Summary
MORTALITY Low threshold compared with high threshold phototherapy: Low threshold phototherapy (initiation of phototherapy at enrolment, serum bilirubin [SBR] expected to be 85 micromol/L and recommencement of phototherapy if SBR >85 micromol/L in first 7 days of life or SBR >137 micromol/L from day 7–14 of life) and high threshold phototherapy (initiation of phototherapy at 137 micromol/L and recommencement if SBR >137 micromol/L in first 7 days and SBR >171 micromol/L from day 7–14 of life) seem equally effective at improving mortality before day 15, mortality before discharge, and mortality or the composite outcome of mortality and neurodevelopmental impairment at 18 to 22 months in extremely low birth weight infants 12 to 36 hours old with non-severe haemolytic disease and absence of major congenital abnormality ( moderate-quality evidence ). NEUROLOGICAL/NEURODEVELOPMENTAL Conventional phototherapy compared with no treatment: We don't know whether conventional phototherapy is more effective than no treatment at reducing cerebral palsy or other motor abnormalities (clumsiness, hypotonia, abnormal movement) at 1 or 6 years in infants with hyperbilirubinaemia as we found insufficient evidence from one RCT. The study did not use very intensive phototherapy ( low-quality evidence ). Prophylactic phototherapy compared with threshold phototherapy: We don't know whether prophylactic phototherapy (commencement of phototherapy within 12 hours of birth) is more effective than threshold phototherapy (commencement once SBR >150 micromol/L) at reducing the proportion of infants with either cerebral palsy, with the composite outcome of cerebral palsy or death, or with an abnormal developmental index score at 18 months, in infants of birth weight <1500 g within 12 hours of birth without isoimmunisation or major life-threatening anomaly ( very low-quality evidence ). Low threshold compared with high threshold phototherapy: Low threshold phototherapy (initiation of phototherapy at enrolment, SBR expected to be 85 micromol/L and recommencement of phototherapy if SBR >85 micromol/L in first 7 days of life or SBR >137 micromol/L from day 7–14 of life) seems more effective than high threshold phototherapy (initiation of phototherapy at 137 micromol/L and recommencement if SBR >137 micromol/L in first 7 days and SBR >171 micromol/L from day 7–14 of life) at reducing the proportion of infants with neurodevelopmental impairment, profound impairment, and severe hearing loss at 18 to 22 months in extremely low birth weight infants 12 to 36 hours old with non-severe haemolytic disease and absence of major congenital abnormality, but we don't know about cerebral palsy or blindness (moderate-quality evidence). NEED FOR EXCHANGE TRANSFUSION Conventional phototherapy compared with no treatment: Conventional phototherapy may be more effective than no treatment at reducing exchange transfusion in infants with established hyperbilirubinaemia and birth weight of 2000–2499 g, but we don't know about in infants of 2500 g or over. Subgroup analysis suggests that conventional phototherapy may be more effective than no treatment at reducing exchange transfusion in infants with established hyperbilirubinaemia with non-haemolytic jaundice, but we don't know about in infants with haemolytic jaundice. The study did not use very intensive phototherapy, which may explain the lack of effect of phototherapy in some subgroups (low-quality evidence). Conventional phototherapy compared with fibreoptic phototherapy: We don't know whether conventional phototherapy and fibreoptic phototherapy differ in effectiveness at reducing the proportion of infants who require exchange transfusion (very low-quality evidence). Double compared with single phototherapy: We don't know whether fibreoptic plus conventional phototherapy is more effective than conventional phototherapy alone at reducing the rate of exchange transfusion as we found insufficient evidence from one small RCT (low-quality evidence). Low threshold compared with high threshold phototherapy: Low threshold phototherapy (initiation of phototherapy at enrolment, SBR expected to be 85 micromol/L and recommencement of phototherapy if SBR >85 micromol/L in first 7 days of life or SBR >137 micromol/L from day 7–14 of life) and high threshold phototherapy (initiation of phototherapy at 137 micromol/L and recommencement if SBR >137 micromol/L in first 7 days and SBR >171 micromol/L from day 7–14 of life) seem equally effective at preventing the need for exchange transfusions in extremely low birth weight infants 12 to 36 hours old with non-severe haemolytic disease and absence of major congenital abnormality (moderate-quality evidence). Phototherapy compared with immunoglobulin: Phototherapy (continuous or intermittent blue light exposure in analysis) may be less effective than intravenous immunoglobulin for 3 consecutive days at reducing the rate of exchange transfusion (further details not reported) in neonates diagnosed with ABO haemolytic disease of the newborn (ABO-HDN), mother with blood type O, anti-A/B valence >1:128, blood type A or B of the infant, positive Coombs' test and/or positive free antibody test and/or positive antibody release test, and with clinical symptoms of haemolysis, jaundice, and anaemia (very low-quality evidence). DURATION OF TREATMENT Fibreoptic phototherapy compared with no treatment: We don't know whether fibreoptic phototherapy is more effective than no treatment at reducing the proportion of infants who require conventional phototherapy in term infants with haemolysis excluded (low-quality evidence). Conventional phototherapy compared with fibreoptic phototherapy: Conventional phototherapy may be more effective than fibreoptic phototherapy at reducing the duration of phototherapy treatment and the use of additional phototherapy treatment in term and preterm infants analysed as a group, but we don't know about duration of phototherapy or additional phototherapy in preterm infants alone (very low-quality evidence). Double compared with single phototherapy: Double conventional phototherapy (using daylight fluorescent lamps) may be more effective than single conventional phototherapy at reducing the duration of treatment in term infants of birth weight 2500 g or above with haemolysis included. Conventional phototherapy plus fibreoptic Wallaby phototherapy may be more effective than Wallaby, BiliBlanket, or conventional phototherapy alone at reducing mean duration of treatment in preterm infants of <31 weeks' gestation with haemolytic jaundice excluded. We don't know whether fibreoptic plus conventional phototherapy is more effective than single conventional phototherapy in reducing the need for additional phototherapy or repeat phototherapy for rebound jaundice in term and preterm infants. We don't know whether double fibreoptic phototherapy (infants wrapped in 2 BiliBlankets) is more effective than single conventional therapy at reducing duration of treatment or use of repeat phototherapy in term infants with haemolysis excluded (very low-quality evidence). Intermittent phototherapy compared with continuous phototherapy: We don't know whether intermittent equal duration phototherapy (4 hours on, 4 hours off), intermittent short duration phototherapy (1 hour on, 3 hours off), and continuous phototherapy differ in effectiveness at reducing duration of phototherapy treatment in term infants who are 2500 g or above with physiological jaundice (low-quality evidence). Close phototherapy compared with distant light-source phototherapy: Close light-source conventional phototherapy from a distance of 20 cm above the neonate seems more effective than more distant light-source phototherapy at 40 cm above the neonate at reducing mean duration of treatment in infants with hyperbilirubinaemia not severe enough to require exchange transfusion and with absence of congenital metabolic disorders (moderate-quality evidence). Increased skin exposure compared with standard skin exposure phototherapy: We don't know whether conventional phototherapy in partially clothed infants (disposable nappy only) and conventional phototherapy in naked infants differ in effectiveness at reducing the proportion of infants still requiring phototherapy at 24 to 48 hours and 48 to 72 hours in preterm infants of 1500 g or more and 36 weeks' gestation or less with non-haemolytic hyperbilirubinaemia with total serum bilirubin in the range for phototherapy (low-quality evidence). Fluorescent compared with blue fluorescent lamps: We don't know whether daylight fluorescent lamps, standard blue fluorescent lamps, and blue fluorescent lamps with a narrow spectral emission differ in effectiveness at improving the proportion of infants who discontinue phototherapy after 1 to 3 days in infants with hyperbilirubinaemia in the first 72 hours of life (non-haemolytic) as the RCT did not test the significance of differences between groups (low-quality evidence). Blue fluorescent compared with green fluorescent lamps: Blue fluorescent lamps and green fluorescent lamps seem equally effective at reducing the duration of phototherapy in term and preterm infants with non-haemolytic jaundice (moderate-quality evidence). Blue-green fluorescent compared with blue fluorescent lamps: Blue-green fluorescent light may be more effective than blue fluorescent light at reducing the proportion of infants requiring phototherapy after 24 hours in healthy low birth weight infants with hyperbilirubinaemia in the first 4 days of life, but we don't know whether 6 focused arrays of blue-green LED phototherapy is more effective than 6 focused arrays of blue LED phototherapy at increasing the mean rate of serum bilirubin decline in jaundiced but otherwise healthy term infants (very low-quality evidence). Blue LED compared with conventional quartz-halogen: We don't know