India specific intrauterine growth curves were constructed by many authors [2-5]. Changes in the parity, socioeconomic and environmental conditions necessitate an update in the existing growth charts. We planned the present study to derive a gestation specific reference growth chart for our local population. The objective was to construct centile charts for birthweight, length and head circumference for infants born from 24 to 42 weeks of gestation.

This was a medical record review of the period from July 1999 to October 2009 at a maternity and newborn level III care hospital in South India. All consecutively live born singleton infants from 24 to 42 weeks of gestational age during the study period were eligible. The gestational age had been calculated in completed weeks from the findings of the early dating ultrasound or from the last menstrual period. Infants with major congenital anomalies and those with uncertain gestational age were excluded. The weights had been measured on a digital weighing machine with a variability of ± 10 grams. The head circumference had been measured using a non- stretchable tape and recorded in centimetres, and the length of the infant with an infantometer in centimetres.

All the relevant risk factors [hypertensive diseases of pregnancy, maternal medical disorders, severe anemia (Hb <7g/dL) in the mother] were noted from the maternal and neonatal database and in select cases, from the case records. The perinatal details were obtained from the neonatal database.

Six sets of charts were constructed defined by the sex of the infant for the birthweight, length and head circumference. The normal distribution of the Z scores resulting from the fitting of the LMS models within each group was verified by obtaining normal probability plots (Q-Q plots). Centiles were calculated using the LMS Chart Maker software and the other analyses was carried out using SPSS, version 16.

Thirty three thousand and seven hundred and forty infants were born alive in the hospital during the study period. 787 infants were products of multiple pregnancies and in 28 infants gestation was either less than 24 weeks or more than 42 weeks. From the eligible 32,925 infants, gestation was not certain in 1444 infants. Among the remaining 31,481, the data was available for 31,391 (males, 16,054), 28,812, (males, 14,730), and 28,790 (males, 14,724) infants for the plotting of weight, length and head circumference centiles, respectively. After removing the outliers, the data was available for 30,016, 27,228 and 26,974 infants for the final plotting of weight, length and head circumference centiles, respectively (Table I). Sixty percent of the pregnancies were booked in the hospital before 20 weeks of pregnancy and 40% of the infants were born to primiparous mothers. The number of infants with birthweight less than 2500 grams were 5,921 (19%) and those with birthweight less than 1500 g and 1000 g were 1,023 (3.3%) and 282 (1%) infants, respectively. The incidence of PIH was 6.5% (n=2060) and that of diabetes and severe anemia was 5.5% (n=1736) and 1% (n=313), respectively among the mothers of infants included in the study.

TABLE I	Infants at Each Gestational Age After Removing the Outliers

The Q-Q plot for z scores of birth weight, length and head circumference displayed good correlation for the data between 5th and 95th centiles. The crude curve for the raw data of male infants displayed a bump in the 97th percentile from 30 to 32 weeks. The crude curves of female infants were relatively smoother. Female infants were lighter than the male infants especially from 35 weeks onwards (Fig. 1). However, there was not much difference in the length and head circumference (Fig. 2,3).

(a)

(b)

Fig. 1 Smoothened centile curves for the birthweight of (a) male and (b) female infants (without outliers).

(a)

(b)

Fig. 2 Smoothened curves for length of (a) male and (b) female infants (outliers removed).

(a)

(b)

Fig. 3 Smoothened curves for head circumference of (a) male and (b) female infants (outliers removed).

The study met most standards required to obtain an ideal reference growth chart [11]. The data included singleton babies born from a monoethnic group and most importantly the best method for gestational age estimation (first trimester based ultrasound dating) was used in more than 60% of the included infants as they were born of pregnancies registered early in our hospital. Gestational assessment was mandatory for our hospital records. The gold standard LMS method [6] was used for estimating the centiles and for the smoothening of the centile curves.

Most of the recent studies have shown a secular trend of increasing birthweight at higher gestational ages. This is also evident in our cohort. The mean birthweights after 34 weeks were higher in our cohort compared to the previous studies [2-5]. At term gestational age the mean birthweights were almost 100 to 200 grams higher in our present cohort. However, at lower gestations, our birth weights were similar or lower (Web Table I). This may be attributed to improved survival of smaller babies at lower gestations over the last two to three decades. Similar trends were observed for the length at birth. There was a trend of increased length at higher gestations. However, there was no change in head circumference centiles across all gestations.

Inspite of the best effort in calculating the centiles using the large cohort and also best statistical methods, this study still has few limitations. Gestational age estimation was not ultrasound based in all the included infants. Anthropometric measurements were done by different nurses and this may have contributed to the inter-observer variability. We did not construct separate curves for parity and maternal height. This is a single centre cross-sectional study spanning 10 years. This growth data is from all the socioeconomic strata and hence might not reflect the ideal potential for intrauterine growth. An ideal chart should be multicentred and involve data from 2 to 5 years. Ideally, at each gestation, we would require 100 babies for estimating all the centiles. Our data lacks this parameter at lower gestations (less than 32 weeks). We did not include stillbirths in our data for the ease of data collection, as it was retrospective.

Contributors: SM and GK designed the study. HK collected the data. SM and SS monitored the data collection. HK and SM analyzed data and wrote the manuscript with inputs from SS.

Funding: None; Competing interests: None stated.

1. Lubchenco L, Hansman C, Dressler M, Boyd E. Intrauterine growth as estimated from live born birth weight data at 24 to 42 weeks of gestation. Pediatrics. 1963;32:793-800.

2. Ghosh S, Bhargava SK, Madhavan S, Taskar AD, Bhargava V, Nigam SK. Intrauterine growth of north Indian babies. Pediatrics. 1971;47:826-32.

3. Singh M, Giri SK, Ramachandran K. Intrauterine growth curves of live born single babies. Indian Pediatr.1974;11:475-9.

4. Mohan M, Shiv Prasad SR, Chellani HK, Kapani V. Intrauterine growth curves in north Indian babies: weight, length, head circumference and ponderal index. Indian Pediatr.1990;27:43-51.

5. Mathai M, Jacob S, Karthikeyan NG. Birthweight standards for south Indian babies. Indian Pediatr.1996;33:203-9.

6. Cole TJ. Fitting smoothed centile curves to reference data. J R Stat Soc. 1988;151:385-418.

7. Kramer MS, Platt RW, Wen SW, Joseph KS. A new and improved population-based Canadian reference for birthweight for gestational age. Pediatrics. 2001;108;e35.

8. Bonellie S, Chalmers J, Gray R, Greer I, Jarvis S, Williams C. Centile charts for birthweight for gestational age for Scottish singleton births. BMC Pregnancy and Childbirth. 2008;8:5.

9. Olsen IE, Groveman SA, Lawson ML, Clark RH, Zemel BS. New intrauterine growth curves based on United States data. Pediatrics. 2010;125;e214-24.

10. National Neonatology Forum. National Neonatal-Perinatal Database (2002-03). Available from www.newbornwhocc. org/pdf/nnpd_report_2002-03.PDF. Accessed on 1 March, 2011.