Management of late-preterm and term infants with hyperbilirubinaemia in resource-constrained settings

Guidelines for the management of hyperbilirubinaemia in high-income countries are unlikely to address the peculiar challenges in LMICs without appropriate modification [17,22,23]. Adaptation of existing evidence-based guidelines from one geographical, economic and socio-cultural context to another is an internationally accepted alternative to the more costly, time-consuming, de novo guideline development for improved health care delivery [24]. In this report, we followed relevant protocols of the WHO Handbook for Guideline Development [25], ADAPTE Guideline Adaptation Toolkit [26], and AGREE II-Global Rating Scale [27], for the adaptation of clinical practice guidelines. Based on prior in-depth reviews of the literature from 1970 to 2013 on the burden of neonatal hyperbilirubinaemia and current management practices in LMICs [17,22], we identified four major themes for improving the care of affected infants namely: primary prevention, early detection and monitoring, treatment and follow-up. We then undertook a review of existing guidelines (see Additional file 1: Table S1) and relevant literature from both high-income and LMICs to identify key issues relevant to improved care at primary, secondary and tertiary levels in LMICs [28,29]. The existing guidelines were rated individually by the core working group (CWG) and one external content methodologist who had no prior involvement with generating these guidelines. The overall average score for each guideline was computed based on the seven components of the AGREE-II instrument: methodology, presentation, completeness, appropriateness, overall quality, disposition for personal use and likelihood of recommending the guideline to others [27]. We developed a practice framework for different levels of newborn care based on essential tools and skills considered appropriate for each level of care. The proposals aimed at balancing the safe, effective, patient-centred, timely, efficient and equitable components of quality care enunciated by the Institute of Medicine [30], as well as minimising the risk of unintended harms such as costly, unnecessary overtreatment or increased parental anxiety. Where scientific evidence was lacking or limited, proposed actions were based on consensus among the CWG using the Delphi process [31]. The draft and final proposals were critically reviewed by an international panel of experts for scientific soundness and practicality. The experts were identified and agreed by the CWG based on their independently verifiable work on the subject-matter and with a view to achieving a fair representation from all world regions. While the expert panel review was not intended as individual endorsement of the entire framework, all comments and queries were carefully addressed by the CWG in subsequent revisions. Authors made reasoned judgment where contradictory views were expressed by panel members on an issue [25-27]. For clarity and consistency the key terminologies and definitions used in this report are summarised in Table 1 (also Additional file 2).

The average AGREE II-GRS ratings of the 21 guidelines reviewed ranged from 41% for Ghana and 99% for UK’s NICE (see Additional file 1: Table S1). Priority was given to guidelines with high quality scores (≥70%) except for issues pertinent to clinical practice in LMICs but not explicitly addressed by these guidelines such as factors accounting for delays in seeking and receiving appropriate care [17]. Only four guidelines (Ghana, India, Kenya and WHO) were from eligible LMICs, and two (Ghana and WHO) did not meet the 70% quality rating threshold. The American Academy of Paediatrics (AAP) guideline was the benchmark for the majority of high scoring national guidelines including NICE [4,5]. The interventions and tools proposed for each level of health care delivery are summarised in Table 2.

The contribution of maternal/family knowledge gaps regarding the importance of neonatal jaundice commonly manifesting in late presentation of infants with severe hyperbilirubinaemia to health services in LMICs is well documented [17]. Mothers and families are able to detect jaundice from yellowish discolouration of the skin in their newborns accurately, if appropriately educated [4]. Educating pregnant women, especially primigravidae during antenatal clinics, on the risks and adverse consequences of severe hyperbilirubinaemia, avoidance of potentially harmful traditional/herbal therapies and the mis(use) of haemolytic agents should be a priority [17]. Routine determination of mother’s blood type and timely provision of anti-D globulin should be widely promoted to prevent Rh and neonatal jaundice due to other haemolytic diseases [14,32]. Educational interactions with mothers and families must also recognise and seek to address common barriers to appropriate health-seeking behaviour for childhood illnesses [17]. Although most infants are born outside hospitals in many LMICs, pre-discharge counselling of mothers who deliver in hospitals on the risks of hyperbilirubinaemia after discharge should be considered.

Infants that are exclusively breastfed have an increased risk for severe hyperbilirubinaemia in the first 2 to 5 days of life compared to formula-fed infants [33]. It is therefore, essential to provide good lactation support to all mothers at all levels of care to increase successful breastfeeding, at least 8–12 times a day, as breast-milk benefits outweigh the risk [4,5,33]. Mothers, families and their jaundiced infants will also be best served by information provided during antenatal care, about hospitals in their communities that are able to provide requisite support for neonatal hyperbilirubinaemia [17].

Mothers who deliver at home, especially those who do not attend antenatal clinics present a special challenge that must be appropriately addressed in various communities. The inclusion of neonatal jaundice in the WHO recommended training on essential newborn care for traditional/home birth attendants, community and lay health workers should be considered in such settings [34-36]. The training should also be geared towards avoidance of haemolytic agents or traditional therapies, early recognition of the onset of jaundice by mothers and care givers, and surveillance for timely presentation to the nearest health facility [17].

