A Longitudinal Study to Establish the Normative Value and to Evaluate Perinatal Factors Affecting in Preterm Infants

Pak Cheung Ng,1 Barbara Sau Man Tam,2 Cheuk Hon Lee,1 Samuel Po Shing Wong,3 Hugh Simon Lam,1 Alvin Kwan Ho Kwok,2 and Tai Fai Fok1

PURPOSE. To establish a normative range of intraocular pressure vulnerable, high-risk infants after birth, and intraocular pres- (IOP) in preterm infants and to identify important perinatal sure (IOP) can be easily measured by a handheld instrument factors that could affect the IOP during the early weeks of during routine ophthalmic examination. An understanding of neonatal life. the natural maturation process of the , establish- METHODS. The IOP of 104 preterm infants, with a median ment of a normative range of values, and identification of (interquartile range) gestational age of 29.8 (28.7–30.9) weeks important intrinsic or extrinsic factors that may affect various and birth weight of 1208 (1049–1370) g, were assessed in a ocular measurements, including IOP, are essential in assisting clinicians in assessing developmental abnormalities and con- university-affiliated tertiary neonatal center. These infants had 4 IOP measured by a handheld tonometer at 1, 4, 6, 8, and 10 genital diseases involving the eyes. To date, relatively few weeks of postnatal age. The mixed-effects models were used to studies have reported IOP in preterm infants. Of these studies, evaluate the longitudinal IOP measurements and to identify most obtained the IOP measurements in a cross-sectional man- ner.4–10 In only one small trial involving 20 preterm infants critical perinatal factors that would significantly affect the 11 ocular pressure. was IOP measured weekly during the first month of life. Although investigators performing earlier studies have been RESULTS. A percentile chart of IOP in preterm infants was meticulous in their examination techniques and in the mea- constructed, and the median (10th–90th percentile) IOP surement of ocular pressure, relatively little clinical informa- ranged from 16.9 (12.3–21.5) to 14.6 (10.1–19.2) mm Hg at tion concerning the characteristics and outcomes of patients 26.1 and 46.4 weeks of postconceptional age, respectively. were described.4–12 Important information included the mode The IOP was significantly and negatively associated with Ͻ ϭ of mechanical ventilation, presence or absence of intracranial postconceptional age (P 0.001), mean blood pressure (P complications, and interventions or drugs that may influence 0.01), Apgar score at 1 minute (P ϭ 0.04), and use of inhaled ϭ the IOP. However, the interrelationship and complexity of corticosteroids (P 0.03), but was positively correlated with different factors affecting the IOP renders normative data dif- the commencement of high-frequency oscillatory ventilation ϭ ficult to determine. Thus, the objectives of this study were (1) (P 0.01). to determine a normative range of IOP by longitudinally mea- CONCLUSIONS. A quantitative statistical model has been devel- suring the ocular pressure at standardized time intervals (1, 4, oped and a percentile chart of IOP constructed for preterm 6, 8, and 10 weeks) after birth and (2) to determine important infants that could be used for future reference. Pediatric oph- perinatal factors, both physiologic and environmental, that thalmologists and neonatal clinicians can compare the IOP of may affect the IOP during the early weeks of neonatal life. A preterm infants against this chart and make relevant quantita- percentile chart of IOP in preterm infants will provide neonatal tive adjustments for critical perinatal factors so that the IOP clinicians and pediatric ophthalmologists with the ability to may be properly evaluated, both in healthy and ill infants. compare ocular pressures between well and sick infants and to (Invest Ophthalmol Vis Sci. 2008;49:87–92) DOI:10.1167/ monitor infants who receive specific medications, such as iovs.07-0954 systemic corticosteroids, or undergo eye surgery and/or laser photocoagulation treatment. reterm, very-low-birth weight (VLBW; 1500 g) infants are Pfrequently exposed to high concentrations of oxygen and are at risk of development of of prematurity METHODS (ROP).1–3 Regular screening for ROP is necessary for these Patients Preterm infants admitted consecutively into the neonatal intensive care unit (NICU), at the Prince of Wales Hospital, Hong Kong, who met the From the Departments of 1Pediatrics, 2Ophthalmology and Visual Science, and 3Statistics, Prince of Wales Hospital, The Chinese Univer- screening criteria for ROP were prospectively enrolled over a period of Ͻ sity of Hong Kong, Hong Kong. 32 months. The inclusion criteria were: (1) all infants 32 weeks of Supported by the Departmental Fund from the Department of gestational age or birth weight Ͻ1500 g, (2) moderately preterm Pediatrics, The Chinese University of Hong Kong. infants (32–36 weeks gestational age) receiving supplemental oxygen Submitted for publication July 27, 2007; revised August 28, 2007; Ͼ7 days, and (3) parental consent to participate in the study. Infants accepted November 12, 2007. were excluded if they had undetermined gestational age; multiorgan Disclosure: P.C. Ng, None; B.S.M. Tam, None; C.H. Lee, None; dysfunction and were expected to die imminently; chromosomal ab- S.P.S. Wong, None; H.S. Lam, None; A.K.H. Kwok, None; T.F. Fok, normalities and dysmorphic syndromes; or congenital ocular abnor- None malities, such as corneal clouding or . Patients who received The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertise- postnatal systemic (i.e., oral or intravenous) corticosteroids treatment ment” in accordance with 18 U.S.C. §1734 solely to indicate this fact. were investigated separately, as previous trials in older children and Corresponding author: Pak Cheung Ng, Department of Pediatrics, adults have found that the use of systemic corticosteroids as well as Level 6, Clinical Sciences Building, Prince of Wales Hospital, Shatin, topical steroids applied locally to the eye are important factors associ- N.T., Hong Kong; [email protected]. ated with the development of ocular hypertension.13–16 Further, we

