Article Text

Download PDFPDF

Dry eye symptoms align more closely to non-ocular conditions than to tear film parameters
  1. Anat Galor1,2,
  2. Elizabeth R Felix3,4,
  3. William Feuer2,
  4. Nabeel Shalabi1,2,
  5. Eden R Martin5,6,
  6. Todd P Margolis7,
  7. Constantine D Sarantopoulos1,8,
  8. Roy C Levitt1,6,7,8
  1. 1Department of Ophthalmology, Miami Veterans Administration Medical Center, Miami, Florida, USA
  2. 2Bascom Palmer Eye Institute, University of Miami, Miami, Florida, USA
  3. 3Research service, Miami Veterans Administration Medical Center, Miami, Florida, USA
  4. 4Department of Physical Medicine and Rehabilitation, University of Miami Miller School of Medicine, Miami, FL, USA
  5. 5John P. Hussman Institute for Human Genomics, University of Miami Miller School of Medicine, Miami, Florida, USA
  6. 6John T Macdonald Foundation Department of Human Genetics, University of Miami Miller School of Medicine, Miami, Florida, USA
  7. 7Ophthalmology Department, Washington University School of Medicine, St. Louis, Missouri, USA
  8. 8Department of Anesthesiology, Perioperative Medicine and Pain Management, University of Miami Miller School of Medicine, Miami, Florida, USA
  1. Correspondence to Dr Anat Galor, Bascom Palmer Eye Institute, University of Miami, 900 NW 17th Street, Miami, FL 33136, USA; agalor{at}


Objective To evaluate the relationship between dry eye symptoms, non-ocular conditions and tear film parameters.

Methods Design: Cross-sectional study. Participants/setting: The study population consisted of patients who were seen in the Miami Veterans Affairs eye clinic. Patients filled out standardised questionnaires assessing dry eye symptoms (dry eye questionnaire 5 (DEQ5) and ocular surface disease index (OSDI)), non-ocular pain, depression and post-traumatic stress disorder (PTSD), and also underwent measurement of tear film parameters. Main outcome measures: Correlations between dry eye symptoms and non-ocular conditions as compared with tear film parameters.

Results 136 patients with a mean age of 65 (SD 11) years participated in the study. All correlations between the dry eye questionnaire scores (DEQ5 and OSDI) and (A) self-reported non-ocular pain measures (numerical rating scale and pain history), (B) depression and (C) PTSD were significant and moderate in strength (Pearson's coefficient 0.24 to 0.60, p<0.01 for all). All correlations between the dry eye questionnaires and tear film measures were weak (Pearson's coefficient −0.10 to 0.18) and most were not significant. Multivariable linear regression analyses revealed that PTSD and non-ocular pain more closely associated with dry eye symptoms than did tear film parameters. Specifically, non-ocular pain and PTSD accounted for approximately 36% of the variability in DEQ5 scores (R=0.60) and approximately 40% of variability in OSDI scores (R=0.64). Of note, none of the tear parameters remained significantly associated with dry eye symptoms in either model.

Conclusions Dry eye symptoms more closely align to non-ocular pain, depression and PTSD than to tear film parameters.

  • Ocular surface

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Dry eye affects the quality of life of millions of people worldwide1 but developing effective therapies for those with disease has been challenging. There are probably a number of reasons for this, but in part this is because dry eye represents a heterogeneous group of ocular surface disorders caused by different pathophysiological mechanisms with poor correlation between signs and symptoms.2–5

Most dry eye treatments attempt to improve the tear film through the use of artificial tears, punctal occlusion, topical anti-inflammatory agents, dietary supplements and environmental modifications. Although some of these therapies lead to improvements in tear parameters, many patients report little or no improvement in ocular symptoms. Thus, additional treatments are needed.