whether blue LED and conventional phototherapy (using halogen-quartz bulbs) differ in effectiveness at reducing the mean number of hours spent under phototherapy as we found insufficient evidence (low-quality evidence). Blue-green LED compared with conventional quartz-halogen: We don't know whether 6 focused arrays of blue-green LED phototherapy and conventional phototherapy consisting of three halogen-quartz bulbs differ in effectiveness at reducing the mean number of hours of phototherapy in jaundiced but otherwise healthy term infants (low-quality evidence). Prophylactic phototherapy compared with threshold phototherapy: We don't know whether prophylactic phototherapy (commencement of phototherapy within 12 hours of birth) is more effective than threshold phototherapy (commencement once SBR >150 micromol/L) at reducing the mean number of hours of phototherapy in infants of birth weight <1500 g within 12 hours of birth without isoimmunisation or major life-threatening anomaly (low-quality evidence). Low threshold compared with high threshold phototherapy: Low threshold phototherapy (initiation of phototherapy at enrolment, SBR expected to be 85 micromol/L and recommencement of phototherapy if SBR >85 micromol/L in first 7 days of life or SBR >137 micromol/L from day 7–14 of life) seems less effective than high threshold phototherapy (initiation of phototherapy at 137 micromol/L and recommencement if SBR >137 micromol/L in first 7 days and SBR >171 micromol/L from day 7–14 of life) at reducing the duration of phototherapy in extremely low birth weight infants 12 to 36 hours old with non-severe haemolytic disease and absence of major congenital abnormality. We don't know whether commencing phototherapy once SBR is >13 mg/dL, commencing phototherapy once SBR is 10 mg/dL or above and treating for 12 hours, and commencing phototherapy once SBR is 10 mg/dL or above and treating for 24 hours differ in effectiveness at reducing the proportion of infants who receive phototherapy at <72 hours or >72 hours in infants of birth weight <2500 g (moderate-quality evidence). Compact fluorescent light phototherapy compared with standard length tube light phototherapy: We don't know whether compact fluorescent light phototherapy is more effective than standard length tube light phototherapy at reducing the total duration of treatment required in infants above 34 weeks' gestation with haemolytic jaundice excluded (low-quality evidence). Phototherapy compared with immunoglobulin: Phototherapy (continuous or intermittent blue light exposure in analysis) may be less effective than intravenous immunoglobulin for 3 consecutive days at reducing the duration of phototherapy in neonates diagnosed with ABO-HDN, mother with blood type O, anti-A/B valence >1:128, blood type A or B of the infant, positive Coombs' test and/or positive free antibody test and/or positive antibody release test, and with clinical symptoms of haemolysis, jaundice, and anaemia (low-quality evidence). SERUM BILIRUBIN LEVEL Conventional phototherapy compared with no treatment: Conventional phototherapy may be more effective than no treatment at reducing the proportion of infants with maximal serum bilirubin levels and may be more effective at reducing mean serum bilirubin levels in infants with hyperbilirubinaemia (low-quality evidence). Continuous phototherapy compared with no treatment: We don't know whether continuous phototherapy is more effective than no treatment at reducing the proportion of infants with serum bilirubin levels >12 mg/dL and >15 mg/dL in preterm infants of birth weight 1250–2000 g who were Coombs' negative with no haemolytic anaemia as the RCT did not test differences between groups. However, absolute rates were lower in the continuous phototherapy group (low-quality evidence). Intermittent phototherapy compared with no treatment: We don't know whether intermittent phototherapy (12 hours on, 12 hours off) is more effective than no treatment at reducing the proportion of infants with serum bilirubin levels >12 mg/dL and >15 mg/dL in preterm infants of birth weight 1250–2000 g who were Coombs' negative with no haemolytic anaemia as the RCT did not test differences between groups. However, absolute rates were lower in the intermittent phototherapy group (low-quality evidence). Fibreoptic phototherapy compared with no treatment: Fibreoptic phototherapy may be more effective than no treatment at increasing the percentage change in serum bilirubin per hour and the percentage change after 24 hours of treatment in term infants with haemolysis excluded (low-quality evidence). Conventional phototherapy compared with fibreoptic phototherapy: Conventional phototherapy may be more effective than fibreoptic phototherapy at increasing the percentage change in serum bilirubin at 24 hours and 48 hours in term and preterm infants analysed as a group, but we don't know about percentage change in serum bilirubin at 24 hours in preterm infants alone (very low-quality evidence). Double phototherapy compared with single phototherapy: Double conventional phototherapy (using daylight fluorescent lamps) may be more effective than single conventional phototherapy at increasing the rate of reduction of serum bilirubin in term infants of birth weight 2500 g or above with haemolysis included. Conventional phototherapy plus fibreoptic Wallaby phototherapy may be more effective than Wallaby, BiliBlanket, or conventional phototherapy alone at reducing the increase in bilirubin levels over the first 24 hours in preterm infants of <31 weeks' gestation with haemolytic jaundice excluded. We don't know whether fibreoptic plus conventional phototherapy is more effective than single conventional phototherapy at improving the percentage change in serum bilirubin levels after 24 or 48 hours in term and preterm infants. We don't know whether double fibreoptic phototherapy (infants wrapped in 2 BiliBlankets) is more effective than single conventional therapy at improving the percentage change in serum bilirubin per hour or per day in term infants with haemolysis excluded. Double surface phototherapy may be more effective than single surface phototherapy at increasing the total decline in serum bilirubin levels after 48 hours in term infants of 2500 g or above with non-haemolytic hyperbilirubinaemia who were exclusively breastfed, but we don't know whether it is more effective at 24 to 48 hours (very low-quality evidence). Triple phototherapy compared with double phototherapy: We don't know whether triple phototherapy (2 single fluorescent lamps 25 cm above bed plus third fluorescent lamp 35 cm from bed) is more effective than double phototherapy (2 single fluorescent lamps 25 cm above bed) at improving mean bilirubin levels at 8, 16, or 24 hours in infants of 2500 g or more and of 37 weeks' gestation or above with non-haemolytic jaundice (low-quality evidence). Intermittent phototherapy compared with continuous phototherapy: We don't know whether intermittent phototherapy (1 hour on, 1 hour off) is more effective than continuous phototherapy (2 hours on, 30 minutes off) at improving mean serum bilirubin levels at 12, 24, 36, and 48 hours in infants above 2000 g with hyperbilirubinaemia not exceeding the range for exchange transfusion nor requiring high intensity phototherapy. We don't know whether continuous phototherapy and intermittent phototherapy (12 hours on, 12 hours off) differ in effectiveness at reducing the proportion of infants with serum bilirubin levels of >12 mg/dL or >15 mg/dL in preterm infants with birth weight 1250–2000 g who were Coombs' negative with no haemolytic anaemia as the trial did not test the significance of differences between groups. We don't know whether intermittent equal duration phototherapy (4 hours on, 4 hours off), intermittent short duration phototherapy (1 hour on, 3 hours off), and continuous phototherapy differ in effectiveness at slowing the rate of increase in bilirubin levels or improving the rate of decrease of bilirubin levels in term infants of 2500 g or above with physiological jaundice (low-quality evidence). Increased skin exposure compared with standard skin exposure phototherapy: We don't know whether conventional phototherapy in partially clothed infants (disposable nappy only) and conventional phototherapy in naked infants differ in effectiveness at improving the mean percentage decline in serum bilirubin levels or the absolute change in mean serum bilirubin levels in preterm infants of 1500 g or more and of 36 weeks' gestation or less with non-haemolytic hyperbilirubinaemia with total serum bilirubin in the range for phototherapy (low-quality evidence). Fluorescent lamps compared with halide lamps: We don't know whether fluorescent and halide lamps differ in effectiveness at improving serum bilirubin levels at 24 hours in infants with non-haemolytic hyperbilirubinaemia of 40 hours of age or more (low-quality evidence). Fluorescent compared with blue fluorescent lamps: We don't know whether daylight fluorescent lamps, standard blue fluorescent lamps, and blue fluorescent lamps with a narrow spectral emission differ in effectiveness at improving the mean decrease of serum bilirubin levels at 1 to 3 days in infants with hyperbilirubinaemia in the first 72 hours of life (non-haemolytic) as the RCT did not test the significance of differences between groups (low-quality evidence). Blue fluorescent compared with green fluorescent lamps: We don't know whether blue fluorescent lamps and green fluorescent lamps differ in effectiveness at increasing the rate of fall of serum bilirubin in term and preterm infants with non-haemolytic jaundice or in low birth weight infants with non-haemolytic jaundice stratified by initial serum bilirubin levels (21–16.1 mg/dL; 16–12.1 mg/dL; 12–9.0 mg/dL) (low-quality evidence). Blue-green fluorescent compared with blue fluorescent