Evaluation of the Accuracy of Transcutaneous Bilirubinometry on the Forehead, Sternum, Interscapular Region, Ear Cartilage, and Wrist and Comparison with Total Serum Bilirubin in Newborns with GA Above 30 W in Shahid Beheshti Hospital, Isfahan, Iran

Abstract

Background:

Prompt and precise bilirubin assessment is critical for the appropriate intervention in neonatal jaundice management. Various healthcare settings have adopted transcutaneous bilirubin (TcB) devices for this purpose. This study aimed to analyze the relationship between TcB readings obtained from five distinct body regions and total serum bilirubin (TSB) levels.

Materials and Methods:

The research encompassed 202 neonates exhibiting clinical signs of jaundice, admitted to the neonatal intensive care unit (NICU) and general wards of Beheshti Hospital in Isfahan. The cohort included neonates of both genders with a gestational age of 30 weeks or more. Measurements of TcB were performed using the MBJ20 jaundice detector probe on the forehead, sternum, auricle, wrist, and interscapular region, and these were then compared with the TSB levels.

Results:

The study included a balanced sample of 202 newborns, with an equal distribution between males and females. The mean gestational age was 34.19 ± 2.63 weeks; the average birth weight was 2083 ± 692.18 grams, and the TSB level was 9.95 ± 2.68 mg/dl. The findings revealed a robust, direct, and meaningful correlation between the TcB and TSB levels across all the sites examined. Notably, the correlation was statistically more pronounced at the sternum and forehead.

Conclusion:

The study confirms that noninvasive TcB assessment at two specific sites—the sternum and forehead—using the MBJ20 bilirubinometer, provides significant congruence with the TSB levels determined by clinical laboratory techniques in both term and preterm neonates. Measurements at the sternum are marginally more precise than those at the forehead.

INTRODUCTION

Jaundice is a prevalent neonatal condition marked by an excess of bilirubin, causing yellow discoloration in the skin and sclera. It is notably widespread, affecting a significant majority of full-term and nearly all premature infants. While commonly a benign and temporary condition, it can assume a physiological role in the development of neonates; however, excessive bilirubin levels beyond the physiological limit can accumulate in the brain, potentially resulting in neurological impairment. Consequently, the prompt detection of hyperbilirubinemia is of paramount importance and warrants thorough examination.[1,2,3]

Historically, total serum bilirubin (TSB) testing has been the benchmark for diagnosis, yet this traditional method is invasive and can be painful, stressful, and time-intensive for neonates. Alternatively, transcutaneous bilirubin (TcB) measurement offers a swift, uncomplicated, and noninvasive approach. It has demonstrated a superior capacity for visual evaluation in identifying infants at elevated risk of hyperbilirubinemia, serving effectively as a screening tool.[4,5,6]

There is established evidence of a linear relationship between TcB readings and serum bilirubin (SBR) levels, with several studies advocating for the use of TcB as a reliable screening modality to identify significant jaundice, potentially reducing the frequency of blood draws required.[7,8]

Clinically, the forehead is commonly selected for TcB measurements. Our research, however, seeks to assess the TcB values across five distinct cutaneous regions—forehead, sternum, auricle, wrist, and interscapular area—and their correlation with TSB to ascertain the most reliable measurement site.[9]

MATERIALS AND METHODS

This investigation took place from May 2022 through January 2023 within the neonatal intensive care unit (NICU) and general wards of Shahid Beheshti Hospital in Isfahan, Iran. Ethical approval for the study was granted by the local ethics committee (IR.MUI.MED.REC.1400.177), and parental consent was acquired for every participating infant.

A total of 202 clinically jaundiced neonates, who had not yet started phototherapy, were enrolled in this study. The inclusion criteria consisted of neonates of either sex, with a gestational age of 30 weeks or more, an Apgar score of 7 or above at 5 minutes post-birth, absence of hemolytic disease or skin disorders, no clinical signs of sepsis, and no history of exchange transfusions or phototherapy in the 24 hours preceding the study. Infants presenting with significant congenital or skin anomalies were excluded.

