Chapter  65:  Management of Neonatal Hyperbilirubinemia

Association of Neonatal Hyperbilirubinemia with Neurodevelopmental Outcomes

1. 1. What is the relationship between peak bilirubin levels and/or duration of hyperbilirubinemia and neurodevelopmental outcome?

2. 2. What is the evidence for effect modification of the results in Question 1, by gestational age, hemolysis, serum albumin, and other factors?

Treatments for Neonatal Hyperbilirubinemia

1. 3. What are the quantitative estimates of efficacy of treatment at (1) reducing peak bilirubin levels (e.g., number needed to treat at 20 mg/dl to keep TSB from rising); (2) reducing the duration of hyperbilirubinemia (e.g., average number of hours by which time TSB greater than 20 mg/dl may be shortened by treatment); and (3) improving neurodevelopmental outcomes?

Diagnosis of Neonatal Hyperbilirubinemia

1. 4. What is the efficacy of various strategies for predicting hyperbilrubinemia, including hour-specific bilirubin percentiles?

2. 5. What is the accuracy of transcutaneous bilirubin (TcB) measurements?

Patient Population and Settings

The target population included infants of at least 34 weeks gestational age. Based on the findings of an earlier National Institute of Child Health and Human Development (NICHD) study, in which none of the 1,339 infants greater than or equal to 2,500 grams were less than 34 weeks, the EPC felt justified in grouping infants weighing greater than or equal to 2,500 grams with those greater than or equal to 34 weeks gestation.

Literature Search and Review Parameters

Methodological Quality

Methodological quality or internal validity addresses the design, conduct, and reporting of the study. Some of the items belonging to this domain are widely used in various “quality” scales and for randomized controlled trials, and usually include items such as concealment of random allocation, treatment blinding, and handling of dropouts. Because different types of study designs are used to address different questions and for consistency in the interpretation across different designs, we defined a three-category scale to report the methodological quality of the studies in the evidence report: A (least bias), B (susceptible to some bias), or C (likely to have large bias).

Statistical Analysis

The number needed to treat (NNT), expressing the benefit of an active treatment over a control, was calculated to quantify the efficacy of treatment for neonatal hyperbilirubinemia. For Question 3 in this report, NNT can be interpreted as the number of newborns needed to be treated at 20 mg/dl to keep the TSB in one newborn from rising.

A meta-analysis of Question 4 was conducted using the summary ROC method to combine studies, which evaluated diagnostic test performance. A meta-analysis of correlation coefficients was conducted to correlate performance of transcutaneous bilirubin measurements with serum bilirubin.

Results

Search Strategies

We searched the MEDLINE® database on September 25, 2001 for publications from 1966 to the present using relevant MeSH terms (“hyperbilirubinemia”, “hyperbilirubinemia, hereditary”, “bilirubin”, “jaundice, neonatal”, “kernicterus”) and text words (“bilirubin”, “hyperbilirubinemia”, “jaundice”, “kernicterus”, “neonatal”). The abstracts were limited to human and English studies focusing on newborns between birth and one month of age. In addition, the same textwords used for the MEDLINE® search were used to search the Pre- MEDLINE® database. The strategy yielded 4,280 MEDLINE® and 45 Pre-MEDLINE® abstracts. We consulted domain experts and examined relevant review articles for additional studies. A supplemental search for case reports of kernicterus in reference lists of relevant articles and reviews was also performed.

Screening and Selection Process

In our preliminary screening of abstracts, we identified over 600 potentially relevant articles for questions 1, 2, and 3. To handle this large number of articles, we devised the following scheme to address the key questions and to ensure that the report was completed within the time and resources constraints. We included only studies that measured neurodevelopment or behavioral outcomes (except for question 3 part-1, for which we evaluated all studies addressing the number-needed-to-treat (NNT) issue regardless of whether the study reported these outcomes). For the specific question on quantitative estimates of treatment efficacy, all studies concerning therapies designed to prevent hyperbilirubinemia (generally defined as bilirubin ≥ 20 mg/dl) were included in the review. The inclusion and exclusion criteria for the systematic review were discussed in several teleconferences of the EPC evidence review team and technical experts. The criteria underwent several revisions before their final acceptance by the panel members. The final screening criteria for inclusion and exclusion of articles are described below.

Inclusion Criteria

The target population of this review was healthy, full-term infants. For the purpose of this review, we included articles concerning infants who were at least 34 weeks estimated gestational age (EGA) at the time of birth. From studies that reported birthweight rather than age, infants whose birthweight was greater or equal to 2,500 grams were included. This cut-off was derived from findings of the National Institute of Child Health and Human Development hyperbilirubinemia study (Bryla, 1985), in which none of the 1,339 infants greater than or equal to 2,500 grams was less than 34 weeks EGA. Articles were selected for inclusion in the systematic review based on the following additional criteria:

• Key Question 1 or 2 (risks association)

  • Population: Infants ≥ 34 weeks EGA or ≥ 2,500 grams

  • Sample Size: More than 5 subjects per arm

  • Predictors: Jaundice or hyperbilirubinemia

  • Outcomes: At least one behavioral/neurodevelopmental outcome reported in the article

  • Study Design: Prospective cohorts (more than 2 arms), prospective cross-sectional study, prospective longitudinal study, prospective single-arm study, or retrospective cohorts (more than 2 arms).

  Case reports of kernicterus:

  • Population: Kernicterus case

  • Study Design: Case reports with kernicterus as a predictor or an outcome

  • Definitions of kernicterus: Acute phase of kernicterus (poor feeding, lethargy, high-pitched cry, increased tone, opisthotonus, seizures), kernicterus sequelae (motor delay, sensorineural hearing loss, gaze palsy, dental dysplasia, cerebral palsy, mental retardation), necropsy finding of yellow-staining in the brain nuclei.

• Key Question 3 (treatments)

  Number-Needed-to-Treat (NNT) question:

  • Population: Infants ≥ 34 weeks EGA or ≥ 2,500 grams

  • Sample Size: More than 10 subjects per arm

  • Treatments: Any treatment for neonatal hyperbilirubinemia

  • Outcomes: Serum bilirubin level ≥ 20 mg/dl or frequency of exchange transfusion specifically for bilirubin level ≥ 20 mg/dl

  • Study Design: Randomized or non-randomized controlled trials

  All other issues:

  • Population: Infants ≥ 34 EGA or ≥ 2,500 grams

  • Sample Size: More than 10 subjects per arm for phototherapy; any sample size for other treatments

  • Treatments: Any treatment for neonatal hyperbilirubinemia

  • Outcomes: At least one neurodevelopmental outcome was reported in the article

• Key Question 4 or 5 (diagnosis)

  • Population: Infants ≥ 34 EGA or birthweight ≥ 2,500 grams

  • Sample Size: More than 10 subjects

  • Reference Standard: Laboratory-based serum bilirubin

Exclusion Criteria

Case reports of kernicterus were excluded if they did not report serum bilirubin level, or gestational age and birthweight.

Results of Screening of Titles and Abstracts

Preliminary screening identified 663 out of a total of 4,560 abstracts located through the literature search described above. There were 158, 174, 99, 153, and 79 abstracts for question 1, 2, 3, 4 and 5 respectively.

Screening of Full-Text Articles

After full-text screening (according to the inclusion and exclusion criteria described above), 138 of total 253 retrieved articles were included in this report. There were 35 articles in the correlation section (questions 1 and 2), 28 articles of kernicterus case reports, 21 articles in the treatment section (question 3), and 54 articles in the diagnosis section (questions 4 and 5). There were inevitable overlaps because treatment effects and neurodevelopmental outcomes were inherent in the study designs. Below is a summary of the four-step selection process.

Summarizing the Evidence of Individual Studies

Grading of the evidence can be useful for indicating the overall methodological quality of a study. While a simple evidence grading system using a single scale may be desirable, the “quality” of evidence is multi-dimensional, and a single metric cannot fully capture information needed to interpret a clinical study. We believe that information on individual components of a study contribute more to the evaluation of evidence by deliberating bodies than a single summary score. The evidence-grading scheme used here assesses four dimensions that are important for the proper interpretation of the evidence:

• study size

• applicability

• summary of results

• methodological quality

Applicability, also known as generalizability or external validity, addresses the issue of whether the study population is sufficiently broad to be generalizable to the population at large. Individual studies are often unable to achieve broad applicability due to restricted study population characteristics and a small number of study subjects (Lau, Ioannidis, and Schmid, 1997). In this evidence report, because of the relative homogeneity (primary focus being healthy newborns) of the study populations, applicability is not explicitly graded for each study. Instead, where applicable, studies are grouped together to form more similar subgroups for analyses or discussion.

Study Size

The study sample size is used as a measure of the weight of the evidence. A large study provides a more precise estimation of the treatment effect but does not automatically confer broad applicability unless the study included a broad spectrum of patients. Very small studies, taken individually, cannot achieve broad applicability. But several small studies that enrolled diverse populations, taken together, may have broad applicability. The study size is included as a separate dimension used to assist the assessment of applicability. For summarizing all studies, this would be the number of studies and the total number of patients in these studies.

Methodological Quality

Methodological quality or internal validity addresses the design, conduct, and reporting of the study. Some of the items belonging to this entity are widely used in various “quality” scales and for randomized controlled trials. They usually include items such as concealment of random allocation, treatment blinding, and handling of dropouts. Because different types of study designs are used to address different questions and for consistency in the interpretation across different designs, we defined a three category scale to report the methodological quality of the studies in the evidence report: A (least bias), B (susceptible to some bias), or C (likely to have large bias). These criteria are described below.

• Criteria for evaluating the methodological quality of studies are that assess association (questions 1 and 2):

  1. Prospective. Complete methods and results (including inclusion/exclusion criteria). Proper control/comparison group, correct analyses performed.

  2. Prospective or retrospective. Not all criteria of A. Some deficiencies; however, unlikely to cause major bias.

  3. Prospective or retrospective. Significant design or reporting errors, large amount of missing information or potential bias.

• Criteria for evaluating the methodological quality of studies that assess effects of treatments (question 3):

  1. Randomized controlled trial. Complete methods and results (including inclusion/exclusion criteria) described. Proper randomization and/or blinding, and correct analyses performed.

  2. Non-randomized controlled trial or other prospective design (prospective cohort or case-control study). Proper selection of control group. Not all criteria of A. Some deficiencies; however, unlikely to cause major bias.

  3. Retrospective or no control group. Significant design or reporting errors, large amount of missing information or significant potential bias.

• Criteria for evaluating the methodological quality of studies that assess diagnostic test performance (questions 4 and 5):

  1. Prospective. Complete methods and results (including inclusion/exclusion criteria) described. Proper reference standard used and correct analyses performed.

  2. Prospective or retrospective. Not all criteria of A. Some deficiencies; however, unlikely to cause major bias.

  3. Prospective or retrospective. Significant design or reporting errors, large amount of missing information or bias.

Definition of Terminology in This Report

• Confounders (for Key Question 1 only): (1) An ideal study design to answer Question 1 would be to follow two groups, jaundiced and normal infants, without treating any infant for a current or consequent jaundice condition, and observe their neurodevelopmental outcomes. Therefore, any treatment received by the subjects in the study was defined as a confounder. (2) If subjects had known risk factors of jaundice, such as prematurity or low birth weight, the risk factors were defined as confounders. (3) Any disease condition other than jaundice was defined as a confounder. (4) Since bilirubin level is the essential predictor, if the study did not report or measure bilirubin levels for the subjects, lack of bilirubin measurements was defined as a confounder.

• Acute phase of kernicterus: poor feeding, lethargy, high-pitched cry, increased tone, opisthotonus, seizures.

• Chronic kernicterus sequelae: motor delay, sensorineural hearing loss, gaze palsy, dental dysplasia, cerebral palsy, mental retardation.

Brainstem Auditory Evoked Potential (BAEP) or Brainstem Auditory Evoked Response (BAER)

• This test is generally used for screening of newborn hearing and the report uses the following definition:

  “It is recorded as five to seven waves. Waves I, III and V can be obtained consistently in all age groups. Waves II and IV appear less consistently. The latency of each wave (time of occurrence of the wave peak after stimulus onset) increases, and the amplitude decreases with reductions in stimulus intensity or loudness. Developmental change occurs in the latency of the various waves; latency decreases with increasing age, with the earliest waves reaching mature latency values earlier in life than the later waves. It is used as an audiometric test, providing information about the ability of the peripheral auditory system to transmit information to the auditory nerve and beyond, and it is used in the monitoring of central nervous system pathology. It is not accurate in predicting neurologic recovery and outcome. (Behrman, Kliegman, Jenson, 2000).”

Statistical Analyses

In this report, two statistical analyses were performed when there was sufficient data: the number needed to treat (NNT) and receiver operating characteristics (ROC) curve.

Meta-analyses of Diagnostic Test Performance

Meta-analyses were performed to quantify the transcutaneous bilirubin measurements where the data were sufficient. We used three complementary methods for assessing diagnostic test performance: summary receiver operating characteristics (SROC) analysis, independently combined sensitivity and specificity values, and meta-analysis of correlation coefficients. All meta-analysis schematic and statistics were reported in the Meta-Analyses chapter.

Meta-analysis of Correlation Coefficients

Correlation coefficients measure the correlation of one diagnostic test to another, but do not provide any information about the clinical utility of the diagnostic test. Also, they are inadequate for measuring the accuracy of a test in estimating serum bilirubin levels for two reasons. Although correlation coefficients (r) measure the association between transcutaneous bilirubin and “standard” serum bilirubin measurements, the correlation coefficient is highly dependent on the distribution of serum bilirubin in the study population selected. Second, correlation measures ignore bias and measure relative rather than absolute agreement (Bland and Altman, 1986). Nonetheless, a large number of studies continued using the correlation coefficient but only some of them did sensitivity and specificity analyses for examining the test accuracy.

Meta-analyses were performed to compare transcutaneous bilirubin (TcB) against laboratory essay of total serum bilirubin (TSB) measurements, where the data were sufficient. These studies provided the correlation coefficient for TcB vs. TSB, but several analytic issues arose regarding potential duplication of information. To address these issues, the following rules were developed and applied:

Rule 1. Whenever a study provided not only data for the whole population but also data splitting the population into subgroups of patients, only the one entry about the whole population was used to avoid duplicating the data. However, if the subgroups of patients did not meet our inclusion criteria described before, the whole population data was dropped and only the qualifying subgroup data was entered.

Rule 2. When a study gave data for separate conditions on the same subjects and then also an aggregate of all conditions, only aggregate data was used to avoid duplication.

Rule 3. Studies were indexed by different measurement sites or measurement protocols (metrics). When several different metrics were provided for a specific comparison, all were retained and entered in the respective metric-specific subgroup analyses. However, only one metric was retained for the overall synthesis to avoid duplication. The order of preference was: forehead, sternum, other sites (preferring the most often used site).

We combined the correlation coefficient across studies using a random-effects model. The random-effects model incorporates both within and between study variations. Such models yield more “conservative” (i.e. wider) confidence intervals. Subgroup analyses on different factors that might affect the correlation coefficient of the test and gold standard were also performed.

Since using the repeated measurements in one subject would overestimate the correlation coefficient, the number of measurements was replaced by the number of patients whenever it was available. We report 95 percent confidence intervals for all estimates.

Software and Statistical Analyses

Statistical analyses using the SROC curve method and combining sensitivity and specificity using the random-effects model was performed using “Meta-Test” version 0.6. The computer program was developed by the EPC director (Dr. Lau) and is available to the public. Meta-analysis of correlation coefficient using a random-effects model was performed using Comprehensive Meta-Analysis™ version 1.0.23 (Biostat™ Inc.). We report 95 percent confidence intervals along with all estimates.

Analyses for Kernicterus Case Reports

Where data was available, we extracted serum bilirubin measurements on admission or at the time of diagnose of kernicterus. We also extracted the peak serum bilirubin levels. From these data, we created histograms on the distribution of serum bilirubin levels. We performed subgroup analyses of these cases to see if there are factors related to the patterns of serum bilirubin distribution.

Case Reports of Kernicterus

Demographics of Kernicterus Cases

Table 3.1 Feeding and Gender Status in Newborns Who (more...)

Table 3.1 Feeding and Gender Status in Newborns Who Had Kernicterus

	Sex	Feeding			Total
		Breast	Formula	Unknown
Newborns with idiopathic jaundice with kernicterus	M	7	1	6	14
	F	2	0	4	6
	?	11	0	4	15
Subtotal		20	1	14	35
Newborns with comorbid factors with kernicterus	M	8	0	17	25
	F	1	1	19	21
	?	6	0	36	42
Subtotal		15	1	72	88
Total		35	2	86	123

Table 3.2 Racial or Ethnicity Compositions in Newborns (more...)

Table 3.2 Racial or Ethnicity Compositions in Newborns Who Had Kernicterus

	Race/Ethnicity
	AA	AF	Chinese	Malay	White/NE	Pal	Unknown	Total
Newborns with idiopathic jaundice with kernicterus	0	0	6	2	9	1	17	35
Newborns with comorbid factors with kernicterus	4	2	14	1	6		61	88
Total	4	2	23	3	15	1	78	123

AA = African American; AF = African; NE = Northern European; Pal = Palestinian

Articles we found span from 1955 to 2001 with a total of 123 kernicterus cases. Twelve cases in two studies were reported before 1960. However, some studies reported cases that spanned almost two decades. Data on subjects' birth year were only reported in 55 cases. Feeding status, gender, racial background and ethnicity were not noted in the majority of the reports. Of studies that reported the characteristics, almost all of the subjects were breastfed and the majority were males (see Tables 3.1 and 3.2).