Investigative & Visual Science, January 2008, Vol. 49, No. 1 Copyright © Association for Research in Vision and Ophthalmology 87

Downloaded from iovs.arvojournals.org on 10/01/2021 88 Ng et al. IOVS, January 2008, Vol. 49, No. 1

monitored the latter group of infants differently, with IOP checked TABLE 1. Clinical Characteristics of the Study Population according to the dose regimen of the dexamethasone course received ؍ by the patient. Characteristics n 104 Gestational age (wk) 29.8 (28.7–30.9) Methods Birth weight (g) 1208 (1049–1370) Gestational and Postconceptional Ages. The gestational Length (cm) 37.6 (36.5–39.0) OFC (cm) 26.5 (25.8–27.7) age of an infant was obtained on the basis of maternal obstetric history Sex or early perinatal ultrasound, and was subsequently confirmed after Male (n) 52 (50%) birth by the New Ballard score.17 The postconceptional age was cal- Female (n) 52 (50%) culated by adding the postnatal age of the infant at the time of Apgar scores ophthalmic examination to the gestational age. 1 min 7 (6–9) Measurement of IOP. The IOP of preterm infants was longitu- 5 min 9 (8–10) dinally measured at standard time intervals at 1, 4, 6, 8, and 10 weeks Arterial cord blood pH 7.30 (7.23–7.34) after birth. The first eye examination at week 1 was voluntary and the Base excess (mmol/L) Ϫ4.0 (Ϫ7.2–Ϫ2.2) other four measurements coincided exactly with the timing of routine Mode of delivery ROP examinations provided by the NICU for at-risk infants. A single Cesarean section (n) 66 (63%) ophthalmologist (BSMT) performed the ocular examination in all stud- Vaginal (n) 38 (37%) ied infants. In each examination, 1 drop of 1% preservative-free ameth- Maternal diabetes (n) 14 (13.5%) ocaine hydrochloride (Chauvin Pharmaceuticals Ltd., Surrey, UK) was Prolonged rupture of membrane (n) 16 (15.4%) instilled before the were opened with a Cook’s infant specu- Histologic chorioamnionitis* (n) 19 (23%) lum. The infant was then held supine by an experienced neonatal Pre-eclampsia of pregnancy (n) 29 (28%) nurse. It usually took the patient 30 to 90 seconds become accustomed Antenatal dexamethasone (mg) 12 (6–24) Duration between the last dose of antenatal to the speculum and settle down. IOP was always measured before the dexamethasone and delivery (h) 6 (3–43) dilatation of and indirect ophthalmic examination for ROP and CRIBS score 1 (1–2) never in irritable or crying infants. None of the patients received Oxygenation index (12 h of age) 3.7 (1.9–6.6) sedative drugs or muscle relaxants just before or during the eye RDS examination. IOP was measured with a handheld applanation tonom- ϽStage 2 60 (57%) eter (Tonopen II; Oculab, La Jolla, CA). Other methods of IOP mea- ՆStage 2 44 (43%) surement such as Goldmann applanation tonometry and noncontact Mechanical ventilation (week 1) (n) tonometry are impossible to perform at the bedside in preterm new- HFOV 11 (11%) borns, and Perkin’s tonometry is also not suitable because of the IPPV 5 (5%) CPAP 45 (43%) relatively large contact area on the . Previous studies have No ventilatory support 43 (41%) reported good agreement in IOP measurement between the Tonopen Use of surfactant (n) 86 (83%) 18–20 and Goldmann applanation tonometers. In addition, applanation Duration of mechanical ventilation (days) 1.3 (0.4–6.3) tonometry was chosen over indentation tonometry because the latter Duration of O2 requirement (days) 17 (5–34) 21 method is substantially affected by scleral rigidity. Thus, the hand- FiO2max 0.30 (0.23–0.40) held Tonopen represents the most accessible tool that can be reliably Mean airway pressuremax (cmH2O) 9 (8–11) used for IOP assessment in this age group of patients. The reported IOP in each eye represented an average value of three consecutive mea- Results are median (interquartile range) or number (%). surements. Triplicate measurements with discrepancy greater than * Data were available in 82 infants. 10% were considered unreliable and the outlier readings were ex- cluded from the study. number (%). As multiple measurements were obtained for each infant, Data Collection. During each IOP assessment, detailed informa- mixed-effects models22 were used to assess the association between tion was recorded regarding the health status of the subject. The IOP and gestational/postconceptional/postnatal ages, anthropometric information included postconceptional and postnatal ages; anthropo- parameters, and the important physiologic or environmental factors metric parameters such as body weight, body length, and head circum- listed in Tables 1 and 2 that may affect the ocular pressure. The ference; systolic, mean, and diastolic blood pressures; respiratory data mixed-effects models were used to adjust for longitudinal and cross- comprising the mode of mechanical ventilation at the time of IOP sectional (i.e., left-and-right eye correlation) random factors, and the measurement: high-frequency oscillatory ventilation ([HFOV] model covariates were chosen by a stepwise selection method consisting of a 3100A; Sensormedics Co., Yorba Linda, CA), intermittent positive- forward addition step and a backward elimination step. The forward pressure ventilation (IPPV; Infant Star; Infrasonics Inc., San Diego, CA), addition step started with the null model, which included the intercept continuous positive airway pressure (CPAP; Infant Flow System MDE/ term and the random individual effect only. The most significant M672B; Electro Medical Equipment Ltd., Brighton, UK), and the max- covariate is then added in a stepwise fashion into the model until all

