-
Acta Pdiatrica ISSN 08035253
REGULAR ARTICLE
Therapeutic effect of turquoise versus blue light with equal
irradiance inpreterm infants with jaundiceFinn Ebbesen
([email protected])1, Poul Madsen2, Sren Stvring1, Heidi Hundborg3,
Giovanni Agati4
1.Department of Pediatrics, University Hospital of Aalborg,
Aalborg, Denmark2.Department of Clinical Biochemistry, University
Hospital of Aalborg, Aalborg, Denmark3.Department of Clinical
Epidemiology, University of Aarhus, Aarhus, Denmark4.Institute of
Applied Physics-CNR, Florence, Italy
KeywordsNewborn, Jaundice, Phototherapy, Blue light,Turquoise
light
CorrespondenceFinn Ebbesen, Department of Pediatrics,Aalborg
University Hospital, DK 9000 Aalborg,Denmark.Tel: +4599321310 |
Fax: +4599321895 |Email: [email protected]
Received8 February 2006; accepted 30 January 2007.
DOI:10.1111/j.1651-2227.2007.00261.x
AbstractAim: To compare the efficiency of turquoise light with
that of TL52 blue in treatment of preterm
infants with jaundice at the same level of body irradiance.
Methods: Infants with gestational age 2837 weeks and
non-haemolytic hyperbilirubinemia were
treated for 24 h with either turquoise light (OSRAM L18W/860
fluorescent lamps) or blue light
(Philips TL20W/52 fluorescent lamps). The concentrations of
serum total bilirubin and bilirubin
isomers were measured by the Vitros routine method and by HPLC,
respectively.
Results: The decrease in serum concentrations of total
bilirubin, total bilirubin isomers and the toxic
Z,Z-bilirubin was greatest for infants treated with turquoise
light. Further, the increase in Z,E-bilirubin
was smaller and there was a trend towards a higher rise in
E,Z-bilirubin.
Conclusions: Turquoise light has a greater bilirubin reducing
effect than TL52 blue light with equal irradiance,
expressed both by serum total bilirubin, total bilirubin isomers
and Z,Z-bilirubin, i.e. the turquoise spectral range
is more efficient than the blue. This is in accordance with
deeper penetration into the skin, lower production of
the Z,E-bilirubin and greater production of E,Z-bilirubin and
lumirubin, in infants under turquoise light. This
suggests, given equal irradiances, that light in the turquoise
spectral range is preferable to the TL52 blue in
treatment of newborn jaundiced infants.
INTRODUCTIONPhototherapy is the most widespread treatment for
lower-ing the bilirubin concentration in neonates. Light
absorptionin the skin converts the toxic Z,Z-bilirubin molecules
intomore easily excretable compounds (photoisomers) (1):
theconfigurational isomers Z,E-bilirubin and E,Z-bilirubin, andthe
structural isomers Z-lumirubin and E-lumirubin (Fig. 1).The
Z,Z-bilirubin Z,E-bilirubin process is reversible andquite
efficient. However, Z,E-bilirubin is very slowly elimi-nated by the
liver and therefore accumulates in plasma (2).Configurational
photoisomerization also produces the E,Z-bilirubin isomer, which is
a precursor for the lumirubin.The Z,Z-bilirubin lumirubin process
is irreversible andless efficient than the Z,E-bilirubin production
(3), but lu-mirubin (throughout the article, the term lumirubin
includesboth Z- and E-lumirubin) is quickly eliminated by the
liver(4). Thus, the excretion of lumirubin is considered
quantita-tively to be more important during phototherapy than
excre-tion of Z,E-bilirubin (4). The photoisomers are
presumablyless toxic than Z,Z-bilirubin (5). Besides
photoisomerizationZ,Z-bilirubin undergoes photooxidation with a
very smallyield (6).
Blue fluorescent lamps with emission peak wavelengthmatching the
absorption maximum of the plasma bilirubin-albumin complex (at 462
nm) have been used worldwide,in particular Philips TL20W/52 lamp
with peak emission at452 nm and bandwidth of 55 nm (Fig. 2), but in
photother-
Figure 1 Photoisomerization processes of bilirubin:
configurational isomeriza-tion of Z,Z-bilirubin produces
Z,E-bilirubin and E,Z-bilirubin. Formation of thestructural
photoisomer Z-lumirubin proceeds via cyclization either directly
fromZ,Z-bilirubin or with E,Z-bilirubin as intermediate product.
