Effect of Cleft Lip With or Without Palate on Pediatric General and Dental Growth in Korea: a Nationwide Population-based Cohort Study

There are very few nationwide studies discussing the height, weight, head circumference, and dental screening of children with cleft lip with or without palate (CL/P), with most reports on this subject based on a 1900s cohort. This study aimed to characterize CL/P children in the 2000s in terms of height, weight, head circumference, and dental screening. This nationwide population-based study evaluated the National Health Insurance Service-Infants and Children’s Health Screening (NHIS-INCHS), specically the height, weight, and head circumference of millions of children. Dental screening data, including the status of each tooth and comprehensive dental judgment, were also evaluated. Syndromic and nonsyndromic CL/P children had lower height, weight, and head circumference than no CL/P children until the age of 66–71 months. Children with cleft palate only or both cleft lip and palate showed similar results. Regarding dental screening, the primary teeth of CL/P children erupted later and fell out faster than no CL/P children. Dental caries was also more common in CL/P children. Children with CL/P had inferior general growth, regardless of palatoplasty surgery. More aggressive dental treatment was required for CL/P children due to the instability of primary teeth and tendency for caries. divided into 1000 percentiles for the entire cohort. To exclude technical or input errors for each numerical value, numerical values less than the 1st percentile and more than the 999th percentile were excluded. Mean values and standard deviation of the general growth data were calculated for CL/P and no CL/P children according to subgroups and the presence of associated syndromes. Analysis of variance with post hoc test (Scheffe) was used for statistical comparison. Regarding dental and oral examination, the differences regarding eruption, noneruption, erupting, demineralization, caries, restoration, pit and ssure sealing, and suspicious caries were compared. The categorical variables were analyzed using Chi-squared test. Categorical dependent variables, which included the needed treatment, prevention, plaque, malocclusion, parafunction, caries, proximal caries, restoration, risk for caries, and total judgment of the teeth, were also assessed. Logistic regression tests were performed for these categorical dependent variables. All statistical analyses were conducted using STATA MP, version 16.1 (StataCorp, College Station, Texas, USA). Statistical signicance was set at p < 0.05.


Validation for diagnostic accuracy
For diagnostic accuracy, a total of 437 children who had congenital facial anomalies and visited a single medical center from January 2006 to December 2018 were analyzed by reviewing the medical records.
Two plastic surgeons reviewed all medical records, including medical photographs and 3-dimensional facial computed tomography (CT) ndings. Cleft lips were identi ed by medical photographs and cheiloplasty (cleft lip repair). Cleft palates were diagnosed by medical photographs of the palate, 3dimensional facial CT, and palatoplasty (cleft palate repair). To validate the subgroups of CL/P, CLO patients were de ned as only those that were tagged for Q36 and not Q35. Cleft palate only (CPO) was tagged only with code Q35 and not Q36. Patients tagged with both Q35 and Q36 or Q37 were identi ed as cleft lip and palate (CLP). These algorithms were also used during the review of medical records in a single medical center.
Among all 437 children, 221 diagnosed with CL/P were analyzed for sensitivity, whereas 216 with facial congenital anomalies other than CL/P were analyzed for speci city. The sensitivity of these diagnostic algorithms was assessed by checking whether CL/P infants had satis ed the diagnostic criteria, and the speci city of these algorithms was assessed by determining why infants with congenital facial anomalies other than CL/P did not satisfy the diagnostic criteria. The results revealed that CL/P infants had 99.10% sensitivity and 99.07% speci city. The diagnostic algorithm for subgroup analysis showed 95.65% sensitivity and 97.83% speci city for CLO, 97.39% sensitivity and 99.13% speci city for CPO, and 96.67% sensitivity and 98.33% speci city for CLP.

