OBJECTIVES: Our aim was to determine and compare the incidences of sensory impairments among very preterm (VP) (<32 + 0/7 weeks), moderately preterm (MP) (32 + 0/7–33 + 6/7 weeks), late preterm (LP) (34 + 0/7–36 + 6/7 weeks), and term infants (≥37 weeks) and to establish risk factors of neurosensory disabilities.
METHODS: This national register study included all live-born infants in Finland between 1991 and 2008. Infants who died before the age of 1 year, who had any major congenital anomaly, or had missing data were excluded (n = 21 007; 2.0%). A total of 1 018 256 infants were analyzed. Incidences of hearing loss, visual disturbances or blindness, other ophthalmologic disorders, and retinopathy of prematurity were determined for gestational age (GA) groups. Risk factors of hearing loss and visual disturbances or blindness were analyzed.
RESULTS: The incidences of sensory impairments decreased with advancing GA at birth (P < .001). The most prominent factors associated with increased risks of hearing loss and visual impairment were intracranial hemorrhage and convulsions. VP (odds ratio [OR] 2.34; 95% confidence interval [CI] 1.75–3.14) and LP (OR 1.26; 95% CI 1.04–1.52) births were associated with an increased risk of hearing loss, and VP (OR 1.94; 95% CI 1.55–2.44), MP (OR 1.42; 95% CI 1.11–1.80), and LP (OR 1.31; 95% CI 1.16–1.49) births predicted an increased risk of visual impairment.
CONCLUSIONS: Incidences of sensory impairment decreased with increasing GA at birth. The most prominent risk factors predictive of sensory disabilities were intracranial hemorrhage and convulsions. VP and LP births were associated with an increased risk of hearing loss, and VP, MP, and LP births were associated with an increased risk of visual impairment.
- CI —
- confidence interval
- GA —
- gestational age
- HDR —
- hospital discharge register
- ICD —
- International Classification of Diseases
- ICD-9 —
- International Classification of Diseases, Ninth Revision
- ICD-10 —
- International Classification of Diseases, 10th Revision
- LGA —
- large for gestational age
- LP —
- late preterm
- MBR —
- medical birth register
- MP —
- moderately preterm
- OR —
- odds ratio
- ROP —
- retinopathy of prematurity
- SGA —
- small for gestational age
- SII —
- Social Insurance Institution
- VP —
- very preterm
What’s Known on This Subject:
Very preterm birth is associated with an increased risk of sensory disabilities, and the risk increases with decreasing gestational age at birth. There are few reports concerning sensory impairment among moderately and late-preterm children.
What This Study Adds:
The incidences of hearing loss and visual impairment were increased among moderately and late-preterm infants compared with term-born children. The most prominent risk factors predictive of these disabilities were convulsions during the neonatal period and intracranial hemorrhage.
An estimated 19 million children have impaired vision, and 1.4 million children of the world are blind.1 Hearing loss and blindness or low-level vision have been considered as some of the major impairments associated with preterm birth.2 Childhood visual and hearing impairments constitute a major burden to children, families, and society. Children who are blind are at increased risks of socioeconomic problems, developmental delay, more frequent hospitalization, and death compared with children with sight.3 Very preterm (VP) (<32 + 0/7 weeks) children in particular are more likely to have ophthalmic impairments compared with those born at term. Children born VP have 3 to 4 times poorer visual acuity and nearly 10 times more strabismus than term-born children.4 The results of a few studies have also revealed increased ocular morbidity in infants born moderately preterm (MP) (32 + 0/7–33 + 6/7 weeks) and late preterm (LP) (34 + 0/7–36 + 6/7 weeks).5,6
The incidence of congenital hearing impairment has been estimated to be 1 to 2 per 1000 newborns, and, in risk populations, the incidence has been reported to be increased 10- to 50-fold.7,8 VP birth, admission to a NICU, and low birth weight have been reported to increase the risk of hearing disabilities in childhood.9–11
Our aim with this study was to determine the incidences of sensory impairments in a large national birth cohort and to establish prenatal and neonatal risk factors predictive of these disabilities. To our knowledge, there are no large population-based reports that are focused on sensory impairments in MP and LP infants.