The transcutaneous jaundice detector MBJ20 was employed to measure bilirubin levels. This device calculates bilirubin concentration by processing the reflection of visible light from the skin and factoring out the absorption by other skin components, such as hemoglobin and melanin, mathematically to ascertain the bilirubin concentration in the capillary and subcutaneous tissue. The TcB readings are displayed in a range from 0.0 to 32.0 mg/dl, in line with the manufacturer’s specifications.

For each neonate, the following measurements were conducted:

• TcB levels were assessed using the MBJ20 detector.

• SBR levels were quantified through direct spectrophotometry in a clinical laboratory setting.

• Within this research, the MBJ20 Bilichek instrument was applied to five distinct locations on the neonate’s body—specifically the forehead at a midpoint between the hairline and glabella, the sternum, auricle, wrist, and the interscapular area. Subsequently, a digitized display indicated the level of TcB in milligrams per deciliter at each site. Before each session, the apparatus was adjusted in accordance with the guidance provided by its manufacturer.

• For the assessment of newborn TSB, heel punctures were performed to draw blood, which was then collected in heparinized capillary tubes, shielded from light, and forwarded to a laboratory for evaluation.

Entries were made on a standardized recording form for both TcB and TSB levels, as well as for other essential information, such as the neonate’s gestational age, sex, and body weight. Treatment and management of the patients followed the established World Health Organization guidelines for handling neonatal jaundice.

The analysis incorporated a computation of the mean difference (MD) between the bilirubin levels as projected from TSB readings. A paired t-test was conducted to ascertain the presence of any notable differences. Additionally, the standardized MD (SMD) was determined to assess the extent of discrepancy between the two methods. To evaluate the consistency across TSB and alternative testing methods, Bland–Altman plots were employed. Furthermore, the degree of congruence between these methods was quantified through the concordance correlation coefficient (CCC), which measures the extent to which observation pairs align along the line at a 45-degree angle originating from the zero point. The interrelation of TSB and alternative methods was visualized through scatter plots with a regression line fitted based on the Deming regression methodology, where an ideal correlation is indicated by a slope close to unity and an intercept approximating zero. All statistical examinations were executed utilizing MedCalc version 20 and Statistical Package for the Social Sciences (SPSS) version 27 software, with significance determined by a P value of less than 0.05.

RESULTS

The research encompassed 202 infants, split almost evenly between 104 females and 98 males. Their average weight at birth was documented at 2083.41 grams, and the gestational duration averaged at 34.19 weeks. A detailed overview of these primary attributes is illustrated in Table 1.

Table 1.

Characteristics of infants

Variable		Mean
Gestational age (week)		34.19±2.63
Preterm		160 (79.2)
Term		42 (21.8)
Infant gender
Male		98 (48.1)
Female		104 (51.9)
Weight (gr)		2083.41±692.18

Within the study, newborn TSB levels were assessed, ranging from 2.5 to 20 mg/dl, with an average concentration of 9.9 mg/dl. Table 2 encapsulates the associated TcB figures, captured via the MBJ20 bilirubinometer, revealing both the mean variance and the concordance with TSB across different testing locations. The average discrepancy between TSB and sternum-based readings was a negligible 0.14 mg/dl. The mean bilirubin level measured at the forehead was 0.34 mg/dl less than TSB (P = 0.044), indicating a statistically significant variance. Moreover, average readings from the shoulder (MD = 1.50), wrist (MD = 3.54), and auricle (MD = 0.90) exceeded TSB, each distinction reaching statistical significance (P <0.05).

Table 2.