One relatively unexplored approach to therapy is the use of agents that target dry eye symptoms instead of targeting the tear film. There is growing evidence that some patients who have been diagnosed with dry eye have dysfunction of the sensory pathways that innervate the ocular surface (ie, neuropathic pain).6 Shared features between dry eye symptoms and neuropathic pain include the presence of spontaneous and persistent pain, allodynia (ie, pain to normally non-noxious stimuli), and hyperalgesia (ie, increased pain associated with a noxious stimuli). Furthermore, some patients diagnosed with dry eye have been found to have abnormal corneal nerve morphology and physiology,7–9 much like that observed in patients with non-ocular neuropathic pain.10

Furthermore, patients with dry eye have been found to have greater sensitivity to pain in non-ocular sites (ie, forearm), a finding suggestive of alterations in central pain processing.11 In addition, depression and post-traumatic stress disorder (PTSD), conditions that frequently coexist with chronic pain,12 ,13 are more common in patients with dry eye than in patients without.14 In our studies, veterans with a dry eye diagnosis had a twofold higher risk of carrying a diagnosis of depression or PTSD.15 ,16 Similarly, in a study carried out in the UK, dry eye in a female cohort was significantly associated with depression, pelvic pain and chronic widespread pain syndrome.17

Studies in non-US populations have suggested that dry eye symptoms may more closely align with systemic rather than tear film parameters.18 ,19 In our previous study of Miami veterans, individuals with depression and/or PTSD had greater dry eye symptoms but similar tear film parameters as compared with those without psychiatric comorbidities.20 In this previous study psychiatric disease was defined using international classification of disease-9 codes in the medical record and not through standardised questionnaires. The goal of the current study, therefore, was to build on our initial observations and examine associations between dry eye symptoms and non-ocular and ocular parameters. We hypothesised that dry eye symptoms would more closely relate to non-ocular parameters than to tear metrics. Such a relationship would imply that these comorbidities are important in understanding the mechanism of dry eye pain and in treating patient symptoms; perhaps even more so than conventional tear metrics.


Study population

Patients with normal eyelid and cornea anatomies were prospectively recruited from the Miami VAMC eye clinic between October 2013 and November 2014 and underwent a complete ocular surface examination. Patients were prescreened by various eye care practitioners and eligible subjects were informed about an opportunity to participate in a 1-day research study with the purpose of studying mechanisms underlying dry eye symptoms. Patients were excluded from participation if they wore contact lenses, had a history of refractive, glaucoma or retinal surgery, used ocular medications with the exception of artificial tears, had HIV, sarcoidosis, graft-versus host disease or a collagen vascular disease, had an active external ocular process, or had cataract surgery within the last 3 months. The study was conducted in accordance with the principles of the Declaration of Helsinki.

Data collected

For each individual we collected demographics, past ocular and medical histories, and medication use. Patients filled out standardised questionnaires regarding ocular conditions including the dry eye questionnaire 5 (DEQ5)21 (score 0–22) and ocular surface disease index (OSDI) (score 0–100). Patients were then asked to specifically consider their non-ocular pain which was assessed with a number of questionnaires including a (1) numerical rating scale (NRS) (How would you describe the overall intensity of your non-ocular pain on average during the past 1 week? score 0–10); (2) number of chronic (≥3 months) pain conditions including arthritis, burn pain, headaches, diabetic neuropathic pain, tendonitis, central pain syndrome, muscle pain, causalgia, back pain, cancer pain, trigeminal neuralgia pain, sciatica, shingles, surgical pain, temporomandibular disorder and fibromyalgia; and (3) number of pain areas (via pain drawing). Psychiatric status was assessed with the use of the patient health questionnaire 9, a validated depression scale (score 0–27) and the PTSD Checklist—Military Version (PCL-M) (score 17–85).

Furthermore, all patients had a tear film assessment including (1) tear osmolarity (TearLAB, San Diego, California, USA) (once in each eye), (2) tear breakup time (TBUT) (5 µL fluorescein placed, three measurements taken in each eye and averaged), (3) corneal staining (national eye institute scale, five areas of cornea assessed (score 0–4 in each)),22 Schirmer's testing with anaesthesia, and eyelid and Meibomian gland assessment.2 Eyelid vascularity was graded on a scale of 0 to 3 (0 none; 1 mild engorgement; 2 moderate engorgement; 3 severe engorgement) and meibum quality on a scale of 0 to 4 (0=clear; 1=cloudy; 2=granular; 3=toothpaste; 4=no meibum extracted). Data were entered into a standardised database.

