You are on page 1of 11

A Clinical Prediction Rule for

Rebound Hyperbilirubinemia
Following Inpatient Phototherapy
Pearl W. Chang, MD,a Michael W. Kuzniewicz, MD, MPH,b,c Charles E. McCulloch, PhD,d Thomas B. Newman, MD, MPHb,c,d

OBJECTIVES: The American Academy of Pediatrics provides little guidance on when to abstract
discontinue phototherapy in newborns treated for hyperbilirubinemia. We sought to
develop a prediction rule to estimate the probability of rebound hyperbilirubinemia after
inpatient phototherapy.
METHODS: Subjects for this retrospective cohort study were infants born in 2012 to 2014 at
35 weeks gestation at 16 Kaiser Permanente Northern California hospitals who received
inpatient phototherapy before age 14 days. We defined rebound as the return of total serum
bilirubin (TSB) to phototherapy threshold within 72 hours of phototherapy termination. We
used stepwise logistic regression to select predictors of rebound hyperbilirubinemia and
devised and validated a prediction score by using split sample validation.
RESULTS: Of the 7048 infants treated with inpatient phototherapy, 4.6% had rebound
hyperbilirubinemia. Our prediction score consisted of 3 variables: gestational age <38
weeks (adjusted odds ratio [aOR] 4.7; 95% confidence interval [CI], 3.07.3), younger age
at phototherapy initiation (aOR 0.51 per day; 95% CI, 0.380.68), and TSB relative to the
treatment threshold at phototherapy termination (aOR 1.5 per mg/dL; 95% CI, 1.41.7).
The model performed well with an area under the receiver operating characteristic curve
of 0.89 (95% CI, 0.860.91) in the derivation data set and 0.88 (95% CI, 0.860.90) in the
validation data set. Approximately 70% of infants had scores <20, which correspond to a
<4% probability of rebound hyperbilirubinemia.
CONCLUSIONS: The risk of rebound hyperbilirubinemia can be quantified according to an
infants gestational age, age at phototherapy initiation, and TSB relative to the treatment
threshold at phototherapy termination.

aDepartment
WHATS KNOWN ON THIS SUBJECT: There are no
of Pediatrics, Seattle Childrens Hospital, Seattle, Washington; bDivision of Research, Kaiser
Permanente Northern California, Oakland, California; and Departments of cPediatrics, and dEpidemiology & standards and little evidence to support decisions
Biostatistics, University of California, San Francisco, California about when to discontinue phototherapy in
newborns being treated for hyperbilirubinemia.
Dr Chang conceptualized and designed the study, carried out statistical analysis and
interpretation of data, and drafted the initial manuscript; Drs Kuzniewicz and McCulloch assisted WHAT THIS STUDY ADDS: We describe a model to
with study design; Dr Newman conceptualized and designed the study, obtained funding, and quantify the risk of rebound hyperbilirubinemia.
guided statistical analysis and interpretation of data; and all authors revised and reviewed the This model will enable clinicians to discontinue
manuscript and approved the nal manuscript as submitted. phototherapy when the risk of rebound
DOI: 10.1542/peds.2016-2896 hyperbilirubinemia reaches a suitably low level.
Accepted for publication Dec 19, 2016
Address correspondence to Pearl W. Chang, MD, Seattle Childrens Hospital, M/S FA.2.115, PO Box
5371, Seattle, WA 98145-5005. E-mail: pearlchangmd@gmail.com To cite: Chang PW, Kuzniewicz MW, McCulloch CE, et al. A
PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275). Clinical Prediction Rule for Rebound Hyperbilirubinemia
Following Inpatient Phototherapy. Pediatrics. 2017;139(3):
Copyright 2017 by the American Academy of Pediatrics e20162896

