Background/aims To determine if significant progression of disease occurs in older, non-contact lens wearing, subjects with keratoconus and to identify potential predictive factors.
Methods Clinical and computerised corneal topography records of subjects with keratoconus attending a specialist optometry practice were retrospectively analysed to identify those aged ≥30 years, with ≥2 consultations ≥12 months apart, no contact lens wear and no corneal scarring, surgery or corneal hydrops. Topographic parameters assessed included: maximum keratometry (Kmax), steep keratometry (Ksteep), flat keratometry (Kflat), inferior-superior (I-S) ratio and the surface asymmetry and regularity (surface asymmetry index and surface regularity index) indices.
Results Of the 449 subjects with keratoconus assessed, 43 eyes of 27 patients (6.01%) met inclusion criteria, with median age 38.45 (12.86) years at baseline and median follow-up 4.36 (8.68) years. There was a significant increase in Kmax (0.30 (1.21) D), Ksteep (0.27 (0.90) D), Kflat (0.34 (1.12) D) and I-S (0.26 (0.82) D) between baseline and final review, p<0.05. Notably, 18.6%–25.6% of eyes demonstrated ≥1.00 D increase in one or more of four principal topographic parameters (Kmax, Ksteep, Kflat, I-S ratio), while 18.5%–37.0% of subjects had ≥1.00 D increase in the aforementioned parameters in at least one eye over the study period. However, <10% of eyes exhibited ≥1.00 D increase/year in all topographic parameters. The only significant predictor of progression was follow-up time.
Conclusions This study confirms that keratoconus may continue to progress beyond age 30. Older subjects with keratoconus should be monitored for progression, particularly with respect to possible corneal collagen cross-linking or astigmatic correction in cataract surgery.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Keratoconus is an ectatic, thinning disorder in which the cornea takes up a conical shape with the increased curvature inducing irregular astigmatism, resulting in visual impairment.1 Typically, as keratoconus progresses the corneal curvature increases. Traditionally, corneal curvature is measured using manual keratometers; however, these devices use just four measurements localised to the paracentral cornea and have largely been superseded by computerised corneal topography/tomography devices, which use several thousand data points over much of the corneal surface, providing a more detailed assessment.2
The traditional view of the natural history of keratoconus is onset in early teenage years with progression up until the age of 30 or 40 years when it generally arrests.1 However, the full natural history of keratoconus is yet to be determined definitively, particularly beyond age 30. Previous studies have generally involved participants with a wide age range and included many subjects wearing rigid contact lenses (CLs), the latter compromising accurate, longitudinal measurement of corneal shape.3 ,4 Keratoconus is a relatively common disease affecting 8.6–2340/100 000 of the population5 ,6; therefore, there is a considerable number of middle-aged and older subjects with mild-to-moderate keratoconus currently managed with spectacles or CLs. Better knowledge of the risk of significant disease progression in this age group would be beneficial for both those affected and healthcare providers, in terms of planning follow-up, prognosis and surgical management, including the use of collagen cross-linking (CXL).7
The aims of this study were to determine (1) if significant progression of manifest keratoconus occurs beyond age 30 in non-CL wearers, based on longitudinal changes in topographic parameters and (2) if topographic disease severity and subject-specific parameters can be used to predict the likelihood of progression.
Subjects and methods
In conjunction with the Department of Ophthalmology, University of Auckland, a retrospective review was performed of the clinical records of subjects with keratoconus who underwent consultation between 2002 and 2015 at a subspecialty optometry practice with more than 60 years of specialised provision of eye care for keratoconus (Mortimer Hirst, Auckland, New Zealand).
The investigation was focused on patients with manifest keratoconus, thus all subjects diagnosed with keratoconus, on the basis of examination by an experienced clinician using a combination of computerised corneal topography, slit-lamp biomicroscopy and refraction, were identified (N=449). The following inclusion criteria were applied: confirmed diagnosis of manifest keratoconus, age ≥30 years and at least two consultations ≥12 months apart. Exclusion criteria included forme fruste keratoconus, any previous CL wear, a history of hydrops corneae, corneal scarring, previous ocular surgery or other ocular pathology.
