A comparative study of transcutaneous bilirubin measurement and total serum bilirubin in neonatal hyperbilirubinemia

INTRODUCTION

Neonatal hyperbilirubinemia (NH), commonly referred to as neonatal jaundice, is a prevalent condition in the neonatal period, primarily resulting from the adaptation of bilirubin metabolism during this time. Bilirubin is produced as a byproduct of haemoglobin degradation, with approximately three-quarters of bilirubin generated through oxidation-reduction reactions from haemoglobin breakdown, further assisted by the breakdown of myoglobin, cytochromes and catalase.^[1-3] Neonatal jaundice is a significant global health concern, affecting around 60% of term and 80% of preterm neonates to varying degrees. It is estimated that 1.1 million babies develop severe hyperbilirubinemia annually, particularly in sub-Saharan Africa and South Asia.^[4] In India, the incidence of neonatal jaundice ranges from 54.6% to 77%.^[5] Neonatal jaundice is identified as the seventh most common cause of neonatal mortality globally.^[6]

Clinically, neonatal jaundice is characterised by the yellowish discoloration of the skin and sclera due to elevated bilirubin levels.^[7] Hyperbilirubinemia typically begins in the eyes and progresses to the face, chest, abdomen and legs.^[8] Given these serious risks, accurate measurement of bilirubin levels is crucial for early diagnosis and intervention. The gold standard for diagnosing hyperbilirubinemia remains the measurement of total serum bilirubin (TSB) levels, which reliably indicate the severity of the condition, as outlined by the National Institute for Health and Clinical Excellence guidelines.^[8]

However, conventional TSB testing is invasive, requiring blood sampling that can cause discomfort, stress and a risk of infection to the neonate.^[9] Delays in obtaining bilirubin test results may also postpone the start of treatment for hyperbilirubinemia. Several methods are used for measuring TSB levels, with the Jendrassik-Grof diazo method being the most widely used. This technique involves reacting bilirubin with diazotised sulfanilic acid and an accelerator at acidic pH, with absorbance measured at 600 nm, accurately quantifying both conjugated and unconjugated bilirubin, making it suitable for diagnosing jaundice in neonates and adults.^[10,11] Other methods, such as high-performance liquid chromatography, offer detailed separation of bilirubin fractions but are less commonly used due to their complexity.^[12] Direct spectrophotometry measures bilirubin absorbance around 454 nm, using a two-component system to improve accuracy by subtracting haemoglobin interference, though it can be affected by other pigments.^[13] While enzymatic methods, which use bilirubin oxidase, are less prone to haemolysis interference and can differentiate between bilirubin forms based on PH.^[10,12]

In light of these challenges, transcutaneous bilirubin (TCB) measurement has emerged as a widely used, non-invasive alternative for screening jaundice in neonates. TCB involves the use of a bilirubinometer, which measures bilirubin levels by directing light into the skin and assessing the intensity of light returned. It is particularly beneficial as it eliminates the need for repeated blood samples and provides quick results.^[9,14] The measurement is usually taken by gently pressing the meter against the sternum or forehead. TCB measurement provides an immediate (less than a minute) result of bilirubin levels. Using this point-of-care device saves time compared to measuring serum bilirubin and may reduce costs associated with measuring serum bilirubin in newborns. However, the accuracy of TCB results may be affected by gestational age, body weight and skin colour.^[15]

TCB has been shown to correlate well with TSB values, making it a valuable screening tool for hyperbilirubinemia management worldwide. The American Academy of Paediatrics (AAP) recommends pre-discharge evaluation of bilirubin levels in all neonates using either TSB or TCB measurements.^[9] Studies, including a meta-analysis by Okwundu et al., report that the sensitivity of TCB for detecting significant hyperbilirubinemia varies widely, ranging from 74% to 100%, while specificity ranges from 18% to 89%.^[15] In addition, Tiwari and Pise concluded that there is a strong correlation between serum and TCB levels both before and after phototherapy. TCB leads to early intervention and timely management of hyperbilirubinemia, potentially reducing the need for more invasive procedures.^[14]

Given these findings, NH remains a significant global health concern, with early detection and accurate diagnosis being critical to prevent severe complications. While TSB is the gold standard, its invasive nature and potential delays highlight the need for non-invasive alternatives. TCB measurement offers a promising solution, providing quick, reliable results without the discomfort of blood draws. Thus, this study aims to compare the accuracy of TCB with TSB in neonates with hyperbilirubinemia.

MATERIAL AND METHODS

This was a cross-sectional study performed over a period of 15 months, that is, from August 2023 to November 2024.

All neonates presenting with neonatal jaundice in the Department of Paediatrics, SNR District Hospital Kolar, Karnataka, over a period of 15 months were enrolled.

The study protocol was approved by the Institutional Ethics Committee and written informed consent was obtained from the parents or legal guardians of the neonates before study initiation.

RESULTS

Out of 189 neonates, the majority were in the age group of ≥37 weeks (51.32%) followed by 34–36 (+6 days) weeks (35.98%), while the least number of neonates were in the age group of 28–33 (+6 days) weeks (12.70%). The age of the neonates ranged from 28 to 40 weeks, with a mean of 36.18 ± 2.51 weeks.