Mean differences and concordance TSB with sternum, forehead, shoulder, wrist, and auricle

		Mean±SD		Mean difference±SD		P a		SMD (95% CI)		Concordance correlation coefficient
TSB with		9.95±2.68
Sternum		10.09±2.58		-0.14±2.42		0.404		-0.06 (-0.20,0.08)		0.58 (0.48,0.66)
Forehead		10.29±2.54		-0.34±2.38		0.044		-0.14 (-0.28,0.00)		0.58 (0.48,0.66)
Shoulder		8.45±2.53		1.50±2.49		<0.001		0.60 (0.45,0.75)		0.46 (0.37,0.55)
Wrist		6.40±2.33		3.54±2.61		<0.001		1.36 (1.17,1.55)		0.23 (0.16,0.30)
Auricle		9.05±2.75		0.90±2.96		<0.001		0.30 (0.16,0.44)		0.39 (0.27,0.49)

SMD=Standardized mean difference. a=resulted from paired t-test

The stratified concordance of TSB with alternative assessment methods based on gestational age—term versus preterm—is presented in Table 3. It was observed that full-term infants exhibited a higher correlation between the measurement techniques and TSB in comparison with preterm infants. Nonetheless, the performance of the measurement methods did not significantly differ between full-term and preterm infants’ outcomes. The highest levels of agreement with TSB were seen at the forehead (CCC = 0.55 in preterms and CCC = 0.65 in terms) and at the sternum (CCC = 0.55 in preterms and CCC = 0.65 in terms), while the wrist displayed the least concordance (CCC = 0.22 across both categories).

Table 3.

Mean differences and concordance TSB with sternum, forehead, shoulder, wrist, and auricle among term and preterm infants

		Mean±SD		Mean difference±SD		P a		SMD (95% CI)		Concordance correlation coefficient
Preterm infants (n=160)
TSB with		9.64±2.45
Sternum		10.06±2.53		-0.42±2.36		0.025		-0.18 (-0.33,-0.02)		0.55 (0.43,0.64)
Forehead		10.31±2.40		-0.67±2.25		<0.001		-0.30 (-0.45,-0.14)		0.55 (0.44,0.65)
Shoulder		8.36±2.48		1.28±2.44		<0.001		0.53 (0.36,0.69)		0.44 (0.33,0.55)
Wrist		6.36±2.20		0.51±2.76		0.021		0.19 (0.03,0.34)		0.22 (0.14,0.29)
Auricle		9.13±2.64		3.29±2.49		<0.001		1.32 (1.10,1.53)		0.41 (0.28,0.53)
Term infants (n=42)
TSB with		11.11±3.19
Sternum		10.20±2.79		0.92±2.40		0.018		0.38 (0.07,0.69)		0.65 (0.44,0.79)
Forehead		10.21±3.06		0.90±2.49		0.024		0.36 (0.05,0.67)		0.65 (0.45,0.80)
Shoulder		8.79±2.71		2.33±2.54		<0.001		0.92 (0.55,1.27)		0.49 (0.29,0.65)
Wrist		6.58±2.78		2.36±3.27		<0.001		0.72 (0.38,1.06)		0.25 (0.11,0.39)
Auricle		8.75±3.14		4.53±2.84		<0.001		1.60 (1.13,2.05)		0.36 (0.13,0.55)

SMD=Standardized mean difference. a=resulted from paired t-test

Figure 1a-e present the Bland–Altman plots to visualize the agreement across the various TSB measurement approaches. The plots utilize the MDs (vertical axis) against the average bilirubin estimates (horizontal axis). The discrepancy in values between the sternum and TSB and the forehead and TSB readings is minor, displaying no progressive deviation in MD, regardless of the increase in the estimated bilirubin average. In contrast, the shoulder and wrist techniques frequently produced overestimated bilirubin levels. The CCC, signifying the extent of agreement between TSB and the alternative methods, ranked highest for the sternum (CCC = 0.58) and the forehead (CCC = 0.58), with the wrist recording the lowest (CCC = 0.23). The shoulder and auricle registered CCCs of 0.46 and 0.39, respectively.

Figure 1.

