Foot Length for Gestational Age Assessment and Identification of High-Risk Infants: A Hospital-Based Cross-Sectional Study

Abstract

The knowledge of the gestational age of the newborn is essential for management. In the absence of a dating scan, the postnatal assessment scores have drawbacks of being difficult to learn and administer in the community. The measurement of the foot length is easy, reproducible and offers an objective assessment. The objective of this study was to determine the correlation of postnatal (<48 h) foot length measurement (with calipers) with gestational age as determined by antenatal dating ultrasound, create a predictive model for the same and propose foot length measurement cutoffs for <37 and <34 weeks of gestation. Secondary objectives were to assess the correlation between foot length as measured with calipers and that measured with a ruler and a paper footprint. This was a hospital-based cross-sectional study. Among the 520 babies assessed, the correlation of foot length with gestational age was 0.89. Operational cutoffs for the categories of <37 and <34 weeks at a sensitivity of 95% were <70 and <65 mm, respectively. The Pearson’s correlation between foot length as measured by caliper and ruler was 0.95 and between caliper and paper footprint was 0.87. This study correlating foot length and gestational age has the potential to help neonatal care providers make informed management decisions, particularly in resource-limited settings.

foot length in neonates, postnatal assessment of gestational age, assessment of gestational age in the community

INTRODUCTION

The knowledge of the gestational age of the newborn is the bedrock of neonatal management. The reference standard for assessment of gestational age is the first trimester dating ultrasound; however, in many parts of the world this is not available. While mothers’ last menstrual period has traditionally been the benchmark for gestational age assessment, this can vary with irregular menstrual cycles and in some cases, may not be known.

Postnatal gestational assessment scores in the newborn are often lengthy and require technical training, which are drawbacks when trying to use these in the community. Foot length measurement is easy to learn, reproducible and non-invasive, can be done irrespective of the clinical status of the baby, with no technical training required in the measurement. Studies have been done across the world and in different communities that have looked at using the foot length measurement to evaluate and identify neonates at risk. These however have largely focused on birth weight or birth weight as a surrogate for gestational age. The correlation between gestational age and foot length to establish operational cutoffs to identify high-risk neonates based on gestational age have been less well described [1] (Table 1). However, in most of these studies, the standard for assessment of gestational age used is postnatal assessment scores and not the gold standard of the dating ultrasound. The technique for the measurement of foot length is also variable, with various studies using rulers, measuring tape and paper footprint. The measurements have also been taken by a spectrum of assessors ranging from community health workers to medical professionals. Consistency in measuring techniques and a method that is accurate, consistent and easily reproducible is essential.

Table 1

Operational cutoffs for the identification of preterm babies from foot length measurement

Study	Standard used for assessment of gestational age	Cutoffs proposed (cm)	Sensitivity (%)	Specificity (%)	PPV (%)	Foot length measured using
Nabiwemba et al. [13]	Eregie score	7.6	96	76	NA	Ruler
Mukherjee et al. [14]	New Ballard score	7.7	92.3	86.3	NA	Ruler
Kc et al. [12]	Last menstrual period	7.8	76.9	53.9	10.6	Ruler
Marchant et al. [15]	Eregie score	8	93	58	43	Ruler
Lee et al. [10]	Dating ultrasound	7.5	64	35	8	Ruler
Fawziah et al. [16]	NA	7	75	98.1	94.3	Ruler
Thi et al. [18]	New Ballard score	7.3	80	81	79	Ruler
Dagnew et al. [17]	New Ballard score	7.3	98.5	96.3	93	Calipers

In 2019, up to 20% of births across the world occurred in the absence of a skilled birth attendant [2]. It is also estimated that up to 10 million preterm births occur at home or in primary health care centres in low- to middle-income countries [3]. Non-medically trained community health workers are often the first and sometimes only point of contact for these babies. Lack of recognition and appropriate referral of preterm babies due to an inability to assess gestational age or incorrect assessment can delay potentially lifesaving therapy, particularly when decisions regarding transport to a higher care centre need to be made. Clinically, neonates at gestational ages <37 weeks and particularly when born at a gestational age of <34 weeks are at higher risk and are more likely to require referral and intensive care. In the community, identification and classification of neonates into these gestational age strata will aid in decision making.

OBJECTIVE

The objective of this study was to determine the correlation of postnatal (<48 h) foot length measurement (with calipers) with gestational age as determined by antenatal dating ultrasound to create a predictive model for the same, and propose foot length measurement cutoffs for <37 and <34 weeks of gestation.

Secondary objectives were to assess the correlation between foot length as measured with calipers and that measured with a ruler and a paper footprint, measured by non-medically trained professionals.

