Dry Eye Parameters and Lid Geometry in Adults Born Extremely, Very, and Moderately Preterm with and without ROP: Results from the Gutenberg Prematurity Eye Study

Background/Aims: This study aimed to analyze the effects of perinatal history on tear film properties and lid geometry in adults born preterm. Methods: The Gutenberg Prematurity Eye Study (GPES) is a German prospective examination of adults born preterm and term aged 18 to 52 years with Keratograph® 5M and Schirmer test I. Main outcome measures were first non-invasive tear film break-up time (F-NITBUT), bulbar redness (BR), Schirmer test, and nasal palpebral angle measurement. The associations with gestational age (GA), birth weight (BW), and BW percentile, retinopathy of prematurity (ROP), ROP treatment, and other perinatal factors were evaluated using regression analyses. Results: 489 eyes of 255 preterm and 277 eyes of 139 full-term individuals (aged 28.6 +/− 8.8 years, 220 females) were included. Of these, 33 participants (56 eyes) had a history of spontaneously regressed ROP and 9 participants (16 eyes) had a history of ROP treatment. After adjustment for age and sex, lower F-NITBUT (<20 s) was associated with ROP treatment (OR = 4.42; p = 0.025). Lower GA correlated with increased bulbar redness (B = −0.02; p = 0.011) and increased length of wetting in the Schirmer test (B = −0.69; p = 0.003). Furthermore, low GA was associated with narrowing of the nasal palpebral angle (B = 0.22; p = 0.011) adjusted for age and sex, but not when considering ROP in the multivariable model. Conclusion: Our analyses indicate that perinatal history affects ocular surface properties, tear production and lid geometry in adults born term and preterm. This might indicate that affected persons have a predisposition to diseases of the corneal surface such as the dry eye disease.


Introduction
Preterm delivery leads to an abrupt change in the surrounding fetal environment and may result in ocular changes, especially in those with immediate exposure to the environment such as the ocular surface. Indeed, children born preterm have altered ocular morphology, including a steeper corneal geometry [1], a smaller anterior chamber depth [2], thicker lens [2], shorter axial length [1,2], and altered posterior pole [3][4][5]. Similar results were reported in adults indicating that low birth weight (<2500 g) as a surrogate − What is already known on this topic? Prematurity is associated with altered ocular morphology and functioning in childhood and adulthood, so we investigated whether perinatal factors have long-term effects on the ocular surface and lid configuration in adulthood. − What does this study add? ROP treatment is linked to reduced tear film break-up time later in life. Furthermore, low gestational age is associated with increased bulbar redness, longer Schirmer strip measurement, and a narrower nasal lid angle. − How might this study affect research or practice? Perinatal history affects the ocular surface, tear production, and lid geometry in adults born term and preterm, which might predispose affected persons to diseases of the corneal surface and dry eye disease in later life.

Study Population
The Gutenberg Prematurity Eye Study (GPES) is a single-center cohort study at the University Medical Center of the Johannes Gutenberg University Mainz in Germany (UMCM) that recruits individuals that (i) have been born preterm or at term between 1969 and 2002 and (ii) were between 18 and 52 years of age at study enrolment. According to these design elements, the study is a retrospective cohort study with a prospective acquisition of follow-up data. For the GPES, every preterm newborn with a GA ≤ 32 weeks and every second randomly chosen preterm newborn with GA 33-36 weeks was contacted and invited to participate. From each month from 1969 to 2002, 6 (3 males and 3 females) randomly selected full-term subjects with a birth weight between the 10th and 90th percentile were also invited to serve as controls as reported earlier [20][21][22][23][24].
The study examinations were performed between 2019 and 2021. The flow chart for eligibility and recruitment efficacy proportion is shown in Figure S1. Every participant underwent a detailed ophthalmological examination including measurement with a Keratograph ® 5M (Oculus, Wetzlar, Germany) and a medical history interview. Furthermore, their medical records documenting the perinatal and postnatal history were assessed.
Written informed consent was obtained from all participants before their entry into the study and the GPES complies with Good Clinical Practice (GCP), Good Epidemiological Practice (GEP), and the ethical principles of the Declaration of Helsinki. The study protocol and documents were approved by the local ethics committee of the Medical Chamber of Rhineland-Palatinate, Germany (reference no. 2019-14161; original vote: 29 May 2019, latest update: 2 April 2020).