imum mean airway pressure (MAPmax) and maximum oxygen concen- covariates that are not selected have P Ͼ 0.05. Thereafter, the back- tration (FiO2max) during the assessment period; medications, especially ward elimination step removes the most nonsignificant covariate from inhaled corticosteroids (fluticasone propionate; Flixotide; GlaxoSmith- the model until all covariates remaining have P Ͻ 0.05. In this exercise, Kline Ltd., Brentford, UK) 100 ␮g administered every 12 hours via a the first forward step indicated that postconceptional age, postnatal holding chamber (MV15s Aerochamber; Trudell Medical International, age, and body length were the most significant covariate factors. As London, ON, Canada), diuretics (hydrochlorothiazide; Apotex Inc., these factors were highly intercorrelated and the postconceptional age Weston, ON, Canada, 1 mg/kg administered every 12 hours or spirono- was considered most important and clinically meaningful among them, lactone; Pharmacia Ltd., New York, NY, 1 mg/kg administered every the forward addition procedure, therefore, started from the model that 12 hours); and serious complications associated with prematurity—in included the intercept and postconceptional age as the fixed effects. particular, adverse intracranial events such as intraventricular hemor- The random effect was always included to explain the individual rhage, periventricular leukomalacia, and cerebral atrophy; stages of variation. All tests were performed with commercial software (Windows ROP; bronchopulmonary dysplasia; and clinical sepsis. version of S-Plus 2000; MathSoft Inc., Seattle, WA). The mixed-effects Statistical Analysis. The descriptive statistics on demographic models were fitted using the function LME in the software. All compari- data and outcomes are expressed as median (interquartile range) or sons were two-tailed, and the level of significance was set at 5%.

Downloaded from iovs.arvojournals.org on 10/01/2021 IOVS, January 2008, Vol. 49, No. 1 IOP in Preterm Infants 89