METHODOLOGY

Setting

The study was carried out in the babies admitted to the Neonatology unit of the Christian Medical College, Vellore over a period of 6 months. This is a tertiary level neonatology unit situated in the Vellore District of Tamil Nadu, South India and includes wards for the admission of the babies roomed in with their mothers. The unit catered to 14 139 deliveries in 2019.

Study population

This was a hospital-based cross-sectional study that included inborn singleton babies born during the study period (August 2016 to April 2017) who had an in-house first trimester dating ultrasound. Babies with foot deformities—e.g. congenital talipes equinovarus, rocker bottom foot, vertical talus and major congenital anomalies were excluded. Neonates born in multiple gestational pregnancies were excluded.

In low- to middle-income countries, ∼20% of babies are born small for gestational age (SGA). This proportion is highest in South Asia, with estimates as high as 42% [4, 5]. In order to be representative of current realities, it was decided to not exclude on the basis of SGA, defined as weight <10th percentile for the GA as per the Fenton’s growth chart [6].

From a previous study that involved term appropriate for gestational age neonates [1], the correlation between foot length and gestational age was found to be 0.88. Since we were planning to include SGA neonates, it was possible that our expected correlation would be less. So with a minimum expected correlation that would have clinical significance in our sample to be 0.84 with power at 80% and alpha error at 5% for a two-sided test at least 332 babies were required to be assessed.

In order to get a sufficient number of preterms, as well as babies who were term SGA, it was decided to assess babies in a ratio of 1 : 1 : 3—with 100 babies to be assessed that were <34 weeks, 100 that were between 34 and 36 + 6 weeks and 300 term babies, which would include term SGA babies. Thus, a total of 500 babies were planned to be assessed.

Recruitment was prospective and in consecutive order of birth. All babies were assessed within 48 h of delivery. Antenatal history, risk factors such as pregnancy-induced hypertension, gestational diabetes mellitus and history suggestive of an intrauterine infection were collected from the antenatal mothers’ folders once informed consent was given to be included in the study. All procedures were approved by the Institutional Review Board and the Institutional Ethics Committee, CMC Vellore.

Methods

Anthropometric measurements of the baby were taken as per standard protocol of the nursery. The weight of the neonate was measured at birth (within 1 h) using a digital weighing scale manufactured by Essae, Model No. 852 with an accuracy of ±5 g [7]. Length was measured using an infantometre, with an accuracy of 1 mm as per standard recommendations [8].

Neonatal footprints which are recorded for reasons of identification as a protocol in the hospital were taken at birth within 1 h of life. This involved pressing the newborns’ foot into a stamp pad and making the impression of the same onto paper cards.

The foot length was measured using a metal Vernier calipers (Mitutoyo, Model 530-101, with a range of 0–150 mm and accuracy of 0.05 mm) by Neonatology Residents and the measurement of the foot using the ruler and the measurement of the paper footprint was done by social workers who were not medically trained. Measurements were carried out within 48 h of life. Standard protocols described in the literature [1] were used. For reasons of uniformity, it was decided to measure the left foot. In case this was not possible (e.g. IV line with splint in situ) the right foot was measured using both calipers and ruler. The investigators were instructed to hold the caliper in the right hand and the foot in the left. The sole of the foot was placed on the calliper bar, the head of the first metatarsal and medial aspect of the heel were aligned parallel with the edge of the instrument and the sliding measure was manipulated using the thumb of the right hand to bring it to the edge of the heel. Three measurements were taken in millimetres, the average of which would serve as the reference standard for comparison. Foot length measurement by a hard, plastic, commercially available ruler was measured similarly and recorded in centimetres to one decimal place. The measurement of the foot from the paper footprint was also from the heel to the longest toe, using the same ruler. Only complete footprints that involved both the heel and big toe were taken, with minimal smudging of the ink involved in the creation of the paper footprint. Interobserver variability was calculated every 100 babies, the techniques used to measure foot length reassessed and the correct method emphasized.

The result of the dating ultrasound and therefore the gestational age was known only to the principal investigator. The assessors of the foot length via the calipers, ruler and paper footprint and ruler were blinded to the gestational age and the birth weight.

Data entry was done using Epidata. Statistical analysis was done using SPSS 16.0. Pearson’s correlation was used to assess the relation between foot length and gestational age of the baby based on antenatal dating ultrasound. Simple linear regression was used to predict the gestational age based on the caliper foot length. From the derived model, cutoff for foot length was established using receiver operating characteristics curve (ROC) for preterm babies for both <37 and <34 weeks. Intraclass correlation (ICC) was performed to assess the reliability of measuring Foot length using calliper when compared with a ruler or paper footprint.

