Data Collection Procedures
The primary outcome measure for both trials was change in spherical equivalent refractive error as determined by cycloplegic autorefraction. After instillation of a topical anesthetic (proparacaine), cycloplegia was achieved with two drops of tropicamide 1% spaced 5 min apart. Thirty minutes after the second tropicamide instillation, refractive data for right and left eyes were obtained by recording the mean of five consecutive readings taken with an open-field instrument (NVision-K5001, Shin-Nippon, Japan). The secondary outcome measure for both trials was the change in axial length as measured by partial coherence interferometry (IOLMaster, Carl Zeiss Meditec, Germany).
Cycloplegic subjective refraction formed the basis for prescribing a full correction of each subject's myopia, and occurred at the first visit, along with acquisition of the primary and secondary outcome measures. Assigned spectacles were dispensed at visit 2, typically 2 to 3 weeks later, and during this interval, the majority of subjects wore their own single-vision spectacles. At the second visit, baseline data not requiring cycloplegia (e.g., phorias and peripheral refraction with and without assigned spectacles) were collected. The target dates for the 6- and 12-month visits, each with a window period of ±14 days, were calculated from the day of visit 2. This structure was adopted to minimize the number of times cycloplegic drops were administered in an effort to limit subject dropouts. However, differing times between visits 1 and 2, plus variations caused by attending the 6- and 12-month visits earlier vs. later in the window periods resulted in some variation in the true duration of each “6-month” period.
To address the problem of variations in the precise duration of the first and second “6-month” periods, a procedure was used similar to that previously reported.9,10 For each of the subjects, rates of myopia progression for the visit 1 baseline to 6-month visit, and for the 6-month to 12-month visit were calculated by dividing the spherical equivalent changes in autorefraction by the actual number of days in the “6-month” periods, and then multiplying the result by 182.5.
Progression rates were defined for the first and second 6 months of the 12 months studied. Only right eye data were used. Data were classified in terms of “summer,” “autumn,” “winter,” or “spring,” based on the mid-point of the 6-month period between visits. Periods with mid-points falling in the months of June through August were classified as summer, September through November as autumn, December through February as winter, and March through May as spring. As may be seen by inspecting Fig. 1, trial A commenced in mid-autumn and the 6-month data were collected in the following spring. All of winter was included in this period. All of summer was included in the second 6-month period of trial A. Trial B commenced in mid-summer, and the 6-month data were collected in the following winter. Similar proportions of summer and winter were represented in both the first and second 6-month periods of the trial, with mid-points of each period falling in autumn and spring, respectively. Difference in progression between seasons was analyzed using linear mixed models with subject random intercepts to account for repeat observations within a subject arising from multiple visits and eye-specific data. Age at baseline was added as a covariate in all analyses. Post hoc multiple comparisons were adjusted to an overall 5% level of significance using Bonferroni correction.
Change in Spherical Equivalent
As shown in Table 3, control subjects in trial A exhibited a significantly greater mean myopia progression than those in trial B at the 6-month visit (p = 0.011). However, in the ensuing 6 months, mean progression slowed markedly in trial A, while continuing at a similar rate in trial B. At the 12-month visits, mean myopia progression from baseline levels was similar for both trials (p = 0.680).
Mean, normalized, 6-month spherical equivalent myopia progression in terms of summer, autumn, winter, and spring months for trial A, trial B, and combined trial A and trial B subjects are shown in Table 4. Note that the bulk of the summer/winter data were obtained from trial A, and the bulk of the autumn/spring data were obtained from trial B. In trial A, mean progression for summer was significantly less than that for winter (p < 0.001). In trial B, there was no difference between progression in the four seasons (p = 0.238); however, the sample sizes for winter, and particularly for summer, were small. For combined trial A and trial B data, post hoc analysis showed that mean myopia progression in summer was significantly less than for winter months (p < 0.001), but there was no difference in progression between that of autumn and that of spring.
Change in Axial Length
As shown in Table 5, there was no significant difference between mean axial length increase in trial A and trial B at the 6-month (p = 0.190) or 12-month (p = 0.639) visits.
Mean, normalized, 6-month increase in axial length in terms of summer, autumn, winter, and spring months for trial A, trial B, and combined trial A and trial B subjects are shown in Table 6. Again, most of the summer/winter data were obtained from trial A, and most of the autumn/spring data from trial B. In trial A, mean axial elongation for summer was significantly less than that for winter (p < 0.001). In trial B, contrasting with the similar changes in spherical equivalent during autumn and spring, post hoc analysis showed mean increase in axial length during spring months was significantly less than that during autumn (p < 0.001) and winter (p = 0.036). Post hoc analysis for combined trial A and trial B data again showed mean increase in axial length during spring months was similar to that during summer (p = 1.000), but significantly less than that during autumn (p < 0.001) and winter (p < 0.001).