Statistical analysis
This study compared the general growth (i.e., height, weight, and head circumference) of CL/P and no CL/P children as well as compared their teeth by individual tooth number. These comparisons were performed according to sex, subgroups, and presence of associated syndromes. For general growth, all seven screenings were included in the analysis. Height, weight, and head circumference were divided into 1000 percentiles for the entire cohort. To exclude technical or input errors for each numerical value, numerical values less than the 1st percentile and more than the 999th percentile were excluded. Mean values and standard deviation of the general growth data were calculated for CL/P and no CL/P children according to subgroups and the presence of associated syndromes. Analysis of variance with post hoc test (Scheffe) was used for statistical comparison. Regarding dental and oral examination, the differences regarding eruption, noneruption, erupting, demineralization, caries, restoration, pit and ssure sealing, and suspicious caries were compared. The categorical variables were analyzed using Chisquared test. Categorical dependent variables, which included the needed treatment, prevention, plaque, malocclusion, parafunction, caries, proximal caries, restoration, risk for caries, and total judgment of the teeth, were also assessed. Logistic regression tests were performed for these categorical dependent variables. All statistical analyses were conducted using STATA MP, version 16.1 (StataCorp, College Station, Texas, USA). Statistical signi cance was set at p < 0.05.

Results
A total of 5,234,695 newborns born during 2007-2018 were registered in the NHI database. Among them were 10,455 children with CL/P. There were 5,292 males and 5,163 females. There were 2,462 cases of syndromic CL/P and 7,993 cases of nonsyndromic CL/P. There were 2,140, 5,538, and 2,777 children further classi ed into the CLO, CPO, and CLP subgroups, respectively (Table 1).
General growth (height, weight, and head circumference) Table 2 shows the mean heights of the study population analyzed via the post hoc test. At 66-71 months, the mean heights of nonsyndromic and syndromic CL/P children were lower than controls (by 0.78 cm and 3.13 cm, respectively). In the subgroup comparisons, there were no signi cant differences between CLO and controls and between CLP and CPO children in terms of height. However, at 66-71 months, CPO and CLP children were signi cantly shorter than controls (by 1.64 cm and 1.45 cm, respectively). This trend in height difference was consistent throughout all screenings (Fig. 2). Table 3 shows the mean weights of the study population. At 66-71 months, nonsyndromic and syndromic CL/P children had lower mean weight than controls (by 0.52 kg and 1.71 kg, respectively). In the subgroup comparisons, there were no signi cant differences between CLO and controls and between CLP and CPO children. However, at 66-71 months, CPO and CLP children weighed signi cantly less than controls (by 1.06 kg and 0.81 kg, respectively). This trend in weight difference was consistent throughout all screenings (Fig. 3). Table 4 shows the mean head circumferences of the study population. At the nal screening, nonsyndromic and syndromic CL/P children had lower mean head circumference than controls (by 0.09 cm and 0.79 cm, respectively). During subgroup comparison, at 66-71 months, CLP children had lower mean circumference than controls, whereas CPO children had lower mean circumference than CLP children. CLO children and controls were not signi cantly different (Fig. 4).

Dental and oral screening
Dental and oral screenings were compared between no CL/P and CL/P children (Supplementary Table 2 -25). CL/P children had slower eruption of primary teeth than no CL/P children, except for the primary second molars (#55, #65, #75, #85). However, all the permanent teeth of CL/P children erupted faster than controls. CL/P primary teeth had more demineralization, caries, and restoration than controls, except for the upper primary incisors (#51, #52, #61, #62), wherein they had less demineralization and caries, but higher restorations than controls. Compared to controls, CL/P children had no difference in the upper central incisors (#11, #12), but increased demineralization, caries, and reduced restoration in the lower central incisors (#31, #41). There was increased caries, reduced demineralization, and reduced restoration in the lateral incisors (#12, #22), and increased caries, increased demineralization, and reduced restoration in the lower lateral incisors (#32, #42). The rst upper teeth molars (#16, #26) also had increased demineralization, increased caries, and decreased restoration. The rst lower teeth molars (#36, #46), were not different between CL/P and no CL/P children. The statistical differences of each tooth between CL/P and no CL/P children are shown in both the upper and lower teeth (Supplementary Table 26). CL/P children had signi cantly increased risks for malocclusion, parafunction, caries, risk for caries, and poor total judgment than controls at 18-29 months. At 42-53 months, CL/P children had increased risk for malocclusion and parafunction but decreased risk for needing treatment. At 54-65 months, CL/P children had increased risk for malocclusion, caries, restoration, and poor total judgment and a decreased risk for needing treatment (Table 5).