lamps: We don't know whether blue-green fluorescent and blue fluorescent phototherapy differ in effectiveness at improving serum bilirubin levels. There were conflicting results between trials depending on the population studied, the exact intervention used, and the analysis undertaken (very low-quality evidence). Blue LED compared with conventional quartz-halogen: We don't know whether blue LED and conventional phototherapy (using halogen-quartz bulbs) differ in effectiveness at improving the rate of serum bilirubin decline as we found insufficient evidence (low-quality evidence). Blue LED compared with blue fluorescent lamps: We don't know whether blue LED (overhead neoBLUE LED plus either BiliBlanket or Wallaby system underneath) and phototherapy with blue fluorescent lights (8 overhead blue fluorescent lights plus either BiliBlanket or Wallaby system underneath) differ in effectiveness at improving the decline of serum bilirubin levels in infants of 35 weeks' gestation or more (very low-quality evidence). Blue-green LED compared with conventional quartz-halogen: We don't know whether 6 focused arrays of blue-green LED phototherapy and conventional phototherapy consisting of three halogen-quartz bulbs differ in effectiveness at improving the mean rate of serum bilirubin decline in jaundiced but otherwise healthy term infants (low-quality evidence). Prophylactic phototherapy compared with threshold phototherapy: We don't know whether prophylactic phototherapy (commencement of phototherapy within 12 hours of birth) is more effective than threshold phototherapy (commencement once SBR >150 micromol/L) at reducing peak unconjugated serum bilirubin levels in infants of birth weight <1500 g within 12 hours of birth without isoimmunisation or major life-threatening anomaly. Subgroup analysis suggests that prophylactic phototherapy may be more effective than threshold phototherapy at reducing peak unconjugated serum bilirubin levels in infants with a birth weight <1000 g, but not in infants with a birth weight of 1000–1499 g (low-quality evidence). Low threshold compared with high threshold phototherapy: Low threshold phototherapy (initiation of phototherapy at enrolment, SBR expected to be 85 micromol/L and recommencement of phototherapy if SBR >85 micromol/L in first 7 days of life or SBR >137 micromol/L from day 7–14 of life) seems more effective than high threshold phototherapy (initiation of phototherapy at 137 micromol/L and recommencement if SBR >137 micromol/L in first 7 days and SBR >171 micromol/L from day 7–14 of life) at decreasing the level of serum bilirubin at day 5 in extremely low birth weight infants 12 to 36 hours old with non-severe haemolytic disease and absence of major congenital abnormality (moderate-quality evidence). Compact fluorescent light phototherapy compared with standard length tube light phototherapy: Compact fluorescent light phototherapy may be more effective than standard length tube light phototherapy at reducing the mean total serum bilirubin over 12 hours in infants above 34 weeks' gestation with haemolytic jaundice excluded (low-quality evidence). Phototherapy compared with immunoglobulin: Phototherapy (continuous or intermittent blue light exposure in analysis) may be less effective than intravenous immunoglobulin for 3 consecutive days at lowering serum bilirubin levels and increasing the drop in serum bilirubin levels at 3 days in neonates diagnosed with ABO-HDN, mother with blood type O, anti-A/B valence >1:128, blood type A or B of the infant, positive Coombs' test and/or positive free antibody test and/or positive antibody release test, and with clinical symptoms of haemolysis, jaundice, and anaemia (low-quality evidence).
Benefits
Conventional phototherapy versus no treatment:
We found one systematic review (search date 2001), which included two RCTs,
although their results were not combined statistically. Neither of these studies assessed the primary outcomes of this review.
The largest RCT identified by the review compared conventional phototherapy using daylight fluorescent lamps versus no treatment in three birth weight groups (<2000 g, 2000–2499 g, and 2500 g and over). Exchange transfusion was given at predetermined serum bilirubin levels in each group. The RCT examined prevention of hyperbilirubinaemia in the lowest birth weight group, and treatment of established hyperbilirubinaemia in the remaining two groups. Only the results of treatment of established hyperbilirubinaemia are reported here. The RCT found that in the 2000–2499 g birth weight group (141 infants, serum bilirubin [SBR] 171 micromol/L or more, average 212 micromol/L), phototherapy significantly reduced the proportion of infants with higher maximal serum bilirubin levels compared with no treatment (SBR 257 micromol/L or more: 18.6% with phototherapy v 42.3% with no treatment; P = 0.002). For this group, it found that phototherapy significantly decreased the proportion of infants who needed exchange transfusion compared with no treatment (4.3% with phototherapy v 25.4% with no treatment; P <0.001). The RCT found that, in the 2500 g or over birth weight group (276 infants, SBR 222 micromol/L or more, average 267–268 micromol/L), phototherapy significantly reduced mean serum bilirubin levels until 24 hours after stopping treatment compared with no treatment (results presented graphically; P value not reported). The RCT found no significant difference between phototherapy and no treatment in the proportion of infants who needed exchange transfusion (10.0% with phototherapy v 16.9% with no treatment; reported as not significant). In both birth weight groups, subgroup analysis suggested that, in infants with non-haemolytic jaundice, phototherapy significantly decreased exchange transfusion compared with no treatment (infants 2000–2499 g: 1.9% with phototherapy v 27.5% with no treatment; P = 0.0002; infants 2500 g or more: 2.9% with phototherapy v 17.3% with no treatment; P = 0.05), but there was no evidence of effectiveness in preventing exchange transfusion in infants with haemolytic jaundice. A subsequent report of the RCT noted that there were two deaths before hospital discharge (2000–2499 g group: 1 with phototherapy v 1 with no treatment; 2500 g or more group: none). A further follow-up report of the RCT found no significant difference in cerebral palsy or other motor abnormalities (clumsiness, hypotonia, abnormal movement) after 1 and 6 years in either of the birth weight groups.
The smaller RCT identified by the review compared the effect of 4 interventions on hyperbilirubinaemia (SBR >291 micromol/L) on 125 term breastfed infants. The 4 interventions were: continue breastfeeding and observe (25 infants), substitute formula feed (26 infants), continue breastfeeding and administer phototherapy (36 infants), and substitute formula and administer phototherapy (38 infants). Phototherapy resulted in a smaller proportion of infants whose serum bilirubin rose above 342 micromol/L (8.1% of 74 infants receiving phototherapy v 21.6% of 51 infants not receiving phototherapy; RR 0.34, 95% CI 0.15 to 0.95; see comment).
Continuous phototherapy versus no treatment:
We found one RCT comparing continuous phototherapy versus no treatment.
The three-armed RCT (120 preterm infants, birth weight 1250–2000 g, Coombs' negative, no haemolytic anaemia, no gross congenital anomalies, no severe respiratory distress syndrome) compared continuous phototherapy versus no treatment versus intermittent phototherapy (12 hours on, 12 hours off) for 5 days. We only report the data for continuous phototherapy compared with no treatment here. The RCT found that a smaller proportion of preterm infants treated with continuous phototherapy had serum bilirubin levels >12 mg/dL and >15 mg/dL (note: 1 mg/dL = 17.1 micromol/L [SI unit]) compared with no treatment (>12 mg/dL: 2/40 [5%] with continuous phototherapy v 14/40 [35%] with no treatment; >15 mg/dL: 0/40 [0%] with continuous phototherapy v 5/40 [13%] with no treatment; P values not reported).
Intermittent phototherapy versus no treatment:
We found one RCT comparing intermittent phototherapy versus no treatment.
The three-armed RCT (120 preterm infants, birth weight 1250–2000 g, Coombs' negative, no haemolytic anaemia, no gross congenital anomalies, no severe respiratory distress syndrome) compared continuous phototherapy versus no treatment versus intermittent phototherapy (12 hours on, 12 hours off) for 5 days. We only report the data for intermittent phototherapy compared with no treatment here. The RCT found that a smaller proportion of preterm infants treated with intermittent phototherapy had serum bilirubin levels >12 mg/dL and >15 mg/dL compared with no treatment (>12 mg/dL: 3/40 [8%] with intermittent phototherapy v 14/40 [35%] with no treatment; >15 mg/dL: 1/40 [3%] with intermittent phototherapy v 5/40 [13%] with no treatment; P values not reported).
Fibreoptic phototherapy versus no treatment:
We found one systematic review (search date 2000; term and preterm infants; randomised and quasi-randomised trials; see comment below). The review identified one RCT (46 term infants, haemolysis excluded), which compared fibreoptic phototherapy (Wallaby system) versus no treatment. Conventional phototherapy was commenced if the serum bilirubin reached predetermined levels. The review found that, compared with no treatment, fibreoptic phototherapy significantly increased the percentage change in serum bilirubin per hour (WMD –0.44%, 95% CI –0.67% to –0.21%) and the percentage change after 24 hours of treatment (WMD –10.70%, 95% CI –18.14% to –3.26%). It found that infants in the fibreoptic phototherapy group were less likely to require conventional phototherapy, but this did not reach significance (0/23 [0%] with fibreoptic phototherapy v 3/23 [13%] with no treatment; RR 0.14, 95% CI 0.01 to 2.62).
