Influence of bilirubin photoisomers on unbound bilirubin measurement in clinical settings

Abstract

Background

Measured unbound bilirubin concentration is influenced by bilirubin photoisomers. Bilirubin photoisomers are produced even with only a slight light exposure, and clinical samples are inevitably exposed to light. The objective of the study was to evaluate the influence of bilirubin photoisomers on the measurement of unbound bilirubin using serum of jaundiced neonates during blue light phototherapy.

Methods

Five neonates treated with phototherapy for hyperbilirubinaemia were enrolled. The samples were taken 12 h after initiation of phototherapy. Samples were processed by irradiation with blue light, by indoor ceiling light, by both blue light and indoor ceiling light or shaded. Bilirubin subfractions, total bilirubin and unbound bilirubin were measured.

Results

Compared with the non-irradiated samples, the (EZ)-cyclobilirubin concentration and (ZE)-bilirubin/(ZZ)-bilirubin ratio significantly increased in the blue light-irradiated samples, the (ZE)-bilirubin/(ZZ)-bilirubin ratio significantly increased in the indoor ceiling light-irradiated samples, and the (EZ)-cyclobilirubin, (EZ)-bilirubin and (ZE)-bilirubin/(ZZ)-bilirubin ratio significantly increased in the samples irradiated with both lights. No change was noted in unbound bilirubin in any group.

Conclusions

We consider that changes in bilirubin photoisomers induced by light exposure during clinical practice do not influence the measured unbound bilirubin concentration.

Introduction

Unbound bilirubin is unconjugated bilirubin which does not bind to human serum albumin (HSA) or other proteins, and is a useful marker in addition to total bilirubin for evaluating bilirubin neurotoxicity. ^1,2 Unbound bilirubin is measured automatically by glucose oxidase and peroxidase with dilution in a UB-Analyzer (UA-2; Arrows Co Ltd, Osaka, Japan), which is widely used in Japan among other places. The equilibrium unbound bilirubin measured by this method was lower ^3,4 than unbound bilirubin measured by the two-fold sample dilution method, but it significantly correlated with it. ⁴ Also, unbound bilirubin measured by this method is influenced by bilirubin photoisomers. ^5–7 In spite of these limitations, unbound bilirubin measured by this method has correlated with various indicators of bilirubin toxicity. ^1,2,8–11

The serum (EZ)- and (EE)-cyclobilirubin concentrations in neonatal hyperbilirubinaemia were very low even during phototherapy, but the unbound bilirubin increased when (EZ)- and (EE)-cyclobilirubin increased considerably in our previous studies. ^5,6

(ZE)-bilirubin is a photoisomer abundantly present in serum exposed to indoor ceiling light. According to the photochemical characteristic of (ZE)-bilirubin, (ZE)-bilirubin attains photoequilibrium with (ZZ)-bilirubin during short-time light exposure. ¹² The (ZE)-bilirubin/(ZZ)-bilirubin ratio is determined by the type of light source, but it is different between bilirubin irradiated with the same light source in blood and serum. ¹³ Itoh et al. ⁵ reported that (ZE)-bilirubin IXα and (EZ)- and (EE)-cyclobilirubin IXα did not influence the unbound bilirubin measurement within the clinically observed range. McDonagh et al. ⁷ reported that short-time light irradiation of 427.5 μmol/L of (ZZ)-bilirubin IXα combined with HSA and adult serum created (ZE)-bilirubin in vitro, and the influence of (ZE)-bilirubin on the unbound bilirubin measured by the UB-Analyzer (UA-1; Arrows Co Ltd) was considered. As a result, a 0 to 10–25% rise of (ZE)-bilirubin was accompanied by a remarkable decline of unbound bilirubin: 136.8–109.4 nmol/L in bilirubin/HSA solution and 35.9–22.2 nmol/L in bilirubin/human adult serum solution.

In clinical practice, light exposure of samples from a jaundiced neonate occurs unintentionally since the samples are usually obtained and analysed in a lighted room. They will contain varying degrees of bilirubin photoisomers; certainly not zero.

This study was carried out in order to evaluate the influence of bilirubin photoisomers on unbound bilirubin measurement under the various photoisomers condition. Clinical samples were used to evaluate the degree of influences on measurement values in a state as close as possible to the clinical settings.

