Placental changes and obstetric outcomes among women with diabetes mellitus in pregnancy in Kano, Nigeria: a case control study

Abstract

Background

Diabetes mellitus may induce recognizable changes in the placenta and may lead to adverse obstetric outcomes. Placental examination may identify these changes.

Aim

To compare ultrasonographic and histo-morphological changes in the placentae of diabetic with normo-glycaemic women, as well as determine and compare the obstetric outcomes of the women.

Methods

This was a case-control study of 65 pregnant diabetic women matched with 65 non-diabetic pregnant women from 34 weeks to 40 weeks’ gestation. All the participants had ultrasound scanning (USS) for placenta changes done within the last seven days of delivery and histo-morphological placental examination postpartum. Obstetric outcomes and relationship between ultrasonographic and histo-morpholgical changes were also determined.

Results

The diabetics compared to non-diabetic had significantly higher mean ultrasonographic placental weight (610.0 ± 81.54 versus 513.2 ± 74.25, p < 0.001), volume (1001.5 ± 359.80 versus 721.7 ± 186.16, p < 0.001), diameter (14.6 ± 2.20 versus 13.7 ± 2.08, p = 0.015), central thickness (4.8 ± 1.18 versus 4.2 ± 1.34, p = 0.015), umbilical cord coiling index (0.6 ± 0.23 versus 0.3 ± 0.13, p < 0.001), vascularisation index (0.6 ± 0.09 versus 0.5 ± 0.08, p < 0.001), flow index (1.0 ± 0.30 versus 0.8 ± 0.20, p < 0.001) and vascularisation-flow (2.6 ± 0.61 versus 2.1 ± 0.38, p < 0.001) index. Histomorphology showed diabetics had significantly higher placental weight (764.9 ± 159.26 versus 570.0 ± 92.73, p < 0.001), volume (737.7 ± 179.85 versus 536.9 ± 83.00, p < 0.001), diameter (19.6 ± 1.75 versus 18.3 ± 2.05, p < 0.001), central thickness (2.9 ± 0.65 versus 2.4 ± 0.58, p < 0.001), villous immaturity, villous edema, chorangiosis, fibrinoid necrosis and stromal fibrosis. The diabetics were also significantly more associated with caesarean deliveries, fetal macrosomia, still births, poor first minute Apgar score, hypoglycaemia at birth, respiratory distress syndrome and SCBU admissions compared to non-diabetics. In both diabetics and non-diabetics, there was a significant relationship between the placental weight (P = 0.002), volume (P = < 0.001), diameter (P = 0.02) and calcifications (P = < 0.001) compared with the actual histomorphological assessment.

Conclusion

Placental changes and adverse obstetric outcomes occurred more significantly in the diabetic group. Some of the histo-morphological and ultrasonographic changes were relatable. Sonographic placental examination can, therefore, be useful in the management of diabetes in pregnancy.

Introduction

Diabetes mellitus (DM) is an endocrine disorder in pregnancy that results from chronic hyperglycaemia due to absolute or relative insulin deficiency [1]. Hyperglycaemia in pregnancy may be either diabetes in pregnancy (DIP) – when the patient is a known diabetic predating pregnancy or when the conditions for the diagnosis of diabetes outside pregnancy are met; or gestational diabetes mellitus (GDM) – when hyperglycaemia is first detected during pregnancy and does not meet the diagnostic criteria for DIP [2]. The global prevalence of DM in pregnancy ranges from 3–10%. [1] In Nigeria, reported values are 13.8% in the North-west3, 16.7% in the north-central [4] and 5.2% in the South-south [5]. The reported prevalence of GDM is 11.9% in Europe [6].

Women with DM in pregnancy have a 12.7% and 44.3% increase in the rates of occurrence of pre-eclampsia and caesarean section respectively [7]. They also have a 5–20% increase in the risk of infections, birth injuries and postpartum haemorrhage as compared to non-diabetics [1]. Similarly, congenital anomalies, intrauterine fetal death, and recurrent miscarriages seen commonly in pregnancies complicated by DM, can affect the emotional status of the mothers of affected fetuses [1]. Also, intrauterine growth restrictions and fetal macrosomia might have long term implications - notably an increased risk for type-2 DM, cardiovascular diseases and other metabolic disorders in affected fetuses [1].

The placenta is a vital organ of intrauterine life interphasing fetal and maternal circulations [8]. It serves as a conduit for transfer of nutrients and waste product between the fetus and the mother [9–11]. Poor glycaemic control in early gestational age affects the development and damages the structure of the placenta [8, 9–11]. However, when the insult occurs later in pregnancy, placental function is mostly affected rather than its structure [10–12]. The placentae of diabetic patients are reported to be heavier and larger than that of appropriately matched normoglycaemic patients [13–15]. Microscopic changes reported include villous oedema and immaturity, chorioangiosis, fibrinoid necrosis, amongst others [13, 14].

The clinical importance of postpartum histo-morphological placental changes are well established however, those of antenatal ultrasonographic placental changes are yet to be fully known. Placental changes elicited at postpartum placental examination serves more in the prevention of adverse obstetric outcome in future pregnancies [14, 16]. To protect an index pregnancy from adverse obstetric outcomes, there is a need for such modalities such as ultrasonography that can monitor the developing fetus in-utero, and also provide an important means of pre-empting for any eventualities related to the fetus and mother [17–19]. These ultrasonographic changes may also relate to histomorphological changes. For example, placental calcification and lakes correlate with villous immaturity and increased placental thickness [17, 18].