Discussion
This study compared the general growth and orodental conditions of CL/P and no CL/P children. Both syndromic and nonsyndromic CL/P children were generally shorter than controls at any age. At the nal follow-up (66-71 months), nonsyndromic and syndromic CL/P children were shorter than controls. The CL/P children never had more height or weight than no CL/P children. At 66-71 months, nonsyndromic and syndromic CL/P children weighed less than no CL/P children. Contrarily, head circumference showed a slightly different trend. At 1 year of age, both nonsyndromic and syndromic CL/P children had smaller head circumference than no CL/P children. At 2 and 3 years, syndromic CL/P children still had relatively low head circumference, but there were no signi cant differences between nonsyndromic CL/P and no CL/P children. However, at 4 years, both nonsyndromic and syndromic CL/P children had smaller head circumference than no CL/P children. At the nal screening, nonsyndromic and syndromic CL/P children had smaller head circumferences than no CL/P children by 0.09 cm and 0.79 cm, respectively.
At 66-71 months, nonsyndromic CL/P children were shorter by < 1 cm, but syndromic CL/P children were shorter by > 3 cm, indicating a big difference. Syndromic CL/P is closely correlated with chromosomal anomalies, and these patients may have developmental delay, cardiac anomalies, or a severely shortened life span. 3 Thus, these growth disorders in syndromic CL/P children may have been caused by their associated congenital anomalies and intrinsic factors. Interestingly, children with nonsyndromic CL/P also had lower height and weight, but they did not differ in intrinsic factors. Thus, the difference may be due to the environmental factors caused by CL/P. However, this raises a few questions: 1) What is the normal height? 2) Can nonsyndromic children 0.78 cm shorter than no CL/P children be considered abnormal? Since the average height did not differ by even 1 cm, it is di cult to consider nonsyndromic CL/P children as abnormal. Nevertheless, it was a clear nding that nonsyndromic CL/P children are shorter than no CL/P children. According to a textbook of plastic surgery, CL/P and no CL/P children were not different at birth, but growth retardation appeared in CL/P children upon reaching the appropriate age for undergoing palatoplasty. Moreover, after palatoplasty, growth was recovered at 4 years, and the difference with no CL/P children disappeared. CL/P and no CL/P children did not differ in growth by 6 years of age. However, in late childhood, CL/P children had a greater average decrease in height and weight. 3 On subgroup analysis, CLO and CL/P children were not different from controls, but those with CPO or CLP had lower height and weight until 66-71 months of age. In terms of head circumference, only CLO was not different from any CL/P children at 4-6 months. From 1-3 years, CPO patients had a smaller head circumference than no CL/P patients. At around 4 years, head circumferences were similar with those at 4-6 months. At 5 years, CLP children had smaller head circumference than those without CL/P, and CPO children had smaller head circumference than CLP and CLO children. CLO children had no differences in head circumference from no CL/P and CLP children. Previously, there was a study regarding the length and body weight of millions of newborns with CL/P. At birth, compared to controls, the body dimensions of infants with CLO (n = 865) were not different, whereas those with CPO and CLP had lower weight and height. However, in that study, long-term follow-up was lacking, and the subjects were infants born during 1973-1992. 6 In our study, most cases (8,095 out of 8,315 cases; 97.35%) requiring palatoplasty underwent primary palatoplasty. Conversely, our study found that growth in terms of height and weight could not catch up with that of normal children after palatoplasty surgery.
In dental and oral screenings, primary teeth of CL/P children erupted slower than those without CL/P, except for the primary second molars (#55, #65, #75, #85). However, all permanent teeth of CL/P children erupted faster than those without CL/P. The rapid eruption of the permanent teeth means that the primary teeth were quickly dislodged, suggesting greatly reduced stability in the primary teeth of CL/P children. Interestingly, even though CL/P is associated with the upper teeth, similar results were found in the lower teeth of CL/P children. Some believe that these delays in growth are because many CL/P infants are born prematurely. 17 Although we did not describe this in the results section, in our cohort, only 4% of no CL/P children and 7% of CL/P children were premature. Furthermore, if these results were obtained because there were many premature infants among those with CL/P, the permanent teeth would have erupted later than in those without CL/P. In our study, CL/P children never caught up with the growth of no CL/P children, except in head circumference. Unfortunately, the current evidence based on textbook descriptions is from a population cohort in the 1900s, and thus it is not representative of the recent growth of children. Because "catch up" is a relative concept, no matter how well CL/P children eat after palatoplasty, catch up growth can still be demonstrated if no CL/P children in the 2000s have better growth indicators than those in the 1900s. Moreover, although efforts are made to improve feeding in children with cleft palate, these may still be insu cient in early infancy, especially because nonsyndromic cleft palate infants do not differ in digestive ability compared to normal infants. The low levels of general growth in syndromic and nonsyndromic CL/P, CPO, and CLP children suggest anatomical de ciencies in oral intake. Therefore, we believe that short-term tube feeding assistance or palatal splits may be helpful for increasing oral intake during early infancy.
