Health Status Among Adults Born With an Oral Cleft in Norway

Key Points

Question  What is the association between isolated oral clefts and the risk of adverse health outcomes in the future?

Findings  In this registry-based cohort study, little excess morbidity or mortality was found among children born with an isolated cleft lip with or without cleft palate. Children born with cleft palate only had increased mortality and increased morbidity across a range of conditions.

Meaning  Children born with isolated cleft lip defects have a low risk of adverse health outcomes, whereas children with isolated cleft palate should just be monitored more closely by caretakers.

Importance  Parents regularly express concern about long-term health outcomes for children who are born with an oral cleft.

Objective  To assess whether oral clefts affect the health and ability to work of young adults.

Design, Setting, and Participants  A population-based cohort study was conducted on all individuals born in Norway between calendar years 1967 and 1992 (n = 1 490 401). All patients treated for clefts in Norway during the study period were invited to participate (n = 2860). This study used population-based, long-term follow-up data from national registries to focus on the future health outcomes of individuals with cleft and no additional chronic medical conditions or congenital anomalies. A total of 523 individuals were excluded from the study cohort because they declined participation, could not be reached by mail, or had birth defects other than clefts. The final cohort, consisting of 2337 cases with isolated clefts and 1 413 819 unaffected individuals, was followed up until December 31, 2010, using compulsory national registries and clinical data. Data analysis was conducted from February 13, 2014, to April 18, 2016.

Exposures  Oral clefts.

Main Outcomes and Measures  Death, intellectual disability, schizophrenia, mood affective disorders, anxiety disorders, autism spectrum disorders, attention deficit/hyperactivity disorder, severe learning disability, cerebral palsy, epilepsy, muscle or skeletal disorders, trauma, and episodes of reduced health.

Results  Of 2860 individuals born with an oral cleft, 2337 were included in the analysis; of these, 1401 were male (59.9%). Mean (SD) age in 2010 was 30.6 (7.7) years. Compared with unaffected individuals, no increased risks were found regarding morbidity or mortality among persons with isolated cleft lip only. Among individuals with isolated cleft lip and cleft palate, increased risks of intellectual disability (relative risk [RR], 2.2; 95% CI, 1.2-4.1) and cerebral palsy (RR, 2.6; 95% CI, 1.1-6.2) were found. Individuals with isolated cleft palate (ie, without cleft lip) had increased mortality (hazard ratio, 3.4; 95% CI, 2.1-5.7) in addition to an increased risk of intellectual disability (RR, 11.5; 95% CI, 8.5-15.6), anxiety disorders (RR, 2.9; 95% CI, 1.3-6.5), autism spectrum disorders (RR, 6.6; 95% CI, 2.8-15.7), severe learning disabilities (RR, 10.6; 95% CI, 5.5-20.2), cerebral palsy (RR, 4.8; 95% CI, 2.3-10.0), epilepsy (RR, 4.9; 95% CI, 2.2-10.8), and muscle or skeletal disorders (RR, 2.7; 95% CI, 1.4-5.4).

Conclusions and Relevance  Young adults who were born with isolated cleft lip only did not differ significantly from unaffected individuals in their risk of health problems. However, individuals with isolated cleft palate had increased health risks and mortality. This information should be provided to genetic counselors, parents of children with clefts, and health care workers involved in the treatment or follow-up of these children.

Oral clefts are a relatively common category of birth defects, with a prevalence of about 2.1 per 1000 live births in Norway.¹ Oral clefts may be divided into 3 major subgroups: cleft lip only (CLO), cleft lip and cleft palate (CLP), and cleft palate only (CPO).

As do most other parents, parents of children with oral clefts have a well of feelings at the time of birth of their child, including a mixture of expectations, worries, and wishes for their newborn.² Even though surgery and accompanying treatment may correct the appearance and restore crucial functions, parents may feel anxious about their child’s future.^2-4 Such concerns have some support in the scientific literature.^5-7 However, most research to date has focused on children and adolescents, and our knowledge of long-term health status is more limited.⁸ An additional, widely acknowledged, methodologic challenge in cleft research is recruitment of large and representative samples from which reliable and valid conclusions can be drawn. Population-based studies comparing long-term health status among unaffected individuals and those born with a cleft may address some of these issues.

