatively homogenous.

Among human and nonhuman primates, investigations into causes of human stillbirth and neonatal death are ongoing. In utero infection [11], growth aberrations [12], maternal health (specifically diabetes) and behavioral aspects [13-15] are all known to play important roles. Abnormalities in glucose metabolism, particularly small deviations that indicate insulin resistance or pre-diabetes, are becoming more recognized as a factor that leads to adverse pregnancy outcomes [16, 17]. The vervet monkey, like many nonhuman primate species, spontaneously develop insulin resistance and hyperglycemia [18] and thus may have reproductive consequences. The goals of the current study were to examine pregnancies in a breeding colony of vervet monkeys and develop a growth nomogram across the three trimesters to serve as a baseline for the comparison of infants of differing survivorship. Early neonatal assessments were initiated to attempt to associate growth and survivorship. We also examined social and metabolic maternal factors that had potential influence on pregnancy outcome. Our objective was to determine similarities and differences between the vervet monkey and what is known about human pregnancy outcomes with the hypothesis that subclinical gestational diabetes may be contributing to macrosomia and reduced survivorship.

Materials and methods

Subjects

The study population consisted of 345 adult African green or vervet monkeys (Chlorocebus aethiops sabaeus) that form a multigenerational pedigreed colony (Vervet Research Colony). Animals are descendants of 57 original founders imported from St. Kitts, West Indies, and have remained a closed colony since 1985. During study, animals were housed as 16 matrilineal social groups in corrals with approximately 300 square feet indoors and 1,200 square feet outdoors. Both indoor and outdoor sections were fitted with elevated perches, platforms and climbing structures. Monkeys were fed commercial primate laboratory chow (Lab Diet 5038, Purina, St Louis MO) supplemented with fresh fruits and vegetables. All animals had ad libitum access to food, water and opportunities to exercise. Group sizes ranged from 11 to 39 animals, with 2 intact adult males included in each group. Unfamiliar males are rotated into each group every 3 - 5 years. After detecting a high infant mortality rate (Figure 1), strategies to combat this had included behavioral interventions and dietary modification to a lower glycemic index food (Lab Diet 5052) [2]. Behavioral interventions aimed to contravene kidnapping or abandonment of infants resulted in moderate improvements in postnatal survival, while dietary modification had no significant effect.

Figure 1.

Pregnancy outcomes in the vervet colony over the decade preceding and including the current study year. The outcomes were classified as abortions (■), neonatal death if the infant died at day 0 or 1 of life and includes the stillborn fetuses (), postnatal death if the infant died between day 2 and 30 days of life (□), or surviving if the infant was alive after 30 days (■). Colony size was controlled through vasectomy of various breeding groups, such that only 2008 represents all groups actively breeding. High infant mortality rates of 30% or higher were consistently seen prompting initiation of behavioral interventions to improve maternal-infant bonding in 2003, and a dietary intervention to increase fiber content and improve glycemic control in 2004. Interventions had small impact on overall breeding success, with minor improvements in postnatal death rates.

Pregnancy Detections

All females over 3 years of age had pregnancy status determined ultrasonographically (SonoSite 180, SonoSite Inc., Bothell, WA) at colony-wide health screenings done in April or August of 2008. The mean age of this population was 9.4 years (standard deviation of 4.8 years, range 3-24 years). Depending on time of conception in relation to the colony-wide evaluations, some females were not examined during pregnancy (n=2) and others had ultrasound examinations once or twice during gestation (n=164).

Vervet monkeys have a gestation period of approximately 157-168 days [9, 10, 19]. As the exact day of conception was unknown, all full term infant birth dates were considered day 163 of gestation [20]. Approximate day of gestation during ultrasound was calculated backwards from infant date of birth. Pregnant animals had gestational cavity diameter [19] and biparietal diameter measured (distance from parietal bone to parietal bone, Figure 2) [21]. Fetal femur length was measured if an adequate image was obtained [21]. The ultrasonographic measures were reviewed for accuracy using saved images (KK).

Figure 2.

Example of fetal biparietal diameter measurement by ultrasound. These measures were collected on all pregnancies at least once for the development of a fetal nomogram. This example is from a fetus that is 50-60 days estimated gestational age.