Materials and methods

Five neonates with hyperbilirubinaemia treated with phototherapy were enrolled in this study. They were born in Handa Hospital between February and April 2009. Their mean gestational age was 40 weeks (range 39.4–40.7 weeks), mean birth weight was 3274 g (range 2864–3762 g) and median 5-min Apgar score was 10 (range 8–10). They did not have haemolytic anaemia, obstructive jaundice or other anomalies.

We obtained informed consent from all their parents. They received blue light phototherapy using six fluorescent blue lights (FL20SBW-NU; Atom Medical Co., Tokyo, Japan) with a light intensity of 9.25 μW/cm²/nm (Minolta Fluoro-Lite Meter 451; Minolta Air-Shields, Osaka, Japan) and 48.6 μW/cm²/nm (Atom Phototherapy Analyzer) at the skin surface.

We collected 0.5 mL blood from patients for this study at 12 h when performing routine total bilirubin measurement after starting phototherapy. For sample collection, all indoor lighting was turned off. Serum was separated by centrifugation (1000 rpm for 10 min) under dark conditions and was divided in half into tubes protected by aluminum foil. One sample was irradiated by the same blue light for one minute with a light intensity of 9.25 μW/cm²/nm (Minolta Fluoro-Lite Meter 451) and 48.6 μW/cm²/nm (Atom Phototherapy Analyzer) at the same location as where the patient was treated and the other sample was not processed. Both samples of serum were frozen at −20°C until the next process. Each sample was divided in half again and one of each pair of samples was irradiated by indoor ceiling light (FLR40S-EXN/M/36-A; Hitachi, Tokyo, Japan) with a light intensity of 0.213 μW/cm²/nm (Minolta Fluoro-Lite Meter 451) and 0.5 μW/cm²/nm (Atom Phototherapy Analyzer) for 10 min. The others were kept in the dark. There were ultimately four processed samples: non-irradiated, blue light-irradiated, indoor ceiling light-irradiated and blue light and indoor ceiling light-irradiated (Figure 1).

Figure 1

Sample processing

We analysed total bilirubin and unbound bilirubin with a UB-Analyzer (UA-2; Arrows Co Ltd) ¹⁴ in the dark room under red light. In this method, 20 μL serum was added to 1 mL phosphate buffer containing 1 mg glucose. After recording the absorbance at 460 nm for the calculation of total bilirubin concentration, 25 μL of an enzyme solution containing peroxidase (3.2 U/3 mL) and glucose oxidase (3.2 U/3 mL) in phosphate buffer was added to the mixture. Peroxidase oxidizes unbound bilirubin to HSA to colourless products in the presence of hydrogen peroxide formed by glucose and glucose oxidase. After adding the enzyme solution, the unbound bilirubin concentration was automatically determined from the velocity of decrease in absorbance at 460 nm as described by Cashore et al. ¹⁵ Bilirubin photoisomers and (ZZ)-bilirubin were measured by using high-performance liquid chromatography (HPLC). ¹⁶ A Shimazu LC-3AFG liquid chromatograph (Shimazu Co., Kyoto, Japan) with an SPD-6AV detector and Chromatopac C-R1A was used for HPLC procedures in combination with a radical compression unit (Waters RCM 8 × 10 Module; Waters Corporation, Milford, MA, USA). Separations were performed using a Nova-Pack C18, 5 mm, radical compression cartridge (Waters C18; Waters Corporation). In brief, a gradient elution technique was performed using acetonitrile:0.01 mol/L phosphate buffer pH 5.5:dimethyl formamide (all from Wako, Osaka, Japan) (5:30:65 v/v/v) as the primary eluent and acetonitrile:0.01 mol/L phosphate buffer pH 5.5:dimethyl formamide (20:15:65) as the secondary eluent. Each sample was mixed with an equal volume of dimethyl sulphoxide (Dojindo, Kumamoto, Japan) and acetonitrile, and centrifuged, after which 25 μL of supernatant were injected into an HPLC in dark laboratory.

Statistical analysis

The two-way analysis of variance (ANOVA) was used for comparison among groups. Post hoc paired t-test with Bonferroni correction was used for comparison between any two groups. The level of statistical significance was set as P < 0.05.

