Diurnal Variation of Axial Length, Intraocular Pressure, and Anterior Eye Biometrics

Abstract

purpose. To investigate the diurnal variation in axial length and anterior eye biometrics, while simultaneously measuring intraocular pressure (IOP) with dynamic contour tonometry in human subjects.

methods. Fifteen young adult near-emmetropic subjects had axial length, anterior eye biometrics (central corneal thickness and anterior chamber dimensions), and IOP measured at six different times across a 24-hour measurement period. Repeated-measures ANOVA and sine curve fitting were used to analyze the diurnal rhythms in each measured parameter.

results. Axial length was found to undergo significant diurnal variation (P = 0.0006). The mean amplitude of axial length change was 0.046 ± 0.022 mm. The mean peak in axial length was found to occur at 1113. Intraocular pressure and ocular pulse amplitude were also found to undergo significant diurnal change (P < 0.0001 and 0.0006, respectively). The variation in axial length exhibited a significant association with the change in IOP (r = 0.37, P = 0.001). No significant difference was found between the mean peak times of axial length and IOP. Anterior eye biometric measures of central corneal thickness and anterior chamber depth were also found to undergo significant diurnal changes (P < 0.0001 and 0.0368, respectively).

conclusions. Axial length undergoes significant variation over a 24-hour period. Associations exist between the change in axial length and the change in IOP, as measured by dynamic contour tonometry. These results may have significant implications for the role of ocular diurnal rhythms in emmetropization.

It has been well established that axial length undergoes a significant diurnal variation in both animal ^{1 2 3 4 5 6 7 8 9 10} and human subjects. ^{11 12} Evidence from animal studies indicates that the amplitude and phase of these daily fluctuations in axial length may play an important role in the control of eye growth and refractive error development. ^{3 5 9} While the pattern of these daily fluctuations in ocular dimensions has been well documented, the physiological causes and mechanisms underlying these changes are less well understood. 

As intraocular pressure (IOP) is also known to undergo diurnal variation in animals ^{4 7 13 14} and in humans, ^{15 16 17 18 19 20 21 22 23 24 25 26 27} it is conceivable that mechanical expansion and contraction of the globe as a result of changes in IOP may be responsible in part for the daily rhythms observed in axial length. It has also been found that large alterations in IOP (due to mechanical or surgical interventions) can cause significant and predictable changes in axial length consistent with the expansion and contraction of the globe in response to the IOP. ^{28 29 30 31 32} However, the exact role of natural IOP variations in the diurnal variation of axial length is still unclear. Phase shifts observed between the diurnal rhythms of the axial length and IOP in chicks suggest a more complex relationship between the two variables than a simple passive stretching of the globe in response to IOP. ³ Autonomic denervation in chicks has also been found to disassociate the rhythms of axial length and IOP. ⁶ A recent study investigating the diurnal variations of IOP and axial length in human subjects found similar mean peak times for the two rhythms, but no significant correlation was observed for the amplitude, phase, or period of the IOP and axial length rhythms in individual subjects. ¹²  

Anterior eye dimensions, such as corneal thickness ^{33 34 35 36 37} and anterior chamber depth, ^{38 39 40} are also known to undergo significant diurnal change. As axial length is typically defined as the distance from the anterior corneal surface to the retina, changes in these anterior eye dimensions may influence the observed changes in axial length. IOP measures with most modern tonometric techniques are also known to be influenced by the thickness of the cornea. ^{41 42 43 44 45 46} Given that both axial length and IOP measures may be influenced by the dimensions of the anterior eye, it is conceivable that changes in anterior eye biometrics could potentially confound studies conducted to investigate the relationship between IOP and axial length. 