Assessment of Pre-, Peri-and Postnatal Medical History
The medical histories were assessed from medical records stored at the UMCM. Data were collected regarding GA (weeks), birth weight (kg), presence of ROP, stage of ROP, ROP treatment, placental insufficiency, preeclampsia, maternal smoking during pregnancy and breastfeeding. Birth weight percentiles were also calculated according to Voigt et al. [25].

Categorization
For descriptive analysis, participants were allocated to group 1: full-term participants (GA ≥ 37 weeks), group 2: preterm participants with GA between 33-36 weeks without ROP, group 3: preterm participants with GA between 29-32 weeks without ROP, group 4: preterm participants with GA ≤ 28 weeks without ROP, group 5: preterm participants with GA ≤ 32 weeks with postnatal ROP without ROP treatment and group 6: preterm participants with GA ≤ 32 weeks with postnatal ROP and ROP treatment as reported earlier [21,22]. In the case that only one eye of a participant had ROP, the other non-ROP eye was excluded from the analysis.

Ophthalmological Examination
Each participant was examined with a Keratograph ® 5M (Oculus, Wetzlar, Germany) to measure the non-invasive tear film break-up time (first: F-NITBUT) and grade of breakup time. Furthermore, the software of the Keratograph ® 5M measures the nasal and temporal conjunctival areas, detects conjunctival blood vessels and calculates redness grade. Values for the mean global bulbar redness consist of values from temporal and nasal bulbar redness. The limbal nasal and limbal temporal redness values are subcategories of the nasal and temporal bulbar redness, respectively. Lid geometry was determined by analyzing photographs of the anterior segment and measuring the palpebral fissure and the nasal palpebral angle. Each parameter was controlled for outliers.
Tear production was measured by the Schirmer test. In every participant, the distance of the tears to travel along the length of a paper test strip was measured [1] without anesthesia of the ocular surface (Schirmer test I). Furthermore, the study participants completed the ocular surface and disease questionnaire, and the Ocular Surface Disease Index (OSDI) was calculated [26].

Inclusion/Exclusion Criteria
Only participants with successful measurement of the tear film break-up time were included. Each measurement was checked for correctness and centration and excluded if considered invalid. Participants with a history of corneal or cataract surgery were excluded as this may have contributed to an altered ocular surface.

Statistical Analysis
The main outcome measures were non-invasive tear film break-up time, bulbar redness, length of wetting in the Schirmer test I, and nasal palpebral angle. Descriptive statistics were computed for these measures stratified by clinical group. Absolute and relative frequencies were calculated for dichotomous parameters; the mean and standard deviation were calculated for approximately normally distributed variables, otherwise median and interquartile range. Categorical data were compared with the chi-square test, with not-normally distributed continuous parameters compared with the Mann-Whitney U-test. Linear regression models in the case of normal distribution with general estimating equations (GEE) were used to assess associations and account for correlations between corresponding eyes. For non-normally distributed parameters, quantile regression was performed including only right eyes (Schirmer I test). Binary logistic regression analyses were conducted for reduced F-NITBUT (<20 s). First, univariate analyses of the main outcome measures and sex (female), age (years), GA (weeks), birth weight (kg), birth weight percentile, ROP (yes), ROP treatment (yes), placental insufficiency (yes), preeclampsia (yes), breastfeeding (yes) and maternal smoking during pregnancy (yes) were computed. Then, only parameters associated in the univariate analyses were included in model #1 with additional adjustment for sex and age. In a further model, the potential effect of ROP occurrence (yes) and/or ROP treatment (yes) were analyzed. Birth weight was excluded in the multivariable models to avoid collinearity which was strong between GA and birth weight. This was an explorative study, so there was no adjustment for multiple testing. Calculations were performed using commercial software (IBM SPSS 20.0; SPSS, Inc., Chicago, IL, USA; R version 4.0.0 (24 April 2020), R Core Team (2020), R: A language and environment for statistical computing and R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/; package "quantreg" (accessed 10 January 2022).

