Indian Pediatrics 1998; 35:733-743

 

D.K. Agarwal, K.N. Agarwal, K. Satya and S. Agarwal

From the Maternal Child Health Unit, Department of Pediatrics, Institute of Medical Sciences, Banaras Hindu University, Varanasi 221 005, India.

Reprint requests: K.N. Agarwal, Professor of Pediatrics, University College of Medical Sciences,
Delhi-110 095, India.

Manuscript received: February 25, 1997; Initial review completed: April 11, 1997;
Revision accepted: February 5,1998

 

Objective: To identify risk factors for high perinatal (PMR) and infant (IMR) mortality in a rural area. Design: In 49 randomly selected villages from two adjoining blocks of rural Varanasi, all pregnant women and live births were followed for perinatal and infant mortality, during the years 1988-1992. Subjects: 6790 births and their 6649 live births. Results: The PMR was 90.7 per thousand births and IMR was 98.6/1000 live births. These mortalities were significantly higher if weight gain during pregnancy was less than 7.0 kg. Low weight gain during pregnancy was also associated with significantly higher low birth weight deliveries and to some extent increased still birth rate. PMR and IMR decreased with higher levels of hemoglobin in third trimester and socioeconomic index; however, the calculated RR were not significant. Conclusion: Low weight gain during pregnancy is an important risk factor for PMR and IMR.

Key words: Perinatal mortality, Pregnancy weight gain, Infant mortality.

PERINATAL and infant mortality rates (PMR and IMR) are very sensitive indicators of maternal and child health services in the population. Maternal age at conception, birth order, birth interval, type of birth attendant, maternal nutrition (short stature, poor pre-pregnancy weight, inadequate weight gain during pregnancy and anemia), and reproductive tract infections significantly influence the pregnancy outcome. In addition education, occupation, income, socia-economic status, and available health facilities particularly perinatal care are also important factors related to maternal and infant health(1-4).

In 1987-88 in Uttar Pradesh (UP), the rural IMR was 133 per 1000 live births as compared to national figure of 101. The corresponding figures for rural IMR in 1994 were 88 and 79(5). The rural perinatal mortality rates in 1994 were 43.4 and 41.8 per 1000 births, for the country and UP, respectively(5). The reduction in PMR and IMR remain strong determinants of accepting fertility control measures(6). In the present study pregnancies with 6790 births were observed to collect perinatal and infant mortality data, with the objective to identify the likely maternal nutritional and sociodemographic factors responsible for higher mortality rates. An attempt was also made to study the differences for the study parameters in the ICDS and non-ICDS blocks after regulating the visits of the workers.

Subjects and Methods

The study was carried out in randomly selected 28 villages of Harahua (ICDS population 33770) and 21 villages of Kashi Vidyapeeth (non-ICDS population 32307), the adjoining blocks of district Varanasi during 1988-92. The sociodemographic and biological characteristics are summarized in Table I. In order to assess the overall socio-economic status of the families of the study subjects, a linear composite score was obtained by combining the rank order scores of 6 variables, namely, parents education and occupation, caste and per capital income. The linear sum of socio-economic index scores ranged from 6-33(7). All the information, follow-up measurements for weight and mid arm at 16±2, 28± 2 and 36±2 wk of gestation were taken by the nutritionist. These workers have been undertaking anthropometry for growth studies for over five years(8). The hemoglobin was estimated according to method of Crosby et al.(9) in 1954 women who allowed estimation in the third trimester. The number of pregnant women has varied for measurements, particularly weight gain in pregnancy. These women delivered at home, some needing assistance were attended to in the University Hospital of the Institute Medical Sciences, Varanasi.
 

TABLE I

Sociodcmographic and Biological Charactcristics of Study Women.

Characteristics	Harahua-ICDS	Kashi Vidyapeeth Non-ICDS
Total villages	172	121
Selected for stud	28	21
Population	33770	32307
Households	4494	4582
Target women	4694	4326
Excluding family planning acceptors and menopausal
Birth rate*	34.7	37.1
Fertility rate*	179.9	186.5
Death rate*	11.4	12.1
Maternal socio-demographic features'(Mean±SE)
Age (yr)	26.0 ± 0.1	25.2 ± 0.1
Parity	3.6 ± 0.04	3.5 ± 0.03
Gravida	3.6 ± 0.04	2.9 ± 0.03
Inter pregnancy interval (mo)	21.2±0.29	23.0±0.19
Socioeconomic index	1O.7±0.04	1O.9±0.04
Percapita income (Rs.)	178±1.3	159±1.2
Illiterate (%)	88	84
Occupation (House wife %)	99	99
Castes Scheduled (%)	23.4	18.4
Backward (%)	58.2	67.2
Upper (%)	12.4	12.3
Others (%)	6.0	2.0

* Per 1000 population as per survey done for these blocks by the Department of Preventive and Social Medicine, Institute of Medical Sciences in 1986 (CBD-Report).