Configurational iso-merization of Z-lumirubin gives
E-lumirubin.
apy of jaundiced neonates, the longer wavelength
turquoisespectral band was expected to be more efficient than the
bluespectral band in the clearance of bilirubin (7): (1)
turquoiselight penetrates deeper into the skin than blue light; (2)
thelight-induced yield of lumirubin from Z,Z-bilirubin is greaterat
longer than shorter wavelengths, while (3) the oppositeoccurs for
the Z,Z-bilirubin Z,E-bilirubin process. Points(2) and (3) are
expected to produce in vivo higher concentra-tion of lumirubin and
lower concentration of the competingZ,E-bilirubin. Furthermore,
presumably (4) the absorbanceof bilirubin in vivo in the
long-wavelength range is greaterthan that in vitro (8,9).
On the basis of these effects, an increase of photother-apy
efficiency was predicted for a lamp emission spectrumpeaked at
about 490 nm (7). This spectral region was clin-ically tested with
successful results by using non-standard
C2007 The Author(s)/Journal Compilation C2007 Foundation Acta
Pdiatrica/Acta Pdiatrica 2007 96, pp. 837841 837
-
Effect of turquoise light Ebbesen et al.
50
40
30
20
10
0
irr
adia
nce
(W
/cm
2 /nm
)
650600550500450400wavelength (nm)
1.2
1.0
0.8
0.6
0.4
0.2
0.0
Z,Z-
bilir
ubin
/HSA
ab
sorp
tion
(a.u.)
Turquoise OSRAM Blue Philips/52 Z,Z-bilirubin/HSA
Figure 2 Emission spectra of the turquoise (OSRAM L18W/860)
(solid line)and blue (Philips TL20W/52) (dashed curve) flourescent
lamps measured atthe distance of 32 and 41 cm, respectively, to
have the same total irradiance.Narrow peaks superimposed on the
broad fluorescence bands are due to themercury emission lines (at
413, 436, 546 and 577 nm). Dashed-dotted curverepresents the
absorption spectrum of Z,Z-bilirubin. a.u.: arbitrary units.
Humanserum albumin (HSA).
cold-cathode fluorescent lamps at 490 nm (10). There-after,
turquoise standard fluorescent lamps (L18W/860)with emission peak
at 490 nm and bandwidth of 65 nm weredeveloped by OSRAM (Fig.
2).
We compared the effectiveness of these turquoise lampswith
Philips TL20W/52 blue lamps in terms of reducing theplasma total
bilirubin concentration in preterm infants (11).The distance from
the lamps to the surface of the infants wasidentical for the two
lamps. There was no statistically sig-nificant difference in the
percentage reduction in the serumtotal bilirubin concentration
between the two lamps; but thelight irradiance for the blue lamps
was 30% higher than thatfor the turquoise lamps. Thus, it was
concluded, that (1) thephototherapeutic efficiency of the turquoise
lamps was equalto that of the blue lamps, and that (2) the
efficiency of theturquoise light is equal to or higher than that of
the bluelight.
In the present study, we performed a more accurate com-parison
between the efficiency of turquoise and TL-52 bluelight in reducing
the serum bilirubin concentration. Sinceboth the irradiance and the
colour (spectral range) of theemitted light affect the
effectiveness of the treatment, to com-pare the turquoise and blue
spectral emission regions weneeded to keep the irradiance on the
infants at the same levelfor the two lamps. Furthermore, we
quantified the serumconcentrations of the invidual bilirubin
isomers before andafter 24 h of phototherapy with the two types of
light, toobtain a more comprehensive evaluation of the efficiency
ofthe two treatments.
MATERIAL AND METHODSThe study period was March 1st 2003 to
September 30th2004. The inclusion criteria were preterm infants
with agestational age 196258 days, a postnatal age >24 h,
non-haemolytic hyperbilirubinemia and no previous photother-
apy. The infants participated in the study for 24 h, and
wereincluded consecutively.