Conventional versus fibreoptic phototherapy:
We found one systematic review (search date 2000; term and preterm infants; randomised and quasi-randomised trials; see comment below) and two subsequent RCTs.
The review found that conventional phototherapy significantly increased the percentage change in serum bilirubin after 24 and 48 hours of treatment compared with fibreoptic phototherapy (24 hours: 5 trials, 203 infants; WMD 3.59%, 95% CI 1.27% to 5.92%; 48 hours: 4 trials, 183 infants; WMD 10.79%, 95% CI 8.33% to 13.26%). It also found that fibreoptic phototherapy significantly increased the use of additional phototherapy compared with conventional phototherapy (8 trials: 52/366 [14%] with fibreoptic v 35/390 [9%] with conventional; RR 1.68, 95% CI 1.18 to 2.38), and also resulted in an increase in duration of phototherapy treatment (6 trials, 562 infants: WMD +13.6 hours, 95% CI +10.1 hours to +17.1 hours). It found no significant difference between fibreoptic and conventional phototherapy in the use of exchange transfusion (3 trials: 4/97 [4%] with fibreoptic v 3/117 [3%] with conventional; RR 1.62, 95% CI 0.38 to 6.93). In a subgroup analysis of preterm babies only, the review found no significant difference between fibreoptic phototherapy and conventional phototherapy in the duration of phototherapy, use of additional phototherapy, percentage change in serum bilirubin after 24 hours of treatment, percentage change in serum bilirubin after 24 hours of treatment, and repeat phototherapy for rebound jaundice (duration of phototherapy: 3 trials, 232 infants; WMD +2.00 hours, 95% CI –3.50 hours to +7.52 hours; use of additional phototherapy: 5 trials; 3/148 [2.0%] with fibreoptic v 3/156 [1.9%] with conventional; RR 1.07, 95% CI 0.27 to 4.27; percentage change in serum bilirubin after 24 hours of treatment: 1 trial, 20 infants; WMD +1.7%, 95% CI –2.65% to +6.05%; repeat phototherapy for rebound jaundice: 3 trials; 10/122 [8%] with fibreoptic v 5/121 [4%] with conventional; RR 2.00, 95% CI 0.71 to 5.63).
The first subsequent RCT (109 term infants, birth weight 2500 g or more, infants with haemolytic jaundice excluded) found that conventional daylight phototherapy significantly increased the rate of decline of serum bilirubin, and decreased treatment duration compared with fibreoptic phototherapy (bilirubin decline rate: 2.6 ± 1.0 micromol/L/hour with conventional v 1.7 ± 0.9 micromol/L/hour with fibreoptic; P <0.05; duration of phototherapy: 49.4 ± 14.4 hours with conventional v 61 ± 13.1 hours with fibreoptic; P <0.05).
The second subsequent RCT (140 preterm infants, gestation <31 weeks, infants with haemolytic jaundice excluded) compared single conventional phototherapy, fibreoptic Wallaby phototherapy, fibreoptic BiliBlanket phototherapy, and combined conventional plus fibreoptic Wallaby phototherapy. We only report data on the conventional, fibreoptic Wallaby and fibreoptic BiliBlanket groups here. The RCT found no significant difference in the duration of treatment required for either Wallaby and BiliBlanket fibreoptic phototherapy compared with conventional phototherapy alone (92 hours with Wallaby v 95 hours with BiliBlanket v 90 hours with conventional: reported as not significant; P value not reported).
Double versus single phototherapy:
We found one systematic review (search date 2000; term and preterm infants; randomised and quasi-randomised trials; see comment below) and three subsequent RCTs.
The systematic review included one RCT (86 term infants, haemolysis excluded) comparing double fibreoptic phototherapy (infants wrapped in 2 BiliBlankets) versus single conventional phototherapy. The RCT included in the review found no significant difference between groups in duration of treatment, percentage change in serum bilirubin per hour, percentage change in serum bilirubin per day, and the use of repeat phototherapy for rebound jaundice (duration of treatment: WMD +2.24 hours, 95% CI –10.68 hours to +15.16 hours; percentage change in serum bilirubin per hour: WMD –0.04%, 95% CI –0.17% to +0.09%; percentage change in SBR per day: WMD +2.82%, 95% CI –1.84% to +7.48%; and the use of repeat phototherapy for rebound jaundice: RR 1.05, 95% CI 0.07 to 16.22). The review also compared double phototherapy using a combination of fibreoptic plus conventional phototherapy versus conventional phototherapy alone. It found no significant difference between fibreoptic plus conventional phototherapy and single conventional phototherapy in exchange transfusion, additional phototherapy, and percentage change in serum bilirubin after 24 or 48 hours, although it noted a trend favouring the fibreoptic plus conventional group (exchange transfusion: 1 trial; 0/19 [0%] with fibreoptic plus conventional v 2/23 [8%] with conventional alone; RR 0.24, 95% CI 0.01 to 4.72; additional phototherapy: 1 trial; 0/90 [0%] with fibreoptic plus conventional v 4/90 [4%] with conventional; RR 0.11, 95% CI 0.01 to 2.02; percentage change in SBR after 24 hours: 1 trial, 26 infants; WMD –3.2%, 95% CI –17.2% to +10.8%; percentage change in SBR after 48 hours: WMD –9.2%, 95% CI –25.02% to +6.62%). It found no significant difference between fibreoptic plus conventional phototherapy and single conventional phototherapy in repeat phototherapy for rebound jaundice (6 trials; 36/232 [16%] with fibreoptic plus conventional v 30/240 [13%] with conventional; RR 1.29, 95% CI 0.85 to 1.95).
The first subsequent RCT (51 term infants, birth weight 2500 g or more, haemolysis included) compared double conventional phototherapy using daylight fluorescent lamps versus single conventional phototherapy. It found that double conventional phototherapy reduced serum bilirubin at a significantly higher rate during the first 24 hours compared with single conventional phototherapy (3.8 ± 2.1 micromol/L/hour with double v 2.4 ± 1.7 micromol/L/hour with single; P = 0.02). It found a trend for double conventional phototherapy to reduce bilirubin at a higher rate on the second day, but this did not reach significance (P = 0.06). It found that double conventional phototherapy significantly reduced duration of treatment compared with single conventional phototherapy (34.9 ± 12.6 hours with double v 43.7 ± 17.5 hours with single; P = 0.039). It did not report on kernicterus or other long-term outcomes.
The second subsequent RCT (140 preterm infants, gestation <31 weeks, infants with haemolytic jaundice excluded) compared single conventional phototherapy, fibreoptic Wallaby phototherapy, fibreoptic BiliBlanket phototherapy, and combined conventional plus fibreoptic Wallaby phototherapy. It found that the combined phototherapy reduced mean duration of treatment required compared with either of the treatments used alone (Wallaby: 92 hours; BiliBlanket: 95 hours; conventional: 90 hours; combined Wallaby and conventional: 75 hours; P <0.05 for combined Wallaby and conventional v either Wallaby or BiliBlanket alone; P <0.01 for combined Wallaby and conventional v conventional alone). It also found that the combination of conventional phototherapy plus Wallaby fibreoptic phototherapy produced a smaller increase in bilirubin levels over the first 24 hours compared with conventional phototherapy alone (16% with conventional plus Wallaby fibreoptic v 27% with conventional alone; P <0.01).
The third subsequent RCT (60 term infants 37–42 weeks, birth weight 2500 g or more, exclusively breastfed, 1- and 5-minute Apgar scores >6, total SBR 13.0–19.9 mg/dL, with non-haemolytic hyperbilirubinaemia) compared double surface phototherapy (4 deep blue and 2 daylight fluorescent lamps at least 30 cm above the baby plus 4 deep blue fluorescent lamps 25 cm below the baby) with single surface phototherapy (4 deep blue and 2 daylight fluorescent lamps at least 30 cm above the baby). The RCT found no significant difference in the mean serum bilirubin levels between double surface phototherapy compared with single surface phototherapy at 24 hours (10.3 ± 1.9 mg/dL with double surface phototherapy v 11.3 ± 2.1 mg/dL with single surface phototherapy; P = 0.05). However, the RCT found that compared with single surface phototherapy, double surface phototherapy significantly increased levels of decline in serum bilirubin after 24 hours (5.4 ± 2.0 mg/dL with double surface v 3.5 ± 1.7 mg/dL with single surface; P <0.001). The RCT found no significant difference between groups in the total declined serum bilirubin levels between 24 and 48 hours (3.1 ± 1.7 mg/dL with double surface v 3.0 ± 1.8 mg/dL with single surface; P = 0.9). The RCT also found that compared with single surface phototherapy, double surface phototherapy significantly increased total decline in serum bilirubin levels after 48 hours (8.4 ± 1.7 mg/dL with double surface v 6.5 ± 2.3 mg/dL with single surface; P = 0.001). No exchange transfusions were performed in either group.
