Severe neonatal hyperbilirubinemia, defined as total serum bilirubin (TSB) ≥20 mg/dl, is associated with a higher risk of permanent neurological sequelae and death. Jaundice can and should be promptly diagnosed and treated. Reliable methods for TSB assay are not always readily available, particularly in low- and middle-income countries, making the true incidence of severe neonatal jaundice (NNJ) difficult to estimate. To gather a more comprehensive picture, a symposium addressing NNJ worldwide was organized during the 2015 Don Ostrow Trieste Yellow Retreat. Data collected by several researchers in different regions of the world were presented and differences/similarities discussed. This report points out the need for: (1) a coordinated worldwide effort to define the burden and the causes of severe NNJ and its consequences; (2) aggressive educational programs for families and health personnel to facilitate timely care-seeking, and (3) accurate diagnostics and effective phototherapy.
Over 60% of all newborns develop neonatal jaundice (NNJ), a physiologic condition characterized by yellowish discoloration of the skin and conjunctiva as a consequence of increased levels of serum bilirubin during the first week of life [1,2,3]. NNJ is usually benign, but in some cases it can progress to severe hyperbilirubinemia, acute bilirubin encephalopathy (ABE) and kernicterus/chronic bilirubin encephalopathy (CBE) [4,5,6,7,8]. ABE and CBE are largely preventable if severe hyperbilirubinemia is identified early and treated promptly with effective phototherapy or, for hazardous cases, exchange transfusion. Guidelines for managing jaundice have been proposed by the American Association of Pediatrics (AAP), the UK National Institute for Health and Care Excellence (NICE) and others [8,9,10,11,12,13,14,15,16,17,18,19,20].
With improvements in prevention and treatment, the number of cases of severe hyperbilirubinemia in high-income countries (HICs) has decreased markedly since the 1990s [21,22]. As assessed by population-based studies and registries, the incidence of severe hyperbilirubinemia in HICs is currently estimated to be about 31.6/100,000 live births (95% CI 11.8-51.3) [23,24,25,26,27], while the incidences of ABE and CBE have been estimated as being in the range of 1.0-3.7 and 0.4-2.7/100,000 live births, respectively [28,29,30].
The situation is completely different in low- and middle-income countries (LMICs). No harmonized protocols for hyperbilirubinemia classification and management have been implemented in most LMICs, leading to wide variations in protocols and rendering difficult if not impossible comparisons between different locations. The classification of hyperbilirubinemia in the countries included in this article was usually established at a local level, with the exception of Malaysia and Egypt which adopted the AAP guidelines for NNJ management. Despite these limitations, the prevalence is said to be high in LMICs, where records and documentation of the incidence of NNJ, ABE and CBE are usually poor and variable [1,20,31].
A recent Child Health Epidemiology Reference Group (CHERG) modeling study used country-specific and regional estimates of the prevalence of Rhesus (Rh)-positive babies born to Rh-negative mothers, G6PD deficiency, moderate-to-late preterm birth (i.e. 32-36 weeks gestation) and infants with none of these 3 factors, to estimate the risk of neonatal mortality and/or survival with kernicterus worldwide and by geographical region . The CHERG study estimated that Rh disease and/or extreme hyperbilirubinemia (EHB; TSB >25 mg/dl) due to other causes, were responsible for a mortality rate of 119/100,000 live births in Eastern Europe/Central Asia, Latin America, sub-Saharan Africa and South Asia, compared with 1/100,000 live births in HICs . In these same 4 geographical regions, the prevalence of kernicterus was estimated at 73/100,000 live births, compared with 10/100,000 live births in HICs .
The Don Ostrow Trieste Yellow Retreat (DOTYR-15; fig. 1) is a meeting where basic and clinical aspects of bilirubin are discussed by international experts. In 2015, for the first time, a session was devoted to reports from LMICs from Africa and Asia. This article summarizes what was reported at this meeting on the prevalence of NNJ, ABE and CBE in these countries.