Some children with clefts have other underlying medical conditions, including genetic syndromes,^9,10 that may explain future health-related problems. Such underlying conditions may be hard to detect both early in life and later. Risks related to underlying conditions may be difficult to separate from risks directly related to the oral cleft itself. As an attempt to address this challenge, it is common to focus on a group of individuals with isolated oral cleft (ie, no other congenital anomalies). Still, evaluation of the prospects of a child with no congenital anomalies other than an oral cleft is complicated.^11,12 With these considerations in mind, it is nonetheless valuable for parents as well as clinicians to be informed about future health risks for a child born with an isolated oral cleft.

We had access to detailed information about health outcomes and background characteristics for a cohort of approximately 1.5 million individuals born in Norway between 1967 and 1992. This cohort included a group of 2860 individuals born with an oral cleft, who were followed up until 2010, when all participants were between age 18 and 43 years. A total of 2337 persons in the cleft group had no other registered congenital anomaly. The aim of the present study was to compare children who were born with an isolated oral cleft with unaffected children regarding long-term health status and survival patterns.

The 3 cleft subgroups (CLO, CLP, and CPO) were studied separately because the etiology of the cleft differs between the groups and the consequences regarding treatment and health outcomes are likely to be affected by cleft type. The total cohort consisted of all 1 490 401 live-born individuals born in Norway from calendar year 1967 to 1992. Follow-up was performed until December 31, 2010, when all participants had reached the age of majority (18 years), and the oldest participants were 43 years. All patients in the source population were invited to participate in the study, although some could not be contacted (n = 166). Information obtained from local patient registries, the Medical Birth Registry, the Education Data Base, the National Insurance Scheme, the Cause of Death Registry, and the National Registry formed the basis of our analyses. All Norwegians have a unique personal identification number, which facilitates the merging of the registries. Individuals who were not present in all the registries or had unknown survival status (usually emigrated) were excluded from the study. This exclusion also applied to individuals with any congenital malformation other than cleft, hydrocele, and hip dysplasia. Chromosomal anomalies and genetic syndromes were counted among the congenital anomalies if diagnosed within the first days of life. After exclusions, the cohort consisted of 1 416 156 individuals, of whom 2337 had isolated oral clefts; the remaining 1 413 819 individuals were in the reference group. The exclusion procedure is outlined in Figure 1.

The study was approved by the Regional Committee for Medical Research Ethics of Western Norway and required patients referred for cleft treatment (or, when too young, their parents or guardians) to provide written informed consent to participate in the study. There was no financial compensation; data were deidentified.

Data from local patient registries were available for all patients referred to cleft treatment in the study period. In Norway, treatment is coordinated by 2 centers (Oslo University Hospital and Haukeland University Hospital, Helse-Bergen, Bergen). These 2 facilities have documented medical information about all patients with oral cleft since the early 1960s. Detailed information about the cleft diagnoses was obtained from these data.

All health registries used in this study classify diagnoses according to the International Classification of Diseases.¹³ Since 1967, it has been mandatory to register all births in Norway in the Medical Birth Registry. This registration includes information on congenital anomalies, syndromes, and other medical conditions diagnosed within the first week after birth; year of birth; sex; parity; maternal age; and marital status.

The National Insurance Scheme provides health insurance for all residents of Norway, including benefits for people with medical conditions deemed sufficiently severe to cause an economic burden and disability pensions for people whose ability to work is permanently reduced by at least 50%. The same person may be granted more than 1 benefit. Diagnoses were based on medical examinations. Information about morbidity and absence due to sickness was retrieved from the National Insurance Scheme. Data from the National Insurance Scheme were available only from 1992 to 2008. A cleft diagnosis alone did not qualify for social security benefits.