Downloaded from by guest on August 12, 2017


PEDIATRICS Volume 139, number 3, March 2017:e20162896 ARTICLE
Phototherapy is a widely the benefit of earlier phototherapy nor order, we used the 1 hour before
used treatment of neonatal discontinuation outweighs the risk of discharge time (34%).
hyperbilirubinemia. The American rebound hyperbilirubinemia.
Academy of Pediatrics (AAP) TSB levels were measured via the
provides guidelines1 for when to start Vitros BuBc Neonatal Bilirubin
phototherapy, but little guidance is METHODS method (Ortho Clinical Diagnostics,
available for when to discontinue Raritan, NJ); 75.0% of study infants
phototherapy. A 1991 review Design, Subjects, and Human were born after a significant
recommended stopping birth hospital Subjects Committee Approval recalibration of these instruments.11
phototherapy after 2 consecutive To estimate the coefficient of
We selected subjects for this
total serum bilirubin (TSB) levels of variation of TSB measurements
retrospective cohort study from
<11 mg/dL.2 Although the appendix across hospitals, we used a linear
the population of 105808 infants
mixed effects regression with facility
to the AAP hyperbilirubinemia born at 35 weeks gestational age
guideline states that phototherapy as a random effect to model the first
at 16 Kaiser Permanente Northern
may be discontinued when the TSB TSB level as a function of facility
California hospitals between January
level falls to <14 mg/dL in infants of birth, age, and other covariates
1, 2012 and December 31, 2014. We
readmitted for phototherapy, no (adjusted R2 = 0.69). At a mean first
included subjects who underwent
guidance is provided for when to stop TSB of 8.9 mg/dL, the SD of the first
their first inpatient phototherapy
birth hospitalization phototherapy. TSB level across the 16 facilities
before age 14 days (n = 7202). We
In contrast, the United Kingdoms was 0.327 mg/dL (coefficient of
excluded infants (n = 126) who
variation = 3.7%).
National Institute for Health and Care did not have 2 TSB levels before
Excellence guideline recommends phototherapy termination (both not We subtracted the AAP phototherapy
stopping phototherapy 3 mg/dL required to be after phototherapy threshold from the last TSB before
below their treatment threshold, initiation) and infants (n = 28) with the start of phototherapy to obtain
which could be a TSB as high as a conjugated bilirubin level of 2 the relative TSB at phototherapy
17.5 mg/dL in a full-term infant.3 mg/dL before or during their first initiation. To estimate the TSB
The main risk of discontinuing phototherapy admission. After these at the time of phototherapy
phototherapy too early is exclusions, the study cohort had discontinuation, we used the TSB
the potential for rebound 7048 infants. level closest to phototherapy
hyperbilirubinemia, necessitating This study was approved by the termination, if it was measured
reinitiation of phototherapy and Kaiser Permanente Northern between 3 hours before and 1 hour
possible rehospitalization. Most California Institutional Review Board after termination (24.6% of the
previous studies of rebound and the University of California, San cohort). If no TSB level was measured
hyperbilirubinemia examined Francisco Committee on Human in this time window, we estimated a
rebound within a short time frame Research. TSB at 3 hours before phototherapy
of 24 hours and varied in their termination by linear extrapolation
definitions of rebound (returning Predictor Variables using the last 2 TSB levels before
to the bilirubin level at phototherapy termination. We then subtracted the
We derived all predictor variables AAP phototherapy threshold from
termination48 versus returning to
from electronic data sources. We the estimated TSB at termination
treatment threshold9,10). In addition,
determined the number of formula to obtain the relative TSB at
these studies examined a limited
feedings during phototherapy phototherapy termination.
number of risk factors, and sample
hospitalization from nursing
sizes were modest.
flowsheets. For phototherapy We assumed infants were continued
In this study, we used a large data start time, we used the time of the on home phototherapy after
set from an integrated health care first order for phototherapy. We inpatient phototherapy if there was
system to evaluate predictors of estimated the time of phototherapy an order for a home unit during
rebound hyperbilirubinemia in termination by using1 nursing hospitalization, and we considered
neonates after their first episode flowsheet documentation of infants to have restarted on home
of inpatient phototherapy. We discontinuation of phototherapy phototherapy after discharge if there
sought to create a clinical prediction (25% of cohort).2 If no such was an order for home equipment 1
rule to estimate the probability of documentation was available, calendar day before the day of our
rebound hyperbilirubinemia so that we used the time stamp of the primary outcome (defined below)
clinicians and parents can make more discontinuation order (41%).3 If or within 2 calendar days of initial
informed decisions about whether there was neither a nursing flowsheet inpatient phototherapy termination.