It therefore follows that a combination of in utero and postpartum placental examination can provide valuable information that can be useful in prevention and management of complications in both the mother and the fetus. Also, as there are inconsistencies on the clinical significance of ultrasonographic changes of placenta in diabetic women, and in their relationships to histo-morphological changes, studies that explore these subject fields will be invaluable. The objectives of this study were to compare ultrasonographic and histo-morphological changes in the placentae of diabetic with normo-glycaemic women, as well as determine and compare the obstetric outcomes of the two groups of women.

Methodology

This was a prospective case-control study conducted at the department of Obstetrics and Gynaecology of Aminu Kano Teaching Hospital (AKTH) and Murtala Mohammed Specialist Hospital (MMSH) situated in Kano, Kano state, Northwest, Nigeria. Booking and deliveries at the Obstetrics and Gynaecology department (AKTH) are about 6347 and 2999 per year respectively [20] while 24,000 deliveries was reported at MMSH per annum [21].

The cases comprised of diabetic pregnant women who delivered in the Department of Obstetrics and Gynaecology of Aminu Kano Teaching Hospital (AKTH) and Murtala Muhammed Specialist Hospital (MMSH). Controls were non-diabetic pregnant women who delivered at AKTH and MMSH during the period of the study (December 2018 to June 2019).

Consented booked and unbooked diabetic and non-diabetic pregnant women between 34 and 40 weeks of gestation using last normal menstrual period or ultrasound scan done before 20 weeks; who delivered at AKTH or MMSH irrespective of the fetal condition at the time they presented for delivery were included in this study.

Pregnant women who were of uncertain gestational age, less than 34 weeks or postdated or who had any of the following medical condition were excluded from the study; post-date pregnancies, hypertensive disorders in pregnancy, rhesus iso-immunisation, anaemia in pregnancy, twin pregnancy, antepartum haemorrhage, ssickle cell anaemia in pregnancy, pregnant women with documented antiphospholipid syndrome, pregnant women with documented two or more episodes of malaria infestation and who were not on malaria prophylaxis.

Over 30 weeks period, women were recruited via simple random sampling technique until the sample size was met.

The booked patients were recruited from the antenatal clinics. During each of the antenatal clinic days, the case folders of the women who presented for ANC that day were obtained from the records and checked. The case folders of those that met the inclusion criteria identified were separated into diabetic (cases) and non-diabetic (control) pregnant women. Women who have been diagnosed diabetic with evidence of being treated by any of these: diet, exercise, oral glucose lowering agents and insulin, were regarded as having diabetes in pregnancy (cases). For those that were not known to be diabetic, a blood test was done. The non-diabetic status was based on either a fasting blood sugar of less than 5.6 mmol or random blood sugar less than 7.8 mmol [1]. A history of the patient’s last meal was used to determine the suitable test. The value of blood glucose was considered as fasting if the last meal of the patient was eight hours or more and random if her last meal was less than eight hours [22].

They were counselled on the study protocol and a written consent was obtained in the clinic as soon as their consent was given. As the cases were recruited, the next non-diabetic woman, matched for age (± 1 year), parity ((± 1) and gestational age (± 1 week) was also recruited as control after a normal fasting blood sugar (FBS) or random blood sugar (RBS) was obtained.

Their antenatal care at recruitment and subsequent visits as well as labour were according to the protocol for the management of DM and non-diabetic women in pregnancy in each of the hospitals. All the booked study participants were followed up via phone call to ascertain when they go into labour. Caesarean section was performed for obstetric indication.

For the unbooked participants, counselling on the study protocol and obtaining informed consent were done at presentation in labour ward when they were not having labour pains. They had packed cell volume and indirect coombs test done (for those who are Rhesus negative) to exclude anaemia and rhesus isoimmunisation respectively since there were no prior antenatal record to identify such in them. The same definitions stated above used to group booked patients into cases and controls were used for the unbooked.

A coded proforma written in both English and Hausa languages was administered to the patients’ to record socio-demographic and reproductive characteristics (excluding names) obtained from each woman recruited during the study period. The placental changes, obstetric outcomes and glycated haemoglobin (HbA1C) were recorded and added to the proforma for the final statistical analysis.

When the booked and unbooked diabetic pregnant women presented for scanning or in labour ward before delivery respectively, 5 ml of venous blood was taken into a heparinized sample bottle for estimation of glycosylated haemoglobin (HbA1C) by a chemical laboratory Scientist.

Scanning was done within 14 days preceding the expected date of delivery by a consultant radiologist. For unbooked patients, this was done at recruitment after consent was obtained. Those that did not deliver within 7 days of the initial scanning had a repeat scan. Scanning was done with participants lying supine with left lateral tilt using 3.5 megahertz probe with Doppler facilities. Ultrasonographic outcome of interest included placental texture, weight, diameter, central thickness, volume, calcification, umbilical cord coiling index, placental lakes and shape as well as placental vascularisation, flow and vascularisation- flow indices.

Within a minute of delivery of the baby, the umbilical cord was clamped. Placenta with cord and membranes of each participant were collected without trimming off the membranes and gross features of interest were assessed as follows:

• Shape: The shape was categorized as discoid, oval or irregular. Discoid placentas were considered normal whereas irregular and oval placentas were considered abnormal.