In order to further investigate the factors influencing the diurnal fluctuation in axial length, we measured axial length and anterior eye biometrics (central corneal thickness and anterior chamber depth) over a 24-hour period (but not during the subjects’ sleep periods) in a group of healthy young adult subjects while simultaneously assessing IOP using a tonometric technique (Pascal dynamic contour tonometry) that is thought to provide measurements of IOP independent of corneal thickness and biomechanical properties. ^{47 48 49 50}  

Methods

Subjects and Procedure

Fifteen young, healthy, near-emmetropic adult subjects aged between 20 and 27 years (mean age, 22 years) were recruited for the study. The subjects were primarily recruited from the students and staff of the QUT School of Optometry. Eight of the subjects were men. All subjects were free of any ocular or systemic disease and had no history of ocular surgery or significant trauma. None of the subjects were contact lens wearers. Before the study, each subject underwent an initial ophthalmic examination to ensure good ocular health and to determine their refractive status. All subjects had normal visual acuity of logMAR 0.00 or better. The subjects’ mean ± SD best sphere refraction (DS) was found to be −0.3 ± 0.4 DS (range, +0.25 to −1.125), with a mean ± SD cylinder (DC) refraction of −0.2 ± 0.3 DC (range, 0.00 to −1.00). No subject exhibited anisometropia greater than 0.75 DS. Slit lamp biomicroscopy revealed no evidence of narrow anterior chamber angles in any subject. 

Approval from the university human research ethics committee was obtained before commencement of the study, and subjects gave written informed consent to participate. All subjects were treated in accordance with the tenets of the Declaration of Helsinki. 

The study took place over five separate measurement days, with two to four subjects participating on each day. On each measurement day, the axial length, anterior eye biometrics, and IOP of each subject’s right eye were measured with the subject in the sitting position every 3 to 7 hours at six separate measurement sessions over a 24-hour period. The initial measurement for all subjects took place in the morning, at least 2 hours after their reported time of awaking on that day. After the initial measurement, subjects undertook their regular daily activities, and returned to the research laboratory for each measurement session. Over the course of the study, measurement sessions occurred at the following mean times: session 1: 0940 (range: 0835–1100), session 2: 1300 (1200–1410), session 3: 1730 (1700–1830), session 4: 2230 (2200–2320), session 5: 0600 (0500–0640), and session 6: 0920 (0800–1020). One subject was unable to attend for one of the scheduled measurement sessions. After session 4 (mean time, 2230), the subjects went to sleep in individual darkened rooms within the research laboratory. To ensure that postural variations in IOP ^{27 51 52} did not influence the results, the next morning, the subjects were awakened and instructed to sit for 5 minutes with their eyes closed before the beginning of measurement session 5. At each measurement session, the time taken to perform the entire series of measurements on each subject was approximately 20 minutes. To make sure that the axial length and anterior eye biometric measurements were not influenced by any corneal epithelial disruption brought about by contact tonometry or local anesthetic instillation, the tonometry was always the final measurement performed at each session. 

Axial length (defined in this case as the distance from the anterior corneal surface to the retinal pigment epithelium) was measured with an optical biometer (IOLMaster; Carl Zeiss Meditec, Inc., Jena, Germany). The biometer is a noncontact instrument based on the principles of partial coherence laser interferometry (PCI) ⁵³ and has been found to provide highly precise measurements of axial length. ^{54 55 56} For each subject, a total of five measurements of axial length were taken at each measurement session, and the mean of these readings was calculated. Any measurements of axial length from the biometer with a reported signal-to-noise ratio of less than 2.0 were repeated until five valid readings were attained. 

Anterior eye biometrics were also measured using a noncontact rotating Scheimpflug camera (Pentacam HR system; Oculus Inc., Wetzlar, Germany). Previous studies have shown the system to have excellent repeatability for measuring both central corneal thickness ^{57 58 59 60} and anterior chamber depth. ^{60 61 62 63} We used the system’s 50-picture, 3-D scan measurement mode for all measures. At each measurement session five scans were performed on each subject. Any measurements flagged by the instrument’s quality specification as unreliable were repeated until five valid measures were obtained. For this study, central corneal thickness (CCT; centered on the corneal apex), anterior chamber depth (ACD; the axial distance from the corneal endothelium to the anterior lens surface), and anterior chamber volume (ACV; calculated for a 12-mm diameter around the corneal apex) from each measurement were all recorded, and the mean of each parameter was calculated for each subject at each measurement session. 