Downloaded from by guest on August 12, 2017


2 CHANG et al
Outcome Variables by using area under the receiver hyperbilirubinemia. In adjusted
operating characteristic (AUROC) analysis, lower gestational age
Our primary outcome was rebound
curve. We performed analyses by was associated with higher odds
hyperbilirubinemia, defined as the
using Stata version 14 (Stata Corp, of rebound hyperbilirubinemia,
return of TSB to or above the AAP
College Station, TX). especially infants born at 35 to
phototherapy threshold within 72
37 weeks gestation, for whom
hours of phototherapy termination.
the OR was >10 compared with
We chose 72 hours as a time RESULTS infants born at 40 weeks gestation.
frame that could be reasonably
Cohort Characteristics In addition, the odds of rebound
attributed to the same episode of
hyperbilirubinemia were higher
hyperbilirubinemia. We anticipated Characteristics of the study cohort for Asian infants and infants with a
that there were infants readmitted are shown in Table 1. The average higher relative TSB at phototherapy
for inpatient phototherapy before gestational age was 38.0 1.7 termination.
reaching treatment threshold who weeks. The mean postnatal age at
would be censored by our definition phototherapy initiation was 2.3 The odds of rebound
of rebound hyperbilirubinemia. (SD 1.3) days, and more than half hyperbilirubinemia were lower
Therefore, we performed sensitivity of infants (61.9%) underwent their in African American infants and
analyses in which we included these first inpatient phototherapy during infants at an older postnatal age at
infants in our primary outcome. the birth hospitalization. The TSB phototherapy initiation. Compared
In addition, we examined rebound at phototherapy initiation was close with infants with birth weights of
of TSB to 1 and 2 mg/dL above to the AAP phototherapy threshold 3000 to 3499 g, infants weighing
phototherapy threshold within 72 (mean difference 0.1 2.0 mg/dL). <2500 g had lower odds of rebound
hours of phototherapy termination as The mean estimated TSB at hyperbilirubinemia. Furthermore,
our secondary outcomes. phototherapy termination was 9.8 infants who received 4 formula
2.7 mg/dL, which averaged 7.4 feedings had lower odds of rebound
Statistical Analyses 3.1 mg/dL below the phototherapy hyperbilirubinemia, as did those who
threshold. Although ~75% of the TSB continued on home phototherapy
We obtained bivariate and
levels at phototherapy termination after inpatient treatment (OR 0.62;
multivariate odds ratios (ORs) by
were extrapolated, the average 95% CI, 0.410.94) (Table 3).
logistic regression, with standard
errors adjusted for clustering by difference between an infants
extrapolated value and his or her last Prediction Rule and Score
hospital. We devised a clinical
prediction rule and score by measured TSB was 0.4 1.1 mg/dL. After stepwise selection, the
using a split sample approach, prediction rule consisted of 3
Predictors of Rebound
with a random one-half of the Hyperbilirubinemia predictors: gestational age, age
cohort used for derivation of the at phototherapy initiation (as a
prediction rule and the other half Of the 7048 included infants, 324 continuous variable; OR 0.51; 95% CI,
for internal validation. To generate (4.6%) met our definition of rebound 0.380.68; logistic coefficient 0.68),
a parsimonious rule, we included hyperbilirubinemia (Table 1). The and relative TSB at phototherapy
variables if they were significant average age at the time of rebound termination (OR 1.5; 95% CI, 1.41.7;
(P < .01) by using backward stepwise hyperbilirubinemia was 3.3 1.5 logistic coefficient 0.42). For ease
logistic regression. We confirmed days. Thirty-four infants (0.5% of of clinical use, we dichotomized
these results by using best subsets the cohort) were readmitted for gestational age at <38 weeks (OR 4.7;
variable selection.12 To formulate inpatient phototherapy below the 95% CI, 3.07.3; logistic coefficient
the score, we summed the 3 highest treatment threshold. There were 715 1.5). Multiplying the logistic
ranked predictor variables, each infants in the cohort (10.1%) who did coefficients by 10 and adding 50, as
multiplied by 10 times its logistic not have a TSB measurement after previously described, resulted in the
regression coefficient (to avoid inpatient phototherapy termination, following equation for the score:
decimals), and added 50 to the total and 462 infants in the cohort (6.6%)
were later restarted on home Score = 15 (if gestational age <
(to avoid negative scores). Because
phototherapy, of whom 34 infants 38 weeks) 7 (age in days at
the logistic coefficients are equal to
(7.4%) met criteria for the primary phototherapy initiation) 4 (AAP
the logarithms of the ORs, summing
outcome. phototherapy threshold TSB at
them is equivalent to multiplying
phototherapy termination) + 50
their ORs. We assessed goodness of Table 2 shows the bivariate
fit by using the HosmerLemeshow predictors and Table 3 shows the For example, an infant born at 37
test (10 groups) and discrimination multivariate predictors of rebound weeks gestational age who was 2.5