Geographic Distribution of Kernicterus Cases

Figure 3.1 Geographic distribution of kernicterus cases (more...)

   Figure 3.1 Geographic distribution of kernicterus cases

The 28 case reports with a total of 123 cases were from 14 different countries. The number of kernicterus cases in each study ranged from one to 12. The distribution of kernicterus cases geographically is shown in Figure 3.1.

Kernicterus has been defined by pathological findings, acute clinical findings and chronic sequelae (like deafness or athetoid cerebral palsy). Because of the small number of subjects, all definitions of kernicterus have been tabulated together in the analysis. Exceptions will be noted in the following discussion.

Kernicterus Cases with Unknown Etiology

Figure 3.2 Distribution of Peak Total Serum Bilirubin (more...)

   Figure 3.2 Distribution of Peak Total Serum Bilirubin Level in Term Newborns with Idiopathic Jaundice Who Had Kernicterus

Among infants ≥ 34 weeks or who had birthweight ≥ 2,500 grams and had no known explanation for kernicterus, there were 35 infants with peak bilirubin ranging from 22.5 mg/dl to 54 mg/dl. Fifteen had no information on their gender; 14 were males and six were females. Fourteen provided no information on feeding; 20 were breast-fed and one was formula-fed. The majority of the infants with kernicterus had bilirubin greater than 25 mg/dl: 25 percent of the kernicterus cases had peak total serum bilirubin (TSB) level up to 29.9 mg/d; 50 percent of the kernicterus cases had peak TSB up to 34.9 mg/dl (Figure 3.2). There was no association between the bilirubin levels and birthweights.

Table 3.3 Summary of 35 Case Reports of Term Infants (more...)

Table 3.3 Summary of 35 Case Reports of Term Infants with Idiopathic Jaundice Who Had Kernicterus

	Definitions of KI	N	Mean Peak TSB±SD, mg/dl (range)	Mean BW±SD†, g (range)	Sex
Term infants w/ idiopathic Jaundice	Acute phase of KI w/o long term follow-up	9	33.4±11.4 (22.5–54.0)	3496±535 (2700–4500)	3 Females; 2 Males; 4 ?
	Acute phase of KI but normal long-term follow-up	5	32.3±3.6 (27.0–36.0)	3088±575 (2400–3742)	5?
	Chronic KI sequelae	17	37.0±5.9 (29.5–49.7)	3271±442 (2700–4280)	1 Females; 11 Males; 6 ?
	Death	4‡	36.9±11.0 (23.0–48.0)	2902±503 (2324–3357)	2 Female; 1 Male; 1 ?
Total		35	35.4±8.0 (22.5–54.0)	3251±512 (2324–4500)	6 Females; 14 Males; 15 ?

KI = Kernicterus

†

Contain some missing data

? = Unknown

‡

One infant had acute phase of KI and chronic KI sequelae; then died at age of 5 months. Necropsy was not done. One infant had acute phase of KI and died on the 5^th day of life; necropsy found yellow-staining in the brain nuclei.

Table 3.4 Summary of 88 Case Reports of Term Infants (more...)

Table 3.4 Summary of 88 Case Reports of Term Infants with Comorbid Factors Who Had Kernicterus

		Diagnosis of KI
		Clinical				Total (N)	Mean Peak TSB±SD, mg/dl (range)	Mean BW±SD†, g (range)	Sex
	Comorbidity	Acute phase w/o follow-up	Acute phase but normal follow-up	Chronic sequelae	Autopsy
Term infants with cormobid factors	ABO incompatibility	6	0	12	1‡	19	31.6±8.2 (19.0–51.0)	3118±680 (2270–4313)	6 Females; 11 Males; 2 ?
	Rhesus incompatibility	3	2	27	1	33	32.1±7.1 (17.7–46.0)	3063±387 (2300–3969)	5 Females; 4 Male; 24 ?
	G6PD deficiency	10	0	1	2	13	31.8±8.5 (23.0–50.0)	3353±437 (2700–4100)	2 Males; 11 ?
	Sepsis or Infections	3	1	4	5	13	31.8±9.9 (14.5–47.8)	3368±586 (2580–4360)	7 Females; 4 Males; 2 ?
	Multiple conditions	3	0	4	3	10	29.1±16.1 (4.0–49.2)	2913±750 (1780–3686)	3 Female; 4 Males; 3 ?
Total		25	3	48	12	88	31.6±9.0 (4.0–51.0)	3155±534 (1780–4360)	20 Females; 25 Males; 42 ?

KI = Kernicterus

? = Unknown

†

Contain some missing data

‡

This infant had acute phase of KI and chronic KI sequelae; then died at age 19 months.

Figure 3.3 Cumulative Percentage of Peak Total Serum (more...)

   Figure 3.3 Cumulative Percentage of Peak Total Serum Bilirubin Levels in Term Newborns with Idiopathic Jaundice Who Had Kernicterus

Four infants died (Boon, 1957; Ebbesen, 2000). Four infants who had acute clinical kernicterus had normal follow-up at 3 to 6 years by telephone (Harris, Bernbaum, Polin, et al., 2001). One infant with a peak bilirubin of 44 mg/dl had a flat brain stem auditory evoked response (BAER) initially but normalized at 2 months of age; that infant had normal neurologic and developmental examinations at 6 months of age (Johnson, 1991). Ten infants had chronic sequelae of kernicterus when followed up between 6 months and 7 years of age. Seven infants were noted to have neurologic findings consistent with kernicterus but the ages of diagnosis were not provided. Nine infants had diagnosis of kernicterus with no follow-up information provided. To summarize, 11 percent of this group of infants died; 14 percent survived with no sequelae and at least 46 percent of infants had chronic sequelae (Table 3.3). The distribution of peak TSB levels was higher when only infants who died or had chronic sequelae were included (Figure 3.3).

Kernicterus Cases with Comorbid Factors

Figure 3.4 Distribution of Peak Total Serum Bilirubin (more...)

   Figure 3.4 Distribution of Peak Total Serum Bilirubin Level in Term Newborns with Comorbid Factors Who Had Kernicterus

In the 88 term and near-term infants diagnosed with kernicterus and who had hemolysis, sepsis and other neonatal complications, the bilirubin level ranged from 4.0 mg/dl to 51.0 mg/dl. Forty-two provided no information on their gender; 25 were males and 21 were females. Seventy-two had no information on feeding; 15 were breast-fed and 1 was formula-fed. The majority of those infants with kernicterus had bilirubin greater than 20 mg/dl: 25 percent of the kernicterus cases had peak total serum bilirubin (TSB) level up to 24.9 mg/dl; 50 percent of the kernicterus cases had peak TSB up to 29.9 mg/dl (Figure 3.4). In this group, there was no association between the bilirubin levels and birthweight.

Figure 3.5 Cumulative Percentage of Peak Total Serum (more...)

   Figure 3.5 Cumulative Percentage of Peak Total Serum Bilirubin Levels in Term Newborns with Comorbid Factors Who Had Kernicterus

Five infants without clinical signs of kernicterus were diagnosed with kernicterus by autopsy. Eight infants died of kernicterus. One infant was found to have a normal neurological exam at 4 months of age (Hanko, Lindemann, Hansen, 2001). Another one with galactosemia and a bilirubin of 43.6 mg/dl who had acute kernicterus was normal at 5 months of age (Ebbesen, 2000). Forty-nine patients had chronic sequelae ranging from hearing loss to athetoid cerebral palsy; the follow-up age reported ranged from 4 months to 14 years. Twenty-one patients were diagnosed with kernicterus with no follow-up information. To summarize, not including the autopsy diagnosed kernicterus, 10 percent of these infants died (8/82); two percent of infants were found to be normal at 4 to 5 months of age and at least 60 percent had chronic sequelae (Table 3.3). The distribution of peak TSB levels was slightly higher when only infants who died or had chronic sequelae were included (Figure 3.5)

Evidence Associating Bilirubin Exposures with Neurodevelopmental Outcomes in Healthy Term or Near-Term Infants

This section examines the evidence associating bilirubin exposures with neurodevelopmental outcomes primarily in subjects without kernicterus. Studies that were designed specifically to address the behavioral and neurodevelopmental outcomes in healthy infants greater or equal to 34 weeks of gestation will be discussed first, as that group of infants is the subject of interest in this review. With the exception of the results from the Collaborative Perinatal Project (CPP, see below), the rest of the studies comprising mixed subjects (preterm and term, diseased and non-diseased) will be categorized and discussed by outcome measures which include behavioral and neurologic outcomes, hearing impairment including sensorineural hearing loss, and intelligence measurements. CPP, with the largest number of subjects (54,795) in this review, has generated many follow-up studies with smaller number of subjects and those studies will be discussed together in a separate section.

Table 3.5 Effect of Serum Bilirubin Levels on Neurological (more...)

Table 3.5 Effect of Serum Bilirubin Levels on Neurological or Behavioral Outcomes in Healthy Infants (GA ≥ 34 weeks of gestation)

Author, Year, UI #	Subjects N (Control)	Peak bilirubin Level (range) mg/dl	Outcomes	Confounders	Quality
Rapisardi, 1989 89213178	29	8.8 (2.3–18.5)	@50–71 hrs of life: CV of cry frequencies, as an indicator of “neurophysiological functioning”	PhotoRx, Age	B
			The correlation between the CV of F1 during phonation (PF1) and level of bilirubin was significant (r=0.37, p<0.05), suggesting that more variability in F1 was related to higher concentrations of bilirubin.
Escher-Gräub, 1986 87136731	76 ( 401)	11.7–14.6	@day 4–5 of life: BNBAS	None	B
			Jaundiced infants not requiring PhotoRx showed a less mature behavioral organization than the infants of the control group.
Soorani-Lunsing, 2001 11726727	20 (20) b	18.3 (>12.9)	@3–8 days, 3 and 12 months of life: Neurological exams	PhotoRx; No TSB level available for control group	C
			Infants with TSB > 12.9 mg/dl had significantly higher prevalence of mild neurologic dysfunction.Controlling for other factors, the odds of developing mild neurologic dysfunction were 6.62 (95%CI 1.59–27.51), 3.75 (95%CI 1.03–14.53), and 9.47 (95%CI 1.67–53.65) at 3–8 days, 3 and 12 months of life respectively.
			@12 months of life: Behavioral Outcomes
			It didn't differ between study and control subjects according to the parental questionnaire.
Grimmer, 1999 99345621	16 (18)	23.0 (20.1–28.69)	@61–187 months: Neurological Outcome (Touwen sub-scales)	PhotoRx	C
			None of the children studied had medical problems or overtly abnormal neurological findings.
			Cluster profiles did not differ significantly between formerly jaundiced children and controls, except for worse scores of formerly jaundiced children on the choreiform dyskinesia scale (P=0.028).
Paludetto, 1986 86296856	17 (17) a	10.6 (8.4–14.3 )	@3rd, 4th days of life, and 1 month of age: BNBAS	Not all controls had TSB measured	C
			There was no difference in any behavioral item of the BNBAS score at every observation.

a

Only 4 controls had bilirubin measurement.

b

No TSB level available for control group

BET = exchange transfusion; PhotoRx = phototherapy; BNBAS = Brazelton Neonatal Behavioral Assessment Scale; CV = coefficient of variation; F1 = formant frequency 1

Rapisardi, Vohr, Cashore, et al. (1989) studied and divided 29 healthy full-term infants into three different groups according to bilirubin level (less than or equal to 6 mg/dl, 6 to 12 mg/dl, greater than 12 mg/dl). Some of these infants received phototherapy. No further details were provided. There were no differences in mean values of acoustic measures (cry fundamental frequency, resonant frequency, energy and duration). There was a statistically significant difference between the three groups in the coefficient of variation (CV) of formant frequency-1 (frequency of resonance as determined by the filtering of vocal tract that is affected by the cross-sectional area and motor innervation of the supraglottal airway). Infants with the highest levels of bilirubin showed a greater CV of formant frequency-1 during phonation than infants with lower or medium levels of bilirubin (P<0.05). The authors suggested that “an increase in the variability of the first formant could be due to an increase in instability of the neural mechanisms controlling the cross-sectional area of the supraglottal airways.”

Escher-Graub and Fricker (1986) compared 76 term infants with maximum bilirubin of 11.7 to 14.6 mg/dl to 401 control term infants with bilirubin of less than 5.8 mg/dl using the Brazelton Neonatal Behavioral Assessment Scale (BNBAS). The BNBAS was done either on the fourth or fifth day of life. These term infants had no evidence of hemolysis and did not require phototherapy. The results showed that the two groups differed significantly (p<0.01 to 0.05) on 12 of the 28 scale items. Five of the six dimensions of newborn behavior were affected: habituation, orientation, motor performance, and regulation of state and autonomic stability. “Infants of the jaundice group showed a less mature behavioral organization than the infants of the control group.”

Soorani-Lunsing, Woltil, and Hadders-Algra (2001) assessed the neurologic condition of 20 healthy non-hemolytic term newborns with peak bilirubin of 13.6 to 26 mg/dl and 20 clinically non-jaundiced control infants. Ten of the jaundiced infants received phototherapy. A standardized neurologic examination similar to Prechtl's was carried out at the age of 12 months. “The examiner occasionally knew group membership of the children, but was never aware of the degree of hyperbilirubinemia of the study children.” Moderate hyperbilirubinemia was associated with a significant increase in minor neurologic dysfunction throughout the first year of life: newborn, 3 months of age and 12 months of age. Examples of minor neurologic dysfunction were mild abnormalities in postural behavior, the presence of high frequency tremors, mild deviancies in muscle tone regulation and mild asymmetries in infantile reactions and tendon reflexes. The odds of developing minor neurologic dysfunction at 12 months were 9.47 (95% CI 1.67 – 53.65). Neurologic outcome at 12 months was strongly related to the peak bilirubin level. The authors suggested that “total serum bilirubin levels of 335 micromole/liter (19.6 mg/dl) should be avoided.” The authors also noted that the behavior of the study and control children at age 12 months did not differ according to the parental questionnaires

Grimmer, Berger-Jones, Buhrer, et al. (1999) identified 29 non-hemolytic healthy term newborns with a total bilirubin between 20 and 30 mg/dl based on ten-year records of a referral center: Sixteen were available for a standardized, age-adjusted neurological examination (according to Touwen) at 61 to 187 months of age. All of them received phototherapy. Compared to 18 case controls with bilirubin less than 12 mg/dl, jaundiced children scored significantly (P=0.028) worse only on the choreiform dyskinesia scale. There was no difference in the other neurological subscales (fine and gross motor, sensorimotor apparatus and others).

Paludetto, Mansi, Rinaldi, et al. (1986), also using the Brazelton Neonatal Behavioral Scale, studied 17 jaundiced term 3-day-old infants (bilirubin of 8.4 to 14.3 with a mean of 10.6 mg/dl) not treated with phototherapy and 17 clinically non-jaundiced matched subjects; 14 four-day-old infants (8.4 to 12.9 with a mean of 10.4 mg/dl) and 14 clinically non-jaundiced matched subjects; 10 “ex-jaundiced” 1-month-old infants (9.1 to 15.9 with a mean of 11.5 mg/dl) and 10 clinically non-jaundiced matched subjects. No differences were found. The authors concluded “that moderate levels of hyperbilirubinemia not treated with phototherapy do not influence neonatal behavior.” Bilirubin levels were not measured in the control group.

The following group of studies consists of subjects who, in addition to healthy term newborns, might include newborns younger than 34-week gestation, neonatal complications such as sepsis, respiratory distress, hemolytic disorders and other factors. Nevertheless, some of the conclusions drawn might be applicable to a healthy term population. In these studies, greater emphasis will be placed on the reported results for the group of infants who were greater or equal to 34-week gestation or have birthweight greater or equal 2,500 grams.

Studies Measuring Behavioral and Neurological Outcomes in Infants with Hyperbilirubinemia

Table 3.6 Effect of Serum Bilirubin Levels on Neurological (more...)

