Age-specific incidences are shown in 1-year intervals for congenital esotropia, exotropia, partially accommodative esotropia, and fully accommodative esotropia. The number of cases (including imputed cases) and person-years included in the calculations for each age group are shown below the plot.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Torp-Pedersen T, Boyd HA, Skotte L, et al. Strabismus Incidence in a Danish Population-Based Cohort of Children. JAMA Ophthalmol. 2017;135(10):1047–1053. doi:https://doi.org/10.1001/jamaophthalmol.2017.3158
What are the cumulative incidences and age-specific incidence profiles of the major strabismus subtypes in Danish children?
In a Danish population-based cohort of 96 842 children, 1309 cases of strabismus were identified, giving a cumulative incidence of 2.56% at 7 years. The 4 major strabismus subtypes in this population (partially accommodative esotropia, congenital esotropia, exotropia, and fully accommodative esotropia) had different age-specific patterns of incidence.
These data suggest that strabismus is a common, heterogeneous disorder featuring subtypes with different ages at onset in Denmark.
To our knowledge, there have been few population-based studies of strabismus incidence conducted. Our population-based study provides valuable data for health services planning and identifying research needs.
To determine the incidence and age distribution of strabismus, overall and by subtype, among children 7 years or younger.
Design, Setting, and Participants
This population-based cohort study was conducted with data from 96 842 children enrolled in the Danish National Birth Cohort.
Main Outcomes and Measures
Age-specific incidence and cumulative incidence and median age at the detection of strabismus, overall and by subtype.
The study cohort included 96 842 children born between 1996 and 2008 who are predominantly Caucasian and is composed of approximately 30% of births in Denmark, with a boy-girl ratio of 51:49. Overall, 1309 cases of strabismus were identified in the cohort. We found an overall cumulative strabismus incidence of 2.56% (95% CI, 2.42-2.69) at 7 years. The overall incidence was similar among boys and girls. Two hundred sixteen participants (16.5%) (95% CI, 14.5-18.6) had congenital esotropia, 177 (13.5%) (95% CI, 11.7-15.5) had fully accommodative esotropia, 252 (19.3%) (95% CI, 17.1-21.5) had partially accommodative esotropia, and 181 (13.8%) (95% CI, 12.0-15.8) had exotropia. The esotropia:exotropia ratio was 5.4:1 (95% CI, 3.4:1 to 7.5:1). Age-specific incidence curves for congenital esotropia, fully accommodative esotropia, partially accommodative esotropia, and all exotropia revealed interactions between strabismus subtype and age, suggesting that the different subtypes had different age-specific patterns of incidence (P < .001 for all comparisons between pairs of curves). The median age at detection for the 4 subtypes was 0, 32.0, 26.1, and 16.6 months, respectively.
Conclusions and Relevance
In a national, population-based cohort study, we found a cumulative incidence of strabismus consistent with those reported in smaller European and American cohorts, but a somewhat higher esotropia:exotropia ratio than those that, to our knowledge, are typically reported by English and American studies. Patterns of incidence by age differed for different strabismus subtypes, indicating differences in age at onset and thereby implying differences in the underlying etiology.
Strabismus has a reported overall prevalence of 2% to 6% among children in Western countries.1-6 The consequences of strabismus can include psychosocial effects and reduced visual acuity that is caused by amblyopia. Strabismus may also be associated with a loss of stereoacuity and subsequent poor hand-eye coordination7 that limit both professional and recreational opportunities.
Most previous studies were cross-sectional1,2,4-6 and yielded information on strabismus prevalence. Another study that followed up a birth cohort prospectively did not report ages at disease detection or age-specific incidences.3 Investigators affiliated with a population-based cohort of North American children have reported cumulative incidences of esotropia, exotropia, and hypertropia among children up to age 18 years,8-10 as well as subtype-specific incidences and median ages at onset,11,12 but these data were limited by the small size of the population studied.
Denmark’s national health registers contain almost 40 years of high-quality data on contacts with the health care system and resulting disease diagnoses. Because health care in Denmark is available to all citizens and free of cost, the health registers cover the entire population, and because reimbursement for medical services is linked to the reporting of data, data are regularly updated. By linking Danish register data with data from a well-defined, population-based birth cohort (the Danish National Birth Cohort), we were able to study strabismus incidence in a very large (n = 96 842), unselected population of children. Our objective was to determine the cumulative incidence of strabismus up to 7 years, overall and by sex and strabismus subtype.