Downloaded from by guest on August 12, 2017


PEDIATRICS Volume 139, number 3, March 2017 3
TABLE 1 Cohort Characteristics (N = 7048) 0.895 in the derivation sample.
Characteristic n % Therefore, we elected to include
Gestational age, wk, mean (SD) 38 (1.7)
only the original 3 variables in our
<38 wk 2791 39.6 prediction rule.
Birth wt, g, mean (SD) 3237 (577)
We also performed sensitivity
Male sex 3830 54.3
Infant race and ethnicity analyses with inclusion of gestational
White 2106 29.9 age as a categorical variable in the
Asian 2357 33.4 predictive model, which minimally
Hispanic 1607 22.8 affected the AUROC (0.891 as
African American 356 5.1
a categorical versus 0.887 as a
Other 453 6.4
Unknown 169 2.4 dichotomous variable). Inclusion
DAT of the 34 infants who underwent
Positive 1019 14.5 recurrent phototherapy below
Negative 3340 47.4 treatment threshold did not alter
Not done 2689 38.2
Feeding during phototherapy hospitalization
the predictors selected by stepwise
Breast milk only 2051 29.1 regression or the equation or have
13 formula feedings 994 14.1 a marked effect on the Hosmer
410 formula feedings 2605 37.0 Lemeshow goodness of fit (P = .22)
11 formula feedings 1107 15.7 and AUROC (0.877) in the derivation
Unknown 291 4.1
Phototherapy during birth hospitalization 4364 61.9
data set. In the subset of infants
Initiation of phototherapy with measured TSB at phototherapy
Age, d, mean (SD) 2.3 (1.3) termination (n = 1737), AUROC was
Relative TSB, mg/dL, mean (SD) 0.1 (2.0) slightly higher at 0.901 (95% CI,
Termination of phototherapy 0.8710.930). For our secondary
Age, d, mean (SD) 3.6 (1.3)
Estimated TSB, mg/dL, mean (SD) 9.8 (2.7)
outcome of TSB rebound to 1
Estimated relative TSB, mg/dL mg/dL above treatment threshold,
<13 197 2.8 application of the above score gave
13 to <10 1153 16.4 an AUROC of 0.888 (95% CI, 0.866
10 to <7 2677 38.0 0.909), and for TSB rebound to 2
7 to <4 2114 30.0
4 to <0 839 11.9
mg/dL above treatment threshold,
0 68 1.0 the AUROC was 0.901 (95% CI,
Home phototherapy order 0.8690.932).
During inpatient phototherapy 310 4.4
After inpatient phototherapy termination 462 6.6 The probability of rebound
Rebound hyperbilirubinemia within 72 h 324 4.6 hyperbilirubinemia was <10% with
0 to <1 above AAP phototherapy threshold 159 2.3 a prediction score of <30 (87% of
1 to <2 above AAP phototherapy threshold 86 1.2 the derivation group, 86% of the
2 above AAP phototherapy threshold 79 1.1
validation group) and <4% with
Recurrent inpatient phototherapy below threshold 34 0.5
a prediction score of <20 (71% of
, not applicable.
the derivation group, 69% of the
validation group) (Table 4 and Fig 2).
days (60 hours) old at phototherapy of freedom) was 7.7 (P = .47) and There were 3023 infants who had
initiation and whose TSB was 6 mg/ the AUROC was 0.887 (95% CI, 2 measured TSBs after the start
dL below the treatment threshold at 0.8640.910). The validation data of phototherapy. Of these infants,
phototherapy termination (eg, TSB set (n = 3530) had a similar AUROC 1025 (34%) had a score of <20 at
11.5 mg/dL at 96 hours) would have (0.881) and 95% CI (0.8590.903) the penultimate TSB measurement
a score of 15 (7 2.5) (4 6) + (Fig 1). With the best subsets variable and were continued on inpatient
50 = 23.5, whereas a score for the selection, the best model by Bayesian phototherapy for another average
same baby at 38 weeks gestation information criterion consisted of the 23 9 hours.
would be 15 fewer, or 8.5. same 3 predictors. The best model
The discrimination and fit of by Akaikes information criterion
the predictive model using the consisted of the same 3 predictors DISCUSSION
generated score were excellent. In plus formula use and race and We used a large cohort to
the derivation data set (n = 3518), ethnicity. With a 5-variable model, examine predictors of rebound
the HosmerLemeshow 2 (8 degrees the AUROC minimal improved to hyperbilirubinemia, defined