It is important to note that information reported from LMICs is largely drawn from tertiary hospitals that, overwhelmingly, treat infants admitted after the onset of hyperbilirubinemia in the community [1,20]. The denominator, in the studies described here, is usually the number of infants admitted with NNJ, and thus overestimates the population-based prevalence. If referral hospital(s) were able to capture a whole catchment area where the number of live newborns was known, the incidence of severe jaundice, ABE and CBE might be significantly underestimated. The limited information we report here is therefore reflective of failures in the system of jaundice management, diagnosis and documentation in many LMICs.
NNJ in Selected African Countries
Recent studies have reported important data on severe hyperbilirubinemia, ABE and CBE in African countries (table 1). An Egyptian study of 247 babies with TB ≥25 mg/dl admitted in 2008 to the Cairo University Children Hospital, a referral children's hospital, reported that 44 (17.7%) presented with moderate or severe ABE and 26 (10.4%) died, leading to a case fatality rate (CFR) of 56.8% (26/44) . A separate study in the same hospital over 15 months in 2009-2010 reported on 193 infants requiring intensive phototherapy or exchange transfusion; 58 (30.0%) had moderate or severe ABE at presentation, with kernicterus diagnosed in 35 (60.3%), and 13 with severe ABE died, giving a CFR of 22.4% among those with ABE .
A Kenyan study at a pediatric referral center reported that a total of 306 infants were admitted in the year 2000, with 106 (34.4%) being diagnosed with jaundice; 24 of the jaundiced infants died, giving a CFR of 22.7% . A separate study in a Kenyan district hospital over 19 years (1990-2008) reported that out of 8,756 neonatal admissions, 811 (9.2%) had severe NNJ and 116 of these died, giving a CFR of 14.3% .
A recent comprehensive review of 198 studies (1990-2014) based primarily on single-hospital experiences in Nigeria concluded that little progress has been made over the last 50 years . Severe NNJ and kernicterus remain highly prevalent and continue to be associated with a high CFR [20,37,38,39,40,41,42,43,44,45,46].
As part of an ongoing ‘Saving Lives at Birth (SLAB)' project, a collaborative study of severe NNJ and ABE prevalence was performed at 9 hospitals in 6 regions of Nigeria in 2014-2015. The prevalence of severe hyperbilirubinemia was high: 26.9% of admissions for jaundice had bilirubin levels >20 mg/dl and 14.9% of hospitalized neonates developed ABE, with a CFR of 13.0% based on infants with NNJ (unpubl. data). A delay in seeking care is common in newborns with severe NNJ . As a consequence, advanced ABE that benefits little from therapy is often present at the time of admission. A high prevalence of G6PD deficiency, combined with exposure to oxidants (e.g. mothballs and menthol creams), is a major cause of hemolytic jaundice, followed by ABO incompatibility and Rh isoimmunization [1,20].
NNJ in Selected Asian Countries
The prevalence of serious outcomes associated with NNJ in Asian countries varies widely (table 2). The Chinese Medical Association conducted a survey of all infants discharged from 86 general and maternity hospitals in 2005 and found that 49.1% had NNJ; 0.9% developed bilirubin encephalopathy, with a higher rate in the term neonates (0.9%) than in the preterm neonates (0.5%) . The authors suggest that the higher rate in the term infants could be due to an assumption that jaundice in this group is physiological in origin, leading to a failure to consider pathological risk factors.
Published data on NNJ in Bangladesh is limited. A prospective cohort of neonatal admissions to the Khulna Medical College Hospital over 36 months in 2005-2008 reported that 15.7% of infants presented with NNJ, with 2.8% of these developing kernicterus and 5 of them dying (CFR 55.6%) . In another study, 5.9% of the neonates admitted developed severe jaundice while 0.2% had jaundice-associated deaths (CFR 3.9%) . It is important to note that G6PD deficiency is common in Bangladesh and may be an important contributor to the development of severe jaundice and kernicterus .