Quiz Ref IDOur study cohort included 96 842 children in the Danish National Birth Cohort (DNBC) born between 1996 and 2003 (30% of all deliveries in Denmark in that period).13,14 Pregnant women received written material about the DNBC from their general practitioner and subsequently mailed a signed consent form to the DNBC’s offices, at which point they were enrolled in the cohort. The consent covered using the child’s medical record information for research purposes in DNBC-based studies, including ours. Approval for this study was obtained from the Danish National Ethics Committee.
Using the personal identification number assigned to all Danish residents, we linked information from the National Discharge Register and the Health Security System to the children’s DNBC information, allowing us to identify all children in this cohort who were evaluated for strabismus between their birth date and March 2006. The National Discharge Register, to which the reporting of data on hospital admissions and outpatient visits is mandatory, includes dates of hospitalization/outpatient visits, discharge diagnosis codes, and codes for any surgical procedures.15 The Health Security System registers reimbursements to specialists in private practices for services rendered and contains information on procedures and examinations performed (but not on end diagnoses). We searched the National Discharge Register for strabismus diagnosis codes (International Statistical Classification of Diseases and Related Health Problems, Tenth Revision [ICD-10] codes H49-H51) and strabismus surgery codes (Nordic Medico-Statistical Committee Classification of Surgical Procedures codes KCEA-KCEW) to identify children with a strabismus diagnosis that was confirmed by a hospital and children who had undergone a surgical procedure to treat strabismus. We searched the Health Security System to identify children who had been evaluated for strabismus (procedure code 19.2001) by a private ophthalmologist.
The National Discharge Register yielded 447 children who were seen for strabismus in hospital ophthalmology and pediatric departments. The search of the Health Security System found 8674 children who were evaluated for strabismus by 186 private ophthalmologists. We requested the ophthalmologic records for all 8783 children identified by these searches (the total number of children identified was less than the sum of the children identified in the 2 searches because some children were seen both by an ophthalmologist and at a hospital). We obtained medical records for 5652 children (64%). Only 1 of the 3131 missing records was a hospital record. Our failure to collect the remaining 3130 records from 66 private ophthalmologists was mainly because the ophthalmologist refused to participate (n = 31, 47%), we received no response (n = 23, 35%), or because the ophthalmologist retired (n = 10, 15%).
Two orthoptists evaluated the ophthalmologic records, determined whether strabismus had been diagnosed, and recorded the specific strabismus subtype (if it was noted), the date of diagnosis, the date of first suspicion, and any developmental or neurological diseases present that were relevant to strabismus. If there were discrepancies between the recorded strabismus subtype and recorded orthoptic measurements, the diagnosis was assigned based on the measurements. The orthoptists used a classification scheme with 28 strabismus diagnoses (Table). Congenital esotropia (ET) and congenital exotropia (XT) were defined as constant nonaccommodative ET or XT that was present by 6 months of age, of more than 8 prism diopters (PDs), and was not reduced by 10 or more PDs with hyperopic refractive correction. Fully accommodative ET was defined as ET that could be reduced by 10 or more PDs to fewer than 10 PD with hyperopic refractive correction, while partially accommodative ET was defined as ET that could be reduced by 10 or more PDs with hyperopic correction, with a residual deviation of 10 or more PDs. Microesotropia was defined as constant ET of 8 or fewer PDs (without hyperopic refractive correction and/or a fewer than 10-PD reduction with hyperopic correction). Acquired nonaccommodative ET was defined as a constant ET of more than 8 PDs that developed after 6 months of age and could not be reduced by 10 or more PDs with hyperopic correction. Sensory ET and XT were defined as ET and XT with a reduced visual acuity because of other organic eye diseases. Paralytic ET and hypertropia were defined as esotropia and hypertropia, respectively, with limited ductions, in which there was no other evidence of restrictions. Intermittent XT was defined as an XT of 10 or more PDs at a distance that was not constant and was categorized into 3 subtypes: convergence insufficiency (a ≥10-PD greater deviation for near viewing than distance viewing), divergence excess (a ≥10-PD greater deviation with distance viewing than with near viewing), and basic (similar deviations for both near and distance viewing). Esotropia, XT, and hypertropia that were associated with central nervous system (CNS) diseases, based on information in the ophthalmologic records, were assigned to their own categories. We did not classify pure phorias as strabismus. We classified strabismus only according to the deviation present before the patient underwent any surgical procedure; we did not report consecutive, residual, or recurrent strabismus following surgical procedures. Age at disease detection, used in all analyses, was defined as the age that strabismus was first reported by parents or a health professional or, if such data were not reported in the medical record, the age at which the patient received a diagnosis.