RESULTS

A total of 520 inborn singleton babies were recruited into the study between August 2016 and April 2017 with gestational ages ranging from 27 to 41weeks and birth weights ranging from 880 to 4720 g. The mean gestational age was 36.5 ± 3 weeks, with mean birth weight of 2552.3 ± 7 g (Table 2).

Table 2

Demographic characteristics of the study population

	n = 520 (%)
Gender
Male	261 (50.2)
Female	259 (49.8)
Gestational age
≥37 weeks	319 (61.3)
34–36 + 6 weeks	100 (19.2)
<34 weeks	101 (19.4)
Birth weight
≥2500 g	287 (55.2)
1500–2499 g	176 (33.8)
<1500 g	57 (10.9)

Of the term babies, 24% (78 babies) were term SGA. Eighty-four (16.15%) mothers were diagnosed with pregnancy-induced hypertension and 95 (18.26%) of mothers had gestational diabetes mellitus. None of the mothers had proven intrauterine infections. The mean foot length as measured by calipers was 66.97 mm (with a range of 48.33–88 mm) (Table 3).

Table 3

Measurement of caliper foot length for each gestational age

Gestational age (weeks)	n	Foot length (mm), mean ± SD
27	1	56
28	4	53.25 ± 1.89
29	8	57 ± 3.70
30	12	56.80 ± 3.65
31	17	58.96 ± 4.30
32	29	62.59 ± 6.22
33	30	66.07 ± 4.13
34	19	68.28 ± 4.77
35	31	70.41 ± 3.37
36	50	71.42 ± 4.39
37	60	74.25 ± 4.22
38	83	76.97 ± 3.85
39	101	78.43 ± 3.76
40	74	79.11 ± 3.76
41	1	75

Foot length did not vary significantly based on gender (p = 0.34). Interobserver correlation between the investigators (using calipers) was 0.95 and interobserver correlation of the social workers using the plastic ruler and paper footprint was 0.92 and 0.91, respectively.

The correlation between the measurements done by the investigators using calipers and the measurements made by the social workers using the ruler and the paper footprint were assessed. The Cronbach’s alpha for foot length as measured by the ruler was 0.97, with ICC coefficient of 0.97 (0.96–0.97). The Cronbach’s alpha for foot length as measured by footprint was 0.92, with ICC coefficient of 0.92 (0.90–0.93) indicating high levels of internal consistency.

Correlation of foot length with gestational age

The correlation between gestational age and calliper foot length was 0.89. Based on the correlation between the foot length measurements using calipers and the gestational age, a predictive model was created and the following equation derived.

GA = 12.582+ (0.329) caliper foot length (mm).

The equation thus derived was externally validated on a separate cohort of 100 neonates of varying gestational ages (<34 weeks: n = 30, 34–36 weeks: n = 30 and >37 weeks: n = 40) that were admitted in the nursery.

The Pearson’s correlation between foot length as measured by calliper and ruler was 0.95 and between calliper and paper footprint was 0.87.

The correlation between foot length and birth weight was 0.84.

Use of foot length measurements to identify neonates postulated to be at high risk

ROC curves were created for foot lengths for gestational ages of <37 and <34 weeks. Sensitivity of operational cutoffs was set at 95%. Thus, <70 mm to identify <37 week babies and <65 mm to identify <34 week babies were chosen. Using these cutoffs, specificity, positive and negative predictive values were calculated.

ROC curves were created for foot lengths for birth weights of <2.5 and <1.5 kg. At a sensitivity of 95%, operational cutoffs for these categories were 71.5 and 65 mm.

Using the cutoffs of 70 mm for the categories of <37 weeks and <2500 g and 65 mm for the categories of <34 weeks and <1500 g, specificity, positive predictive values and negative predictive values were calculated (Table 4).

Table 4

Specificity, positive and negative predictive values of the determined cutoffs for the categories of <37 and <34 weeks and <2500 and <1500 g

	70 mm		65 mm
	<37 weeks	<2500g	<34 weeks	<1500 g
Specificity (%)	72.4	98.2	75.2	92.6
Positive predictive value (%)	84.1	97.1	94.2	62.2
Negative predictive value (%)	84.2	80.7	84.4	99.8

When the cutoff of 70 mm was applied to the term SGA babies, it was found that 25% of these babies were classified preterm.

DISCUSSION

An estimated 1 million babies die globally annually due to prematurity related complications [9]. Early identification of preterm babies at the community level in order to appropriately monitor and refer relevant high-risk babies to suitable facilities for appropriate care is of paramount importance. The ability to identify a preterm baby even among ‘small’ or low birthweight babies has an advantage as prematurity carries a greater risk than growth retardation in most babies. With difficulties in motivation of families for referral and in providing transport, proper assessment may influence the decision for the care to be planned.