Bland–Altman plots comparing total serum bilirubin (TSB) measurements with sternum (a), forehead (b), shoulder (c), wrist (d), auricle (e)

In Figure 2a-e, we demonstrate the relationships between TSB values and those obtained through alternative methods, alongside the regression line estimated using the Deming method. Ideally, a slope approaching 1 and an intercept close to zero suggest a more accurate concordance between the two methodologies. There was a significant positive association between TSB and the bilirubin levels ascertained by all alternative methods. Notably, the most robust correlation and the minimal bias relative to TSB were observed, in descending order of accuracy, at the sternum, forehead, shoulder, auricle, and wrist.

Figure 2.

Correlation between total serum bilirubin (TSB) levels and readings obtained from sternum (a), forehead (b), shoulder (c), wrist (d) and auricle (e), as well as the corresponding regression line calculated via Deming regression analysis. A slope approaching unity and an intercept close to the origin are indicative of a high degree of concordance between the two measurement methods

DISCUSSION

Neonatal jaundice is the predominant condition necessitating medical intervention in newborns. Thus, the application of a noninvasive, pain-free, and trustworthy technique for ascertaining bilirubin levels is critical in all environments, whether inpatient or outpatient. This approach mitigates the risks linked to clinical judgments and subsequent complications.

In our investigation, we have established a notable correlation between TcB measurements using the MBJ20 jaundice detector and TSB levels in infants exhibiting jaundice across five different skin sites, with the sternum and forehead sites showing the highest concordance, respectively. Our findings also indicated uniformity in measurement outcomes between full-term and preterm infants, irrespective of the assessment technique or site employed.

Building upon the work of El-Kabbany et al.,[5] which examined 316 neonates matching our study’s gestational age and criteria, our results were consistent, with slight variations potentially due to the increased sample size and the employment of an alternate bilirubinometer (JM103) in their study.

Further comparative analyses with different bilirubinometer models[10,11,12,13] have corroborated our study’s outcomes. Research by Chimhini GLT et al. involving 283 Zimbabwean neonates, both term and preterm, utilizing the Draeger JM103 bilirubinometer,[10] showed a strong correlation for both sternum and forehead readings. The sternum was revealed to be the superior site for detecting jaundice in infants when compared to the forehead.

A similar study by Kumar D et al. involving 276 Indian neonates with a gestational age of 35 weeks or more[11] discovered substantial linear correlations between TSB and TcB, specifically before commencing phototherapy. In another study, Mohamed M et al. utilized the Jaundice Meter JM105 on 130 Malay neonates, also confirming the linear relationship between TSB and TcB measurements, favoring the sternum site as per Bland–Altman analyses.[12]

Lilly Gwendoline et al.’s study of 283 term and preterm infants revealed greater precision in TcB readings at the sternum compared with the forehead.[13] Moreover, the research conducted by Yang ST et al.[14] concluded that the accuracy of TcB measurement in preterm infants was equivalent to that in full-term infants.

Our study reinforces the reliability of transcutaneous bilirubinometry as an effective screening tool to identify infants with jaundice, potentially reducing the frequency of invasive TSB assessments.

The data our research yielded align with the findings of Wong, Quintas, and Conceição et al., which affirmed a heightened precision in TcB readings at the sternum in contrast to the forehead.[15,16,17]

This concurrence in TcB levels across different studies may stem from the gestational age parity among the neonates selected for these analyses. For instance, Cat FC et al. examined 105 term and early-term Turkish infants within the gestational window of 37 to 42 weeks, employing the same bilirubinometer model (MBJ20). Their study highlighted a superior correlation in TcB measures at the forehead as opposed to the sternum.[18]

Conversely, Jeon J et al. undertook a comprehensive study of 1084 infants, averaging a gestational age above 35 weeks and a mean fetal weight of 2400 grams. Their findings suggest that transcutaneous bilirubinometry yields more accurate results in infants who have not undergone phototherapy, with the forehead proving more reliable than the sternum for those who have.[19]

In research conducted by Agrawal G et al., a cohort of neonates with gestational ages under 34 weeks was scrutinized at three different skin sites: the interscapular region, sternum, and forehead. Their findings indicated that while all sites showed a correlation with SBR levels, the interscapular region presented the highest accuracy.[20]