TABLE 2. Clinical Outcomes of the Study Population IOP of preterm infants was significantly (1) decreased by Ϫ0.11 (0.04) mm Hg for each week in postconceptional age P Ͻ 0.001; Fig. 1); (2) increased by 1.86 (0.68) mm Hg in) 104 ؍ Outcomes n infants on HFOV compared with those who were breathing Use of inhaled corticosteroids (n) 40 (38%) spontaneously without assisted ventilation (P ϭ 0.01; Fig. 1). Use of diuretics (n) 42 (40%) There was, however, no significant influence of conventional Clinical sepsis (n) 31 (30%) ϭ ϭ ROP (left eye) IPPV (P 0.63) or CPAP (P 0.35) on IOP; (3) decreased by ϽStage 2 79 (76%) Ϫ0.05 (0.02) mm Hg for each 1-mm Hg increment in mean ՆStage 2 25 (24%) blood pressure (P ϭ 0.01); (4) decreased by Ϫ0.70 (0.32) mm ROP (right eye) Hg after beginning of treatment with inhaled corticosteroids ϽStage 2 79 (76%) (P ϭ 0.03; Fig. 1). In addition, in 40 infants treated with inhaled ՆStage 2 25 (24%) fluticasone propionate, the IOP before and after treatment was Intraventricular hemorrhage (n) 16.7 (14.9– 19.8) mm Hg (median [interquartile range]) and ϽGrade 2 81 (78%) Ͻ Ն 15.0 (13.5–18.0) mm Hg, respectively (P 0.05); and (5) Grade 2 23 (22%) Ϫ Cerebral atrophy with ventricular dilatation (n) 2 (2%) decreased by 0.23 (0.11) mm Hg for each unit increase in Apgar score at 1 minute (P ϭ 0.04). In the third stage with the O2 requirement at 36 weeks’ postconceptional age (n) 5 (5%) Duration of hospitalization (days) 75 (60–97) backward elimination step, all nonsignificant parameters were successfully eliminated, whereas factors (1) through (5) were Infants requiring postnatal systemic corticosteroids treatment retained in the statistical models. were excluded from this study. The median (interquartile range) IOP ranged from 16.9 (14.5–19.3) to 14.6 (12.2–17.1) mm Hg at 26.1 and 46.4 weeks of postconceptional age, respectively. Figure 1 summarizes the Approvals. Approval of the study ethics was obtained from the 10th, 50th, and 90th percentile distribution of the normative Research Ethics Committee of The Chinese University of Hong Kong. IOP after adjustment for the significant perinatal factors. Written informed parental consent was obtained for each infant before commencement of the examination, in accordance with the Declara- tion of Helsinki. DISCUSSION

RESULTS We designed this study to establish a normative range of IOP in preterm infants and to monitor longitudinal changes of ocular Of 192 preterm infants eligible for ROP screening during the pressure during this vulnerable period of eye development. We study period, 131 were longitudinally assessed for IOP. Patients also determined critical perinatal factors, both physiologic and excluded were infants (1) with undetermined gestational age environmental, that may affect IOP during the early weeks of because their mothers did not receive antenatal care (n ϭ 15), postnatal life. This study, however, was not designed to iden- (2) whose parents refused consent or could not provide con- tify the underlying mechanisms of action of these influential sent (n ϭ 21), (3) who were recruited but subsequently re- perinatal factors. The latter investigation is beyond the scope ceived systemic corticosteroids for treatment of chronic lung of the present study and should be performed after such disease (n ϭ 27), (4) who were outborn admissions referred to factors have been identified. As far as we are aware, this is the Prince of Wales Hospital for emergency surgery and subse- first large-scale longitudinal study of IOP in preterm newborns quently transferred back to the referring hospitals (n ϭ 11), (5) that has taken into account the clinical factors and outcomes with early neonatal death (n ϭ 6), (6) with multiorgan failure that may affect the ocular pressure. The most significant and (n ϭ 5), or (7) with dysmorphism or chromosomal abnormal- consistent finding was that the IOP of preterm infants de- ities (n ϭ 3). In total, 104 infants (excluding those treated with creases with increasing postconceptional age. The IOP was systemic corticosteroids) participated in the study, and 208 also negatively associated with the mean blood pressure, Apgar eyes were examined with 2868 measurements performed. The score at 1 minute, and use of inhaled corticosteroids, but clinical characteristics and outcomes of the study infants are correlated positively with HFOV. In addition, the study pro- summarized in Tables 1 and 2, respectively. vides a range of normative IOP after adjustment for influential There was no significant difference in longitudinal measure- perinatal factors (Figs. 1). There is, however, no significant ments of IOP between the left and right eyes (P ϭ 0.27), and association between IOP and stages of ROP. Pediatric ophthal- the mean (SE) discrepancy of measurement between the two mologists and neonatal clinicians equipped with such informa- eyes was 0.22 (0.20) mm Hg. All the results were therefore tion may now compare their clinical measurements against the pooled and analyzed together. In the first stage of the analysis, percentile chart (Fig. 1) and make appropriate adjustments for all clinical characteristics and outcome parameters listed in perinatal factors so that the IOP of these vulnerable infants can Tables 1 and 2 were subjected to a univariate analysis. The be properly assessed. Thus far, neonatal clinicians have not results indicate that the following factors were significantly been routinely monitoring IOP in serious neonatal conditions associated with IOP: postconceptional, gestational, or postna- with intracranial complication or the effects of drug treat- tal age (P Ͻ 0.0001); longitudinal anthropometric parameters, ments, in particular, systemic corticosteroids usage. The per- including body weight, body length, and occipitofrontal head centile chart should therefore be useful in assisting surveillance circumference (P Ͻ 0.0001); Apgar score at 1 minute (P ϭ and revealing clinical conditions or therapeutic agents that may 0.03); treatment interventions such as the mode of mechanical result in ocular hypertension and adversely affect vision. Ͻ ϭ ventilation (P 0.0001), FiO2max (P 0.02), and the use of To our knowledge, there has been no normative range of inhaled corticosteroids (P ϭ 0.0003); systolic, mean and dia- IOP established in preterm newborns in the literature. In the stolic blood pressures (P Ͻ 0.0001); stages of ROP in both the present study, none of our subjects was sedated or given left and right eyes (P ϭ 0.03 and P ϭ 0.04, respectively); and muscle relaxants during the examination, as these mediations severity of intraventricular hemorrhage (P ϭ 0.002). In the could have interfered with the results by artificially lowering second stage, these factors were subjected to the mixed-effects the IOP. Thus, our measurements are intended to reflect the models for assessment (i.e., forward addition of parameters to IOP of calm and resting preterm infants in the normal arousal the mixed-effects models). The results indicated that the mean state. Two previous cross-sectional studies in 1950s suggested

