In our current study, using a prospective study design with a higher number of included patients, we confirm our previous findings of a stable vr-QoL in KC-patients after CXL [11]. These results are in line with the findings of other studies analyzing the vr-QoL in progressive KC after CXL using the NEI-25 [20, 21]. Kandel et al published a comprehensive review about the measurement of QoL in KC clearly emphasizing the importance of using a comprehensive and high-quality patient-reported outcome to achieve reliable results. The NEI-25, developed in 2001, “was designed to measure the dimensions of self-reported vision-targeted health status that are most important for persons who have chronic eye diseases. The content was identified during a series of condition-specific focus groups with patients who had age-related cataracts, glaucoma, age-related macular degeneration, diabetic retinopathy, or CMV retinitis” [19, 22]. Next to the content-related inaccuracies when using a questionnaire originally designed by analyzing patients with different pathologies, also the method of designing the questionnaire is important to achieve reliable results.
Many questionnaires like the NEI-25 use traditional Likert (summary) scoring without the application of statistical validation derived from item response theory [23]. Important components of these statistical validation options include Rasch weighting, which converts standard Likert scales into linear data and helps to remove redundant items, or principal component analysis, which helps to establish whether items on a scale all measure the same trait, or whether a scale is unidimensional [24, 25]. In 2017, based on these principles (importance of pathological specific content and statistical validation implementation), Khadka et al. developed a keratoconus outcome research questionnaire (KORQ), which has been termed by Kandel et al as “the only validated keratoconus- specific questionnaire”, which provides “the most superior psychometric properties” [8, 10]. Consequently, in more recent studies, the KORQ has been predominantly used to analyse vr-QoL in KC patients [6, 7, 26, 27].
To ensure the reliability of our results, we not only used the NEI-25 as we did in our previous study, but also the KORQ and did additional Rasch analysis to validate the precision of the KORQ when applied to our study population. Despite single items of the KORQ display statistically significant differences before – after CXL (in total 3 of 29, whereby 2 items displayed improvement of VA-related tasks and one a worsening of dry eye symptoms), neither the combined “activity” or “symptoms” scores demonstrated differences of statistical significance. The Rasch analysis revealed a good measurement precision using the KORQ in our population.
The two studies we found analysing the vr-QoL in KC patients pre and post CXL using the KORQ, reporting either an improvement or an equal vr-QoL [6, 7]. As done in our study, Kandel et al. and Ferrini et al included 39 or 38 patients with a follow up of 6 months. However, other than in their studies, we couldn`t find any correlation of visual- or morphology– parameters with the KORQ-scores. This was completely unexpected because most studies report, as expected, high correlation between vr-QoL scores and the visual acuity [6, 7, 10]. Looking at the data given in the respective studies, the most obvious biasing factors leading to different correlation-results might be a distinct heterogeneity of the age and the used VA-data: The mean age of the patients analyzed in the study by Khadka et al. was 45 years of age, which at least almost double the mean age of the patients included in our and the studies by Ferrini et al. (20.5) and Kandel et al. (24,2). Regarding the influence of the visual acuity data, Ferrini et al report a mean BCVA of the treatment eye of 0.2 before the treatment without giving information about the BCVA of the fellow eye. Their correlations analysis, conducted for both eyes of the patient, revealed significant correlations of the KORQ scores and the treated eye before, but not 6 months after the treatment. The untreated eye didn`t show any significant correlations between the BCVA and the KORQ scores “symptoms” or “activity”. Kandel et al. reporting a mean BCVA of the treatment eye of 0.4 before the treatment, without giving data of the untreated eye. The initial BCVA of 0.4 is much lower than in similar studies. Khadka et al. analyzing KC patients, but not before and after CXL, reported about the worse-, and better-eye, whereby the “better eye” of the patient was used for statistical analysis (mean BCVA better eye: 0.2, worse eye: 0.3).
Independently of the heterogeneity of the BCVA reported in the different studies, it seems to be not completely clear which eye should be used for correlation analysis: The one of the treatment eye? The one of the “better” eye or the binocular visual acuity? A recent study by Kandel et al., conducted to get more insight in the relationship between the QoL scores and the standard clinical variables, reported that the correlations of QoL scores with most clinical parameters and visual acuity were similar when considered either better eye or worse eye [28]. They also demonstrated statistically significant correlations between clinical and QoL scores, but only with “low magnitudes” suggesting a more complex relationship between clinical parameters and patient-reported outcomes [28].
A potential drawback of our study could be the heterogeneity of the VA of the included patients. Usually, we perform CXL in patients with a “still good enough vision for daily tasks”. Otherwise, we use a combined approach (PRK/ CXL), implant rings (MyoRing®) or do keratoplasty. However, some patients didn`t follow our recommendation and decided for the “more basic” (CXL) treatment which is usually also the only treatment option fully covered by their insurance company. On the one hand we could have decided to excluded these patients, on the other hand, we intentionally wanted to include all the patients treated in our facilities to avoid unrecognized bias and to get a “unfiltered” insight in the “general” vr-QoL of (our) KC patients.