1. Common clinic scenarios
'Doctor, my eyes really feel awful — dry, stinging, sometimes like there's sand. I've been to two ophthalmologists, both said my tear volume and cornea looked fine, just use artificial tears. But this is miserable every day. Am I just too sensitive?'
'Doctor, my mom is 70, she says her eyes feel fine. But check-ups have called her dry eye severe and suspected Sjögren's. She figures if it doesn't hurt or itch, why treat it?'
These two scenarios sound opposite, but they highlight one of the trickiest problems in dry-eye care: what the patient FEELS rarely matches what the doctor SEES on exam. For doctors, decision-making gets harder (which side do I trust?). For patients, it's deeply frustrating — 'I feel terrible and the doctor says nothing's wrong' or 'I feel fine but they say I'm severe.'
A February 2026 secondary analysis of the DREAM trial in American Journal of Ophthalmology (Matar et al.) put numbers on this. Across 535 moderate-to-severe dry-eye patients, the rank-based method showed 77% of patients had symptoms that didn't match signs. This article walks through the methods, the 4 phenotypes, who falls into each, why it matters for treatment choice, and the current Taiwan NHI coverage rules for artificial tears and Cyclosporine eye drops (Restasis).
2. Bottom line: what you really need to know
30-second takeaways
- Symptom-sign discordance is the rule, not the exception — in 535 DREAM patients, 77% had mismatched subjective and objective severity
- 4 phenotypes: symptom-dominant 37.8%, sign-dominant 39.3%, concordant 23.0% (mild + severe combined)
- 'Discomfort' barely correlates with the exam — Spearman r ≤ 0.12 against TBUT, Schirmer, corneal staining, meibomian assessment
- Who falls where: symptom-dominant → younger, female, ↑ bodily pain; sign-dominant → Sjögren's, older; antidepressant use → more concordant
- Clinical implication: symptom-dominant may involve neurosensory amplification — artificial tears alone may not be enough; sign-dominant needs adherent treatment and follow-up — 'no symptoms' ≠ 'no damage'
- Taiwan NHI: artificial tears (since July 2024) need Schirmer <5mm or TBUT <5s; Cyclosporine (Restasis) restricted to Level 3+ severe DED with pre-authorization
- Evidence level: moderate — single-study secondary analysis, no control arm, excluded post-LASIK patients, no neuro assessment, lead author has substantial industry ties. Not yet adopted by TFOS DEWS II
- Patient message: your subjective experience is real — discordance is a documented, common phenomenon, not malingering. Discuss with your ophthalmologist which phenotype you fit and what treatment fits
3. Common patient Q&As
4. What is symptom-sign discordance?
Symptom-sign discordance is a long-recognized challenge in DED care: a patient's subjective severity and the clinician's objective findings often don't align. Nichols et al. (Cornea, 2004) first highlighted that DED symptoms and signs "lacked association," and subsequent large cohort correlations have consistently sat below 0.2 (Pult 2011, Sullivan 2014, TFOS DEWS II 2017).
① Why the mismatch? Three proposed mechanisms
The literature proposes three main mechanisms, which can coexist in one patient:
- Tear film instability + ocular surface inflammation (traditional mechanism) — the TFOS DEWS II core definition: tear instability, hyperosmolarity, inflammation, neurosensory abnormalities feed into each other
- Neurosensory amplification / neuropathic ocular pain — corneal nerves over-respond to minor stimuli, or central pain processing is altered (central sensitization); this subgroup keeps hurting even when the ocular surface 'looks' normal
- Corneal hypoesthesia — chronic inflammation, diabetes, autoimmune disease (especially Sjögren's), neuropathy, chronic medications reduce corneal nerve density; this subgroup feels little even with substantial surface damage
These mechanisms explain why symptoms and signs don't move in parallel — each reflects a different pathology, and conventional 'measure tear volume / look at corneal staining' tools only capture part of the picture.
② Why does this classification matter clinically?
If different phenotypes involve different mechanisms, treatment strategy shouldn't be one-size-fits-all:
- Symptom-dominant: possibly neurosensory → beyond basic tears, evaluate neuropathic pain, consider neuromodulators, autologous serum
- Sign-dominant: possibly corneal hypoesthesia → anti-inflammatory (steroids, Cyclosporine), Sjögren's workup, punctal plugs
- Concordant-severe: classic DED → standard stepwise therapy
5. How was the study conducted?
① About the DREAM trial
DREAM (Dry Eye Assessment and Management) was an NIH/NEI-funded multicenter RCT testing oral omega-3 supplementation for moderate-to-severe DED (Asbell et al., NEJM 2018 — primary result was negative; omega-3 no better than placebo). The trial enrolled 535 patients across 27 US sites, randomized to omega-3 or olive-oil placebo, followed 12 months.
The 2026 Matar paper is a secondary analysis of DREAM — re-using the trial's symptom questionnaires and clinical exam data to study symptom-sign discordance, which wasn't DREAM's original research question.