• Diameter: The placenta was gently placed on a flat tray in such a way that cotyledons were visible and did not rest on one another. The first maximum diameter was measured and the second maximum diameter was taken at right angles to the first one using a 60 cm flexible non-stretchable tailoring tape. The mean of the two measurements was considered as the diameter of the placenta and expressed in centimetres.

• The thickness of the placenta was measured from its centre by piercing a silvery coloured knitting needle and the embedded part which represents the thickness was measured with a tape measure and expressed in centimetres.

• Weight: The placenta was placed gently and weighed on the balance using a weighing machine (SALTER, MODEL 180, Year 2014, Sensitivity 1 g= 1 g + 0.1), in grams, made in India.

• Volume: The volume was obtained by water displacement method in a graduated two litre transparent plastic beaker. One thousand millilitres of water was placed in the beaker, then the placenta was put inside the water. The water level rose and the maximum point was noted. 1000 ml was then subtracted from new volume. The volume of water displaced by the placenta which became the placental volume in millilitres was noted.

Placenta was cut along the maximum diameters into two halves and a tissue specimen (2 × 2 cm) was collected in a tissue cassette in such a way that it incorporated the whole central thickness, membrane and the cord. The specimen was preserved with 10% neutral buffered formalin fixative and immediately sent to the histopathology laboratory for analysis. The fixative was thrice the volume of the specimen, [13, 16, 23] and the specimen bottle was coded serially to ensure the Pathologist was blinded. Thereafter, routine paraffin embedded sections were cut at 3 microns on glass slides using a microtome and slides were stained with Haemotoxylin & Eosin at the Histopathology department of AKTH.

Microscopic study for the presence of villous immaturity and oedema, chorioangiosis, infarction, calcification, fibrinoid necrosis, stromal fibrosis, syncytial knots and increased membrane thickness were carried out and photo micrographs were taken.

Perinatal outcome measures

• A.
  Primary

  • First and fifth minutes Apgar scores: these were scored at the first and fifth minutes of delivery as described by Virginia Apgar. For analysis, this was considered a categorical variable with scores of 7 and above considered normal, while scores less than 7 considered below normal.
  • The birth weight: Within the first 30 min of birth, the baby was placed on a weighing balance and the weight was measured in kilograms. This was analysed as a normally distributed continuous variable. Abnormal birth weight = Macrosomia (≥ 4 kg), low birth weight (< 2.5 kg).
  • Fetal condition: Stillbirth is any baby without gross body and breathing movements and who had no heartbeat at birth. This was analysed as a categorical variable.
• B.
  Secondary

  • Respiratory distress syndrome: A Paediatrician made the diagnoses using both clinical and radiological assessments. This was analysed as a categorical variable.
  • Fetal blood sugar: Abnormal blood sugar level at delivery = hyperglycaemia (> 5.8mmol/L) + hypoglycaemia (< 2.2mmol/L). This was analysed as a categorical variable.

Babies admitted to special care baby unit were followed up for 7 days via phone calls and these outcomes that were exclusively from diabetes mellitus as stated by the paediatrician were recorded by the principal researcher and/ or a trained assistant on a proforma.

Maternal outcome measures

• Mode of delivery: This was either via vaginal route or Caesarean section. This was analysed as a categorical variable.

• Preterm delivery: Delivery before 37 completed weeks but at or greater than 34 weeks using the last normal menstrual period (LNMP) or ultrasound scan (USS) done before or at 20 weeks of gestation. This was analysed as a categorical variable.

• Glycaemic control: Good glycaemic control- HbAIc ≤ 6.5% while poor glycaemic control- HbAIc ≥ 6.5%. [1] This was analysed as a categorical variable.

• Unbooked patients: Any woman who did not attend antenatal clinic in either AKTH or MMSH on more than two occasions.

Statistical analysis

Data obtained were analysed using the statistical product and service solution 23 Inc. (SPSS Inc, Chicago, IL USA, 2015). Chi-square or fishers exact tests (with Monte carlo simulation) and student t- test were used to determine association between categorical and continuous variables respectively. Odd ratios (OR) with 95% CI was calculated to assess strength of associations. Pearson’s correlation and Cramer’s V test were used to determine the relationship between ultrasound scan and the corresponding histo-morphological placental measurements that were continuous and categorical variables respectively. P - value less than or equals 0.05 was taken as statistically significant. The accuracy of the ultrasound scan assessment of each variable was determined by the total number of correctly identified placentas as normal and abnormal, that is true positives and true negatives divided by the total number of placentas and expressed as a percentage. Results were presented in tables and text forms.

Results

The 30–39 age range accounting for 46.2% of the diabetic group and 44.6% of the control group was the most common age range in both groups. Grandmultiparity (46.2% versus 44.6%) was also the most common parity in the study and control groups respectively. Miscarriages were reported significantly more in the diabetes group (p < 0.001). Women in the diabetic group reported more miscarriages 24 (36.9%) compared to non-diabetics 8 (12.3%). Over three quarter of the diabetic patients had poor glycaemic control (Table 1).

Table 1.

