Monica Sarin, Sourabh Dutta and Anil Narang
From the Division of Neonatology, Department of
Pediatrics, Postgraduate Institute of Medical Education and Research,
Chandigarh 160 012, India.
Correspondence to: Dr. Sourabh Dutta, Assistant
Professor, Department of Pediatrics, PGIMER, Sector 12, Chandigarh 160
012, India. E-mail:
Manuscript received: September 2, 2004, Initial review
completed: November 30, 2004,
Revision accepted: January 18, 2006.
Background: Special blue tube lights of standard
length are used in most neonatal units to deliver phototherapy. Of late,
special blue compact fluorescent lamp phototherapy equipments have been
introduced in India, which are claimed to be better than standard tube
lights. Aim: To compare special blue compact fluorescent lamp (CFL)
phototherapy with special blue standard-length tube lights (STL).
Methods: This randomized, controlled trial was conducted in a level
III NICU. Neonates, otherwise healthy, of gestation >34 weeks with
hyperbilirubinemia requiring phototherapy, were included. Rh iso-immunized
babies, those who underwent prior exchange transfusion and whose parents
declined to consent were excluded. By stratified block randomization,
babies were allocated to receive phototherapy by CFL or STL. CFL and STL
were both special blue lights with irradiance maintained above 15
µW/nm/cm2. Total serum bilirubin (TSB) was measured 12 hourly till
phototherapy was stopped or an exchange transfusion was done.
Temperature and clinical and laboratory parameters of dehydration were
recorded 12 hourly till 72 hrs. Nursing staff answered an objectivized
proforma about the disadvantageous effects on nurses. Results:
Fifty babies were enrolled in each group. Baseline characteristics,
causes of jaundice, hemolysis, baseline TSB and irradiance were similar
in both groups. Mean duration of phototherapy (P = 0.98) was similar in
both groups. Kaplan-Meier analysis of phototherapy duration showed no
difference in the survival curves of the 2 groups (P = 0.6). Axillary
temperature was similar in both groups and no baby was dehydrated.
Nursing staff reported no significant differences between CFL and STL
vis-à-vis glare hurting the eyes, giddiness and headache.
Conclusions: CFL phototherapy has no superiority over STL
phototherapy in terms of efficacy and adverse effects on the neonate and
effects on nursing staff.
Keywords: Jaundice, Neonate, Phototherapy.
About 3% of all hospital-born babies in India
develop significant jaundice, with Total Serum Bilirubin (TSB) levels
more than 15 mg/dL(1). Phototherapy is a useful method for treating
neonatal hyperbilirubinemia because it is easily available and devoid of
all complications of double volume exchange transfusions. The efficacy
of phototherapy depends on the dose and wavelength of light used and the
surface area exposed(2). Phototherapy is associated with side-effects
like dehydration and hyperthermia. Glaring, giddiness and headache
caused to nursing staff and difficulty in clinical monitoring are dis-advantages
associated with blue light photo-therapy(3).
Over the last 2 decades, there has been a constant
endeavour to develop ways to increase the efficacy of phototherapy and
at the same time reduce the side-effects and disadvantages to nursing
personnel. Recently, compact fluorescent special blue lamp phototherapy
units have been introduced in the Indian market. The manufacturers of
these units claim that they are more efficacious and more acceptable to
nursing staff compared to conventional units. These advantages are said
to be due to the smaller size, focused area, lower scatter and higher
irradiance. A Medline search did not yield any information regarding the
use of compact fluorescent lamps. To test the hypothesis that compact
fluorescent lamps are superior to conventional phototherapy units, we
performed a randomized, controlled trial.
Subjects and Methods
This study was conducted on inborn neonates with
jaundice admitted to a level III neonatal unit in Northern India. The
study protocol was approved by the Institute’s Ethics Committee.
Neonates with gestational age more than 34 completed weeks, with
hyperbilirubinemia requiring phototherapy (as per Cockington’s charts)
and who were other-wise "healthy" were included in the study(4). For the
purpose of the study, "healthy" was defined as "an active baby
exclusively on oral feeds with normal physiological vital parameters".