The estimation of incidence was performed using a survival analysis to take into account that not all children were observed up to their seventh birthday. Thus, for each child, the study period extended from birth until the first of the following events: the receipt of any strabismus diagnosis, the patient’s seventh birthday, emigration, death, or December 31, 2005.
Because not all ophthalmologic records were available for evaluation, we used multiple imputations to estimate the number of cases expected among the children whose records could not be retrieved. Each child was randomly assigned a strabismus status (no strabismus or a specific strabismus subdiagnosis) based on the sex- and birth year–specific probabilities of having strabismus observed among the children whose ophthalmologic records were evaluated. Similarly, children designated as strabismus cases were randomly assigned a year of detection based on the sex- and birth year–specific distributions observed among children whose records were evaluated. All estimates were calculated as the mean of estimates from 100 independent imputations. Standard errors were calculated from variance estimates from each imputation, correcting for variances among the imputations.16 The number of cases and years of follow-up were calculated using the mean of the imputations.
Age-specific incidences of strabismus were estimated as the number of cases divided by the follow-up time within each 1-year age stratum. To summarize the incidence of strabismus in the first 7 years, we calculated the cumulative incidence at 7 years, defined as the sum of the age-specific incidences for ages 0 to 6 years. Confidence intervals for cumulative incidences were calculated using the approach described by Waltoft.17 Because strabismus is not a very common disorder, the cumulative incidence can be interpreted as a risk. Incidences among boys and girls were compared using a log-linear Poisson regression adjusting for age (1-year groups). Age-specific incidences for selected pairs of strabismus subtypes were compared using a competing risk approach and evaluating the multiplicative interaction between age (1-year groups) and subtypes in a log-linear Poisson regression model.18 We used SAS, version 9.1.3 (SAS Institute) for data analysis.
The 96 842 DNBC children included in our study cohort contributed 492 810 person-years of follow-up time, with a median follow-up time of 5.2 years per child. During this period, 8783 children (9%) were evaluated for strabismus. Medical record review for 5652 children (6%) with available ophthalmological records yielded 1309 strabismus cases; the Table shows the number of medical record–confirmed cases for each strabismus subtype. In addition, 705 cases were imputed among the 3131 children (3%) whose records were not available for review, yielding 2014 strabismus cases (medical record-confirmed and imputed).
Quiz Ref IDOverall, the cumulative incidence of strabismus at 7 years was 256 (95% CI, 242-269) cases per 10 000 children. The estimated cumulative incidence at 7 years and median age at disease detection for each strabismus subtype are shown in the Table. Cumulative incidences were 201 cases (95% CI, 188-213) and 37 cases (95% CI, 31-43) per 10 000 children for ET and XT, respectively, yielding an ET:XT ratio of 5.4:1 (95% CI, 3.4:1 to 7.5:1). The overall incidence of strabismus did not differ by sex (P = .35). Incidence differed by sex for only 3 strabismus subtypes (accommodative ET, 31 [95% CI, 24-38] per 10 000 boys vs 42 [95% CI, 34-50] per 10 000 girls, P = .03; microesotropia, 16 [95% CI, 9.5-22] per 10 000 boys vs 8.1 [95% CI, 3.6-13] per 10 000 girls, P = .02; and intermittent ET, 14 [95% CI, 9.7-18] per 10 000 boys vs 20 [95% CI, 15-26] per 10 000 girls, P = .02). However, these differences were only nominal; there were no differences in sex-specific incidence when multiple comparisons were considered.
The Figure presents age-specific incidence up to 7 years for congenital ET, fully accommodative ET, partially accommodative ET, and overall XT (the 4 strabismus subtypes with 150 or more cases). Pairwise comparisons of these curves revealed interactions between strabismus subtype and age (congenital ET vs fully accommodative ET, P = 2.0 × 10-34; congenital ET vs partially accommodative ET, P = 1.7 × 10-36; congenital ET vs XT, P = 1.1 × 10-15; fully accommodative ET vs partially accommodative ET, P=1.4 × 10-4; fully accommodative ET vs XT, P = 1.9 × 10-7; partially accommodative ET vs XT, P=2.6 × 10-5), implying that the major strabismus subtypes have different ages at onset. The median age at disease detection for the 4 subtypes varied from 0.0 months for congenital ET to 32.0 months for fully accommodative ET (Table).