All IOP measures were performed using the Pascal dynamic contour tonometer (DCT; Ziemer Ophthalmic Systems, Port, Switzerland). The DCT is a contact tonometer that works on the principle of contour matching. The instrument outputs mean IOP and ocular pulse amplitude (OPA; defined as the difference between the diastolic and systolic IOP over the measurement time) as well as a quality score (where a score of 4 or 5 indicates an unreliable result) for each measurement. The DCT has been found to exhibit good inter- and intraobserver repeatability, comparable to Goldmann applanation tonometry. ⁴⁷ Furthermore, the DCT has been found to provide IOP measures that are closer to true manometric pressures than is Goldmann applanation tonometry ⁶⁴ and also less influenced by corneal thickness than are other tonometric techniques. ^{47 48 49 50} Measurements with the DCT were taken according to the manufacturer’s instructions, after the instillation of a drop of local anesthetic (0.4% oxybuprocaine hydrochloride). Three DCT measurements were taken for each subject at each measurement session, and the mean IOP and OPA of the three measurements were calculated. Any measurement displaying a quality score of 4 or 5 was repeated until three valid measures were obtained. All DCT measurements were taken by one clinician experienced in the use of the instrument. In one subject at one measurement session, valid readings with quality better than 4 could not be obtained, and therefore no IOP or OPA measures were recorded for this session for this subject. After the IOP measurements, a careful slit lamp examination was performed to ensure that no substantial epithelial disruption occurred as a result of the contact tonometry procedure or anesthetic drops. 

Data Analysis

After data collection, the mean axial length, DCT measures (IOP and OPA), and anterior eye biometric measures (CCT, ACD, ACV) were calculated for each subject for each measurement session. Based on the data collected from all subjects across all sessions, the average coefficient of variation was calculated for each of the measurements, and was found to be 0.08% for axial length, 4.40% for IOP, 10.39% for OPA, 0.69% for CCT, 0.63% for ACD, and 2.14% for ACV. 

Each individual subject’s data were combined, and the group mean for each parameter at each measurement session was then calculated. To investigate the significance of changes occurring in the group mean values of each of the measured variables over the 24-hour measurement period, a repeated-measures ANOVA was performed with one within-subject factor (time of day). The two subjects who did not have complete sets of data from all six measurement sessions were not included in this ANOVA, since it required all subjects to have six complete data sets for all variables. Each subject’s average daily axial length, IOP, OPA, CCT, ACD, and ACV were also calculated as the mean of all measurements across all sessions for each parameter. The difference from each subject’s daily mean for each variable at each measurement session was then calculated. The amplitude of change (the difference between the maximum and minimum change from the mean) in each variable for each subject over the 24-hour period was also calculated, as well as the group mean amplitude of change in each variable. 

To investigate associations between the variations in axial length and the variation occurring in the other measured variables, we performed an analysis of covariance (ANCOVA), as described by Bland and Altman, ⁶⁵ for the analysis of repeated observations. This analysis provides a correlation coefficient (r) and a regression coefficient (slope) that describe the relationship between axial length and each of the other parameters. The significance of the association is determined based on the F statistic from the ANCOVA. Similar analyses were also performed, to investigate the relationship between the change in IOP and the change in anterior eye biometrics. 