Main outcome measures
The following corneal topographic parameters and keratoconus indices were obtained from Medmont-E300 (Medmont, Camberwell, Australia, Medmont studio V.126.96.36.199—all captured exams were analysed with this version), Placido ring-based axial power maps, for the first (baseline) and most recent (final) reviews for all eyes that met the inclusion/exclusion criteria; apical keratometry (Kmax), steep simulated keratometry (Ksteep), flat simulated keratometry (Kflat), inferior-superior dioptric asymmetry (I-S),8 surface asymmetry index (SAI)9 and surface regularity index (SRI).10
To determine the repeatability of the parameters obtained from the Medmont-E300 in patients with keratoconus, three consecutive scans were carried out on 20 randomly selected patients with keratoconus who met the inclusion criteria of the study. However, CL wearers were included if no CL was worn for ≥2 days. The median and IQR, within subject SD (Sw), precision (1.96 Sw) and repeatability (2.77 Sw) was calculated for all parameters.
The Kolmogorov-Smirnov test revealed that all study parameters were non-parametric. The difference for all outcome measures between baseline and final reviews, for all eyes, were compared using the Wilcoxon signed-rank test (Table 1A). The Mann-Whitney U test was used to investigate differences in all outcome measures between the repeatability and main study groups to assess differences in disease severity. Eyes were subsequently grouped according to degree of progression based on the repeatability of the topographic parameters (table 1B); no progression (<1.00 D change) and significant progression (≥1.00 D increase). The rate of change in the variables that demonstrated statistically significant differences between baseline and final review was assessed and the number of eyes with significant rates of progression was determined, defined using similar criteria to progression; no progression (<1.00 D change/year) and significant rate of progression (≥1.00 D increase/year).
Univariate associations between potential predictors and changes in outcome measures were evaluated; Spearman's correlation test for continuous predictors and outcome measures and Mann-Whitney U test for dichotomous predictors; gender, age at baseline and follow-up time. The Spearman's correlation test was also used to determine if the changes in outcome measures were correlated with each other.
Factors associated with an increase of ≥1.00 D in at least one eye for all outcome measures were performed on a subject-specific basis, using the between-eye mean of eye-specific predictors, such as Kmax at baseline, in the analyses. The initial analyses included the Mann-Whitney U test for continuous predictors and the χ2 test for dichotomous predictors. All variables that were significant were then included in a logistic regression analysis to produce ORs and 95% CIs. Age at baseline and follow-up time were assessed as continuous and dichotomous predictors; age ≥40 and <40 years and follow-up time <5 and ≥5 years. A p value <0.05 was considered to be significant throughout.
Forty-three eyes of 27 subjects met the study criteria and were included. The median age at baseline was 38.45 (12.86) years, 13 (48.1%) subjects were aged ≥40 years and 14 subjects (52.0%) were male. The median age at final review was 43.21 (14.86) years. The median follow-up time was 4.36 (8.68) years with 11 (40.7%) subjects having ≥5 years follow-up.
All topographic parameters demonstrated a significant median increase between baseline and final review (p<0.05), except for SAI and SRI, which did not change significantly (p>0.600) (table 1). The repeatability analysis of the topographic parameters was conducted on 37 eyes of 20 patients and suggests that a change of ≥1.00 D is indicative of significant disease progression for all parameters except for SRI (table 1). Kmax, Ksteep and SRI were significantly higher in the repeatability cohort.
The overall median rate of change (D/year) was: 0.06 (0.28), 0.04 (0.19), 0.08 (0.15) and 0.04 (0.20), for Kmax, Ksteep, Kflat and I-S, respectively. The majority of eyes (93.0%–100.0%) exhibited no progression (<1.00 D change/year) for all topographic parameters (figure 1A). A significant rate of change (≥1.00 D/year) was observed in Kmax and Ksteep, 7.0% and 2.3% of eyes, respectively.
The greatest proportion of eyes (74.4%–81.4%) exhibited no progression over the follow-up period, with an overall change <1.00 D between baseline and final reviews for all topographic parameters. However, a total increase of ≥1.00 D in corneal topographic parameters was observed in 18.6%–25.6% of eyes (figure 1B).
Table 2 illustrates associations between potential predictors and progression in computerised corneal topographic parameters. The change in Ksteep and Kflat demonstrate significant positive correlations with follow-up time, while Kmax and Ksteep increased significantly in female subjects. There were significant moderate-to-strong positive correlations between the changes in all variables; r>0.4, p<0.01 in all cases.
Figure 2 and table 2 highlight the relationship between Ksteep and follow-up time as a dichotomous variable. A greater proportion of eyes experienced no progression (<1.00 D increase) in the group with <5 years follow-up (88.9% vs 56.3% for <5 and ≥5 years follow-up, respectively), while a greater proportion of eyes with ≥5 years follow-up experienced significant progression (≥1.00 D increase, 11.1% vs 43.7%).
Table 3 shows the predictors for ≥1.00 D increase in a topographic parameter in at least one eye. Follow-up time as a continuous variable and gender were significantly associated with Ksteep (p=0.031) and Kmax (p=0.004), respectively.
The results of the logistic regression analysis demonstrate that follow-up time is a significant predictor for ≥1.00 D increase in Ksteep in at least one eye (OR 1.38; 95% CI 1.04 to 1.84, p=0.026), while male gender was significantly associated with a reduction in the risk of ≥1.00 D increase in Kmax in at least one eye (OR 0.048, 95% CI 0.005 to 0.490, p=0.010).
The pathogenesis, mechanisms and rate of progression of keratoconus are not completely understood.11 It has been noted that there is an increase in the amount of non-enzymatic cross-linking with increasing age12 likely due to exposure to ultraviolet radiation throughout life, which is a possible mechanism by which keratoconus progression is slowed or halted with increasing age.13 Additionally, the Collaborative Longitudinal Evaluation of Keratoconus (CLEK) reported that fewer patients progressed to transplantation later in life (12%–20% aged 10–40 years vs 8%–3% aged >40 years) and that older age at baseline was protective against requiring transplantation (OR 0.72).14
The current study assessed the progression of keratoconus using quantitative computerised corneal topography, and we believe this study is unique in that it is focused on the natural history of subjects with keratoconus: (1) that had never worn CLs, (2) assessment focused on computerised topography parameters at baseline and final review (median follow-up 4.36 (8.68) years) and (3) that were typically a little older than those in investigations of keratoconus. The lack of previous CL wear avoided the potentially confounding effects of CL wear on corneal shape and possibly the keratoconus disease process. The majority of variables investigated increased significantly between the baseline and final reviews (Kmax 0.30 D, Ksteep 0.27 D, Kflat 0.34 D and I-S 0.26 D), suggesting that corneal curvature in patients with keratoconus over age 30 continues to increase. Additionally, changes in each parameter had a moderate-to-strong positive correlation with the changes in all other parameters, thus an increase in one parameter may be accompanied by a similar increase in others, suggesting that all topographic parameters investigated could potentially be used to monitor progression.
The results of this study are somewhat comparable to the landmark CLEK, which prospectively tracked changes in corneal curvature over 8 years in 1032 subjects with mean age 38.9 years at enrolment.3 Measured using manual keratometry, the CLEK reported that Kflat increased at a mean rate of 0.18±0.60 D/year.3 However, the authors suggested that progression is non-linear with the greatest rate of progression occurring earlier in life; average rate of increase in Kflat of 1.0 D/year age <20, decreasing to approximately 0.4 D/year age 20–29, 0.2 D/year age 30–39 and 0.1 D/year age ≥40.3 The median rates of change in our cohort (median age 38.5 years) were between 0.06 and 0.08 D/year for all outcome measures, similar to that reported in the CLEK for patients aged ≥40 years. Notably, over 90% of eyes in our study cohort had <1.00 D change/year, thus change occurs at a slow pace.
An increase of ≥3.00 D in Kflat over the duration of the CLEK was considered significant and occurred in 39.1% of patients aged <30 years, but 19.8% aged ≥30 years.3 The authors of the CLEK hypothesised that the natural history of keratoconus is dichotomous; the majority of progression occurring before age 35, which revealed that patients aged <35 years are more likely to have an increase of ≥3.00 D in Kflat in at least one eye (OR 3.12).3 Interestingly, in the current study using computerised corneal topographic measures, 21% of eyes in this older cohort demonstrated an increase in Kmax approaching 3.00 D.
The Medmont-E300 has been shown to be highly repeatable on normal corneas,15 however, unlike other topography instruments,16 it has not been verified on keratoconic corneas. The surface irregularity of keratoconic corneas has been demonstrated to reduce repeatability of topography devices.17 The current investigation performed a repeatability analysis on keratoconic corneas in order to produce categorical criteria for progression and found that ≥1.00 D increase in parameters indicated significant progression. A recent investigation of the natural history of corneal curvature with increasing age in normal subjects, using the Medmont-E300, found no trend towards change in corneal curvature up to age 69.18 Thus, utilisation of the repeatability to form criteria for progression in eyes with keratoconus is justified.
The criterion for progression (≥1.00 D increase) in this study was robust as the repeatability group had a greater average Kmax, which may reduce the repeatability of topography devices,19 thus, the change required for progression to be detected in the main study group was likely less than a 1.00 D increase. The current investigation elucidated that approximately 20% of eyes underwent significant progression (≥1.00 D increase in a parameter) and 80% no progression (<1.00 D increase in a parameter), while a third of patients experienced significant progression in at least one eye. While the changes in the anterior topographical parameters documented in this investigation have been verified as a means of detecting progression,20 recent evidence suggests that changes in the posterior surface may occur prior to the anterior surface.21 The Placido-based Medmont-E300 does not enable evaluation of the posterior corneal surface or corneal thickness, thereby limiting our definition of progression to changes in anterior topographical parameters.
The characteristics of subjects with keratoconus and risk factors for progression have been established by prospective observational studies such as CLEK and the Dundee University of Keratoconus Study.3 ,4 The CLEK in particular established that the greatest predictors for progression were younger age and greater disease severity at baseline.3 However, the CLEK quantified corneal shape by keratometry, not computerised corneal topography and a large number of participants wore rigid CLs, which may have caused corneal distortion/warpage. In the current study, the only significant predictor of disease progression was follow-up time; therefore, it is possible that given enough time, additional subjects in the study cohort may experience significant progression. However, the large range of follow-up durations poses a limitation in estimations of rate of progression, as theoretically eyes with longer follow-up duration might have experienced progression following the baseline visit but progression had ceased before the final visit.
In the current study, male gender appears to be protective against ≥1.00 D increase in Kmax in at least one eye (OR 0.048), although this may reflect a limitation of the sample size. Additionally, all baseline topographic parameters were higher in the group with ≥1.00 D change in at least one eye at baseline, implicating greater disease severity as a potential predictor of progression, despite none of these baseline parameters being statistically significant. Once more this may be a limitation of the sample size as a number of prior studies have reported steeper keratometry to be a significant predictor of progression.22 ,23
The sample size in the current study was limited largely due to strict inclusion/exclusion criteria limiting it to subjects with keratoconus who had never worn CLs; generally, these are subjects with mild-to-moderate rather than severe disease. Indeed, presuming that disease severity may be a predictor for progression, subjects with greater disease severity may have already progressed to CL wear or even corneal transplantation by this age, so it would be difficult/impossible to accurately assess disease progression by computerised topography. In this respect, it is notable that of 449 subjects with keratoconus assessed for this study, 94% had worn CLs, had developed corneal scarring or had progressed to corneal transplantation and only 6% of subjects met the inclusion criteria. Interestingly, keratoconus has been the most common indication for corneal transplantation in New Zealand over the last two decades.24 ,25 Changes in visual acuity, spectacle refraction and the effect of commonly known, non-corneal topographic, associations of keratoconus including family history, eye rubbing and atopy could not be investigated due to variations in recorded history, the recording of spectacle prescription and the occasional provision of reduced cylindrical correction to maximise spectacle tolerance.
Overall, this study highlights that keratoconus may continue to progress in apparently clinically stable non-CL-wearing subjects beyond the fourth decade; however, <10% of eyes may progress at ≥1.00 D/year. These data suggest that subjects with keratoconus, over the age of 30 years, should still be monitored for progression long term, as on average one in three may experience significant progression in at least one eye, although this proportion might be higher in patients with more severe disease and those with longer follow-up. Indeed, 21% of eyes exhibited a mean increase of 2.73 D in Kmax over the duration of the study. This has significant implications for the consideration of CXL, particularly in older subjects with keratoconus progressing at a significant rate, as well as the use of toric intraocular lenses in the context of cataract surgery in keratoconus.7 However, CXL may not yield a significant benefit in patients with very slowly progressing disease.
AG received support for his doctoral studies through a University of Auckland Doctoral Scholarship and a New Zealand Association of Optometrists Post Graduate Scholarship.
Contributors All the authors contributed to the design of the study. AG and GAW carried out the study, analysing patient records and collecting data. AG performed the statistical analysis of the data. All the authors contributed to writing the paper. AG produced the draft manuscript which was edited and critiqued by DVP, GAW and CNJM. All authors approved the final version.
Competing interests None declared.
Ethics approval Ethical approval for the study was obtained from the University of Auckland Human Participants Ethics Committee (approval number: 010547).
Provenance and peer review Not commissioned; externally peer reviewed.