Neonatal Assessment

Physical examinations with body weight and head circumference measurements were conducted on infants at 4-10 days of age (with the exception of one animal that was evaluated at 13 days of age). If an infant died prior to the initial physical exam, then body weight and head circumference were measured at necropsy. All dead infants underwent a full diagnostic necropsy by a board certified veterinary pathologist within 4 hours of being found. Aborted and stillbirth infants were confirmed by lung flotation. Placentas were collected, if available, for histopathological analysis. Infant deaths were classified as dystocia, dehydration and malnutrition, inflammation, or miscellaneous (congenital defects, undetermined etc.) based on necropsy results.

Newborn infants were visually monitored twice daily for the first two weeks of life to assess maternal care. Allomothering is common in this species thus behavior interventions (reuniting the mother and infant outside of the social group for up to 16 hours) were conducted only if the health of the infant was at risk as a result of excessive allomothering. After two weeks of age all infants were monitored routinely with the rest of the colony.

Maternal assessment

Monkeys were fasted overnight, prior to sedation with intramuscular ketamine (8 -10 mg/kg) to facilitate the collection of blood samples, ultrasonography and morphometric measurements (n=266 adult females). Bodyweight for each monkey was obtained (kg), and sitting height was measured as the distance from the crown to the bottom of the pubic bone (cm). An index of body mass index (BMI) was calculated from the weight (in kg) divided by length (in meters) squared. A flexible tape measure was placed around the monkey’s abdomen at the level of the umbilicus to measure waist circumference (cm). Blood samples were collected from the femoral vein and placed on ice until samples were processed. Plasma and whole blood were stored at −80C prior to analysis. Percent glycation of hemoglobin in whole blood (A1c) was measured by HPLC borate affinity column (Primus PDQ, Kansas City, MO) to assess long-term glycemic control. Fasting blood glucose was measured colorimetrically using the glucose oxidase method (Roche, Basel, Switzerland) and fasting plasma insulin concentrations were measured by enzyme-linked immunosorbent assay (ELISA) (Mercodia, Uppsala, Sweden). Social status of females were assessed by focal observations and scoring, as has been described [1].

Study procedures had been approved by the Wake Forest University Institutional Animal Care and Use Committee. All procedures adhered to the USDA guidelines for the care and use of nonhuman primates.

Data Analysis

Infants were classified into three groups based on the number of days they survived. The surviving group lived >30 days, the postnatal death group died between 2-30 days, and the neonatal death group died between 0-1 day. Historically, only a small percentage of infants died after 30 days of age [3]. Ultrasound measurements of biparietal diameter of surviving infants were used to develop a fetal growth nomogram by non-linear modeling of data (GraphPad Prism 4.03, GraphPad Software Inc, La Jolla, CA). In creating a simple table, the upper confidence limits were used for upper gestational age boundary and lower confidence limits were used for lower gestational age boundary. Measured endpoints conferring high risk were defined as a value > 80^th percentile for surviving infants, or pregnant females. Group differences were assessed by one way analysis of variance, and association analyses by Pearson’s correlation coefficients. Sex differences in categories of mortality were compared by the mean of the chi-squared test. Standard linear regression modeling was used for ultrasonographic measures of gestational age.

For maternal analysis, females were additionally categorized as having a surviving infant (>30 days) or a non-surviving infant (<30 days). Along with metabolic characteristics above the 80^th percentile, a low social rank score was considered a maternal risk factor (scored as 3=low, 2=mid-ranking and 1=high ranking)[3]. Odds ratios were used to compare the two maternal groups to each risk factor (Epi Info 3.5.1, Centers for Diseases Control and Prevention, Atlanta, GA). Maternal age (<4, 4-6, 7-9, or 10+ years) and parity (at least 1 prior successful birth surviving >30 days) were also compared between the two groups.

All analyses were computed using Statistica 8 software (StatSoft, Tulsa OK) unless otherwise indicated above, and significance for all tests was determined with alpha < 0.05.

Results

The historical infant mortality rate was on average 39% (Figure 1) despite environmental changes. During the study period, 153 births were recorded with a 34% mortality rate. The temporal breakdown of the deaths revealed primarily neonatal deaths (n=29, 56%), then postnatal deaths (n=18, 35%) and abortions (n=5, 9%). Abortions were determined at necropsy as stillborn and developmentally premature and are omitted from further maternal and fetal analyses.