All live births in Finland between 1991 and 2008 were collected from the medical birth register (MBR) (n = 1 039 263), as described with covariate definitions in detail in our earlier publications.12–14 The MBR contains data on background factors of mothers, pregnancies, and live births with a birth weight ≥500 g or gestational age (GA) of at least 22 weeks, and the data have been validated and shown to be reliable.15,16 Infants with missing data on GA (n = 5520; 0.53%), with at least 1 major congenital anomaly (n = 13 007; 1.25%), and those who died before the age of 1 year (n = 2659; 0.26%) were excluded. Data on major structural anomalies were derived from the register of congenital malformations and were excluded from the risk factor analysis as a significant confounding factor.17 The remaining population (n = 1 108 265; 98.0% of all newborns) constituted the cohort for analysis. Children were analyzed in the following subgroups according to GA at birth: VP (<32 + 0/7 weeks; n = 6329), MP (32 + 0/7–33 + 6/7 weeks; n = 6796), LP (34 + 0/7–36 + 6/7 weeks; n = 39 928), and term (≥37 weeks; n = 965 203). The term group also included postterm infants (≥42 + 0/7 weeks; n = 47 318). GA was based on early pregnancy ultrasonography and was corrected according to whether there were marked discrepancies (5–7 days) between the initial estimation made by the last menstrual period. Three different time periods (1991–1995, 1996–2001, and 2002–2008) were compared. These periods were defined because Finland changed the classification system of diagnoses from the International Classification of Diseases, Ninth Revision (ICD-9) to the International Classification of Diseases, 10th Revision (ICD-10) in 1996, and the MBR changed the data collection forms in 1990 and 1996. The children were analyzed up to the age of 7 years or up to 2009.
Diagnoses of sensory disturbances were obtained from the hospital discharge register (HDR) and the register of the Social Insurance Institution (SII). The HDR contains data on inpatient and outpatient visits in all hospitals in Finland, and the data are considered to be reliable.18 Diagnoses were coded according to the ICD-9 between 1991 and 1995 and according to the ICD-10 between 1996 and 2009. The SII register has data on granted medicine reimbursements and disability allowances. Data from the MBR, HDR, the register of congenital malformations, and the SII register were linked (between mothers and infants) via anonymized codes by the register-keeping authorities. Hearing loss (H90-H91/389) and visual disturbances or blindness (H53-54/368-369) were considered as major sensory disabilities, and the data were obtained from registers via International Classification of Diseases (ICD) codes that have also been used in other register studies.19,20 Hearing loss included conductive and sensorineural hearing loss as well as other causes of hearing deficit according to ICD codes. Minor sensory defects were other ophthalmologic disorders including disorders of ocular muscles, binocular movement, refraction and accommodation (H49-H52/367, 378), and retinopathy of prematurity (ROP) (H35.1/362.22-27). In the Finnish health care system, these diagnoses are made in secondary or tertiary units by specialists. According to national recommendations, the otoacoustic emission was recommended to use for hearing screening of all newborns in Finland in 2004.
Infants born small for gestational age (SGA) were defined as those with a birth weight ˃2 SDs below the mean weight for GA (less than the fifth percentile), and infants born large for gestational age (LGA) were defined as those with a birth weight ˃2 SDs over the mean weight for GA (greater than the 95th percentile) according to Finnish sex-specific fetal growth curves.21 The stratification of 1-minute Apgar scores to 0 to 3 and 4 to 10 was based on ICD-code definition of severe birth asphyxia as 1-minute Apgar scores 0 to 3.