The indications for phototherapy followed the guidelinesof the
Danish Paediatric Society (12). Phototherapy was ad-ministered
continuously, except during feeding, nursing careand blood
sampling. The number of hours of photother-apy was measured for
each infant. The infants were treatednaked except for diapers
(Pampers, 6 1421 cm) andeye pads. They were treated either in an
incubator or in abassinet. The average distance from the
phototherapy ap-paratus to the surface of the infants was 32 cm for
infantstreated with turquoise light (normal distance) and 41 cm
forinfants treated with TL52 blue light, so that the average
lightirradiance over the whole spectrum would be the same
forinfants treated with turquoise and blue light.
The infants were randomised by sealed envelopes to oneof two
phototherapy regimens: either eight of the turquoisefluorescent
lamps (18 Watt, 60 2.6 cm), or eight of thePhilips TL20W/52 blue
fluorescent lamps (20 Watt, 60 3.7 cm). The geometries of the two
light systems were iden-tical. The lamps were replaced after 1000 h
of operation.In Figure 2, the calibrated emission spectra of the
turquoiseand TL52 blue lamps are shown, along with the
absorptionspectrum of Z,Z-bilirubin (20 mol/L), in 0.1 mol/L
sodiumphosphate buffer, pH 7.4, containing human serum albumin(40
mol/L), normalised to its maximum value.
The light irradiance of the two phototherapy units wasmeasured
three times during the 24-h study period on the in-fants front or
back. The light irradiance was measured with abroadband (380780 nm)
photodiode power meter (EG&G,mod. 460, Salem, MA, USA),
previously calibrated at thetwo emission bands, 452 and 490 nm,
against a thermopile(Nova power meter, 2A thermal head, Ophir
Optronics Ltd.,Jerusalem, Israel). The spectral irradiance was
measured bya calibrated SpectraScan PR-704 spectroradiometer
(PhotoResearch, Catsworth, CA, USA).
The serum concentrations of total bilirubin and biliru-bin
isomers were measured on capillary blood drawn by aheel prick 0 and
24 h after start of phototherapy. The bloodwas drawn into tubes
wrapped with tin foil to avoid lightirradiation.
The serum total bilirubin concentration was measured onthe
Vitros 950 analyser [Vitros 950, Johnson & Johnson,Rochester,
NY, USA, using the Vitros BuBc slide (13)]. Dur-ing the study
period, the imprecision expressed as coefficientof variation was
1.9% at a total bilirubin level of 225 mol/L.The Vitros method is
our routine method.
The serum concentrations of the bilirubin isomers wereanalysed
twice by isocratic reversed phase high performanceliquid
chromatography (HPLC) (8).
The imprecision expressed as coefficient of variation was2.7%
for the serum concentration of total bilirubin isomers,2.7% for
Z,Z-bilirubin, 4.2% for Z,E-bilirubin, 6.4% for E,Z-bilirubin, 7.1%
for Z-lumirubin and 12% for E-lumirubin.
The serum concentration of conjugated bilirubin is in-corporated
in the measurement of total bilirubin by Vitrosmethod, but was not
measured by the HPLC method.The concentration of conjugated
bilirubin in plasma, in
838 C2007 The Author(s)/Journal Compilation C2007 Foundation
Acta Pdiatrica/Acta Pdiatrica 2007 96, pp. 837841
-
Ebbesen et al. Effect of turquoise light
Table 1 Demographic and clinical characteristics of the
patients
Turquoise light TL52 blue light
Number of infants 72 69Gestational age (days) (mean SD) 234 17
237 18Birth weight (g) (mean SD) 2061 579 2095 635Postnatal age (h)
(mean SD) 74 30 74 34Gender female/male 29/43 32/37Respiratory
distress syndrome (number) 13 12Hypoglycaemia (blood glucose 14
15
< 1.8 mm/L) (number)Septicaemia (number) 3 4Asphyxia (Apgar
score 5 min < 8) (number) 2 0Incubator/bassinet (number) 39/33
42/27Percentage of study period in phototherapy 0.85 0.05 0.85
0.06
(mean SD)Serum total bilirubin conc. before 221 60 221 61
phototherapy (mol/L) (mean SD)
infants with uncomplicated hyperbilirubinaemia is
negligible(14).