Triple versus double phototherapy:
We found one RCT (40 infants >37 weeks' gestation, >2500 g, with no medical problems and non-haemolytic jaundice) comparing triple phototherapy (2 single fluorescent lamps 25 cm above bed plus third fluorescent lamp 35 cm from bed) with double phototherapy (2 single fluorescent lamps 25 cm above bed). The RCT found no significant difference between triple and double phototherapy in the length of hospital stay (41.5 ± 17.7 hours with triple v 34.6 ± 16.5 hours with double; P = 0.211). The RCT also found no significant difference in mean bilirubin levels between triple and double phototherapy at 8, 16, or 24 hours (8 hours: 14 ± 1.7 mg/dL with triple v 13.7 ± 1.8 mg/dL with double; P = 0.59; 16 hours: 12.4 ± 1.7 mg/dL with triple v 12.2 ± 1.7 mg/dL with double; P = 0.76; 24 hours: 10.9 ± 1.0 mg/dL with triple v 10.3 ± 2.0 mg/dL with double; P = 0.37).
Intermittent versus continuous phototherapy:
We found three RCTs comparing intermittent versus continuous phototherapy.
The first RCT (114 infants >2000 g, absence of concomitant disease, hyperbilirubinaemia not exceeding the range for exchange transfusion or requiring high intensity phototherapy) compared intermittent phototherapy (phototherapy on for 1 hour then off for 1 hour) versus continuous phototherapy (2 hours on, 30 minutes off). The RCT found no significant difference between intermittent compared with continuous phototherapy in mean serum bilirubin level at 12, 24, 36, and 48 hours (12 hours: 13.57 ± 2.30 mg/dL with intermittent v 13.73 ± 1.89 mg/dL with continuous; P = 0.6; 24 hours: 10.86 ± 2.13 mg/dL with intermittent v 11.06 ± 2.06 mg/dL with continuous; P = 0.6; 36 hours: 9.02 ± 1.94 mg/dL with intermittent v 9.17 ± 1.83 mg/dL with continuous; P = 0.7; 48 hours: 9.30 ± 1.43 mg/dL with intermittent v 8.93 ± 1.26 mg/dL with continuous; P = 0.7).
The second three-armed RCT (120 preterm infants, birth weight 1250–2000 g, Coombs' negative, no haemolytic anaemia, no gross congenital anomalies, no severe respiratory distress syndrome) compared continuous phototherapy for 5 days versus intermittent phototherapy (12 hours on, 12 hours off) for 5 days or no treatment. We only report the data on continuous and intermittent groups here. The RCT found no difference between groups in the proportion of preterm infants who had a serum bilirubin level >12 mg/dL (3/40 [8%] with intermittent phototherapy v 2/40 [5%] with continuous phototherapy; P value not reported). However, the RCT found that compared with intermittent phototherapy, a smaller proportion of preterm infants treated with continuous phototherapy had serum bilirubin levels >15 mg/dL (1/40 [3%] with intermittent phototherapy v 0/40 [0%] with continuous phototherapy; P value not reported).
The third RCT (34 term infants >2500 g, physiological jaundice) compared intermittent equal duration therapy (4 hours on, 4 hours off) versus intermittent short duration phototherapy (1 hour on, 3 hours off) versus continuous phototherapy. The RCT found no significant difference in the duration of phototherapy required or total hours of irradiation (duration of phototherapy required: 86.7 ± 28.9 hours with intermittent equal v 100.0 ± 61 hours with intermittent short v 89.9 ± 54.2 hours with continuous; P >0.05; total hours of irradiation: 43.4 ± 14.5 hours with intermittent equal v 25.0 ± 15.3 hours with intermittent short v 89.9 ± 54.2 hours with continuous; P <0.05). The RCT also found no significant difference between groups in the rate of increase of serum bilirubin levels or in the rate of decrease in serum bilirubin levels (rate of increase of serum bilirubin levels: 1.25 ± 0.66 micromol/L/hour with intermittent equal v 0.89 ± 0.65 micromol/L/hour with intermittent short v 0.82 ± 0.43 micromol/L/hour with continuous; P >0.05; rate of decrease in serum bilirubin levels: 1.49 ± 0.87 micromol/L/hour with intermittent equal v 1.09 ± 0.56 micromol/L/hour with intermittent short v 1.08 ± 4.10 micromol/L/hour with continuous; P >0.05).
Close versus distant light-source phototherapy:
We found one RCT (774 infants, hyperbilirubinaemia not severe enough to require exchange transfusion, absence of history of traditional herbal treatment, absence of treatment with phenobarbitol, absence of septicaemia, absence of hepatomegaly regardless of cause, absence of suspected congenital metabolic disorders) comparing conventional phototherapy given from a distance of 20 cm (close light-source) above the neonate versus 40 cm (distant light-source) above the neonate. The RCT found that close light-source phototherapy significantly reduced the mean duration of treatment compared with distant light-source phototherapy (66 ± 22 hours with close light-source v 81.6 ± 24.6 hours with distant light-source; P <0.001).
Increased skin exposure versus standard skin exposure phototherapy:
We found one RCT (59 preterm infants at least 36 weeks' gestation, >1500 g birth weight, non-haemolytic hyperbilirubinaemia with total serum bilirubin in range for phototherapy, absence of congenital anomaly, absence of need for respiratory support, absence of co-existing pathology) comparing conventional phototherapy in partially clothed infants (disposable nappy only) versus naked infants. The RCT found no significant difference between groups in the number of infants still requiring phototherapy between 24 and 48 hours (13/30 [43%] with partial clothing v 13/29 [45%] with naked infants; P = 0.9). The RCT also found no significant difference between groups in the number of infants still requiring phototherapy between 48 and 72 hours (2/30 [7%] with partial clothing v 4/29 [14%] with naked infants; P = 0.4). The RCT found no significant difference between groups in the mean percentage decline in serum bilirubin levels (15.4% ± 18% with partial clothing v 19.0% ± 15% with naked infants; P = 0.4). There was also no significant difference between groups in the absolute change in mean serum bilirubin levels (37.6 ± 41 micromol/L with partial clothing v 46.4 ± 37 micromol/L with naked infants; P = 0.4).
Fluorescent lamps versus halide lamps:
We found one three-armed RCT (101 infants, at least 40 hours of age, non-haemolytic hyperbilirubinaemia, not on antibiotics) comparing treatment with 6 standard fluorescent lamps versus fluorescent lamps plus white reflecting curtains versus a halide lamp. We only report the data from the fluorescent compared with halide lamps here. The RCT found no significant difference between groups in reduction in bilirubin levels per 24 hours (3.3 ± 1.5 mg/dL with fluorescent lamps v 3.2 ± 1.3 mg/dL with halide lamps; P >0.05).
Fluorescent versus blue fluorescent lamps:
We found one three-armed RCT (72 infants, hyperbilirubinaemia in first 72 hours of life, SBR >8.6 mg/dL, no sepsis, no respiratory distress, non-haemolytic) comparing daylight fluorescent lamps versus standard blue fluorescent lamps versus blue fluorescent lamps with a narrow spectral emission. Phototherapy was discontinued when serum bilirubin concentration had declined steadily for at least 12 hours, and had reached a level of at most 8 mg per 100 mL. The RCT found that a smaller proportion of infants discontinued phototherapy after 1 day with daylight fluorescent lamps compared with blue fluorescent lamps (with or without narrow spectral emission) (2/24 [8%] with daylight v 6/24 [25%] with standard blue v 12/24 [50%] with narrow spectrum blue; P values not reported). However, the RCT found similar rates in the number of infants discontinuing phototherapy on days 2 and 3 (day 2: 6/24 [25%] with daylight v 14/24 [58%] with standard blue v 8/24 [33%] with narrow spectrum blue; P values not reported; day 3: 7/24 [29%] with daylight v 4/24 [17%] with standard blue v 4/24 [17%] with narrow spectrum blue; P values not reported). However, the RCT found that compared with daylight fluorescent lamps, blue fluorescent lamps (with or without narrow spectral emission) increased mean decreases in serum bilirubin levels after the first 24 hours, the second 24 hours, and the third 24 hours of phototherapy (mean decrease in the first 24 hours: 0.96 mg/dL with daylight v 2.17 mg/dL with standard blue v 3.52 mg/dL with narrow spectrum blue; P values not reported; mean decrease in the second 24 hours: 0.38 mg/dL with daylight v 1.38 mg/dL with standard blue v 2.32 mg/dL with narrow spectrum blue; P values not reported; mean decrease in the third 24 hours: 1.46 mg/dL with daylight v 1.72 mg/dL with standard blue v 1.82 mg/dL with narrow spectrum blue; P values not reported).
Blue fluorescent versus green fluorescent lamps:
We found two RCTs comparing blue fluorescent versus green fluorescent light.