In a study performed in rural Bangladesh using community health workers to assess various methods of gestational age, the Ballard score as well as the simplified scores—Capurro, Bhagwat and Eregie had poor sensitivity in identification of neonates <37 weeks gestation. This study showed that neonatal anthropometry had an ROC of 0.52–0.80 for the classification of preterm infants [10]. Anthropometry has been evaluated previously in view of its technical ease to teach and reproducibility. Equations thus generated have showed a sensitivity of 75%, for detecting preterm neonates [11]. However, considerable mathematical calculation has made practical application of this unpopular in a community setting.

In the community, health workers or volunteers often lack tools and clinical skill to identify newborns as high risk. The measurement of foot length has multiple advantages—it is easy to teach, reproducible, requires less manipulation and only minimal exposure of the baby.

All assessments in our study were carried out within 48 h of delivery. Assessment of foot length at a later point—at 5 days of life has not shown a significant difference [12]. The relatively poorer correlation of foot length measurement (and other clinical methods) in the identification of preterm babies in the cohort from Bangladesh was postulated to be due to the delay in assessment and survival bias [10].

Foot length models and recommended operational cutoffs for the identification of preterm babies have been proposed [10, 12–18]. However, this is the first study that reports cutoffs based on gold standard assessments of the dating ultrasound against the reference standard of the caliper foot length with measurements taken by trained medical professionals. The cutoffs proposed by our study also have a higher positive predictive value, when compared with the cutoffs proposed by other studies. Our study cutoff of 7 cm for the identification of preterm babies is similar to that of Fawziah et al. [16] and is lower than that proposed by other studies. Genetic, racial and geographic differences, the inclusion of SGA babies in the study, as well as variation secondary to the different instruments used are likely to play a role.

At <34 weeks, the neonate is more likely to require extensive medical intervention. The only other study that reports results specifically to identify neonates at <34 weeks was carried out in 1988, from India. Similar to our findings, it was postulated that a cutoff of 6.5 cm would correlate to a gestational age of 34 weeks. Gestational age assessment was based on the last menstrual period and foot length was measured from a paper footprint [19]. Our cutoff of 6.5 cm correlates with both the gestational age of <34 weeks and the birth weight of <1500 g, both of which are significant clinical end points for management and referral.

The effect of foetal growth restriction and the baby being born SGA on foot length has been debated. A study done to correlate foot length (using a ruler) and gestational maturity (gestational age assessed by New Ballard scoring system) in neonates showed good correlation among term as well as preterm appropriate and SGA neonates [20]. However, this study did not propose cutoffs. In our study, a quarter of the babies were term SGA, and using our study cutoffs would lead to classification of a quarter of these babies as preterm, which is a clinically acceptable outcome.

In our study, both ruler and paper footprint were found to have good correlation with calipers, with ruler better than paper footprint. In the community, the use of a ruler would be both acceptable and inexpensive. A transparent box with slider attached has been proposed as an alternative means to measure foot length in the community [21]. This however has not been tested. Foot length measurement from footprint has been used as a surrogate for birthweight and an indicator for referral in the Rural Neonatal Care Project in Danahu, Maharashtra, India [19]. In southern Tanzania, a card that showed a scale picture of a foot with classifications as very short (<7 cm), short (7–7.9 cm) and not short (>8 cm) was tested among community volunteers and found to be moderately reliable, but overestimated the babies requiring referral and extra care [22]. A foot length caliper that was colour coded to indicate actions to be taken such as home care; care at a rural health care centre or referral to a tertiary care centre has been tested. Gestational age was calculated from last menstrual period. The cutoff values for gestational age <34 weeks and birth weight >2.5 kg used were 6.8 and 7.3 cm [23].

CONCLUSION

The results of this study are encouraging and could serve as the basis for guiding treatment of babies born in communities where healthcare reach is limited. The study is adequately powered and used the current gold standard for assessment of gestational age (the antenatal dating ultrasound). The reference standard of caliper foot length measurement had significant correlation with gestational age, high interuser correlation and high positive predictive values. The correlation of the foot length measurements using the ruler and the paper footprint with that of calipers shows significant promise for translation into use in the community after evaluation in a community-based study. Lifesaving decisions regarding intervention and referral could potentially be made easier and more efficient if proven that using these measurement methods, easily taught and requiring minimal technology can decrease morbidity and mortality on a community level.

ACKNOWLEDGEMENTS

The authors thank Ms. Shreya Mondal and Mrs. Indira Balan and Drs. Avadesh Joshi and Abhilipsa Acharya, Department of Neonatology, CMC Vellore.

FUNDING

This work was supported by a fluid research grant from the Institutional Review board, CMC Vellore.