Results

The photoisomers, total bilirubin and unbound bilirubin concentration are shown in Table 1 and Figure 2. Significant between-group differences were noted in (EZ)-cyclobilirubin (F 20.23, P < 0.0001, two-way ANOVA), (EZ)-bilirubin (F 4.14, P = 0.031) and (ZE)-bilirubin/(ZZ)-bilirubin ratio (F 40.13, P < 0.0001). No significant differences were noted in total bilirubin or unbound bilirubin among the groups. Compared with the non-irradiated samples, the (EZ)-cyclobilirubin and (ZE)-bilirubin/(ZZ)-bilirubin ratio significantly increased in the blue light-irradiated samples. In the indoor ceiling light-irradiated samples, no change was noted in the bilirubin photoisomers, but the (ZE)-bilirubin/(ZZ)-bilirubin ratio significantly increased. (EZ)-cyclobilirubin, (EZ)-bilirubin and (ZE)-bilirubin/(ZZ)-bilirubin ratio significantly increased in the samples irradiated with both lights (P < 0.05, paired t-test with Bonferroni correction). No significant changes were noted in any item in the samples irradiated with both lights following blue light irradiation.

Figure 2

Comparison of effect of irradiation on (EE)- and (EZ)-cyclobilirubin, the (ZE)-bilirubin/(ZZ)-bilirubin ratio and unbound bilirubin. The bars indicate the mean value. Significant between-group differences were noted in (EE)- and (EZ)-cyclobilirubin (F 20.23, P < 0.0001, two-way analysis of variance) and (ZE)-bilirubin/(ZZ)-bilirubin ratio (F 40.13, P < 0.0001). No significant differences were noted in unbound bilirubin (F 3.27, P = 0.059) among the groups

Table 1

Comparison of effect of different irradiation schemes on bilirubin subfractions, total serum bilirubin and unbound bilirubin

	No irradiation	Blue light irradiation	Indoor ceiling light irradiation	Blue light and indoor ceiling light irradiation	P values
(EE)-cyclobilirubin (μmol/L)	2.87 (0.66)	3.73 (1.13)	3.28 (0.37)	4.21 (1.28)	0.088
	(2.22–3.76)	(2.91–5.64)	(2.74–3.76)	(2.74–6.16)
(EZ)-cyclobilirubin (μmol/L)	5.23 (0.99)	11.56 (2.62)*	5.34 (1.04)†	11.39 (2.44)*	<0.001
	(3.76–6.33)	(8.55–15.56)	(4.10–6.33)	(9.06–14.36)
(ZE)-bilirubin (μmol/L)	47.03 (9.61)	55.85 (5.92)	53.32 (13.52)	53.83 (3.61)	0.12
	(35.74–59.34)	(45.83–60.53)	(36.08–67.55)	(48.56–57.63)
(EZ)-bilirubin (μmol/L)	5.54 (1.23)	7.35 (1.3)	7.42 (1.76)	8.34 (1.41)*	0.031
	(3.93–6.84)	(5.64–9.23)	(4.96–9.23)	(6.50–10.09)
(ZZ)-bilirubin (μmol/L)	196.4 (23.74)	179.7 (3.68)	191 (33.77)	177.2 (4.77)	0.32
	(167.8–231.7)	(175.8–184.3)	(147.9–229.0)	(173.6–185.2)
Total serum bilirubin (μmol/L)	292.8 (18.55)	292.4 (20.87)	280.1 (10.36)	280.4 (8.11)	0.30
	(278.7–324.9)	(277.0–328.3)	(268.5–294.1)	(266.8–287.3)
Unbound bilirubin (nmol/L)	13.00 (2.26)	12.79 (2.17)	12.18 (1.78)	12.31 (1.93)	0.059
	(10.09–16.25)	(10.26–16.07)	(10.09–14.71)	(10.60–15.39)

Total bilirubin and unbound bilirubin was measured with a UB-analyzer (UA-2) (1:50 sample dilution). Values are mean (SD) and range (minimum–maximum). Blue light irradiation to sample for one minute, indoor ceiling light irradiation to sample for 10 minutes. The two-way analysis of variance was used for comparison among groups

*Significant difference from no irradiation

^†Significant difference from blue light irradiation and from both light irradiation (P < 0.05: paired t-test with Bonferroni correction)

Discussion

To avoid bilirubin toxicity, neonates with hyperbilirubinaemia are treated by phototherapy. The mechanism of phototherapy for neonatal hyperbilirubinaemia involves the production and excretion of (ZE)-bilirubin and (EZ)- and (EE)-cyclobilirubin. ¹⁷