Downloaded from by guest on August 12, 2017


4 CHANG et al
as returning to phototherapy TABLE 2 Bivariate Predictors of Rebound Hyperbilirubinemia
treatment threshold within Variable N Total N % OR 95% CI P
72 hours of termination of Rebound Rebound
an infants first inpatient Sex .43
phototherapy. We found that Female 3218 142 4.4 Reference
rebound hyperbilirubinemia can Male 3830 182 4.8 1.08 0.891.31
Gestational age, wk <.001
be predicted well from the infants
35 643 28 4.4 1.90 1.302.78
gestational age, age at phototherapy 36 834 59 7.1 3.18 1.825.57
initiation, and relative TSB at 37 1314 120 9.1 4.20 2.636.71
phototherapy termination. 38 1152 34 3.0 1.27 0.881.84
39 1615 50 3.1 1.34 0.971.84
In our study cohort, 4.6% 40 1070 25 2.3 Reference
(324 out of 7048) had rebound 41 420 8 1.9 0.81 0.391.69
hyperbilirubinemia and 0.5% Birth wt, g .002
<2000 131 1 0.8 0.15 0.030.69
(34 out of 7048) were readmitted
20002499 554 16 2.9 0.60 0.360.98
for inpatient phototherapy below 25002999 1674 92 5.5 1.17 0.901.51
treatment threshold. This is lower 30003499 2446 116 4.7 Reference
than what Barak et al10 found in 35003999 1620 74 4.6 0.96 0.691.34
their randomized controlled trial, 40004499 490 19 3.9 0.81 0.391.69
4500 132 6 4.5 0.96 0.382.41
in which 19% of infants needed
Infant race and ethnicity <.001
reinitiation of phototherapy for White 2106 82 3.9 Reference
a TSB level that returned to AAP Asian 2357 137 5.8 1.52 1.132.05
phototherapy thresholds 24 hours Hispanic 1607 75 4.7 1.21 0.771.90
after treatment termination. African American 356 13 3.7 0.94 0.561.56
Other 453 8 1.8 0.69 0.331.45
However, phototherapy in that study
DAT <.001
was discontinued at an average Negative 3340 135 4.0 Reference
of 1.7 mg/dL or 4.1 mg/dL below Positive 1019 91 8.9 2.33 1.763.08
treatment threshold (depending on Not done 2689 98 3.6 0.90 0.641.25
the randomization group), much Feeding during <.001
phototherapy
higher TSB levels at termination
hospitalization
than the average in this study. In Breast milk only 2051 131 6.4 Reference
contrast, in the Bansal et al9 study, 13 formula feedings 994 46 4.6 0.71 0.471.07
7.3% of 232 infants who had a 410 formula feedings 2605 91 3.5 0.53 0.430.66
postphototherapy TSB returned to 11 formula feedings 1107 37 3.3 0.51 0.320.80
Age at phototherapy 0.47 0.390.56 <.001
phototherapy threshold within
initiation, per day
24 6 hours of treatment Age at phototherapy 0.46 0.350.60 <.001
termination, although that study termination, per day
included infants born at <35 weeks Relative TSB at 1.55 1.411.70 <.001
gestation. phototherapy
termination, per mg/dL
We found that 3 of the Home phototherapy at <.001
best predictors of rebound discharge
No 6738 290 4.3 Reference
hyperbilirubinemia were an Yes 310 34 11.0 2.74 1.844.08
infants gestational age, age at Born before BuBc .17
phototherapy initiation, and TSB recalibration
level at phototherapy termination No 5273 253 4.8 Reference
relative to the AAP treatment Yes 1775 71 4.0 0.83 0.631.08
threshold. Gestational age is , not applicable.

a well-established risk factor


for neonatal jaundice1 and was Maisels et al13 found that a second hours at phototherapy initiation
also found to be a risk factor for course of phototherapy occurred was a risk factor for rebound,
rebound hyperbilirubinemia and more commonly in infants who although rebound in that study
recurrent phototherapy in previous received phototherapy during their was defined as rising above the TSB
studies.8,9 Previous studies have birth hospitalization than infants level at phototherapy termination.
also found postnatal age to be a risk readmitted for phototherapy, and Most previous studies have not
factor for recurrent phototherapy. Kaplan et al8 found that age <72 examined TSB level as a risk factor