The Education Database covers all Norwegians born after 1967. It is updated annually and contains 1 record for each individual every year since birth. This source provided data regarding both the educational level of each individual and the parents’ level of education when members of the study cohort were aged 16 years. For individuals who died before that age, the parents’ level of education was collected 16 years after birth. The parents’ level of education was categorized according to the parent with the highest educational level: 5 years or more at the university level (master’s or doctoral degree), 3 years at college or university (bachelor’s degree), completed high school, or did not complete high school.

The Cause of Death Registry provided information regarding survival status for all individuals in the cohort and cause and date of death for those who died during the study period. Immigration status (0, 1, or 2 parents born in Norway) was provided by the Central Population Registry. Population characteristics and important health metrics are summarized in Table 1.

The first step in the analyses was to estimate risk increase for several different medical conditions and disorders (the eTable in the Supplement provides details) for each of the 3 cleft groups (CLO, CLP, and CPO) relative to the reference group. These conditions and disorders included the most common reasons for obtaining support from the National Insurance Scheme. Poisson regression analyses with robust error variance¹⁴ were conducted as a numerically robust alternative to log-binomial regression. Adjustments were done for parents’ educational level, sex, birth year, immigration status, maternal marital status, parity, and maternal age. When fewer than 4 individuals in one of the cleft groups had a condition or disorder, we performed an exact logistic regression analysis using the cleft group and a sample of 0.1% from the reference group. Adjustments were done for the parents’ educational level, sex, birth year, and maternal marital status. This procedure of random selection and analysis was repeated 20 times, and the median odds ratios (ORs) and median 95% CIs were calculated.

Available information regarding congenital anomalies and underlying conditions at birth in the cleft group was incomplete. Undiagnosed underlying conditions, such as syndromes, may cause intellectual disability.¹⁵ Therefore, additional sensitivity analyses were performed in which we excluded all individuals who at some point in life had been granted benefits for a diagnosis of intellectual disability. This exclusion was done to reduce the chance of including individuals with nonisolated clefts in the study cohort; it does not adhere to a stringent prospective cohort design, since information obtained during follow-up was used to define the cohort.

Subsequently, Cox proportional hazards regression was performed with the same adjustment and study variables as described above and survival curves were plotted. The outcome was risk of death. These analyses were performed separately for nonmedical (eg, accident, murder, and suicide) and medical (eg, infection and stroke) causes of death. When considering medical deaths, nonmedical deaths were regarded as censoring and vice versa.

Survival curves for the risk of an episode of reduced health (ie, an individual had been granted a disability pension or had ≥8 weeks of continuous absence due to sickness) were estimated for all cleft groups. Individuals who had already experienced an episode of reduced health at the age of 18 years (24 632 individuals, with 61 in the cleft group) were excluded from the analyses. This cutoff was chosen because few Norwegians start their working careers before this age. Deaths were censored. Hazard ratios (HRs) were calculated using Cox proportional hazards regression, and the analyses were adjusted for the same variables as before. The assumption of proportionality for the Cox models was assessed by creating log − log plots.

Data preparation was conducted using Stata, version 13,¹⁶ and statistical analyses were performed in R, version 3.2.2.¹⁷ Robust error variance was obtained using sandwich and lmtest packages,^18-20 whereas Cox regressions relied on the survival package.²¹

Among 2521 live-born infants with an oral cleft, 184 were excluded because they were registered as having congenital anomalies other than cleft, hydrocele, and hip dysplasia. These exclusions were 20 individuals (2.8%) with CLO, 52 (5.4%) with CLP, and 112 (13.4%) with CPO.