The study was approved by the ethics committee for NorthJutland
County. Both verbal and written informed consentwas obtained from
the parents.
The statistical analyses were performed using t-test onStata
version 8.0.
RESULTSThe demographic and clinical characteristics of the
infantstreated with turquoise light (n = 72) were compared
withthose of the infants treated with TL52 blue light (n = 69)
inTable 1. Two infants did not complete the study, one fromeach
group, either due to withdrawal of all treatment ordeath.
The spectral distribution of total irradiance is reportedin
Figure 2. For the turquoise light the average irradiancewas 7.43
W/cm2/nm over whole spectrum (380780 nm),14.6 W/cm2/nm in the
420480 nm and 29.9 W/cm2/nmin the 460520 nm range. In the same
ranges for theTL-52 blue light, it was 7.33 W/cm2/nm, 37.5
W/cm2/nmand 17.4 W/cm2/nm, respectively. The difference of0.11
W/cm2/nm in the whole spectrum is not statisti-cally significant
[95% confidence interval (95% CI): 0.55;0.33] p = 0.59 by
t-test.
The decrease in serum total bilirubin concentration in in-fants
treated 24 h with turquoise light, average 92 mol/L(17 mol/L = 1
mg%), was statistically significantly greaterthan that of infants
treated with TL-52 blue light, average78 mol/L (Table 2).
In 127 infants (90%), measurement of the bilirubin iso-mers was
performed. In the remaining infants, blood for theHPLC measurement
had not been taken.
The serum concentrations of total bilirubin isomers andthe
individual bilirubin isomers, before phototherapy, areshown in
Table 3. Both in infants treated with turquoise andTL52 blue light,
there was a decrease in the serum values oftotal bilirubin isomers
and Z,Z-bilirubin, and an increase in
Table 2 Decrease in serum total bilirubin concentration,
measured by Vitrosmethod, in infants treated with turquoise and
TL52 blue light
Turquoise light TL52 blue light Difference p Value(n = 72) (n =
69) meanmean SD mean SD (95 % CI)
Decrease in serum total 92 31 78 34 15 (4; 25) 0.008bilirubin
conc. (mol/L)
Statistical analysis: t-test.
Table 3 Serum concentrations of total bilirubin isomers and
individual isomers,measured by HPLC method, before phototherapy
with turquoise and TL52 bluelight
Serum conc. (mol/L) Turquoise light TL52 blue light(25%;75%
percentiles)] (n = 65) (n = 62)
Total bilirubin isomers 218.8 (174.3; 259.9) 201.4 (160.9;
270.3)Z,Z-bilirubin 197.8 (163.8; 244.6) 182.8 (145.4;
251.4)Z,E-bilirubin 12.8 (9.1; 18.7) 15.8 (11.5; 19.9)E,Z-bilirubin
1.4 (1.0; 1.8) 1.5 (1.0; 2.2)Z-lumirubin 0.5 (0.0; 1.0) 0.5 (0.0;
1.0)E-lumirubin 0.2 (0.0; 0.3) 0.2 (0.0; 0.3)
Z,E-bilirubin, E,Z-bilirubin, Z-lumirubin and E-lumirubin.In
Table 4, the changes in the concentrations of these iso-mers are
compared for infants treated with the two lightregimens. The
average decrease in total bilirubin isomers forinfants treated with
turquoise light, 82.3 mol/L, was sta-tistically significantly
greater than that for infants treatedwith blue light, 62.1 mol/L.
Further, the average decreasein Z,Z-bilirubin during turquoise
light, 95.9 mol/L, wasstatistically significantly greater than
during TL52 blue light,81.6 mol/L. In infants treated with
turquoise light, the in-crease in Z,E-bilirubin was statistically
significantly less thanthat for infants treated with blue light,
and there was a trendtowards the higher rise in E,Z-bilirubin. As
regards the in-crease in the Z- and E-lumirubin, the two groups did
notdiffer.