Correlation between Change in Spherical Equivalent Refraction and Axial Length
Figures 2a and b are scatterplots of 6-month change in axial length against 6-month change in spherical equivalent for the first and second 6-month periods of each trial. The coefficients of determination (R2, the increase in spherical equivalent myopia that is statistically explained by the increase in axial length) for the first and second 6-month change in trial A were 48% and 33%, respectively. Both were statistically significant (p < 0.001). For trial B, the corresponding coefficients were 57% (p < 0.001) and 8% (p = 0.064). It was also observed that the slopes of the relationship between change in spherical equivalent and change in axial length were different between the two 6-month periods for trial B (p < 0.001), but not for trial A (p = 0.218).
Our finding of increased mean myopia progression rates in winter months as compared with those in summer in conventionally corrected, myopic Chinese schoolchildren is in agreement with recent studies on North-American school-age children wearing single-vision spectacles.9–11 Although hours of sunlight, ambient temperature, or some other environmental factors that vary with the seasons could contribute to this phenomenon, the authors of those reports suggest that the most likely explanation was the difference in the amounts of close work performed during these periods. However, the true impact of near work alone is difficult to assess, because most near work is performed indoors, and at illumination levels far lower than those generally encountered during outdoor activities.
Summer school vacation in North America typically lasts 12 weeks, with a much shorter, 2-week Christmas break during winter. The academic year is broken up somewhat differently in Guangzhou, with 2 months of summer vacation during June and July, and 1 month of winter vacation from mid-January to mid-February. In comparison with the higher latitudes of Indiana and northeastern Oklahoma where the aforementioned North-American studies were conducted, there are comparatively smaller differences in temperature extremes and hours of daylight between the seasons in Guangzhou (Tables 7 and 8). Guangzhou has a more temperate climate in mid-winter, and spending time outdoors in Guangzhou may even be more appealing in the winter vacation than in the summer break. Guangzhou has more daylight hours in mid-winter, with the maximum difference in daylight hours between January and June in Guangzhou being roughly half that of Bloomington, with Tahlequah's difference closer to that found in Bloomington.
Myopia progression data from all three North-American studies were presented for 6-month periods classified as “summer” in which all of the summer vacation was included, or as “winter” or “school” in which the children attended school for the entire 26 weeks, with the exception of the 2-week break at Christmas. Separate studies reported progression rates for one summer and two winters,9 for two summers and three winters,11 and for a variable number of follow-up seasons.10 For the purposes of comparison with our Guangzhou data in which a single winter was followed by a single summer, Table 9 presents data only for the first winter and first summer of the three North-American papers. The Goss and Rainey results are from a single report10 based on pooled data from two longitudinal studies, both of which included some subjects for whom the first 6 months were classified as winter, whereas for other subjects the first period was summer. Data for those subjects for whom the first winter preceded the first summer are tabulated in the “WS” column, whereas those for which the first summer preceded the first winter are listed under “SW.” The Fulk and Cyert (1996) and Goss and Rainey papers reported summer and winter rates in terms of diopters per year. To facilitate comparisons between studies in Table 9, the reported rates for these studies were divided by a factor of 2.
The percentage calculated for Goss and Rainey “WS” shows substantially less difference in myopia progression for summer vs. winter than do the other North-American results. The reasons for this apparent anomaly are not clear, but possible explanations include both subject and methodological differences. The progression rates for our Chinese children were greater in absolute terms than the rates exhibited by the North American children. In this respect, it may require stronger protective stimuli to slow the faster rates of progression. In addition, the small sample size of the “WS” data and the methods used to monitor progression may have contributed to the observed differences. In the two studies reported by Goss and Rainey, progression data were based on serial changes in subjective refraction, rather than by changes in the means of multiple autorefraction measurements, as was used in the other two North-American studies.9,11 Moreover, the decision to not use cycloplegia for the subjective refractions might have been a contributing factor, although for some subjects, cycloplegic retinoscopy was used to verify the subjective findings.
Although our results suggest a decrease in the difference between summer and winter myopia progression rates in our Guangzhou subjects compared with those found in the North-American studies, this does not necessarily strengthen the case for illumination being a factor in larger seasonal differences in myopia progression. A recent report on education in China noted that, as well as in-school education, supplementary academic coaching was becoming the norm, and in 2006, 80% of elementary school students in Guangzhou were receiving extracurricular education.12 Data on time spent indoors vs. outdoors, or time engaged in near work that relates specifically to the summer and winter school vacations, were not collected for the subjects in our Guangzhou studies. Accordingly, we cannot rule out the possibility that a substantial part of the winter vacation was spent performing indoor, close work, and that our Chinese children spend proportionately the same time outdoors in winter vs. summer as do their North-American counterparts.
A further issue that cannot be addressed by our data is precisely when the pace of myopia progression is decreasing during the period classified as summer. Although it may be intuitive to assume that it occurs during the 2-month summer vacation, a Singaporean study13 casts doubt on this notion. Based on cycloplegic autorefraction performed five times over a 10-month period, it was found that, in 7-year olds, there was a significant increase in the rate of myopia progression after the academically important final examinations. This period of increased progression coincided with the summer vacation at the end of the academic year. The authors speculate that the enhanced stimulus for myopia progression generated by the increased pre-examination study load requires time before it manifests as an escalation in the rate of ocular growth. It is reasonable to surmise that the heavy close work load also means that students are spending very little time outdoors while studying, and it is possible that it is this factor, rather than close work per se, that is the catalyst for subsequent accelerated myopia progression. Thus, it is possible that most of our reported reduction in the summer rate of myopia progression actually occurs during the first month of school after the summer vacation, rather than during the vacation. An inspection of Fig. 1 confirms that such a scenario would still be consistent with our finding of reduced myopia progression in summer months.