To investigate further the 24-hour rhythms in each of the measured variables with a mathematical model, we modeled the change from the daily mean of each variable at each session for each subject with sine curve fitting. The best-fitting sine curve was calculated for each subject from the six measurements taken of each of the measured variables over the 24-hour measurement period. The following equation was used to fit the data for each variable:  

\[y{=}\ \frac{a}{2}\ \mathrm{sin}(2{\pi}\ \frac{\mathrm{time}}{24}{+}c)\]

The fitted curve therefore had a fixed period of 24 hours, and is defined by terms a (peak-to-trough difference) and c (phase). The parametersa and c were fit to the data by a linear least-squares method. For each variable, the acrophase (i.e., the time at which the peak value occurred) was also determined based on the model and expressed as the actual clock time of the occurrence of the fitted peak. The group mean peak-to-trough difference and acrophase were also calculated. To investigate whether the distribution of acrophases across our population of subjects for each measured variable was statistically different from a Gaussian distribution, the Rayleigh statistical test was used. ⁶⁶ The Wilcoxon signed rank test was used to determine whether there were significant differences between the group mean acrophase for each of the measured variables. 

Results

Significant diurnal variation was found to occur in axial length, DCT measures, and anterior eye biometrics in this population of near emmetropic young adult subjects. Repeated-measures ANOVA revealed a significant effect of time of day (P < 0.05) for all measured variables. Table 1displays the mean values (as measured for all subjects across all time points), measured amplitude of change, and repeated-measures ANOVA results for all the measured variables. Pair-wise comparisons (with Bonferroni adjustment for multiple comparisons) revealed no significant difference (P > 0.05) between the two morning measurement sessions performed at similar times (i.e., session 1 compared with session 6) for all measured variables, indicating minimal day-to-day variation in the measured parameters. 

The group mean axial length calculated from all measurements from all subjects was found to be 23.77 ± 0.7 mm. Axial length displayed a relatively consistent pattern of change over the measurement period for all subjects, with the measured maximum in mean axial length occurring at measurement session 2 (mean time of measurement, 1300) and the minimum in the evening at measurement session 4 (mean time of measurement, 2230). The mean measured amplitude of change in axial length (maximum to minimum difference) for all subjects was 0.046 ± 0.022 mm (range, 0.020–0.092). Repeated-measures ANOVA revealed the diurnal variation in axial length to be highly significant (P = 0.0006). Figure 1illustrates the mean change in axial length over the study period. 

Consistent diurnal variations in DCT measures of IOP and OPA were also observed. The group mean IOP was found to be 14.49 ± 1.7 mm Hg and the mean OPA was 2.10 ± 0.79 mm Hg. Both IOP and OPA measures exhibited a significant change over the 24-hour measurement period (ANOVA P < 0.0001 and 0.0006 for IOP and OPA, respectively). The mean measured amplitude of change in IOP over the study was 3.12 ± 0.94 mm Hg (range, 1.97–4.97) and in OPA was 1.27 ± 0.44 mm Hg (range, 0.43–2.50). Both IOP and OPA exhibited their mean maximum values at measurement session 1 (0940) and mean minimum values at measurement session 4 (2230; Fig. 1 ). 

Anterior eye biometrics were also found to undergo significant diurnal change. The group mean CCT was 0.532 ± 0.029 mm, with a mean amplitude of change of 0.018 ± 0.008 mm (range, 0.007–0.03) over the 24-hour study period. The mean maximum CCT was observed to occur at measurement 5 (0600) immediately after waking, and the mean minimum in CCT occurred at measurement session 4 (2230), just before the subjects went to sleep. Repeated-measures ANOVA revealed the change in CCT over the 24-hour study period to be significant (P < 0.0001). The group mean ACD was found to be 3.16 ± 0.27 mm with a measured mean amplitude of change of 0.073 ± 0.037 mm (range, 0.034–0.178). The mean peak in ACD (i.e., deepest ACD) occurred at measurement session 4 (2230), and the mean trough in ACD (i.e., shallowest ACD) occurred at measurement session 5, immediately after waking (0600). The change in ACD over the course of the study also just reached statistical significance (repeated measures ANOVA, P = 0.036). Figure 2illustrates the mean change observed in CCT, ACD, and axial length over the course of the 24-hour study period. ACV exhibited a pattern of change similar to that observed in the ACD throughout the day. The mean ACV was found to be 187.1 ± 30.1 mm³, with a mean amplitude of change of 15.10 ± 6.95 mm³ (range, 5.6–27.3). The diurnal change in ACV was highly significant (repeated-measures ANOVA P < 0.0001). 