From India, Dutta et al. [50 ]reported that severe jaundice represented 15.3% of neonatal admissions, with a CFR of 6.7%, leading to 4.4% of the deaths related to jaundice. An observational study by Bang et al. [51 ]reported that severe jaundice had a mortality rate of 7.3/1,000 live births in Indian rural villages.
In 2013, statistics from Myanmar government hospitals reported that NNJ was responsible for 46% of hospital admissions country-wide and was a major cause of neonatal morbidity and death . A study of neonates treated with phototherapy in 2 specialized pediatric referral hospitals identified home-births, self-referrals and G6PD screening status as important risk factors for presentation with ABE .
In 1999, a preliminary report of the National Perinatal Health Conference stated that about 25-30% of babies admitted to selected hospitals in Malaysia were severely jaundiced with serum bilirubin of >20 mg/dl . The incidence of severe jaundice reported from 4 Malaysian major centers in 2012 varied widely. In 3 of these centers, severe jaundice (bilirubin levels >20 mg/dl) represented 1.7-3.5% of NICU admissions, while data reported by Selayang Hospital at DOTYR-15 showed that severe jaundice was present in 15.8% of its NICU admissions. Data on ABE and kernicterus have not been routinely or systematically collected, so their incidence in Malaysia is unknown.
A recent study analyzing NNJ in Indonesia with SLAB funding assessed the prevalence of severe jaundice (TSB >20 mg/dl) and ABE at 8 hospitals in 3 regions of Indonesia [unpubl. data]. The prevalence of severe jaundice and ABE in babies admitted during the study was 6.8 and 2.2%, respectively, leading to a CFR caused by NNJ of 24.2 and 74.9%, respectively, for ABE. However, most of the deaths from ABE were associated with neonatal sepsis. The data showed different patterns of severe hyperbilirubinemia and ABE. Of note is the observation that the majority of ABE cases were in hospitals located in remote regions, where fewer resources for NNJ management are available.
Future Data Needs and Directions
The data reported above (graphically represented in fig. 2) are not based on a systematic review of the literature; rather, they reflect a convenience sample of data familiar to the DOTYR-15 attendees (for the list of authors who participated in the collection of the data, see the Appendix). Nevertheless, it is clear that there is a need for a data collection strategy that can inform the development of a worldwide approach for the early diagnosis and appropriate management of NNJ, in order to prevent the tragedy of bilirubin-induced neurological damage. Such a strategy could combine internationally developed frameworks for data collection that are adopted and adapted to local conditions and capacities. The process of local adaptation could commence with national scoping studies like that undertaken in Nigeria  that identify and review relevant data already published in peer-reviewed and grey data. This information can then be used as a basis for identifying the research priorities and routine collecting of information that are needed to inform national policy and practice.
The Present and the Future in Preventing Serious NNJ
The CHERG modeling study estimated that 78% of cases of EHB are attributable to Rh disease, 6% to G6PD, 2% to moderate/late preterm birth and 15% to other causes; 80% of the affected neonates are in countries with a neonatal mortality rate ≥15/1,000 live births with the distribution of risks similar to that found worldwide .
Although the incidence of severe hyperbilirubinemia due to Rh disease in sub-Saharan Africa appears to be less than that due to G6PD deficiency , it can be prevented by anti-D prophylaxis . The CHERG model estimates that there is no EHB attributable to Rh disease in HICs thanks to effective prevention and treatment, but at the same time notes that the high cost of immunoprophylaxis needs to be addressed in LMICs . The importance of Rh disease varies widely with the prevalence of Rh-negative maternal status, but the issue of affordable prophylaxis is clearly a priority for many LMICs. The lack of routine maternal and neonatal blood testing in many LMICs is also a barrier, and also impacts the management of ABO incompatibility as a risk factor for serious NNJ [7,32,53,56,57,58,59].