Fetal Growth

A fetal growth nomogram was developed to allow detection of in utero growth differences between surviving and non-surviving infants. Gestational cavity diameters (n=36; Figure 3a) was only useful in categorical detection of pregnancy with no differences seen with respect to infant outcome. Femur length could potentially be useful in determining gestation; however, the data is limited from this study for appropriate statistics, and there appeared to be no apparent pattern between infant outcome (n=23; Figure 3b). The small sample size for fetal femur measurements was due to time constraints during ultrasound which did not allow optimization of images. Biparietal diameter was most reliable (n=95) and adequately related size to gestational age within a 10 day window (Figure 4, Table 1). Post-study assessments in prospective pregnancy detections of 25 females confirmed this window of accuracy. No ultrasonographic measure was found to be significantly different between surviving and non-surviving infant groups.

Figure 3.

a Scatterplot of gestational cavity diameter by approximate gestation period in days. Fetal outcome (neonatal death including stillborn and death in the first 48 hours of life [], or postnatal death being infants that died between 2 and 30 days of life [○]) could not be discriminated by size of the gestational cavity from infants that survived [●] due to large amounts of variability. First reliable detection of pregnancy is possible between 15-20 days with a cavity size of 1-2mm diameter. Regression analysis resulted in r²=0.44, p<0.001 (n=36).

b Scatterplot of femur length by approximate gestation period in days. Fetal outcome (neonatal death including stillborn and death in the first 48 hours of life [], or postnatal death being infants that died between 2 and 30 days of life [○]) could not be discriminated by femur size from infants that survived [●]. Linearity of size versus age was excellent (r²=0.80, p<0.001) however the required time to generate quality images was a limiting factor in large colony screening situations (n=23).

Figure 4.

Scatterplot of biparietal diameter by approximate gestation period in days for surviving infants (n=95 data points, r²=0.95, p<0.001). Linearity of size versus age was excellent in the first 2 trimesters, with increasing variability in the third trimester.

Table 1.

Chart of predicted biparietal diameter measures based on regression modeling of data presented in Figure 2

Gestational Age (days)			Biparietal Head Diameter (cm)
30	to	40	0.07	to	0.78
40	to	50	0.46	to	1.25
50	to	60	1.00	to	1.70
60	to	70	1.49	to	2.14
70	to	80	1.94	to	2.55
80	to	90	2.34	to	2.92
90	to	100	2.71	to	3.23
100	to	110	3.02	to	3.54
110	to	120	3.34	to	3.82
120	to	130	3.65	to	4.03
130	to	140	3.87	to	4.24
140	to	150	4.08	to	4.43
150	to	160	4.23	to	4.61
160	to	170	4.33	to	4.70

Neonatal characteristics

Infants that died in the first 2 days of life (neonatal death) were found to be of comparable size to the postnatal or surviving infant groups (Table 2). In contrast, infants dying in the postnatal period were significantly smaller in bodyweight than those that survived (p=0.003) or those that died in the first 2 days of life (p<0.001). This growth retardation among animals dying postnatally was reflected also in their significantly smaller head sizes as compared to neonatal death despite their older age (p<0.05, Table 2). Using the surviving infants as the reference group, the 80^th percentile for bodyweight was 380g and 420g for females and males, respectively. All infants above the 80^th percentile in the neonatal death group were stillborn and necropsy reports were consistent with macrosomia, as 72% of neonatal deaths listed dystocia as having a causative role. Thirty eight percent of neonatal deaths, 5.6% of postnatal and 16% of surviving infants were over the 80^th percentile for sex-specific body weight.

Table 2.

Mean (standard error of the mean) values for available characteristics from infants that were classified as neonatal death (died between 0 and 2 days), postnatal death (2 – 30 days) or survived (>30 days). Different letter superscript indicates statistically significant differences between groups

	Surviving n=101	Neonatal Deaths n=29	Postnatal Death n=18
Infant Bodyweight (g)	371 (4.97) a	388 (16.2) a	320 (12.9) b
Infant Head Circumference (cm)	17.7 (0.07) ab	17.8 (0.22) a	17.2 (0.17) b
Sex Distribution (ratio M/F)	1.1 : 1a	3.14: 1b	1.25 : 1a

Male sex conferred significantly increased risk for neonatal death (p<0.05; Table 3) with a ratio of more than three males to every female in this category. The sex difference in survivorship was particularly apparent in the stillborn animals. Male infants comprised 75% of this population, and 39% of these males were >80^th percentile for body weight. This led to 79% of stillborn monkeys having evidence of dystocia at necropsy. Gross and histopathological assessments indicated 21% had inflammatory infiltrates of fetal or placental tissues and only 2 cases of confirmed infection were seen in this group, of which neither were identified as agents such as listeriosis or brucellosis, known to induce abortion or stillbirth [22, 23].