ANCOVA revealed several significant associations between the changes occurring in axial length and the changes in the other measured variables (Table 2) . A significant positive correlation was found between the change in axial length and the change in IOP (r = 0.370, P = 0.001). The regression coefficient for these two variables was 0.0059, indicating a change of 5.9 μm in axial length for every 1-mm Hg change in IOP. Figure 3illustrates the relationship between the change in IOP and the change in axial length. A relatively weak but significant correlation was also found between the change in axial length and change in OPA (r = 0.300, P = 0.009), and a weak correlation that just reached significance was also found between the change in axial length and the change in ACD (r = −0.232, P = 0.045). The change in IOP was also found to have a strong positive correlation with the change in OPA (r = 0.659, P < 0.0001) and a weak negative correlation with the change in ACD (r = −0.271, P = 0.02). 

Sine curve fitting was performed for each subject to provide a mathematical model for the 24-hour rhythm occurring in axial length, IOP, OPA, CCT, ACD, and ACV data. This modeling provided estimates for each subject of the mean peak-to-trough difference and acrophase for each variable. Table 3displays the mean peak-to-trough difference and acrophase for each of the measured variables as well as the results of the Rayleigh statistical test. The Rayleigh test revealed that the timing of the peak (acrophase) was significantly different from a random Gaussian distribution for all variables measured (P < 0.05), suggesting that significant synchronized 24-hour rhythms occurred in each of the measured variables across our population of subjects. The Wilcoxon signed rank test revealed no significant difference (P > 0.05) between the mean acrophases of axial length (mean acrophase, 1113), IOP (mean acrophase, 1023), and OPA (mean acrophase, 1130), suggesting that the mean timing of the peaks in the rhythms of these three variables coincided. Ten of the 15 subjects exhibited peak times for axial length occurring within 4 hours of the peak of IOP. Significant differences were found between the mean acrophase of the axial length and CCT (mean acrophase, 0449; P = 0.001) and axial length and ACD (mean acrophase, 1913; P = 0.002) indicating significant phase differences between axial length and these other rhythms. 

Discussion

We have shown that significant variation occurs in axial length over a 24-hour period in our population of young adult, near-emmetropic subjects. Although the diurnal rhythms in axial length in animals have been well studied, there have been relatively few studies exploring the diurnal variation of axial length in human subjects. ^{11 12} The mean amplitude of change (0.046 mm) and timing of the peak axial length (1113) found in our present study are in relatively close agreement with the results from the two previous studies of the diurnal variation in axial length in human subjects who also used PCI techniques to measure axial length. Differences in the age of subjects tested, and/or differences in subjects’ refractive status may account for some of the small differences between our present study and previous investigations. 

We also found that a significant association exists between the variations occurring in axial length and the variations of IOP as measured by DCT. The mean phase timing of the peak of these two rhythms also appeared to be similar. The association observed between IOP and axial length is consistent with the hypothesis of passive expansion and contraction of the globe in response to IOP. Previous studies have found significant associations between IOP and axial length when large changes in IOP are surgically or mechanically induced, ^{28 29 30 31 32} but to our knowledge, this is the first study to show that associations exist between the natural, physiological changes in IOP and those of axial length in human subjects. The precise measurements of axial length with PCI and the fact that our IOP measures were taken with DCT (and are therefore unlikely to be confounded by concurrent changes in corneal thickness) have revealed the association between these two physiological rhythms in this study. 

The association found between the change in axial length and IOP, although statistically significant, was not strong (r = 0.370), indicating that only 14% of the variation in the change in axial length could be accounted for by the change in IOP. The regression coefficient for these two variables suggests approximately 5.9 μm of change in axial length per 1-mm Hg change in IOP. The mean amplitude of change in IOP was 3.12 mm Hg, which, based on this regression analysis, leads to 18 μm of change in axial length. As the measured mean amplitude of change in axial length was 46 μm, the total change observed in axial length cannot be explained completely by the change in IOP. Changes in IOP therefore may be involved in the diurnal variation of axial length in human subjects, but are not the sole reason for the changes observed. Changes in choroidal thickness, ^{3 7 8 9} and/or scleral proteoglycan synthesis ⁶⁷ as noted in previous animal studies may also be involved in the diurnal variation of axial length in human subjects. Because the ocular biometer (IOLMaster; Carl Zeiss Meditec, Inc.) measures from the anterior cornea to the retinal pigment epithelium, it cannot differentiate choroidal thickness changes from scleral changes, and therefore further research is necessary to characterize comprehensively the origins of the axial length changes found. 

That the diurnal change in IOP and axial length exhibit a significant association may have important implications for eye growth and refractive error development. Liu et al. ⁶⁸ found that young subjects with moderate levels of myopia exhibited differences in both the amplitude and phase of their 24-hour rhythms of IOP compared with age-matched emmetropic or mildly myopic subjects. As we have found that associations exist between the change in IOP and the change in axial length in our population of emmetropic subjects, it is plausible that a population of young myopic subjects may exhibit differences in their pattern of diurnal axial length change compared with our study population. Studies with animals have shown that there are significant differences in the phase of axial length rhythms in chicks undergoing myopic eye growth, and it has been suggested that these rhythms play an important role in the control of eye growth in these animals. ^{3 5 9} Further research to characterize the diurnal rhythms occurring in axial length and IOP in human myopic subjects may therefore help to clarify the etiological factors involved in the development of myopia and the control of eye growth in humans. 

The diurnal variation in IOP has been the focus of numerous investigations. In studies in which the diurnal variation of IOP was investigated, several different tonometric techniques have been used, including Goldmann applanation tonometry, ^{18 22} pneumotonometry, ^{25 27} noncontact air-puff tonometry, ^{19 20 24} and a handheld tonometer (Tonopen; Medtronic, Jacksonville, FL). ²³ All these tonometric techniques have been found to be influenced to different degrees by corneal thickness. ^{41 42 43 44 45 46} Our present study is one of the first to report on the diurnal variation of IOP with DCT, a tonometric technique that is not influenced by corneal thickness. Our results, however, showed the same general trend as several of the previous studies in healthy adult subjects using older tonometric techniques, with most subjects exhibiting their peak in IOP in the morning, and their minimum or trough in IOP observed in the afternoon/evening. ^{15 19 22 24 25 27 69} The magnitude of change that we found in IOP (mean amplitude, 3.12 mm Hg) is also consistent with previous studies into the diurnal variation of IOP in similar populations of healthy young adult nonglaucomatous subjects. ^{25 27 69} Patients with glaucoma have been found to exhibit differences in their diurnal pattern of IOP change, ^{15 22 70 71 72} and diurnal variations in IOP may also be an important risk factor in the development and progression of glaucoma. ^{73 74 75} As the use of DCT helps to remove some of the variability in IOP measures associated with corneal biometric parameters, investigation of the diurnal variation in IOP with DCT in glaucomatous subjects may help to further the understanding of the role of these IOP variations in glaucoma. 

A recent study by Hamilton et al. ⁷⁶ reported significant associations between the change in IOP and the change in CCT after waking, suggesting that peaks in IOP measured on waking may relate to errors in IOP estimates due to the overnight swelling in corneal thickness (due to the associations between IOP and corneal thickness with applanation tonometers). Our present study with the DCT generally did not find peaks in IOP to occur on waking (in most of our subjects), and also found no significant association between the change in CCT and IOP (r = 0.14 P = 0.23). As DCT measures are less influenced by corneal thickness measures than other tonometric techniques, our findings are in general agreement with the suggestion of Hamilton et al. ⁷⁶ that peaks in IOP on waking may relate to errors in IOP estimates due to corneal swelling. It is also possible that any peaks in IOP as a result of sleep may have subsided before our DCT measures were performed (the data collection procedures at each measurement session took approximately 20 minutes to complete after subjects awoke), as it has been shown previously that peaks in IOP during sleep return rapidly to normal levels within 15 minutes of waking. ^{19 20} We did not want to interrupt our subjects’ sleep patterns with the relatively lengthy measurement protocol (i.e., 20 minutes), so we therefore did not take any measurements with the DCT during the subjects’ sleep period. However, it may aid in the understanding of the “true” IOP during sleep, to investigate the IOP with DCT during the nocturnal sleep period. 