Table 2 reports the relative risks (RRs) of certain medical conditions and disorders for each cleft group. The CLO group did not differ significantly in risk from the reference group for any of the disorders or conditions. The CLP group had an increased risk of intellectual disability (RR, 2.2; 95% CI, 1.2-4.1) and cerebral palsy (RR, 2.6; 95% CI, 1.1-6.2). The CPO group had an increased risk of intellectual disability (RR, 11.5; 95% CI, 8.5-15.6), anxiety disorders (RR, 2.9; 95% CI, 1.3-6.5), autism spectrum disorders (RR, 6.6; 95% CI, 2.8-15.7), severe learning disabilities (RR, 10.6; 95% CI, 5.5-20.2), cerebral palsy (RR, 4.8; 95% CI, 2.3-10.0), epilepsy (RR, 4.9; 95% CI, 2.2-10.8), and muscle or skeletal conditions (RR, 2.7; 95% CI, 1.4-5.4). In the sensitivity analyses, in which individuals who had a later diagnosis of intellectual disability were excluded, the results were mostly similar (Table 2). There were no significant risk increases in the CLO group. For the CLP group, the risk of cerebral palsy was similar in both analyses but no longer statistically significant (RR, 2.5; 95% CI, 0.9-6.6). The CPO group still had increased risks for several conditions, but the risk of epilepsy was halved and not statistically significant (RR, 2.1; 95% CI, 0.5-8.3).

Figure 2 shows survival curves from age 1 year stratified on medical vs nonmedical causes of death. Regarding individuals who died from medical causes (top panel), only the CPO group had a risk that was significantly increased (HR, 3.4; 95% CI, 2.1-5.7). For nonmedical deaths (bottom panel), no risks were elevated compared with the reference group.

Estimated curves for time to first episode of reduced health are shown in Figure 3. The isolated CPO group had an increased risk (HR, 1.2; 95% CI, 1.1-1.4). For the 2 other groups, risks were not increased.

Analyses show that children born with apparently isolated CLO or CLP had morbidity and mortality rates that were similar to those of the reference group, whereas children with isolated CPO had increased morbidity as well as increased mortality (Table 2, Figure 2, and Figure 3). Knowledge about long-term health risks for children born with oral clefts is limited. Nonetheless, professionals and parents have reasonable concerns regarding the risk of adverse outcomes.^5-7 A substantial number of these children have underlying conditions that could be expected to affect their future health. It is therefore crucial for a meaningful prediction of future health to attempt to identify and isolate cleft cases without such conditions. By doing so, we can distinguish between challenges that are purely associated with the cleft and those associated with other underlying conditions. However, such conditions may be hard to detect early in life. As an example, developmental delay is often not diagnosed until after several years of observation. We therefore applied a common strategy of removing cases with other congenital anomalies to define groups of isolated CLO, CLP, and CPO. Still, many congenital anomalies are not as apparent as an oral cleft at birth and are often not diagnosed until much later. The Medical Birth Registry is not updated with any diagnoses after the first year of life; hence, the registration of congenital anomalies is not complete. In an attempt to minimize the impact of unobserved conditions, sensitivity analyses were conducted in which all individuals who had received a diagnosis of intellectual disability later in life were excluded from the study cohort (Table 2). These sensitivity analyses did not change the main conclusion of the study. In both approaches, the isolated CLO and CLP groups were similar to the reference group, whereas the isolated CPO group had increased health problems in several domains.

The main strength of the present study was its prospective and population-based cohort design, in which information from comprehensive and compulsory registries was linked for all children born in Norway within the study period. Thanks to the unique personal identification number, all individuals were easily tracked across the registries, reducing loss to follow-up to a minimum. The addition of clinical data ensured that the same inclusion and exclusion criteria were used throughout the study period for all children with clefts. This also holds for the classification of cleft subtypes. Furthermore, treatment for clefts is centralized and free of charge in Norway, with almost complete birth cohorts as a consequence. Therefore, the included cleft cohorts can be expected to be highly representative of the Norwegian population. In addition, the availability of information about social confounders was advantageous. Such information is often difficult to obtain but may be important when studying birth defects. Because of the large sample size, most analyses were conducted with reasonable power.

