Hypoxic ischaemic encephalopathy (HIE) is a heterogeneous syndrome characterised by neurologic dysfunction in an infant born at or beyond 35 weeks of gestation, with failure to initiate and maintain respiration. The syndrome is manifested by a reduced level of consciousness or seizures and is frequently associated with reduced reflexes and tone (Coetzee 2018). It is one of the commonest causes of neonatal mortality and severe neurological deficit in children and contributes to ever-increasing medico-legal litigation (Bhorat et al. 2023). In low- and middle-income countries, neonatal encephalopathy accounts for approximately 1 million deaths annually (Pauliah et al. 2013). Globally, 287 000 infants with HIE were estimated to have died in 2010 (Lee et al. 2013). The incidence of HIE differs significantly, with the highest burden prevailing in sub-Saharan Africa (14.9 per 1000 live births) and substantially reduced rates (1.5 per 1000 live births) in high-income countries (Ballot et al. 2020). Charlotte Maxeke Johannesburg Academic Hospital, a tertiary referral centre in South Africa, had HIE rates of 2.3–13.3 per 1000 live births in 2016 (Ballot et al. 2020). In 2011, Chris Hani Baragwanath Academic Hospital (CHBAH) reported the incidence to be 8.7–15.2 per 1000 live births (Bruckmann & Velaphi 2015). A study in the Southern Cape Peninsula reported the incidence to vary from 2.3 to 4.3 per 1000 live births (Horn et al. 2013). Nakwa et al. (2023) reported an incidence of 8.8 per 1000 live births with a mortality of 29%. More than 400 000 infants each year develop neurodevelopmental impairment (NDI) after HIE. In 2010, about 233 000 infants survived with moderate or severe NDI (Lee et al. 2013).

The rate of mortality and the severity of NDI correspond to the severity of HIE, using the Thompson score (Thompson et al. 1997) and Sarnat scoring systems (Sarnat & Sarnat 1976). Neurodevelopmental impairment includes cerebral palsy (CP), cognitive impairment, visual and hearing impairment, epilepsy, attention deficit hyperactivity disorder, autistic spectrum disorders and subtle deficits affecting learning later in childhood (Azzopardi et al. 2014). Infants with mild HIE have been revealed to have few motor or intellectual deficits at preschool age, while those with severe HIE are more heterogeneous in terms of outcomes, having higher mortality and rates of NDI (Van Handel et al. 2007). The risk of severe impairment in at least one domain is high after intrauterine and neonatal insults (Mwaniki et al. 2012).

Therapeutic hypothermia (TH) remains one of the most effective neuroprotective interventions available for HIE (Coetzee 2018). Management with TH has been demonstrated to improve survival and neurodevelopmental outcome at 18 months of age in infants with stage 2 and stage 3 encephalopathy (Mathew, Kaur & Dsouza 2022). This neuroprotective effect has been seen to persist at 6–7 years of age (Coetzee 2018).

Very few studies have examined the developmental profiles of HIE infants. The majority of the studies look at the HIE complications but not at the individual domains affected. Stark, Van der Vyver and Gretschel (2020) managed 30 HIE children from 3 months to 5 years at a secondary hospital in South Africa. Gross motor ability exceeded both cognitive and fine motor skills at 5 years of age. These findings differ with respect to research findings from high-income countries, where fine motor ability supersedes gross motor ability (Mc Guiness et al. 2012). It has been postulated that in less-advantaged communities, children engage in more outdoor play as access to formal schooling and technology is less (Stark et al. 2020). In this study, children with mild HIE developed appropriately, while CP was associated with severe HIE. Those infants with HIE and no major motor disability had a heightened risk of long-term verbal, motor and intellectual deficits at 5 years of age (Stark et al. 2020). In this research, we aimed to describe the range of NDI and the developmental profiles at 1 year of age in infants with stage 2 and 3 HIE, who were treated with and without TH.

A total of 355 infant files were reviewed. One hundred and sixteen were excluded (details in Figure 1). Two hundred and thirty-nine infants were included; 211 (88.3%) were stage 2, and 28 (11.7%) were stage 3. Therapeutic hypothermia was received by 166 (69.5%) infants, 152 were stage 2 and 14 were stage 3. Therefore, 72% of stage 2 and 50% of stage 3 infants received TH (Figure 1).

The mean gestational age and birth weight at delivery were 38.8 (± 1.7) weeks and 3116 g (± 503.4), respectively. One hundred and thirty-nine infants (58.2%) were male, and more than half of the infants 136 (56.9%) were delivered vaginally. Most infants (86.2%) were born to human immunodeficiency virus (HIV)-negative mothers. There were significantly lower appearance, pulse, grimace, activity, and respiration (APGAR)-scores and higher Thompson scores (TS) in the stage 3 HIE group (Table 1). A greater number of stage 2 HIE patients received TH (p = 0.001).

Neurodevelopmental impairment was observed in 17.1% (41) of the whole cohort, of which the majority (63.4%) had severe NDI (GQ < 70), while 17.1% had moderate NDI (GQ 70–79). NDI occurred in 12.7% (27) of the stage 2 (22.2% mild, 22.2% moderate and 55.6% severe NDI) and in 50% (14) of the stage 3 HIE group (14.3% mild, 7.1% moderate and 78.6% severe) as shown in Figure 2. The severity of NDI was not significantly different between the TH and non-TH groups (p = 0.25).

Just under half (n = 20; 48.8%) of those with NDI had not received TH (Table 2). Sixteen stage 2 NDI infants and five stage 3 received TH. Infants with stage 3 had higher rates of NDI than those with stage 2 (14[50%] vs 27 [12.8%]; p < 0.001). Infants with stage 3 who received TH had fewer NDI (5 [35.7%] vs 9 [64.3%]; p = 0.005). There were no significant differences in GQ and Griffith subscales between Sarnat stages 2 and 3 in those with NDI, as shown in Table 2. NDI and CP were less frequent in the group that received TH compared to the one that did not (NDI: 21 [12.1%] vs 20 [27.4%]; p = 0.005 and CP: 9 [5.4%] vs 13 [17.8%]; p = 0.002).

Twenty-two (9.2%) of the infants had CP: 16 (72.7%) Sarnat stage 2 and 6 (27.3%) stage 3. Eight (50%) with stage 2 received TH, and only one (16.7%) infant with stage 3 received TH. More than a third (36.3%) of infants with CP had seizures, six (27.3%) were blind, and three (18.7%) had a hearing impairment. Out of the 9 CP patients that received TH, two (22.2%) were blind, one (11.1%) had hearing impairment, and four (44.4%) had epilepsy.

There was a significantly higher percentage of infants with visual impairment, epilepsy and CP in the stage 3 group versus stage 2 HIE as shown in Table 3. More than a fifth (21.4%) of stage 3 HIE infants were diagnosed with CP; 21.4% had visual impairment, and a quarter (25%) had epilepsy. Stage 2 infants had fewer complications, with 7.5% diagnosed with CP, 0.9% with visual impairment and 6.6% with epilepsy.

Overall median GQ scores were within normal range for Sarnat stage 2 and 3 infants. However, stage 3 GQ scores (89.5 [69–105]) were significantly lower than stage 2 (GQ 103 [95–109], p = 0.002). Stage 2 infants outperformed stage 3 infants in all domains. This difference was significant in all subscales, except in the language domain (subscale C) in Table 4. Therapeutic hypothermia resulted in significantly higher GQ (p = 0.002) and subscale scores in Sarnat stages 2 and 3 (data not shown). This effect was also clearly shown in the stage 3 group. Those that did not receive TH had median GQ and subscale scores in the severe NDI range (GQ 69 [43–100] vs 101 [83–107]; p = 0.02) as shown in Table 5.

This study revealed that infants with stage 2 and stage 3 HIE followed up at CHBAH NNDC over a 6-year period had NDI in 17.1% and CP in 9.2% at 1 year of age. Neurodevelopmental impairment was present in half of the stage 3 and in 12% of stage 2 HIE infants. CP was diagnosed in 21.4% of stage 3 infants and in 7.6% of stage 2 infants. Epilepsy, visual impairment, lower APGAR scores and higher TS were also more frequent in the stage 3 group. This finding is consistent with a vast amount of research globally, which demonstrates that the incidence of long-term disability is related to the severity of HIE.

A Cochrane review in 2007 reported that in infants who survive severe HIE, as many as 80% of them develop significant complications, about 10% – 20% have moderate disabilities, and about 10% of them are healthy. In infants who survive moderate HIE, 30% – 50% of them possibly have significant lasting consequences, and 10% – 20% have slight neurologic disabilities (Evans, Levene & Tsakmakis 2007).

Reported rates of CP after HIE differ but are typically around 10% – 13% among survivors of moderate-to-severe encephalopathy in high-income countries, with dyskinetic CP and spastic quadriplegia being the most common subtypes. Rates of hearing loss are reported to be as high as 17.1% in those with other persistent neurological deficits. Up to 41% of infants diagnosed with HIE have some form of visual impairment in the first year of life (Ahearne et al. 2016).

Mbatha et al. (2021) reported a 32% impairment in 2-year-old infants who received TH at CHBAH, and 6% showed an impairment at 1 year of age. Studies have reported NDI rates between 34.3% and 36% in infants with stage 2 and 3 HIE (Carli, Reiger & Evans 2004; Kachhwaha et al. 2023). In Kathmandu, 45% of the neonatal encephalopathy patients died in the neonatal period, 20% had NDI in survivors at 1 year, and a physical disability rate of 25% was predicted in those with moderate neonatal encephalopathy (Ellis et al. 1999). In Bangladesh, 21.6% of children were found to have normal development and mild disability, all were either stage 1 or stage 2 HIE. Moderate and severe disability were observed in 33% and 25% of the children, respectively (Thayyil et al. 2021). In this study, more infants with stage 3 had NDI than those with stage 2. This finding is similar to that of a report in 2013 at CHBAH, by Sukha, in which infants with stage 3 (100%) had substantially more developmental impairment (p = 0.01) than those with stage 2 HIE (52%). More infants with stage 3 had both gross and fine motor delays (Sukka 2013).

The results of this research support the use of TH as a neuroprotective strategy in both stage 2 and stage 3 HIE. Therapeutic hypothermia was associated with fewer NDI (p = 0.005) and CP (p = 0.002) in these infants. General quotients and SQ median scores were significantly higher in those who received TH and particularly in the stage 3 group.

Many other studies have reported on the benefit of TH in reducing NDI. A systematic review reported that TH decreases developmental impairment and CP in infants with stage 2 and stage 3 (Mathew et al. 2022). A randomised clinical trial by Laptook et al. (2017) showed a 76% probability of any decline in death or disability in the TH compared to the non-TH group. Azzopardi et al. (2014) found that infants with HIE who received TH had markedly reduced risk of CP (21% vs 37%; p = 0.03) and that the frequency of moderate-to-severe disability was lower in the TH group (22% vs 37%; p = 0.03). Edwards et al. (2010) found that TH improved survival with normal neurological function and decreased the rates of severe neurological disability in survivors. Kachhwaha et al. (2023) found that 85.7% of the infants in the TH group were neurodevelopmentally normal. In the non-TH group, 15 (42.8%) were normal, and 20 (57.2%) had some neurodevelopmental abnormality. Gano et al. (2014) reported that the proportion of patients with NDI at 12 months was significantly lower (p < 0.05) in the TH group (9.43%) than in the non-TH group (36%). Ellis et al. (1999) found that the majority with NDI had severe NDI, and 78% of those children had CP. In this study, infants who received TH were less likely to develop NDI and CP. The rates of CP were lower as the moribund infants were excluded from receiving TH, and the more severe infants died in the neonatal period.

General quotients and sub-quotients were in the normal range for the group as a whole at 1 year. All group comparisons revealed differences in all subscales, except for subscale C. Language testing at 12 months may be less sensitive than later testing, as the first year milestones mostly relate to prelinguistic skills and communicative intent. The complexity of language, both vocabulary and receptive language, increases rapidly during the second year. Vocabulary development presents a lexical spurt at around 17 months (Serrat-Sellabona et al. 2021). Language delays, especially subtle deficits, are more likely to be detected later when standardised testing is performed. Items recorded on the GMDS at 12 months include understanding 1 to 2 single words, singing, and observing the response to name and inhibitory words, and some rely on the maternal report. Mbatha et al. reported higher rates of NDI on GMDS testing at 18–24 months than at 12 months. Therefore, standardised testing at 2 years and beyond provides a deeper understanding of subtle deficits. Language is therefore unlikely to be a sensitive indicator of neurodevelopmental delay at 1 year of age using the GMDS. Specific standardised language tools should be considered for testing at 12 months in future studies.

A third of stage 2 and stage 3 infants who received TH in the same institution had moderate-to-severe NDI at 18–24 months. A greater number survived with mild or no impairment at 2 years (Mbatha et al. 2021). These findings highlight the significance of continuous developmental assessment beyond the first year of life. Normal scores on the GMDS at 1 year or 2 years do not always predict a good outcome at 6 years and 7 years. An abnormal score is most likely to be associated with substandard academic performance at school age (Barnett et al. 2004). Early neurodevelopmental assessment in the first 2 years of life does not rule out neurological, cognitive or perceptual motor abnormalities at school-going age. The sensitivity values for the GMDS were marginally higher at 2 years than at 1 year (90% vs 76%) (Barnett et al. 2004). An understanding of the outcome of this category of children is essential because these subtle deficits can affect academic achievement and cause behavioural problems in the future. Failure to identify such children in infancy results in a missed opportunity for early intervention, which may help them cope better at home and at school (Barnett et al. 2004).

The strength of this study is that it investigated data which was collected over a 6-year period at the largest hospital in sub-Saharan Africa. The sample was large, and standardised GMDS testing was routinely performed in all patients, resulting in high-quality data. The limitations of this study were that this was a retrospective descriptive study and, therefore, could not evaluate causation, but only associations. The severe HIE group had far smaller numbers than the moderate group had; 50% had NDI, and 21.4% had CP. These results, although encouraging when compared to other studies, which report rates of NDI around 80% for severe HIE, should be interpreted cautiously. These infants represent a small sample in that the most severe HIE had died. They also had lower rates of TH in this group as compared to those in the moderate HIE sample. The infant GMDS at 12 months may not detect subtle language deficits and therefore, analysis at later ages or alternative standardised tools should be considered.

Neurodevelopmental impairment and CP were observed in 17.1% (41) and 9.2% (22) of the cohort, which compares favourably with the rates in other countries. The GQ and SQ scores were in the average range for the cohort with equivalent profiles across the five domains. Therapeutic hypothermia and moderate HIE were associated with decreased CP and NDI. Neurodevelopmental assessments at 2 years and beyond are necessary to determine longer-term outcomes and subtle deficits.