A WHO Working Group recommended in 1989 that, in all regions with a prevalence of G6PD deficiency of 3-5% or more in males, there should be universal screening accompanied with an education campaign for parents and health workers [60,61]. A recent review noted the lack of empirical evidence of the efficacy of such programs in reducing the risk of severe hyperbilirubinemia, but nevertheless echoed this recommendation on the basis of the available evidence and understanding [1,9,10,31,53,56,62,63,64,65,66]. While screening of cord-blood for G6PD deficiency is of relatively low cost, the experience of the DOTYR-15 participants is that there would need to be a substantial investment in infrastructure to support universal screening programs.
Where routine maternal and neonatal blood testing and universal screening for G6PD deficiency are currently unavailable, a number of low-cost interventions remain immediately feasible. Extremely important are: the education of parents and health care workers on the risks of bilirubin-induced neurological damage, the avoidance of exposure to triggers of hemolysis and the appropriate follow-up of newborns after delivery [67,68,69,70]. Public awareness and parental training for the identification of the signs of NNJ [68,69,70,71,72,73,74,75], together with proper training for health care providers [10,75,76,77], can contribute significantly to reducing the high prevalence of severe hyperbilirubinemia and ABE in low-resource settings.
Furthermore, the lack of affordable tools for the real-time objective measurement and monitoring of bilirubin levels continues to be a challenge [1,31]. Health facilities often face long delays for TSB results and in primary health care settings or remote areas, testing is usually unavailable. The Bilistick System (Bilimetrix S.R.L., Trieste, Italy) is a low-cost, point-of-care bilirubin assay able to provide early NNJ diagnosis by determining TSB concentration from a tiny drop of blood, and is currently in advanced testing . If successful, it could fulfil the need for rapid results and predischarge screening in already-equipped facilities and for decentralized testing and referral to support visual assessments.
The lack of effective phototherapy units, especially those capable of providing intensive irradiance (>30 µW/cm2), contributes to the incidence of severe NNJ [1,45,79,80,81]. A recent report described a canopy for providing filtered-sunlight phototherapy, which was found to be effective in reducing TSB in infants with mild-to-moderate hyperbilirubinemia . In addition, we have seen rapid growth in the availability of robust LED phototherapy, both single- and double-sided, designed specifically for use in low-resource settings.
Conclusions and Perspectives
It is clear that severe NNJ remains a life-threating condition in many areas of the world, though the true dimension of the problem is largely unknown. Severe NNJ has different etiologies, dependent on variable genetic backgrounds and geographical location, even within regions of the same country. The identification of needs and a concerted effort to improve management at different levels of the health system can significantly reduce ABE and improve opportunities for thousands of newborns around the world.
The prime importance of educational programs for families and health care personnel to achieve early identification of the signs of NNJ and seek prompt treatment should be highlighted. Moreover, we stress the need for introducing newborn screening methods for G6PD deficiency in LMICs, increasing the number of regular check-ups of ABO-incompatible and Rh-negative pregnant women and the application of specific Rh disease protection as important public health perspectives.
We recommend the introduction of inexpensive and simple-to-use devices, for measuring bilirubin and effectively treating NNJ, as essential tools required in LMICs.
Last, but not least, is the need of support from national health systems and international agencies to recognize NNJ as a priority and to invest resources to address this tragic, and largely preventable, condition.
List of Authors Who Participated in the Collection of the Data
Egypt: R. Gamaleldin, S. El Houchi and I. Seoud (Cairo University Children Hospital, Cairo, Egypt).
Indonesia: A. Tiurmaida Hutapea (Cengkareng District Hospital, Jakarta); M. Heidy Limanto and L. Rundjan, (Dr. Cipto Mangunkusumo, Jakarta); D. Iriani (Koja General Hospital, Jakarta); E. Sianipar (Pasar Rebo General District Hospital, Jakarta); O. Dyah Paramita (Tarakan General Hospital, Jakarta); W. Indri Padmosiwi Purba and M. Kristi Daradjati Saudale (WZ Johannes Hospital, Kupang); J. Rompis and R. Wilar (Prof. Dr. RD Kandou Hospital, Manado); M. Bahar Zulkifli and R. Sihombing (Budhi Asih District Hospital, Budhi Asih).
