Bilateral In Vivo Confocal Microscopic Changes of the Corneal Subbasal Nerve Plexus in Patients with Acute Herpes Zoster Ophthalmicus

Varicella-zoster virus (VZV), also known as human herpesvirus 3, is a double-stranded DNA virus from the alpha herpesvirus family [1]. The primary infection, transmitted through highly contagious respiratory droplets, causes varicella (chickenpox), a common childhood illness that usually resolves without complication. Following the initial infection, the virus remains latent in the sensory ganglia [2‐5]. Reactivation of VZV and its spread from a single ganglion to the corresponding dermatome and neural tissue leads to herpes zoster (HZ, or shingles), a painful, vesicular eruption, whose incidence increases with age and immunosuppression, reaching in Canada, America, Europe, Asia, and Australia, a median incidence of 4–4.5 cases per 1000 person-years [6‐8].

Herpes zoster ophthalmicus (HZO) occurs when the ophthalmic branch of the trigeminal nerve is affected, which is reported in 10–20% of HZ cases. Indeed, the trigeminal ganglion is the most common site of latency [9]. The extension of the rash to the tip of the nose, called Hutchinson's sign, indicates the involvement of the nasociliary branch and is associated with a higher risk of ocular symptoms [10]. Complications can affect any ocular or orbital tissue. The cornea, the body's most densely innervated tissue, is involved in nearly 66% of cases [11]. Corneal nerves preserve integrity and transparency through a complex mechanism involving sensation, blink reflex, homeostasis, and trophic factor release [12]. They form a dense network of unmyelinated nerve fibers between the Bowman's membrane and the basal epithelium, known as the subbasal nerve plexus (SNP) and extend into the epithelium [13]. Its damage can affect all corneal layers and cause epithelial and stromal keratitis, endotheliitis, and neurotrophic keratitis (NTK), characterized by reduced or absent corneal sensation. This degenerative disease may result in epithelial and stromal ulceration and, in severe cases, corneal perforation, severely impacting visual acuity and overall quality of life. Its complex management, particularly in advanced stages, highlights the importance of prevention and early diagnosis. Despite the availability of effective recombinant, adjuvanted, and live-attenuated vaccines, the incidence of HZO continues to rise, probably due to insufficient vaccination coverage [14‐17].

In vivo confocal corneal microscopy (IVCM) is a rapid, non-invasive technique that enables detailed observation and analysis of the corneal structure at the cellular level. Since its first application for corneal imaging by Masters and Thaer in 1994 [18], its performance has significantly increased, particularly in real-time SNP imaging. Alterations in this network of thin nerve fibers have been observed in various corneal and systemic conditions, including keratoconus, dry eye disease, corneal dystrophy, diabetes mellitus, and infectious keratitis [19‐22]. IVCM has demonstrated its utility in the early detection of subclinical keratoconus, assessing flap-related complications after refractive surgery, and the diagnosis of filamentous fungal keratitis [23‐25]. Ongoing research explores its potential in diagnosing early-stage diabetic peripheral neuropathy [26, 27]. In a study by Hamrah et al., bilateral corneal nerve loss was found in patients with unilateral chronic HZO and corneal involvement, which correlated strongly with reduced corneal sensation [28]. More recently, Mok et al. investigated early-stage corneal nerve changes in patients with unilateral HZO and detected corneal nerve degeneration in the ipsilateral eyes of HZO as early as 2 months post-reactivation. Among patients enrolled in the study, 26.7% had clinical corneal involvement [29].

We aimed to assess whether acute unilateral HZO in patients without corneal involvement commonly results in acute and bilateral corneal denervation. Therefore, we evaluated IVCM parameters in patients with acute unilateral HZO and no active or inactive signs of keratitis in either eye.

This is the first study to assess acute and bilateral corneal nerve changes following HZO, exclusively in patients with no clinical sign of keratitis. It is also the largest sample to date that analyzes corneal nerve morphology after HZO. In both eyes of patients with HZO, all seven corneal nerve parameters measured with ACCMetrics were significantly reduced compared with controls. In the HZO ipsilateral eye group, significantly more patients had a decreased corneal sensitivity. Furthermore, a positive Hutchinson’s sign was associated with lower CNFD, CNBD, CNFL, CTBD, and CFracDim values. These results suggest that all patients with HZO present bilateral subclinical signs of NTK during the acute phase of the disease.