No side effects were observed with the exception of loosegreen
stools.
DISCUSSIONWe found that treatment of preterm jaudiced infants
withturquoise light, compared with TL52 blue with the same
ir-radiance, gave a greater reduction in the serum concentra-tions
of total bilirubin measured by Vitros method and to-tal bilirubin
isomers measured by HPLC. The decrease wasabout 25% greater for the
turquoise light than that for theblue light, which is suggested to
be of clinical importance.Concerning the concentration of
Z,Z-bilirubin, the decreasewas greater when treating with turquoise
light than that withblue light.
According to the overlapping of the lamp emissions withthe in
vitro Z,Z-bilirubin absorption spectrum (Fig. 2), wewould not
predict a greater efficiency of the turquoise lightthan the TL52
blue in reducing the serum bilirubin levels.
C2007 The Author(s)/Journal Compilation C2007 Foundation Acta
Pdiatrica/Acta Pdiatrica 2007 96, pp. 837841 839
-
Effect of turquoise light Ebbesen et al.
Table 4 Changes in serum concentrations of total bilirubin
isomers and individual isomers, measured by HPLC method, in infants
treated with turquoise light andTL52 blue light
Change in serum conc. (mol/L) Turquoise light (n = 65) TL52 blue
light (n = 62) Difference p Valuemean SD mean SD (mean; 95% CI)
Decrease in total bilirubin isomers 82.3 35.9 62.1 38.1 20.2
(7.3; 33.2) 0.003Decrease in Z,Z-bilirubin 95.8 36.4 81.6 38.2 14.3
(1.2; 27.4) 0.033Increase in Z,E-bilirubin 9.5 10.2 15.7 13.9 6.2
(1.9; 10.5) 0.005Increase in E,Z-bilirubin 0.7 1.2 0.4 0.9 0.3
(0.1: 0.7) 0.121Increase in Z-lumirubin 2.9 2.0 2.9 2.9 0.0 (0.9;
0.9) 0.941Increase in E-lumirubin 0.4 0.5 0.4 0.4 0.0 (0.1; 0.2)
0.966
Statistical analysis: t-test; based on unequal variance.
Therefore, the superior phototherapy efficiency of theturquoise
light indicates that the other factors mentionedin the introduction
(light penetration, photoisomerizationyields, in vivo absorbance)
were also important in determin-ing the efficiency of the light.
Accordingly, the increase inthe serum concentration of
Z,E-bilirubin was lowest for theinfants treated with turquoise
light, and there was a trendtowards a greater increase in
E,Z-bilirubin. We found nodifferences in the increase of the serum
concentrations ofZ- and E-lumirubin between the groups. This may be
dueto the low concentrations and the great variations, makingit
difficult to detect the presence of small differences.
Alsophotooxidation of Z,Z-bilirubin (not measured in the study)may
have influenced the light-induced reduction of
serumZ,Z-bilirubin.
In 10% of the infants, blood sampling for the determi-nation of
bilirubin isomers was not done. This dropout wasrandom between the
two groups, and is not expected to haveany influence on the
interpretation of the results.
Since there were no excluded infants, the present infantsare
representative for the whole population of preterm in-fants with a
gestational age above 28 weeks, i.e. the resultsof the study must
be regarded generalizable.
The strengths of the study are a relatively high numberof
infants, highly significant results, reliable analytical meth-ods,
few dropouts and a great homogeneity in the technicalaccomplishment
(single center study).
Side effects may be less serious with longer
wavelengthphototherapy, regarding cytotoxicity (15), mutagenic
risks(16), as well as photodegradation processes (17).
CONCLUSIONSTurquoise light has a greater bilirubin reducing
effect thanTL52 blue light with equal irradiance, expressed bothby
serum total bilirubin, total bilirubin isomers and Z,Z-bilirubin,
i.e. the turquoise spectral range is more efficientthan the blue.
This is in accordance with deeper penetra-tion into the skin, lower
production of the Z,E-bilirubin andgreater production of
E,Z-bilirubin and lumirubin, in infantsunder turquoise light. This
suggests, given equal irradiances,that light in the turquoise
spectral range is preferable to theTL52 blue in treatment of
newborn jaundiced infants.
