Study design, duration and study area
A parallel randomized controlled trial to test three treatment groups was conducted in the neonatal care unit (NCU) at the Kilimanjaro Christian Medical Centre (KCMC) in Northern Tanzania from January 2019 to May 2019. The NCU has a bed capacity of 62, with an average nurse to patient ratio of 1:10 per shift. The babies are nursed in locally made cots which are heated by incandescent bulbs. In our neonatal care unit, conventional phototherapy has been the only treatment modality used for unconjugated hyperbilirubinemia.
Sixty-five term neonates (>37 weeks of gestation), less than 7 days of age, admitted with jaundice, were identified by clinicians working in the neonatal unit and screened for eligibility. We included those with both hemolytic and non-hemolytic unconjugated hyperbilirubinemia with a total serum bilirubin level that has reached phototherapy threshold values as per the American Academy of Pediatrics (AAP) nomogram which is a validated tool used for making decision regarding phototherapy in infants with unconjugated hyperbilirubinemia. The threshold for initiating phototherapy in term neonates with no risk factors (risk factors = isoimmune hemolytic disease, asphyxia, significant lethargy, temperature instability, sepsis) was the low risk line on the AAP nomogram. Whereas the threshold for initiating phototherapy in term neonates with a risk factor was the medium risk line on the AAP nomogram. Excluded from the study were: neonates receiving phenobarbitone, neonates with bilirubin levels that have reached exchange transfusion levels on the AAP nomogram, neonates with conjugated hyperbilirubinemia, neonates who have already received phototherapy prior to enrollment, and those whose parents refused to consent.
Sample size estimation
In order to obtain our sample size, we estimated a 34µmol/L effect size in the three arms, with a level of significance of 5%, power of 85% and an estimated SD of 24µmol/L for the level of serum bilirubin after 24 hours. The minimal expected sample size in addition to non-response rate of 20% was 39. We included in our final analysis a total of 41 term neonates who met the inclusion criteria.
Study variables and data collection method
Written consent was obtained from parents of eligible participants. A questionnaire was used to collect demographic data such as sex, gestational age at birth (determined according to the maternal history and Ballard’s scoring system), mode of delivery, birth weight in kilograms, age at enrollment in days, and mode of feeding.
Simple randomization was performed by drawing a paper from a container containing 45 folded papers: 15 marked ‘FB’ (Fiberoptic BiliBlanket), 15 marked ‘B’ (Bluelight PT) and 15 marked ‘W’ (White light PT). Participants were then allocated to intervention groups by the research assistants. This was an open label trial whereby both the parent of the participants and the researchers knew which intervention the neonate was receiving.
Before phototherapy initiation, a clinical assessment was done on the neonate, and vitals were recorded. Laboratory characteristics such as total serum bilirubin and direct serum bilirubin were recorded. Phototherapy was employed as per our standard nursery protocol and equipment. The neonate was nursed in an open cot wearing only a diaper which was folded to allow maximum skin exposure to phototherapy, and was turned every two hours from prone to supine positions.
The white light PT unit “Atom model PIT- 220 TL” (Atom Medical), consists of six white fluorescent bulbs (PHILIPS, TL-D 18W/64-765) and was placed 35 cm above the neonate with an illuminated cabinet of 66.5cm x34.5cm and a mean irradiance of 8 µW/cm2/nm. The blue light PT unit “Olympic Bili-Lite model 66” (Olympic Medical), consisted of four Olympic blue fluorescent bulbs placed 20cm above the neonate with an illuminated cabinet of 66cm x 27.9cm and a mean irradiance of 27 µW/cm2/nm. Eye pads were used for neonate on conventional phototherapy to prevent damage to the retina. The BiliBlu LED fiberoptic phototherapy unit (Martand Medical Services) with a body surface area of 25cm x 40cm was wrapped around the neonate with a mean setting of 34 mW/cm2/nm irradiance. Irradiance of the phototherapy units was measured using the GE Healthcare BiliBlanket Light Meter II. Irradiance in all phototherapy units was measured from the skin surface of the neonate, 3 measurements were taken: one measurement centrally and two measurements at the two most extreme peripheries exposed to the PT light, subsequently an average irradiance was calculated for each participant. Phototherapy was administered continuously except during minor procedures such as feeding, diapering, physical examination and capillary blood sampling.
Serum bilirubin reduction rate was assessed as the decline in the TSB levels for the duration of exposure to phototherapy, expressed as a percentage of decline per hour. This was monitored through daily total serum bilirubin levels. Treatment failure was defined as the need for additional phototherapy units (‘double phototherapy ‘) determined by serial serum bilirubin response while on phototherapy; a rise in serum bilirubin level more than 9μmol /L per hour after phototherapy initiation was an indicator for double phototherapy. Phototherapy was stopped when serum bilirubin levels were 50mmol/l below the phototherapy threshold value on the AAP nomogram. Treatment duration was then assessed as the duration of time of exposure to phototherapy in hours.
Blood samples were analyzed at the KCMC clinical laboratory. The Cobas Integra 400 Plus (Roche Company) was used to analyze the serum bilirubin levels. The sample reached the laboratory within 15min of collection. A minimum of 1ml of blood was required. Care was taken to prevent exposure to light by placing the samples in a dark blood transportation box.
Neonates were closely monitored for short term side effects of treatment which included: loose, greenish stools; hydration status: assessed as the difference in daily body weight after starting PT (5% body weight loss was considered as mild dehydration, 5-10% body weight loss was considered as moderate dehydration and more than 10% body weight loss was considered as severe dehydration); skin rashes and brownish discoloration of skin assessed by a dermatologist. Participant’s vital signs including temperature, respiratory rate, heart rate were monitored every 4 hours throughout phototherapy. Weight was checked daily.
Feeds were calculated based on the daily kilocalories requirement for age. At enrollment, mothers were asked to express breast milk to quantify the milk production. When necessary, neonates were supplemented with infant formula milk to maintain the total daily feeds requirement. Neonates were fed every 3 hours in all PT groups. Those who were not maintaining adequate oral intake were supplemented with IV fluids to maintain the total daily fluid requirement, considering the postnatal age, and clinical and laboratory findings.
Research assistants recruited for the study were registered nurses working in the NCU. Research assistants enrolled the participants for the study. They were trained on how to randomize and assign participants to intervention. They were familiarized with the study protocol for PT, how to correctly place the PT unit, when to interrupt PT with clear documentation of duration of PT interruption, and measurement of the body weight & vital signs (axillary temperature, heart rate, respiratory rate). A pilot study was conducted to assess the feasibility of the fiberoptic PT unit and to familiarize the research assistants with the study.
Data analysis was done using Statistical Package for Social Sciences (SPSS) version 23. Mean and standard deviation was used for descriptive statistics. The data was analyzed with the independent t-test to compare means of bilirubin reduction rate (%/hr) and treatment duration between the treatment groups. Analysis of variance was used to compare the baseline and clinical characteristics between the three treatment groups. A p-value of ≦0.05 was regarded as statistically significant.