If past is epilogue, making sense of dry eye disease (DED) will remain a distant hope for the foreseeable future. Considering the vast intellectual and financial resources that have been invested in the continuing search for clarity in our understanding of this disorder, one would imagine that the seemingly intractable enigmas posed by the tear-centered model should raise questions about its soundness. It seems astounding that they are nowhere to be found.
The data is out there for everyone to see. Insufficient tears and excessively rapid tear evaporation provoke dry eye symptoms. Increasing Meibomian glands productivity can mitigate them. Chronic dry eye symptoms are also reported by patients with adequate (even overflowing) tears, healthy Meibomian glands and pristine corneal surfaces. Nevertheless, despite this cognitive dissonance the luster of the tear deficiency/hyperevaporation model continues to shine with blinding luminance. My polemic represents an effort to initiate an open discussion of the limitations of the current model as a preamble to considering alternative paradigms that reflect the real world of these disorders. Our ultimate goal is to assemble all of the established clinical observations into a cohesive whole.
Designing a new theoretical model
Like all pain, that provoked by corneal surface drying signals impending or actual tissue damage. However, I suggest that in a teleological context, the dry eye-like pain (DELP) alarm evolved to protect and sustain the most powerful focusing lens of the human eye; the optical tear layer. As further support of this suggestion consider the unusually high density of nociceptor terminals located in the superficial corneal epithelial layers (far greater than needed to protect the tissue) and the specialized sensors embedded in their terminals designed to be activated when the thickness of the tear film falls below a critical level and threatens its break up. When triggered, they initiate the cascade of events culminating in the restoration of its thickness and termination of the alarm process that preempts its escalation to conscious DELP.
How do these sensors sense when the tear film is about to fragment? As thermoreceptors, they are programmed to discern the rate of temperature drop that occurs as the tear film undergoes evaporative thinning. How thin? This is determined by their sensitivity. As it increases, the tear layer thickness required to maintain the alarm in an inactive state also increases. Sensitized corneas feel dry with normally asymptomatic tear metrics.
Corneal inflammation and sensitization
Physiological nociceptor sensitization is induced by elevated levels of proinflammatory cytokines released during inflammation and is clinically manifested as hypersensitivity to noxious stimuli (hyperalgesia). Corneal hyperalgesia is typically expressed as hypersensitivity to tear evaporation (evaporative hyperalgesia).
Hyperosmolar tears caused by its insufficiency/hyperevaporation triggers corneal inflammation that promotes corneal sensitization and DELP which is consistent with the traditional DED model. Nevertheless, it fails to explain the poor correlation between symptoms and tear metrics among individuals in this cohort or why chronic DELP can be experienced by patients with a healthy supply of iso osmolar tears.
Let’s do the unthinkable and begin with a clean slate by removing the term dry from its current position of primacy and open our minds to the possibility that we are dealing with a spectrum of disorders that share the common features of chronic, inadequately explained symptoms of DELP and inflammation. These include age-related chronic DELP associated with corneal nerve attrition, reduced tear metrics, depressed corneal tactile sensitivity and hypersensitivity to tear evaporation (all of which are consistent with denervation sensitivity), a younger cohort having identical symptoms but reported as being associated with normal tear metrics and corneal nerve densities, patients with chronic DELP carrying a diagnosis of complex multisystem disease such as fibromyalgia and Sjogren’s syndromes and certain neurological disorders including small fiber neuropathies. Nor can it ignore tear deficiency due to suppression of the neurosecretory activity of the lacrimal functional loop, destroyed lacrimal gland acini or both. These disparate features caution us to resist the seductive appeal of a reductionist approach.
It seems reasonable to begin the exercise by listing the major clinical features of this spectrum of diseases. In this regard it makes sense to assign greater weight to the immutable ones (such as pain unsupported by external signs and inflammation) and less to the variable features (such as tear metrics, Meibomian gland productivity and tear film disruption). I argue that it is not by chance that the constant features are identical to those of painful peripheral neuropathies. Their shared disparity between symptoms and signs is striking. Moreover, the elevated levels of proinflammatory cytokines reported in the tears of patients without supporting signs of ocular surface desiccation can be explained by neuroinflammation generated during prolonged intense nociceptor activity. (The typically subclinical nature of this inflammation also supports the plausibility of this theory.)
What about their variable properties? It seems reasonable to consider them as representing downstream sequelae which is consistent with my informal observations of a few patients whom I followed from the early stages of their disease in whom reduced tear metrics followed the onset of symptoms by periods measured in years. In this context it seems reasonable to suspect that an asynchronous onset of complications could account for the otherwise enigmatic reports of Sjogren’s and non-Sjogren’s patients with chronic DELP whose tear metrics were statistically similar to those in the normal population despite robust evidence that tear deficiency is characteristic of the cohort as a whole. The neuropathic paradigm also predicts that treatments targeting complications such as tear deficiency and DELP are no more than palliative if the upstream dysfunctional mechanisms that drive them (those of corneal neuropathy?) remain unchecked. In other words, despite their having significant interactive relationships, DELP and tear deficiency are no more than fellow travelers.
Centralized corneal neuropathic pain
There is no pain without the brain. This self-evident statement serves to remind us that studies of chronic pain are incomplete without taking into account the powerful influences of their central signal processing circuitry. This is especially true of the cornea considering the strikingly disproportionate large size and complexity of its trigeminal nociceptive circuitry relative to its receptive field.