Similarly, Lucanoval LC et al. assessed 102 full-term Slovak infants, determining that TcB measurements were dependable when performed on the forehead, sternum, and abdomen, with the foremost results observed at the forehead.[21]

In a Japanese context, Yamana K et al. evaluated 82 infants with gestational ages exceeding 35 weeks. Their investigation concluded that the scaphoid fossa and conchal cavity offer TcB measurements that are more accurate and closely aligned with TSB than readings from the ear lobule area.[22]

Our study posits that the discrepancies in TcB values noted at the forehead versus the sternum could be a function of differing light exposure on the skin—where the chest area is typically shielded by clothing, and the forehead is more light-exposed. Additionally, variations in skin properties, such as collagen and melanin concentration, may influence the wavelength absorption by the TcB photometry, engendering the observed variances in TcB estimation. Furthermore, the sparser hair on the chest area relative to the forehead in neonates may contribute to this disparity.

Other investigations have highlighted a propensity of bilirubinometer devices to either overestimate or underestimate SBR levels.[15] In our focused study on anatomical sites, we discerned that the average TcB levels were generally higher at the sternum and forehead but lower at other regions (auricle, wrist, and interscapular area) compared with the mean TSB levels.

In an investigation by Mohamed M et al. employing the Jaundice Meter JM105, it was observed that TcB readings at the forehead and sternum tended to report lower levels than those of newborn TSB.[12] Complementary findings by Simsek FM et al. in Turkey and Engle WD et al. in Hispania also indicated that TSB values were consistently underestimated when using the transcutaneous bilirubinometer in their respective studies.[23,24]

Concurrently, research by Jandial S et al. in India and Olusanya BO et al. in Africa found the opposite, with TcB readings generally exceeding TSB levels when using devices, such as the Drager JM103 in India and both BiliCheck and JM103 in neonates of black African descent.[25,26]

These variations in TSB estimations, as observed by our study in comparison with previous ones, could be attributable to ethnic and skin tone differences across populations. TcB levels have been noted to be higher than TSB, particularly in populations with darker skin, such as those from Indian and African backgrounds, potentially leading to unnecessary interventions for neonatal jaundice. Therefore, in instances where TcB readings and clinical assessments of jaundice do not align, a confirmatory TSB test is recommended.

Our study was limited to a single institution, Shahid Beheshti Hospital, which predominantly serves an Iranian patient base, a group with less ethnic variety. It becomes imperative to conduct further research encompassing a broader array of ethnic groups and skin tones, as these factors can influence the accuracy of transcutaneous bilirubinometry.

There is a call for additional research on the correlation of TcB values to TSB pre- and post-phototherapy to ascertain the efficacy of TcB devices in our demographic. Such studies could also assist in the creation of a tailored TcB nomogram, facilitating more precise prediction of hyperbilirubinemia within our community.

CONCLUSION

Our analysis confirmed a substantial linear correlation between TcB and TSB values. Evaluations of TcB at the sternum and forehead provided satisfactory diagnostic precision in determining the need for phototherapy in neonates. Furthermore, our research indicated that there was no notable disparity between full-term and preterm infants in terms of measurement accuracy or preferred assessment locations.

It was also observed that TcB could either understate or overstate the TSB values depending on the site of measurement, with the sternum and forehead typically showing higher readings and considerably lower readings at the auricle, wrist, and interscapular region.

While a bilirubinometer serves as a practical, noninvasive screening option for mild hyperbilirubinemia in newborns, especially when used at the sternum or forehead, reliance on TSB measurements is advised in severe cases of hyperbilirubinemia or when there is a significant discrepancy between TcB levels and clinical evaluations of jaundice.

Ethics approval and consent to participate

The study protocol was approved by the ethical committee of Isfahan University of Medical Sciences (code: IR.MUI.MED.REC.1400.177).

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Isfahan University of Medical Sciences, Isfahan, Iran, provided the financial resources for the present project.