Patients were included by consecutive sampling. The patients in the bag trial were newborns admitted to the neonatal care unit (NCU) of the Hospital Universitario San Ignacio (HUSI) of Bogota, Colombia, between March 1, 2016 and October 5, 2017. Patients in the blanket trial were enrolled in the KMC programme of the HUSI between February 8, 2016 and November 15, 2017. In the NCU, the need for phototherapy was established based on serum bilirubin levels, and in the KMC it was established using the Bilicheck® transcutaneous bilirubinometer with subsequent confirmation by measurement of serum bilirubin.

We calculated the necessary sample sizes for 2 groups with a 1:1 allocation ratio for a statistical significant threshold of 0.05 and a power of 80 % for detection of a difference in the mean decrease in bilirubin of at least 0.15 mg/dL/h assuming a standard deviation of 0.2 mg/dL based on the current evidence.11,12 The sample for the phototherapy bag trial was of 29 patients per group, and the sample for the bag trial, adjusted for losses of 5 %, of 32 patients per group.

The exclusion criteria for the bag trial were: patients requiring an incubator (risk of hyperthermia), patients with skin lesions that complicated administration of phototherapy or patients requiring transfer to another hospital. For the blanket trial, the exclusion criteria were onset of jaundice before 48 h post birth, bilirubin level greater than18 mg/dL, diagnosis of diseases requiring hospital admission, extensive skin lesions or parents that could not stay at the KMC programme a minimum of 6 h.

We randomised patients to different groups using computer software and assigned groups to a given intervention using opaque, sealed envelopes.

After obtaining the signed informed consent of the parents, the intervention in the bag trial consisted of conventional phototherapy with blue light emitted by fluorescent tubes or phototherapy with the NeoMedLight Bilicocoon Bag®, with patients in both groups receiving continuous treatment while wearing eye protectors.

In patients with risk factors such as haemolytic disease, glucose 6-phosphate dehydrogenase deficiency, asphyxia, significant lethargy, unstable body temperature, sepsis or acidosis, we measured serum bilirubin levels every 8 h until 2 consecutive values were the same or showed a decrease, followed by measurement of serum bilirubin every 12 h. In patients without risk factors, we measured serum bilirubin levels every 24 h. The treatment was discontinued once the serum bilirubin was below the threshold for indication of phototherapy. In patients with haemolytic disease, we collected blood samples 6–12 h after discontinuation of phototherapy to ensure that bilirubin levels did not increase again.

In the blanket trial, patients received phototherapy with the NeoMedLight Bilicocoon Nest® or the Ohmeda BiliBlanket® in Kangaroo position. The administration of phototherapy was intermittent, given for 2-h intervals with 60 min of rest in between. Levels of bilirubin were measured with the Bilicheck® bilirubinometer and with blood tests at the beginning and end of treatment. If the levels were below the threshold established for treatment or the patient completed three 2-h sessions, a second blood sample was obtained and treatment discontinued, discharging the patient with a follow-up the following day, or the patient was admitted to hospital.

In inpatients, the axillary temperature was measured at start, at 30 min, 1 h, 2 h and thereafter daily and at the end of phototherapy; in outpatients, it was measured at start, at 30 and 60 min and at 2, 4 and 6 h. Patients underwent a physical examination at admission and after completion of treatment to ensure that they did not become dehydrated or develop skin lesions. In the bag trial, patients underwent physical examinations with measurement of weight and testing of fluid and electrolyte balance at baseline, every 24 h and after completion of phototherapy. In both trials, we assessed for the development of side effects.

We surveyed parents and health care staff at the end of treatment to assess the perceived comfort and ease of use of the devices, with items rated on a Likert scale.

In the inpatient trial, we used a phototherapy unit with fluorescent tubes as the source of light (Medix LU-6T®) in the control group and the NeoMedLight Bilicocoon Bag® in the treatment group. The latter is a sleeping bag with ventral and dorsal light exposure, that allows insertion of the infant in a plastic sheath, with two 20 × 30 cm pads of fibre optic fabric connected to a light box, that allows programming the duration of phototherapy. The emission spectrum ranges from 400 to 500 nm with peak wavelengths between 458 and 460 nm. The irradiance is of 35 µW/cm2/nm. The 2 phototherapy pads envelop the baby and are held by a disposable bag that is adjustable to the infant based on weight and height, containing and soothing the infant and preventing the pads from sliding off during use. The bag has flaps on the side that can be opened to allow air circulation, thus preventing hyperthermia.13,14 In the outpatient trial, the device used in the control group was the Ohmeda Medical Biliblanket Plus®, de Ohmeda Medical, which has a high-output light box that delivers light through a fibre optic blanket. The patient is exposed to light in the 400-to-550 nm spectrum. The phototherapy system consists of a light source, a 121.9 cm long fibre optic cable and a lightweight pad measuring 11.12 × 23.83 cm. The light unit houses a high-intensity tungsten halogen lamp with a built-in reflector. In the treatment group, the device used was the NeoMedLight Bilicocoon Nest®, a LED-lit pad with a surface area of 40.5 × 30 cm with a 20 × 30 cm fibre optic fabric swath encased in a protective layer of vinyl with a disposable cover. The emission spectrum is of 400–550 nm, with peak wavelengths between 456 and 460 nm and a mean irradiance of 35 µW/cm2/nm ± 15 %, exceeding the minimum recommended by the AAP for intensive phototherapy13,15 (Fig. 1).

