Correlación y validación de criterios diagnósticos y escala de severidad en la insuficiencia limbar asistida con microscopia confocal

Abstract

INTRODUCTION Limbal stem cell deficiency (LSCD) is an eye disease caused by a critical decrease in limbal stem cells responsible for the renewal of the corneal epithelium and form a barrier to the conjunctival epithelium, opaque and vascularized epithelium. It is known the relationship between the corneal epithelium and neural subasal corneal nerve plexus, having a feedback circuit. The lack of healthy corneal epithelium and corneal neovascularization surface relates to the absence of corneal nerve plexus reciprocally. It has been associated with decreased density of the corneal nerve plexus nerves (SND) and basal cell density (BCD) with early stages of IL. HYPOTHESIS Analysis of cell density and subasal corneal epithelial nerve plexus using confocal microscopy help to early diagnosis and classify LSCD based on documentable and non-invasive tests. MATERIAL AND METHODS In our work, we have analyzed 97 eyes of 49 patients, of whom 85 eyes were diagnosed clinically and the rest of IL acted as controls. They underwent complete eye examination, including photographs of the anterior pole, vital stains, Oxford and Van Bijsterveld scales, OSDI and VFQ 25 questionnaires, Pentacam® tomography, Keratograph 5M® (K5M) and HRT II confocal microscopy with Rostock® cornea module. We classified these patients through Rama, Lopez-Garcia, Deng clinic, confocal Deng and Schwartz classifications. SPSS 22 was used for statistical analysis and nQuery Advisor Release 7 for sample size. To make the ordered logistic regression model, multivariate analysis was done by a non-automated method of direct introduction of variables, measuring the reliability with Hosmer-Lemeshow index. Agreement of classifications was measured through Cohen's kappa coefficient. RESULTS SND was analyzed with confocal microscopy in patients with IL: 6469 (6295) microns / mm2; lower than our healthy controls, 21766 (4970) microns / mm2) and reference numbers described in the literature,19000 microns/ mm2 (p <0.0005). Cell densities were lower compared to controls (statistically significant results). The logistic regression model showed that the SND is a protective factor against the fibrovascular pannus (Odds Ratio = 0.069; p = 0.017), but not the BCD. By the other hand, absence of corneal nerve plexus subasal, evidenced with confocal microscopy, and the presence of corneal pannus, have been positively related (p <0.05) with the involvement of central cornea, late fluorescein staining , persistent epithelial defect, decreased TBUT, Oxford scale and Van Bijsterveld, temporal limbar redness (K5M) and high-order aberrations (Pentacam®). Finally, osmolarity did not provide diagnostic information in our study. CONCLUSIONS We accept the hypothesis that confocal microscopy is a good method for the study, diagnosis and classification of LSCD, showing that the presence of corneal nerve plexus subasal is an independent protective factor for corneal neovascularization. We do not accept the hypothesis that basal cell density is a good diagnostic criterion for LSCD. We propose a new classification of LSCD, adding the presence of the central corneal nerve plexus as a protective factor.