When we included only children born after singleton pregnancies in the analyses, we obtained results similar to those including all children (data not shown). When comparing age-specific incidences of strabismus subtypes, excluding congenital XT (n=11) from the group with XT yielded similar results to those including all XT (data not shown).
In a population-based cohort of 96 842 children, we found an overall cumulative strabismus incidence at 7 years of 256 cases per 10 000 children, which is consistent with incidences reported by previous smaller studies1-6 in European and American cohorts. Our cumulative incidence of ET (201 per 10 000 children at age 7 years) was comparable with the cumulative incidence of 2.13% (n = 350) reported in children 18 years or younger in a population-based cohort in Minnesota.8 However, our cumulative incidence of XT (37 per 10 000 children at age 7 years) was only approximately half of the 0.69% (n = 114) found in the same Minnesota population.9 Thus, in our study, ET occurred more than 5 times more frequently than XT, which is a somewhat higher ET:XT ratio than found in recent studies in the United States and England, which reported ratios of 1.8:1 to 3.4:1.3,6,12
Intermittent exotropia accounted for 133 cases (73%) of XT at 7 years in our study but for only 106 XT cases (52%) in the Minnesota cohort.9 Accommodative ET (fully and partially accommodative combined) was the most frequent strabismus subtype in both the Minnesota cohort and our study (27% [n = 175] and 32% [n = 429] of all strabismus, respectively).11 Other frequent subtypes from the Minnesota cohort were intermittent XT (17% [n = 106] of all strabismus) and acquired nonaccommodative ET (10% [n = 64]), compared with only 133 (11%) and 66 (5%), respectively, of all strabismus cases in our study. Conversely, congenital ET accounted for 216 all strabismus cases (15%) in our study, compared with only 66 (5%) in the Minnesota cohort.
Esotropia and XT that were associated with CNS diseases occurred more frequently in the Minnesota cohort (44 [6.7%] and 30 [4.9%] of all strabismus, respectively) than in our study (16 [1.0%] and 7 [0.5%] of all strabismus, respectively).11 Two studies featuring consecutive patients from Tennessee with ET19 and XT,20 respectively, also found less congenital ET and more strabismus related to CNS diseases than our study. The discrepancies in congenital ET and CNS disease-associated strabismus incidences in our study and the American studies may be partially explained by the different criteria for ET that was associated with CNS diseases. In our study, the orthoptists reviewing ophthalmologic records for children with suspected strabismus evaluated whether strabismus was associated with CNS diseases or was simply an incidental finding in a child who also had a CNS disease. In contrast, the American studies defined strabismus associated with CNS disease as strabismus in children with any neurologic impairment other than isolated speech delay. In our study, 42 children with esotropia and co-occurring developmental or neurological disorders were not classified as having ET that was associated with a CNS disease, but were assigned to other subgroups of ET according to the specific ET characteristics they had experienced. However, even if these 42 children had been classified as having CNS-associated strabismus, the proportion of strabismus associated with CNS diseases reported in our study would still be less than half of that found in the American studies.
Quiz Ref IDWe found evidence of interaction between the 4 most common strabismus subtypes (congenital esotropia, fully accommodative esotropia, partially accommodative esotropia, and exotropia overall) and age. The median age at disease detection differed considerably for these 4 subtypes (0.0, 32.0, 26.1, and 16.6 months, respectively). These findings provide evidence that these strabismus subtypes have differing ages at onset, which argues favorably for differences in the underlying etiology, with different neural pathways being affected at different stages in postnatal development.
Although all general practitioners (approximately 3500) in Denmark were invited to recruit pregnant women for the DNBC, not all chose to participate. Composing approximately 30% of births during the period, 101 042 pregnant women chose to participate in the study. A woman’s decision to participate in the cohort may have depended on her behavior before and during pregnancy, which was potentially also associated with her offspring’s risk of developing strabismus. Nevertheless, a 2006 study concluded that participating women were only marginally healthier than the background population,14 indicating that health-related risk factors were similarly distributed among the general Danish population and DNBC participants.
