1. Low intensity versus high intensity CXL
Arthur Cummings FRCSEd
Wellington Eye Clinic, Dublin, Ireland

2. Financial disclosure:
I have no financial interest in the materials being presented in this presentation

3. Purpose
To review the clinical data of corneal cross-linking with the IROC UV-X 2000 To determine the performance and safety of corneal cross-linking with the UV-X 2000 compared to the IROC UV-X 1000

4. Important factors in CXL
Light sources
Wavelength
Time
Energy dose
Riboflavin
Concentration
Spatial distribution
Oxygen
Reactive environment
Type of collagen
Radical absorbers
Additives

5. Main principles of today‘s photooxidative CXL
The photooxidative cross-links are induced by a photosensitizer (Riboflavin) and UV-light
UV-light
3O2
1O2 +
Tissue molecule +
Cross-linked tissue molecule
Riboflavin
3RF*
triplet
Riboflavin
singlet
1RF*
Riboflavin RF

6. Induction of Cross-links
Activation of photosensitizer
induce
1O2
3O2
decay
Molecule products
interaction RF
Additives
Generation of singlet oxygen
The induction of cross-links depends on:
Amount of singlet oxygen molecules
Probability of inducing cross-links
Induction of cross-links

7. Rate equation for generation of cross-links
Maximize
Minimize
Is it possible to shorten the CXL-treatment?
Is it possible to make the CXL-treatment more effective?

8. Rate equation for generation of cross-links
Maximize
Minimize
Factors which can be influenced
Concentration
Intensity
Energy dose (time)
Additives
Glucose
D2O
Sodium acids 11

9. Intensity
CXL-Congress 2008, R. Krueger & E. Spörl
Schumacher et al., IOVS, 2011
10 mW/cm2
9 min
3 mW/cm2
30 min
Control
2 mW/cm2
45 min
3 mW/cm2
30 min
10 mW/cm2
9 min
15 mW/cm2
6 min
Further increase possible ?
CXL induces an 1.3/1.5 fold increase in corneal stiffness. No statistical differences between doses. 12

10. Oxygen depletion
Higher intensities lead to a faster depletion of oxygen
Diffusion of oxygen into the cornea takes several minutes
Missing reactive partners in short time might lead to a reduced effect
Are we inducing hypoxia with CXL?

11. Summary CXL is a complex physical process Consider concentration of
Photons
Riboflavin molecules
Oxygen
Reactive partners (collagen, …)

12. Biomechanical changes in standard and rapid CXL
Jeremy Wernli1
Silvia Schumacher1
Eberhard Spörl2
Michael Mrochen1
Biomechanical changes in standard and rapid CXL

13. Purpose What is the problem? Progressive keratectasia
UV-light corneal-crosslinking (CXL)
Standard treatment parameters
3 mW/cm2 for 30 minutes
Disadvantages
Patient’s comfort
Surgeon’s patient throughput
Goal
Shorter treatment time!
Question
How far can we go?
VISION - OPTICS - LIGHT we focus your ideas!

14. Bunsen reciprocity law: Energy dose of standard protocol:
Purpose
Bunsen-Roscoe reciprocity law:
Bunsen reciprocity law:
Photochemical processes depend
on the absorbed energy dose
Energy dose:
Energy dose = Intensity x Time
Energy dose of standard protocol:
3 mW/cm2 x 30 min = 5.4 J/cm2
3mW/cm2
30 min
10 min
9mW/cm2
30mW/cm2
3 min
90mW/cm2
1 min
Same CXL effect in theory…really?
VISION - OPTICS - LIGHT we focus your ideas!

15. Purpose …really? ok ok ? Established standard protocol
3mW/cm2
30 min
10 min
9mW/cm2
30mW/cm2
3 min
90mW/cm2
1 min
Same CXL effect in theory… really? ok ok ?
Established standard protocol
Scientific proof of biomechanical strengthening effect is insufficient in the literature
Schumacher et al., IOVS, 2011
Let’s prove! (ex vivo first)
VISION - OPTICS - LIGHT we focus your ideas!

16. Methods Experimental study design
Cross-linking of ex vivo porcine corneal tissue
In compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research
180 porcine eyes:
Randomly assigned to 10 different treatment groups with different CXL illumination intensitites
ranging from 3 to 90 mW/cm2 and correpsonding illumination times from 30 minutes to 1 minute
constant energy dose of 5.4 J/cm2
To each of theses groups control eyes of the same date were tested, which were not irradiated with UV light, but otherwise underwent the same procedure
Two main steps:
Cross-linking procedure
Stress-strain measurement
Stiffness (Young‘s modulus) as a derivative of the stress-strain curve at 10% strain
Nonparametric Kruskal-Wallis test for the comparison between groups (not normally distributed population)
VISION - OPTICS - LIGHT we focus your ideas!

17. Comparison of the different groups to the control group
Results
Comparison of the different groups to the control group
Table 2 - Post hoc analysis of Kruskal-Wallis test. All treatment groups (‘Group B’) were compared against the control group (‘Group A’). Q represents the calculated values which are compared to Q(0.01) from the Kruskal-Wallis test.
VISION - OPTICS - LIGHT we focus your ideas!

