To the Editor:
In previous articles,1–3 it was concluded that the ultraviolet (UV) light (365 nm) absorption of riboflavin depends at first linear (0.0% to 0.2%) on the concentration. The linear coefficient β, which couples the absorption coefficient μ(c)=β*c with the concentration c (in mg/mL, or %) of the riboflavin, was found to be nearly the same.1–3 However, at a specific concentration, depending on the measurement method, the linear relation turned into a constant value. This fact initiated a systematic error within the measurement methods that is present in all data sets. We describe the underlying problem and provide corrective data for the absorption of UV light depending on riboflavin concentration.
The error is caused by the dynamic range of the UV light. The light intensity that is transmitted through the absorbing layers of riboflavin (cuvette filled with riboflavin solution or tissue samples soaked with riboflavin) is, at a certain level, too small to be detected by the sensors mentioned in previous publications. The transmitted light intensity thereby depends solely on the Lambert-Beer-Law:
Considering a linear dependence of the absorption coefficient μ(c) for all concentrations of riboflavin, one can estimate the light intensity behind the tissue sample using the linear coefficient β=560 (mL/mg) cm−1 (β=560 (%) cm−1) obtained from the data of Spoerl et al.1
For a sample saturated with riboflavin solution of 0.02%, the absorption coefficient would be μ(0.02%)=11.2 cm−1 and for a solution concentration of 0.2% μ(0.2%)=112 cm−1. Assuming a UV-light source of I0=3 mW/cm2 is used for cross-linking of the cornea and an absorbing riboflavin layer of 1 mm, the resulting transmitted intensities are I(0.02%, 1 mm)=0.33*I0 and I(0.2%,1 mm)=0.000014*I0. In the second case, the resulting intensity 42*10−6 mW/cm2 cannot be detected with the light detectors used in previously published experimental setups. It may be that the ambient light, even in a dark room (1 to 10 μW/cm2), is detected and causes the constant absorption coefficient. Considering the lowest intensity that can be detected is 5 μW/cm2 and an incident light intensity of 3 mW/cm2 (I/I0= (5/3)*10−3), the maximal detectable concentration is 0.11%. Investigating higher concentrations, the light detector finds ambient light or noise, because the transmitted light is too small to be detected.
To detect transmitted light for higher concentrations (eg, 0.5% of riboflavin), one must change the experimental setup. Using a high intensity light source (365 nm, up to 90 mW/cm2) and a special designed cuvette, which allows only a fluid thickness of 100 μm of riboflavin solution, we repeated the measurements and obtained a linear relationship of concentration and absorption coefficient over the whole concentration range from 0.0% to 0.5% riboflavin with a linear coefficient of β=469 (mL/mg) cm−1 (β=469 (%) cm−1) (Fig).
Figure. A) Normalized ultraviolet light transmission for different riboflavin concentrations and B) the resulting absorption coefficient.
These results demonstrate that the absorption co-efficient depends linearly on the concentration up to 0.5% riboflavin. Previous publications may suffer from the described systematic error.
We apologize for the initial errors in our experimental setups.
Silvia Schumacher, PhD
Michael Mrochen, PhD
Eberhard Spoerl, PhD
- Spoerl E, Mrochen M, Sliney D, Trokel S, Seiler T. Safety of UVA-riboflavin cross-linking of the cornea. Cornea. 2007;26(4):385–389. doi:10.1097/ICO.0b013e3180334f78 [CrossRef]
- Wollensak G, Aurich H, Wirbelauer C, Sel S. Significance of the riboflavin film in corneal collagen crosslinking. J Cataract Refract Surg. 2010;36(1):114–120. doi:10.1016/j.jcrs.2009.07.044 [CrossRef]
- Iseli HP, Popp M, Seiler T, Spoerl E, Mrochen M. Laboratory measurement of the absorption coefficient of riboflavin for ultraviolet light (365 nm). J Refract Surg. 2011;27(3):195–201.