Malaysia: S.C. Chee (Department of Paediatrics, Selayang Hospital, Selangor, Malaysia).
Myanmar: A.A. Thein (Department of Neonatology, University of Medicine (1), Yangon); H.M. Nwe (Department of Paediatrics, University of Medicine (1), Yangon); D. Trevisanuto (Amici della Neonatologia Trentina, Trento, Italy; Children and Women's Health Department, Medical School University of Padua, Padua, Italy); A.A. Thin (Mandalay Children's Hospital (300), Mandalay); N. Aung and N.S.S. Aye (Central Women's Hospital, Mandalay); T. Defechereux (Department of Surgery, Liege University Hospital, Liege, Belgium); D. Kumara (Thrive Networks, Oakland, Calif., USA); L. Moccia (Thrive Networks, Oakland, Calif., USA and Amici della Neonatologia Trentina, Trento, Italy).
Nigeria: T.M. Slusher, Z.L. Farouk, B.W. Jibir, I. Muhammed; A. Kulya-Gwarzo, F. Usman, H.U. Ashiru, S.U. Abdullahi, F.I. Tsiga-Ahmed and L. Umar (Amino Kano Teaching Hospital and Murtala Mohammed Specialist Hospital, Kano); C. Isichei, F. Bode-Thomas, S. Oguche, B. Toma, V.C. Pam, C.S. Yilgwan, D.D. Shwe, Z.I. Hassan, H. Abdu, A.O.D. Ofakunrin, U.M. Diala, J.O. Abba, R. Johnson, S.N. Attah, E. Olagbaju and J. Angyu (Jos University Teaching Hospital, Jos); A. Emokpae, Z. Imam, C. Mabogunje, A. Odunsi and S. Olaifa (Massey Street Children's Hospital, Lagos); C. Ezeaka, I. Fajolu, P. Akitan and B. Ezenwa (Lagos University Teaching Hospital, Lagos); B.O. Olusanya (Centre for Healthy Start Initiative, Lagos), I. Abdulkadir, W.N. Ogala, L. Hassan, F. Abdullahi and S. Purdue (Ahmadu Bello University Teaching Hospital, Shika, Zaria); E. Omoyibo and U.O. Chima (Federal Medical Centre, Asaba, Delta State).
DOTYR-15 Contributors: M. Trip (Shoklo Malaria Research Unit, Mae Sot, Tak, Thailand); S. Riordan, J.B. Le Pichon and D. Bittel (Children's Mercy Hospital and Clinics, Kansas City, Miss., USA).
This study and the 2015 session of the Don Ostrow Trieste Yellow Retreat were made possible by an in-house grant from the Fondazione Italiana Fegato (FIF). We are grateful to Lucia Napoli and Sabrina Corsucci of the FIF for the support in organizing the event and to the AREA Science Park for providing logistic help.
The studies on NNJ performed in Nigeria and Indonesia were possible due to the generous support of the Saving Lives at Birth partners: the US Agency for International Development (USAID), the government of Norway, the Bill & Melinda Gates Foundation, Grand Challenges Canada and the UK government (FOG award No. AID-OAA-F-13-00072).
We would like to thank Bolajoko Olusanya, PhD Centre for Healthy Start Initiative, Lagos, Nigeria, and Tina M. Slusher and Tara Zamora, University of Minnesota and Hennepin County Medical Center, Minn., USA, for their help in collecting the supporting data for the global data presented. Dr. Marcelina Heidy Limanto at the Dr. Cipto Mangunkusumo Hospital, Jakarta, is acknowledged for her very efficient assistance during the Indonesian data collection.
C.G. and C.D.C.Z. contributed equally to this article; the other authors are listed in alphabetical order.