18. Results Relative stiffness increase
Average increase in stiffness of all groups follows a typical threshold function (Boltzmann function)
Function dips at about 40 mW/cm2
50% limit is associated at approximately 47mW/cm2 ± 1.5mw/cm2
Figure 4 – Stiffness increase of all treatment groups compared to control group. The second x-axis at the top indicates the corresponding irradiation times to maintain a constant energy dose of 5.4 J/cm2
VISION - OPTICS - LIGHT we focus your ideas!

19. Discussion Why does the Bunsen-Roscoe law fail?
We can only speculate at this time
Hypothesis:
Suppression of initiating processes at high intensities
Oxygen availability for reaction
Riboflavin bleaching
Activation of termination processes at high intensities
Reaction of activated Riboflavin
Oxygen radical
Other substances pre-existing in the cornea (e.g. Vitamin C as quencher)
VISION - OPTICS - LIGHT we focus your ideas!

20. Optimized beam profile
Peripheral intensity 30% higher than ENERGY DOSE (not intensity) =
Increased biomechanical strengthening of the cornea
Improved corneal flattening

21. Corneal cross-linking procedure
Abrasion of corneal epithelium
Application of Riboflavin solution to cornea
Illumination of cornea with UV-light

22. Standard and high intensity procedure
Standard protocol UV-X 1000
3 mW/cm2 for 30 minutes
5 min min min 30 min
Diffusion of Riboflavin
Illumination with UV light
Abrasion
Check diffusion
High intensity protocol UV-X 2000
9 mW/cm2 for 10 minutes
Abrasion
Check diffusion
Diffusion of Riboflavin
Illumination with
UV light
5 min min min min

23. UV-X 2000: Optimized beam profile
Hypothesis: more flattening of the cornea due to optimized beam profile

24. UV-X 2000: Optimized Beam Profile
Crosslinking of more corneal tissue
Post OP comparison of demarcation line shows more tissue crosslinked in the periphery with the UV-X 2000

25. Center treatment on thinnest point
Treatment should be centered on the thinnest point with UV-X 2000
to avoid direct illumination of limbus
Protect limbus
When doing a treatment it is important to center the the treatment on the thinnest point and not on the center of the pupil. In keratoconic eyes the thinnest point is mostly not the center of the pupil
Dislocation of the cone and the thinnest point
Seiler et al. (2010)

26. Respect corneal thickness profile
Corneal thickness profiles are different for normal and keratoconus cornea. Ambrosio et al. (2006)

27. Results for endothelium safety (human donor eyes)
10 mW / cm2; 9 min
Pair
Untreated
Treated
24 hours before
24 hours after
M1529B
2800
2900
M15309
2600
2400
M15457
3150
3000
No statistical difference in endothelium cell count or apoptosis
Apoptotic cell appear green
Only the anterior region of the stroma showed apoptosis
No apoptosis was found in the endothelium
Untreated controls had no UV but otherwise were treated in the same way as the treated corneas.
Hilarby at al. 2011, University of Manchester

28. Comparing different UV-A Intensities
IOP = 8mm Hg
Control
3mW
10mW
30mW
100mW
Backward SH Imaging

29. Comparing different UV-A Intensities
IOP = 11mm Hg
Control
3mW
10mW
30mW
100mW
Backward SH Imaging

30. Comparing different UV-A Intensities
IOP = 16mm Hg
Control
3mW
10mW
30mW
100mW
Backward SH Imaging

31. Comparing different UV-A Intensities
No IOP measurement
Control
3mW
10mW
100mW
Backward SH Imaging

32. Comparing different UV-A Intensities
No IOP measurement
Control
3mW
10mW
Forward SH Imaging

33. Clinical Outcomes Ethics Committee Approved Study
Collecting UV-X 2000 data
Comparing to database of UV-X 1000 cases and new UV-X 1000 cases performed during the clinical trial
Recruitment has stopped
Analysis ongoing

34. Outcome measures Performance: Safety: Adverse events Change in K-Max
Change in BSCVA
Endothelial cell loss
Adverse events

35. Data source Clinical data collected by:
A. Cummings, Wellington Eye Clinic, Ireland
T.S. Seiler / T. Koller, IROC Clinic, Switzerland
F. Raiskup, TU Dresden, Germany
June 2012:
 92 eyes treated
 33 eyes available for interim data analysis
Up to 12 months follow-up

36. Demographic data Number of eyes (OD/OS): 33 (18/15) (55%/45%)
Gender: % male 18 % female
Age (years):
mean: 26.2 ± min: 15 max: 40
Kmax (D) (Patients with pre-op Kmax > 65 D were excluded)
mean: 52.9 ± 5.8 min: 41.8 max: 64.0
Sphere (D) (11 eyes):
mean : 0.6 ± 1.7 min: max: 3.75
Cylinder (D) (11 eyes):
mean: -3.4 ± 2.4 min: max: -0.5
Corneal thickness (micron) (21 eyes):
mean: 477 ± 39 min: 392 max: 533