Table 3.

Maternal risk factors measured in August from females with surviving (> 30 days) and non-surviving (< 30 days) infants. Data shown as the proportion of at-risk females/total females with calculated odds ratio (95% confidence interval) for survival outcome

	Risk definition	Mothers at-risk Surviving	Mothers at risk Non-surviving	Odds Ratio (95% CI)	p
Glycosylated A1c %	> 3.8	8/81	12/40	3.72 (1.58 – 10)	p=0.02
Fasting blood glucose mg/dL	> 99	14/80	6/40	1.8 (0.72 – 4.5)	p=0.31
Fasting blood insulin uIU/L	> 118	19/81	6/40	0.69 (0.25 – 1.9)	p=0.68
Body mass index kg/m2	> 30	17/81	10/40	1.24 (0.5 – 3.1)	p=0.81
Social Rank (3=low, 2=mid, 1= high)	3	22/122	8/34	2.27 (0.86-5.97)	p=0.04

Maternal Characteristics

Maternal metabolic data was collected on all pregnant animals only once during gestation (n=121). During data collection 7.4% of animals were in first trimester, 25.6% were in second trimester, and 67% were in third trimester gestation. Average maternal hemoglobin A1c was significantly higher in females with non-surviving infants (n=40) compared to those with surviving infants (n=81), with a biologically significant difference in value of about 0.3% (p=.04; Figure 5). However glucose, insulin, and BMI were not different between categories. Maternal hemoglobin A1c above the 80^th percentile was also significantly associated with increased risk of infant death (OR = 3.72, [CI 1.58-10] p=.018) while insulin, glucose, and BMI above the 80^th percentile were not (Table 3). Although the maternal hemoglobin A1c risk factor conferred significant risk of infant death, elevated maternal hemoglobin A1c that reached the risk definition was not seen in the majority of infant deaths (30%). Maternal hemoglobin A1c was significantly and negatively correlated with the head circumference of dead infants (r=−0.31, p<0.05, n=35), but not with surviving infants (p=0.08, n=86). Maternal glucose and insulin were not significantly correlated with infant head circumference. No correlations were seen between infant body weight and maternal hemoglobin A1c, glucose, or insulin. Maternal weight did correlate with infant weight (r=0.21, p=0.03) as expected based on prior studies [18].

Figure 5.

Long-term glycemic control of the mother as assessed as maternal A1c% was significantly different (p=0.04). Lower average blood glucose levels reflected as lower A1c% was seen in females that produce surviving infants as compared to females that did not produce a surviving infant.

Cumulative risk scores were calculated from the assessment of the mother’s A1c, glucose, insulin, and BMI. A score of 1 was given if the data point was > 80^th percentile of the total population, such that the total possible was a score of 4. Proportions of mothers with a score of ≤ 2 were not different between surviving and non-surviving infant groups. However mothers with 3 metabolic risk factors clustered in the non-surviving group (12.5% vs. 2.5%) such that compounded metabolic ill health conferred significant risk for infant death (p=0.03). Metabolic characteristics of the mothers by infant outcome category are shown in Table 4, demonstrating the most pronounced changes in glycemic status in the neonatal death category.

Table 4.

Breakdown of maternal risk factors by group, showing the mean (± SEM) value with p-values for overall analysis of variance. Unlike superscripted letters indicate significant group differences resulting of post-hoc testing when overall p was ≤ 0.05. .Data reported from April assessments, when females were in earlier stages of pregnancy with the exception of A1c which measured in August reflects a 3 –month retrospective average glucose. Categories are mothers with infants that underwent neonatal death (died between 0 and 2 days), postnatal death (2 – 30 days) or survived (>30 days)

	Neonatal Death	Postnatal Death	Surviving	p
Glycosylated A1c %	3.87 (0.21) b	3.67 (0.20) ab	3.48 (0.03) a	p=0.02
Fasting blood glucose (mg/dL)	86 (20) a	53 (2.6) b	79 (4.8) a	p=0.05
Fasting blood insulin (uIU/mL)	20.4 (6.8)	23.4 (5.33)	50.5 (7.23)	p=0.07
Bodyweight (kg)	5.04 (0.22)	4.73 (0.14)	4.96 (0.09)	p=0.43
Body mass index (kg/m2)	23.5 (2.26)	24.4 (0.67)	25.4 (0.38)	p=0.22
Waist Circumference (cm)	31.7 (0.81)	32.5 (0.71)	32.1 (0.37)	p=0.76

