An epidemiologic analysis of the association between eyelid disorders and ocular motility disorders in pediatric age

Aim of the study was to assess: (a) the prevalence and type of strabismus, ptosis and eyelid dynamic disorders features, (b) the prevalence of refractive errors, amblyopia and, (c) their association with ocular/systemic syndromes in a cohort of patients. This is a retrospective observational multicenter cohort study. Patients with coexisting ocular motility disorders, comitant and incomitant strabismus, ptosis and dynamic eyelid disorders who have never undergone surgery were enrolled throughout a 3-years a study period. 137 out of 19,089 patients were enrolled, of which 97 with uniocular and 40 with binocular disease. Isolated congenital ptosis was observed in 84 patients. A polymalformative syndrome was present in almost one third of cases, whilst among strabismus type, esotropia was slightly more prevalent. Most patients were hypermetropic. In monocular disease, myopia mainly affected older patients, who were characterized by a worse ptosis margin reflex distance and levator function, and significantly higher astigmatism. Amblyopia occurred in 67.4% of the study sub-population. Of note, in monocular disease this was mild in 25.8%, moderate in 24.2% and severe in 11.3% of cases, whilst in binocular disease it was mild in 25%, moderate in 41.7% and severe in 16.7%. All patients with coexisting eyelid and ocular motility dysfunctions in pediatric age need ophthalmologic and systemic evaluation to accurately assess amblyopia, refractive errors and systemic/ocular disorders.


Results
General characteristics of the study cohort. During the 3-year study period 137 out of 19.089 patients (0.72%) afferent at the two Ophthalmological Pediatric Centers met the inclusion criteria.
In the same observation period, overall, patients with eyelid static and dynamic disorders were 292 (1.52%) and isolated congenital ptosis were 84 (0.44%).
Ninety-seven out of 137 patients had a monocular disease, affecting either right or left eye, whereas in 40 cases the disease affected both eyes. Patients were also homogeneously distributed for sex (47.4% M vs. 52.6% F), both overall and between groups, with an overall median age of 12 years (IQR 7-16 years).
Patients with disease affecting both eyes were characterized by a significantly higher strabismus deviation angle for distance (p = 0.039) and for near (p = 0.023). As well, we observed a significantly higher prevalence of comitant strabismus deviation (p = 0.021) and ptosis familiar history (p = 0.008) in this subgroup. Levator function was instead significantly higher among patients with single-eye disease (median 17 vs 8; p < 0.001). All characteristics of the study population are described in Table 1.
Each outcome of interest was further separately compared between the two subgroups.
Patients with monocular disease showed that ptosis onset occurs significantly later in those with vertical squint (median 6 months vs.1 in exotropic and esotropic patients, p = 0.010). A suggestive association instead emerged as for what concerns strabismus onset (p = 0.083), towards a later occurrence in those with vertical squint.
For what concerns ptosis features, all patients with vertical squint had unilateral ptosis, with the congenital ones significantly more prevalent (83.9% vs. 51.4% in esotropia and 75.9% in exotropia, respectively; p = 0.010).
Findings were similar also in patients with binocular disease as for vertical strabismus (median 18 months in those with vertical squint vs. 0 in both other strabismus types, p < 0.001), whilst ptosis familiar history was significantly prevalent in those with exotropia (42.9% vs. 14.3 in esotropia and 0 in vertical squint, p = 0.019). A suggestive association instead emerged towards a younger age (median 10 years vs. 16 years in exotropia and 13 years in vertical squint; p = 0.073) and higher MRD values (3 mm vs 1 mm in the other two categories; p = 0.081) among patients with esotropia. All data are reported in Table 2.
Ptosis features. We thus analysed patients according to ptosis features, i.e. congenital ptosis or eyelid asymmetry, either unilateral or bilateral, according to the mono-or bin-ocular disease.
In patients with monocular disease, vertical strabismus emerged as significantly higher among those with unilateral congenital ptosis (p = 0.011). Moreover, we observed a significantly higher prevalence of esotropia in patients with unilateral eyelid asymmetry (60% vs. 28.4%; p = 0.010). In the same subgroup, also MRD (median 4 mm vs. 3 mm; p < 0.001) and levator function (median 18 mm vs. 16 mm; p < 0.001) were significantly higher than among those with congenital ptosis.
Looking at patients with both eyes affected, findings were almost similar, with MRD (median 4 mm vs. 1 mm; p < 0.001) and LF (median 18 mm vs. 8 mm; p < 0.001) both significantly higher in patients with bilateral eyelid asymmetry as compared to congenital. Conversely, astigmatism value was significantly lower in this subset of patients (median 0.75 vs. 2; p = 0.026). A suggestive association instead emerged towards a slightly higher refractive error among those with eyelid asymmetry (p = 0.060), and towards a higher strabismus deviation angle for near in those with congenital ptosis (p = 0.093). All data are reported in Table 3.