Method of Registration and Data Collection

The target women were contacted by the village female worker (non-ICDS) and anganwadi worker (ICDS). These female workers had education upto 8-10th class and were trained for the project work. The number of visits were similar in both the blocks, thus ICDS had extra input of irregular and insufficient food supplement to 33.6% pregnant women, only. They did regular LMP monitoring and on each contact weighed the woman (provided the pregnancy weight). The height and mid-arm circumference, were also measured in the first contact.

Women who missed 2-3 consecutive menstrual periods were registered. The information of these study women was collected by the team of nutritionist and social scientists having postgraduate qualifications. The female village level worker assisted. The information was collected in precoded and pre-tested proforma for village house hold, pregnant women's house hold, obstetric and present pregnancy history, physical examination, pregnancy outcome and infant death record.
 

Statistical Methods

The followings indices were calculated:

                                     Number of still births and infants death of less than 7 days during the year
Perinatal mortality rate = ------------------------------------------------------------------------------------------------------------ x 1000
                                                  Number of live births and still births during the year

                                      Number of infant deaths during the year
Infant mortality rate =  ----------------------------------------------------------------------- X 1000
                                        Number of live births during the year


                           Number of still births during the year
Still birth rate =   -------------------------------------------------------------------------  = x 1000
                            Number of live births and still births during the year

The values of Chi square (X²), "relative risk (RR) and 95% confidence interval and Z proportions were calculated. The 't' test was used for group comparison.


The female village workers and anganwadi workers were encouraged to record birth weight within 24 hours and positively within 48 hours after birth. They were provided scales for weighing the mothers on each visit and the baby at birth. All births were checked by the senior team member every week, who supervised five villages. The team had transport facilities. Women were weighed in standing position with normal clothing on Chattilon platform weighing scale (M/s John Chattilon and Sons, USA), accuracy upto 20g. Neonates were weighed on a modified Tansi scale with accuracy of 109. The women were measured for height by using a metal calibratedrod and for mid-arm circumference by fibre glass tape to the nearest of 0.1 cm, for both(8).

Results

The sociodemographic and biological characteristics are presented in Table 1. These are similar in the ICDS and non- ICDS blocks. As the PMR and IMR for the ICDS and non-ICDS blocks were similar (Table II), the pattern of change with maternal indices Was analyzed for all the villages together (Table III). However, the RR and confidence intervals are calculated separately for the two blocks (Table IV & V).

In Table II, data for Harahua and Kashi Vidhyapeeth blocks for births in 6790 women showed that early, late and po~t neonatal deaths and infant mortality data are not different; however, the still birth rate was significantly lower (p < 0.01) in the ICDS areas as compared to the non-ICDS.

 

In Table III, PMR and IMR in relation to prepregnancy weight, height, mid-arm circumference, weight/height index and body mass index groups, hemoglobin in third trimester, weight gain during pregnancy, per capita income' and socio-economic index are presented. The PMR did not show any significant relationship with maternal anthropometric parameters. However, there was a significant reduction with higher weight gain during pregnancy, higher hemoglobin level in 3rd trimester and better socio-economic index. A higher maternal pre-pregnancy height, body mass index, hemoglobin (third trimester), weight gain in pregnancy and socio-economic index were significantly associated with the IMR.

In Tables IV and V, RR was significantly higher if weight gain during pregnancy was < 7.0 kg, for PMR as well as IMR, for both the blocks. For PMR, pre-pregnancy height in the ICDS and child's sex in the non-ICDS areas were also significant risk factors. In Table VI, still birth rate and low birth weight (LBW) prevalence decreased with increase in maternal weight gain during pregnancy. However, the risk of LBW deliveries was only significantly reduced in women gaining more weight during pregnancy. The regression equation was LBW = 8.138 x weight gain + 75.61 (p < 0.022); showing that with every kg increase in weight gain during pregnancy there was 8% reduction of low birth weight deliveries.

The correlation coefficients for birth weight with weight gain during pregnancy were 0.366 and 0.289, for the ICDS and non-ICDS blocks; (p < 0.01) for both. The multiple regression analysis performed on the data of 2450 women in ICDS area showed that weight gain during pregnancy was significantly influenced by early pregnancy maternal height, weight and mid-arm circumference. Abdominal girth, blood pressure and hemoglobin in the first trimester studied in 1503 women did not influence the weight gain in pregnancy (Table VII).