The first RCT (262 infants, non-haemolytic jaundice) compared treatment with blue fluorescent lamps versus green fluorescent lamps. The RCT also reported a planned subgroup analysis on term (at least 37 weeks' gestation) versus preterm (<37 weeks' gestation) infants. The RCT found no significant difference in the duration of phototherapy in the term infants or preterm infants with blue fluorescent light compared with green fluorescent light (term infants: 49.88 ± 3.02 hours with blue light v 42.68 ± 2.74 hours with green light; P >0.05; preterm infants: 53.29 ± 5.90 hours with blue light v 53.26 ± 5.52 hours with green light; P >0.05). The RCT also found no significant difference in the rate of fall of serum bilirubin in term infants or preterm infants with blue light compared with green light (term infants: 2.86 ± 0.17 micromol/hour with blue light v 3.27 ± 0.22 micromol/hour with green light; P >0.05; preterm infants: 2.50 ± 0.39 micromol/hour with blue light v 2.91 ± 0.38 micromol/hour with green light; P >0.05).
The second RCT (84 low birth weight infants, non-haemolytic jaundice, no respiratory distress, no sepsis, no post-phototherapy rebound) compared treatment with blue fluorescent lamps versus green fluorescent lamps. The RCT reported a planned subgroup analysis, in which three groups were compared based on initial serum bilirubin levels (group 1: 21–16.1 mg/dL; group 2: 16–12.1 mg/dL; group 3: 12–9.0 mg/dL). The RCT found no significant difference in the percentage decrease in serum bilirubin levels for any of the subgroups with blue light compared with green light at 24 and 48 hours (group 1: 24 hours: 31.7% decrease with blue light v 31% decrease with green light; 48 hours: 36% decrease with blue light v 46% decrease with green light; group 2: 24 hours: 22% decrease with blue light v 20.3% decrease with green light: 48 hours: 27% decrease with blue light v 22% decrease with green light: group 3: 24 hours: 20% decrease with blue light v 19% decrease with green light: 48 hours: 16% decrease with blue light v 10.5% decrease with green light; P >0.5 for all comparisons).
Blue-green fluorescent versus blue fluorescent lamps:
We found 4 RCTs comparing blue-green fluorescent light versus blue fluorescent light.
The first RCT (85 infants, preterm with a gestational age 196–258 days, postnatal age >24 hours, non-haemolytic hyperbilirubinaemia) compared treatment with 6 turquoise (blue-green) fluorescent lamps plus two daylight fluorescent lamps versus 6 blue fluorescent lamps plus two daylight fluorescent lamps. The RCT found no significant difference in total serum bilirubin level decrease after 48 hours of treatment with turquoise light compared with blue light (P = 0.36; absolute data presented graphically).
The second RCT (141 infants, preterm with a gestational age 196–258 days, postnatal age >24 hours, non-haemolytic hyperbilirubinaemia and no previous phototherapy) compared phototherapy with 8 turquoise fluorescent lamps at an average distance of 41 cm versus phototherapy with 8 blue fluorescent lamps at an average distance of 32 cm. The RCT found that compared with blue fluorescent light, turquoise fluorescent light significantly increased the mean decrease in serum bilirubin levels after 24 hours of treatment (mean decrease: 92 ± 31 micromol/L with turquoise light v 78 ± 34 micromol/L with blue light; mean difference 15 micromol/L; P = 0.008).
The third three-armed RCT (114 jaundiced, but otherwise healthy, term infants) compared 6 focused arrays of blue LED phototherapy, 6 focused arrays of blue-green LED phototherapy, and conventional phototherapy consisting of three halogen-quartz bulbs. We only report the data for the blue LED phototherapy compared with blue-green LED phototherapy arms here (47 infants). The RCT found no significant difference in the mean number of hours spent under phototherapy with blue-green LED phototherapy compared with blue LED phototherapy (39.2 ± 25.5 hours with blue-green LED v 31.6 ± 19.6 hours with blue LED; P value reported as not significant). The RCT also found no significant difference between groups in the mean rate of serum bilirubin level decline (–1.55 ± 3.54 micromol/L with blue-green LED v –2.82 ± 2.44 micromol/L with blue LED; P value reported as not significant).
The fourth RCT (40 infants, low birth weight, hyperbilirubinaemia in first 4 days of life, healthy) compared treatment with blue-green fluorescent lights versus treatment with blue fluorescent lights. The RCT found that compared with blue fluorescent light, blue-green fluorescent light significantly reduced the number of infants still requiring phototherapy after 24 hours of treatment (1/20 [5%] with blue-green light v 10/20 [50%] with blue light; P <0.0001). The RCT also found that blue-green fluorescent light significantly increased the mean percentage decrement in serum bilirubin levels after 24 hours of treatment compared with blue fluorescent light (46.4% with blue-green light v 22.6% with blue light; P <0.0001).
Blue LED versus conventional quartz-halogen:
We found three RCTs comparing blue LED with conventional phototherapy.
The first RCT (69 infants, jaundiced but otherwise healthy, gestational age >37 weeks) compared phototherapy with 6 focused blue gallium nitride LEDs versus conventional phototherapy with three halogen-quartz bulbs. The RCT found no significant difference in the number of hours spent under phototherapy with blue LED compared with conventional phototherapy (31 ± 17 hours with blue LED v 32 ± 17 hours with conventional phototherapy; P = 0.93). The RCT also found no significant difference in the rate of serum bilirubin level decline with blue LED compared with conventional phototherapy (–2.87 ± 2.44 micromol/L/hour with blue LED v –2.07 ± 3.03 micromol/L/hour with conventional phototherapy; P = 0.94).
The second three-armed RCT (114 jaundiced, but otherwise healthy, term infants) compared 6 focused arrays of blue LED phototherapy, 6 focused arrays of blue-green LED phototherapy, and conventional phototherapy consisting of three halogen-quartz bulbs. We only report the data for the blue LED phototherapy versus conventional phototherapy comparison here (82 infants). The RCT found no significant difference between groups in the mean number of hours spent under phototherapy (31.6 ± 19.6 hours with blue LED v 35.4 ± 20.2 hours with conventional phototherapy; P value reported as not significant), or in the mean rate of serum bilirubin level decline (–2.82 ± 2.44 micromol/hour with blue LED v –2.42 ± 3.03 micromol/hour with conventional phototherapy; P value reported as not significant).
The third RCT (88 preterm infants, birth weight >1000 g, non-haemolytic jaundice, no ecchymosis, no malformations, no congenital infection) compared 5 blue LEDs versus single quartz-halogen phototherapy. The RCT found that compared with conventional phototherapy, blue LED phototherapy significantly decreased the mean time spent under phototherapy (36.8 ± 21 hours with blue LED v 63.8 ± 37 hours with conventional phototherapy; P <0.01). The RCT also found that blue LED significantly reduced mean serum bilirubin levels after 8, 16, and 24 hours of treatment compared with conventional phototherapy (8 hours: 9.3 ± 2.5 mg% with blue LED v 10.5 ± 2.1 mg% with conventional; P <0.05; 16 hours: 8.1 ± 2.7 mg% with blue LED v 9.4 ± 1.8 mg% with conventional; P <0.01; 24 hours: 7.2 ± 2.5 mg% with blue LED v 9.6 ± 2.4 mg% with conventional; P <0.01).
Blue LED versus blue fluorescent lamps:
We found one RCT (66 healthy infants, at least 35 weeks' gestation) that compared phototherapy using blue LED (overhead neoBLUE LED plus either BiliBlanket or Wallaby system underneath) versus phototherapy with blue fluorescent light (8 overhead blue fluorescent lights plus either BiliBlanket or Wallaby system underneath). The RCT found no significant difference in the mean rate of serum bilirubin level decline with blue LED compared with blue fluorescent phototherapy (0.35 ± 0.25 mg/dL/hour with blue LED v 0.27 ± 0.25 mg/dL/hour with blue fluorescent phototherapy; P = 0.20).
Blue-green LED versus conventional quartz-halogen:
We found one three-armed RCT (114 jaundiced, but otherwise healthy, term infants) comparing 6 focused arrays of blue LED phototherapy, 6 focused arrays of blue-green LED phototherapy, and conventional phototherapy consisting of three halogen-quartz bulbs. We only report the data for the blue-green LED phototherapy versus conventional phototherapy comparison here (79 infants). The RCT found no significant difference in the mean number of hours spent under phototherapy (39.2 ± 25.5 hours with blue-green light v 35.4 ± 20.2 hours with conventional phototherapy; P value reported as not significant) or in the mean rate of serum bilirubin level decline (–1.55 ± 3.54 micromol/hour with blue-green light v –2.42 ± 3.03 micromol/hour with conventional phototherapy; P value reported as not significant) with blue-green LED phototherapy compared with conventional phototherapy.