The Mann–Whitney U test, the χ2 test, independent samples Kruskal–Wallis test, or Fisher’s exact test were used in group comparisons, as appropriate, and P values <.001 were considered statistically significant in group comparisons (Tables 1 and 2). A sensitivity analysis of incidences of impairments was done for those children who survived the first month of life to take into account the greater likelihood of survival with increasing GA. Risk factors of hearing loss and visual disturbances or blindness were sought by using a generalized linear mixed model with an lmer function as regards data recorded between 1996 and 2008 (Tables 3 and 4). A binary response (disability yes versus no) was used as a dependent variable, and the results of analysis were expressed as odds ratios (ORs) with 95% confidence intervals (CIs). All included explanatory variables shown in Tables 3 and 4 were modeled as fixed variables. The deliveries of any 1 mother constituted a potential source of variation; therefore, this subject-specific effect was included as a random effect in the models.22 The generalized linear mixed model analyses were performed by using Statistical Package R, version 3.3.0 package lme4 (www.r-project.org), and the remaining analyses by using IBM SPSS Statistics version 23.0 software (IBM SPSS Statistics, IBM Corporation). Values of P < .05 (2-tailed) were considered statistically significant in mixed models.
Characteristics of the infants and mothers are shown in Table 1. The rate of VP birth was 0.62%, that of MP birth 0.67%, that of LP birth 3.92%, and the proportions of preterm births remained constant during the study. The MP and LP groups together constituted 88% of all preterm births and made up 28% of all admissions to neonatal units.
The incidences of sensory disabilities are presented in Table 2 and Fig 1. The incidence of hearing loss in the VP group was sevenfold greater, over twofold greater in the MP group, and, in the LP group, 1.5-fold greater than in the term group. Similarly, the incidences of visual disturbances or blindness and other ophthalmologic disorders decreased with advancing GA at birth. ROP was mostly presented in the VP group of infants. The incidences were mainly similar either after exclusion of infants who died before the age of 1 month or of infants who died before the age of 1 year (Supplemental Table 5).
Maternal smoking during pregnancy, being born SGA, an Apgar score <4 at 1 minute of age, admission to a neonatal unit, mechanical ventilation, intracranial hemorrhage, and convulsions during the neonatal period were associated with increased risks of both hearing loss and visual disturbances or blindness (Tables 3 and 4). Being born in other than level III hospitals and receiving antibiotic treatment during the first week of life predicted an increased risk of hearing loss. There was an association between primiparity and a decreased risk of hearing loss. Maternal age of ≥40 years was associated with an increased risk, and birth in other than a level III hospital was associated with a decreased risk of visual disturbances or blindness.
LP and VP births predicted an increased risk of hearing loss. LP, MP, and VP births were associated with an increased risk of visual disturbances or blindness after adjusting for background factors.
In this large national population-based cohort, the incidence of sensory impairments decreased with advancing GA at birth. Preterm birth was associated with an increased risk of visual disturbances or blindness, and VP and LP births predicted an increased risk of hearing loss. The most prominent risk factors associated with an increased risk of sensory impairment were intracranial hemorrhage and convulsions during the neonatal period.
Strengths and Limitations
The strength of this study is that it involves a population-based, complete national cohort with large numbers and reliable and validated health-register data. All children with neurodevelopmental disabilities are examined and diagnosed within the public health care system in Finland, which is easily accessible to all irrespective of socioeconomic status, employment, or family income. All children under school age undergo routine regular physical and developmental assessments in child health centers, and, if any developmental problems are detected, children are referred to public special health care. Diagnoses in special health care are made according to multidisciplinary evaluations and clinical investigations, as appropriate according to national clinical guidelines. Diagnoses are immediately reported to register-keeping authorities by health care providers, this being obligatory according to Finnish legislation. We used linked data from several national health registers to obtain as complete a picture as possible, and this has been shown to be a reliable method.23 Children with major congenital anomalies were excluded, because such anomalies represent a significant confounder when establishing risk factors of disabilities related to the perinatal period and prematurity.24 Most of the children were analyzed up to early school age. Sensory impairments can reliably be diagnosed by that age.