Those with Rh isoimmunization, those who had undergone a double volume
exchange transfusion for any reason prior to enrollment and whose
parents declined to give informed consent were excluded. Babies who
satisfied the eligibility criteria were randomly allocated to the 2
intervention groups: Compact Fluorescent Lamp (CFL) and Standard-length
Fluoresent Tube Light (STL) phototherapy. Stratified, block
randomization procedure was used, with stratification for gestational
age (<37 weeks and ³37 weeks). The random numbers were computer
generated, and slips bearing the allocated group were placed in serially
numbered, opaque, sealed envelopes.
The CFL equipments (supplied by Phoenix India Ltd)
comprised of 4 special blue CFL’s and 2 white CFL’s (Dulux®L 18W/71,
Osram, Italy) mounted on metal frames with adjustable heights. The lamps
were covered by a special transparent sheet that focuses and prevents
scattering of light. The STL equipments were indigenously made, and
comprised of 4 special blue tube light (TL 20 W/52BB®, Phillips India)
and 2 white tube lights mounted on metal frames with adjust-able
heights. Before starting phototherapy on a subject, the irradiance was
checked by a photoradiometer (Fluoro-lite 451®, Minolta/Air Shields,
USA). Our target was to maintain irradiance above 15 µW/nm/cm2 at all
times and lamps were replaced whenever necessary, to maintain this
irradiance. Our unit policy is to place the baby as close as possible to
the light source, provided it does not interfere with nursing care and
does not cause temperature instability. Except for an eye pad, all areas
are left uncovered. There is no specified protocol for position changes.
After enrolment, information was recorded regarding
demographic data and causes of jaundice. Blood group, direct Coomb’s
test (DCT), Glucose-6-phosphate dehydrogenase (G6PD) levels, peripheral
blood smear and reticulocyte counts were performed. Total Serum
Bilirubin (TSB) was measured every 12 hours by direct spectrophotometry
(Twin Beam Micro-bilimeter, Ginevri, Italy). The rate of decline of
phototherapy and the duration of phototherapy were calculated. The time
for which the babies were taken out of phototherapy for feeding, nappy
changes etc were also recorded to give an estimate of the actual number
of hours under phototherapy. Phototherapy was discontinued after 2
consecutive TSB values below phototherapy zone were obtained. Exchange
transfusions were done based on Cockington’s charts. The urinary
frequency, clinical signs of dehydration, weight, serum sodium and
axillary temperature were recorded every 12 hours. Mothers were
instructed to feed their babies every 2-3 hours.
To test the acceptability of the equipments by
nursing staff, we randomly selected 50 nurses, of the 64 nurses working
in the neonatal unit, and administered to them a questionnaire. They had
to answer the following 3 questions: Does the light (a) cause a
glare and hurt the eyes? (b) cause giddiness? (c) cause a
headache? They selected answers from 4 options: ‘no’, ‘minimally’,
‘moderately’ and ‘extremely’. Separate answers were given for CFL and
The primary outcome was the total duration of
phototherapy. Secondary out-comes were: rate of decline of TSB, need for
exchange transfusion, temperature instability, evidence of dehydration
and nurses’ reports of ‘moderate’ or ‘extreme’ distress.
A convenience sample of 100 babies was recruited,
since there were no prior data on this subject. Data of the 2 groups
were analyzed by Chi square test for categorical variables, Student’s
t test for normally distributed numerical variables and Mann Whitney
test for variables with skewed distributions. Babies who underwent a
double volume exchange were not included in the analysis of the duration
of phototherapy. For calculating the rate of decline of TSB, babies who
underwent an exchange transfusion were included till the point of time
when they underwent this procedure. Thereafter, they were excluded from
Of 293 consecutive babies with gestation more than 34
weeks who received photo-therapy, 100 babies, who satisfied the
eligibility criteria were enrolled and randomized (Fig. 1). There
were 50 babies in each group. Baseline demographic characteristics,
causes of hyperbilirubinemia and incidence of hemolysis were balanced
between the 2 groups (Table I). The mean TSB before starting
phototherapy was comparable in both groups. There was no difference in
the mean irradiance of the equipments at onset of phototherapy between
the CFL and STL groups [Mean irradiance (± SD): 18.39 ± 2.38 vs 17.77 ±
1.81 respectively, P >0.05].
