What are the costs, capacity, and clinical implications of 'waiting for documented progression' in young West of Scotland patients prior to collagen cross linking?

Collagen crosslinking (CXL) has emerged as the standard intervention to arrest progression of corneal ectasia [1-5]. Since 2015, young patients with keratoconus in NHS Greater Glasgow and Clyde (NHSGGC), Scotland, have been monitored every 4–6 months in a dedicated optometrist-led early keratoconus clinic. Those with documented evidence of progression are offered CXL, a cost-effective procedure intended to stabilise the cornea and avoid future corneal transplantation [6, 7]. CXL has been shown to be effective in arresting the progression of corneal ectasia in this younger age group at 2 and 5 years follow up [1, 8–10]. Progression of keratoconus is known to be more aggressive in the younger age group, so CXL services were protected and prioritised in many corneal units during the COVID-19 pandemic [11]. These measures were necessary as rapid topographic changes may occur and result in corneal thinning progressing beyond standard treatment protocols (corneal thickness <400 μm). [12] In light of this, the advocacy for early intervention with CXL at presentation rather than ‘wait for documented progression’ is growing for the younger age group. We wished to identify our younger patients with keratoconus and model the real-world clinical, financial costs, and service capacity implications of ‘waiting for documented progression’ prior to CXL.

Methods

Electronic records of patients <21 years of age at referral attending our early keratoconus clinic in NHSGGC between January 2015 and September 2019 with a minimum of 1 year follow up were retrospectively reviewed. Demographic, topography, visual acuity, clinic attendance, and intervention data were analysed.

Results

We identified 76 patients (56 males and 20 females) with 142 eyes which had not previously undergone CXL. Mean age was 18 years (median 18, SD 1.91, range 14–21). Six patients had advanced disease with corneas considered too thin for conventional CXL at presentation (thinnest pachymetry <400 μm). Sixty (41%) eyes of 46 patients demonstrated documented progression. One patient’s cornea progressed during monitoring intervals such that CCT was too thin for conventional CXL. Three patients declined treatment, leaving 56 eyes (39%) of 42 patients proceeding to epithelium-off, accelerated, pulsed CXL. Twenty-nine right eyes and 27 left eyes were treated. Fourteen patients had each eye treated sequentially. There was no statistical difference in the mean age of patients with progression leading to CXL compared to those that did not receive CXL in the study period (17.8 vs 18.2 years, p = 0.48). Those that went on to require CXL had significantly more extreme topographic measurements at first presentation compared to those who simply required monitoring (mean CCT 470.71 um vs 500.00 um (p < 0.01) and mean Kmax 57.64D vs 52.06D (p < 0.01)).

The median time interval from initial presentation to CXL treatment was 40 weeks (mean 57; SD 48, range 3–214). In this time period, there was evidence of significant visual loss (p = 0.03) and topographic progression (p < 0.01) [See Table 1]. Mean length of time from listing to receiving CXL was 93 days (median 77.5, range 13–386). These 42 patients (56 eyes) attended an overall total of 196 clinic appointments prior to first eye CXL treatment (median 2; mean 3.5; range 1–7). There were 117 appointments for the 34 patients that did not receive CXL in either eye during this study period.

Table 1

Presenting Pentacam topographic characteristics and visual acuity of 56 young eyes with keratoconus that underwent CXL following documentation of progression, with initial presentation, prior and post CXL treatment measurements.

	Presentation (±SD)	Prior to CXL (±SD)	P value	Post CXL (±SD)	P value
Mean Cyl (D)	2.80 ± 1.85	3.84 ± 2.04	<0.1a	3.60 ± 2.19	0.19a
Mean spherical equivalent (D)	−0.53 ± 1.76	0.18 ± 1.82	0.32a	−0.09 ± 1.96	0.39a
Mean Anterior K2 (D)	49.01 ± 4.04	50.17 ± 4.60	<0.01a	51.57 ± 5.15	<0.01a
Mean Posterior K2 (D)	−7.41 ± 0.77	−7.53 ± 0.80	<0.01a	−7.75 ± 0.88	<0.01a
Mean CCT (μm)	470.71 ± 42.5	457.64 ± 29.04	<0.01a	450.19 ± 32.7	0.03a
Mean Kmax (D)	57.64 ± 7.36	58.72 ± 7.13	<0.01a	60.31 ± 7.96	<0.01a
Visual acuity
Median LogMAR	0.2	0.3		0.3
Mean LogMAR	0.16 ± 0.19	0.29 ± 0.24	<0.01a	0.28 ± 0.24	0.01b
Mean change in lines of Snellen VA		−0.79 ± 0.89	<0.01b	−0.50 ± 0.79	0.01b

SD standard deviation, CXL collagen cross linking, D dioptre, CCT central corneal thickness, BCVA best corrected visual acuity.

aPaired two-sided t-test.

bOne-sample t-test.

Discussion

Early intervention with CXL at first presentation remains controversial, as the procedure carries the potential for complications that may result in reduction in vision. These include development of corneal stromal haze, microbial keratitis, scar formation, and failure of treatment resulting in continued ectasia progression [13, 14]. In our cohort, two eyes out of 56 (3.6%) developed keratitis following CXL. However, these uncommon risks need to be balanced with the potential benefits of arresting progression and eventual avoidance of corneal transplantation. The average duration between initial presentation and CXL treatment was over 1 year in our cohort, and we also identified significant time delays between time of listing to delivery of treatment, which is a further source of unrecognised progression. Variability in, or unreliable, topographic readings can also delay detecting progressive disease, which increases the risk of ‘silent’ progression during monitoring and can influence the decision to treat [13, 15, 16].

We found 41% of our eyes had documented progression over the 4-year period, comparable to a large Dutch study where 230 of their 504 monitored eyes (46%) underwent CXL [17]. Young patients with keratoconus are likely to have unrecognised visual and topographic changes prior to initial ophthalmic referral. For this reason, previous publications have proposed CXL treatment at time of initial presentation for young patients with keratoconus [5, 18, 19]. Similar to the modern glaucoma management philosophy (treatment before damage), we agree that patients under 21 years of age with keratoconus should be offered CXL at initial presentation, rather than delaying treatment until evidence of disease progression. Applying this policy change to our young Glasgow cohort, the 56 eyes that underwent CXL could have avoided significant topographic deterioration or mean VA loss of 0.79 lines (median loss of 1 line). If all 56 eyes requiring CXL were treated at first presentation without monitoring for progression, 140 consultations would have been avoided in this cohort, with a conservative potential cost saving of £18,900 (clinic appointment cost £135; local NHSGGC estimate). However, this policy would have resulted in 28 potentially unnecessary CXL procedures (and exposure to the 3% risk of complications) if our whole cohort had their first/worse eye treated at presentation [13]. We have shown that young patients who progress have more extreme topography at presentation. More validated research is required to identify parameters at presentation to better identify those patients who are likely to experience progression and who require immediate CXL. Regardless of any change in treatment policy, the impact of stopping eye rubbing behaviour should not be underestimated and is a viable treatment option during any monitoring period [19-21]. Maximising clinical capacity has been a priority during the present COVID19 restrictions and will remain a challenge for the foreseeable future. In light of this, we propose that the financial costs, clinical capacity, and visual implications of requiring documentation of keratoconus progression prior to CXL should be revisited in the younger population.