Table 3.6 Effect of Serum Bilirubin Levels on Neurological or Behavioral Outcomes in All Infants

Author, Year, UI #	Subjects N (Control)	Peak bilirubin Level (range) mg/dl	Outcomes	Confounders	Quality
Vohr, 1989 89328704	23 a (27)	14.3±2.8 (10–20)	On day 1 to 2: BNBAS	PhotoRx	A
Vohr, 1990 90339225			Jaundiced infants had significantly lower BNBAS scores in every behavioral item except autonomic stability.
			Significant correlations were found in increased levels of TSB with decreased scores on the individual BNBAS items. After controlling for PhotoRx by partial correlations, most correlations remained significant except for state regulation and autonomic stability.
Valaes, 1980 81031849	44 c (445)	47% 16–20	@61–81 months of age: Visual-Motor integration test (VMI)	Prematurity; BET; Phenobarbital	B
		29% 20–25	Degree of jaundice was found not to be associated with neurologic scores.
		24% >25
Hyman, 1969 69132491	405 b	≥ 15	@4 years of age: CNS abnormalities	BET; streptomycin Rx; BW; Selection bias	B
			No significant difference in incidence of CNS abnormalities among infants exposed to different levels of bilirubin below 20 mg/dl. However, with TSB> 20 mg/dl, incidence of CNS abnormalities increased sharply.
Yilmaz, 2001 21125314	87 g	11% 23-20	@10–72 months: Neurological exam	PhotoRx; BET; Duration of jaundice	C
		49% 20–23.9	Subjects with maximum TSB levels < 20 mg/dl didn't show any neurological abnormalities, whereas 9.3% of the subjects with TSB 20–23.9 mg/dl (N=43) had only dysconjugate gaze and 17.6% of the subjects with TSB ≥ 24 mg/dl had neurological manifestations.
		39% ≥ 24
Agrawal, 1998 99232291	30 f (25)	22.4±2.7	@1 year of age: DDST	PhotoRx; BET; lost to follow-up	C
			Neurological development was normal in all infants with TSB 15–20 mg/dl, in 89% of infants with TSB 21–25 mg/dl, and in 67% infants with TSB > 25 mg/dl.
Wolf, 1997 98030356	45 e @4 months	28.5±6.3	@4 months of age: Infant Motor Screening (IMS)	PhotoRx; BET; Prematurity; Illness	C
Wolf, 1999 99156393	35 @1 year		Linear correlation showed an association of maximum TSB and test rating at 4 months (r=0.32, p<0.05) and test scores at 4 months (r=0.44, p<0.03)
			Mean TSB in the normal IMS group was 27.3±5.3 mg/dl; in the suspect and abnormal IMS group the mean TSB was 28±4 mg/dl and 33.7±10.3 mg/dl, respectively. These differences were not significant (p=0.06)
			@1 year of age: Bayley (BSID - PDI)
			8 (23%) term infants, with mean TSB 33.4 mg/dl, had abnormal and suspect BSID and clinical diagnosis.
			27 (77%) term infants, with mean TSB 26.5 mg/dl, scored normal on the BSID.
			“Correlation between Bayley motor score and TSB was 0.59 (p<0.01)”
Chen, 1995 96017937	72 (22)	39% 10–15	@1 year of age: DDST and Neurological exams	PhotoRx; BET	C
		32% 15–20	None (0%) of the infants w/ TSB 10–20 mg/dl showed any abnormality in the DDST and neurological examination.
		29% >20	Among infants w/ TSB > 20 mg/dl, 4 (22%) were abnormal in gross motor and fine motor skills of DDST.
Grunebaum, 1991 91231060	46	12.06±2.6	@age 31.1±16.6 months: Growth, Neurological exam, DDST	PhotoRx; Lost to follow-up	C
			Growth and neurological examination were normal in all infants
			2 (4%) had abnormal DDST, but all had normal DDST repeated 1 month later
Holmes, 1968 68310019	63 (17) d	6.4–24 (46% 6.4–12.5; 54% 15–24)	@4 years 7 months - 7 years 8 months: Neurologic examination;s Motor development (Oseretsky test); Audiometric exams	Age; BET; Streptomycin	C
			In normal FT babies, mild to moderate hyperbilirubinemia is not associated with findings on Oseretsky motor development test.
			All had normal hearing, including subset with streptomycin Rx.

a

60% ABO incompatibility [56% of controls has ABO incompatibility as well]

b

Including 10% preemies and some hemolytic diseases

c

29% ABO incompatibility; 2% G6PD deficiency; 69% unknown cause of jaundice

d

All causes of jaundice were included in the study

e

43% term+ 57% preterm (mean GA 36.6±3.5 wk.). 44% LBW; 16% ABO incompatibility; 16% sepsis; 6% congenital syphilis

f

26.7% had ABO incompatibility and 63.3% had idiopathic hyperbilirubinemia

g

44% ABO incompatibility, 16% Rh incompatibility; 40% non-hemolytic jaundice

Vohr, Lester, Rapisardi, et al. (1989) or Vohr, Karp, O'Dea, et al. (1990) divided 50 term infants into either a group with low bilirubin of <8mg/dl or a group with moderate bilirubin of 10 mg/dl to 20 mg/dl. The Brazelton Neonatal Behavioral Assessment Scale (BNBAS) was administered between the second day and fifth day of life (mean ± S.D.; 65± 22 hours in moderate bilirubin group; 57 ± 13 hours in low bilirubin group, differences not significant). Some of these infants had positive Coombs test. None of the patients in the low bilirubin group received phototherapy. Seventeen of 23 patients in the moderate group received phototherapy. Partial correlation analysis controlling for phototherapy revealed that increased bilirubin concentration correlated negatively with BNBAS in state range (r= -0.28), orientation clusters (r= -0.42), and individual Brazelton items that involve auditory processing (r = -0.39 to -0.58).

Valaes, Kipouros, Petmezaki, et al. (1980), in a selective analysis of his maternal phenobarbital study, studied 44 term and preterm infants with peak bilirubin greater than 16 mg/dl. This group consisted of infants with ABO incompatibility, G6PD deficiency and nonspecific hyperbilirubinemia. Nine of them received exchange transfusion. These subjects had neurologic examination, Visual-Motor Integration test, Draw-a-Woman test and Georgas Vocabulary test at 61 to 82 months of age. No relationship between the level of bilirubin (from 16 mg/dl to >25 mg/dl) and test results was found.

Hyman, Keaster, Hanson, et al. (1969), from an original sample of 742 infants with hemolytic disease or indirect hyperbilirubinemia of ≥ 15 mg/dl (49 of whom died in the neonatal period), followed and studied 405 infants until 4 or more years of age. Analysis of the original group of 742 neonates as compared with the study group of 405 subjects showed that 70 percent of those with bilirubin ≥ 20 mg/dl as compared with 55 percent of those with bilirubin < 20 mg/dl were followed. Three hundred sixty-eight were term. Three hundred forty-eight had Rh and 48 had ABO incompatibility. Three hundred fifty-one received streptomycin and exchange transfusion. There was no significant difference in incidence of CNS abnormalities (sensorineural hearing loss, athetosis, strabismus, seizures, minimal cerebral dysfunction) among infants exposed to different levels of bilirubin below 20 mg/dl. Infants exposed to bilirubin above 20 mg/dl had a 20 percent incidence of CNS abnormalities vs. 7 percent incidence in the group with exposure below 20 mg/dl.

Yilmaz, Karadeniz, Yildiz, et al. (2001) studied 87 (out of a possible 755) children who were term newborns with jaundice. There were no control subjects in this study. Fourteen subjects had Rh and 38 had ABO incompatibility and 35 had non-hemolytic jaundice. 11.4 percent of subjects were treated with phototherapy; 11.5 percent were treated with phototherapy and exchange transfusion. Follow-up examinations were done from 10–72 months. Neurological dysfunction was observed in 11.5 percent (10/87) of cases. These patients had significantly higher peak bilirubin levels (26.1 ± 4.9 mg/dl vs. 23.2 ± 3.4 mg/dl) and longer duration of jaundice (6.1 ± 1.9 days vs. 4.7 ± 2.4 days) than those who were neurologically normal (p<0.05). There was no statistically significant relationship between the etiology of hyperbilirubinemia and the neurological prognosis.

Agarwal, Narayan, Kumari, et al. (1998) studied 30 term newborns with bilirubin > 15 mg/dl. Ten percent had Rh incompatibility, 26.7 percent had ABO incompatibility and 63.3 percent had idiopathic hyperbilirubinemia. Fourteen received exchange transfusion; the rest had phototherapy. Two subjects had kernicterus. At one year of age, screening by DDST showed that development was normal in 100 percent of infants with bilirubin of 15–20 mg/dl, 88.9 percent of infants with bilirubin of 21–25 mg/dl, and 66.7 percent of infants with bilirubin >25 mg/dl. No further statistical details were provided.

Wolf, Beunen, Casaer, et al. (1997) and Wolf, Wolf, Beunen, et al. (1999) studied the effect of maximum serum bilirubin in excess of 23.4 mg/dl on the neurodevelopmental outcome of 50 Zimbabwean neonates at four months and at one year of age. All received phototherapy, twenty-six were preterm infants, seven had sepsis without meningitis and seven received exchange transfusion. Out of 50 infants with serum bilirubin greater than 23.4 mg/dl, forty-five were evaluated at 4 months using the infant motor screen; six infants scored abnormal, six scored suspect. Forty-three were followed until one year of age, two died and five were lost to follow-up. Eleven scored within abnormal range on the Bayley Scales of Infant Development at 1 year of age and five developed the choreo-athetoid type of cerebral palsy. All of those infants were either preterm babies or had hemolytic diseases. All term infants without hemolysis and with bilirubin levels between 23.4 mg/dl and 29.2 mg/dl were normal at one year of age.

Chen and Kang (1995) studied Visual Evoked Potentials (VEP) in 72 term infants with hyperbilirubinemia and 22 control infants in four sessions for 8 weeks after birth. Reasons for jaundice were not stated. Infants with bilirubin >15 mg/dl were treated with phototherapy. Those with bilirubin ≥ 20 mg/dl also had exchange transfusion. The levels of maximum bilirubin found positively related to the wave latencies of first VEP. Within 8 weeks after birth, the wave latencies were significantly prolonged in infants in the severe (20 to 29.8 mg/dl) and moderate (15 to 18.6 mg/dl) groups than in the controls (4.8 to 8 mg/dl). Neurological examination and the Denver Developmental Screen Test (DDST) were performed on 62 of the term infants with neonatal hyperbilirubinemia at 1 year of age. Of the sixty-two, 24 infants belonged to the low (11.6 to 14.4 mg/dl), twenty to the moderate and 18 to the severe groups. None of the infants in the low and moderate groups showed any abnormality in the neurological examination and DDST. In the severe group, four of the 18 infants were abnormal in gross motor and fine motor skills of the DDST. One of these four infants showed general hypotonia, and his VEP's showed markedly low amplitudes and prolonged peak latencies for 8 weeks. Statistical analyses of the neurodevelopmental findings were not presented.

Grunebaum, Amir, Merlob, et al. (1991) enrolled 60 term and pre-term infants at 3 weeks of age into a breast milk jaundice study. Most infants were seen at out-patient clinic during the end of the second week of life (25 percent) or during the third week of life (28.3 percent). 13.3 percent were seen at an age of 5 weeks after birth or later. Infants, in whom the jaundice resolved within 3 weeks, were excluded from the study. Range of bilirubin at presentation was 4.2 mg/dl to 16.8 mg/dl. Fifteen infants received phototherapy. In 14 infants there was a brief interruption of nursing (in 13 cases for 24 hours and in one case for 48 hours). Forty-six subjects had neurodevelopmental examination between 10 and 60 months (mean age 31.1 ± 16.6 months). Their neurological examination was normal. All children had normal DDST except for two children. One had a delay in language ability and the other a delay in fine motor ability. A repeat DDST one month later was normal in both children.

Holmes, Miller, and Smith (1968) studied 63 children with neonatal bilirubin of 6.4 mg/dl to 24 mg/dl and 17 control children without neonatal clinical jaundice. These 80 children were selected out of a total of 195 available subjects. Only babies ≥ 2.5 Kg and who received more than one bilirubin determination were included in the experimental group. All causes of jaundice were included in the study. One case was excluded because he and his father displayed bizarre behavior. Another was excluded because he was mentally retarded; his bilirubin was in “low” range of 6–12 mg/dl. Eleven of 63 received exchange transfusion and 13 of 63 received streptomycin. A modified Oseretsky motor test and audiometry were performed in these children from 4 years 7 months to 7 years 8 months. The authors found no relationship between hyperbilirubinemia and the Oseretsky motor development test. Authors further noted that none of the subjects developed hearing impairment.

Effect of Bilirubin on Brainstem Auditory Evoked Potential (BAEP)

Table 3.7 Effect of Serum Bilirubin Levels on Auditory (more...)

Table 3.7 Effect of Serum Bilirubin Levels on Auditory Brainstem Response & Hearing deficit in All Infants

Author, Year, UI #	Subjects N (Control)	Peak Bilirubin Level (range) mg/dl	Outcomes	Confounders	Quality
Tan, 1992 92139200	30 (31)	16.7	@0–3 days of Rx: BAER	PhotoRx	A
			In PhotoRx group, there was significant difference between the preexposure and postexposure values of latency V and IPL III–V and I–V.
			The correlation between the bilirubin levels and the BAER values during PhotoRx was significant r=0.967 and p<0.001 for latency V; r=0.999 and p< 0.001 for IPL I–V; and r=0.963 and p< 0.001 for IPL III–V.
Valaes, 1980 81031849	44 d (445)	47% 16–20	@61–81 months of age: Sensorineural hearing defect	Prematurity; BET; PB exposure	B
		29% 20–25	Sensorineural hearing defect was significantly more common in children with moderate (bilirubin >12 mg/dl) or marked (bilirubin > 16 mg/dl) neonatal jaundice than in the children with slight or no jaundice (9.2% vs. 1.1%).
		24% >25
Hyman, 1969 69132491	351 b	≥ 15	@4 years of age: Hearing loss	BW	B
			Infants with TSB ≥ 20 mg/dl (N=108) had significantly higher prevalence of sensorineural hearing loss (10%), than infants with TSB < 20 mg/dl (N=243) (2%).
Johnston, 1967 67081206	129 c (95)	>20	@5–6 years of age: Audiological assessment	BET; Prematurity	B
			Infants with TSB > 20 mg/dl had significantly more sensorineural hearing loss (7%), compared to infants with “low” bilirubin levels (0%).
Suresh, 1997 98374505	42 g	Neonatal: 26.6±5.8	@2 months to 21 years old: Hearing loss	Rx for hyperbiliru-binemia; Prematurity	C
		Postnatal: 20.5±5.5	Hearing was normal in 94% patients. No sensorineural hearing loss (reported in 36 patients).
Boo, 1994 95018454	28 e (100)	21.6±7.5 [18.4±7.4]	@11.6±3.9 days: Hearing loss	PhotoRx	C
			16% (13/83) of subjects with bilirubin < 19.8 mg/dl vs. 33% (15/45) of subjects with bilirubin > 19.8 mg/dl had hearing loss. The difference was not significant.
			Logistic regression analysis showed that the significant risk factors associated with hearing loss in the term jaundiced neonates were exchange transfusion and age of onset of hyperbilirubinemia.
Esbjorner, 1991 91281442	9 a	20.3 (16.4–25.6)	@4 (3.5–6) days, post Rx: ABR	PhotoRx	C
			Significant correlation between the latencies to wave I, III, V, IPL I–V and the reserve albumin concentration (r=-0.89, p<0.01), but no significant correlation between the latency to wave V and the total bilirubin concentration was found (r=0.02, p=0.95)
Rosta, 1971 72185033	84 f	≥ 15	@8 years of age: Hearing loss	BET, Streptomycin Rx	C
			Severe sensorineural hearing loss were detected in 5 (6%), a mild to moderate one in 9 (11%) patients. “The above incidence was surpassing the frequency of similar alterations in the normal population.”

a

One infant had ABO incompatibility

b

Including 10% preemies and some hemolytic diseases

c

71% iso-immunization; 20% ABO incompatibility

d

29% ABO incompatibility; 2% G6PD deficiency

e

Non-physiologic causes of jaundice were known in 39 subjects but details were not provided

f

“Nearly equal distribution Rh and ABO immune-hemolysis, and hyperbilirubinemia of unknown origin”

g

All patients, age 2 months to 21 years old, were Crigler-Najjar syndrome type 1, resulting in life long jaundice.

Table 3.8 Mean Latencies and Interpeak Latencies of (more...)

Table 3.8 Mean Latencies and Interpeak Latencies of Brainstem Auditory Evoked Response (BAER) in All Infants (GA ≥ 34 weeks)

Author, Year, UI #	Sample N (Control)	Peak Bilirubin Level (range) mg/dl	Mean Latency (msec)			Mean Interpeak (msec)			Confounders	Quality
			I	III	(IV),V	I–III	III–V	I–V
Tan, 1992 92139200	30 (31)	16.7	1.75 [1.77]	4.56 [4.54]	7.00 [6.75]	2.80 [2.77]	2.44 [2.25]	5.25 [5.00]	None	A
Vohr, 1989 89328704	25 a (25)	13.5±2.6 (10.2–19)	1.8 [1.8]	4.5 [4.3]	7.5 [7.3]	2.7 [2.6]	3.0 [3.0]	5.8 [5.5]	PhotoRx	A
Vohr, 1990 90339225
Gupta, 1998 98355379	15 (15)	21.98 (18.4–26.4)	-	-	-	2.94 [2.48]	3.65 [3.20]	6.00 [5.12]	sedation; hemolytic disease; (Unknown Rx effect)	B
	30 (15)	16.6 (15.8–18.0)	-	-	-	2.65 [2.48]	3.34 [3.20]	5.83 [5.12]
	45 c (15)	18.4 (15.8–26.4)	-	-	-	2.88 [2.48]	3.50 [3.20]	5.92 [5.12]
Sabatino, 1996 97171477	48 (42)	4–25.8	1.9 [1.8]	4.8 [4.0]	8.1 [7.0]	3.0 [2.3]	3.1 [2.9]	6.1 [5.2]	None	B
Perlman, 1983 84041257	24 b (19)	18.7±2.7	-	-	-	2.77 [2.67]	3.20 [3.09]	5.97 [5.75]	PhotoRx; BET	B
Agrawal, 1998 99232291	30 e (25)	22.4±2.7	1.8 [1.7]	4.8 [4.5]	7.3 [6.7]	3.0 [2.8]	2.4 [2.2]	5.4 [5.1]	None	C
Gupta, 1990 91244386	25 d (20)	20–35	1.98 [1.96]	5.22 [5.01]	7.34 [7.10]	-	-	5.36 [5.14]	BET;	C

Bold = significantly different (p < .05)

n.s. = not significantly different

* = significantly different between pre- and post- treatment

a

60% ABO incompatibility [56% of controls has ABO incompatibility as well]

b

40% Rh incompatibility; 25% ABO incompatibility

c

38% G6PD deficiency; 20% ABO incompatibility

d

28% ABO incompatibility, 40% Rh incompatibility; 5% Subaponeurotic bleed

e

10% had Rh incompatibility, 26.7% had ABO incompatibility and 63.3% had idiopathic hyperbilirubinemia

[ ] Controls' value; Abn = abnormal = IPL I–V values > 4.34±0.15 msec; VEP =Visual Evoked Potential

Tan, Skurr, and Yip (1992) presented similar findings in a group of 30 healthy non-hemolytic term infants with mean bilirubin of 15 mg/dl and who had undergone phototherapy. When compared to control (bilirubin <8.2 mg/dl), this group of infants who had prolonged BAEP latency V (p<0.02) and interpeak latencies III–V (p<0.05). Concomitant improvement occurred in these values as the bilirubin levels declined in response to phototherapy.