Downloaded from by guest on August 12, 2017


PEDIATRICS Volume 139, number 3, March 2017 5
TABLE 3 Multivariate Predictors of Rebound Hyperbilirubinemia supplementing with formula,
Variable OR 95% CI P discharging an infant
Male sex 1.19 0.951.49 .124 with home phototherapy (if
Gestational age, wk <.001 available), or lowering the relative
35 10.63 5.4820.62 TSB by an additional 1 mg/dL at
36 11.39 6.5119.92 phototherapy termination with
37 11.68 6.9519.64
similar efficacy.
38 2.70 1.883.89
39 1.91 1.342.74 Although the OR for a positive
40 Reference direct antiglobulin test (DAT) was
41 0.76 0.341.70
significant in unadjusted analyses,
Birth wt, g <.001
<2000 0.12 0.040.37 the adjusted OR was only 1.37
20002499 0.33 0.190.56 (95% CI, 0.902.07), probably
25002999 0.83 0.571.20 because a positive DAT moves a
30003499 Reference baby to a higher risk group, with
35003999 1.40 0.991.97
a lower phototherapy threshold.
40004499 1.43 0.633.25
4500 1.74 0.605.01 This change in turn reduces
Infant race and ethnicity <.001 the difference between the last
White Reference measured TSB and the threshold.
Asian 1.62 1.112.36 Thus, the effect of the DAT is
Hispanic 1.02 0.671.54
captured by the difference between
African American 0.46 0.260.81
Other 0.53 0.251.12 the last TSB and the phototherapy
DAT .34 threshold.
Negative Reference
Positive 1.37 0.902.07
Not done 0.86 0.591.27 In our cohort, 34% of infants may
Feeding during phototherapy .001 have been able to discontinue
hospitalization
Breast milk only Reference
inpatient phototherapy a day
13 formula feedings 0.92 0.621.36 earlier with <4% risk of rebound
410 formula feedings 0.63 0.480.83 hyperbilirubinemia. The decision to
11 formula feedings 0.43 0.270.68 discontinue phototherapy is based
Age at phototherapy initiation, 0.38 0.330.44 <.001 on balancing the risks and costs
per day
Age at phototherapy termination, 1.15 0.831.57 .40
of prolonging treatment against
per day the benefit of reducing the risk
Relative TSB at phototherapy 1.48 1.321.66 <.001 of rebound hyperbilirubinemia.
termination, per mg/dL Hospitalization is burdensome
Home phototherapy at discharge 0.62 0.410.94 .02 for families, and phototherapy
, not applicable. can disrupt breastfeeding and
infant bonding. In addition,
for rebound hyperbilirubinemia. were formula feeding during the there is emerging evidence
In the aforementioned Bansal phototherapy hospitalization and that phototherapy may have
et al9 trial, there was no significant continuing on home phototherapy potential associations with
difference in the occurrence of after discharge. A higher number melanocytic nevi14 and infantile
repeat phototherapy between of formula feeds was associated cancer, especially acute myeloid
infants randomly assigned to with lower odds of rebound leukemia.15,16 Our prediction
discontinue phototherapy at a TSB hyperbilirubinemia. Interestingly, score quantifies the probability
of 1 mg/dL versus 3 mg/dL the OR for 4 to 10 formula of rebound hyperbilirubinemia
below the AAP phototherapy feedings (0.63) was similar to to help physicians and parents
threshold (5 out of 25 vs 5 out that for continuation on home decide, based on their level of
of 27, respectively, P = .58), but phototherapy (OR = 0.62) and for acceptable risk for rebound
the trial was small, consisting of a 1mg/dL decrease in relative TSB hyperbilirubinemia, when to
52 infants born at >36 weeks at phototherapy termination (OR = discontinue phototherapy. For
gestation. 0.68). These results suggest that a example, consider a 37-week
clinician aiming to reduce the risk gestational age infant who starts
Additional significant predictors of rebound hyperbilirubinemia phototherapy at 4 days of age.
of rebound hyperbilirubinemia further could consider Discontinuing phototherapy at a

Downloaded from by guest on August 12, 2017


6 CHANG et al
FIGURE 1
Receiver operating characteristic curve of the prediction score in the validation data set. Area under receiver operating characteristic curve = 0.8808.

FIGURE 2
Probability of rebound hyperbilirubinemia by score. Score = 15 (if gestational age <38 weeks) 7 (age in days at phototherapy initiation) 4 (AAP
phototherapy threshold TSB at phototherapy termination) + 50.