37. Comparison UV-X 1000 and UV-X 2000
Interim data analysis for UV-X 2000 at 6 months
Status
Kmax (D)
Number of eyes (%)
Flattening
< -1 D
14 (44.8)
Same
from -1 D to 1 D
15 (46.9)
Steepening
> 1 D
(9.4)
UV-X 2000 (32 eyes):
After 6 months, progression is halted in 90.6 %.
Status
Kmax (D)
Number of eyes (%)
Flattening
< -1 D
39 (37.1)
Same
from -1 D to 1 D
58 (55.2)
Steepening
> 1 D
(7.6)
UV-X 1000 (117 eyes):
After 12 months, progression is halted in 92.4 %.
Comparable results
Higher percentages of eyes are flattened with UV-X2000
Equal rate to stop progression after 6 months UV-X2000 compared to
12 months UV-X 1000
We expect further improvement of flattening over time for UV-X 2000

38. UV-X1000 improvements over time
-0.75D
-1.5D
-2.0D
-2.5D
We expect further improvement for UV-X 2000 from 6 to 12 months as known from UV-X1000 results
Raiskup et al. 2008

39. Performance – Change in Kmax
Decreased
Unchanged
Increased
7 eyes show a strong flattening effect at 6 months

40. Performance – Change in Kmax
More flattening for higher Kmax values

41. Performance – Change in Kmax
Progression of corneal steepening
6 month follow-up in 32 eyes
Mean Kmax = -1.0 D
Status
Kmax (D)
Number of eyes (%)
Flattening
< -1 D
14 (44.8)
Same
from -1 D to 1 D
15 (46.9)
Steepening
> 1 D
(9.4)
Approx. 45% of the eyes are significantly flattened
In more than 90 % the progression of ectasia has stopped

42. Safety – Change in BSCVA
Mean increase in BSCVA by 1 or more lines: 36%
No eye lost more than 2 lines
Safety is comparable to published data. Koller / Raiskup
80% of eyes are equal or improved in BSCVA compared to pre-OP
3 eyes lost two lines of BSCVA at 6 months
None of the eyes lost more than 2 lines

43. Safety – Change in BSCVA
UV-X 2000: 0 eyes lost more than 2 lines (this data summary)
min: 1-2% lost more than 2 lines (Raiskup et al. 2008)
UV-X 1000: 1-2 % lost more than 2 lines (Hersh et al. 2011)
Comparable safety for UV-X 2000

44. ECC: Comparison UV-X 1000 vs UV-X 2000
Number of eyes
Follow-up in months
Cell loss?
Statistitical significance of endothelial cell loss
Percentage of endothelial cell loss 8 6
N.S. No
UV-X 2000:
After 6 months:
No statistically significant cell loss reported
Number of eyes
Follow-up in months
Endothelial cell loss in % or reported damages?
Statistitical significance of endothelial cell loss
Percentage of endothelial cell loss 88 52
Yes 2% 51 12
No cell loss No
N.S. 24 66 36 30 18 14
Standard:
After 12 to 52 months:
Percentage of endothelial cell loss is less than naturally occurring 1% per year

45. AE: Comparison UV-X 1000 vs UV-X 2000
UV-X 2000 (92 eyes): 3.3% of cases
Corneal edema: 1
Corneal edema with Descemet folds: 2
Standard (more than 2800 eyes): 3.5% of cases
Corneal scarring
Sterile infiltrates
Delayed epithelium healing
Keratitis
Anterior chamber reaction
Corneal edema
Comparatively low incidence of adverse events

46. Risk of UV damage
UV-X 2000 has equal total energy dose compared to UV-X 1000
Thus, lens and retina get the same theoretical low UV light dose

47. Add……..
Graphs from WEC comparing AXL to CXL off IBRA

48. WEC Clinical Data
AXL
CXL

49. Typical clinical observation UV-X 2000
Corneal haze ring can be found in the mid-periphery of the cornea.
Similar to UV-X 1000, the haze is reduced within the first 6 months

50. Summary Performance: Strong flattening effect
More flattening for steeper corneas
Approx. 45% of the eyes are significant flattened
In more than 90 % the progression of ectasia has stopped
Further improvement expected in corneal flattening from 6 to 12 months follow-up
Safety:
Comparable safety regarding vision & endothelium
Comparable theoretical risk for UV damage of lens & retina
Adverse events:
Comparable low rate of adverse events

51. Conclusion The UV-X 2000 provides: better performance,
equal safety and
a shorter treatment time
compared to standard corneal cross-linking procedures.

52. Ongoing Studies at the WEC
Epi-ON CXL (CXLO)
Conductive Keratoplasty (CK) + CXL

53. 4th Keratoconus Experts Meeting
Pre-ESCRS
One item on agenda
Define parameters to measure keratoconus progression and regression
John Kanellopoulos
ISV
IHD

54. 9th CXL Meeting
6-7 December 2013
Dublin, Ireland

55. Thank you for your attention