Vohr, Lester, Rapisardi, et al. (1989), in a separate study of 50 term infants (some with ABO incompatibility, none with Rh incompatibility), found that moderate bilirubin group had significantly prolonged brain-stem conduction time. Correlation analysis showed that serum bilirubin values correlated positively with brain-stem conduction time (r = 0.36, p<0.01).

Vohr, Karp, O'Dea, et al. (1990), in the same study mentioned above, found increased bilirubin concentration correlated with an increased latency of brain-stem auditory evoked response in wave 4, 5 (latency of 7.50 ± 0.29 msec in the moderate bilirubin group vs. 7.32 ± 0.26 msec in the low bilirubin group, p<0.05).

In a similar study, Gupta and Mann (1998) performed BAEP's on 60 neonates (37–40 week gestation). Forty-five patients with bilirubin levels ranging from 15.8 mg/dl to 26.4 mg/dl were compared to 15 control patients who were either clinically non-jaundiced or had bilirubin < 12 mg/dl. Causes of hyperbilirubinemia included G6PD deficiency, ABO incompatibility and unknown reasons. Treatment information was not provided. A positive correlation was found between increasing levels of serum bilirubin and brain stem conduction time (P<0.01). Retest at one month of age showed 33.33 percent of cases with a mean serum bilirubin of 19.46 mg/dl and 80 percent with a mean bilirubin of 15.97 mg/dl showed total recovery. After 6 months, three cases with a mean bilirubin of 26.3 mg/dl and one case with a mean bilirubin of 17.7 mg/dl failed to show any improvement. The authors also concluded that neonates with distortion of normal wave patterns on BAEP's were found to have a poorer prognosis compared with those with delayed inter-peak (I-P) latencies.

Sabatino, Verrotti, Ramenghi, et al. (1996) studied brain stem auditory evoked potentials (BAEP) in 90 full-term newborns with jaundice. After enrolling only the subjects that showed the three major BAEP components (waves I, III and V), 48 full-term subjects with bilirubin from 13.9 to 25.8 mg/dl and without evidence of hemolysis were followed over a 3-year period. BAEP's were performed on day 3 and repeated 5–7 days post phototherapy or exchange transfusion until the bilirubin had decreased below 8 mg/dl. On day 3, latencies of waves III and V were significantly increased compared to control subjects. Recordings performed 5 to 7 days post therapy showed no significant differences during subsequent session. Serial neuropsychological evaluations at 6, 18 and 30 months using the Brunett-Lezine test showed no neurodevelopmental abnormality. Statistical analyses of the neuropsychological evaluations were not given.

Perlman, Fainmesser, Sohmer, et al. (1983) studied BAEP's in 24 neonates (37–41 week gestation) with bilirubin levels of 15 mg/dl to 25 mg/dl. Six had ABO incompatibility, one had Rh incompatibility, two had hemorrhage and the remainder had unknown reasons for jaundice. All 24 received phototherapy; three infants also received exchange transfusion. Nineteen neonates who were either clinically non-jaundiced or had bilirubin levels <12 mg/dl served as control. Hyperbilirubinemic infants had prolonged brainstem transmission time (P<0.01). Wave complex IV–V was present in all 19 control infants; it was absent in at least one recording of 10 of 24 jaundiced infants (P<0.001). Nine of 13 infants who had sequential studies over several days showed apparent improvement manifested by the subsequent appearance of waves that were initially absent or by shortening of initially prolonged latency.

Agrawal, Shukla, Misra, et al. (1998) in the previously mentioned study of 30 term newborns with bilirubin > 15 mg/dl, recorded BAEP's in these infants before and after phototherapy or exchange transfusion and at the age of 2 to 4 months. Seventeen of 30 infants showed abnormalities on initial BAEP's; these abnormalities correlated significantly with bilirubin level. After therapy, abnormalities reverted back to normal in 10 cases but persisted in seven out of 17. Follow-up BAEP's were done in 21of 30 cases at 14.4 ± 2.2 weeks. Of the 15 patients who had an initial abnormal BAEP, in only three were there persistent abnormalities. Two of them had kernicterus.

Gupta, Raj, and Anand (1990) studied 25 term infants with bilirubin > 20 mg/dl and who received exchange transfusion. Seven had Rh and 10 had ABO incompatibility; one had a subaponeurotic bleed; seven had undetermined cause for jaundice. Twenty normal term infants without jaundice served as control. BAEP's were performed at 10 ± 3 days, just before hospital discharge. Mean latencies (wave III, V) and I–V interpeak latency (brainstem conduction time) were significantly prolonged in jaundiced neonates as compared with controls (P<0.01). BAEP abnormalities were found with greater frequency in jaundiced neonates requiring a repeat exchange transfusion (P <0.001). However, all infants had normal BAEP's on follow-up retesting at 3 months.

Valaes, Kipouros, Petmezaki, et al. (1980) followed up 415 children (233 exposed to prenatal phenobarbital, 182 were not) at 61 to 82 months of age in a prenatal phenobarbital study. None of them received phototherapy. An unspecified number had exchange transfusion. Ten children were diagnosed with sensorineural hearing defect. It was significantly more common in children with moderate (bilirubin >12 mg/dl) or marked (bilirubin > 16 mg/dl) neonatal jaundice than in the children with slight or no jaundice (9.2 percent vs. 1.1 percent; P<0.05).

Hyman, Keaster, Hanson, et al. (1969), in the previously mentioned study of patients with hemolytic disease or hyperbilirubinemia, reported that in the 351 patients who received streptomycin, there was a significant association of sensorineural hearing loss and bilirubin ≥ 20 mg/dl as tested by audiometry after age 4 years. Eleven out of 108 patients (10.2 percent) with bilirubin ≥20 mg/dl vs. Five out of 243 patients (2.1 percent) with bilirubin <20 mg/dl developed sensorineural hearing loss (P<0.001).

Johnston, Angara, Baumal, et al. (1967) studied 129 children at 5 to 6 years of age with neonatal bilirubin >20 mg/dl. Ninety-two babies had Rh incompatibility, twenty-six had ABO incompatibility and 11 had other reasons for their hyperbilirubinemia. There were 24 premature babies in this group. Ninety-five percent of the cases received exchange transfusion. Seven children in this group (five of them are full term) had sensorineural hearing loss. In the control group of 95 children with low bilirubin (11 were premature babies), there was not a single case of sensorineural hearing loss (p=0.00128). Both groups of children received streptomycin when they were neonates.

Suresh and Lucey (1997) obtained information via returned questionnaires on 42 patients (2 months to 21-years-old) with Crigler-Najjar syndrome type-1 via returned questionnaires. Mean bilirubin was 19.8 ± 4.5 mg/dl (typical) in the neonatal period. The mainstay of therapy in these patients with Crigler-Najjar was home phototherapy for 10 to 16 hours a day. Hearing was reported to be normal in 34/36 patients. The two children with hearing loss had conductive loss from otitis media.

Boo, Oakes, Lye, et al. (1994) studied brain stem evoked response in 128 jaundiced term infants at a mean age of 11.6 days (S.D.=3.9). One hundred thirty-six were originally admitted to the nursery. Eight were excluded because of aminoglycoside treatment, cyanotic heart disease and congenital infection. Non-physiologic causes of jaundice were identified in 39 subjects but details were not provided. All subjects had phototherapy. Forty-three also had exchange transfusion. Sixteen percent (13/83) of subjects with bilirubin < 19.8 mg/dl vs. 33 percent (15/45) of subjects with bilirubin > 19.8 mg/dl had hearing loss. The difference was not significant. Logistic regression analysis showed that the significant risk factors associated with hearing loss in the term jaundiced neonates were exchange transfusion and age of onset of hyperbilirubinemia.

Esbjorner, Larsson, Leissner, et al. (1991) studied 9 healthy neonates (35 to 42 week gestation) with bilirubin from 16.4 mg/dl to 25.6 mg/dl. One case of jaundice was caused by ABO incompatibility; no underlying disease was found in the rest. All received phototherapy. No significant correlation between the BAEP latency to wave V and the total bilirubin concentration was found at 4 days of age. The authors did find a statistically significant correlation between the reserve albumin concentration (as measured by monoacetyldiaminodiphenyl sulphone analysis) and the latency to wave V (r = -0.89, p<0.001).

Rosta, Makoi, Bekefi, et al. (1971) studied 84 children with neonatal bilirubin of >15 mg/dl at 7 to 9 years of age. Etiologies of hyperbilirubinemia included ABO or Rh incompatibilities, and hyperbilirubinemia of unknown origin. All subjects had birthweight ≥2,500 grams. The authors did not specify exactly how many subjects received exchange transfusion and/or streptomycin. Five out of 84 had a severe sensorineural hearing loss and nine had a mild to moderate sensorineural hearing loss. There was no control group in this study.

Effect of Bilirubin on Intelligence Outcomes

Table 3.9 Effect of Serum Bilirubin Level on Intelligence (more...)

Table 3.9 Effect of Serum Bilirubin Level on Intelligence Outcome in All Infants (GA ≥ 34 weeks)

Study, Year, UI #	Subjects N (Control)	Bilirubin Level (range) mg/dl	Outcomes	Confounders	Quality
Bengtsson, 1974 74170112	111 g (115)	20.0–51.0	@6.5–13 years old: IQ; Sensorineural hearing loss	BET	B
			No statistically significant correlation was found between I.Q. and maximal bilirubin values
			3 (7%) children with ABO-incompatible hyperbilirubinemia had neurogenic hearing loss, while none was found in control (P>0.05).
Rosta, 1971 72185033	84 I	≥ 15	@8 years of age: IQ	BET, Streptomycin Rx	C
			Distribution of IQ was the same as in the normal population. 4 (5%) children had IQ below 0.70, 11 (13%) had an IQ between 0.70 and 0.80.
Hyman, 1969 69132491	29 h	≥ 15	@4 years of age: IQ; Auditory rote memory difficulties	BET; streptomycin Rx; BW; Selection bias	C
			No association was found between high bilirubin exposure and low IQ (N=38).
			No significant difference in the prevalence of auditory rote memory difficulties between infants with TSB ≥ 20 mg/dl (N=17) and infants with TSB < 20 mg/dl (N=21)

g

All had negative Coombs' tests

h

from original 405 subjects, Including 10% preemies and some hemolytic diseases

i

“Nearly equal distribution of Rh and ABO immune-hemolysis, and hyperbilirubinemia of unknown origin”

Table 3.10 Mean Intelligence Quotient of Children with (more...)

Table 3.10 Mean Intelligence Quotient of Children with a History of Hyperbilirubinemia

Author, Year, UI #	Sample N (Controls)	Bilirubin Level (range) mg/dl	Mean IQ			Confounders	Quality
			Full	Verbal	Performance
Seidman, 1991 91375839	308 c (1,496)	Var1	103	-	-	PhotoRx; BET	B
	144 d (1,496)	Var2	103
Bengtsson, 1974 74170112	66	20–35	107	-	-	BET	B
	45 a	20–51	106	-	-
Culley, 1970 70266548	73 b (48)	12–16	106	-	-	Estimated TSB levels	B
		> 16	107	-	-
Ozmert, 1996 97154941	13 (27)	17–20	102	102	102	Possible selection bias (Unknown Rx effect)	C
	26 (27)	20–22	103	106	101
	27 (27)	22–25	99	101	95
	19 (27)	> 25	102	103	101
	17 g (27)	> 20	87	95	93
Odell, 1970 70075541	8 (6) f	> 19	94	-	-	BET	C
			25% < 80
Johnston, 1967 67081206	129 e (82)	> 20	105	-	-	BET; Prematurity	C

a

100% ABO incompatibility

b

Healthy infants (no hemolytic disease; no congenital malformations; no perinatal asphyxia; no neonatal illness or infections)

c

Including some infants with Coombs positive tests

d

Including infants with hemolytic disease (2% subjects had Coombs positive tests)

e

71% iso-immunization; 20% ABO incompatibility

f

43% Rh incompatibility; 14% ABO incompatibility; 14% Other blood incompatibility; 21% unknown cause of jaundice

g

65% Rh and 35% ABO incompatibility

st

Var1 = 5–8 mg/dl on 1 day of life, 10–15 mg/dl on the 2^nd day of life, or 13–20 mg/dl thereafter

st

Var2 = >8 mg/dl on the 1 day of life, >15 mg/dl on the 2^nd day of life, and >20 mg/dl thereafter

Bengtsson and Verneholt (1974) studied 111 children with neonatal bilirubin of 20 mg/dl to 51 mg/dl and 115 controls without visible jaundice. All were healthy full-term infants with negative Coombs' test. Forty-four of the hyperbilirubinemic group received exchange transfusion. Follow-up examinations were done between 6½ to 13 years of age. No statistically significant correlation was found between IQ and maximal bilirubin values. Out of the 111 children with hyperbilirubinemia, two had athetotic coordination disturbance and four had sensorineural hearing loss, while none of the 115 controls had such disorders. The differences were not statistically significant.

Rosta, Makoi, Bekefi, et al. (1971), in the previously mentioned uncontrolled study of 84 children with bilirubin >15 mg/dl and in whom follow-up examinations were done 7 to 10 years later, found that “the distribution of IQ was the same as in the normal population. Four children had IQ below 0.70; 11 had an IQ in the range between 0.70 –0.80.” Actual comparative data to the normal population was not provided.

Hyman, Keaster, Hanson, et al. (1969), in the previously mentioned study of infants with hemolytic disease and hyperbilirubinemia of other etiology, selected 29 available patients with either sensorineural hearing loss or Minimal Cerebral Dysfunction and nine control patients without CNS abnormality for further testing after age 4 years by one psychologist. These tests included Stanford-Binet, Wechsler Intelligence Scale for Children and others. “Based on the tests used, 29 patients (22 term and seven premature) were found to have normal mentality, five (four term and one premature) had subnormal mentality, and four (three term and one premature) were of superior mentality. Patients exposed to high bilirubin [≥ 20 mg/dl] and low bilirubin [< 20 mg/dl] during the neonatal period were almost equally represented in each mentality rating. Further analyses of the five patients with subnormal mentality showed that three had very high bilirubin exposure, 30 mg/dl or greater. No other correlation between bilirubin and mentality could be made.”

Seidman, Paz, Stevenson, et al. (1991) analyzed 1,948 17-year-olds (out of a total of 2,044 from a single hospital) who had neonatal records. Bilirubin values were available in 585 (30 percent of the 1,948) infants; subjects who had no clinical jaundice and for whom bilirubin levels were not obtained were included in the mild bilirubinemia group. Subjects' transformed intelligence test scores (from verbal Otis test and nonverbal matrices test results to values that correlated with Wechsler Adult Intelligence Scale) and medical examination findings performed at the military draft board at 17 years of age were stratified according to serum bilirubin levels. No direct linear association was found between neonatal bilirubin levels and intelligence test scores. However, the risk for IQ score less than 85 was found to be significantly higher (p=0.014) among full-term Coombs-negative male subjects with serum bilirubin greater than 20 mg/dl (odds ratio, 2.96; 95% CI 1.29–6.79).

Culley, Powell, Waterhouse, et al. (1970) studied Stanford-Binet intelligence quotient at 6 years of age in 169 neonates with non-hemolytic jaundice and 159 control neonates. These neonates' birthweight ranged from <1500 g to >3000 g. Treatment information was not provided. After adjusting for either birth weight or gestation in each jaundiced group, authors concluded that “in non-hemolytic jaundice there is no evidence of reduction in intelligence quotients with increasing jaundice up to the permitted level of around 20 mg/dl.”

Ozmert, Erdem, Topcu, et al. (1996) studied 102 (out of 416) children between 8 and 13 years of age. All these children had neonatal bilirubin >17 mg/dl (17–48 mg/dl), ≥ 37 week gestation and birthweight ≥ 3000 g. Seventeen children had positive direct Coombs' test (11 Rh, six ABO incompatibility). Treatment information was not provided. The control group consisted of 27 children of the same age but without neonatal hyperbilirubinemia. The Wechsler Intelligence Scale Revised for Turkish Children and other tests were performed in these children. Children who had positive direct Coombs tests and bilirubin >20 mg/dl had lower IQ scores and more prominent neurological abnormalities than the control group (P<0.05). IQ scores and prominent neurological abnormalities did not differ among the other groups of children with negative Coombs test and hyperbilirubinemia. The authors also noted an important risk factor for prominent neurological abnormalities was the duration that the infant's serum indirect bilirubin remained > 20 mg/dl.