Fig. 1. The Study Flow Chart
(n = 50)
(n = 50)
Reticulocyte count % (mean ± SD)
1.03 ± 1.45
0.69 ± 1.02
Hemolysis on peripheral smear
Baseline irradiance in (µW/nm/cm2 (mean ± SD)
18.39 ± 2.38
17.77 ± 1.81
Baseline TSB in mg/dL (mean ± SD)
16.25 ± 1.52
15.38 ± 1.95
All figures in parentheses are percentages
The total number of hours of phototherapy prescribed
were similar in both groups (P = 0.98) as was the time for which
subjects were actually kept under phototherapy (P = 0.96) (Table II).
The number of subjects who underwent a double volume exchange after
treatment allocation was higher in the CFL group, but this did not
achieve statistical significance.
Phototherapy and Exchange Transfusion
(n = 50)
(n = 50)
Hours of phototherapy from starting till stopping
40.66 ± 23.94
40.78 ± 21.83
Mean + SD [range]
Actual time under phototherapy*
29.86 ± 17.57
29.32 ± 16.00
Mean ± SD [range]
* Excludes the time for which babies were taken out of phototherapy for feeding, nappy
change and other aspects of nursing care.
The hour-specific TSB at 12 hourly intervals after
starting phototherapy were compared. There were no differences in the
percent change in each 12 hour interval, nor in the percent change in
TSB from baseline.
Repeated measures ANOVA was per-formed to analyze the
effect of the intervention on TSB. All subjects had at least 2 values of
TSB, i.e., baseline and 12 hours, but many did not have
subsequent values because of stoppage of phototherapy. Therefore, only
baseline and the 12-hr values were included in the repeated measures
ANOVA because inclusion of subsequent values would make the analysis
unbalanced. We observed an association between the 12-hr TSB level and
the type of phototherapy source. The estimated marginal mean TSB in the
CFL group was significantly higher than in STL group [15.53 (95%
confidence interval: 15.01, 16.05) versus 14.55 (95% confidence
interval: 14.03, 15.08) respectively; F value = 6.8; P = 0.01]. There
was a decline in mean TSB between baseline and 12 hours [15.86 (95%
confidence interval: 15.5, 16.2) versus 14.23 (95% confidence interval:
13.7, 14.8); F value = 42.4; P <0.001)].
Fig.2. Kaplan-Meier analysis comparing duration of phototherapy in
2 groups. Subjects undergoing exchange transfusions were censored
after the transfusion.
Kaplan-Meier analysis of the duration of phototherapy
in the 2 groups was performed (Fig. 2). The median duration in
the CFL group [36.0 hrs (95% confidence interval: 29, 43)] was
comparable to that in the STL group [42.0 hrs (95% confidence interval:
31.3, 52.7)], the difference being non-significant (P value of log rank
test = 0.62).
Sub-group analyses were performed to compare the
effects of CFL versus STL among various sub-groups: preterms, babies
with ABO incompatibility and those with G6PD deficiency. No
statistically significant differences were found.
Adverse effects were compared between the CFL and STL
groups. The axillary temperatures were similar at 12, 24, 48 and 72
hours (all P values >0.05). There were no cases of hyperthermia or
hypothermia in either group. There were no cases of dehydration among
our subjects, and serum sodium, urine specific gravity and weight
changes were similar till 72 hours (all P values >0.05).
The analysis of the reports of the nursing staff
showed a slight trend towards more disadvantages associated with STL,
but it did not reach statistical significance. Eyes feeling hurt by the
glare of CFL and STL was reported to a ‘moderate or extreme’ degree by
38% and 48% respectively (P = 0.16). Giddiness was reported as ‘moderate
or extreme’ by 14% and 20% respectively (P = 0.42) and headaches were
reported as ‘moderate or extreme’ by 6% and 8% respectively (P = 1).
The results of our randomized controlled trial showed
that, overall, CFL has no advantage over STL phototherapy in reducing
the duration of phototherapy, and both have no adverse effects on
temperature maintenance and hydration. The need for exchange
transfusions, in fact, showed a higher trend in the CFL group and on
repeated measures ANOVA the mean TSB was higher in the CFL group. In
these respects, CFL was an inferior modality.