Quiz Ref IDWe were unable to retrieve ophthalmologic records for 3131 children (36%) identified as potentially having strabismus. Because the study was conducted in a population-based cohort of children who were identified at birth, estimates of incidence based only on medical record-confirmed cases (ie, only for children with strabismus for whom ophthalmologic records could be retrieved) would have underestimated the true strabismus incidence in the cohort. We could have tried to correct for this bias by removing from the cohort a corresponding proportion of the children who were never examined for strabismus. However, this could have had an unpredictable effect on our age-specific incidence estimates. Instead, we chose a more sophisticated statistical approach in which we imputed cases among children with missing records based on the knowledge we did have (the distribution of birth year and sex) and our knowledge of these distributions among medical record-confirmed cases. It is unlikely that this imputation resulted in a biased estimate, because we have no reason to suspect that a private ophthalmologist´s decision to participate in the study was associated with the rate or distribution of strabismus in that practice or that retired ophthalmologists diagnosed strabismus at a different rate from practicing ophthalmologists.
Ethnically, the Danish population is highly homogenous: in 2003 (when the last DNBC children were born), 90.9% of children younger than 7 years born in Denmark were ethnically Danish/Scandinavian.21 The DNBC was likely even more heavily Danish, because women had to be able to read and write Danish to participate. This homogeneity may account for some of the differences between our findings and findings from nonwhite populations or more ethnically diverse countries, such as the United States. To the extent that strabismus incidence varies with ethnicity, our results may not be generalizable to non-European populations.
Our study has multiple strengths, particularly the large size and population-based nature of our cohort. The structure of Denmark’s health care system ensured that all of the children in the study population had equal access to medical care, which minimized threshold for parents who were seeking an ophthalmologic evaluation for their children and maximized the detection of strabismus early in life. The Danish health registers, with their nationwide coverage and high-quality data, ensured that identifying children who were evaluated for strabismus did not depend on recall, and the final strabismus classifications were made by experienced orthoptists who were reviewing ophthalmologic records.
In a national population-based cohort study, we found a cumulative overall incidence of strabismus of 256 cases per 10 000 children at 7 years, which is consistent with the findings of other studies conducted among European and North American populations. The observed distribution of strabismus subtypes differed somewhat from the distributions reported elsewhere; in particular, we found a larger ET:XT ratio than previously observed in other populations, a lower incidence of strabismus associated with CNS diseases, and a higher incidence of congenital ET. Age-specific strabismus incidences did not differ for boys and girls. We found evidence of interactions between strabismus subtype and age for congenital ET, fully accommodative ET, partially accommodative ET, and XT overall, indicating different patterns of age at onset for the different subtypes, which in turn implied etiologic differences among subtypes. Potential study limitations concern the degree to which DNBC children were representative of Danish children as a whole and our inability to retrieve ophthalmologic records for one-third of the children who were evaluated for strabismus. However, the former is unlikely to have influenced our results substantially, and we addressed the latter issue using statistical imputation techniques. Because Denmark is ethnically very homogeneous, our results may not be generalizable to nonwhite populations.
Corresponding Author: Tobias Torp-Pedersen, MD, PhD, Department of Ophthalmology, Rigshospitalet Glostrup, Nordre Ringvej 57, DK-2600 Glostrup, Denmark (email@example.com).
Accepted for Publication: July 10, 2017.
Published Online: August 31, 2017. doi:10.1001/jamaophthalmol.2017.3158
Author Contributions: Drs Torp-Pedersen and Skotte had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Torp-Pedersen, Boyd, Haargaard, Wohlfarht, Holmes, Melbye.
Acquisition, analysis, or interpretation of data: Torp-Pedersen, Boyd, Skotte, Haargaard, Wohlfarht, Holmes.
Drafting of the manuscript: Torp-Pedersen, Skotte.Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Skotte, Wohlfarht.
Obtained funding: Torp-Pedersen, Boyd, Melbye.
Administrative, technical, or material support: Torp-Pedersen, Holmes, Melbye.
Supervision: Boyd, Haargaard, Wohlfarht, Holmes, Melbye.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Funding/Support: This study was supported by grant 271-05-0406 from the Danish Medical Research Council and grants from the Danish Eye Association, the Velux Foundation, the Margrethe and Johs. F. la Cour Foundation, the Dagmar Marshall Foundation, and Research to Prevent Blindness (an unrestricted grant to the Department of Ophthalmology, Mayo Clinic).
Role of the Funder/Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.
Create a personal account or sign in to: