Letters to the Editor

Indian Pediatrics 2000;37: 1404-1407

Non-Invasive Estimation of Total Serum Bilirubin

Serial estimation of serum bilirubin levels is the hallmark of management of neonatal hyperbilirubinemia, which however requires repeated venepunctures/heel pricks for draw-ing blood, that has several hazards in sick and preterm neonates(1). Pediatricians have always wished to have some non-invasive mechanism of surveillence and inter-vention in jaundiced neonates. Since 1980, when Yamanouchi et al.(2) first reported Trans-cutaneous Bilirubinometry (TcBm) as a non-invasive screening device for monitoring neonatal bilirubinemia, several authors all over the world including from India have evaluated the transcutaneous bilirubinometer in preterm and term, low birth weight and intro-uterine growth retarded babies and even after phototherapy and consistently showed high degree of linear correlation between total serum bilirubin (TSB) and transcuta-neous bilirubin index (TcBI). The TSB was measured by standard Diazo method in almost all previous studies. However, there was a wide difference between TSB level and TcBI. The TSB level was not expected to be the same as TcBI because transcutaneous bilirubinometer employs a technique where scattered light from the photoprobe is divided and measured as the difference between optical densities of blue and green light and translated into an arbitrary digital unit referred to as the TcBI(2). The light is scattered from the tissue bilirubin, melanin, oxyhemoglobin, fat, connective tissue and vasculature. Hence, no non-invasive technique of transcutaneous bilirubin estimation will reproduce the serum bilirubin levels as obtained by Diazo method or microbilirubin analysis. It was, therefore, recommended to prepare nomograms for a particular geographic area and use them as screening charts, thus reducing the need for vene-punctures.

We read with interest the results of TcBm by BilicheckTM as reported by Lodha et al.(3). This instrument is claimed to be of a superior technique since it subtracts the spectral contribution of other factors except tissue bilirubin. The authors found a good linear correlation between TSB and TcBI using Minolta joundice meter as has been reported previously(4,5). The authors also found that there was poor agreement between TSB and TcBI values. This was not surprising because this non-invasive technique of billirubin estimation also depends upon light reflectance which can not eliminate spectral reflectance from all cutaneous factors, some known, some unknown. It reflects TcBI values which comprise billirubin and cutaneous tissues like fat, a connective tissue, vasculature, neural tissue, etc. Hence, ‘TSB’ can not be estimated by BilicheckTM transcutaneous bilirubino-meter. However, it appears that the differ-ence between TSB and TcBI is lesser by BilicheckTM as compared to that reported by Minolta Jaundice Meter. The clinical applica-tion of BilicheckTM remains the same as other bilirubinometers. However, multi wave spec-tral reflectance (MWSR) bilirubinometer will be better choice than other transcutaneous bilirubinometer.

We would also like to comment that while comparing the TSB and TcBI as measured by other bilirubinometers, the authors(3) have inadvertently given an impression that whereas in western studies there was good agreement between TSB and TcBI, Indian authors did not find the same degree of accuracy. This statement is not correct. All the authors from India and abroad have found that TSB and TcBI are not same, have wide difference but have very high linear correlation (r = 0.933)(2), (r = 0.999)(4). In fact linear correlation between TSB and TcBI is maintained even after phototherapy(4,5).

Thus, we strongly feel that TcBm can be effectively used in the NICU and at a primary health care level if MWSR bilirubinometer and conventional jaundice meter specific nomograms for TcBI at different serum bili-rubin levels, both before and after photo-therapy can be prepared. The TcBm can be done as frequently as desired and may substantially replace the conventional method of bilirubin estimation in jaundiced newborn infants.

M.M.A. Faridi,
Deepa Gupta,
University College of Medical Sciences and
Guru Teg Bahadur Hospital,
Shahdara, Delhi 110 095, India.
Fax: 91-11-2290495

1. Gupta P, Sharma KN. Neonatal pain. Indian Pediatr 1995; 30: 1263-1269.

2. Yamanouchi I, Yamachi Y, Igarshi I. Transcutaneous bilirubinometry: Preliminary studies of non-invasive transcutaneous bili-rubinometer in the Okayama National Hospi-tal. Pediatrics 1980; 65: 195-202.

3. Lodha R, Deorari AK, Jatana V, Paul VK. Non-invasive estimation on total serum bili-rubin by multi length spectral reflectance in neonates. Indian Pediatr 2000; 37: 771-775.

4. Kumar A, Faridi MMA, Singh N, Ahmed SH. Transcutaneous bilirubinometry in the management of bilirubinemia in term neonates. Indian J Med Res 1994; 99: 227-250.

5. Harish R, Sharma DB. Transcutaneous bilirubinometry in neonates: Evaluation of Minolta Air Shields Jaundice Meter. Indian Pediatr 1998; 35: 264-267.

We agree with Drs. Faridi and Gupta’s comments that transcutaneous bilirubinometer (TcBm) may not accurately reflect serum bilirubin levels. The level of jaundice (yellowness of skin) does depend on the total serum bilirubin (TSB), but this is not the sole factor. However, the need to reduce invasive sampling for TSB estimation has prompted investigators into the use of yellowness of skin as a surrogate marker of TSB.

The authors contradict themselves in this letter. They initially state that ‘no non-invasive technique of transcutaneous bilirubin estimation will reproduce the serum bilirubin level as obtained by diazo method or microbilirubin analysis’. However, later, they ‘feel strongly that TcBm can be effectively used in the NICU and at primary health care level if specific nomograms for transcuta-neous bilirubin index (TcBI) at different serum bilirubin can be prepared’. They them-selves had suggested creation of nomograms way back in 1994(1). But it looks impossible and unethical as this will involve estimating total serum bilirubin in normal newborns prospectively. A predictor model for develop-ment of pathological jaundice by prospective cross sectional serum bilirubin estimation has been reported elsewhere(2).

One of the messages of our paper was to emphasize that the use of correlation (r value) as a sole marker of agreement may not be good enough for practical purposes, as suggested earlier(3). While the correlation may be good, the difference between the two values may be too large to be acceptable clini-cally (see 95% CI reported in our article)(4). In their own study, Faridi et al.found a good linear correlation (r = 0.909 and not 0.999 as mentioned in the letter) between serum bilirubin and TcBI(1). But a close look at the graph suggests that for a given serum bilirubin value, the TcBI could vary as much as 10 units. This difference could pose a difficult clinical problem. Other Indian studies reported correlation coefficient of 0.65 to 0.83(5,6).

For the transcutaneous bilirubino-meters to be accurate and reliable enough for use in decision making, the agreement with TSB has to be much better. At present it appears that there cannot be a simple formula to predict TSB from TcB. We, therefore, maintain that the utility of transcutaneous bilirubinometers is only for screening. At present these devices cannot be a substitute for actual TSB determination.

R. Lodha,
A.K. Deorari,
V.K. Paul,
Neonatal Division,
Department of Pediatrics,
All India Institute of Medical Sciences,
New Delhi 110 029, India.

1. Kumar A, Faridi MMA, Singh N, Ahmad SH. Transcutaneous bilirubinometry in the management of bilirubinemia is term neo-nates. Indian J Med Res 1994; 99: 227-230.

2. Bhutani VK, Johnson L, Sivieri MS. Predictive ability of a predischarge hour-specific serum bilirubin for subsequent significant hyperbilirubinemia in healthy term and near term neonates. Pediatrics 1999; 103: 6-14.

3. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1: 307-310.

4. Lodha R, Deorari AK, Jatana V, Paul VK. Non-invasive estimation of total serum bili-rubin by multi-wavelength spectral reflect-ance in neonates. Indian Pediatr 2000; 37: 771-775.

5. Harish R, Sharma DB. Transcutaneous bilirubinometry in neonates: Evaluation of Minolta Air Shields Jaundicemeter. Indian Pediatr 1998; 35: 264-267.

6. Bhat V, Srinivasan S, Usha TS, Puri RK. Correlation of transcutaneous bilirubinometry with serum bilirubin in South Indian neonates. Indian J Med Res 1987; 86: 49-52.