Prophylactic phototherapy versus threshold phototherapy:
We found one RCT (95 infants, birth weight <1500 g, within 12 hours of birth, without isoimmunisation or major life-threatening anomaly) that compared prophylactic phototherapy (commencement of phototherapy within 12 hours of birth) versus threshold phototherapy (commencement once SBR >150 micromol/L). The RCT found no significant difference between groups in the combined outcome of death and cerebral palsy at 18 months (5/40 [13%] with prophylactic phototherapy v 10/43 [23%] with threshold phototherapy; OR 1.86, 95% CI 0.58 to 5.92; P = 0.4). The RCT also found no significance between groups in the incidence of cerebral palsy or in the incidence of abnormal developmental index score at 18 months (cerebral palsy: 2/37 [5%] with prophylactic phototherapy v 5/38 [12%] with threshold phototherapy; OR 2.43, 95% CI 0.44 to 13.34; P = 0.44; abnormal developmental index score <84: 9/37 [24%] with prophylactic phototherapy v 8/38 [21%] with threshold phototherapy; OR 0.87, 95% CI 0.30 to 2.48; P = 0.78). The RCT found no significant difference in the mean number of hours of phototherapy or in the mean total of days in neonatal intensive care unit with prophylactic compared with threshold phototherapy (mean number of hours of phototherapy: 85 hours with prophylactic phototherapy v 68.5 hours with threshold phototherapy; P >0.05; mean total of days in neonatal intensive care unit: 82.3 ± 37.9 days with prophylactic phototherapy v 82.7 ± 38.9 days with threshold phototherapy; P >0.05).
The RCT found no significant difference in the peak unconjugated serum bilirubin levels between groups (170 ± 26.0 micromol/L with prophylactic phototherapy v 183.5 ± 28.0 micromol/L with threshold phototherapy; P >0.05); however, a subgroup analysis of infants with a birth weight <1000 g found that compared with threshold phototherapy, prophylactic phototherapy significantly reduced peak unconjugated serum bilirubin levels (139.2 ± 46.0 micromol/L with prophylactic phototherapy v 171.2 ± 26.0 micromol/L with threshold phototherapy; P <0.02). A further subgroup analysis of infants with a birth weight of 1000 g to 1499 g found no significant difference between groups in peak unconjugated serum bilirubin level (190.6 ± 40.8 micromol/L with prophylactic phototherapy v 191.9 ± 26.5 micromol/L with threshold phototherapy; P >0.05). The RCT found that compared with threshold phototherapy, prophylactic phototherapy significantly reduced the number of infants whose peak serum bilirubin level was reached before 48 hours of age (1/45 [2%] with prophylactic phototherapy v 14/47 [30%] with threshold phototherapy; P <0.001).
Low threshold versus high threshold phototherapy:
We found two RCTs comparing low threshold versus high threshold phototherapy.
The first RCT (1974 infants, extremely low birth weight, 12–36 hours old, absence of terminal condition, absence of previous phototherapy, absence of major congenital anomaly, non-severe haemolytic disease, absence of congenital non-bacterial infection) compared low threshold phototherapy (initiation of phototherapy at enrolment, SBR expected to be 85 micromol/L and recommencement of phototherapy if SBR >85 micromol/L in first 7 days of life or SBR >137 micromol/L from day 7–14 of life) versus high threshold phototherapy (initiation of phototherapy at 137 micromol/L and recommencement if SBR >137 micromol/L in first 7 days and SBR >171 micromol/L from day 7–14 of life).
The RCT found no significant difference between groups in mortality before day 15 or mortality before discharge (mortality before day 15: 96/990 [9.7%] with low threshold v 95/984 [9.7%] with high threshold; RR 1.00, 95% CI 0.88 to 1.30; mortality before discharge: 209/990 [21%] with low threshold v 201/984 [20%] with high threshold; RR 1.03, 95% CI 0.88 to 1.21). The RCT also found no significant difference between groups in mortality or mortality and neurodevelopmental impairment at 18 to 22 months (mortality: 230/946 [24%] with low threshold v 218/944 [23%] with high threshold; RR 1.05, 95% CI 0.90 to 1.22; mortality and neurodevelopmental impairment: 465/902 [52%] with low threshold v 493/902 [55%] with high threshold; RR 0.94, 95% CI 0.87 to 1.02).
The RCT found that compared with high threshold phototherapy, low threshold phototherapy significantly reduced the risk of neurodevelopmental impairment and profound impairment at 18 to 22 months (neurodevelopmental impairment: 235/902 [26%] with low threshold v 275/902 [30%] with high threshold; RR 0.86, 95% CI 0.74 to 0.99; MDI score for profound impairment: 22 infants with a score of 50, 121 infants with a score of <50) (80/895 [9%] with low threshold v 119/896 [13%] with high threshold; RR 0.68, 95% CI 0.52 to 0.89). However, the RCT found no significant difference between groups in cerebral palsy at 18 to 22 months (mild/moderate or severe: 81/929 [9%] with low threshold v 91/924 [10%] with high threshold; RR 0.89, 95% CI 0.67 to 1.18).
The RCT found that low threshold phototherapy significantly reduced the risk of severe hearing loss compared with high threshold phototherapy at 18 to 22 months (9/925 [1%] with low threshold v 28/922 [3%] with high threshold; RR 0.32, 95% CI 0.15 to 0.98). It found no significant difference in blindness compared with high threshold phototherapy at 18 to 22 months (2/928 [0.2%] with low threshold v 7/924 [0.8%] with high threshold; RR 0.28, 95% CI 0.06 to 1.37).
The RCT found that low threshold phototherapy significantly increased the duration of phototherapy compared with high threshold phototherapy (88 ± 48 hours with low threshold v 35 ± 31 hours with high threshold; P <0.001). However, it found no significant difference between groups in the number of exchange transfusions or length of hospital stay (exchange transfusions: 2 with low threshold v 3 with high threshold; P = 0.69; length of hospital stay: 97 ± 43 days with low threshold v 100 ± 47 days with high threshold; P = 0.11). The RCT found that low threshold phototherapy significantly decreased the serum bilirubin level at day 5 compared with high threshold phototherapy (0.33 ± 0.25 mg/dL with low threshold v 0.48 ± 0.33 mg/dL with high threshold; P <0.001).
The second RCT (78 infants, birth weight <2500 g) compared starting phototherapy once serum bilirubin levels were >13 mg/dL versus starting phototherapy once serum bilirubin levels were 10 mg/dL or more and treating for 12 hours versus starting phototherapy once serum bilirubin levels were 10 mg/dL or more and treating for 24 hours.
The RCT found no difference in the number of infants requiring phototherapy at <72 hours of age (14/26 [54%] with higher threshold v 17/29 [59%] with lower threshold and 12 hours of phototherapy v 9/23 [39%] with lower threshold and 24 hours of phototherapy; P values not reported).
The RCT also found no difference in the number of infants who received phototherapy in the first 72 hours of life and then required additional phototherapy (8/14 [57%] with higher threshold v 6/17 [35%] with lower threshold and 12 hours of phototherapy v 2/9 [22%] with lower threshold and 24 hours of phototherapy; P values not reported).
The RCT found no difference in the number of infants requiring phototherapy at 72 hours of age or more (12/26 [46%] with higher threshold v 12/29 [41%] with lower threshold and 12 hours of phototherapy v 14/23 [61%] with lower threshold and 24 hours of phototherapy; P values not reported).
The RCT found no difference in the number of infants who received phototherapy after 72 hours of age and required additional phototherapy (3/12 [25%] with higher threshold v 1/12 [8%] with lower threshold and 12 hours of phototherapy v 2/14 [8%] with lower threshold and 24 hours of phototherapy; P values not reported).
Compact fluorescent light phototherapy versus standard length tube light phototherapy:
We found one RCT (100 infants >34 weeks' gestation, haemolytic jaundice excluded), which found no significant difference between compact fluorescent light (CFL) and standard length tube light (STL) phototherapy in total duration of treatment required (40.66 hours with CFL v 40.78 hours with STL; P = 0.98). It found that CFL significantly reduced the mean total serum bilirubin over 12 hours compared with STL (15.86 mg/dL, 95% CI 15.5 mg/dL to 16.2 mg/dL with CFL v 14.23 mg/dL, 95% CI 13.7 mg/dL to 14.8 mg/dL with STL; P <0.001).
Phototherapy versus immunoglobulin:
See benefits of immunoglobulin.
Harms
Conventional phototherapy versus no treatment:
The review found no evidence that phototherapy for neonatal hyperbilirubinaemia has any long-term adverse neurodevelopmental effects.