Odell, Storey, and Rosenberg (1970) located 32 children between the ages of 4 and 7 years of age who had a history of neonatal jaundice. These 32 children represented only a third of the total population on whom bilirubin studies were performed during neonatal life. Twenty-four had exchange transfusion. Sixteen had either ABO or Rh incompatibility. Birthweight ranged from 1400 grams to 3827 grams. Fifteen patients had birth weight >2,500 grams. Maximum bilirubin level ranged from 11.4 mg/dl to 31.6 mg/dl with a median of 19 mg/dl. Stanford-Binet or Wechsler Intelligence Scale, and other visual-motor and short-term memory tests were given to this group of children. Six chi-square analyses were performed using IQ <80, ≥ 80 vs. maximum bilirubin of 20, 21, 22, 23, 24 mg/dl and values of bilirubin <19 mg/dl and ≥ 19 mg/dl. No statistical significant relationship was found. The authors further reported that there was a significant correlation (P <0.01) between the presence or absence of brain damage (measured by psychometric testing) and the degree to which the patient's sera had been saturated with bilirubin (as determined by the salicylate displacement technique).

Johnston, Angara, Baumal, et al (1967), in the previously mentioned study of 129 children with neonatal bilirubin >20 mg/dl, found the average IQ in this group was 104.8. The control group of 82 children had an average IQ of 102.8. No further statistical analysis was provided.

Studies from Collaborative Perinatal Project (CPP)

Table 3.11 Effect of Serum Bilirubin on Neurodevelopmental (more...)

Table 3.11 Effect of Serum Bilirubin on Neurodevelopmental Outcomes in Collaborative Perinatal Project Subjects

Author, Year, UI #	Subjects (N)	Outcomes	Confounders	Quality
Newman, 1993 94021217	CPP subjects, born from 1959 to 1966, BW>2500 g (N=41,324; 33,272 had 7-year outcome data)	@7 yr. of age: Adjusted IQ, Neurologic exams; Hearing loss	BET	A
		No significant difference in adjusted IQ between children with TSB < 20 mg/dl and ≥ 20 mg/dl in black and white subjects.
		There was no significant association between peak TSB and IQ (r=0.13 in whites; r=0.34 in blacks, p>0.05).
		Risk of abnormal or suspicious neurologic exam findings increased in stepwise manner from TSB <10 mg/dl to TSB ≥ 20 mg/dl. (P<0.001) Neurologic exam items most associated with bilirubin levels were mild and nonspecific motor abnormalities.
		2.2% of CPP subjects had hearing loss. Adjusted OR of hearing loss was 1.08 (95% CI 0.3–3.5)
Rubin, 1979 79153498	Subjects (N=241) both in CPP and Educational Follow-up Study( EFS); plus a control group from EFS population (N=125)	@8 months of age: Bayley MDI & PDI	Lost to follow-up; BET	B
		Infants with TSB 16–23 mg/dl (N=153) had significantly lower Bayley MDI and PDI scores at 8 months of age, compared to infants with TSB 11–15 mg/dl (N=66) and those with TSB ≤ 10 mg/dl (N=120). Controlling for GA, there was no longer a difference on MDI, but PDI remained significant.
		@1 year of age: Neurological exams
		Infants with TSB 16–23 mg/dl (N=153) had significantly more neurologic abnormalities at 1 year of age, compared to infants with TSB 11–15 mg/dl (N=69) and those with SB ≤ 10 mg/dl (N=120). Controlling for GA, the difference still remained significant.
		@5 & 6 years of age: MRT & ITPA
		No significant differences were found between groups.
		@7 years of age: IQ; Neurological exams
		No significant difference was found between groups.
Scheidt, 1977 77209363	CPP subjects, born from 1959 to 1964, only term and FT infants (GA≥36 wk and BW>2500 g) were included for current review (original sample size=24,524)	@8 months of age: Bayley MDI & PDI (White infants):	None was found (Unknown Rx effect)	B
		PDI score was significantly decreased with elevated TSB levels (≥10 mg/dl). However, significant decrease in Bayley MDI score only occurred in white term infants with TSB≥15 mg/dl.
		@8 months of age: Bayley MDI & PDI (Black infants):
		For black term (GA ≥ 36 wk and BW > 2500 g) infants, both Bayley MDI and PDI scores were significantly decreased with elevated TSB levels (≥10 mg/dl).
		No association between TSB levels and Bayley MDI and PDI scores was found in black FT subjects (GA ≥ 37 wk and BW > 2500 g).
Boggs, 1967 67261091	CCP subjects, born from 1959 to 1966, BW > 2500 g (N=18,484)	@8 months of age: Bayley MDI & PDI	Hemolytic disease; BET; Congenital malformation	B
		For infants with BW 2501–3000 g (N=4,916), the estimated increase in percent of infants with a low PDI score was 1.2% for each 1 mg/dl rise in TSB level. The estimated increase in percent of infants with a low MDI score was 0.8% for each 1 mg/dl rise in TSB level.
		For infants with BW ≥ 3000 g (N=13,568), the estimated increase in percent of infants with a low PDI score was 0.5% for each 1 mg/dl rise in total SB level. The estimated increase in percent of infants with a low MDI score was 0.3% for each 1 mg/dl rise in TSB level.
Naeye, 1978 79054333	CPP subjects, born from 1959 to 1966, mixed term and preterm infants (N=41,444)	@4 yr. of age: IQ	Prematurity, Placental infection (Unknown Rx effect)	C
		Increased frequency of mental impairment in subjects with peak bilirubin > 7 mg/dl
		The frequency of low IQ with increasing bilirubin levels increased more rapidly in infected infants.
		@7 yr. of age: Neurologic exam
		An increase in the prevalence of abnormalities started at peak neonatal bilirubin levels of 12 to 13 mg/dl.
		Neurologic abnormalities were more prevalent in subjects who had amniotic fluid infections.

MRT = metropolitan readiness tests

ITPA = Illinois test of psycholinguistic ability

Table 3.12 Mean Intelligence Quotient of Children from (more...)

Table 3.12 Mean Intelligence Quotient of Children from Collaborative Perinatal Project Cohort

Author, Year, UI #	Sample N (Controls)	Peak Bilirubin Level (range) mg/dl	Mean IQ	Confounders	Quality
Newman, 1993 94021217	41,324 a	< 10	White:104; Black: 93	BET	A
		10–15	White: 103; Black: 93
	White: 21,375	15–20	White: 104; Black: 93
	Black: 19,949	20–25	White: 105; Black: 91
		≥ 25	White: 106; Black: 90
Rubin, 1979 79153498	203 b (114)	11–15	102	Lost to follow-up; BET	B
		16–23	102
Upadhyay, 1971 71238892	20 c (20)	> 20	94 (10% < 80)	BET	C

Bold = below normal range

a

Including infants with hemolytic disease (2% subjects had positive Coombs' tests)

b

Subjects were from CPP, likely to include hemolytic jaundice and jaundice due to infections (not explicitly stated)

c

50% ABO incompatibility; 35% Rh incompatibility; 5% Septicemia with a gram-negative organism

From the CPP sample of 54,795 live-born infants, Newman and Klebanoff (1993) excluded infants who were not singletons, with birthweight < 2,500 grams or unknown, with race other than black or white, in whom no bilirubin measurement was recorded and the ones who died before their first birthday. This resulted in a study sample of 41,324 (or 75.4 percent of the original CPP sample) singleton black or white infants with birthweight ≥ 2,500 grams, who survived for at least 1 year and had at least one bilirubin level recorded. Outcome data at 7 years were available on approximately 80 percent (33,272 out of 41,324) of these infants. Neonatal bilirubin levels were not associated with loss to follow-up. IQ was measured using the Wechsler Intelligence Scale for children at age 7.1 ± 0.4 years (mean ± SD) in whites and 7.2 ± 0.8 years in blacks. Neurologic examinations were done at 1 and 7 years of age. Only results at age 7 years were presented; results at age 1 year were similar. The result was classified into normal, suspicious and abnormal. Hearing evaluations were done at 3 years ± 2 months and 8 years ± 3 months using pure-tone audiometry. Hearing examination at 8 years was “discontinued for administrative reasons” at six of the 12 centers. Thus, subjects for whom hearing data were available were not representative of the cohort as a whole. There was no consistent association between peak bilirubin level and IQ. To compare results with those of Naeye (see below) and Seidman et al (see above), IQ scores were dichotomized to <85 or ≥85. There was no association between bilirubin levels and dichotomized IQ in either race. The risk of abnormal or suspicious neurologic examination results increased in a stepwise manner from 14.9 percent (4,346 out of 29,258) in the group whose maximum bilirubin level was <10 mg/dl to 22.4 percent (60 out of 268) in the group whose maximum bilirubin was ≥ 20 mg/dl (P<0.001). Athetosis was not associated with peak bilirubin levels in this study. The proportion of children with sensorineural hearing loss at age 8 years was about 2 percent regardless of bilirubin level. However, authors noted that “because a comparatively small subset [16,886 out of 33,272 subjects at 8 years or 51 percent] of the CPP received hearing examinations and sensorineural hearing loss was uncommon, the power to detect an effect of very high bilirubin levels was limited."

Rubin, Balow, and Fisch (1979) enrolled all 241 subjects with maximum bilirubin of 11 mg/dl and above from the 1,613 participants both in the Educational Follow-up study (EFS) and the Minnesota section of the Collaborative Perinatal Project into a study trying to determine the relationship of neonatal hyperbilirubinemia to cognitive development at ages 4 through 7. One hundred and sixty-four subjects with bilirubin levels from 11 to 15 mg/dl were classified as the moderately elevated group; 77 subjects with bilirubin from 16 to 23 mg/dl were classified as the high bilirubin group. A control group of 125 subjects with bilirubin levels of 0–10 mg/dl was randomly drawn from the remainder EFS population and designated as the low bilirubin group. Nine subjects from the high and moderate bilirubin groups and one from the low bilirubin group had respiratory distress. Fourteen (18.2 percent) of the high bilirubin group had exchange transfusions. Bayley scales of Mental and Motor Development were administered at 8 months. IQ measures were obtained at age 4 using the Stanford-Binet, Short form L-M, and at age 7 using the Wechsler Intelligence Scale for Children. Neurologic examinations were administered at 1 year and 7 years. At age 8 months, high bilirubin group had a lower mean Bayley Motor Score the low bilirubin group (p<0.05). At age one year, high bilirubin group also has more neurologic abnormalities than the low bilirubin group (p=0.007). There were no significant differences found amongst bilirubin groups on the Stanford-Binet at age 4, Wechsler Intelligence Scale or the neurologic examinations at age 7.

Scheidt, Mellits, Hardy, et al. (1977), using the CPP population, analyzed the relationships between maximum bilirubin levels and neurodevelopment at 8 months and at 1 year of age. This study included white and black infants. Infants from multiple births, with Down syndrome, meningomyelocoele, hydrocephalus and congenital heart disease were excluded. Only those infants examined within a narrow time frame were included. 24,524 infants were examined at age 8 months ± 2 weeks and 27,720 were examined at age 12 months ± 4 weeks. The effects of bilirubin were evaluated within five birthweight/gestational age categories suggested by Yerushalmy. The two groups that are of particular interest to the present report are group 4 (>2,500 grams and ≥ 36 weeks) and group 5 (>2,500 grams and ≥ 37 weeks). A significant difference in Bayley motor score occurred with bilirubin ≤ 9 mg/dl vs. 15 to 19 mg/dl for white groups 3, 4 and 5 and for black groups 1, 2 and 4 (P≤0.05). A significant difference in motor scores also occurred with bilirubin ≤ 9 mg/dl vs. 10 to 14 mg/dl for white group 3, 4, 5 and black groups 1, 2, 3, 4 (P ≤ 0.02). Infants in group 2 (1,501 – 2,500 g, ≤ 36 weeks) had most variables in the 8 months Bayley and 1-year neurologic examination associated with increased level of bilirubin.

Boggs, Jr., Hardy, and Frazier (1967) compiled a preliminary report of approximately 23,000 subjects from the CPP and their 8-months standardized Bayley scores. The subjects were singletons; some had Coombs positive hemolysis, some had exchange transfusions and some had congenital malformations (analysis excluding infants with congenital malformations was essentially unchanged). Authors found that there was a significant (p<0.005) linear association between the maximum bilirubin levels and low motor score (<27) at 8 months. This trend was observed in all birthweight groups. The slope of the regression line was steeper in the low birthweight group. Similar findings were reported for the mental score.

From the CPP population, Naeye (1978) excluded subjects with multiple births, major congenital malformation, chromosomal disorders, lead intoxication, metabolic problems, CNS diseases, congenital syphilis and rubella. Subjects with mothers who had severe hypotension during labor or delivery or who were drug or alcohol addicted were also excluded. 45,347 infants were available for analysis regarding the relationship of hyperbilirubinemia, amniotic fluid infections, and neurodevelopmental outcomes at 1 year and 7 years. Neutrophil infiltration of the subchorionic plate of the placenta was taken as evidence of amniotic fluid infection. With or without amniotic infection, authors reported an increased frequency of IQ<90 at age 4 years with peak bilirubin ≥ 7 mg/dl. The incidence of IQ ≤ 90 with rising bilirubin levels increased more rapidly in infants with amniotic infection than in infants without. Authors also reported an increase in abnormal neurologic tests at 7 years of age, starting at peak bilirubin levels of 12 to 15 mg/dl.

In another study relating neonatal bilirubin to neurodevelopmental outcome, Upadhyay (1971), using the CPP subgroup from Johns Hopkins, identified 34 subjects (out of 4,195) with birthweight > 2,500 grams and maximum bilirubin > 20 mg/dl. Twenty subjects were evaluated. Fourteen out of 34 were excluded (five were evaluated prior to standardization of the bilirubin method, six were infants of high risk mothers, two were lost to follow-up and one child died at 14 months of kernicterus). Etiologies of hyperbilirubinemia included ABO and Rh incompatibility, sepsis and unknown reasons. Eleven had exchange transfusions. There were 20 matched control subjects with bilirubin less than 10 mg/dl. There was no difference in mean IQ at age 4 years between the two groups. However, in the psychological evaluation, eight out of 20 in the high bilirubin group vs. three out of 20 in the control were rated as non-normal (P<0.01).

Description of Included Studies

Efficacy of Phototherapy for Prevention of TSB Exceeding 20 mg/dl

Table 3.13 Summary Individual Studies of Efficacy of (more...)

Table 3.13 Summary Individual Studies of Efficacy of Phototherapy for Prevention of TSB Exceeding 20 mg/dl, in Healthy ^a, Term or Near-Term (GA≥34) Infants

Study, Year, UI#	Country	Subject Details	Phototherapy Protocol	Outcome	Study Design	Quality
Brown, 1985 85112402	US	Racially diverse jaundiced infants with BW > 2500 g TSB > 13 mg/dl	Continuously PhotoRx began at the time the SB level reached 13 mg/dl in the first 96 hrs. Duration of PhotoRx was 96 hrs. Infants' eyes were covered.	ET: carried out when SB > 20 mg/dl in both groups	RCT	A
Maurer, 1985 85112404
Martinez, 1993 93141315	US	FT breast-fed infants with SB ≥ 17 mg/dl	(I) Continue breast-feeding, administer PhotoRx	SB ≥ 20 mg/dl	RCT	A
			(II) Discontinue breast-feeding, substitute formula, and administer PhotoRx.
			Infants' eyes were covered.
John, 1975 75183114	Australia	Mature infants who developed unexplained jaundice with SB > 15 mg/dl	Each course consisted of a continuously PhotoRx for 18 hrs, followed by a 6 hour break. 3 or 4 courses were given. Infants' eyes were covered.	ET: carried out when SB > 20 mg/dl in both groups	Non-RCT	C

a

Healthy = no hemolytic disease; no congenital malformations; no perinatal asphyxia; no neonatal illness or infections

Non-RCT = nonrandomized control trial

ET = exchange transfusion

Table 3.14 Number Needed to Treat at TSB Level of 20 (more...)

Table 3.14 Number Needed to Treat at TSB Level of 20 mg/dl to Keep TSB from Rising by Phototherapy in Healthy a Jaundiced Term or Near-term (GA≥34) Infants

Study, Year, UI#	Groups	No. of Infants whose SB>20 mg/dl (%)	Total number of Infants	Mean SB level at entry (range, mg/dl)	Follow-up	ARR	NNT (95%CI)
Brown, 1985 85112402	Ctrl grp I b	10 (17)	60	≥ 13	144 hrs	+0.5%	-
	Rx grp I b	12 (17)	70
*sample from NICHD trial	Ctrl grp II	13 (17)	76	≥ 13	144 hrs	14.3%	7
	Rx grp II	2 (3)	70				(6–8)
Maurer, 1985 85112404	Ctrl grp I c	6 (20)	30	15.8 (13–23)	144 hrs	+0.6%	-
	Rx grp I c	7 (21)	34	16.5 (10–23)
*sample from NICHD trial	Ctrl grp II d	17 (16)	106	15.6 (10–24)	144 hrs	9.4%	11
	Rx grp II d	7 (7)	106	15.5 (10–22)			(10–12)
Martinez, 1993 93141315	Ctrl grp I	6 (24)	25	17.8	48 hrs	10.1%	10
	Rx grp I	5 (14)	36	18.0			(8–11)
	Ctrl grp II	5 (19)	26	17.8	48 hrs	16.6%	6
	Rx grp II	1 (3)	38	17.9			(5–7)
John, 1975 75183114	Ctrl grp	70 (18)	381	15–18	ND	11.2%	9
	Rx grp	8 (7)	111				(8–10)

a

Healthy = no hemolytic disease; no congenital malformations; no perinatal asphyxia; no neonatal illness or infections

b

Hemolytic disease group (infants with hemolysis, high risk group). BET were carried when infants' SB >18 mg/dl in the high risk group.

c

Coombs' test positive (91% ABO incompatibility; 9% Rh incompatibility)

d

Including some infants in the high-risk group. ET were carried when infants' SB >18 mg/dl in the high risk group.