Main outcome measures

The main outcome measures were the correlations between dry eye symptoms, tear film parameters (from the more severely affected eye) and non-ocular pain.

Statistical analysis

All statistical analyses were performed using SPSS V.21.0 (SPSS, Chicago, Illinois, USA). Descriptive statistics were used to summarise patient demographic and clinical information. Pearson's and Spearman's correlations were used to evaluate the strengths of association between the main outcome measures. Linear regression analysis with forward selection was used to evaluate the contribution of the independent variables (ocular and non-ocular findings) on variability of the dependent variables (dry eye symptoms). We justified our use of this parametric methodology as (1) the independence assumption was met in our data, (2) the relatively large sample size promotes normality by the central limit theorem and (3) the use of this methodology to determine percentage variability does not require normality of residuals.


Study population

One hundred and thirty-six patients participated in the study. Demographics of the study population are presented in table 1.

Table 1

Demographic and clinical features of the studied population (n=136)

The mean DEQ5 score for all patients was 11 (SD 5.2, range 0 to 21), of which 87% (n=118) had a score greater than or equal to 6 (indicative of mild or greater dry eye (DE) symptoms). The mean OSDI score was 34 (SD 27, range 0–98), of which 82% (n=111) had a score greater than or equal to 6. The correlation between these two questionnaires was moderate to strong (Pearson's r=0.71, Spearman's ρ=0.73, p<0.0005).

Correlations between dry eye symptoms and non-ocular conditions

As summarised in table 2, correlations between the two dry eye questionnaires and non-ocular parameters, including measures of non-ocular pain, depression and PTSD, were moderate in strength, and all were significant (r=0.24–0.60, p<0.01 for all).

Table 2

Correlation between non-ocular and ocular parameters and dry eye symptoms

Correlations between dry eye symptoms and tear parameters

As summarised in table 2, correlations between the dry eye questionnaires and all but one of the tear film measures (corneal staining) were not significant (r=−0.10–0.18).

Linear regression analysis model for dry eye symptoms

Forward stepwise multivariable linear regression models considering demographic information, medication use, non-ocular parameters (non-ocular pain, depression and PTSD), and tear parameters (osmolarity, TBUT, Schirmer's test, corneal staining, eyelid vascularity, meibum quality) were used to evaluate the relationship between these variables and dry eye symptoms. For DEQ5, non-ocular pain (NRS) (B 0.72 SE 0.14, p<0.0005) and PTSD (PCL-M) (B 0.07 SE 0.02), p<0.0005) remained significantly associated with DEQ5 scores, and accounted for about 36% of the variability in scores (R=0.60). In a similar manner, PTSD (PCL-M) (B 0.60 SE 0.10, p<0.0005) and non-ocular pain (NRS) (B 2.67 SE 0.72, p<0.0005) remained significantly associated with OSDI scores, and accounted for about 40% of variability in scores (R=0.64). Of note, none of the tear parameters remained significantly associated with dry eye symptoms in either model.


We found that dry eye symptoms more strongly correlated with non-ocular conditions rather than with tear film parameters. In fact, in models that evaluated the association of non-ocular conditions and tear film metrics with dry eye symptoms, only non-ocular parameters were significantly associated with dry eye symptoms. Several studies carried out in other countries and climates have reported similar types of relationships between dry eye and non-ocular conditions. In two Chinese studies, patients with dry eye had higher depression and anxiety scores compared with non-dry eye control populations.14 ,19 Among patients with dry eye, anxiety and depression scores significantly correlated to dry eye symptoms but not to dry eye signs (TBUT, Schirmer's test, corneal staining). A similar result was described in a geriatric Korean population.18 As in our study, two of these studies found that the relationship between dry eye symptoms and non-ocular parameters was stronger than between symptoms and traditionally measured dry eye signs.18 ,19 Not all studies, however, have reached similar conclusions. For example, Wen et al,23 reported that in 472 patients with depressive or anxiety disorders, a high frequency of patients had dry eye symptoms and abnormal tear metrics (TBUT, Schirmer's test, corneal staining).