Figure 1.

Top: Bilicocoon Bag®. Bottom left: Ohmeda Biliblanket® and Bilicocoon Nest®. Bottom right: Bilicocoon Nest® during kangaroo care.

(0.28MB).

We recorded the irradiance of each device at the beginning of the trial, the duration of phototherapy until achievement of a bilirubin level below the threshold for treatment (in hours), the decrease in bilirubin measured in blood samples (in mg/dL/h), axillary body temperatures, the frequency of exchange transfusion in the bag trial, the frequency of hospital admission in the blanket trial, and the mortality. In addition, we recorded the incidence of skin lesions, dehydration or diarrhoea attributable to phototherapy. When it came to the questionnaire on the comfort and ease of use of devices used to survey parents and nursing staff, we compared the frequency of established score categories in the Likert scale for each of the items (1–3 = poor; 4 = neither good nor bad; 5–7 = good).

The trials were not blind; however, most endpoints, such as the level of bilirubin or the body temperature, were not susceptible of being influenced by awareness of assignment status.

We compared means for analysis of quantitative data and used the chi square test to analyse qualitative data. We performed a multiple linear regression analysis of the decrease in the level of bilirubin (mg/dL/h) in hospitalised patients. The analysis was by intention to treat.

Both the trial protocols and the informed consent forms were evaluated and approved by the ethics and research committee of the hospital.

Of 86 eligible patients, 78 completed the study. The initial sample size was 84 for 4 paired samples, however, most patients underwent a maximum of 3 measurements. We recalculated the sample size based on the difference in the mean decrease in the level of bilirubin (mg/dL/h). The new sample size was of 58 patients, 29 per group. We randomised an additional 22 patients (Fig. 2, left).

Figure 2.

Left: Bilicocoon Bag®. Right: flujograma Bilicocoon Nest®.

(0.31MB).

The initial sample had 86 patients: 40 assigned to the Ohmeda Biliblanket® and 46 to the NeoMedLight Bilicocoon Nest®. The sample exceeded the calculated size of 64 patients because we initially considered comparison of repeat measurements with the Bilicheck® bilirubinometer, estimating that we needed 94 patients; however, during the study we found a low correlation between the initial and final levels measured with the Bilicheck® and the initial and final serum levels, between 0.43 and 0.53. We decided to assess the decrease in bilirubin levels (mg/dL/h) based on the serum concentration. Of the patients in the initial sample, 7 did not meet the criteria to continue the study, 6 requiring hospital admission the next day due to persistent hyperbilirubinaemia, while phototherapy was effective in 73 (Fig. 2, right).

When it came to the measured irradiance, we found the highest level with the conventional phototherapy device (37.3 ± 10.3 mW/cm2/nm), followed by the bag (36 ± 2.6 mW/cm2/nm) and then the nest (34.8 ± 2.4 mW/cm2/nm). The lowest level corresponded to the Ohmeda Biliblanket® (21.6 ± 8.8 μW/cm2/nm).

Tables 1 and 2 show the characteristics of the patients and the beginning of the trial. When comparing the groups treated with the bag system and conventional phototherapy, we found a greater proportion of patients with early-onset jaundice (<36 h post birth) in the bag group (P = .05). The characteristics of the groups were similar otherwise.

As for the outcomes, we found a similar decrease in bilirubin levels with the bag and the conventional phototherapy device (difference in means, 0.03 mg/dL/h; 95 % CI, −0.04 to 0.09). We found difference in body temperature at the end of treatment that was clinically insignificant. There were no cases of hyperthermia, adverse events, need of exchange transfusion or death (Table 3).

The decrease in bilirubin levels with the new phototherapy blanket (Bilicocoon Nest®) was significantly greater compared to the one achieved by the Ohmeda Biliblanket®. We did not find differences in final body temperature, hours of phototherapy, days of phototherapy, readmission rate, side effects or mortality (Table 4).

The linear regression model in which the bilirubin decrease rate was the dependent variable and the type of device and the presence of early-onset jaundice as predictors, found that the bilirubin decrease rate was not associated with the phototherapy device used (P =  0.8) after controlling for the effect of early-onset jaundice (< 36 h) (Table 5).

In the inpatient trial, the ratings on the Likert scale given by parents and nurses in the satisfaction survey for the bag system showed that parents did not perceive any differences in comfort when it came to kangaroo care, breastfeeding, body temperature or exposure to light. Nurses expressed that infants seemed to be less comfortable in the bag system (Table 6).

In the blanket trial, the survey showed that parents and nurses did not perceive any differences between the compared devices.