Socio-demographic characteristics of the respondents

Variables	Cases =65n(%)	Control=65n(%)	Chi-square/Fisher exact	P-value
Age (years)
20-29	28(43.1)	30(46.2)	0.16	0.94
30-39	30(46.2)	29(44.6)
≥40	7(10.7)	6(9.2)
Educational status
Qur’anic	2(3.1)	5(7.7)	11.72*	0.006**
Primary	18(27.7)	4(6.2)
Secondary	21(32.3)	24(36.9)
Tertiary	24(36.9)	32(49.3)
Religion
Islam	56(86.2)	52(80.0)	0.88	0.35
Christianity	9(13.8)	13(20.0)
Ethnicity
Hausa	50(76.9)	48(73.8)	10.02*	0.028**
Fulani	4(6.2)	8(12.3)
Igbo	5(7.7)	1(1.5)
Yoruba	0(0.0)	3(4.6)
Others***	6(9.2)	13(20.0)
Parity at delivery
Primipara	9(13.8)	10(15.4)	0.70	0.97
Multipara	26(40.0)	26(40.0)
Grandmultipara	30(46.2)	29(44.6)
Past history of miscarriage
Present	24(36.9)	8(12.3)	10.61	< 0.001**
Absent	41(63.1)	57(87.7)
Booking Status
Booked	50(76.9)	51(78.5)	0.04	0.83
Unbooked	15(23.1)	14(21.5)
Gestational age at delivery (weeks)
34-37(or <38)	37(56.9)	35(53.8)	0.13	0.72
38-40	28(43.1)	30(46.2)
Glycaemic control
Good	15(23.1)	-
Poor	50(76.9)	-
Mean ± SD	7.05±1.0

Keys: * = Fisher’s exact test, ** = Statistically significant, ***Others = Tiv, Ibibio, Igala, Orhobo, Gwari, Kanuri

Ultrasonographic placental features in diabetics and non-diabetics pregnant women are as shown in Table 2. Diabetic participants had significantly higher rates of all the abnormal ultrasonographic placental changes except for shape, placental lakes and calcification compared to non-diabetic participants.

Table 2.

Ultrasonographic placental features in diabetic and non-diabetic participants

Ultrasonographic features	Cases=65n(%)	Control=65n(%)	c2/ fishers/t-test	P-value	Odds ratio(95%CI)
Weight(g)
450-600	31(47.7)	59(90.8)	28.31†	<0.0001**	10(4.09-28.47)
<450 and >600	34(52.3)	6(9.2)
Volume(ml)
200-950	42(64.6)	54(83.1)	5.74†	0.02**	2.69(1.18-6.13)
<200 and >950	23(35.4)	11(16.9
Shape
Abnormal	31(47.7)	25(38.5)	1.13†	0.29	0.87(0.41-1.82)
Normal	34(52.3)	40(61.5)
Calcification
Present	4(6.2)	24(36.9)	18.21†	<0.0001**	0.11(0.04-0.35)
Absent	61(93.8)	41(63.1)
Placental lakes
Present	32(49.2)	37(56.9)	0.77†	0.38	0.73(0.37-1.45)
Absent	33(50.8)	28(43.1)
Mean diameter± SD(cm)	14.58 ± 2.2	13.65 ± 2.1	2.47†††	0.02**	____
Mean central thickness± SD(cm)	4.83 ± 1.2	4.28 ± 1.3	2.47†††	0.02**	____
Umbilical cord coiling index
Abnormal	45(69.2)	11(16.9)	36.26 †	<0.0001**	11.05(4.79-25.47)
Normal	20(30.8)	54(83.1)
Mean vascularization   index ± SD	0.62 ± 0.1	0.52 ± 0.1	6.63†††	<0.0001**	______
Flow index
Abnormal	4(6.2)	0(0.0)	††	0.02**	_____
Normal	61(93.8)	65(100.0)
Mean ± SD	1.00±0.3	0.78±0.2	5.80†††	<0.0001**
Vascularization-flow index
Abnormal	18(27.7)	11(16.9)	2.18†	0.14	1.90(0.81-4.38)
Normal	47(72.3)	54(83.1)
Mean ± SD	2.60±0.6	2.12±0.4	5.37†††	<0.0001**
Occurrence of at least 1 ultrasonographic change
Present	58(89.2)	43(66.2)	9.99†	0.002**	4.24(1.66-10.83)
Absent	7(10.8)	22(33.8)

KEYS: ** = Statistically significant, Test statistic: c2 = †, Fisher’s = ††, t-test = †††

Abnormal shape = irregular + oval placental shapes

Abnormal coiling index = hyper (>0.59) + hypo (< 0.21) coiling of umbilical cord

Abnormal flow index: = high (> 1.00) + low flow index (< 0.60)

Abnormal vascularisation index = high (> 0.64) + low vascularisation (< 0.42)

Abnormal vascularisation – flow index = high (>2.72) + low vascularisation - flow index (<1.66)

Compared to the non-diabetic placentas, diabetic placentas had significantly higher proportion of most abnormal (mostly large) histo-morphological changes. There was a significantly higher proportion of diabetic placentas that had villous edema (Fig. 1), chorioangiosis (Fig. 2), fibrinoid necrosis (Fig. 3) and stromal fibrosis (Fig. 4). There were no significant differences between the diabetic and non-diabetic placentas in the rates of syncytial knots (Fig. 5), calcifications (Fig. 6), infarctions (Fig. 7) and abnormal thickening of the basement membrane (Fig. 8) (Table 3).

Fig. 1.

Section shows vascularized but edematous chorionic villi. Hematoxylin & Eosin x200

Fig. 2.

Shows chorangiosis characterized by increase in number of blood vessels(arrows) per chorionic villus. Hematoxylin & Eosin x200

Fig. 3.

Shows areas of fibrinoid necrosis with entrapped red and white blood cells. Hematoxylin and Eosin x200

Fig. 4.