The most effective wavelength within 400–520 nm of visible light in vitro for the geometric photoisomerization from (ZZ)-bilirubin to (ZE)-bilirubin is 400–420 nm, and that for the structural photoisomerization from (ZZ)-bilirubin to (EZ)-cyclobilirubin via (EZ)-bilirubin is 500–520 nm. ¹⁸ If using the same light source, the (ZE)-bilirubin/(ZZ)-bilirubin ratio has the same value. The (ZE)-bilirubin/(ZZ)-bilirubin ratio rises as the light source includes many short wavelengths. ¹² The generation of (EZ)- and (EE)-cyclobilirubin is proportional to the light energy and light irradiation time. ¹⁹

For management of the jaundiced neonate, serum total bilirubin, measurements of serum unbound bilirubin and cutaneous bilirubin are used. A semi-automated instrument for measuring the unbound bilirubin with dilution-peroxidase method is available. A problem with this measurement method depends on the apparent binding affinity of albumin due to sample dilution. ²⁰ In clinical studies, unbound bilirubin determined by this measurement method was correlated with auditory brainstem response (ABR) changes in term infants ^1,2 as well as the probability of bilateral refer automated ABR with jaundiced newborns at more than 34 weeks of gestation. ¹¹ In premature infants, unbound bilirubin correlated with abnormal ABR maturation ⁸ and kernicterus. ⁹ Unbound bilirubin of 17.1 nmol/L carries a risk of kernicterus with sensitivity of 100% and specificity of 98% in low-birth-weight infants, while unbound bilirubin of 13.7 nmol/L has a sensitivity of 100% and specificity of 96% in very low-birth-weight infants. ¹⁰

For evaluation of these factors, it is necessary to accurately measure unbound bilirubin. Haemoglobin, ¹⁴ vitamin C, conjugated bilirubin, ⁶ bilirubin photoisomers ^5–7 and sample dilution ^3,4,21 influence the measurement of the unbound bilirubin by the peroxidase method. The most serious error is the failure to appreciate that the unbound bilirubin determined by the peroxidase test is steady-state unbound bilirubin. ²² Bilirubin binds in the large hydrophobic cavity of sub-domain IIA on HSA. ²³ The more bilirubin that binds to the first class of bilirubin sites in the HSA molecule, the more readily bilirubin reaches a condition of photoequilibrium with (ZE)-bilirubin. (ZE)-Bilirubin IXα co-crystallized with HSA binds to an L-shaped pocket in sub-domain IB. ²⁴ When (ZE)-bilirubin increases, the unbound bilirubin decreases. ⁷ It is believed that (ZZ)-bilirubin changes to (ZE)-bilirubin, and when (ZE)-bilirubin moves to other binding sites on HSA, the unbound bilirubin migrates to the vacated site. However, the proper binding site for (ZZ)-bilirubin on HSA remains unknown.

Since Itoh et al. ⁵ already reported the influence of light irradiation-induced changes in photoisomers on the measured unbound bilirubin concentration in samples before phototherapy, we investigated the influence of light exposure-induced changes in photoisomers on unbound bilirubin, in serum separated during phototherapy, which may occur in clinical practice. The reasons for the use of clinical samples are as follows: during blue light irradiation, the main photoisomer in the in vivo neonate serum is (ZE)-bilirubin, and the major photoisomers in the in vitro bilirubin/HSA complex are (ZE)-bilirubin and (EZ)- and (EE)-cyclobilirubin. Another reason is, as shown by this study, the (ZE)-bilirubin/(ZZ)-bilirubin ratio was increased by irradiation with the same light source in samples from patients treated with phototherapy. This may have been due to attenuation of some wavelengths by substances in the body. Unlike samples simply irradiated with the light source, reproduction of this condition in vitro is difficult. The influence of environmental light exposure on bilirubin photoisomers in samples after serum separation may be clinically problematic. The measured unbound bilirubin concentration did not alter with changes in bilirubin photoisomers induced by various combinations of light exposure.

We conclude that major changes in the proportions of the different photoisomers produced by the three different irradiation schemes produced no alterations in measurement of unbound bilirubin with the UB-analyser (UA-2); unbound bilirubin values that are unchanged are unlikely to be clinically significant, but this needs to be confirmed by prospective studies of outcomes with different unbound bilirubin concentrations measured with the UB-analyser.

DECLARATIONS

Competing interests: None.

Funding: None.

Ethical approval: Informed consent was obtained from the mothers of all patients.

Guarantor: SI.

Contributorship: KK and KK obtained samples. KK and KK obtained informed consent from parents. TK and JK processed samples. SY carried out data analysis. KI and SI carried out the protocol. TI measured samples. HO wrote the manuscript. All authors reviewed and edited the manuscript and approved the final version of the manuscript.

Acknowledgements: None.