Participant Characteristics
The present analysis included 489 eyes of 255 preterm and 277 eyes of 139 full-term individuals (age 28.6 +/− 8.8 years, 220 females). Overall, there were 277 eyes of 139 participants with GA ≥ 37 weeks (group 1), 256 eyes of 129 participants with a GA between 33-36 weeks without ROP (group 2), 133 eyes of 70 participants with a GA between 29-32 weeks without ROP (group 3), 28 eyes of 14 participants with a GA ≤ 28 weeks without ROP (group 4), 56 eyes of 33 participants with a GA between 24-32 weeks with ROP without treatment (group 5), and 16 eyes of 9 participants with a GA between 24-32 and with postnatal treatment for ROP (group 6). Of the ROP-treated group, five (nine eyes) participants underwent laser coagulation, while four (seven eyes) participants had cryocoagulation. The recruitment efficacy proportion for each group is presented in Supplementary Figure S1. Overall, 7 participants were excluded because of previous corneal refractive surgery, cataract surgery, or ocular trauma, and further 40 participants were excluded because measurement with a keratograph was invalid or not possible. In total, eight eyes without ROP were excluded in which the fellow eye had postnatal ROP. The participants' characteristics are presented in Table 1.

Descriptive Anterior Surface Parameters
Participants treated for ROP showed a lower F-NITBUT (p = 0.032; Table 2) ( Figure 1). Bulbar redness was increased in the preterm group with a GA 33-36 weeks (p = 0.034; group 2) and the group with untreated ROP participants (p = 0.024; group 5) and in the ROP treated participants (p = 0.048; group 6) compared to the full-term control group (group 1) ( Table 2). Furthermore, the measurement in the Schirmer test was significantly increased in the preterm groups with a GA 33-36 weeks (p = 0.001; group 2), GA 29-32 weeks without ROP (p = 0.001; group 3), and GA ≤ 28 weeks without ROP (p < 0.001; group 4), and descriptively increased in participants with GA ≤ 32 weeks with ROP without treatment (p = 0.092; group 5) and with ROP treatment (p = 0.054; group 6) compared to the full-term control group (group 1) (Table 2; Figure 2). Parameters for palpebral fissure and angle are displayed in Table 3 and Figure 3.

Descriptive Anterior Surface Parameters
Participants treated for ROP showed a lower F-NITBUT (p = 0.032; Table 2) ( Figure  1). Bulbar redness was increased in the preterm group with a GA 33-36 weeks (p = 0.034; group 2) and the group with untreated ROP participants (p = 0.024; group 5) and in the ROP treated participants (p = 0.048; group 6) compared to the full-term control group (group 1) ( Table 2). Furthermore, the measurement in the Schirmer test was significantly increased in the preterm groups with a GA 33-36 weeks (p = 0.001; group 2), GA 29-32 weeks without ROP (p = 0.001; group 3), and GA ≤ 28 weeks without ROP (p < 0.001; group 4), and descriptively increased in participants with GA ≤ 32 weeks with ROP without treatment (p = 0.092; group 5) and with ROP treatment (p = 0.054; group 6) compared to the full-term control group (group 1) (Table 2; Figure 2). Parameters for palpebral fissure and angle are displayed in Table 3 and Figure 3.

Uni-and Multivariable Analyses
In the association analyses, first univariable analyses were performed. Associated parameters were then included in model #1 without the inclusion of ROP occurrence and ROP treatment. In a second multivariable model, ROP occurrence and treatment were additionally included if they were associated in the univariable model. Models #1 and #2 were additionally adjusted for sex and age. First tear film break-up time (<20 s) correlated in univariable analyses with ROP occurrence (p = 0.024) and ROP treatment (p = 0.009). After adjustment for sex and age and the additional inclusion of ROP occurrence and treatment, ROP treatment (OR = 4.42 [95% CI: 1.20; 16.28]; p = 0.025) but not ROP occurrence was associated with F-NITBUT < 20 s. In the univariable analyses, Bulbar redness was associated with GA (p = 0.004), birth weight (p = 0.003), and ROP occurrence (p = 0.05). After adjustment for sex and age in model #1, GA (p < 0.001) was still associated with bulbar redness. In model #2, a low GA revealed an association (B = −0.015 [95% CI: −0.027; −0.003]; p = 0.011) with bulbar redness but not ROP occurrence.