Early identification of infants at risk of severe hyperbilirubinaemia is an essential component of newborn care. All newborns at all levels should be examined within 24 hours of birth and in the following two days. Mothers and other care-givers should be encouraged to look for jaundice by blanching the skin (on the nose in particular), looking at the gums and examining the eyes [4]. The use of Kramer’s chart (see Additional file 3: Figure S1) especially in primary care settings remains valuable despite its limitation in correlating with the severity of jaundice [37,38]. So also is blanching of the gums possibly with an icterometer, particularly in dark-skinned babies [39]. Healthcare professionals and parents are capable of recognizing jaundice, but not very good at assessing its severity [4]. Notwithstanding, this visual assessment is generally more reliable and helpful in ruling out hyperbilirubinaemia than estimating bilirubin levels [4].

The suggested pathways of care for all babies, adapted from NICE guidelines, are described in Figure 1 [4]. At a minimum, infants with gestational age <38 weeks, from high-risk racial groups, previous sibling(s) with a history of treated jaundice, visible jaundice in the first 24 hours of birth, family history of G6PD deficiency or blood group incompatibilities, should be considered as high risk and should be monitored for hyperbilirubinaemia at all levels of care [4,5]. Ongoing training of health workers on the signs and symptoms of ABE including the use of a protocol for bilirubin-induced neurologic dysfunction (BIND) is essential for facilitating timely referral and intervention (see Additional file 4: Table S3) [40,41]. Subtle, moderate and occasionally even advanced BIND is likely reversible [42,43]. Timely detection and treatment of BIND as suggested by the AAP and others is therefore, useful in forestalling the progression of this potentially devastating condition.

The objective tests for estimating and monitoring the degree of jaundice are transcutaneous bilirubin (TcB) and/or TSB. TcB is a non-invasive, portable screening tool ideally used to determine the need for the more accurate TSB which requires a venous or capillary blood sample [44-47]. TcB values above 12 mg/dl (205 μmol/L) should be checked where possible with TSB measurement [47]. TcB becomes unreliable after commencement of phototherapy unless measurements are taken from an area of skin that has been shielded from phototherapy with a photo-opaque patch [47]. However, currently these devices, especially TSB, may not be readily affordable in many resource-limited settings. Low-cost and minimally invasive point-of-care tools for plasma/serum bilirubin measurements are currently being piloted and hold promise for LMICs [48]. The interpretation of TSB/TcB and the recommended actions are provided in Table 3 [47-51]. Besides TSB estimation, access to laboratory facilities for real-time clinical investigations should at a minimum include evaluation of blood group incompatibilities and G6PD status [4,5].

For infants delivered in hospitals and discharged before 48 hours, follow up assessment within 1–2 days after discharge with a TcB, or at a minimum, physical examination should be considered. Table 3 may be used to assess the risk of subsequent hyperbilirubinaemia at the time of discharge. Infants who present in the first week of life should be routinely examined for possible jaundice.

Phototherapy and exchange transfusion are well-established as the most effective treatments for severe hyperbilirubinaemia [4,5]. The proposed actionable TSB/TcB levels for phototherapy and exchange transfusion reflecting available evidence on current practices for the care of jaundiced infants in LMICs are presented in Table 3 [17,22,52]. We adopted the tabular format in the Kenya 2013 guidelines (see Additional file 1: Table S1) for ease of reference at all levels of care. These conservative bilirubin levels may be warranted in settings where the incidence of severe hyperbilirubinaemia is high, late presentation common, determination of haemolytic risk (Blood type/Rh/G6PD) is not possible routinely at birth, and quality of phototherapy is sub-optimal. Higher levels for phototherapy and especially exchange transfusions at or near those recommended by the AAP or NICE guidelines should be strongly considered in tertiary care settings with facilities for intensive phototherapy. A centre or hospital at any level that is not appropriately resourced to provide the required treatment should promptly refer the infant to the closest, appropriate health facility.

There are various phototherapy devices using different light sources: fluorescent tubes, halogen lamps and light emitting diodes (LED). An effective phototherapy device should produce specific blue-light wavelengths (peak emission: 450 ± 20 nm), preferably in a narrow bandwidth to about 80% of an infant’s body surface area [18]. Conventional phototherapy (CPT) should have an irradiance of at least 8-10 μW/cm²/nm and intensive phototherapy should have an irradiance of ≥30 μW/cm²/nm (from either a single or multiple phototherapy units). LED devices are as effective as other light sources in decreasing TSB but have special advantages in LMICs [53,54]. They are more power efficient, portable, weigh less, have a longer life span and lower heat production, making them more suitable for intensive phototherapy than fluorescent bulbs. Irradiance meters for monitoring PT units should be readily available, as well as spare replacement bulbs.