The finding that non-ocular measures correlate more closely to dry eye symptoms than tear parameters has a number of different possible explanations, three of which are summarised here. First, patients with non-ocular conditions may be suffering from central sensitisation, a process that affects generalised pain perception and pain related behaviours. Second, patients with underlying psychiatric conditions, including anxiety, may be misinterpreting ocular surface sensations. Third, patients with non-ocular pain and psychiatric conditions may be taking medications that affect their tear film status. Regardless of the underlying cause, our findings suggest that non-ocular factors are more strongly associated with dry eye symptoms than are tear film parameters.

These findings have implications for treatment of dry eye. Current dry eye treatment paradigms focus on treating ocular surface pathology in order to improve dry eye symptoms. Our findings suggest that systemic findings may need to be considered, and perhaps addressed, when developing a personalised dry eye treatment algorithm. For example, if central sensitisation partially underlies dry eye symptoms, agents that modulate neuronal function may have some role in treating symptoms. One demonstration of this approach was a set of prospective, randomised, placebo-controlled studies that demonstrated that gabapentin and pregabalin, modulators of peripheral and central neuronal activity, decreased postoperative pain after photorefractive keratectomy.24 ,25 These limited studies suggest that more research is needed to determine whether these types of medications will be beneficial in treating dry eye symptoms.

As with all studies, the current study has limitations that need to be considered. First, we only used common metrics to study the ocular surface. Thus, we did not include recently introduced technologies, not in standard use, to dynamically study the tear film. Second, our population of mostly male, older US veterans with a high frequency of systemic comorbidities is not representative of other US populations. However, we see this as one of our study strengths. As many veterans have abnormal tear metrics and psychiatric and non-ocular pain,2 we were able to evaluate which more closely aligned with dry eye symptoms. As such, we see our work as an important extension of other population-based studies18 ,19 which evaluated this question in healthier populations in China and Korea. Third, our measurements were obtained using specific scales and techniques and our findings may have been different if alternative measures were used.

With these limitations in mind, this study suggests that dry eye symptoms are more closely associated with non-ocular parameters than with tear film status. This finding has treatment implications as it suggests that different dry eye therapeutics may be considered based on a patient's systemic status. Future studies will be needed to evaluate which agents, in conjunction with local treatments, will be useful in managing disease symptoms.


The authors thank Drs D A Lubarsky and K Candiotti for their generous input and support. The authors also thank Gail Lewis, Allison L McClellan, Mireya Hernandez, Katherine T McMannus and Jerry Kalangara, MD for their help with executing the study.



  • Contributors All authors have made substantial contributions to conception or design of the work, or the analysis or interpretation of data for the work; and drafting of the work or revising it critically for important intellectual content; and final approval of the version to be published; and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

  • Funding Supported by the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Clinical Sciences Research and Development's Career Development Award CDA-2-024-10S (AG), NIH Center Core Grant P30EY014801, Research to Prevent Blindness Unrestricted Grant, Department of Defense (DOD- Grant#W81XWH-09-1-0675 and Grant# W81XWH-13-1-0048 ONOVA) (institutional); NIH NIDCR R01 DE022903 (RCL and ERM), and the Department of Anesthesiology, Perioperative Medicine, and Pain Management, University of Miami Miller School of Medicine, Miami, Florida for funding. The contents of this study do not represent the views of the Department of Veterans Affairs or the US Government.

  • Competing interests None.

  • Ethics approval Miami VA IRB.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

  • At a glance
    Keith Barton James Chodosh Jost Jonas