ARR = absolute risk reduction rate; NNT = number needed to treat

Two studies (Brown, Kim, Wu, et al., 1985; Maurer, Kirkpatrick, McWilliams, et al., 1985) evaluated the same sample of infants. Both reports were derived from a RCT of phototherapy for neonatal hyperbilirubinemia commissioned by the National Institute of Child Health and Human Development (NICHD) between 1974 and 1976.

Because the phototherapy protocols differed significantly in the remaining studies, their results could not be statistically combined, and are reported separately here. A total of 893 term or near-term jaundiced infants (325 in the treatment group and 568 in the control group) were evaluated in the current review.

The development, design, and sample composition of NICHD phototherapy trial was reported in detail elsewhere (Bryla, 1985). NICHD's controlled trial of phototherapy for neonatal hyperbilirubinemia consisted of 672 infants who received phototherapy and 667 control infants. Of the 672 subjects in the treatment group, 140 had a birthweight ≥ 2,500 grams; all had gestational age (GA) ≥ 34 weeks. Of the 667 control infants, 136 had a birthweight ≥ 2,500 grams; all had a GA ≥ 34 weeks. The racial composition of the total 276 term/near-term subjects was 26 percent black, 38 percent Mexican-American, and 36 percent non-black/non-Mexican-American. Males represented 59 percent of the total sample.

The same group of infants, 140 subjects in the treatment group and 136 controls with BW ≥ 2,500 grams and GA ≥ 34 weeks, were evaluated for the effect of phototherapy on the hyperbilirubinemia of Coombs' positive hemolytic disease in the study of Maurer, Kirkpatrick, McWilliams, et al. (1985). Of the 276 infants whose birthweight was 2,500 grams or more, 64 (23.1 percent) had positive Coombs' tests, 58 secondary to ABO and 6 secondary to Rh incompatibility. Thirty-four out of 64 in this group received phototherapy. The other 30 were placed in the control group. Of the 212 subjects who had negative Coombs' tests, 106 were in the treatment group and the same number was in the control group. No therapeutic effect on reducing the exchange transfusion rate was observed in infants with Coombs' positive hemolytic disease, but there was a 9.4 percent absolute risk reduction in babies who had negative Coombs' tests. In this group of infants, the NNT for prevention of the needs for exchange transfusion, or a TSB > 20 mg/dl, was 11 (95% CI, 10 to 12) (Maurer, Kirkpatrick, McWilliams, et al., 1985).

A more recent RCT compared the effect of four different interventions on hyperbilirubinemia (serum bilirubin concentration ≥ 291 μmole/l or 17 mg/dl) in 125 full-term breast-fed infants. Infants with any congenital anomalies, neonatal complications, hematocrit >65 percent, significant bruising or large cephalhematomas, or hemolytic disease were excluded. The four interventions in the study were (1) continue breast-feeding and observe (N=25); (2) discontinue breast-feeding, substitute formula (N=26); (3) discontinue breast-feeding, substitute formula and administer phototherapy (N=38); (4) continue breast-feeding, administer phototherapy (N=36). The interventions were considered failures if serum bilirubin levels reached 324 μmole/l or 20 mg/dl (Martinez, Maisels, Otheguy, et al., 1993). For the purpose of the current review, we re-grouped the subjects into treatment group or phototherapy group, and, control group or no phototherapy group. Therefore, the original group 4 and 3 became the treatment group I and II, and the original group 1 and 2 were the corresponding control group I and II. It was found that treatment I, phototherapy with continuing breast-feeding, had 10.1 percent ARR and the NNT for prevention of a serum bilirubin exceeding 20 mg/dl was 10 (95% CI 9 to 12). Compared to treatment I, treatment II (phototherapy with discontinuation of breast-feeding) was significantly more efficacious. It had 16.6 percent ARR and the NNT for prevention of a serum bilirubin exceeding 20 mg/dl was 6 (95% CI, 5 to 7).

John (1975) reported the effect of phototherapy in 492 full-term neonates born during 1971 and 1972 and who developed unexplained jaundice with bilirubin levels > 15 mg/dl. One hundred and eleven subjects received phototherapy. Three hundred and eighty-one infants did not. The author stated, “The choice of therapy was, in effect, random since two pediatricians approved of the treatment and two did not.” The results showed that phototherapy had a 11.2 percent risk reduction of exchange transfusion, performed in both treatment and control groups when serum bilirubin levels exceeded 20 mg/dl. Therefore, the NNT for prevention of a serum bilirubin exceeding 20 mg/dl was 9 (95% CI 8 to 10).

Effectiveness of Reduction in Bilirubin Level on Brainstem Auditory Evoked Response (BAER) in Jaundiced Infants with GA ≥ 34 weeks

Table 3.15 Effectiveness of Decreasing Bilirubin Levels (more...)

Table 3.15 Effectiveness of Decreasing Bilirubin Levels on Auditory Brainstem Response in Jaundiced Infants (GA ≥ 34 weeks)

Study, Year, UI #	Sample N (Control)	Mean Bili. Levels before Rx → After Rx (mg/dl)	Mean latency (msec) Before Rx → After Rx			Mean Interpeak (msec) Before Rx → After Rx			Follow-up (follow-up rate)	Quality
			I	(II), III	(IV),V8.1→7.0 [7.0]	I–III	III–V	I–V
Phototherapy and/or Exchange Transfusion
Sabatino, 1996 97171477	48 (42)	14–25.8→ < 8	1.9→1.6 [1.8]→[1.6]	4.8→4.0 [4.0]→[3.9]	8.1 →7.0[7.0] [6.9]	3.0→2.3 [2.3]→[2.2]	3.1→2.9 [2.9]→[2.8]	6.1→5.2 [5.2]→[5.0]	All normal (100%)	B
Deorari, 1994 95189280	7a (20)	24.6±1.6→14.8±2.1	2.1→1.8[1.9]	5.5→5.4*[5.1]	8.1→7.4*[7.0]	-	-	6.0→5.6[5.1]	All normal (100%)	C
Agrawal, 1998 99232291	30b (25)	22.4±2.7→ ND	1.8→1.7*[1.7]	4.8→4.6*[4.5]	7.3→6.9*[6.7]	3.0→2.8*[2.8]	2.4→2.3*[2.2]	5.4→5.1*[5.1]	10% ab-normal (100%)	C
Exchange Transfusion
Hung, 1989 90054139	6 c	20.4→ 11.8	2.5→2.2*	5.2→5.1*	7.4→7.3*	-	-	-	n.d.	B
Kuriyama, 1986 89246798	6d (8)	21.5±4.6→11.6±3.5	1.9→1.7[1.8]	5.0→7.8[4.6]	7.4→7.0[6.9]	3.1→3.0[2.8]	2.4→2.3[2.3]	5.5→5.3[5.1]	All normal (83%)	B
Funato, 1996 97104685	8e (16)	25.3±2.6→11.0±2.4	2.0→1.8*[1.8]	-	-	2.9→2.9[2.7]	-	5.3→5.2*[5.0]	All normal (100%)	C
Nwaesi, 1984 85037769	9 f	22.3±1.4→15.3±3.2	1.9→1.9	-	-	3.0→2.8*	3.4→3.1*	6.4→5.8*	n.d.	C
Treatments (not specified)
Bhandari, 1993 94075016	30 g	16.95±2.8→ ND	1.8→1.7	4.6→4.4	6.7→6.7	2.7→2.7	2.2→2.1	4.8→5.0	n.d.	B

Bold = significantly different than controls

* = significantly different between pre- and post- Rx

[ ] controls' value

n.d. = not done

a

All underwent PhotoRx. ET was carried out when ABR was considered abnormal. Two ABO incompatibility, 1 Rh incompatibility, 1 G6PD deficiency; 3 Idiopathic.

b

10% had Rh incompatibility, 26.7% had ABO incompatibility and 63.3% had idiopathic hyperbilirubinemia

c

Causes of jaundice was “mostly ABO, G6PD, and unknown factors”

d

83% Hemolytic disease; 17% Hemoperitoneum

e

One ABO incompatibility, 3 Polycythemia, 1 Cephalhematoma, 5 Idiopathic jaundice

f

33% Rh incompatibility; 67% ABO incompatibility

g

20% ABO incompatibility, 7% Rh isoimmunization; 30% G6PD deficiency

Table 3.16 Follow-up Results of Effectiveness of Decreasing (more...)

Table 3.16 Follow-up Results of Effectiveness of Decreasing Bilirubin Levels on Auditory Brainstem Response in Jaundiced Infants (GA ≥ 34 weeks)

Study, Year, UI #	Follow-up Outcomes	Confounders	Quality
Phototherapy and/or Exchange Transfusion
Sabatino, 1996 97171477	@1, 2, and 3 years of age: Brunett-Lezine test	PhtoRx; ET	B
	Neuropsychological evaluations performed in subjects showed scores similar to those of age-matched controls.
Deorari, 1994 95189280	@1 year of age: DQ; Neurological sequelae	PhotoRx; ET	C
	Mean development motor quotient and mean development mental quotient in study and control groups were comparable and normal at 1 year of age
	None of the study or control group babies developed seizures, motor deficits or features of cerebral palsy
Agrawal, 1998 99232291	@1 year of age: DDST	PhotoRx; ET; Lost to follow-up	C
	Neurological development was normal in all infants with SB 15–20 mg/dl, in 89% infants with SB 21–25 mg/dl, and in 67% infants with SB > 25 mg/dl
Exchange Transfusion
Kuriyama, 1986 89246798	@25 days to 6 months: neurological development; hearing	ET	C
	All are normal development and not hearing abnormalities.
Funato, 1996 97104685	@1.5–6 years of age: DQ/IQ	ET	C
	All had normal DQ/IQ > 85 except for 1 patient who developed borderline intelligence with IQ=77 at 6 years old

Sabatino, Verrotti, Ramenghi, et al. (1996) studied brain stem auditory evoked potentials (BAEP) in 90 full-term newborns with jaundice. After enrolling only the subjects who showed the three major BAEP components (waves I, III and V), 48 full-term subjects with bilirubin from 13.9 to 25.8 mg/dl and without evidence of hemolysis were followed over a 3-year period. BAEP's were performed on day 3 and repeated five to seven days post phototherapy or exchange transfusion until the bilirubin had decreased below 8 mg/dl. On day 3, latencies of waves III and V; and, interpeak latency III–V and I–V were significantly prolonged compared to control subjects. Recordings performed 5 to 7 days post therapy and during subsequent session showed no significant differences. Serial neuropsychological evaluations at 6, 18 and 30 months using the Brunett-Lezine test showed no neurodevelopmental abnormality. Statistical analyses of the neuropsychological evaluations were not given (Sabatino, Verrotti, Ramenghi, et al., 1996).

Infants with ABO incompatibility, Rh incompatibility, G6PD deficiency or idiopathic jaundice were treated with phototherapy and/or exchange transfusion in 7 studies (Agrawal, Shukla, Misra, et al., 1998; Bhandari, Narang, Mann, et al., 1993; Deorari, Singh, Ahuja, et al., 1994; Funato, Teraoka, Tamai, et al., 1996; Hung, 1989; Kuriyama, Tomiwa, Konishi, et al., 1986; Nwaesei, Van Aerde, Boyden, et al., 1984). The results consistently reported that treatments for jaundice significantly reduced the prolonged auditory brainstem response (ABR) latencies and interpeak latencies back to normal. The mean SB levels of the 96 infants also decreased from 20–25 mg/dl before treatments to 11–15 mg/dl after treatments. Of the 7 studies, four studies had follow-up results of neurological and intelligence outcomes. Three studies (Deorari, Singh, Ahuja, et al., 1994; Funato, Teraoka, Tamai, et al., 1996; Kuriyama, Tomiwa, Konishi, et al., 1986) with 14 jaundiced infants in the treatment group and 28 non-jaundiced infants in the control group, showed all normal developments at 25 days to 6 years of age. Moreover, Agarwal, Narayan, Kumari, et al. (1998) studied 30 term newborns with bilirubin > 15 mg/dl, recorded BAEP's in these infants before and after phototherapy or exchange transfusion and at the age of 2 to 4 months. Of the 30 infants, 10 percent had Rh incompatibility, 26.7 percent had ABO incompatibility and 63.3 percent had idiopathic hyperbilirubinemia. Seventeen out of the 30 infants showed abnormalities on initial BAEP's; these abnormalities correlated significantly with bilirubin level. After therapy, abnormalities reverted back to normal in 10 cases but persisted in 7 out of 17. Two subjects had kernicterus with bilirubin > 15 mg/dl. Follow-up BAEP's were done in 21/30 cases at 14.4 ± 2.2 weeks. Of the 15 patients who had an initial abnormal BAEP, in only 3 were there persistent abnormalities. At one year of age, screening by DDST showed that neurological development was normal in 100 percent of infants with bilirubin of 15–20 mg/dl; neurological development was normal in 88.9 percent of infants with bilirubin of 21 to 25 mg/dl; and neurological development was normal in 66.7 percent of infants with bilirubin >25 mg/dl. No further statistical details were provided (Agrawal, Shukla, Misra, et al., 1998).

Effects of Treatments for Hyperbilirubinemia on Neurodevelopmental Outcomes

Table 3.17 Effect of Serum Bilirubin Levels on Neurodevelopmental (more...)

Table 3.17 Effect of Serum Bilirubin Levels on Neurodevelopmental Outcomes in Infants with GA ≥ 34 weeks

Study, Year, UI #	Subject N (Controls)	Mean Bilirubin levels (mg/dl)	Outcomes					Bias/ Confounders	Quality
Behavioral Outcomes
Abrol, 1998 20235719	15 (15)	12.1–15.0 (8.0–12.0)	Orientation cluster items	24–48 h after initiation of Phototherapy		24–48 h after cessation of Phototherapy		ND	A
			Visual animate	P<0.01		P<0.05
			Visual inanimate	P<0.01		P<0.05
			Animate visual and auditory	P<0.01		P<0.05
Ju SH, 1991 91289763	29 (14)	12–15 (<12)	Orientation cluster	Before Rx	After Rx	24 h after Rx	2wks age	Incomplete binding since had mild or no clinically observable jaundice	B
			Mean	NS	P<0.01	P<0.05	NS	Large number of drop-outs at 2 week follow-up in the control group Small sample size
			Animate visual	NS	P<0.05	P<0.05	NS
			Animate visual & auditory	NS	P<0.05	P<0.05	NS		B
			Inanimate visual & auditory	NS	P<0.05	P<0.05	NS
			Alertness	NS	P<0.05	P<0.05	NS
Paludetto, 1983 84057358	30 (30)	13.3	Orientation cluster		3rd day	4th day	1month	Loss of subject and matched control at each observation 30->14->12 at one month of age	B
		8.4–17.5 (9.6)	Animate visualP<0.01			P<0.05	NS
		(3.5–14.5)	Inanimate visual P<0.05			P<0.01	P<0.05
			Inanimate visual P<0.005 and auditory			P<0.01	P<0.05
			Alertness P<0.005			P<0.05	NS
			Motor performance Pull to sit P<0.001			NS	NS
			Regulation of state Cuddliness P<0.001			NS	NS
Telzrow, 1980 80225471	10 (10)	17.07	Mean cluster scores on Brazelton's scale					No bilirubin assessment made in control group	B
					Day 3	Day 6	Day 10
			Orientation		P< 0.05	P<0.05	P=0.05
			Motor		NS	P<0.01	NS
Valkeakari 1981 81156354	41 (42)	-	Vineland social maturity scale		Phototherapy		Control	No information about the bilirubin levels of study or control group. Multiple analyses from small N.	B
			Self help general		12.56 ± 0.50		12.64 ± 0.41
			Self help eating		8.61 ± 0.70		8.67 ± 0.65
			Self help dressing		4.95 ± 1.75		4.44 ± 1.52
			Locomotion		5.76 ± 0.92		5.66 ± 1.07
			Occupation		7.51 ± 0.84		7.31 ± 0.79
			Communication		5.98 ± 0.11		5.94 ± 0.31
			Socialization		5.85 ± 0.79		5.82 ± 0.82
			Social age		4.20 ± 0.51		4.07 ± 0.48
			Bowel control		1.81 ± 0.54		1.65 ± 0.46
			Bladder control		1.93 ± 0.56		1.74 ± 0.47
			Speech score		1.41 ± 0.38		1.45 ± 0.34
Neurological or Motor Outcomes
Seidman, 1994 94223372	42 (42)	20.0 ± 0.4 (18.8 ± 0.3)	The phototherapy group showed an average IQ score of 108 ± 2 at 17 years which was not significantly different from the average IQ score of 107 ± 2 in the control group.					Retrospective study method used	B
Valkeakari 1981 81156354	41 (42)	ND	There were 2 infants with febrile convulsions in the PT group as compared to one in control group. There was only one infant with breath holding spells in PT group and none in control group.					No information about the bilirubin levels of study or control group	B
Misra, 1980 80203306	81 (60)	10–30 (8.5)	Ten of the 81 subjects manifested some degree of mental retardation at some or the other time during follow-up while all the 60 controls were normal. But on long term follow-up 3 of the 31 subjects remained retarded with AQ<90 at their last examination and remaining 7 had become normal.					Large number of patients lost to follow-up 81→ 12	C
								Unable to sort out effect of phototherapy from higher bilirubin as control group had lower bilirubin

Table 3.18 Effect of Serum Bilirubin Levels on Visual (more...)