TSB of 5 mg/dL below treatment for whom follow-up TSB testing This study has limitations. One of our
threshold gives a score of 17 is difficult or readmission for key predictor variables, the TSB at
and an estimated 2.8% probability hyperbilirubinemia presents time of phototherapy termination,
of rebound hyperbilirubinemia. a greater hardship, it may make was estimated by extrapolation
In comparison, the probability sense to continue phototherapy for the majority of our subjects,
of rebound would increase to 6.0% longer. On the other hand, a 10% which presumably worsened the
at a TSB of 3 mg/dL below and or 15% risk may be acceptable for discrimination of the prediction rule.
12.3% at a TSB of 1 mg/dL below a reliable family close to an infant However, we envision clinicians
treatment threshold. In infants care center. using the rule to decide whether to

Downloaded from by guest on August 12, 2017


PEDIATRICS Volume 139, number 3, March 2017 7
continue phototherapy each time a TABLE 4 Risk of Rebound Hyperbilirubinemia by Score
TSB result becomes available. Thus, Infants With Rebound Hyperbilirubinemia
they will know the TSB at the time of Derivation Group (N = 3518) Validation Group (N = 3530)
phototherapy termination and will
Prediction Score N % N %
not need to extrapolate. In the subset
of subjects in whom the TSB at time 9 6/1792 0.3 5/1723 0.3
1019 20/707 2.8 13/708 1.8
of phototherapy termination was 2029 27/568 4.8 38/617 6.1
known rather than extrapolated, the 3039 56/303 18.5 55/316 17.4
AUROC for the clinical prediction rule 4049 36/109 33.0 32/124 25.8
was 0.90. 50 19/39 48.7 17/42 40.5

Additionally, we based our variable externally validate our prediction about when to discontinue
for home phototherapy on equipment rule, a consideration for future phototherapy.
orders, and therefore we did not research.
know precisely whether and when ACKNOWLEDGMENT
home phototherapy was used. There The authors thank Dr Andrea C.
may have been infants whose TSB Wickremasinghe for her critical review
returned to treatment threshold CONCLUSIONS
of the manuscript and invaluable
within 72 hours who did not have Rebound hyperbilirubinemia revisions and suggestions.
a TSB measurement until later. can be predicted with
Given this limitation and the use excellent discrimination by an ABBREVIATIONS
of home phototherapy, the risk of infants gestational age, age at
AAP:American Academy of
rebound hyperbilirubinemia may be initiation of phototherapy, and
Pediatrics
underestimated in our study, which relative TSB at phototherapy
aOR:adjusted odds ratio
may not be generalizable to infants termination. With a prediction
AUROC:area under the receiver
for whom home phototherapy is not score of <20, phototherapy
operating characteristic
an option. We also only examined can be discontinued with <4%
curve
rebound hyperbilirubinemia after probability of rebound. Clinical
CI:confidence interval
infants first inpatient phototherapy, implementation of this prediction
DAT:direct antiglobulin test
and rebound risks may be different rule via a Web-based calculator or
OR:odds ratio
after subsequent phototherapy. integration into electronic medical
TSB:total serum bilirubin
Finally, we were not able to records could help guide decisions

FINANCIAL DISCLOSURE: The authors have indicated they have no nancial relationships relevant to this article to disclose.
FUNDING: Partially supported by grant R01HS020618 from the Agency for Healthcare Research and Quality. The content is solely the responsibility of the authors
and does not necessarily represent the ofcial views of the Agency for Healthcare Research and Quality. The funder played no role in the design and conduct of
the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conicts of interest to disclose.
COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/peds.2016-3832.

REFERENCES

1. American Academy of Pediatrics at: https://www.nice.org.uk/guidance/ 6. Al-Saedi SA. Rebound


Subcommittee on Hyperbilirubinemia. cg98. Accessed February 23, hyperbilirubinemia in term infants
Management of hyperbilirubinemia 2016. after phototherapy. Saudi Med J.
in the newborn infant 35 or more 2002;23(11):13941397
4. Lazar L, Litwin A, Merlob P.
weeks of gestation. Pediatrics.
Phototherapy for neonatal 7. Erdeve O, Tiras U, Dallar Y. Rebound
2004;114(1):297316
nonhemolytic hyperbilirubinemia. bilirubin measurement is not
2. Tan KL. Phototherapy for Analysis of rebound and indications required for hyperbilirubinemia
neonatal jaundice. Clin Perinatol. for discontinuing therapy. Clin Pediatr regardless of the background
1991;18(3):423439 (Phila). 1993;32(5):264267 attributes of newborns. J Trop Pediatr.
3. National Institute for Health and Care 2004;50(5):309
Excellence. Neonatal jaundice (NICE 5. Yetman RJ, Parks DK, Huseby V, Mistry
Clinical Guideline 98). London, England: K, Garcia J. Rebound bilirubin levels in 8. Kaplan M, Kaplan E, Hammerman C,
National Institute for Health and infants receiving phototherapy. et al. Post-phototherapy neonatal
Clinical Excellence; May 2010. Available J Pediatr. 1998;133(5):705707 bilirubin rebound: a potential cause of