Ipsilateral corneal nerve degeneration following HZO may be attributed to neuroimmune interactions, as indicated by the inverse correlation between the density and size of dendritiform immune cells and corneal nerve parameters with IVCM [33]. Researchers investigating the link between the immune and nervous systems proposed that neuroimmune communication occurs via cytokines and interleukins produced by leukocytes. This interaction, mediated by nerve receptors and neuroendocrine cells, enhances the inflammatory response to pathogens and constitutes an essential self-regulating system [34, 35].

Bilateral neuronal alterations and inflammatory response following unilateral nerve damage have already been observed in studies supporting the concept of diaschisis [36‐39]. Moreover, skin biopsies revealed contralateral epidermal nerve damage in patients with unilateral HZ [40]. Concerning the cornea, contralateral mirror-image involvement of clinically unilateral HZO is presupposed to be mediated by two anatomical pathways. First, peripheral trigeminal nerve fibers project to both sides of the brainstem and caudal medulla, potentially enabling the spread of HZ to the contralateral, unaffected eye [41]. Secondly, the symmetrical architecture of the nervous system suggests that primary afferent trigeminal nerve fibers project bilaterally to the central trigeminal nucleus on both ipsilateral and contralateral sides after crossing the pontine tegmentum [42]. Significant bilateral corneal nerve alterations in patients with unilateral HZO have been reported in several studies with smaller sample sizes since the development of IVCM. Using slit-scanning confocal microscopy (SSCM) with an axial resolution of 25 μm, Hamrah et al. identified nerve alterations in both eyes of patients with unilateral HZO, specifically in total nerve length, total number of nerves, main nerve trunks, and branching, 5 years after disease onset [28]. Other studies using laser scanning technology with a higher resolution (1 μm) have found similar findings [33, 43]. However, they all included patients with initial corneal involvement, while our study focused exclusively on patients with an acute diagnosis of HZO and no apparent keratitis. Another interesting finding in our research is that ipsilateral corneal denervation appears more pronounced than in the contralateral eye, as CNFD values show a significant decrease.

Following unilateral herpes simplex keratitis (HSK) infection, bilateral corneal nerve changes were observed within days after disease onset and continued progressively [44]. Surgical axotomy of the ciliary nerve in mice results in bilateral subbasal nerve loss one day post-surgery [45]. In the case of Zoster infection, it has been shown that neuronal and nerve destruction occurs even before clinical signs [46, 47]. Our study detected bilateral corneal denervation at a median time of 5 days after disease onset. The rapidity of the immune-mediated nerve alteration illustrates the involvement of neuronal transmedian signaling. Mok et al. examined the same IVCM parameters in both eyes of patients with HZO at three-time points: within two weeks of symptom onset, at 2 months, and up to 6 months. Alterations in two of the seven corneal nerve parameters—CNFD and CNBD—were first observed in the ipsilateral HZO eye 2 months after the onset of the disease. These changes were resolved by 6 months, suggesting a spontaneous corneal nerve regeneration. In the HZO contralateral eye, rather than a corneal nerve loss, an increase of three corneal nerve parameters—CNFA, CNFW, and CFracDim—was noted at 2 months, possibly as a compensatory response to nerve damage in the ipsilateral eye [29].

It has been demonstrated that patients with acute corneal nerve loss due to infectious keratitis exhibit spontaneous subbasal nerve regeneration within the first 6 months following the resolution of the infection [48]. Neurotrophins, immune modulators, and nerve guidance molecules such as SEMA 7A mediate this reinnervation process [49, 50]. Considering the acute corneal denervation process in HZO disease, preservative-free artificial tears should be considered, as they represent the primary treatment for early stages of neurotrophic keratopathy [51].

Because of the nature of our data, the correlation between corneal nerve alteration and corneal sensitivity could not be performed. However, in the ipsilateral eye of patients with HZO, we observed a further significant reduction in IVCM parameters in those with reduced corneal sensitivity compared to those with normal corneal sensitivity. These findings support previous studies stipulating that the corneal sensation correlates with alterations in the SNP [28, 52, 53]. In the contralateral eyes of patients with HZO and reduced corneal sensitivity, only CNFW was significantly lower compared with the control group (p = 0.04). CNFW is known to be a very discriminative indicator. Indeed, Giannaccare et al. found that CNFW had the highest diagnostic accuracy in differentiating patients with dry eye disease [54].