One previous population-based study²² that reported health outcomes in the cleft group also adjusted for the presence of other congenital anomalies and a number of social confounders. That study found an increased hospitalization rate in the cleft group, particularly for the CLP and CPO groups. Hospitalization is, however, expected for surgical reconstructive procedures of severe clefts, and the study did not separate these expected hospitalizations from other hospitalizations. Another study²³ based on clinical data found cerebral structural differences between healthy controls and patients born with an oral cleft. The authors hypothesized that such differences could be related to an increased prevalence of cognitive and social difficulties in the cleft group.²⁴ One may speculate that pathologic embryologic processes occurring in association with some cleft types (CPO in particular) could affect cerebral development.

Despite the large sample size, 1 limitation of this study was that some analyses broke down because of small subsamples. There were, for example, no autism cases in the CLO group (Table 2). Furthermore, in the CLP group, the lower limits of the 95% CIs for the RR of intellectual disability and cerebral palsy were 1.2 and 1.1, respectively. With adjustment for multiple testing, these RRs would probably be nonsignificant. The present study cohort also had no participants older than 43 years at the end of follow-up. In Norway, death and disability benefits are expected to be low in individuals younger than 43 years. Many diseases have an expected onset when persons are in their 50s or older and were therefore not included in this study. Finally, for some diagnoses, such as schizophrenia, expensive drugs are an important part of treatment, triggering benefits from the National Insurance Scheme. Other diagnoses, such as learning disabilities, trigger benefits only when symptoms are severe enough to cause an economic burden. A possible source of bias was that access to information about morbidity was limited to the Medical Birth Registry and the National Insurance Scheme. The ascertainment of birth defects in the Medical Birth Registry was not complete,²⁵ and children’s chronic medical conditions were registered in the National Insurance Scheme only if the parents had claimed any benefits. The present study could not answer whether children with clefts were more or less likely to receive benefits for certain conditions compared with otherwise healthy children. In addition, bias may have been introduced to the analyses if the 333 individuals in the cleft group who could not be contacted or declined participation had health risk profiles other than those of individuals in the study cohort. We had no opportunity to adjust for potential confounding from maternal lifestyle factors (eg, smoking). The data set was almost complete, with less than 0.3% missing data (Figure 1). Hence, individuals with missing information were excluded from the analyses. Little bias is expected to result from this exclusion. Furthermore, some bias may be introduced by the small sample sizes in some subgroups in Table 2. However, in conducting additional analyses using exact logistic regression, we found that the results were similar for all disorders and conditions.

The present results are good news for parents of children with isolated cleft lip. The study found little evidence that a CLO was associated with increased mortality or morbidity. The present findings could also be relevant for counseling of parents who are concerned about the health of their fetus if a cleft lip is detected on ultrasonography during pregnancy. A fetal CPO is rarely visible on ultrasonography, whereas CLO and CLP are easier to detect.²⁶ The present study confirms previous findings stating that children born with isolated CPO have higher rates of mortality and morbidity than do individuals in a reference group. Thorough screening for other underlying conditions in this patient group is highly recommended from a young age to ensure necessary interventions and treatment as early as possible.

Corresponding Author: Erik Berg, MD, Department of Global Public Health and Primary Care, University of Bergen, Kalfarveien 31, Bergen, Hordaland 3950, Norway (erik.berg@uib.no).

Accepted for Publication: May 31, 2016.

Published Online: September 26, 2016. doi:10.1001/jamapediatrics.2016.1925

Author Contributions: Drs Berg and Haaland contributed equally to the creation of this manuscript. Drs Berg and Sivertsen had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Haaland, Vindenes, Moster, Lie, Sivertsen.

Acquisition, analysis, or interpretation of data: Berg, Haaland, Feragen, Filip, Moster, Lie, Sivertsen.

Drafting of the manuscript: Berg, Haaland, Feragen, Sivertsen.

Critical revision of the manuscript for important intellectual content: Haaland, Feragen, Filip, Vindenes, Moster, Lie, Sivertsen.

Statistical analysis: Berg, Haaland.

Obtained funding: Sivertsen.

Administrative, technical, or material support: Sivertsen.

Study supervision: Haaland, Filip, Vindenes, Moster, Lie, Sivertsen.

Conflict of Interest Disclosures: None reported.