DCT also provides measures of the OPA, a parameter thought to provide information regarding intraocular blood flow as it represents the dynamic changes occurring in IOP with the cardiac cycle (i.e., the change that occurs in IOP when a bolus of blood enters the ocular circulation with the cardiac pulse). We have demonstrated that significant change occurs in this parameter over a 24-hour period, with the highest levels being present in the morning and the lowest in the evening. A previous study in which the Langham ocular blood flow system was used also reported a slight decrease in OPA in the evening compared with daytime measures in normal subjects. ⁷⁷ Other studies, ^{78 79} including a recent study using DCT ⁷⁹ have reported no significant diurnal variation in OPA over their measurement period. However, neither of these two recent studies ^{78 79} measured OPA over a 24-hour period. By collecting OPA data from our subjects over a 24-hour period, we were able to establish that significant diurnal variation occurs in this parameter. The decrease in OPA at night could be indicative of a decrease in ocular blood flow at this time. However, the change in OPA may simply be related to the observed association between OPA and IOP. We found that a significant positive correlation existed between the change in IOP and the change in OPA. Previous cross-sectional studies have also noted a significant positive association between IOP and OPA in normal subjects. ^{80 81 82}  

We have also observed that significant changes occurred in anterior eye biometrics over the 24-hour study period. The significant swelling of the cornea observed in our subjects on waking is consistent with previous studies into the diurnal variation in CCT. ^{33 34 35 36 37} Axial length was also observed to be increased on waking (compared with the previous nighttime measure), which is consistent with the corneal swelling on waking contributing to this axial length change. However, no significant association was found between the change in axial length and the change in CCT (r = 0.07, P = 0.56). 

Although there have been numerous studies investigating the diurnal variation in CCT, there have been only relatively limited studies investigating the diurnal change in ACD. We found a significant change in ACD over the 24-hour study period, with the maximum ACD observed at night and the minimum observed in the morning. The changes found in ACD were out of phase with the changes in axial length and IOP, with peaks in ACD occurring at a time similar to that of the troughs in the other two variables. The narrowing of the ACD observed in the morning coincided with the swelling of the cornea also observed on waking. This ACD change is therefore consistent with a swelling of the cornea in the posterior direction, as has been reported. ⁸³ The amplitude of ACD change in the morning was larger than the amplitude of change in corneal thickness, indicating that some anterior movement of the lens may also be present at first waking. 

Studies with rabbits ^{2 40} and chicks ^{3 10} have reported similar changes in ACD, with increases in ACD also noted at night. Contrary to these studies with animals and to the results from our present study, the two previous studies investigating the diurnal variation of ACD in human subjects (both using relatively low resolution imaging techniques) showed either irregular changes in ACD with the lowest values generally occurring at midday ³⁸ or a decrease in ACD at night. ³⁹ The reason for this discrepancy with the previous studies on human subjects may be due to the measurement techniques used or the data collection protocols used in the different studies. In our present study, we took six measurements over 24 hours, as opposed to two measurements 12 hours apart ³⁹ or serial measurements between 0900 and 1700 ³⁸ in the previous studies. 

Clinically, the measurement of ACD is important for several applications, including planning of cataract surgery and the diagnosis and management of closed-angle glaucoma. All our subjects had wide anterior chamber angles. However, it is known that the physiological properties and anatomic characteristics of the anterior chamber are different in patients with angle-closure glaucoma. ⁸⁴ Investigation of the pattern of diurnal variation in anterior chamber parameters with high-resolution techniques in populations of subjects with narrow anterior chamber angles may be an area worthy of future research and may lead to improved understanding of the pathophysiology underlying closed-angle glaucoma. 