 

TABLE III

PMR and IMR in Relation to Pre-pregnancy Nutritional Status, Weight Gain During
Pregnancy, Hemoglobin in the Third Trimester and Socio-economic Maternal Factors.

1. Maternal nutritional status (Pre-pregnancy)
(a) Weight (kg)	<35.0	35.0-39.0	40-44.9		45-49.9	≥50
PMR (X2 = 3.6, d.f. = 4, P > 0.05)	82.9	78.1	72.9		63.8	52.5
IMR (X2 = 8.2, d.f. = 4, P > 0.05)	122.2	92.3	100.1		88.8	65.4
n	201	1421	2856		1243	279
(b) Height (cm)	< 140	140-144.9	145.0-149.9		≥150
PMR (X2 = 0.859, d.f. = 3, P > 0.05	204.5	75.8	74.1		68.3
IMR (X2 = 9.206, d.f. = 3, P < 0.05)	157.1	105.9	91.0		88.0
n	155	675	2480		2728
(c) Mid-arm circumference (cm)	<20.0	20.0-20.9	21.0-21.9		22.0-22.9	≥23.0
PMR (X2 = 6.605, d.f. == 4, P > 0.05)	77.8	68.8	80.1		68.4	97.8
IMR (X2 = 1.511, d.f. = 4, P > 0.05)	101.2	109.9	101.7		94.9	97.4
n	257	931	1300		1492	1914
(d) Weight/Height index (%)	<90	90-110	> 110
PMR (X2 = 3.233, d.f. = 2, P > 0.05)	76.1	68.3	31.6
IMR (X2= 2.872, d.f. = 2, P > 0.05)	107.9	94.3	96.8
n	3182	2722	96
(e) Body mass index (kg/m2)	<16.0	16.1-17.0	17.1-18.5	18.6-20.0	20.1-25.0	>25.0
PMR (X2 = 4.824, d.f. = 5, P > 0.05)	89.9	63.7	72.1	76.1	65.6
IMR (X2 = 123.298, d.f. = 5, P < 0.001)	123.7	85.8			89.8	65.6
n	202	484	1654	2052	1584	33
(2) Hemolgobin (g/dl) (Third trimester)	<8.0	81.-9.0	9.1-10.0		10.1-11.0	>11.0
PMR (X2 = 11.419, d.f. = 4, P < 0.05)	57.4	38.7	46.8		39.4	10.0
IMR (X2 = 11.824, d.f. = 4, P < 0.05)	164.2	85.7	87.3		77.3	81.6
n	166	616	679		358	135
(3) Weight gain during pregnancy (kg)	<5.0	5.0-6.0	6.1-7.0		7.1-8.0	≥8.1
PMR (X2 = 60.063, d.f. = 5, P < 0.01)	123.2	59.9	35.9		24.9	23.0
IMR (X2 = 63.244, d.f. = 5, P < 0.01)	165.9	118.0	67.9		47.2	47.6
n	276	1339	1818		838	3.4
(4) Per capita income (Rs.)	≤ 100	101-129		130-300		>300
PMR (X2 = 7.34, d.f. = 3, p> 0.05)	93.2	65.4		65.1		94.9
IMR (X2 = 7.34, d.f. = 3, P > 0.05)	119.5	102.7		92.3		92.4
n	1027	1157		3755		288
(5) Socioeconomic Index	< 8	9-12		13-16		17-20
PMR (X2 = 8.953, d.f. = 3, P > 0.05)	86.2	71.5		55.2		56.3
IMR (X2 = 11.339, d.f. = 3, P > 0.01)	118.8	98.8		80.9		79.3
n   1161	1161	3745		1163		206

TABLE IV

Relative Risk (RR), Chi Square (c²) Confidence Interval (CI) and Z Values for Perinatal Mortality Rate
 in the ICDS and Non-ICDS Areas of Varanasi, 1988-92

 

Discussion

In the present study, poor weight gain during pregnancy « 7.0 kg) was an important risk factor for higher PMR and IMR. Weight gain of 7.0-8.0 kg (25% of the study women) in pregnant women with limited health facilities, heavy household work load, illiteracy, low income could achieve acceptable levels of PMR (24.9), IMR (47.3), SBR (2.7) and LBW (7.4%).