Table 3.18 Effect of Serum Bilirubin Levels on Visual Outcomes in Infants with GA ≥ 34 weeks

Study, Year, UI #	Subjects N (Controls)	Mean Bilirubin levels (Range) mg/dl	Outcomes			Bias/ Confounders	Quality
Granati, 1984 84283738	110 (110)	19.2 ± 2.3 (14.2 –24.6)	Findings	PT group	Control group	Basic testing of vision and hearing may not have detected subtle measurements Bilirubin ranged from 14.0 –24.6 in subjects (mean 19.5)	B
		[9.7 ± 1.7] [7.2–12.2]	Normal visual function	77 ( 95.1%)	84(94.5%)
			Reduced visual acuity	3(3.7%)	4(4.4%)
			Strabismus	1(1.2%)	1(1.1%)
Valkeakari, 1981 81156354	41 (42)	-	Findings	PT	Controls	No information about the bilirubin levels of study or control group	B
			Normal vision	33	36
			Vision determination unsuccessful	7	6
			Anisometry	1	-
			Phoria/Trophia	10	12
			Changes in fundus (non specific)	1	1
			changes in lens (physiological)	-	2
			Nystagmus	-	-
Khanna, 1984 85103758	52 (52)	-	Retinoscopic finding	PT Group	Controls	Unclear how the total cohort of 112 was selected.	C
			Emmetropia	7(13.4%)	4(7.8%)	Wide age range at follow-up( 6 months to 7 years)
			Hypermetropia	36(69.2%)	39(76.4%)	Wide bilirubin range (15–69.2 mg/dl)
			Myopia	6(11.5%)	4(17.8%)
			Astigmatism	3(5.7%)	4(7.8%)
			Dyslexia test	-	-
			Fundus examination	5	6
			Myopic discs	3	2
			Pale discs	2	-
			Variations in size	1	-
			Dull macular reflex	-	2
			Mottling/ greying around macula	-	2

( ) Controls' value

Description of Included Studies

Table 3.19 Part 1. Summary of Individual Studies of (more...)

Table 3.19 Part 1. Summary of Individual Studies of Prediction Strategies of Hyperbilirubinemia in Healthy, Term or Near-Term (GA≥34 weeks) Infants

Study,Year UI #	Country	Race (N)	Definition of Hyperbilirubinemia	Reference Standard: Laboratory-based assay of TSB	Quality
Prediction Method: Cord Bilirubin
Carbonell, 2001 21130776	Spain	ND (585)	SB ≥ 17 mg/dl in the first 4 days of life	Bilirubinometer Bil-Red by direct reading at 461 mcm	A
Knudsen, 1992 92306741	Denmark	ND (138)	SB > 11.7 mg/dl at day 3	Standard direct spectroscopic method at 3rd postnatal day	A
Knudsen, 1989 89189829	Denmark	ND (291)	“Clinical jaundice” following by SB > 10 mg/dl	Standard method of Blom and Doumas	B
Risemberg, 1977 77156212	US	ND (91)	SB > 10 mg/dl at 36 hrs of age	Laboratory-based assay	B
Prediction Method: Early Serum Bilirubin (mg/dl)
Awasthi, 1998 20234451	India	ND (274) a	Peak SB > 15 mg/dl	Spectrophotometry on Toyo Bilirubin Analyzer	A
Carbonell, 2001 21130776	Spain	ND (574+865) b	SB ≥ 17 mg/dl in the first 4 days of life	Bilirubinometer Bil-Red by direct reading at 461 mcm	A
Seidman, 1999 20110115	Israel	Multiple race (1177)	TSB > 10 mg/dl at day 2, > 14 mg/dl at day 3; and, > 17 mg/dl at day 4 or 5	Unitstat Bilirubinometer	A
Risemberg, 1977 77156212	US	ND (91)	SB > 16 mg/dl at 36 hrs of age	Laboratory-based assay	B
Prediction Method: End-tidal Carbon-Monoxide Concentration (ETCOc, p.p.m)
Stevenson, 2001 21326594	US + International	Multiple race (1370)	TSB ≥ 95th %tile	Laboratory-based assay. Each sites used its own laboratory and method	A
Okuyama, 2001 21366215	Japan	ND (51)	Peak SB ≥ 15 mg/dl	Laboratory-based assay	B
Prediction Method: Predischarge Risk Index
Newman, 2000 20529193	US	(496)	TSB within the first 30 days of life ≥ 25 mg/dl	Data from the KPMCP integrated laboratory information system	B
Prediction Method: Predischage Risk Zone, Determined by Hour-Specific Bilirubin Percentile
Bhutani, 1999 99117600	US	Multiple race (2840)	Subsequent TSB risk zone ≥ 95th %tile	2,5-DPH diazo mathod	A

SB = serum bilirubin; TSB = total serum bilirubin

All infants were healthy, term (GA≥34) infants, unless noted:

a

28% “Preterm”

b

574 infants in the phase I study; 865 infants in the phase II study. 90% infants were breast-fed exclusively.

Table 3.19 Part 2. Test Accuracy of Prediction Strategies (more...)

Table 3.19 Part 2. Test Accuracy of Prediction Strategies of Hyperbilirubinemia in Healthy, Term or Near-Term (GA≥34 weeks) Infants

Study, Year, UI #	TP (n)	FN (n)	TN (n)	FP (n)	Sens (%)	Spec (%)	Thresholds
							Method	SB (mg/dl)
Prediction Method: Cord Bilirubin (μmol/l)
Carbonell, 2001 21130776	4	17	534	30	22	95	37	17
Knudsen, 1992 92306741	28	0	6	104	100	5	20	11.7
	27	1	23	87	96	21	25	11.7
	25	3	46	64	89	42	30	11.7
	20	8	75	35	71	68	35	11.7
	11	17	94	16	39	85	40	11.7
Knudsen, 1989 89189829	4	7	70	3	36	96	40	10 j
Risemberg, 1977 77156212	12	16	63	0	43	100	68	10.0
Prediction Method: Early Serum Bilirubin (mg/dl)
Awasthi, 1998 20234451	24	11	157	82	69	66	3.99 f	15
Carbonell, 2001 21130776	7	0	74	88	100	46	6 @24 hrs h	17
	11	0	102	56	100	64	9 @48 hrs h	17
	25	0	239	159	100	60	6 @24 hrs i	17
	45	1	348	426	98	45	9 @ 48 hrs i	17
Seidman, 1999 20110115	49	11	-	-	82	-	5 @day 1	HB g
Risemberg, 1977 77156212	13	0	78	0	100	100	10 @12 hrs	16
	13	0	78	0	100	100	15 @24 hrs	16
Prediction Method: End-tidal Carbon-Monoxide Concentration (ETCOc, p.p.m)
Stevenson, 2001 21326594	92	28	635	615	77	51	1.5 @36 hrs	≥ 95th %tile
	-	-	-	-	90	65	Comb	≥ 95th %tile
Okuyama, 2001 21366215	5	2	27	17	71	61	1.6 @36 hrs	15
	6	1	35	9	86	80	1.8 @42 hrs	15
	6	1	32	12	86	73	1.8 @48 hrs	15
	6	1	29	15	86	66	1.8 @60 hrs	15
Prediction Method: Predischage Risk Index
Newman, 2000 20529193	58	1	129	273	98	32	RI > 7	25
	52	7	244	158	88	61	RI >10	25
	25	34	364	38	42	91	RI > 15	25
	8	51	399	3	14	99	RI > 20	25
Prediction Method: Predischage Risk Zone, Determined by Hour-Specific Bilirubin Percentile
Bhutani, 1999 99117600	126	0	1756	958	100	65	≥ 40th %tile	≥ 95th %tile
	114	12	2300	414	91	85	≥ 75th %tile	≥ 95th %tile
	68	58	2610	104	54	96	≥ 95th %tile	≥ 95th %tile

TP = true positive; FN = false negative; TN = true negative; FP = false positive; N = number of subjects; n = number of measurements; Sens = sensitivity; Spec = specificity; SB = serum bilirubin; r = regression coefficient between prediction method and SB; RI = risk index; HB = hyperbilirubinemia

Comb = combined ETCOc and TSB method (either an TSB between 75th and 95th %tile or an ETCOc ≥ population mean or both @ 36±6 hrs). Infants with TSB ≥ 95th %tile @ 30±6 hrs were excluded.

f

mean SB at 18–24 hours of age

g

TSB > 10 mg/dl at day 2, > 14 mg/dl at day 3; and, > 17 mg/dl at day 4 or 5

h

Phase I study

i

Phase II study

j

“Clinical jaundice” following by SB > 10 mg/dl

Accuracy of Various Strategies for Predicting Hyperbilirubinemia

Accuracy of Cord Bilirubin Level as a Test for Predicting Later Development of Hyperbilirubinemia

Carbonell, Botet, Figueras, et al. (2001) studied the predictive value of cord bilirubin in 585 healthy term newborns. A cord bilirubin ≥ 2.2 mg/dl predicted significant hyperbilrubinemia, defined as TSB ≥ 17 mg/dl between the 3^rd and 4^th days of life, with a sensitivity of only 22 percent and a specificity of 95 percent.

Figure 3.6 Test Performance of Cord Bilirubin Levels (more...)

   Figure 3.6 Test Performance of Cord Bilirubin Levels

Data Source: Knudsen, 1992 [#92306741]

Knudsen studied cord bilirubin levels in two studies (Knudsen, 1989; Knudsen, 1992) comprised of 291 and 138 healthy term infants respectively. Different levels of cord blood bilirubin (CB) were tested for prediction of later development of hyperbilirubinemia, defined as TSB > 11.7 mg/dl or 200 umol/l. Phototherapy was required in 57 percent of newborns with CB > 40 umol/l, but only 9 percent if CB was ≤ 40 umol/l. In the second study, Knudsen (1992), the correlation of CB with TSB was 0.5 (Pearson's r). In both studies, sensitivity increased and specificity decreased directly as the threshold declined from CB > 40 umol/l to CB > 20 umol/l. Figure 3.6 shows the ROC curve from the Knudsen (1992) study.

Risemberg, Mazzi, MacDonald, et al. (1977) studied cord bilirubin in 91 neonates from ABO incompatible pregnancies and 30 control infants to determine predictability for severe hyperbilirubinemia, defined as TSB > 16 mg/dl at 12 to 36 hours of age. A cord bilirubin level of greater than 4 mg/dl was found to indicate high risk, requiring frequent re-evaluation and therapy.

Accuracy of Predischarge Risk Index as a Test for Predicting Later Development of Hyperbilirubinemia

Figure 3.7 Test Performance of Predischarge Risk Index (more...)

   Figure 3.7 Test Performance of Predischarge Risk Index

Data Source: Newman, 2000 [#20529193]

Newman, Xiong, Gonzales, et al. (2000) reviewed the medical records of 73 neonates with maximum TSB of ≥ 25 mg/dl, compared with 423 randomly selected controls with TSB <25 mg/dl, in order to determine biological and health service predictors of extreme hyperbilirubinemia. The strongest predictors were found to be family history of jaundice in a newborn (OR=6.0), exclusive breastfeeding (OR=5.7), bruising (OR=4.0), Asian race (OR=3.5), cephalhematoma (OR=3.3), maternal age ≥ 25 years (OR=3.1), and lower gestational age (OR=0.6/wk). The authors developed a risk index with this data, assigning a numerical value to each characteristic. In the 32 percent of neonates with an index score of ≤ 7, the risk of developing a TSB level of ≥ 25 mg/dl was about 1 in 16,000. In the one percent of neonates with an index score of > 20, the risk of developing a TSB of ≥ 25 mg/dl was about two percent. If a score of ≤ 10 was used at a cutoff for low risk, 61 percent of newborns would be in the low risk group and about 1 in 4200 would develop a TSB ≥ 25 mg/dl. Even the remaining 39 percent classified as high risk would have only a 1 in 370 risk (0.27 percent) of developing a TSB ≥ 25 mg/dl. Figure 3.7 shows the ROC curve from the Newman, Xiong, Gonzales, et al. (2000) study.

Accuracy of Predischarge Risk Zone, Determined by Hour-Specific Bilirubin Percentiles, as a Test for Predicting Later Development of Hyperbilirubinemia

Figure 3.8 Test Performance of Hour-Specific Bilirubin (more...)

   Figure 3.8 Test Performance of Hour-Specific Bilirubin Percentile

Data Source: Bhutani and Johnson, 1999 [#99117600]

Bhutani, Johnson, and Sivieri (1999) evaluated 2840 term and near term newborns to determine the efficacy of predischarge TSB in predicting clinically significant hyperbilirubinemia in the first week of life. Predischarge TSB was obtained at the time of routine metabolic screening and post discharge TSB was obtained at home or outpatient follow-up visits. Among the six percent of neonates with predischarge TSB in the high-risk zone (≥95 percentile) on an hour-specific nomogram, nearly 40% remained in that zone postdischarge. Of the 32% with TSB in the intermediate zone (40 to 95 percentile), 6 percent moved into the high-risk zone postdischarge. Among the 62 percent of newborns with TBS in the low-risk zone (<40%-ile), there was no subsequent risk for significant hyperbilirubinemia. The authors advocate predischarge TSB as a universal policy to target neonates for follow-up and intervention. Figure 3.8 shows the ROC curve from the Bhutani, Johnson, and Sivieri (1999) study.

Description of Included Studies

Table 3.20 Description of Study Outcomes

Table 3.20 Description of Study Outcomes

Author, Year, UI #	Correlation Coefficient	Sensitivity and/or Specificity	Subgroup Analyses	Subgroups
AirShields Minolta bilirubinometer
Bhat, 1987 88085341	Y	N	Y	GA/BW
Bhutta, 1991 92015688	Y	Y	N	-
Bilgen, 1998 99123665	Y	Y	N	-
Boo, 1984 85085660	Y	N	Y	Race/Skin Color
Bourchier, 1987 88247126	Y	N	Y	Race/Skin Color
Carbonell, 2001 21130776	Y	Y	Y	Other: time
Christo, 1988 89253855	Y	Y	Y	Phototherapy
Dai, 1996 97093944	N	Y	N	-
Fok, 1986 86268743	Y	Y	Y	Measurement site Phototherapy
Goldman, 1982 82241425	Y	N	Y	Race/Skin Color
Hanneman, 1982 82105249	Y	N	Y	GA/BW Race/Skin Color Phototherapy
Harish, 1998 98373176	Y	Y	Y	GA/BW Phototherapy
Hegyi, 1981 81144434	Y	N	Y	Measurement site
Karrar, 1989 89271698	Y	Y	N	-
Kenny, 1984 89271698	Y	N	Y	Measurement site Other: Feeding
Kivlaha, 1984 84296964	Y	N	N	-
Knudsen, 1989 89189829	Y	N	Y	Measurement sites
Knudsen, 1990 90242782
Knudsen, 1990 91196560	Y	Y	N	-
Knudsen, 1992 92306741	Y	Y	N	-
Knudsen, 1993 93333142	Y	Y	N	-
Knudsen, 1995 93333142	N	Y	N	-
Kumar, 1992 93084314	Y	N	N	-
Laeeq, 1993 93267865	Y	Y	N	-
Lin, 1993 93383693	Y	Y	Y	Measurement site
Linder, 1994 94325008	Y	Y	N	-
Maisels, 1982 82273864	Y	Y	Y	Measurement site
Maisels, 1997 97247177	Y	N	N	-
Schumacher, 1985 85241873	Y	Y	N	-
Serrao, 1989 85241873	Y	N	N	-
Sharma, 1988 89122266	Y	N	Y	GA/BW
Sheridan-Pereira, 1982 83021365	Y	Y	Y	Phototherapy
Smith, 1985 85112373	Y	Y	N	-
Suckling, 1995 95327901	Y	N	Y	GA/BW
Taha, 1984 85070990	Y	Y	N	-
Tan, 1985 86101839	Y	N	Y	Measure site Race/Skin Color
Tan, 1982 83044572
Tan, 1996 97017283	Y	N	Y	Measurement site Race/Skin Color
Tsai, 1988 90177720	Y	Y	Y	Measurement site
Yamauchi, 1989 89348834	Y	N	Y	Measurement site
Yamauchi, 1990 90273849	N	Y	N	-
Yamauchi, 1991 92188736	Y	N	Y	Other: Time; Sunlight
Yamauchi, 1991 92188737	Y	N	Y	Measurement site Other: Time
Yamauchi, 1988 89086035
BiliCheck™
Bhutani, 2000 20381429	Y	Y	Y	Race/Skin Color
Lodha, 2000 20368382	Y	Y	Y	Other: TSB > 13 mg/dl
Rubaltelli, 2001 21283339	Y	Y	Y	Measurement site
Ingram Icterometer
Bilgen, 1998 99123665	Y	Y	N	-
Chaibva, 1974 75018763	Y	N	N	-
Gupta, 1991 92091059	Y	Y	Y	GA/BW
Schumacher, 1985 85241873	Y	Y	N	-
Colormate III
Tayaba 1998 98393764	Y	N	Y	Race/Skin Color Phototherapy

Bold = Study reported more than one instrument

Instruments

Accuracy of Minolta AirShields Bilirubinometer

Table 3.21 Part 1. Summary Studies of Minolta AirShields (more...)