Refractive errors.
Refractive error was classified as myopic, hypermetropic and anisometropic.
In patients with monocular disease, patients were significantly older among those with myopia (median 14.5 years vs. 10 hypermetropic and 11 anisometropic, p = 0.005). Moreover, refractive error was significantly lower in myopic subjects, though such a result has to be confirmed due to the too small sample of anisometropic patients.
For what concerns patients with binocular disease, instead, patients were only either myopic or hypermetropic. In this subgroup of patients, strabismus onset occurred significantly later in myopic subjects (median 6 months vs. 1, p < 0.001). A comparable finding was also observed for ptosis onset. Myopic subjects were also characterized by a significantly higher prevalence of ptosis familiar history (60% vs. 6.7%, p = 0.001) and higher astigmatism value (median 3 vs. 1, p = 0.003), and a lower refractive error (median -3 vs. 2.5, p < 0.001). Conversely, both MRD (median 2 mm vs.1 mm, p = 0.003) and LF (median 14 mm vs. 3 mm, p < 0.001) were significantly higher among hypermetropic patients. All data are shown in Table 4.

Diseases categories.
For what concerns categories, we considered all categories grouped vs. the isolated one, in order to minimize the bias due to small sample sizes of the single subgroups.
Looking first at monocular disease, we observed a significantly higher prevalence of associated polymalformative syndromes in the isolated category (37.8% vs. 13%, p = 0.026). A suggestive association instead emerged Table 1. General characteristics of the study population (n = 137). M male, F female, CNS central nervous system, MRD margin reflex distance, BCVA best corrected visual acuity. *Others: congenital cranial dysinnervation disorders, Genetic, CNS disorders/neurological diseases, Ocular syndromes/orbital diseases. Significant data are in bold, suggestive p-values in italic.   Among patients with binocular disease, indeed, in the isolated category there was a significantly higher prevalence of strabismus familiar history (50% vs. 5.6%, p = 0.043), and a suggestive association towards a higher MRD (median 3 mm vs. 1 mm, p = 0.064). All data are reported in Table 5. Table 5. Association between disease categories and general and clinical characteristics in the study cohort, according to patients' subgroups (n = 137). M male, F female, MRD margin reflex distance, BCVA best corrected visual acuity. *Amblyopia was analysed on the subgroup of patients for which information was available. Significant data are in bold, suggestive p-values in italic.