Downloaded from iovs.arvojournals.org on 10/01/2021 90 Ng et al. IOVS, January 2008, Vol. 49, No. 1

FIGURE 1. A percentile chart to demonstrate the variation of IOP with postconceptional age in preterm infants. The prediction intervals (10th, 50th, and 90th percentiles) were computed by setting the Apgar score at 1 minute at 7 (median) and the mean blood pressure at 48 mm Hg (median) in infants who did not require high-frequency oscillatory ventilation and inhaled corticosteroid treatment. The prediction formula was evaluated using 1,000,000 simulations from the estimated model. The influences of high frequency oscillatory ventilation and inhaled corticoste- roids on IOP are also illustrated in this graph. The use of high-frequency oscillatory ventilation (red line) has the effect of shifting the median (50th percentile) upward, whereas the use of inhaled corticosteroids (green line) has the effect of shifting the median downward.

that the IOP in preterm infants ranged between an average of ment of perinatal factors. In addition, our study provided the 24.5 mm Hg (range: 6.5–33.0 mm Hg) and 35.0 mm Hg.5,9 percentile distribution of IOP between 26.1 and 46.4 weeks of These results are much higher than the IOP of 16.9 (14.5–19.3) postconceptional age (Fig. 1). Although the previous recom- and 14.6 (12.2–17.1) mm Hg obtained at 26.1 and 46.4 mendation proposed by Tucker et al.4 suggested that a mea- postconceptional weeks in our present study, respectively surement Ͼ18.0 mm Hg could probably be considered ele- (Fig. 1). In contrast, our measurements are in good agreement vated, our results suggested that the top (90th percentiles) and and correspond closely with values obtained from more recent bottom (10th percentiles) 10% of measurements were distrib- studies.4,6–8,10–12 The latter clinical trials recorded IOP levels uted between 20.5 and 22.8 and 8.8 and 11.0 mm Hg, respec- ranging between (mean [ϮSD] or median [CI]) 10.1 (Ϯ2.2) tively, depending on the postconceptional age. A review of the mm Hg6; 10.3 (Ϯ3.5) mm Hg,4; 11.0 (Ϯ2.0) to 13.3 (Ϯ2.9) mm indirect ophthalmoscopy findings during ROP screening Hg depending on postnatal/postconceptional age11; 13.7 showed that none of the infants had abnormalities (Ϯ3.28) and 13.8 (Ϯ3.67) mm Hg in the right and left eye, related to raised IOP such as significant cupping of the optic respectively, of premature infants born by in vitro fertilization nerve head, and no study patient received treatment for ocular or natural conception10; 15.7 (Ϯ2.3) and 16.3 (Ϯ3.7) mm Hg hypertension. As our results were derived from a relatively in two separate control groups12; 15.5 (13.5–18.2) mm Hg7; large sample size, with serial measurements, and more impor- and 18.6 (Ϯ2.3) mm Hg.8 The discrepancy of measurements tant, adjusted for critical perinatal factors, the percentile charts between studies performed in the 1950s,5,9 and those in the (Fig. 1) should more accurately reflect the normative variation recent era4,6–8,10–12 is probably attributable to different cate- of IOP within the specified postconceptional age in preterm gories of survivors of preterm infants (i.e., no ventilator-depen- infants. dent survivors in 1950s), as well as the different instrument and In contrast to the findings of most previous studies,4–6,8,12 our techniques of measurement used between studies.15 In fact, in data suggest a strong negative association between IOP and the latest studies by McKibbin et al.6 and Axer-Siegel et al.,12 postconceptional age or anthropometric parameters, includ- the mean IOPs ranging between 15.5 and 16.3 mm Hg are in ing body weight, body length, and head circumference. The good agreement with the results of the present study. Further, postconceptional age was the most consistent and signifi- an analysis of data obtained from all recent studies4,6–8,10–12 cant factor determined by the mixed-effects models. As suggest that the mean/median IOP varies between 10.1 and there was a wide variation in IOP between individual pa- 18.6 mm Hg. Similarly, our cohort of preterm infants had a tients (Figs. 1) and most studies involved only cross-sec- median IOP within this range (14.6–16.9 mm Hg) after adjust- tional measurements,4–6,8,12 our longitudinal data were

Downloaded from iovs.arvojournals.org on 10/01/2021 IOVS, January 2008, Vol. 49, No. 1 IOP in Preterm Infants 91

probably the key element in revealing a significant relation- likely to be of older gestational age at birth. Hence, without an ship between IOP and postconceptional age or maturity. ideal group of normal patients, the only alternative method to The results further suggested that the IOP slowly decreased establish a normative range of IOP is to construct an appropri- at an average rate of 0.11 mm Hg with every week increase ate statistical model and adjust for significant perinatal factors in maturity after birth. In accordance with our findings, the that may affect the ocular pressure. In this respect, we have only other study with serial IOP measurements also demon- successfully developed a quantitative statistical model and con- strated a decrease in the mean IOP during the first month of structed a percentile chart (Fig. 1) that can be used for future postnatal life in preterm infants with gestational age ranged reference purposes. In addition, we have identified specific between 26 and 32 weeks.11 As it is a well-recognized perinatal factors so that the measured IOP can be appropriately phenomenon that the neonatal blood pressure increases corrected. Second, with recent advances in neonatology, seri- with gestational age,23,24 it is not unexpected to find a ous complications of prematurity are relatively uncommon, significant negative correlation between IOP and mean including severe stages of ROP, intracranial hemorrhage, and blood pressure. However, the exact mechanism giving rise hydrocephalus. Thus, the impact of these rare adverse IOP to a decline in IOP with increasing maturity is not fully events may not be adequately reflected in this study. Third, understood and is beyond the scope of this study. Whether artifactual elevation of IOP may occur in infants who are awake this phenomenon represents a programmed maturation pro- and react with vigorous resistance to ophthalmic examination, cess, related to an increase in the dimension of ocular or associates with the use of speculum during the pro- structures, or under the influence of complex neuroendo- cedure. A previous study suggests that the use of an eyelid crine control, requires further investigation. Perhaps, this speculum in anesthetized children may increase the IOP by an negative correlation is related to central corneal thickness, average of 4 mm Hg (mean IOP was 20.16 and 16.44 mm Hg, which in turn influences IOP measurement. with and without the eyelid speculum, respectively).31 The A significant negative association was also demonstrated low scleral rigidity in preterm infants may further exaggerate between IOP and Apgar score at 1 minute. The clinical signif- this effect. Nonetheless, it is not ethical or practical to admin- icance of this finding has not been fully elucidated but suggests ister general anesthesia just for measuring IOP, and the com- that the condition of infants at birth can potentially influence monly used anesthetic gas can inadvertently suppress the oc- the IOP. The poorer the condition at birth (i.e., the lower the ular pressure, giving rise to falsely low readings. Further, it is Apgar score at 1 minute), the higher will be the IOP. Whether not possible to perform an adequate ophthalmic examination this observation is related to a mild degree of cerebral insult or without assistance of an eyelid speculum, as this age group of edema,25 causing a transient increase in intracranial pressure patients is not able to cooperate. Therefore, we consider that that transmits to the eyes, is not certain. Further studies with the best approach in establishing a normative range is to sophisticated radiologic imaging for detection of cerebral simulate the real-life situation and to measure the IOP in resting edema or direct intracranial pressure measurement in animal preterm infants using the most accessible tools and the routine models are warranted to confirm this finding and to delineate procedures for eye examinations. We meticulously minimized the underlying mechanism. We did not expect, however, to the discomfort of infants and undesirable effects of the eyelid find a significant negative association between IOP and the use speculum by (1) applying a local anesthetic eyedrop before the of inhaled corticosteroids. Our previous study suggested that ophthalmic examination, (2) gently opening the eyelid with inhaled fluticasone could be absorbed systemically via the the smallest infant speculum available, and (3) patiently wait- pulmonary circulation, causing moderately severe hypotha- ing for the infant to become accustomed to the procedure. lamic-pituitary-adrenal axis suppression in preterm in- Thus, the IOP obtained in this study represents the standard fants.26 As systemic corticosteroids and steroids applied method of eye assessment normally used in clinical practice for locally to the eye have been shown to induce ocular hyper- this age group. tension in children and adults,13,15–17 a paradoxical de- In summary, our findings suggest that the median (10th– crease in IOP associated with the use of inhaled prepara- 90th percentile) IOP varies between 16.9 (12.3–21.5) and 14.6 tions in preterm infants is unusual. There are a few plausible (10.1–19.2) mm Hg at 26.1 and 46.4 weeks postconceptional explanations. First, the dosage of fluticasone used in the age, respectively. Important physiologic and environmental previous study was very high (1000 ␮g/d),27 being five times factors, including postconceptional age, mean blood pressure, the dosage used in the present study. Hence, its systemic use of inhaled corticosteroids, and Apgar score at 1 minute effect may have been minimized in the present study. Sec- have been demonstrated to be negatively correlated, whereas ond, although regular users of high doses of inhaled corti- HFOV is positively associated with the IOP. Pediatric ophthal- costeroids prescribed for a prolonged period (Ͼ3 months) mologists and neonatal clinicians will be able to compare the are at an increased risk of ocular hypertension,28 short-term IOP measurement against the percentile chart (Fig. 1) and usage with normal therapeutic dosage has not been associ- make appropriate adjustment for critical perinatal factors so ated with a significant increase in IOP.29,30 Third, the mixed- that the ocular pressure of the preterm infant can be properly effects models analysis indicated that the use of HFOV could evaluated. More important, the percentile chart should be significantly increase IOP (Fig. 1). After treatment with in- useful in assisting surveillance and revealing complications of haled fluticasone, all infants on HFOV were subsequently prematurity or therapeutic agents that may cause a sustained weaned off mechanical ventilation, and this factor could and abnormal increase in IOP. have contributed at least partially to the decrease of IOP in treated infants. However, it is difficult to be certain exactly References which component of HFOV (e.g., the positive pressure or constant vibration) causes the increase in IOP. 1. The STOP-ROP Multicenter Study Group. Supplemental Therapeu- tic Oxygen for Prethreshold Retinopathy Of Prematurity (STOP- There are limitations to this study. Ideally, serial IOP mea- ROP), a randomized, controlled trial, I: primary outcomes. Pediat- surements should be measured in normal uncomplicated pre- rics. 2000;105:295–310. term infants. However, all preterm infants are by definition not 2. Phelps DL. Retinopathy of prematurity: clinical trials. Neoreviews. normal. Even relatively uncomplicated infants, including those 2001;2:e167–172. who do not require mechanical ventilation, are not on media- 3. Ng PC, Kwok AK, Lee CH, et al. Early pituitary-adrenal responses tions such as inhaled or systemic corticosteroids, and are free and retinopathy of prematurity in very low birth weight infants. of serious complications of prematurity, are difficult to find and Pediatr Res. 2004;55:114–119.