Social rank also had a significant effect on infant outcome, with low social rank conferring increased risk of infant death (OR=2.27 [CI 0.86 – 5.97], p=0.04). The low social rank of the mother may lead to increased kidnapping events by higher ranking females and a lesser ability for infant retrieval. Kidnapping leads to increased time off the mother and deaths related to dehydration and malnutrition, as assessed at necropsy. In cases where dehydration and malnutrition were listed as the cause of death (89% of cases), 70% of available maternal behavior records had indicated kidnapping episodes. Social rank category and A1c measures were not significantly associated (p=0.25).

No association of maternal age, parity or matrilineal group had significant effects on infant mortality rates.

Discussion

The pregnancy failure rate (34%) in this colony exceeded that typically reported captive vervets, about 16% [9, 24, 25]. However comparable death rates have been occasionally reported in this species (>28% [20, 25] while substantially higher rates have been reported for some colonies of pigtailed macaques [7] and tamarins [26]. Infant deaths in this colony were associated most clearly with maternal physiological (glycemic measures) and behavior (social status) characteristics. In utero morphometric measures, suitable for screening primate populations, were not successful at predicting outcome. The exact causes of infant death remain obscure, which is also the situation in women, particularly in relation to stillbirth [6] and sudden infant death syndrome [12]. Furthermore, as among women, male sex of the fetus was associated with being stillborn and macrosomic [27] while infectious causes led to very few deaths [11]. Only 2 cases of neonatal death were attributed solely to infectious causes. In infants that survived the initial 48 hour period, death was associated with significantly reduced body size measurements.

We found that long-term glycemic control significantly affected infant survivorship. Colony management practice has been to remove overtly diabetic females for more intensive medical management. Only one vervet female was diagnosed with probable diabetes during assessment. Gestational diabetes and more importantly, increases in blood glucose that are considered within the normal reference range for women, are associated with complications at birthing. In studies of women where the highest blood glucose category approximated the risk cut-off used in this study (99 mg/dL), 20% of the study population experienced adverse pregnancy outcomes, but there was no definite threshold for this effect. Incremental glucose concentrations led to increased risk of adverse pregnancy outcomes [15, 17]. Abnormal glucose tolerance also leads to increased risk of being born pre-term and/or hypoxia-induced encephalopathy [16]. Encephalopathic infants are expected to be less vital, have abnormal behavior [28], and thus potentially experience more maternal rejection.

The interaction between environmental stress, including social stress, and physiology is well documented in nonhuman primates [29, 30]. Social status and metabolism are inter-dependent with chronic stress leading to hypercortisolemia from adrenal hyperplasia [31], central obesity, and insulin resistance in macaques [32]. Social status did not associate with A1c values in this study, which may be a function of the relatively small range of A1c values actually observed in the pregnant females. The observation that low social 276 rank resulted in significant risk for having a non-surviving infant is important, with 45% of the low social status females in the colony experiencing a stillbirth or neonatal death, as compared to 22% or 28% in the mid- and high-ranking females, respectively and recapitulates findings in smaller vervet populations [3]. Ovarian dysfunction in vervet monkeys may be common, with up to 25% of habituated captive females demonstrating irregular menstrual cycles [33] and socially subordinate macaques demonstrating as high as 50% prevalence of ovarian dysfunction and reduced circulating estrogen [31, 34]. Interestingly, the risk of pre-term and very low birth weight babies has been significantly and independently related to increased social stresses, assessed as discrimination in women [13, 35].

A unique feature of the vervet colony is the preservation of stable groups that conserves matrilines. This reduces female-female aggression in breeding groups [24] and preserves social relationships, particularly amongst mothers and daughters, which have been shown to enhance infant survival in baboons [14]. These protective effects of the social structure did not seem apparent in our breeding groups and so further consideration regarding vervet-specific behavior was considered. The group sizes in this study range more than 3-fold in size, with the average size approximating 26 monkeys. Observations of wild vervets in their original St. Kitts location estimates natural groups to be on average 15 monkeys (range 5-40), with an estimated density of 1 monkey per 9 acres [36]. The captive conditions increase group size and density, which may be considered an adverse environment and affect reproductive performance according to life history approaches [37]. Socially subordinate females may be at increased risk for stress these captive conditions as greater numbers of higher ranking females are in close proximity. In this bioanthropological concept, reduced fertility or mothering of neonates may be a trade-off to maintain a population density that promotes the preferred social interaction and perceived energy resources by at-risk mothers. This hypothesis has been applied to vervet maturation where resource availability did modulate reproduction, growth, and estrogen profiles [37, 38].