Continuous phototherapy versus no treatment:
The RCT found that a significantly higher proportion of preterm infants in the control group regained or surpassed their birth weight compared with infants treated with continuous phototherapy at postnatal day 7 (44% with continuous phototherapy v 80% with no treatment; reported as significant; P value not reported). However, the RCT found that infants treated with continuous phototherapy gained significantly more weight in the second and third postnatal weeks compared with infants in the control group (second postnatal week: 184.5 ± 55.5 g with continuous phototherapy v 139.0 ± 69.8 g with no treatment; P <0.05; third postnatal week: 225.3 ± 62.5 g with continuous phototherapy v 162.4 ± 62.3 g with no treatment; P <0.05).
Intermittent phototherapy versus no treatment:
The RCT found that a significantly higher proportion of preterm infants in the control group regained or surpassed their birth weight compared with infants treated with intermittent phototherapy at postnatal day 7 (57.6% with intermittent phototherapy v 80% with no treatment; reported as significant; P value not reported). However, the RCT found that infants treated with intermittent phototherapy gained significantly more weight in the second and third postnatal weeks compared with infants in the control group (second postnatal week: 193.6 ± 82.3 g with intermittent phototherapy v 139.0 ± 69.8 g with no treatment; P <0.05; third postnatal week: 195.5 ± 74.2 g with intermittent phototherapy v 162.4 ± 62.3 g with no treatment; P <0.05).
Fibreoptic phototherapy versus no treatment:
The review gave no information on adverse effects for this comparison.
Conventional versus fibreoptic phototherapy:
In the systematic review, one small trial (20 infants) found that transepidermal water loss (sweating) was significantly higher in infants treated with fibreoptic devices compared with conventional phototherapy, and one small trial (34 infants) found no significant difference between fibreoptic and conventional phototherapy in mothers developing migraine during their infants' treatment with phototherapy. However, the clinical significance of this is uncertain. One RCT reported transient erythema and mild watery stools not leading to dehydration (erythema: 1/50 [2%] with conventional v 1/50 [2%] with fibreoptic; mild watery stools: 3/50 [6%] with conventional v 3/50 [6%] with fibreoptic). A subsequent RCT found similar levels of transient erythema between conventional phototherapy, fibreoptic Wallaby phototherapy, and fibreoptic BiliBlanket phototherapy (10/35 [29%] with conventional v 9/35 [26%] with Wallaby v 8/35 [23%] with BiliBlanket; significance not assessed).
Double versus single phototherapy:
One RCT found no significant difference between double conventional and single conventional phototherapy in weight reduction, frequency of stooling, or fever. Another RCT found a small increase in rates of transient erythema using the combination of Wallaby and conventional phototherapy compared with one type of phototherapy (12/35 [34%] with combined v 10/35 [29%] with conventional v 9/35 [26%] with Wallaby v 8/35 [23%] with BiliBlanket; significance not assessed). The third subsequent RCT found that double surface phototherapy significantly increased body temperature compared with single surface phototherapy after 24 hours of treatment (37.1 ± 0.2 °C with double surface phototherapy v 36.9 ± 0.3 °C with single surface phototherapy; P = 0.003). The RCT also found that double surface phototherapy significantly lowered the number of stools per day compared with single surface phototherapy (4.3 ± 3.0 stools/day with double surface phototherapy v 7.2 ± 3.4 stools/day with single surface phototherapy; P = 0.001). The RCT found no significant difference between groups in body weight at 24 and 48 hours after phototherapy commenced, percent body weight change at 24 and 48 hours after phototherapy commenced, or temperature after 48 hours of phototherapy (body weight at 24 hours: 3021.7 g with double surface phototherapy v 2971.7 g with single surface phototherapy; P = 0.52; body weight at 48 hours: 3043.3 g with double surface phototherapy v 3010.7 g with single surface phototherapy; P = 0.69; percent body weight change at 24 hours: 1% with double surface phototherapy v 1.5% with single surface phototherapy; P = 0.46; percent body weight change at 48 hours: 1.7% with double surface phototherapy v 2.3% with single surface phototherapy; P = 0.44; temperature after 48 hours; 36.9 ± 0.2 °C with double surface phototherapy v 36.9 ± 0.2 °C with single surface phototherapy; P = 0.13).
Triple versus double phototherapy:
The RCT gave no information on adverse effects.
Intermittent versus continuous phototherapy:
The RCTs gave no information on adverse effects.
Close versus distant light-source phototherapy:
The RCT found no significant difference in adverse effects such as troublesome skin rashes, burns, clinical dehydration, or lethargy between close and distant phototherapy (no further data reported).
Increased skin exposure versus standard skin exposure phototherapy:
The RCT found no significant difference between phototherapy groups treated partially clothed (disposable nappy only) or naked in the incidence of the following Parenting Stress Index scores: 'parental distress', 'parent–child dysfunction' (1/25 [4%] with partial clothing v 3/25 [12%] with naked infants; P = 0.3), 'parent–child dysfunction', 'difficult child', or 'total stress score' ('parental distress': 6/25 [24%] with partial clothing v 9/25 [36%] with naked infants; P = 0.4; 'difficult child': 4/25 [16%] with partial clothing v 3/25 [12%] with naked infants; P = 0.7; 'total stress score': 7/25 [28%] with partial clothing v 7/25 [28%] with naked infants; P = 1).The RCT also reported no significant difference between groups in the incidence of rebound jaundice requiring phototherapy (7/30 [23%] with partial clothing v 9/29 [31%] with naked infants; P = 0.5). The RCT reported that there were no episodes of patent ductus arteriosus, skin rashes, or dehydration in phototherapy groups treated either partially clothed (disposable nappy only) or naked.
Fluorescent lamps versus halide lamps:
The RCT gave no information on adverse effects.
Fluorescent versus blue fluorescent lamps:
The RCT gave no information on adverse effects.
Blue fluorescent versus green fluorescent lamps:
The RCTs gave no information on adverse effects.
Blue-green fluorescent versus blue fluorescent lamps:
The first two RCTs
reported that there were no adverse effects apart from "loose green stools". However, it is unclear as to the number of babies who had this adverse effect and in what treatment allocation group they were. The third RCT reported that no adverse effects were noted. The fourth RCT gave no information about adverse effects.
Blue LED versus conventional quartz-halogen:
The first and second RCTs reported that there were no adverse effects in either group.
The third RCT reported that no one in either group required exchange transfusion, developed rashes, or had temperature instability. It also found no significant difference between groups in weight loss or in the incidence of rebound jaundice (weight loss: 1.89% of weight loss against initial weight with blue LED and 1.99% of weight loss against initial weight with conventional phototherapy; P = 0.33; rebound jaundice: 26.8% with blue LED v 18.2% with conventional phototherapy; P = 0.43).
Blue LED versus blue fluorescent lamps:
The RCT gave no information on adverse effects.
Blue-green LED versus conventional quartz-halogen:
The RCT reported that no adverse effects were found.
Prophylactic phototherapy versus threshold phototherapy:
The RCT found no significant difference between groups in the incidence of percentage weight loss, mean days to regain birth weight, incidence of intraventricular haemorrhage, incidence of periventricular leukomalacia, incidence of retinopathy of prematurity greater than stage 2, or rebound phototherapy (weight loss: 12.1% ± 4.9% with prophylactic phototherapy v 11% ± 5.0% with threshold phototherapy; P >0.05; mean days to regain birth weight: 11.8 ± 3.9 days with prophylactic phototherapy v 11 ± 4.0 days with threshold phototherapy; P >0.05; intraventricular haemorrhage: 15/43 [35%] with prophylactic phototherapy v 14/44 [32%] with threshold phototherapy; P >0.05; periventricular leukomalacia: 2/43 [4.7%] with prophylactic phototherapy v 2/44 [4.5%] with threshold phototherapy; P >0.05; retinopathy of prematurity >stage 2: 7/43 [16%] with prophylactic phototherapy v 11/44 [25%] with threshold phototherapy; P >0.05; rebound phototherapy (18/45 [40%] with prophylactic phototherapy v 12/47 [26%] with threshold phototherapy; P >0.05).
Low threshold versus high threshold phototherapy:
The first RCT found no significant difference between groups in patent ductus arteriosus (556/990 [56%] with low threshold v 582/984 [59%] with high threshold; RR 0.95, 95% CI 0.88 to 1.02), or necrotising enterocolitis (105/990 [11%] with low threshold v 117/984 [12%] with high threshold; RR 0.90, 95% CI 0.79 to 1.14). The second RCT gave no information on adverse effects.
Compact fluorescent light versus standard length tube light phototherapy:
One RCT found no significant difference in moderate or extreme eye pain caused by glare, moderate or extreme giddiness, and moderate or extreme headache in nursing staff with compact fluorescent light (CFL) compared with standard length tube light (STL; pain from glare: 38% with CFL v 48% with STL; P = 0.16; giddiness: 14% with CFL v 20% with STL; P = 0.42; headache: 6% with CFL v 8% with STL; P = 1).
Phototherapy versus immunoglobulin:
See harms of immunoglobulin.