Discussion
The present study provides new data describing the long-term effects of prematurity, ROP, and perinatal factors on ocular surface health in adults born preterm. Our study shows that prematurity correlated with an increase in bulbar redness, increased length of wetting in the Schirmer test, and a narrower nasal palpebral angle. Furthermore, ROP treatment was associated with a tear film break-up time lower than 20 s. Overall, this data indicates that adults born preterm may have altered ocular surface homeostasis.
This study provided new insights regarding the long-term effects of perinatal history on ocular surface characteristics. It is well known that preterm birth and the postnatal occurrence of ROP are linked to an altered organ morphology in childhood, adolescence, and adulthood. The Wiesbaden Prematurity Study observed that prematurity was linked with a steeper corneal curvature, while postnatal ROP occurrence and treatment were linked to a shorter anterior chamber depth [1]. In congruence, a population-based report analyzing adolescents showed an association between lower birth weight as a proxy for prematurity and steeper corneal curvature [27]. Recent reports indicate that these alterations persist throughout life, as results from the Gutenberg Health Study showed that low birth weight (<2500 g) was associated with a steeper corneal radius, smaller corneal diameter [6], and altered posterior pole [28][29][30][31]. Furthermore, other studies reported a less regular corneal surface in subjects born preterm with low birth weight, as indicated by increased corneal aberrations in childhood and adulthood [9,10]. The present results indicate that prematurity affects not only ocular geometry but also other ocular surface properties such as tear film quantity and stability.
While Raffa and colleagues [32] observed no association between the eyelid aperture in individuals born preterm and term aged 4 to 15 years, the present study found a narrower nasal palpebral angle in individuals born preterm. One may speculate that this is an unknown risk factor potentially influencing ocular refractive and geometric development in individuals born preterm, as previous reports observed that palpebral angles and lid aperture are risk factors for spherical refractive error development [33,34].
Individuals affected by dry eye diseases suffer from ocular discomfort, burning, instable tear film, and visual disturbance that can lead to inflammatory damage of the ocular surface [17]. They may also have reduced quality of life [35] and lower vision-related quality of life [36] with a prevalence up to one in three in the general population [37] with a high economic impact [38]. There are different recognized risk factors, such as environmental factors, endogenous stress, antigens, infections, genetic factors, and autoimmune disorders [16], which only partially explain the occurrence of dry eye disease, so other risk factors may be involved. Our data are the first to suggest that prematurity and associated factors may predispose to dry eye disease later in life. Dry eye disease can be classified as decreased tear secretion or increased tear evaporation. Since the preterm participants revealed significantly increased Schirmer test measures, one may speculate that preterm birth leads to tear overproduction accompanied by a less stable tear film. However, the pathophysiological mechanisms remain unclear, and it remains to be determined whether the increased tear evaporation is caused by dysfunction of the Meibomian glands or by reduced mucin from the goblet cells. Other pathophysiological mechanisms causing long-term effects on the corneal surface might be the lower extrauterine temperature after preterm birth compared to the intrauterine environment [12] and a shorter time in the intrauterine milieu, or different periods of opened eyes in the interval until reaching full GA [13]. There is evidence that dry eye disease is caused, among other reasons, by subacute inflammation and has many features in common with autoimmune disease [16]. Prematurity is linked to a hyper-responsive innate immune system [15]; thus, one may speculate that this pro-inflammatory tendency also influences the ocular surface. In addition, participants being treated for postnatal ROP revealed a shorter tear film break-up time also potentially indicating that absorbed energy during treatment leads to life-long alterations of corneal hemostasis. It is well known that prematurity is a risk factor for psychosomatic diseases, such as depression, which is also a risk factor for dry eye disease [39]. However, the clinical significance of our findings remains unclear because we observed no difference in the OSDI score between the different study groups. It is possible that the observed ocular changes are of subclinical importance and might predispose to dry eye diseases in later life.

Strengths and Limitations
One limitation of the present study is the single-center hospital-based study design. Furthermore, other lid parameters were not investigated, such as frequency of blinking in different environments, horizontal lid aperture, and temporal lid aperture. Invalid or missing keratograph examinations may further limit the present data. Moreover, the number of individuals with a history of ROP and ROP treatment was low, which must be considered during the data interpretation. A further limitation is the missing data about tissue status and function of the meibomian gland in the present study which should be examined in future studies.
The strengths are that this study was one of the largest cohorts of adults born preterm at different GA with and without ROP. Detailed assessment of perinatal history from medical records enabled a comprehensive data analysis of the perinatal effects on the ocular surface, with all measurements and validation steps blinded with respect to GA and other perinatal factors.

Conclusions
In conclusion, our analyses indicate that perinatal history affects the ocular surface later in life. Preterm birth was associated with increased bulbar redness, increased length of wetting in the Schirmer test, and a narrower palpebral angle, while ROP treatment was associated with a reduced tear film break-up time of less than 20 s, suggesting that affected persons may be predisposed to diseases of the corneal surface in later life, such as dry eye disease.