The limitations of this study are consistent with those of all observational register-based studies. Recording practices and clinical assessment tools may vary over time and between hospital districts, and this is considered to be the main limitation of our study. Because of the nature of this study, we had neither detailed information on the severity of disabilities nor on the functional situation of the children. On the other hand, it is obvious that established and recorded diagnoses constitute a burden to the children and their families. The follow-up period ended in 2009, so children born in the latest years of the study had a shorter follow-up time than the others. The follow-up time in the last time period (years 2002–2008) was probably too short for some of the children to allow making further conclusions of the trends of the incidences of impairments between 3 different time periods. A normal neonatal hearing screening result does not guarantee normal hearing later in childhood.25 Thus, there might be some children in this population with as yet undiagnosed disabilities. Gaining access to the national health data is time consuming with a complex permission process. Thereafter, data collection and linkages were performed carefully and took time also. This led to a time gap from the follow-up time of the registers to the present time and can be also considered as a general limitation of register study.
The estimated rate of hearing loss has been reported to be 2 to 3 incidences per 1000 live births in the general population and 2 to 4 incidences per 100 infants in a higher risk population.7,8,11 The overall incidence of hearing loss in our study was 0.37%, and it increased with decreasing GA at birth, being 2.46% in the VP group. These results are in accordance with those of earlier studies. In a nationwide cohort study of 2186 newborns in the Netherlands, the prevalence of hearing loss was 3.2% in infants born at <30 weeks of GA and/or birth weight <1000 g and treated in NICUs.10 In a register-based study of 11 438 infants born before 33 weeks’ gestation in Poland, hearing deficit was diagnosed in 4.2% of children born between weeks 26 and 28 and in 2.3% of children born between weeks 29 and 32.11 In the current study, the incidences were also higher in MP and LP infants compared with term-born infants.
VP and LP births predicted an increased risk of hearing impairment in our study, whereas MP birth did not, compared with the term group. The apparent lack of impairments in the MP group is an unexpected finding, although incidences of hearing loss and visual disability were higher in the MP children compared with the LP and term children. It is unlikely that the MP group really contains risks that are not consistent with the VP and LP groups. The MP group was significantly smaller than the LP and term groups, which led to a decrease of statistical power. Some confounding factors may be used to explain the nonsignificant association of MP birth and hearing loss. In contrast to our findings, authors of a Norwegian case-control register study (327 cases and 391 992 controls) found that the risk of sensorineural hearing loss increased with decreasing birth weight, but length of gestation had no independent impact on the risk.9 Also, in a French study of 1461 infants at risk for hearing impairment, birth before the 34th week of pregnancy did not show any influence on sensorineural hearing loss.8 Contrasting results may be at least partly due to different outcome definitions between studies.
Intracranial hemorrhages and convulsions were the most prominent risk factors associated with hearing loss. These factors are obviously associated with brain injury, leading to worse outcome. Low Apgar scores at 1 minute, admission to a neonatal unit, mechanical ventilation, and antibiotic treatment predicted a risk of hearing loss. It has been suggested that loud noise in the neonatal unit environment and commonly used aminoglycosides cause toxic reactions in the inner ear of prematurely born infants, leading to hearing deficiencies.26 Systemic hypoxia reduces cochlear oxygenation, causing adverse effects in cochlear function.27 Low Apgar scores and mechanical ventilation can be considered as potential markers of hypoxia, and the association of these factors with hearing loss has also been found in other studies on preterm infants.10,11 Intrauterine growth restriction has been considered to be one of the major mechanisms of sensorineural hearing loss.9 In the current study, being born SGA and maternal smoking during pregnancy were predictors of hearing loss, and these factors may constitute a common pathway disturbing normal development of the fetal central nervous system and inner ear. Hyperbilirubinemia has been shown to be a risk factor of auditory nervous system damage and hearing loss.28,29 Some studies have revealed this association, but interestingly, in our study this was not found.11 This is probably due to the fact that we were not aware of bilirubin levels of the cases because of our study design. Thus, the cases with the highest bilirubin levels could not be identified.