On dental examination, CL/P children had higher risks than CL/P children in terms of malocclusion, parafunction, caries, and risk of caries at 18-29 months. In contrast, at 42-53 months, CL/P children had a higher risk than no CL/P children in terms of malocclusion and parafunction only, but not in caries or risk of caries. Additionally, CL/P children needed less dental treatments than no CL/P children, which is likely because they had more frequent dental visits before 42-53 months of age and received dental care more quickly. There also seems to be a difference in sensitivity to dental treatment. Similarly, at 54-65 months, CL/P children had a lower risk of requiring treatment, but a higher risk of malocclusion and caries. Dental restorations that had already been performed were also more common in CL/P. This also supports the high sensitivity of CL/P children to dental treatment. In all age groups, CL/P children had higher rates of malocclusion and poor total judgment, and the odds ratio for these two variables also gradually increased with age. This suggests that CL/P children should receive dental care more actively before entering elementary school.
This study has some limitations. Since diagnoses were based using the ICD-10 code, detailed diagnostic classi cations re ecting severity (i.e., unilateral and bilateral CL/P) were not available. Medical images, photographs, radiologic ndings, or laboratory ndings were also not included in this database. Thus, bias in diagnostic classi cation may exist. However, our authors validated the diagnostic accuracy of the ICD-10 codes. The Korean NHI is a fairly accurate database as well; the insurance review teams of each general hospital in Korea verify the ICD-10 code and surgical fees before hospitals claim medical fees. Afterwards, the Health Insurance Review & Assessment, which is another national public institution separate from the NHI, performs reveri cation. Our data from the Korean NHI went through these two steps of veri cation. Therefore, the bias for misdiagnosis in the present study is low.
This study also has some strengths. First, this is the rst comparative report to investigate the nationwide general growth, dental growth, and dental examinations in Asian CL/P children, to the best of our knowledge. Second, because this study included children during 2007-2018, it re ects the recent growth of children.
Conclusion CL/P children had lower height, weight, and head circumference than no CL/P children before elementary school with or without surgery during 2007-2018. Both syndromic and nonsyndromic CL/P had similar results. These values were also lower in CPO and CLP children, whereas no difference was found between CLO and no CL/P children. All primary teeth in CL/P children, except the second molars, erupted later than in those without CL/P. Contrarily, permanent teeth in CL/P children erupted faster than in those without CL/P, which means that CL/P children lose their primary teeth more quickly. Moreover, CL/P children had a higher risk for caries and malocclusion than those without CL/P and hence need to receive more aggressive dental treatment before entering elementary school.

Declarations
Author contributions JYR, THP, and JSL designed and managed the research, wrote the rst draft of the paper and had primary responsibility for the nal content. JYR, JDY, HYC, BCC, and KYC conducted the study and were involved in providing data. All authors have read and approved the nal manuscript.

Ethical approval
This study was approved by the Institutional Review Board of Kyungpook National University Hospital (IRB No. KNUH 2020-04-049) and performed in accordance with the principles of the Declaration of Helsinki. All personal information was anonymized. Disclosure and sharing of anonymized health insurance data were guaranteed by Korean law, and there was no reason to presume that participants refuse to consent. Because all data was anonymized, the risk of the study due to the waiver of consent was extremely low. Therefore, informed consents for participants waiver were obtained from Institutional Review Board of Kyungpook National University Hospital and Deliberation Committee of National Health Insurance Sharing Service in Korea.

Competing Interests Statement
None of the authors has commercial associations or nancial disclosures that might create a con ict of interest regarding the methods or results presented in this article.

Data availability
The data that supports the ndings of this study is available from the NHI service, but restrictions apply to the availability of data, which was used with permission for the current study and therefore not publicly available. Data is however available upon reasonable request and with permission of NHI.