Shows chorionic villi exhibiting mild stromal fibrosis (blue arrow) and increased vasculo-syncytial membrane (yellow arrow). Hematoxylin & Eosin x200

Fig. 5.

Shows syncytial knots (blue arrow). Hematoxylin & Eosin x200

Fig. 6.

Shows areas of intravillous calcifications (white arrow) within the placenta. Hematoxylin & Eosin x200

Fig. 7.

Shows infarcted chorionic villi (blue arrow) and few viable ones (white arrow). Hematoxylin & Eosin x200

Fig. 8.

Chorionic villus with thickened basement membrane (arrow). Periodic Acid Schiff stain x200

Table 3.

Histo-morphological placental features in diabetic and non-diabetic participants

Histo-morphological features	Cases=65n(%)	Control=65n(%)	c2/Fishers/T-test	P-value	Odds ratio(95%CI)
Weight(g)
450-600	8(12.3)	50(76.9)	54.91†	<0.0001**	23.75(9.23-60.70)
<450 and >600	57(87.7)	15(23.1)
Volume(ml)
200-950	46(70.8)	58(89.2)	6.92†	0.009**	3.42(1.33-8.84)
<200 and >950	19(29.2)	7(10.8)
Diameter(cm)
20-22	62(95.4)	65(100.0)	††	0.04**	______
<20 and >22	3(4.6)	0(0.0)
Central thickness(cm)
2.5	20(30.8)	42(64.6)	14.92†	<0.0001**	4.12(1.98-8.54)
<2.5 and >2.5	45(69.2)	23(35.4)
Shape
Abnormal	19(29.2)	21(32.3)	0.14†	0.70	0.87(0.41-1.82)
Normal	46(70.8)	44(67.7)
Level of villous maturity
Immature	4(6.2)	0(0.0)	††	0.02**	______
Mature	61(93.8)	65(100.0)
Villous edema
Present	32(49.2)	7(10.8)	22.89†	<0.0001**	8.04(3.19-20.22)
Absent	33(50.8)	58(89.2)
Syncytial knots
Present	11(16.9)	14(21.5)	0.45†	0.50	0.74(0.31-1.79)
Absent	54(83.1)	51(78.5)
Chorioangiosis
Present	24(36.9)	8(12.3)	10.61†	<0.001**	4.17(1.70-10.21)
Absent	41(63.1)	57(87.7)
Fibrinoid necrosis
Present	23(35.4)	4(6.2)	16.88†	<0.0001**	8.35(2.69-25.91)
Absent	42(64.6)	61(93.8)
Stromal fibrosis
Present	23(35.4)	6(9.2)	12.83†	<0.0001**	5.39(2.02-14.37)
Absent	42(64.6)	59(90.8)
Calcification
Present	14(21.5)	9(13.8)	1.32†	0.25	1.71(0.68-4.28)
Absent	51(78.5)	56(86.2)
Infarction
Present	2(3.1)	3(4.6)	††	0.65	_______
Absent	63(96.9)	62(95.4)
Membrane thickness
Increased	8(12.3)	7(10.8)	0.08†	0.78	1.16(0.40-3.42)
Not increased	57(87.7)	58(89.2)
Overall histomorphorlogical  placental changes
Yes	65(100.0)	38(58.5)	††	<0.001**	______
No	0(0.0)	27(41.5)

KEYS: ** = Statistically significant

Abnormal shape: irregular +oval placental shapes

Test statistic: c2 = †, Fishers = †† t-test = †††

Table 4 shows comparison of obstetric outcomes amongst diabetic and non-diabetic participants. Compared to non- diabetic participants, the diabetics had higher occurrence of all the adverse obstetric outcomes except for the gestational age at delivery and fifth minute Apgar score.

Table 4.

Comparison of obstetric outcomes amongst diabetic and non-diabetic participants

Obstetric outcomes	Casesn(%)	Controln(%)	c2/ fishers     /t-test	P-value	Odds ratio(95%CI)
Gestational age at delivery
Preterm delivery	22(33.8)	13(20.0)	3.17†	0.08	2.05(0.92-4.54)
Term delivery	43(66.2)	52(80.0)
Mode of delivery
Caesarean delivery	41(63.1)	18(27.7)	16.42†	<0.0001**	4.46(2.13-9.36)
Vaginal delivery	24(36.9)	47(72.3)
Mean birth weight ± SD	3.85±0.6	3.14±0.5	7.08†††	<0.0001**
Fetal Statusat birth
Dead	6(9.2)	0(0.0)	††	0.003**	_____
Alive	59(90.8)	65(100.0)
First minute Apgar
≤7	26(44.1)	12(18.5)	9.54†	0.002**	3.48(1.55-7.83)
>7	33(55.9)	53(81.5)
Modal APGAR score	8	8	28.10††	<0.0001**
Fifth minute Apgar
≤7	3(5.1)	8(12.3)	2.00†	0.16	0.38(0.96-1.51)
>7	56(94.9)	57(87.7)
Modal APGAR score	9	9	12.82††	0.50
Fetal blood sugar level at delivery
Abnormal	14(23.7)	1(1.5)	14.32†	<0.0001**	19.91(2.53-156.89)
Normal	45(76.3)	64(98.5)
Respiratory distress syndrome
Present	10(16.9)	3(4.6)	5.01†	0.03**	4.22(1.10-16.16)
Absent	49(83.1)	62(95.4)
SCBU admission
Admitted	59(90.8)	15(23.1)	60.73†	<0.0001**	32.78(11.83-90.81)
Not admitted	6(9.2)	50(76.9)

KEYS: ** = Statistically significant

Test statistic: c2 = †, Fishers = †† , t-test = †††

Table 5 shows the relationship between ultrasonographic and histomorphological features as well as the accuracy of ultrasound sound in determining these features using histomorphological features as the gold standard. The relationship was determined between variables common to both diagnostic methods. It shows that some relationship existed between all the ultrasonographic and histomorphological placental features except for the central thickness and the shape and that ultrasonography is accurate enough in determining these histomorphological changes.

Table 5.

Relationship between ultrasonographic and histomorphological features; and accuracy of ultrasonography in determining these features

Placental feature			Relationship		Accuracy of ultrasonography
	Ultrasound	Physical or Post delivery	Test statistic	P-value
Weight (g)	561.60±91.6	667.88±162.3	0.38 ††††	0.002**	93.85%
Volume(ml)	861.61±318.1	637.31±172.3	0.46 ††††	<0.001**	69.23%
Diameter(cm)	14.11±2.2	18.97±2.0	0.30 ††††	0.02**	100.00%
Central thickness(cm)	4.55±1.3	2.65±0.7	--0.14 ††††	0.26	67.69%
Shape
Normal	74(56.9)	116(89.2)	0.26 †††††	0.07	53.85%
Abnormal	56(43.1)	14(10.8)
Calcification
Absent	102(78.5)	107(82.3)	0.49 †††††	<0.001**	84.62%
Present	28(21.5)	23(17.7)

KEY: Statistically significant = ** Pearson’s correlation test (r) = ††††, Cramer’s V test (j_C) = ††††

Table 6 shows the histomorphological features of the placentas of diabetic women who had good glycaemic control compared to the histomorphological features of the diabetic women who had poor glycaemic control. Most of the histomorphological features were similar between the two groups, except the mean placental diameter and the presence of stromal fibrosis. The placentas of women with poor control had a significantly higher mean placental diameter (0.003) and a significantly higher proportion of them had stromal fibrosis (p = 0.022).

Table 6.

Comparison of histomorhological findings of placentas of women with good glycaemic control and poor glycaemic control

Histo-morphological features	Good control=20n(%)	Poor control=45n(%)	c2/Fishers/T-test	P-value
Mean placental weight ± SD (g)	710.00 ± 102.1	789.33 ± 174.4	-1.891†††	0.063
Mean placental volume ± SD (ml)	677.50 ± 103.2	764.44 ± 200.0	-1.832†††	0.072
Mean placental diameter ± SD (cm)	18.66 ± 1.2	20.03 ± 1.8	-3.084†††	0.003*
Mean placental central thickness ± SD (cm)	2.85 ± 0.8	2.95 ± 0.6	-0.586†††	0.560
Shape
Abnormal	4 (20.0)	15 (33.3)	1.190†	0.379
Normal	16 (80.0)	30 (66.7)
Level of villous maturity
Immature	2 (10.0)	2 (4.4)	††	0..763
Mature	10 (90.0)	43 (95.6)
Villous edema
Present	9 (45.0)	23 (51.1)	0.207†	0.789
Absent	11 (55.0)	22 (48.9)
Syncytial knots
Present	3 (15.0)	8 (17.8)	††	>0.999
Absent	17 (85.0)	37 (82.2)
Chorioangiosis
Present	8 (40.0)	16 (35.6)	0.117†	0.732
Absent	12 (60.0)	29 (64.4)
Fibrinoid necrosis
Present	7 (35.0)	16 (35.6)	0.002†	0.966
Absent	13 (65.0)	29 (64.4)
Stromal fibrosis
Present	3 (15.0)	20 (44.4)	5.250†	0.022*
Absent	17 (85.0)	25 (55.6)
Calcification
Present	3 (15.0)	11 (24.4)	††	0.521
Absent	17 (85.0)	34 (75.6)
Infarction
Present	0 (0.0)	2 (4.4)	††	>0.999
Absent	20 (100.0)	43 (95.6)
Membrane thickness
Increased	1 (5.0)	7 (15.6)	††	0.417
Not increased	19 (95.0)	38 (84.4)

Key * = significant, Test statistic: c2 = †, Fishers = †† , t-test = †††

Table 7 compared the outcomes of the women with good glycaemic control to the outcomes of the women with poor glycaemic control. Only the mean fetal weight was significantly different, as women who had poor glycaemic control had significantly bigger babies (p = 0.015).

Table 7.

Comparison of obstetric outcomes amongst diabetic women with good glycaemic control and diabetic women with poor glycaemic control

Obstetric outcomes	Good control = 20n(%)	Poor control = 45n(%)	c2/ fishers     /t-test	P-value
Gestational age at delivery
Preterm delivery	10 (50.0)	16 (35.6)	1.204†	0.411
Term delivery	10 (50.0)	29 (64.4)
Mode of delivery
Caesarean delivery	13 (65.0)	28 (62.2)	0.046†	>0.999
Vaginal delivery	7 (35.0)	17 (38.8)
Mean birth weight ± SD	3.58 ± 0.4	3.97 ± 0.6	-2.497†††	0.015**
Fetal Status at birth
Dead	0 (0.0)	6 (13.3)	††	0.166
Alive	20 (100.0)	39 (86.7)
First minute Apgar
≤7	10 (50.0)	16 (41.0)	0.432†	0.585
>7	10 (50.0)	23 (59.0)
Fifth minute Apgar
≤7	1 (5.0)	2 (5.1)	††	>0.999
>7	19 (95.0)	37 (94.9)
Respiratory distress syndrome
Present	3 (15.0)	7 (17.9)	††	>0.999
Absent	17 (85.0)	32 (82.1)
SCBU admission
Admitted	20 (100.0)	39 (86.7)	††	0.166
Not admitted	0 (0.0)	6 (13.3)

Test statistic: c2 = †, Fishers = †† , t-test = †††

KEYS: ** = Statistically significant

Discussions

The present study also observed a greater affectation of diabetes in the 30–39 years old, and grandmultipara. These findings are similar to those of John et al., [24] which may be explained by an increased risk of medical conditions with increasing maternal age and by extension increased parity [1]. Increased risk of miscarriages amongst the diabetic participants may also be related to the effect of chronic hyperglycaemia [1].

The study also noted that ultrasonographic findings correlated significantly with the histo-morphological placental changes except in the measurement of the central thickness of the placenta.

Findings from this study show statistically significant difference in the macroscopic appearance of the placentae from the diabetic group, with the diabetic placentae being 24 times more likely to be heavier, four times more likely to be thicker at the centre, three times more likely to be voluminous and of larger diameter compared to non-diabetic participants. Similar findings have been reported in Europe [6] and India [15,25]. Additionally, the diabetic pregnant women were also observed to exhibit significantly larger ultrasonographically detected placental weight and volume compared to non-diabetics just as described in other studies [6,18]. These findings may be attributable to the compensatory effect of chronic hypoxia and resultant hyperplacentosis common amongst diabetics [9–11].

Microscopic placental changes also occurred significantly more in diabetics in the index study. These included fibrinoid necrosis and villous edema, each likely to be eight times more common, stromal fibrosis and chorangiosis which were likely to be five and four times more common respectively than among non-diabetics. Villous immaturity was predominantly observed amongst the diabetic compared to non-diabetic pregnant women and this was statistically significant. Findings of more microscopic placental changes in the diabetic compared to the non-diabetic pregnant women in this study are in agreement with those from other studies in the literature [25–30]. These microscopic placental changes are thought to result from hypoxia and consequent oxidative stress; both are products of villous dysfunction [10–12].

Although not statistically significant, presence of syncytial knots and placental infarction were observed to be less common amongst diabetic pregnant women on histo-morphology. Similarly, ultrasonographically detected placental calcification was observed to be significantly less frequent in diabetic pregnant women compared to their non-diabetic counterparts. This may be due to delayed maturation of the villous system as observed in this study, making the placenta of the diabetic group less prone to ageing process, as these lesions are manifestations of placental aging [11,30]. These are in contrast with other studies, [26,31] possibly because of inappropriate specimen collection and processing in their study[26]. This may have caused tissue desiccation and may have introduced artefacts, the resultant of which may appear as abnormal placental changes. Also, non- exclusion of patients with hypertensive disorders and anaemia in pregnancy [31] may have also contributed to their findings as both hypertensive disorders and anaemia in pregnancy are known causes of placental aging known to be associated with these placental changes.

In keeping with reports from other studies, [32, 33] the index study also showed that diabetic participants were 11 times more likely to be significantly associated with abnormal umbilical cord coiling index in favour of hyper coiling compared to the non-diabetic counterparts. This may be closely related to the findings that up to 50% of the diabetic participants in this study presented with poor glycaemic control. Poorly controlled diabetes mellitus has been reported to predispose to polyhydramnios with resultant ease of fetal rotatory movement and thus greater risk of coiling of the cord [34].

The vascularization, flow and vascularization-flow indices were observed to be significantly higher in the diabetic compared to the non-diabetic. These may be due to increased angiogenesis and neovascularization seen in the placenta of diabetic patients resulting in vessels with high impedance and increase resistance [10,11]. These findings are similar to findings by Jako et al. [19] but in contrast to findings by Moran et al. [18]. The disparity with Moran et al.¹⁸ may have resulted from their small sample size and possible inclusion of diabetic patients in the control arm of their study since the normoglycaemic status of their control were not proven biochemically.

In this study, the diabetic group were observed to be twice more commonly associated with increased preterm delivery though not statistically significant. They however, significantly had more Caesarean birth, fetal macrosomia, hypoglycaemia, respiratory distress syndrome (RDS) and three times more likely to be significantly associated with poor first minute Apgar score compared to the non-diabetic group. Also, still birth was found more commonly amongst the diabetics compared to non- diabetics and it was statistically significant. These findings are in agreement with those from other studies [2,4,26,27,35]. The fetal macrosomia, hypoglycaemia at birth, and respiratory distress syndrome observed amongst the diabetics may be related to the high rate of poor glycaemic control observed in this study. The consequent chronic hyperglycaemia may lead to excessive fetal growth, delay in the production of type 2 pneumocytes which are involved in surfactants production. The withdrawal of the neonate from the hyperglycaemic environment at delivery may predispose to neonatal hypoglycaemia [1]. Chronic hypoxia resulting from chronic hyperglycaemia and insulinaemia may lead to poor first minute Apgar score as well as still birth [1]. Fetuses suffering from intra-uterine hypoxia may therefore be delivered preterm and via Caesarean sections to avert further adverse events. Increased fetal macrosomia as observed in this study might also explain the increased risk of Caesarean birth amongst diabetic participants.

Non- diabetics however, were observed in this study to have more poor fifth minute Apgar score compared to the diabetics. This was not statistically significant, however it is not surprising as fifth minute Apgar score is a function of effectiveness of resuscitation and babies of diabetic mothers are more likely to adapt to effect of hypoxia than the babies of non-diabetic mothers [36]. The finding of more poor fifth minute Apgar score in babies of non-diabetics in the present study is in contrast to those from other studies [2,4]. The retrospective nature of these studies may explain the disparity.

This study noted moderate but significant positive relationship between the ultrasound measurements and histo-morphological assessments of the placental volume and calcifications. The relationship between ultrasound assessment of placental weight and diameter and their histo-morphological measurements were also significant, though they were low positive relationships. These significant positive relationships may be related to the effectiveness of 2- dimensional ultrasound scan in obtaining these measurements and detecting the changes [6]. However, there was no significant association observed between ultrasound measurement of placental shape and central thickness with their corresponding histo-morphological measurements possibly because placental shape is a 3- dimensional feature which may be better measured with 3-dimensional ultrasound scan, which was not used in this study. Furthermore, there was inverse relationship between ultrasound measurement of placental central thickness and the actual placental central thickness measured histo-morphologically. This finding may be related to the patients’ myometrial echoes which might have made it possible to incorporate other structures within the myometrium like uterine fibroid into the measurement thereby increasing the central thickness ultrasonographically and causing inverse relationship as observed in this study. None of these findings were documented in all the literatures reviewed.

Furthermore, this study observed ultrasound scan to be accurate in the detection of abnormalities of diameter, weight and calcifications while it was less accurate in detection of placental abnormalities of volume, central thickness and shape. This may also be related on the resolution of the 2- dimensional ultrasound scan used in this study in detecting these placental features.

The study found a significant increase in the mean placental diameter and stromal fibrosis among women with poor glycaemic control when compared to women with good glycaemic control. This is similar to the findings of another study that also reported increase in placental diameter and increase in fibrosis among women with poor control [37]. In that study, placental weight, volume, villous abnormalities were also more among the women with poor control, in contrast to this study. These differences might be due because only women with gestational diabetes were included in that study and the sample size was much larger, whereas women with preexisting diabetes and gestational diabetes were included in this study and the sample size was smaller.

When the outcomes were compared between women with good glycaemic control and those with poor glycaemic control, women with poor glycaemic control had significantly bigger babies, while all other outcomes measures assessed were similar. Very few differences were observed between diabetics with good glycaemic control and those with poor control in terms of outcomes and histomorphological changes, while many differences were observed between the diabetic women and the non-diabetic women. Therefore, it is likely that the findings resulted from the diabetes itself rather than the degree of glycaemic control.

Conclusion

Ultrasonographic and histo-morphological placental changes and adverse obstetric outcomes occurred more significantly in the diabetic group. Ultrasonographic placental findings correlated with the histo-morphological changes. Ultrasonographic placental examination can therefore, be incorporated as part of routine care in the management of diabetes in pregnancy in order to prevent adverse maternal and perinatal outcomes observed to be more common in these individuals.

Since these changes were noted either in late pregnancy or postpartum, there is need for further studies on the ultrasonographic changes in earlier gestation so as to allow for measures to improve upon maternal and perinatal outcomes.

Clinical relevance

Given that ultrasound scan can accurately detect placental abnormalities associated with diabetes in pregnancy, it can be used to guide patient counselling and clinical decision making in the management of women with diabetes in pregnancy, allowing timely interventions to prevent adverse outcomes.

Limitations

1. A single FBS or RBS was used to exclude diabetes mellitus in the controls, and OGTT was not done.

2. Glycaemic control was based on a single HbA1c test result, and this may not be reflective of the true glycaemic control throughout the pregnancy.

Recommendation

Further studies can be conducted focusing on the effect of glycaemic control on placental changes and outcomes, after classifying women into women with diabetes in pregnancy and those with GDM, as these two conditions have different pathophysiologies.

Abbreviations

AKTH

Aminu Kano Teaching Hospital

ANC

Antenatal care

CI

Confidence interval

DM

Diabetes mellitus

FBS

Fasting blood sugar

GDM

Gestational diabetes mellitus

HbA1c

Glycosylated haemoglobin

LNMP

Last normal menstrual period

MMSH

Murtala Muhammed Specialist Hospital

OR

Odds ratio

RDS

Respiratory distress syndrome

SCBU

Special care baby unit

SD

Standard deviation

USS

Ultrasound scan

Authors’ contributions

IJO, TIU, and AA conceptualized the study and IJO wrote the initial proposal for the study and obtained the ethical clearance. IJO and TIU recruited patients for the study. AA and IJO analysed all the placental specimen and produced the histology results, AA and IJO performed the ultrasound scans for the sonographic findings in all the patients. IJO, TIU, AA, LFCN and IOO performed al the statistical analysis of the data and produced the final results. IJO and TIU wrote the discussions. All authors read and approved the final manuscript.

Funding

The work was funded by the authors, there was no external funding.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the Declaration of Helsinki. Approval to conduct the study was obtained from the Health Research Ethics Committee of Aminu Kano Teaching Hospital (NHREC/21/08/2008/AKTH) and Murtala Muhammed Specialist Hospital (MOH/Off/797/T.I/693). A written informed consent was obtained from all participants after the study had been thoroughly explained to them.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.