In contrast to all other types of sustained sensory inputs that undergo adaptation, those of pain increase in intensity. This activity-based phenomenon, known as central sensitization, is generated by complex facilitory and inhibitory feed-forward and feed-back systems integrated in their CNS pain-modulating circuitry. Moreover, central pain signals are further modified by higher brain centers that regulate emotions before reaching the somatosensory cortex where they are decoded into conscious sensations. Enter neuroplasticity. The CNS neural system has the unique capability of adapting to prolonged elevated levels of excitability through changes in its molecular, cellular and anatomical structures. Although it is reversible under physiological circumstances, neuroplasticity can become pathologically entrenched. As a result, incoming pain signals can be amplified and distorted (centralized hyperalgesia) and/or they can be generated spontaneously independently of afferent input. Centralized spontaneous pain is referred to their receptive fields (in this case the cornea) as phantom pain and/or expressed as otherwise unexplained photosensitivity (photoallodynia). Although experienced by patients as originating in the cornea, it is in reality projected to the cornea and sometimes to other areas served by the sensory trigeminal nerve. Nevertheless, despite suicide ideation-promoting pain intensities, external signs are virtually absent.
Although corneal evaporative hyperalgesia is typically initiated in the cornea, there is evidence that it can also be induced by neuropathic central sensitization. This suggestion is based on non-corneal related studies reported in the pain literature and anecdotal observations that are discussed in detail in an unpublished opinion/review. Moreover, my experience suggests that centralized corneal evaporative hyperalgesia may differ from that generated solely by the cornea in their higher intensities and, in some cases, association with corneal chemical and (sometimes) cold hyperalgesia. On the other hand, centralized corneal pain is typically spontaneous and differs from DELP in that it is often described as burning (as opposed to dry), hot, pressure, aching, sharp, cutting, needle-like, foreign body, etc. Moreover, spontaneous pain may have a radiating quality and can also be projected to victims’ heads, orbits, face and jaws. It is not surprising that pharmacological corneal anesthesia may not fully suppress eye symptoms.
Although centralized corneal pain may appear to start spontaneously it is more often initiated by corneal injuries. The provoking corneal insult can be as benign as a brief exposure to noxious fumes or triggered by corneal epithelial abrasions and recurrent erosions. Nevertheless, the massive axotomies performed during keratorefractive procedures are a more common cause and, more often than not, the associated centralized pain has a delayed onset of days to years following surgery. (See the case study of 21 post keratorefractory patients posted on this website which would have been suppressed in the pre-internet era.) The mechanisms underlying the susceptibility of certain individuals to developing long-lasting (or even permanent) neuropathic centralized pain following peripheral injuries that appear to have otherwise healed are unclear. Although circumstances beyond my control have prevented me from following up on most of these patients, I know of at least one who subsequently developed symptoms consistent with fibromyalgia. (See following paragraph.)
Chronic DELP is commonly associated with multisystem disorders such as Sjogren’s syndrome and fibromyalgia. Although these cohorts are reported to have lower tear metrics, the relationship between symptoms and Schirmer’s test measurements among their individuals is inconsistent, similar to that of the age related group. Chronic DELP and spontaneous centralized corneal pain may also be experienced by patients with disorders associated with small fiber neuropathy or as a seemingly isolated event. However, I have observed some in whom these symptoms preceded the clinical manifestations of multisystem diseases several years following their onset.
The lack of cures is not surprising considering the dominance of the traditional tears-based paradigm. Symptoms of DELP are mitigated by treatments that reduce tear evaporation, enhance the robustness of the tear films and suppress inflammation, and are completely blocked by scleral lenses. However, the wearing tolerance of these devices can be severely compromised, likely due to the presence of secondary mechanical hyperalgesia associated with the centralized pain syndrome.
Centralized corneal pain is resistant to treatments. Reducing afferent corneal signals can sometimes be helpful in mitigating the evaporative hyperalgesia pain component, when present. Although systemic analgesics are the mainstay of treatments, their typically modest effectiveness is further limited by disabling cognitive side-effects.
Recognition of the peripheral/central corneal nociceptive system as the source of what we currently label as ‘dry eye disease’ is the beginning of a long journey. I believe that the initial goal of future research efforts should be to identify the common upstream precursors of neuropathic pain if they exist and target them to develop therapeutic interventions that either correct their dysfunctional mechanisms, mitigate their harmful effects or both. The recent literature of mitochondrial dysfunction offers hints that that this emerging field may represent a new frontier for the fruitful study of these disorders.
Readers may wonder about how a ‘retired’ foot soldier whose experience was limited to working the trenches can have the chutzpah to question the research strategy of the generals in academia. It is a question that I often ask myself. Nevertheless, it doesn’t require a PhD to be impressed with the similarities between the path that research has taken in this field and an imaginary study of the effects of drinking chemically contaminated water in details down to molecular levels, while ignoring its upstream sources.
Our fealty to the canonical paradigm has led us to portray these victims as transgressors that, in addition to exponentially increasing their already immense burden, deprives them of their family’s support. They are the sacrificial lambs of our devotion to outdated conventional wisdom. Until effective therapeutic interventions are developed we must, at the very least, validate their suffering.
I urge colleagues to send me their comments pro and (especially) con.
*This blog is not (nor was it meant to be) a scientific exposition. A comprehensive opinion/review supported by a robust bibliography has just completed its gestational phase. B4i1L0l2