The majority of the neonates were male (61.38%) with a male-to-female ratio of 1.59.

Among the 189 neonates, the majority (61.91%) had a birth weight of 2.5 kg and (24.34%) had a birth weight between 2.0 and 2.49 kg, and 12.17% of neonates had a birth weight between 1.0 and 1.99 kg, and only 1.59% <1.0 kg.

Among the 189 neonates, 97 (51.32%) were born full term, followed by 24 (12.70%) neonates born early preterm and 68 (35.98%) neonates born late preterm.

The mean platelet count was 3.40 ± 0.74 million/mcL, while the mean reticulocyte count was 4.07 ± 1.22%. The mean red blood cell count was 5.06 ± 0.24 million/mm³, and the mean haemoglobin level was 14.08 ± 2.38 g/dL. In addition, the mean white blood cell count was 28.49 ± 5.00 10³/mm³.

Among the 189 neonates, the majority had physiological jaundice, with 50 (26.46%) neonates. While 43 (22.75%) had idiopathic causes, 33 (17.46%) had Rh incompatibility and 33 (17.46%) had A,B,O Blood group incompatibility. Sepsis and breast-feeding jaundice each affected 11 (5.82%) neonates, and the least were affected by cephalohematoma, comprising only 8 (4.23%).

During comparison between TCB and conventional TSB levels across three stages of PT, before PT, the two methods showed a strong correlation (r = 0.9355), with TCB readings being slightly higher than TSB. During PT, TCB slightly underestimated TSB, with a strong correlation between them (r = 0.9250). After PT, the difference between the two methods became more pronounced; however, there was a strong correlation between the methods (r = 0.9305). Analysis by independent t-test revealed significantly lower levels with TSB than TCB before (17.79 ± 2.89 mg/dL vs. 17.72 ± 4.42 mg/dL), during (16.20 ± 4.07 mg/dL vs. 16.09 ± 4.02 mg/dL) and after PT (14.67 ± 4.12 mg/dL vs. 13.15 ± 3.77 mg/dL) (all P < 0.05).

The Bland-Altman plot for the phase before initiating PT demonstrates good agreement between TCB and TSB measurements and moderate agreement during PT phase and after PT phase. The mean difference b/n TcB and TSB is 0.0175, 0.1129 and 1.2090 units, respectively.

None of the children were given intravenous immunoglobulin and exchange transfusion; all cases were treated with phototherapy alone (single surface, double and triple based on severity). Cases with severe sepsis and multisystem involvement and with bilirubin encephalopathy were referred to a higher centre.

DISCUSSION

Analysis of study findings showed that 61.9% of neonates had a normal birth weight (>2.5 kg). These findings are consistent with Sahota et al., who reported that 54% of 272 cases were born before 37 weeks of gestational age, with 52% delivered by lower-segment caesarean section and 46.7% having a birth weight of 1.5–2.49 kg.^[16]

In the present study, the most common causes of NICU admission among studied patients were physiological jaundice (26.46%), followed by idiopathic causes (17.46%), Rh incompatibility and ABO incompatibility. However, in other studies, the primary cause of NICU admission was ABO/Rh incompatibility, followed by prematurity, respiratory distress, neonatal jaundice and idiopathic causes.^[17,18]

The TCB and TSB measurements may differ due to their distinct measurement principles. Recognising that TCB is not a direct substitute for TSB is crucial. The AAP recommends using TCB devices only for infants with TSB levels <15 mg/dL, with some experts advocating for routine TSB measurement in all newborns.^[16]

TCB functions as a screening tool for NH based on the assumption that serum and tissue bilirubin remain in equilibrium. It operates by assessing the intensity of specific light wavelengths reflected from the neonate’s skin. Various studies have reported a strong correlation (r = 0.87–0.92) between TCB and TSB. Measurements taken over the forehead before phototherapy have demonstrated high agreement, with Surana et al. and Maisel et al. reporting correlation values around r = 0.83.^[19-21]

However, this relationship shifts once phototherapy begins. The correlation weakens as phototherapy alters bilirubin distribution, reducing TCB accuracy. A systematic review by Nagar et al. noted that TCB remains moderately correlated with TSB during phototherapy, but accuracy is affected by light exposure. Shielded sites, such as the forehead, yield more reliable readings.^[22] Ho et al. similarly found that TCB measurements from covered areas were more accurate than those from exposed skin.^[23]

Despite the overall positive correlation between TCB and TSB, it is essential to recognise the inherent differences between these two measurement methods. TCB primarily reflects extravascular bilirubin, which can lead to discrepancies, particularly at higher serum bilirubin levels where TCB tends to overestimate TSB.^[24]

In the present study, a strong agreement with r = 0.9355 observed between TCB and TSB before phototherapy indicates that both methods effectively reflect bilirubin levels in neonates, with TCB readings being slightly higher due to its measurement of extravascular bilirubin, which can be more abundant in the skin.^[25]

During phototherapy, the slight underestimation of TSB by TCB with r = 0.9250 is likely due to the phototherapy effects on bilirubin metabolism and distribution, which can alter the skin bilirubin content and lead to discrepancies in readings.^[26] The underestimation is particularly pronounced in the early hours of phototherapy, where TCB measurements may lag behind rapid changes in serum bilirubin levels.^[27] During phototherapy, studies have consistently shown that TCB tends to underestimate TSB levels, with a correlation coefficient reported at r = 0.72.^[22] Moreover, Eken et al. noted that TCB levels were generally lower than TSB in covered areas during phototherapy, suggesting that the light exposure impacts the accuracy of TCB readings.^[28]

After phototherapy, the more pronounced difference and lower correlation with r = 0.9305 between the two methods may be attributed to the rebound effect of bilirubin levels as the skin equilibrates post-treatment, which can take several hours. Studies have shown that the correlation between TCB and TSB improves marginally after phototherapy, but it remains lower than pre-treatment levels.^[22] Furthermore, Nagar et al. found that the correlation improved to r = 0.72 after discontinuation of phototherapy, but still indicated a significant difference between the two measurement methods.^[22] The significant differences in TSB levels across the three stages of phototherapy further highlight the dynamic nature of bilirubin metabolism during treatment, necessitating careful monitoring and interpretation of both TCB and TSB measurements.^[29]

Factors such as phototherapy duration and TCB measurement site influence this variability. Jeon et al. found that the forehead provided more reliable TCB readings than the sternum during phototherapy, contributing to differences in correlation coefficients across studies.^[26] However, Lucanova et al.^[36] analysed TCB accuracy at uncovered sites 2 h post-phototherapy, reporting mean TCB-TSB differences ranging from −2.9 to −6.7 mg/dL.

Alsaedi et al. noted that TCB measured during phototherapy showed a negative mean bias that worsened as TSB increased, particularly at higher bilirubin levels.^[30] Conversely, Rohsiswatmo et al. found that TCB tended to overestimate TSB after phototherapy, highlighting the complexities of interpreting these measurements.^[31]

The results from the Bland-Altman analysis indicate a strong agreement between TCB and TSB measurements before the initiation of phototherapy, with a mean difference of 0.0175 units and limits of agreement ranging from 4.1493 to −4.1143 units. This suggests that in most cases, the TCB readings are closely aligned with TSB levels, affirming the reliability of TCB as a non-invasive method for assessing bilirubin levels in neonates prior to treatment. Our findings align with previous studies, which reported good agreement between TCB and TSB, indicating TCB measurements as viable options for guiding jaundice treatment in regions where serum bilirubin tests are not available.^[16,32]

The B-A analysis during phototherapy indicates a mean difference of 0.1129 units between TCB and TSB, with limits of agreement ranging from (2.9589-3.1847). This suggests a moderate agreement between the two methods during the phototherapy phase, indicating that while TCB can be a useful tool, its accuracy may vary significantly under these conditions. This aligns with Olusanya and Emokpae, who reported a moderate correlation between TCB and TSB, with pooled estimates of correlation coefficients ranging from 0.64 to 0.71 during phototherapy.^[33] Nagar et al. noted that while TCB had a significantly positive correlation with TSB during phototherapy, B-A plots showed significant bias and imprecision in the TCB readings, reflecting the challenges of using TCB as a reliable measure under phototherapy conditions.^[22] Contrastingly, Jegathesan et al. found that the agreement between TCB and TSB worsened after the initiation of phototherapy, indicating that TCB may not be as reliable in this context.^[34]

The B-A analysis post-phototherapy showed a moderate agreement between the two methods, though TCB accuracy varies in this context. Katayama et al. similarly reported moderate agreement post-phototherapy, noting that measurement timing and site influence discrepancies.^[35] Nagar et al. further observed that TCB tends to underestimate TSB, particularly immediately after phototherapy.^[22] The B-A plot indicated improved TCB-TSB agreement post-phototherapy; however, TCB still underestimated TSB by −10 ± 31, with 95% limits of agreement between 52 and −72 (P = 0.0001). Eken et al. also found that TCB underestimates TSB, especially when taken from covered skin areas, likely due to bilirubin redistribution following light exposure.^[28]

Measurement site plays a critical role in TCB accuracy. Jeon et al. found the forehead to be more reliable than the sternum for TCB readings during and after phototherapy.^[26]

CONCLUSION

The study findings indicate a strong correlation between TCB and TSB levels before (r = 0.9355), during (r = 0.9250) and after phototherapy (r = 0.9305), with statistically significant associations (P < 0.0001). However, TCB values tended to slightly underestimate serum bilirubin during and after phototherapy.

The Bland-Altman analysis supported these findings by demonstrating good agreement between the two methods before phototherapy, and moderate agreement during and after therapy. These results underscore the utility of TCB as a non-invasive monitoring tool, while also highlighting the need for careful interpretation of TCB values, particularly during and after phototherapy, to ensure clinical assessment of NH.