We selected a group of relatively larger and healthy
babies, so that the outcome would not be influenced by co-morbidity.
Preterm and sick babies may have concomitant disturbances in fluid
balance, hemolysis due to sepsis or disseminated intravascular
coagulation, internal bleeds etc., that may affect the outcome,
hence this group was excluded. We excluded babies with Rh iso-immunization
and those who have already received an exchange transfusion, because
they are qualitatively different from other babies with
hyperbilirubinemia: they have different rates of hemolysis and a higher
likelihood of requiring multiple exchange transfusions.
The efficacy of phototherapy is altered by factors
that affect bilirubin production, such as oxytocin use, G6PD deficiency
and ABO incompatibility, all of which were balanced between the 2
groups. Since gestational age is a key determinant of the incidence and
severity of hyperbilirubinemia, we had stratified our population as
preterm and term. To compare the efficacy of 2 modes of phototherapy, it
is essential that the baseline TSB values be comparable, because the
rate of hourly decline of TSB is greater at high baseline TSB levels and
vice versa(5). In our study, baseline TSB values were comparable.
We had to exclude subjects who underwent an exchange transfusion from
the analysis of duration of photo-therapy, because an exchange
transfusion produces a drastic reduction of TSB and this would interfere
with any further estimation of phototherapy duration.
The efficacy of phototherapy depends on the
wavelength and irradiance of the light source and the surface area of
the infant exposed. These determinants of efficacy have been
demonstrated in several controlled trials. Tan reported that the decline
of TSB is most rapid with special blue lights and total duration of
exposure is also less, when compared to green or fluorescent day-light
lamps(6). High intensity blue light phototherapy have been found twice
as effective as standard daylight phototherapy in non-hemolytic
jaundice(7). Double surface phototherapy is more effective than single
phototherapy, where it is delivered by combining overhead phototherapy
with a bili-blanket or with a phototherapy bed(8). A more recent study
concluded that when using low light irradiance, there was no
statistically significant difference in the effectiveness of
photo-therapy using blue-green LEDs, blue LEDs or conventional
halogen-quartz bulbs(9). Light emitting diodes have been found to be as
effective as conventional blue light phototherapy at comparable
There is no previous study comparing CFL and STL
phototherapy. The manufacturers of CFL equipments claim that it provides
focused light with no scattering. This is supposed to increase efficacy
and decrease the glare that hurts the eyes of the nursing personnel.
However, we were unable to detect any significant difference in efficacy
or accept-ability of the CFL equipment. The reason for similar efficacy
may be attributed to the similar irradiance in both groups, and
maintained by us at more than 15 (µW/nm/cm2. We did not test the
wavelengths of the emission spectra of the 2 kinds of lamps. We assume
there was also no difference in the emission wavelength of the two kinds
of lamps, because both were special blue lamps. In our unit, double
surface phototherapy was not practiced during the study period.
An a priori sample size calculation was not
done because there was no existing data on which to base it. However, a
post hoc power calculation showed that with a standard deviation of
phototherapy duration equal to 24 hours (as observed in our study), to
identify an effect size of difference in mean duration equal to 12
hours, our sample size had 70% power. For 80% power we would have
required ~64 in each group.
A drawback of our study was that the intervention was
unblinded. The sample size had insufficient power to detect a clinically
meaningful difference in phototherapy duration (i.e., 12 hours),
although a comparison of the actual data shows that the durations were
so similar in 2 groups, that even an appropriate sample size would have
resulted in a negative study.
The initial cost of the CFL equipment and the
recurring cost of the CFL lamps is higher than STL equipments. We did
not record the useful life-span of each type of lamp, and hence we
cannot comment on the per hour cost efficiency of CFL versus STL lamps.
We conclude that CFL phototherapy equipment is not
superior to STL equipment, provided the irradiance is similar. There are
also no significant differences in adverse effects and effects on
nurses. On the basis of these parameters, there is nothing to choose
between these 2 types of equipment, except for considerations of cost
Contributions: MS collected the data. SD
conceived the idea, planned the study, analyzed the data and wrote the
manuscript. AN was involved in planning, supervising the study and
corrected the manuscript.
Competing interests: None.
Compact fluorescent lamp phototherapy
has no advantage over the conventional standard length
fluorescent tube lights.