In summary, we found diurnal variation to occur in a range of ocular parameters over a period of 24 hours. We also found significant association between the change in IOP and the change in axial length. Although the associations found are consistent with a passive expansion of the globe in response to IOP, they do not prove that IOP changes cause the changes in axial length. These results may have important implications for the role of ocular diurnal variations in emmetropization and ocular growth. 

 Supported by a QUT Institute of Biomedical Innovation 2007 Seeding Grant.

 Submitted for publication February 5, 2008; revised March 3, 2008; accepted 2008.

 Disclosure: S.A. Read, None; M.J. Collins, None; D.R. Iskander, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Summary of Group Mean Levels and Amplitude of Change for Axial Length, DCT Measures and Anterior Eye Biometrics

Variable	Mean ± SD	Mean Measured Amplitude of Change ± SD	P *
Axial length (mm)	23.77 ± 0.7	0.046 ± 0.022	0.0006
IOP (mm Hg)	14.49 ± 1.7	3.12 ± 0.94	<0.0001
OPA (mm Hg)	2.10 ± 0.79	1.27 ± 0.44	0.0006
CCT (mm)	0.532 ± 0.029	0.018 ± 0.008	<0.0001
ACD (mm)	3.16 ± 0.27	0.073 ± 0.037	0.0368
ACV (mm3)	187.1 ± 30.1	15.10 ± 6.95	<0.0001

*  Represents the significance of the F statistic from repeated-measures ANOVA.

Figure 1.

 

Mean change in axial length (top), IOP (middle), and OPA (bottom) over the 24-hour measurement period. All values expressed as the average difference from the daily mean at each measurement session. Vertical error bars, SEM; horizontal error bars, SE in the mean time that the measurement was taken at each session (in hours).

Figure 2.

 

Mean change in CCT, ACD, and axial length over the 24-hour study period. All values expressed as the average difference from the daily mean at each measurement session.

Table 2.

 

Summary of ANCOVAs Investigating the Relationship between the Variation in Axial Length and the Other Measured Variables

Variable	Regression Coefficient (slope)	Correlation Coefficient (r)	P *
ΔIOP	0.0059	0.370	0.001
ΔOPA	0.0109	0.300	0.009
ΔACD	−0.143	−0.232	0.045
ΔCCT	0.180	0.068	0.563
ΔACV	−0.0002	−0.056	0.637

*  Represents the significance of the F statistic from the ANCOVA.

Figure 3.

 

Change in axial length versus change in IOP. Difference in IOP from each subject’s mean daily IOP at each measurement session plotted against the difference in axial length from each subject’s individual mean daily axial length at each measurement session.

Table 3.

 

Summary of Sine Curve Modeling of the 24-Hour Rhythm in Axial Length, IOP, OPA, CCT, and ACD

	Peak-to-Trough Difference			Acrophase (Time)			Rayleigh Test (P)
		Mean ± SD	Range	Mean (±SD, h:min)	Range
Axial Length	0.029 ± 0.015	0.005–0.06	1113 (±3:27)	0550–1637	0.0015
IOP	2.27 ± 0.89	0.47–3.64	1023 (±3:31)	0513–1536	0.019
OPA	1.04 ± 0.41	0.39–1.75	1130 (±3:17)	0703–1520	0.0010
CCT	0.013 ± 0.007	0.001–0.023	0449 (±2:15)	0053–1100	0.00002
ACD	0.048 ± 0.026	0.015–0.099	1913 (±4:04)	1232–0258	0.0077
ACV	10.56 ± 6.15	2.11–22.06	1504 (±3:39)	0319–1832	0.0001

The authors thank Inez Hsing and Andrew Tran for their assistance in the data collection and analysis procedures.