The mixed outcome in correlation (significant for trial A and first 6 months of trial B, but not for second 6 months of trial B) between change in spherical equivalent myopia and change in axial length suggests that there are other factors also involved in myopia progression. This is highlighted by the finding that, although there was no difference in the change in spherical equivalent myopia during autumn and spring, there was a significant difference (p = 0.036) in axial elongation for autumn (0.24 ± 0.09 mm) compared with that of spring (0.15 ± 0.08 mm). The reason for this apparent anomaly is unclear; however, it may be that there is some mechanism, such as crystalline lens thinning,14 that attempts to compensate for the sudden increase in axial growth presumably induced by the return to school after the summer vacation.
Our findings provide additional evidence that temporal variations in the myopia progression of children occur during the course of a school year. Currently, it is unclear whether this phenomenon is a result of seasonal differences in the amount or intensity of close work, or due to factors related to being indoors vs. outdoors, such as differences in ambient light levels. The physiological pathway by which higher ambient light levels might modulate ocular growth is unclear, but increased levels of light-induced retinal dopamine15 and maintenance of adequate concentrations of cutaneously derived vitamin D in the blood have been hypothesized.16 Further work is needed to determine the relative influences of these factors, and to define more precisely the temporal patterns of myopia progression. Regardless, the fact that there are seasonal variations in the rate of myopia progression emphasizes that studies of potential myopia treatment strategies should be at least 12 months in duration to take seasonal variations into account. Moreover, determining the optimum timing and duration of such interventions may increase our ability to reduce myopia progression in schoolchildren.
Brien Holden Vision Institute
Level 4 Rupert Myers Building, Gate 14, Barker Street
Kensington NSW, 2033
The authors thank Judith Flanagan, PhD, for assistance with preparation of the manuscript. This study is funded by grants from the Brien Holden Vision Institute and the Australian Government's CRC scheme through the Vision CRC.
1. Grosvenor TP, Goss DA. Clinical Management of Myopia. Boston: Butterworth-Heineman; 1999.
2. Pärssinen O, Lyyra AL. Myopia and myopic progression among schoolchildren: a three-year follow- up study. Invest Ophthalmol Vis Sci 1993;34:2794–802.
3. Jones LA, Sinnott LT, Mutti DO, Mitchell GL, Moeschberger ML, Zadnik K. Parental history of myopia, sports and outdoor activities, and future myopia. Invest Ophthalmol Vis Sci 2007;48:3524–32.
4. Jones-Jordan LA, Mitchell GL, Cotter SA, Kleinstein RN, Manny RE, Mutti DO, Twelker JD, Sims JR, Zadnik K. Visual activity before and after the onset of juvenile myopia. Invest Ophthalmol Vis Sci 2011;52:1841–50.
5. Luckeish M, Moss FK. Ocular changes in school children during the fifth and sixth grades. Am J Optom 1939;16:443–50.
6. Kephart NC. Visual changes in children associated with school experience. Am J Optom Arch Am Acad Optom 1950;27:195–9.
7. Kephart NC, Unger RM. An investigation of visual performance among selected groups of school children. Australasian J Optom 1954;37:302–7.
8. Hirsch MJ. Effect of school experience on refraction of children. Am J Optom Arch Am Acad Optom 1951;28:445–54.
9. Fulk GW, Cyert LA. Can bifocals slow myopia progression? J Am Optom Assoc 1996;67:749–54.
10. Goss DA, Rainey BB. Relation of childhood myopia progression rates to time of year. J Am Optom Assoc 1998;69:262–6.
11. Fulk GW, Cyert LA, Parker DA. Seasonal variation in myopia progression and ocular elongation. Optom Vis Sci 2002;79:46–51.
12. Li W, Li Y. An analysis on social and cultural background of the resistance for China's education reform and academic pressure. Int Ed Studies 2010;3:211–5.
13. Tan NW, Saw SM, Lam DS, Cheng HM, Rajan U, Chew SJ. Temporal variations in myopia progression in Singaporean children within an academic year. Optom Vis Sci 2000;77:465–72.
14. Zadnik K, Mutti DO, Fusaro RE, Adams AJ. Longitudinal evidence of crystalline lens thinning in children. Invest Ophthalmol Vis Sci 1995;36:1581–7.
15. McCarthy CS, Megaw P, Devadas M, Morgan IG. Dopaminergic agents affect the ability of brief periods of normal vision to prevent form-deprivation myopia. Exp Eye Res 2007;84:100–7.
16. Mutti DO, Marks AR. Blood levels of vitamin D in teens and young adults with myopia. Optom Vis Sci 2011;88:377–82.
myopia; myopia progression; seasonal variation© 2012 American Academy of Optometry