Table 3.21 Part 1. Summary Studies of Minolta AirShields bilirubinometer in Healthy, Term or Near-Term (GA≥34) Infants not on phototherapy or Exchange Transfusion

Study, Year UI#	Country	Race (N)	Reference Standard: Laboratory-based Assay of TSB	Quality
Measurement site: At Forehead
Bhutta, 1991 92015688	Pakistan	ND (63)	Autoanalyser (Astra, Beckman Instruments Inc.) using the modified Jendraasik-Grof method	A
Bilgen, 1998 99123665	Turkey	ND (96)	Bilitron 444 direct spectrophotometric method	A
Knudsen, 1990 91196560	Denmark	ND (207)	Standard diazo method	A
Knudsen, 1992 92306741	Denmark	ND (138)	Standard direct spectroscopic method	A
Lin, 1993 93383693	Taiwan	Chinese (305)	Spectro-photometric method with BB-meter-Model III	A
Schumacher, 1985 85241873	US	White (106)	Dupont ACA III analyzer, a direct spectrophotometry method	A
Tsai, 1988 90177720	Taiwan	Chinese (98)	BB-meter-Model III (O'hara Co. Ltd)	A
Fok, 1986 86268743	Hong Kong	Chinese (202)a	AO Unistat Bilirubinometer (American Optical, USA)	B
Harish, 1998 98373176	India	ND (60)	Malloy and Evelyn method	B
Karrar, 1989 89271698	Saudi Arabia	ND (155)	DuPont automatic clinical analyzer	B
Knudsen, 1993 93333142	Denmark	Danish (73)	Standard diazo method	B
Knudsen, 1995 93333142	Denmark	Danish (110)	Standard spectrophotometric method	B
Maisels, 1982 82273864	US	White (157)	Automated modified diazo method	B
Smith, 1985 85112373	US	Racial diverse (85)b	Method was not report	B
Taha, 1984 85070990	Saudi Arabia	ND (68)	AO Unistat Bilirubinometer (American Optical, USA)	B
Dai, 1996 97093944	Canada	Racial diverse (38)	Jendrassik-Grof method on a Hitachi 717 analyzer (Boehringer Mannheim, Laral, Quebec)	C
Sheridan-Pereira, 1982 83021365	Ireland	White (60)c	Method of Jendrassik Grof.	C
Measurement site: At Mid-Sternum
Lin, 1993 93383693	Taiwan	Chinese (305)	Spectro-photometric method with BB-meter-Model III	A
Fok, 1986 86268743	Hong Kong	Chinese (202)a	AO Unistat Bilirubinometer (American Optical, USA)	B
Linder, 1994 94325008	Israel	Middle-eastern (123)	AO Unistat Bilirubinometer (American Optical, USA)	B
Maisels, 1982 82273864	US	White (135)	Automated modified diazo method	B
Measurement site: Mixed Sites
Carbonell, 2001 21130776	Spain	ND (574+865)	Bilirubinometer Bil-Red by direct reading at 461 mcm	A
Yamauchi, 1990 90273849	Japan	Japanese (78)	AO Unistat Bilirubinometer (American Optical, USA)	A
Christo, 1988 89253855	South India	ND (91)	Technicon RA1000 within 0.5 and 1 hour from serum blood sample drawn	B
Laeeq, 1993 93267865	Pakistan	ND (105)	Standard diazo method	B

SB = serum bilirubin; TSB = total serum bilirubin

All infants were healthy, term (GA≥34), and not on phototherapy or exchange transfusion, unless noted:

a

1.5% G-6-PD deficiency; 11% ABO incompatible, but none of them show a positive Coomb's test

b

5% ABO incompatibility

c

“Most were white and Mexican-American extractions”

Table 3.21 Part 2. Test Accuracy of Minolta AirShields (more...)

Table 3.21 Part 2. Test Accuracy of Minolta AirShields bilirubinometer to Serum Bilirubin Levels in Healthy, Term or Near-Term (GA≥34) Infants not on phototherapy or Exchange Transfusion

Study, Year UI#	r	TP (n)	FN (n)	TN (n)	FP (n)	Sens (%)	Spec (%)	Thresholds
								TcB	SB (mg/dl)
Measurement site: At Forehead
Bhutta, 1991 92015688	0.66	35	5	32	28	88	53	17	12.5
Bilgen, 1998 99123665	0.83	17	0	44	35	100	56	13	12.9
Knudsen, 1990 91196560	0.84	28	0	61	118	100	34	15	13
		28	0	93	86	100	52	16	13
		24	4	113	66	86	63	17	13
		22	6	134	45	79	75	18	13
		15	13	158	21	54	88	19	13
	0.84	91	4	56	56	96	50	15	10.2
		83	12	81	31	87	72	16	10.2
		68	27	91	21	72	81	17	10.2
		56	39	101	11	59	90	18	10.2
		31	64	108	4	33	96	19	10.2
Knudsen, 1992 92306741	0.63	28	0	13	97	100	12	7	11.7
		27	1	29	81	96	26	8	11.7
		21	7	72	38	75	65	9	11.7
		16	12	48	62	57	81	10	11.7
		9	19	102	8	32	93	11	11.7
Lin, 1993 93383693	0.82	12	6	230	57	68	80	Var	12.9
Schumacher, 1985 85241873	0.74	16	1	69	20	94	78	20	12.9
Tsai, 1988 90177720	0.87	19	2	141	16	90	90	16	13
Fok, 1986 86268743	0.86	16	0	122	64	100	66	22	15
Harish, 1998 98373176	0.83	24	1	21	14	96	59	18	13
Karrar, 1989 89271698	0.82	36	13	95	11	74	90	21	12.5
Knudsen, 1993 93333142	0.85	30	0	5	38	100	11	13	13
		30	0	8	35	100	19	14	13
		30	0	15	28	100	35	15	13
		29	1	22	21	97	51	16	13
		28	2	32	11	93	74	17	13
Knudsen, 1995g93333142	ND	9	0	57	44	100	56	9	18
		8	1	67	34	89	66	9.5	18
		7	2	79	22	78	78	10	18
		3	6	93	8	33	92	11	18
Maisels, 1982 82273864	0.93	7	0	145	5	100	97	24	13
		20	2	121	14	91	90	20	10
Smith, 1985 85112373	0.90	6	1	64	14	86	82	Comb	12.9
Taha, 1984 85070990	0.88	11	5	96	8	69	92	22.5	12.9
Dai, 1996 97093944	ND	8	0	21	11	100	67	17	15.2
Sheridan-Pereira, 1982 83021365	0.70	4	0	40	13	100	75	20	14.5
Measurement site: Sternum
Lin, 1993 93383693	0.86	15	3	241	46	84	84	Var	12.9
Fok, 1986 86268743	0.91	10	6	158	28	63	85	22	15
Linder, 1994 94325008	0.96	19	4	136	3	86	96	ND	12.9
Maisels, 1982 82273864	0.93	4	0	126	5	100	96	23	13
		11	0	105	19	100	85	19	10
Measurement site: Mixed sites
Carbonell, 2001 21130776	0.92	15	3	368	188	83	66	11 h	17
		17	1	287	269	94	52	13 j	17
		45	1	262	557	98	32	13 k	17
Yamauchi, 1990 90273849	ND	17	5	35	21	77	63	14 h	15
		20	2	47	9	91	84	21 i	15
Christo, 1988 89253855	0.90	16	0	65	10	100	87	17	13
		9	0	24	2	100	92	23	17
Laeeq, 1993 93267865	0.77	-	-	-	-	97	41	ND	6
		-	-	-	-	91	78	ND	9.92
		-	-	-	-	90	99	ND	15

TP = true positive; FN = false negative; TN = true negative; FP = false positive; N = number of subjects; n = number of measurements; Sens = sensitivity; Spec = specificity; TcB = transcutaneous bilirubin; SB = serum bilirubin; r = regression coefficient between TcB and SB

r = Pearson's correlation coefficient

Comb = Elevated TcB, Elevated ETco, or Elevated both TcB and Etco. “Elevated” was defined as more than 1 SD above the mean

Var = TcB ≥ 11, ≥ 16, or ≥ 20 on the 1st, 2nd, and following days of life

g

Corrected TcB was use: TcB reading taken at median 21 hours of life was “corrected” by the difference in TcB readings obtained 6 hours later.

h

TcB reading at 24 hours of life

i

TcB reading on day 3 of life

j

TcB reading at 48 hours of life in the phase I study

k

TcB reading at 48 hours of life in the phase II study

An additional challenge to comparison across studies is that each instrument must be calibrated on a daily basis as recommended by the manufacturer and each institution must establish the relationship between the jaundice meter index, TcB, and the TSB, in part due to interlaboratory variations of TSB measurements (Schreiner and Glick, 1982). As can be seen in the table, the TcB index that correlates with a given level of TSB varies from center to center, and as discussed above, the manufacturer recommends that each site develop its own correlation curves.

Overall Meta-analysis of Correlation Coefficients for Minolta AirShields Bilirubinometer

In 38 of the 41 studies, linear regression analysis was performed generating Pearson's correlation coefficient (r) for TcB versus TSB. Overall, it does appear that a linear relationship between TcB as measured by Minolta AirShields bilirubinometer and the TSB in predominately term or near-term infants not on phototherapy. The pooled correlation coefficient was 0.84 (95% CI 0.81–0.87) (see Figure S3 in Meta-Analyses).

Factors Affecting Test Accuracy of Minolta AirShields Bilirubinometer

Measurement Sites

Table 3.22 Overall Correlation Coefficient of Minolta (more...)

Table 3.22 Overall Correlation Coefficient of Minolta AirShields Measurements to Serum Bilirubin Levels on Different Measurement Sites in Healthy Term Infants with (GA ≥ 34 weeks) not on Phototherapy

Study, Year UI #	Country	Race	Sample N	Correlation Coefficient (r)							Quality
				Forehead	Sternum	Abdomen	Upper back	Lower back	Palm	Sole
Lin, 1993 93383693	Taiwan	Chinese	202	0.82	0.86	-	-	-	-	-	A
Tsai, 1988 90177720	Taiwan	Chinese	98	0.87	0.78	0.76	0.69	0.69	0.49	0.47	A
Yamauchi, 1991 92188737	Japan	Japanese	336	0.91	0.92	0.89	0.89	0.88	-	0.77	A
Yamauchi, 1988 89086035
Fok, 1986 86268743	Hong Kong	Chinese	202	0.86	0.91	-	-	-	-	-	B
Hegyi, 1981 81144434	US	White	43	0.77	0.90	0.88	0.87	0.86	0.72	0.72	B
Kenny, 1984 89271698	US	White	18	0.85	0.93	-	0.83	-	-	-	B
Knudsen, 1989 89189829	Denmark	ND	76	0.86	-	0.89	-	-	-	-	B
Knudsen, 1990 90242782
Maisels, 1982 82273864	US	White	135	0.93	0.93	-	-	-	-	-	B
Tan, 1985 86101839	Singapore	Chinese	224	0.93	0.94	-	0.86	-	-	-	B
Tan, 1982 83044572		Malay	303	0.87	0.90	-	0.87	-	-	-
Tan, 1996 97017283	Singapore	Racial diverse a	542	0.80	0.75	-	-	-	-	-	B
Yamauchi, 1989 89348834	Japan	Japanese	114	0.90	0.92	-	-	-	-	-	B

a

Chinese + Malay + Indian infants

In the Table 3.22 below summarizing those studies that evaluated multiple measurement sites, the Minolta AirShields bilirubinometer appears to perform similarly well at the forehead site and the sternum, with the sternum site slightly better. Correlation coefficients were lower at the palm and sole. Meta-analysis of correlation coefficient (r) was performed for forehead (Sample N=1,936) and sternum sites (Sample N=1,815). The pooled r at forehead site was 0.87 (95%CI 0.82–0.90) and that at sternum site was 0.90 (95%CI 0.86–0.92) (see Figure S4 in Meta-Analyses).

Gestational Age or Birth Weight

Table 3.23 Summary of Studies on the Effects of Gestational (more...)

Table 3.23 Summary of Studies on the Effects of Gestational Age or Birth Weight to the Accuracy of Minolta TcB measurements at Forehead in Term or Near-Term Infants not on Phototherapy

Study, Year UI #	Country	Race (N)	Subgroups	Pairs (N)	Coefficient (r)	Significance	Quality
Bhat, 1987 88085341	India	ND (100)	GA ≥ 37 wk	63	0.72	P=.123	B
			GA < 37 wk	37	0.52
			BW > 2500 g	37	1.03 ?	n.d
			BW ≤ 2500 g	63	0.94
Hanneman, 1982 82105249	US	White (57)	Infants GA≥38 wks	35	0.90	P=.739	B
			Infants GA 34–37 wks	22	0.88
Harish, 1998 98373176	India	ND (33)	Term infants	60	0.83	P=.363	B
			SGA term + Preterm infants	40	0.76
Sharma, 1988 89122266	India	ND (200)	Term infants	120	0.74	P=.485	B
			Preterm infants	80	0.69
			BW ≥ 2.5 Kg	110	0.85	P=.371
			BW < 2.5 Kg	90	0.81
Suckling, 1995 95327901	Scotland	White (55)	Term AGA	37	0.802	P=.33	C
			Preterm or SGA	39	0.702

n.d. = not done, because the correlation 1.03 is not valid

Bold = significantly different between subgroups (p < .10)

? = Unknown

In the five studies that gave separate correlation coefficients shown in Table 3.23 for test performance between term and near-term infants, there were no significant differences seen in the studies although a trend of lower correlation of the TcB with TSB in the near-term infants is seen. Meta-analysis of correlation coefficient (r) was performed for term infants with gestational age ≥ 37 weeks or infants with birth weight > 2,500 grams, and near-term infants with gestational age 34–37 weeks or term infants with a birth weight < 2,500 grams. Meta-analysis of correlation coefficient (r) was performed for term (N=425 pairs) and sternum sites (N=308 pairs). The pooled r was 0.81 (95% CI 0.75–0.86) and 0.74 (95% CI 0.64–0.82) for term and near-term infants respectively (see Figure S5 in Meta-Analyses).

Race (or Skin Color)

Table 3.24 Summary of Studies on the Effects of Race (more...)

Table 3.24 Summary of Studies on the Effects of Race / Skin Color to the Accuracy of Minolta TcB measurements at Forehead in Term or Near-Term Infants not on Phototherapy

Study, Year UI #	Country	Race (N)	Subgroups	Pairs (N)	Coefficient (r)	Significance	Quality
Boo, 1984 85085660	Malaysia	Mixed (105)	Malay infants	38	0.83	M vs. C: P=.101	B
			Chinese infants	37	0.66	M vs. I: P=.081
			Indian infants	30	0.63	C vs. I: P=.842
Bourchier, 1987 88247126	New Zealand	Mixed (729)	Caucasian infants	412	0.58	P=.842	B
			Maori infants	317	0.57
			Black infants	108	0.52
Hanneman, 1982 82105249	US	Mixed (99)	White Infants GA≥34 wks	57	0.88–0.90	P=.025	B
			Black infants GA≥34 wks	42	0.74
Tan, 1985 86101839	Singapore	Mixed (125)	Chinese infants	102	0.93	P=.018	B
Tan, 1982 83044572			Malay infants	103	0.87
Tan, 1996 97017283	Singapore	Mixed (542)	Chinese infants	253	0.73	M vs. C: P=.003	B
			Malay infants	169	0.84	M vs. I: P=.435
			Indian infants	120	0.81	C vs. I: P=.077
Goldman, 1982 82241425	US	Mixed (84)	White infants	95	0.71	P=.035	C
			Black infants	108	0.52

Bold = significantly different between subgroups (p < .10)

Table 3.25 Summary of Studies on the Effects of Race (more...)

Table 3.25 Summary of Studies on the Effects of Race / Skin Color to the Accuracy of Minolta TcB measurements at Sternum in Term or Near-Term Infants not on Phototherapy

Study, Year UI #	Country	Race (N)	Subgroups	Pairs (N)	Coefficient (r)	Significance	Quality
Bourchier, 1987 88247126	New Zealand	Mixed (729)	Caucasian infants	412	0.67	P=.253	B
			Maori infants	317	0.62
Tan, 1985 86101839	Singapore	Mixed (125)	Chinese infants	69	0.93	P=.115	B
Tan, 1982 83044572			Malay infants	103	0.90
Tan, 1996 97017283	Singapore	Mixed (542)	Chinese infants	253	0.78	M vs. C: P=.004	B
			Malay infants	169	0.87	M vs. I: P=.230
			Indian infants	120	0.83	C vs. I: P=.202

Bold = significantly different between subgroups (p < .10)

Tables 3.24 and 3.25 below summarize those studies that compared correlation coefficients across different races or skin color. Significant differences with p < 0.10 are highlighted in bold print with higher correlation of TcB with TSB in Malay (r=0.83) versus Indian infants (r=0.63) in one study (Boo and Bakar, 1984) but no difference in another (Tan, Chia, and Koh, 1996); white (r=0.88–0.90) versus black infants (r=0. 74) (Goldman, Penalver, and Penaranda, 1982; Hannemann, Schreiner, DeWitt, et al., 1982); Chinese (r=0.93) versus Malay (r=0.87) (Tan, 1982; Tan, 1985) although in the opposite direction in a subsequent study with Malay (r=0.84 at forehead, 0.87 at sternum) versus Chinese (r=0.73 at forehead, 0.78 at sternum) (Tan, Chia, and Koh, 1996).

Meta-analysis of correlation coefficient (r) was performed for different races at forehead. A total of 564, 258, 392, 150, and 310 pairs of measurements were available for white, black, Chinese, Indian and Malay infants. At forehead, the pooled r for white infants was 0.65 (95%CI 0.59–0.69), for black infants was 0.56 (95%CI 0.47–0.64), for Chinese infants was 0.80 (95%CI 0.77–0.84), for Indian infants was 0.78 (95%CI 0.71–0.84), and for Malay infants was 0.85 (95%CI 0.82–0.88) (see Figure S6 in Meta-Analyses).