Cerebral visual impairment is the most important cause of childhood severe visual impairment and blindness in developed countries.3,30 Blindness has been reported as a long-term adverse outcome after extremely preterm birth (≤25 weeks of gestation), affecting 0.7% to 9% of such infants.31 In our study, preterm birth (<37 weeks) was associated with an increased risk of visual disturbances and blindness, compared with term birth. In a cross-sectional study of 182 children and adolescents (≤16 years of age) in New Zealand (143 blind and 39 children and adolescents with low vision), the main causes of blindness were asphyxia (25%), nonaccidental injury (7%), and prematurity (7%).30 An association between smoking and an increased risk of visual disability was found in our study, as elsewhere.32 The mother being ˃40 years of age predicted visual impairment, but antibiotic treatment had no such association. According to national guidelines, all high-risk pregnancies and preterm deliveries before 32 weeks’ GA are centralized to level III hospitals in Finland. This may explain the association with a decreased risk of visual disturbances and blindness and birth in other than level III hospitals. Otherwise, risk factors were similar to those associated with hearing loss, intracranial hemorrhage, and convulsions being the most prominent risk factors of both disabilities.
The incidences of disorders of the ocular muscles, binocular movement, accommodation, and refraction decreased with increasing GA at birth. The incidence of these disabilities was 2.4 times higher in the MP group than in the term group. According to the results of a Swedish study of 78 children born between GAs of 32 and 36 weeks between 2002 and 2004, there was a significant difference in ocular motility deficits and refractive errors compared with term infants in the control group, and birth weight was found to be the strongest predictor of ophthalmologic abnormalities. In accordance with our results, they found that moderate-to-late preterm birth was associated with a 2.4-fold increased prevalence of refractive errors compared with children born full term.6 Globally, ROP is considered to be a major threat to vision in preterm infants, and in our study, it was a problem of VP infants. It is possible that VP infants first diagnosed with ROP may later be diagnosed with blindness or other visual problems.
Preterm birth, including MP and LP births, increases the risk of adverse sensory outcomes. MP and LP groups constitute the majority of all prematurely born children, and impairments among these groups constitute a major burden to society and the health care system. The most prevalent risk factors associated with sensory disabilities were factors indicating brain damage arising in the perinatal period. The deleterious effects of smoking on the fetus should be mentioned in the counseling of pregnant women early in pregnancy. In addition, MP and LP infants with known risk factors of sensory impairment should have a low threshold for referral to further evaluation, to receive diagnosis and treatment as early as possible.
- Accepted May 18, 2018.
- Address correspondence to Mikko Hirvonen, MD, Department of Pediatrics, Central Finland Health Care District, Central Finland Central Hospital, Keskussairaalantie 19, FI-40620 Jyväskylä, Finland. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Funding for full-time research work (Dr Hirvonen) was received from Central Finland Health Care District and the Competitive State Research Financing of the Expert Responsibility area of Tampere University Hospital (grant 9R051) and from the Arvo and Lea Ylppö Foundation (grant 201610031). The funding sources had no role in the study.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- World Health Organization
- Solebo AL,
- Teoh L,
- Rahi J
- Cooke RW,
- Foulder-Hughes L,
- Newsham D,
- Clarke D
- Meyer C,
- Witte J,
- Hildmann A, et al
- European Surveillance of Congenital Anomalies
- Korvenranta E,
- Lehtonen L,
- Peltola M, et al
- Fitzmaurice G,
- Laird N,
- Ware J
- Walden RV,
- Taylor SC,
- Hansen NI, et al; National Institute of Child Health and Human Development Neonatal Research Network
- Adhikari S,
- Shrestha MK,
- Adhikari K,
- Maharjan N,
- Shrestha UD
- Copyright © 2018 by the American Academy of Pediatrics