A positive Hutchinson’s sign indicates a higher risk of ocular inflammation and development of neurotrophic keratitis [55]. Our results support this dogma since it was associated with more significant alterations in corneal nerve parameters, except for CNFA and CNFW. However, it was not associated with an increased corneal sensitivity loss.

Corneal nerve density is reported to be age-dependent, with a linear decrease of 0.9% per year according to Niederer, and a decline of 0.25–0.30% per year reported by Parissi [56‐58]. As a potential confounding variable, we correlated SNP parameters with age and found a significant negative correlation. When adjusted for age, a significant bilateral decrease of all IVCM parameters remained in patients with HZO.

In both eyes of patients with HZO, no correlation was found between corneal nerve alterations and the duration since the onset of the disease or the start of intravenous acyclovir therapy. Similarly, in Hamrah et al.’s study, despite a longer duration since the onset of HZO (5.1 ± 6.4 years), no correlation between IVCM parameters and the time from the first episode of HZO was identified [28].

This study has several strengths and limitations. The subjects included in the study were patients hospitalized for intravenous acyclovir treatment because of the dermatological severity of HZO, which may have introduced a recruitment bias. While no data suggests that intravenous acyclovir could trigger corneal nerve alteration, patients should have been recruited and examined before the initiation of intravenous acyclovir. Moreover, patients using antiglaucoma topical eye drops were not excluded from the study. Alterations in corneal sensitivity due to antiglaucoma topical eye drops have been reported, particularly those containing the preservative benzalkonium chloride [59, 60]. The acquisition and selection of all images were performed by the principal investigator, who had initial access to the patient's clinical status. While the non-blinded study design helped minimize inter-operator measurement bias, it may have introduced observation and image selection biases. Another limitation is that corneal sensitivity and IVCM measurements were only assessed at the central cornea, which may limit the applicability of the results to the peripheral cornea. Although previous studies have shown that the inferior corneal region contains more tortuous, whorl-like nerves than the central region, and that corneal nerve parameters—such as CNFL, CNFD, and CNBD—are less reproducible across different corneal locations, future studies should investigate all four corneal quadrants. The fully automated image analysis quantification with ACCmetrics may have contributed to measurement bias. This software was initially validated for morphometric analysis of diabetic corneal neuropathy [61]. However, Jia Ying Chin et al. validate automated and manual quantitative analyses to assess corneal nerve plexus changes following refractive surgery [62]. The study of Dehghani et al. also demonstrates that automated analyses effectively differentiate corneal nerve plexus differences in individuals with and without diabetes [63]. Albeit six images are generally sufficient to reduce variability, we selected and analyzed a minimum of 15 images per patient to enhance the reproducibility of the parameters [64, 65]. Our study assessed corneal sensitivity with sterile cotton wool, which lacks reliability and only evaluates mechanical nociceptors. Using a Belmonte or Cochet–Bonnet esthesiometer would have provided a more precise measurement. Finally, the lack of longitudinal follow-up prevented the assessment of potential nerve regeneration over time.

IVCM combined with automated software now enables rapid and reproducible quantification of corneal nerve impairment. The results of our study indicate that unilateral HZO results in acute bilateral corneal denervation, regardless of the clinical presence of keratitis. The rapidity of antiviral therapy initiation does not seem to impact the severity of corneal nerve alteration. These findings emphasize the importance of early follow-up in the first months after disease onset and the potential benefit of bilateral preservative-free artificial tear administration to prevent the development of neurotrophic keratopathy (NTK). Raising the patient's awareness of clinical signs that require urgent follow-up, especially in those with a positive Hutchinson's sign, is essential. Together with the systematic use of preservative-free artificial tears immediately after the first symptoms of HZO, this approach could reduce the rate of patients developing a symptomatic form of NTK, estimated at 25% in the literature [66]. Larger, multicenter longitudinal studies are required to confirm the accuracy and robustness of these results.