Discussion
Over a 3-year study period, this multicenter study analyzed 137 out of 19.089 patients (0.72%) afferent at the two Ophthalmological Pediatric Centers. The overall prevalence of isolated congenital ptosis was 0.44% (84 out of 19,089 patients), of whom 27 (32.1%) were associated with ocular motility disorders.
The reported prevalence of strabismus associated with congenital ptosis ranges from 18 to 36% 1-4 , remarkably higher than in the general population 12 . In about one third there was a concomitant diagnosis of polymalformative syndrome. The binocular type shows a significantly higher prevalence of ptosis familiar history, comitant strabismus, higher strabismus deviation angle and lower levator function values. Table 6. Association between amblyopia and general and clinical characteristics in the study cohort according to patients' subgroups. Amblyopia was analysed on the subgroup of patients for which information was available. Qualitative variables are expressed as absolute and percentage frequency quantitative data as median and interquartile range (IQR). M male, F female, MRD margin reflex distance, BCVA best corrected visual acuity. *Others: congenital cranial dysinnervation disorders Genetic CNS disorders/neurological diseases ocular syndromes/orbital diseases. www.nature.com/scientificreports/ Regarding strabismus type (esotropia, exotropia, vertical squint), congenital unilateral ptosis was more prevalent in patients with vertical squint. Some Authors suggested that strabismus could be secondary to a disruption of binocularity by the ptotic eyelid 13 . On the other hand, as for congenital ptosis associated with vertical strabismus, a prenatal insult or maldevelopment of the third cranial nerve might play an important role 14,15 . Moreover, consistent with the current literature, refractive error value was significantly higher in patients with esotropia 16 .
As for ptosis features, we observed a significantly higher prevalence of esotropia and mild ptosis, as well as a good levator function in patients with either unilateral or bilateral eyelid asymmetry. These data are mainly related to the clinical features of Type I Duane Retraction Syndrome 17 .
With regard to refractive error features, patients with monocular disease were significantly older in the myopic subgroup, and they also had a significantly worse ptosis MRD and LF, as well as higher astigmatism values. Several authors reported more severe refractive errors in either moderate or severe ptosis. Zeng et al. reported a higher frequency of hyperopia and a lower axial length-corneal radius (AL/CR) ratio in case of severe ptosis, thus suggesting a delay in the growth of the eye 18 . Huo et al. instead deemed that long-standing congenital ptosis might produce myopia 19 . The refractive modifications were also related to a mechanic effect of the ptotic eyelid on ocular surface which reshapes the corneal surface 20,21 .
Of note, this study presents several limitations, mainly due to the retrospective nature of the study. This has surely affected inclusion of parameters which might have been interesting to be investigated, such as genes and mutations involved in these patients. However, genetic data were not available for all patients, as they are not routine examination. We hope to include such an important issue in future prospective studies on this topic. In addition, the small sample size did not allow to deeply examine each feature, which was in turn characterized by more sub-categories. The concomitant presence of ocular motility disorders and eyelid dynamic disorders establishes a rare condition, which would need the co-participation of several referral centres for a targeted study.
Finally, our data, though limited by the small sample size, confirm a mechanical and deprivation effect of moderate and severe eyelid ptosis, which may lead to refractive changes.
The monocular isolated association, as compared with the other categories, is significantly related to polymalformative syndromes (37.8% vs. 13%, p = 0.026), and more often shows an incomitant strabismus deviation.
Amblyopia occurred in the 67.4% of the assessed subpopulation. In monocular disease the 25.8% was diagnosed as mild, the 24.2% as moderate and 11.3% as severe (vs. the 25%, 41.7% and 16.7% in those with binocular disease). Amblyopia estimated prevalence in the general population ranges from the 2.5% to 5.6% 22,23 . The reported prevalence of amblyopia in patients with ptosis, and both ptosis and strabismus, is 7-8 times higher, and it is associated to a higher prevalence of refractive errors, as well as to a not alternating strabismus 6,12,24,25 .

Conclusions
To the best of our knowledge, this is the first study investigating the relation of eyelids and ocular motility disorders. Eyelid static and dynamic disorders and ocular motility disorders are often related. Ocular motility evaluation and follow-up is especially important in congenital ptosis, where the ptotic lid may in fact precipitate strabismus and amblyopia. The high incidence of polymalformative syndromes in monocular disease should alert the specialist to a possible multidisciplinary assessment. Ultimately these patients thus need a careful ophthalmologic and systemic evaluation, due the high prevalence of amblyopia, refractive errors and systemic/ ocular associated disorders.

Data availability
The datasets generated and/or analyzed during the current study are not publicly available due the common policy of our institution. However, they are available from the corresponding author on reasonable request.