Downloaded from iovs.arvojournals.org on 10/01/2021 92 Ng et al. IOVS, January 2008, Vol. 49, No. 1

4. Tucker SM, Enzenauer RW, Levin AV, Morin JD, Hellmann J. 19. Li J, Herndon LW, Asrani SG, Stinnett S, Allingham RR. Clinical Corneal diameter, axial length, and intraocular pressure in prema- comparison of the Proview eye pressure monitor with the Gold- ture infants. Ophthalmology. 1992;99:1296–1300. mann applanation tonometer and the Tonopen. Arch Ophthalmol. 5. Spierer A, Huna R, Hirsh A, Chetrit A. Normal intraocular pressure 2004;122:1117–1121. in premature infants. Am J Ophthalmol. 1994;117:810–813. 20. Bandyopadhyay M, Raychaudhuri A, Lahiri SK, Schwartz EC, Myatt 6. McKibbin M, Cassidy L, Dabbs TR, Verma D, McKibbin M. Intraoc- M, Johnson GJ. Comparison of Goldmann applanation tonometry ular pressure, pulse amplitude and pulsatile ocular blood flow with the Tonopen for measuring intraocular pressure in a popula- measurement in premature infants screened for retinopathy of tion-based survey in rural West Bengal. Ophthalmic Epidemiol prematurity. Eye. 1999;13:266–267. . 2002;9:215–224. 21. Hartnett ME, Hirose T, Richardson TM, Garsd A, Gilbert MM, Krug 7. Musarella MA, Morin JD. Anterior segment and intraocular pres- J Jr. Intraocular pressure determination in infants with severe sure measurements of the unanesthetized premature infant. Metab retinopathy of prematurity. Arch Clin Exp Ophthalmol. 1992;230: Pediatr Syst Ophthalmol. 1985;8:53–60. 406–410. 8. Axer-Siegel R, Bourla D, Sirota L, Weinberger D, Snir M. Ocular 22. McLeod A. Multivariate multilevel models. In: Leyland AH, Gold- growth in premature infants conceived by in vitro fertilization stein H, eds. Multilevel Modelling of Health Statistics. West Sus- versus natural conception. Invest Ophthalmol Vis Sci. 2005;46: sex, UK: John Wiley & Sons; 2001:59–71. 1163–1169. 23. Zubrow AB, Hulman S, Kushner H, Falkner B. Determinants of 9. Dolcet L. Tension ocular del recien nacido. Arch Soc Oftal His- blood pressure in infants admitted to neonatal intensive care units: pano-Am. 1952;12:1057–1063. a prospective multicenter study. Philadelphia Neonatal Blood Pres- 10. Brockhurst RJ. The intraocular pressure of premature infants. Am J sure Study Group. J Perinatol. 1995;15:470–479. Ophthalmol. 1955;39:808–811. 24. LeFlore JL, Engle WD. Clinical factors influencing blood pressure 11. Ricci B. Intraocular pressure in premature babies in the first month in the neonate. Neoreviews. 2002;3:e145–150. of life. J AAPOS. 1999;3:125–127. 25. Mujsce DJ, Christensen MA, Vannucci RC. Cerebral blood flow and 12. Axer-Siegel R, Bourla D, Friling R, et al. Intraocular pressure vari- edema in perinatal hypoxic-ischemic brain damage. Pediatr Res. ations after diode laser photocoagulation for threshold retinopathy 1990;27:450–453. of prematurity. Ophthalmology. 2004;111:1734–1738. 26. Ng PC, Fok TF, Wong GW, et al. Pituitary-adrenal suppression in 13. Weinreb RN, Polansky JR, Kramer SG, Baxter JD. Acute effects of preterm, very low birth weight infants after inhaled fluticasone dexamethasone on intraocular pressure in glaucoma. Invest Oph- propionate treatment. J Clin Endocrinol Metab. 1998;83:2390– thalmol Vis Sci. 1985;26:170–175. 2393. 14. Tham CC, Ng JS, Li RT, Chik KW, Lam DS. Intraocular pressure 27. Fok TF, Lam K, Dolovich M, et al. Randomised controlled study of profile of a child on a systemic corticosteroid. Am J Ophthalmol. early use of inhaled corticosteroid in preterm infants with respi- ratory distress syndrome. Arch Dis Child Fetal Neonatal Ed. 1999; 2004;137:198–201. 80:F203–F208. 15. Herbert HM, Viswanathan A, Jackson H, Lightman SL. Risk factors 28. Garbe E, LeLorier J, Boivin JF, Suissa S. Inhaled and nasal glucocor- for elevated intraocular pressure in . J Glaucoma. 2004;13: ticoids and the risks of ocular hypertension or open-angle glau- 96–99. coma. JAMA. 1997;277:722–727. 16. BenEzra D, Wysenbeek YS, Cohen E. Increased intraocular pres- 29. Duh MS, Walker AM, Lindmark B, Laties AM. Association between sure during treatment for chronic uveitis. Arch Clin Exp Ophthal- intraocular pressure and budesonide inhalation therapy in asth- mol. 1997;235:200–203. matic patients. Ann Allergy Asthma Immunol. 2000;85:356–361. 17. Ballard JL, Khoury JC, Wedig K, Wang L, Eilers-Walsman BL, Lipp 30. Behbehani AH, Owayed AF, Hijazi ZM, Eslah EA, Al-Jazzaf AM. R. New Ballard Score, expanded to include extremely premature Cataract and ocular hypertension in children on inhaled cortico- infants. J Pediatr. 1991;119:417–423. steroid therapy. J Pediatr Ophthalmol . 2005;42:23– 18. van der Jagt LH, Jansonius NM. Three portable tonometers, the 27. TGDc-01, the ICARE and the Tonopen XL, compared with each 31. Epley KD, Tychsen L, Lueder GT. The effect of an eyelid speculum other and with Goldmann applanation tonometry. Ophthalmic on intraocular pressure measurement in children. Am J Ophthal- Physiol Opt. 2005;25:429–435. mol. 2002;134:926–927.

Downloaded from iovs.arvojournals.org on 10/01/2021