During 1981-84, Tripathi et al. (10) in the non-ICDS area of the present study, demonstrated an average weight gain during pregnancy of 7.1 and 7.4 kg at gestational ages 39 and 40 weeks, respectively. Further if pre-pregnancy weight was < 40 kg and weight gain during pregnancy < 5 kg; 2/3 of the live births were low birth weight. The present study showed that pre-pregnancy maternal nutrition influenced weight gain in pregnancy. The higher weight gain significantly reduced the LBW deliveries. Whether reduction in LBW has influenced the PMR and IMR is difficult to say. The Indian Council of Medical Research National Collaborative Study recorded neonatal mortality of 58.4 and 54.0 for rural and urban slum areas, the most important underlying and contributing cause being a high prevalence of LBW deliveries(11). In black women with pre-pregnancy weight below 46.0 kg who had a weight gain in pregnancy between 6.5 to 8.0 kg, the percentage of LBW infants born was . 15.6%(12). These findings support the present study having a LBW percentage of 18.8 and 7.4 in pregnancy weight gain groups of 6.1-7.0 kg and 7.1-8.0 kg, respec- tively. The observations are also supported by others(13,14) showing a higher preva- lence of LBW infants upto 50% as well as with a perinatal mortality rate as high as 155 per 1000 births in women with low prepregnancy weight and inadequate weight gain in pregnancy. Further, Edwards et al.(15) observed that even if underweight women have an adequate total weight gain as well as adequate rates of gain by trimester, their incidence of LBW is still two fold over women with normal pre- pregnancy weight.

 

TABLE VI

Still Birth Rate and Low Birth Weight Prevalence (%) in Relation to Weight Gain in Pregnancy.

	Weight gain (kg)
	n	<5.0	5.0-6.0	6.1-7.0	7.1-8.0	≥ 8.1
Still birth rate/1000 birth	141	14.5	6.3	6.5	2.7	4.6
* % low birth weight (< 2500 g)	806	48.2	29.4	18.8	7.4	9.8

(* Regression equation LBW = -8.138 x weight gain + 75.617; p < 0.022, significant)

 

TABLE VII

Effect of Various Factors on Weight of Women in the Third Trimester in ICDS Block

Variables	Mean ± SE	Co-efficient	Standard error	I-value
Constant		-0.530	1.232	0.430
Supplemented	0.413±0.013	0.016	0.050	0.313
Measurements in the first trimeste.r
Gestation (weeks)	16.074±0.019	-0.038	0.033	1.153
Height (cm)	149.508±0.106	0.035	0.007	5.362***
Weight (kg)	43.078±0.105	0.973	0.008	120.682***
Mid arm
circumference (cm)	22.194±0.041	0.069	0.019	3.632***
Abdomen girth (cm)	68.465 ±o.075	0.017	0.009	1.924
Systolic BP (mm)	104.109±0.261	-0.0004	0.003	0.144
Diastolic BP (mm)	67.942±0.184	0.001	0.004	0.119
Hemoglobin (g/ dl)	9. 996±0 .026	0.02'1	0.025	1.158

*p < 0.005; ***p < 0.001. **p < 0.01;

The mortality rates of rural India and Uttar Pradesh(5) are compared with the present study in Table VIII. In the present study, neonatal mortality, particularly early (0-7 days) is high (51/1000 live births). The higher SBR than the national and UP figures may have been due to proper recording. This is evident from the contrast of SBR of 9.0 in Kerala and 2.9 in Bihar(5). PMR is influenced by weight gain during pregnancy and maternal nutrition (pre-pregnancy). Amongst all deaths during infancy in the present study, 70% are neonatal deaths. The increased weight gain during pregnancy in the present study possibly resulted in a significant reduction in PMR and IMR by reducing the still births and the LBW deliveries. The mortalities were similar in both the blocks inspite of the fact that the ICDS area has shown a reduction in the still birth rate as well as low birth weight deliveries.

 

TABLE VIII

Comparison of Mortality Rates

Mortality rate	India	Uttar Pradesh	Present study
Neonatal mortality rate per 1000 live births	52.0	56.4	69.8
Post neonatal mortality rate	27.5	35.0	28.7
IMR	80	91	98.5
PMR	43.4	41.8	70-71
Still birth rate (per 1000 births)	7.3	3.8	20.8

It is concluded that better weight gain during pregnancy is associated with a reduced PMR and IMR and lower prevalence of LBW deliveries.

Acknowledgement

This study was partly financed by the Indian Council Medical Research, Ansari Nagar and US-AID, New Delhi. The Banaras Hindu University, Varanasi provided the infrastructural support.