Costs of providing intensive or special care for jaundiced newborns could be prohibitive, next only to that of caring for preterm babies in LMICs [55]. However, the development of affordable phototherapy devices and using simple inexpensive enhancements such as hanging white reflecting material around cots (being careful to avoid overheating particularly from halogen lamps), changing bulbs regularly and reducing the distance between baby and lamps, improve the effectiveness of phototherapy [18,19,56,57]. WHO maintains a valuable compendium of innovative and low-cost technologies including phototherapy devices recommended for LMICs [58]. Whatever the light-source, the effectiveness of phototherapy devices can be compromised by erratic power supply, inadequate skin exposure from overcrowding with multiple infants placed under a single device, sub-optimal irradiance levels, and poor device maintenance [18,19]. Failure to address these issues is likely to increase the frequency of avoidable exchange transfusion. It is, therefore, essential that phototherapy devices are properly monitored, regularly maintained, and staff well trained on how to care for infants receiving phototherapy. If enough effective phototherapy devices are not available in very busy resource-constrained settings, clinical judgement as to which infants should receive priority based on the risk of ABE may be required. Under such conditions, a BIND assessment can aid decisions for intervention [40,59].

Guidelines for hyperbilirubinaemia in many high-income countries prohibit exposure to sunlight as a form of treatment [4,5,60]. This is primarily due to safety concerns regarding potentially harmful infrared and ultraviolet rays and possible sunburn. In Ghana, exposure of jaundiced infants to sunlight is acknowledged as a treatment option [61], but WHO guidelines and some other developing countries like Malaysia, India and Kenya discourage or have not made provisions for sunlight therapy (see Additional file 1: Table S1). However, mothers and caregivers with or without the support of health workers continue to expose their jaundiced babies to direct sunlight even in high-income countries [62-66]. Emerging evidence suggests that the potential risks can be mitigated through specially filtered film canopies which have been successfully piloted in West Africa [67,68]. However, their use is still experimental and limited to daytime care at periods with favourable climatic conditions. In remote tropical locations where access to conventional treatment is not assured, the trade-off between the risk of ABE/kernicterus and of exposure to sunlight remains a challenge that requires individual and informed judgment call in the patient’s best interest.

Immediate exchange transfusion is warranted when phototherapy has failed to effectively curtail the rate of bilirubin rise and the TSB/TcB measurement is near exchange levels or the infant has any of the signs of moderate to advanced ABE regardless of the TSB/TcB levels (Figure 1, Table 3). This treatment is most likely to be available at tertiary hospitals with trained personnel and facilities for special care, including monitoring and resuscitation capabilities. As exchange transfusion is not without risks [69-71], its frequency should be minimised as far as practicable. Exchange transfusion with G6PD deficient donor blood should be avoided where possible as this may prolong time under phototherapy and result in repeat exchange transfusions [72]. Similarly, blood should be screened for HIV and hepatitis. Rh-negative blood should be used for neonates with Rh-isoimmunisation while O group should be used for neonates with ABO incompatibility.

The evidence in support of pharmacotherapies such as D-penicillamine, phenobarbital, metalloporphyrins, clofibrate, bile salts, laxatives and bilirubin oxidase are inconclusive and these interventions have not been recommended [73]. Likewise traditional herbs or medications used to treat newborn jaundice in many home settings are not recommended.

The manifestations of BIND such as cerebral palsy, auditory impairments, epilepsy, gross motor deficits, behavioural problems and intellectual difficulties are not uncommon in LMICs [9-13]. Follow-up evaluation of survivors of severe hyperbilirubinaemia for potential neurodevelopmental sequelae is necessary to facilitate early detection and intervention for the affected infants. This must be considered as an integral part of any clinical protocol for the management of infants who have been treated for severe hyperbilirubinaemia. Because hearing impairment, including auditory neuropathy spectrum disorders in the first year of life is often not clinically apparent, at-risk infants must be objectively evaluated within the first three months and monitored for language development in the first two years of life irrespective of the hearing test result. Referral to the audiology or otolaryngology unit of a tertiary hospital should therefore be considered. Several low-cost and simple-to-use validated tools for early developmental assessment as well as approaches to effective intervention in resource-limited settings have also been documented [74,75].

While we set out to ensure that our recommendations are realistic and consistent with the prevailing conditions in most LMICs, the methodological approach used in this document deserves clarification. We followed standard protocols for guideline adaptation including a comprehensive review of the priority issues to be considered, the formation of a working group, quality rating of existing guidelines and independent review by a panel of leading clinicians and experts on newborn care from different world regions [25-27]. However, we did not grade specific studies in support of various proposals in this paper, more so because the body of evidence was predominantly adapted from existing guidelines or supported by the best available data from LMICs [17,22,76]. Neonatal hyperbilirubinaemia is frequently underpinned by complex interactions of diverse biological and environmental risk factors across populations. It is our view therefore, that each LMIC still needs to develop context-specific guidelines for their own population [77]. Such an initiative should be broadened to involve key stakeholders including parent groups and community leaders. It may be helpful to prepare separate documents for each level of care to avoid information overload especially at the primary care level. While the scope of this framework is not exhaustive, like any of the existing guidelines, we believe that most of the recommendations will still be valuable to LMICs that did not meet our selection criteria. Finally, strong clinical and public health leadership at all levels will be required to surmount the challenges that typically mitigate against initiatives for improved child and newborn care in LMICs [78,79].