Macrosomia was seen in a large proportion of stillborn vervet monkeys, many of which did not have high maternal A1c risk. This is also the pattern in women and baboons who demonstrate high recurrence rates of macrosomia in non-diabetic pregnancies [8, 39]. Pig-tail macaques similarly demonstrate that high risk females have a history of multiple fetal deaths [7]. We note a significant sex difference in proportions of stillbirth with males being 75% of cases, and almost half of stillborn infants being macrosomic. Male sex and abnormal maternal glycemic control is known to increase risk for stillbirth by nearly 50% in women [5] however the exact causes of death remain unexplained in 25-40% of cases [6]. Our pathology assessments corroborate this finding with dystocia representing the majority of stillborn diagnoses however 20% of cases had no clear etiology. Review of male fetal differences indicate higher frequency of renal abnormalities, placentitis, and growth restriction [40]. Feral African vervet monkeys have also shown bias towards loss of male infants when bred in captivity, although the difference was seen in both stillborn and postnatal death categories [25]. Importantly for the perinatal period, male infants are more susceptible to behavioral abnormalities as a result of birthing trauma [40, 41] which in the vervet monkeys may be reflected in the poor mothering behavior.

Another unique characteristic of this vervet colony includes its’ closed nature, such that for up to 8 generations, no new genetic influence has been introduced. Significant heritability exists for many morphological and metabolic endpoints [18] and it might be reasonable to speculate that inbreeding contributed to the high rates of dystocia and stillbirth, as has been observed and quantified in dairy cattle [42]. The mean inbreeding coefficient of the vervet colony has been estimated at 3.3% [43] which is low and, based on the bovine model, would contribute < 1% to the stillbirth and dystocia incidence [42]. Inbreeding coefficients at the <10% level, such as those estimated in the vervet colony are also associated with an increase in male to female sex ratio for offspring [42].

Growth restriction is a clear risk factor for survivorship however we are limited by having variability in regards to age-associated body weights. What can be concluded is that small weight and restricted perinatal growth is a significant factor associated with survival and should be monitored closely. The construction of a nomogram for this population of vervet monkeys allows improved detection of small or large-for gestational age fetuses, and enables estimated parturition dates which are most accurate in the first trimester, and early second trimester. Maternal body mass index change of greater than 25% from this point to third trimester can be useful in predicting macrosomia [44]. The nomogram developed from this dataset indicates that this St. Kitts sourced population is larger than African-sourced vervet monkeys based on comparable biparietal diameter measurements [19]. This supports a founder population effect that may be compounded by the effects of inbreeding.

Low body weight is highly associated with neonatal and postnatal mortality, most commonly associated with poor mothering. We are limited in not having true birth weights on the day of delivery and thus weight is confounded by weight loss post-partum therefore classifications by low birth weight were not done. Our study points to maternal factors being important, however other primate colonies have identified sire effects unrelated to fertility as also being an important determinant in infant outcome and low birth weight [7]. In this study population, the poor infant survival rate has had a long history that spanned multiple male siring periods [3], so we conclude that if sire effects are contributory that they are a result of inbreeding. Data were not closely evaluated for genetic variation that might have contributed to the mortality that we observed. Furthermore, no substantial interventions – such as removal and sequestration of pregnant females – were employed that might have reduced postnatal wastage and tested causal hypothesis. However, behavioral interventions had been attempted for several years and did not appear to have a substantial effect on neonatal death.

In summary, we have demonstrated that development of a fetal growth nomogram is useful for predicting birth date however does not allow the identification of at-risk pregnancies. Maternal characteristics of A1c and social status can be used to classify high-risk pregnancies. Male infants were at greater risk of being stillborn, with dystocia being the predominant cause, while glycemic control and genetic factors are also likely involved. Low social status and the associated chronic stress negatively affected survivorship, and most of these aforementioned factors are also seen in studies of women [5, 35, 45]. This colony demonstrates the complexity of causes associated with infant mortality and will be a useful tool for in-depth investigations into the maternal effects on survivorship.