Downloaded from by guest on August 12, 2017


8 CHANG et al
signicant hyperbilirubinaemia. between laboratory calibration of and melanocytic nevi: a twin study.
Arch Dis Child. 2006;91(1): a serum bilirubin assay, neonatal Pediatrics. 2011;128(4). Available at:
3134 bilirubin levels, and phototherapy www.pediatrics.org/cgi/content/full/
9. Bansal A, Jain S, Parmar VR, Chawla use. JAMA Pediatr. 2016;170(6): 128/4/e856
D. Bilirubin rebound after intensive 557561 15. Newman TB, Wickremasinghe
phototherapy for neonatal jaundice. 12. Lindsey C, Sheather S. Best subsets AC, Walsh EM, Grimes BA,
Indian Pediatr. 2010;47(7): variable selection in nonnormal McCulloch CE, Kuzniewicz MW.
607609 regression models. Stata J. Retrospective cohort study of
2015;4(13):10461059 phototherapy and childhood cancer
10. Barak M, Berger I, Dollberg S,
in northern California. Pediatrics.
Mimouni FB, Mandel D. When should 13. Maisels MJ, Kring E. Rebound in
2016;137(6):e20151354
phototherapy be stopped? A pilot serum bilirubin level following
study comparing two targets of serum intensive phototherapy. Arch Pediatr 16. Wickremasinghe AC, Kuzniewicz
bilirubin concentration. Acta Paediatr. Adolesc Med. 2002;156(7): MW, Grimes BA, McCulloch CE,
2009;98(2):277281 669672 Newman TB. Neonatal phototherapy
11. Kuzniewicz MW, Greene DN, Walsh EM, 14. Csoma Z, Tth-Molnr E, Balogh K, and infantile cancer. Pediatrics.
McCulloch CE, Newman TB. Association et al. Neonatal blue light phototherapy 2016;137(6):e20151353

Downloaded from by guest on August 12, 2017


PEDIATRICS Volume 139, number 3, March 2017 9
A Clinical Prediction Rule for Rebound Hyperbilirubinemia Following Inpatient
Phototherapy
Pearl W. Chang, Michael W. Kuzniewicz, Charles E. McCulloch and Thomas B.
Newman
Pediatrics; originally published online February 14, 2017;
DOI: 10.1542/peds.2016-2896
Updated Information & including high resolution figures, can be found at:
Services /content/early/2017/02/13/peds.2016-2896.full.html
References This article cites 15 articles, 5 of which can be accessed free
at:
/content/early/2017/02/13/peds.2016-2896.full.html#ref-list-1

Citations This article has been cited by 5 HighWire-hosted articles:


/content/early/2017/02/13/peds.2016-2896.full.html#related-u
rls
Subspecialty Collections This article, along with others on similar topics, appears in
the following collection(s):
Fetus/Newborn Infant
/cgi/collection/fetus:newborn_infant_sub
Hyperbilirubinemia
/cgi/collection/hyperbilirubinemia_sub
Permissions & Licensing Information about reproducing this article in parts (figures,
tables) or in its entirety can be found online at:
/site/misc/Permissions.xhtml
Reprints Information about ordering reprints can be found online:
/site/misc/reprints.xhtml

PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly


publication, it has been published continuously since 1948. PEDIATRICS is owned, published,
and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk
Grove Village, Illinois, 60007. Copyright 2017 by the American Academy of Pediatrics. All
rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.

Downloaded from by guest on August 12, 2017


A Clinical Prediction Rule for Rebound Hyperbilirubinemia Following Inpatient
Phototherapy
Pearl W. Chang, Michael W. Kuzniewicz, Charles E. McCulloch and Thomas B.
Newman
Pediatrics; originally published online February 14, 2017;
DOI: 10.1542/peds.2016-2896

The online version of this article, along with updated information and services, is
located on the World Wide Web at:
/content/early/2017/02/13/peds.2016-2896.full.html

PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly


publication, it has been published continuously since 1948. PEDIATRICS is owned,
published, and trademarked by the American Academy of Pediatrics, 141 Northwest Point
Boulevard, Elk Grove Village, Illinois, 60007. Copyright 2017 by the American Academy
of Pediatrics. All rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.

Downloaded from by guest on August 12, 2017

You might also like