ACKNOWLEDGEMENTSHeidi Hundborg performed the majority of the
statisticalanalyses, but due to sickness the analyses were
completedby Henrik Thomsen, Department of Clinical Epidemiol-ogy,
University of Aarhus, Aarhus, Denmark. Gunnar LaugeNielsen is
thanked for review of the manuscript.
References
1. McDonagh AF, Lightner DA. Phototherapy and thephotobiology of
bilirubin. Semin Liver Dis 1988; 8: 27283.
2. Ennever JF, Knox I, Denne SC, Speck WT. Phototherapy
forneonatal jaundice: in vivo clearance of bilirubinphotoproducts.
Pediatr Res 1985; 19: 2058.
3. Agati G, Fusi F. New trends in photobiology (invited
review).Recent advances in bilirubin photophysics. J
PhotochemPhotobiol B:Biol 1990; 7: 114.
4. Ennever JF, Costarino AT, Polin RA, Speck WT. Rapidclearance
of a structural isomer of bilirubin duringphototherapy. J Clin
Invest 1987; 79: 16748.
5. McDonagh AF, Lightner DA. Like a shrivelled blood
orange.Bilirubin, jaundice, and phototherapy. Pediatrics 1985;
75:44355.
6. Lightner DA, Linnane WP, Ahlfors CE. Bilirubinphotooxidation
products in the urine of jaundiced neonatesreceiving phototherapy.
Pediatr Res 1984; 18: 696700.
7. Agati G, Fusi F, Donzelli GP, Pratesi R. Quantum yield
andskin filtering effects on the formation rate of lumirubin.J
Photochem Photobiol B:Biol 1993; 18: 197203.
8. McDonagh AF, Agati G, Fusi F, Pratesi R. Quantum yields
forlaser photocyclization of bilirubin in the presence of
humanserum albumin. Dependence of quantum yield on
excitationwavelength. Photochem Photobiol 1989; 50: 30519.
9. Malhotra V, Greenberg JW, Dunn LL, Ennever JF. Fatty
acidenhancement of the quantum yield for the formation oflumirubin
for bilirubin bound to human albumin. Pediatr Res1987; 21:
5303.
10. Agati G, Fusi F, Pratesi R, Pratesi S, Donzelli GP.
Actionspectrum for neonatal jaundice phototherapy: prediction
andclinical evidence. Proceedings of the CIE Warsaw Session.CIE
1331999 1999; 1: 3324.
11. Ebbesen F, Agati G, Pratesi R. Phototherapy with
turquoiseversus blue light in preterm infants with jaundice. Arch
DisChild Fetal Neonatal Ed 2003; 88: F 4301.
12. http://aasint12.aas.its.nja.dk13. Doumas BT, Kwok-Cheung PP,
Perry BW et al. Candidate
reference method for determination of total bilirubin in
serum:Development and validation. Clin Chem 1985; 31: 177989.
840 C2007 The Author(s)/Journal Compilation C2007 Foundation
Acta Pdiatrica/Acta Pdiatrica 2007 96, pp. 837841
-
Ebbesen et al. Effect of turquoise light
14. Brodersen R, Jacobsen J. Serum bilirubin diglucuronide in
thehuman adult and the newborn child. In: Bouchier IAD, BillingBH,
eds. Bilirubin metabolism. Oxford: Blackwell ScientificPress 1967;
1115.
15. Roll EB, Christensen T. Formation of photoproducts
andcytotoxicity from bilirubin irradiated with turquoise and
bluelight phototherapy. Acta Paediatrica 2005; 94: 11922.
16. Bradley MO, Sharkey NA. Mutagenicity and toxicity of
visiblefluorescent light to cultured mammalian cells. Nature 1977;
2:7246.
17. Gromisch DS, Lopez R, Cole HS. Cooperman JM.
Light(phototherapy)-induced riboflavin deficiency in the neonate.J
Pediatr 1977; 90: 11822.
C2007 The Author(s)/Journal Compilation C2007 Foundation Acta
Pdiatrica/Acta Pdiatrica 2007 96, pp. 837841 841