② Measurement tools used
| Tool | Type | Description |
|---|---|---|
| OSDI | Subjective | Ocular Surface Disease Index, 12 items, 0–100 |
| BODI | Subjective | Brief Ocular Discomfort Inventory, ocular discomfort severity |
| TBUT | Objective | Tear break-up time (sec), tear film stability |
| Schirmer test | Objective | Tear production (mm/5min) |
| 角膜螢光染色 | Objective | 0-3 scale, corneal epithelial damage (NEI grading) |
| 結膜麗絲胺綠染色 | Objective | 0-3 scale, conjunctival epithelial damage |
| 瞼板腺評分 | Objective | MG plugging + secretion quality score |
| SF-36 | QoL | 36-item health survey (overall QoL impact) |
③ Three discordance methods
| Method | Logic | Clinical use |
|---|---|---|
| ① Rank-based | Rank each patient's symptom score and composite sign score among the 535; compute the rank difference | Cross-sectional snapshot — answers 'which phenotype is this patient?' |
| ② Discomfort-based | Use BODI #3 'past-week mean ocular discomfort severity' (0-100) into 4 strata: none/mild, low-mod, high-mod, severe | Functional burden lens — answers 'how much is this patient's suffering affecting daily life?' |
| ③ Temporal | Compare DIRECTION of change in symptoms vs signs between month 3 and month 12 | Treatment-response tracking — answers 'is treatment working?' 'time to adjust?' |
The authors emphasize these 3 methods are complementary, not competing — each answers a different kind of clinical question.
6. Finding 1： cross-sectional — 77% are discordant
With the rank-based method on DREAM baseline data and a ±0.1 threshold for concordance:
| Phenotype | N (%) | Profile |
|---|---|---|
| Symptom-dominant | 202 (37.8%) | Symptoms much worse than signs; younger, female, higher bodily pain, higher reported discomfort |
| Sign-dominant | 210 (39.3%) | Signs much worse than symptoms; older, more Sjögren's, higher reported relief from treatment |
| Concordant (mild + severe) | 123 (23.0%) | Subjective and objective align; spans both ends |
* From Matar et al. 2026, Figure 1 + Table 2. Note DREAM only enrolled patients with moderate-severe symptoms AND signs, so the concordant-mild quadrant is sparse in this sample.
① Which factors associate with discordance?
Univariate linear regression (Table 2) showed:
- Age (β=-0.004, p=0.003): older = more sign-dominant
- BODI discomfort (β=0.006, p<0.001): more discomfort = more symptom-dominant
- BODI relief from treatment (β=-0.002, p=0.007): more relief reported = more sign-dominant
- Sjögren's syndrome (β=-0.175, p=0.002): strongly sign-dominant
- RA, OA: trend toward sign-dominant but not significant (p=0.07)
- Sex, race, smoking, diabetes, depression, fibromyalgia, antidepressant use: no significant association in this analysis
7. Finding 2： discomfort barely correlates with the exam
One of the paper's most striking findings. Spearman correlations between self-reported ocular discomfort (BODI #3) and various measures (Table 3):
| vs Discomfort | Spearman r | p value | Strength |
|---|---|---|---|
| Subj: OSDI total | 0.47 | <0.001 | Moderate |
| Subj: OSDI vision | 0.23 | <0.001 | Weak-mod |
| Subj: OSDI ocular sx | 0.44 | <0.001 | Moderate |
| Subj: OSDI environmental | 0.38 | <0.001 | Weak-mod |
| Obj: Corneal staining | 0.12 | 0.006 | Very weak |
| Obj: Conjunctival staining | 0.04 | 0.36 | None |
| Obj: TBUT | -0.05 | 0.25 | None |
| Obj: Schirmer | -0.10 | 0.02 | Very weak |
| Obj: Meibomian gland | 0.02 | 0.71 | None |
* Convention: r > 0.5 strong; 0.3-0.5 moderate; 0.1-0.3 weak; < 0.1 negligible.
⚠️ What does this mean?
When a patient says 'my eyes feel terrible,' the ophthalmic exam has very limited power to predict 'how terrible' (r ≤ 0.12). This is not because exams are inaccurate — it's that subjective distress and objective ocular damage are partly independent phenomena, driven by neurosensory amplification, central pain processing, and psychosocial factors.
The study also found higher discomfort was significantly associated with worse SF-36 scores across bodily pain, mental health, social functioning, sleep — confirming that the subjective burden of DED affects quality of life broadly, beyond just the eyes.
8. Finding 3： temporal — 46% divergent trajectories over 9 months
The study compared the DIRECTION of change in symptoms vs composite signs between month 3 (M3) and month 12 (M12). M3 (not baseline) was used as reference to avoid early placebo effects observed in DREAM.
| Sign metric | N | Sign-dominant | Symptom-dominant | Concordant |
|---|---|---|---|---|
| Composite | 445 | 22.7% | 23.4% | 53.9% |
| Corneal staining | 454 | 32.2% | 30.0% | 37.9% |
| Conjunctival staining | 458 | 34.5% | 30.6% | 34.9% |
| Schirmer | 449 | 39.2% | 28.1% | 32.7%（最低） |
| TBUT | 446 | 25.8% | 26.7% | 47.5%（最高） |
| MGD | 456 | 35.5% | 26.1% | 38.4% |
* From Matar et al. 2026, Table 4.
Key findings:
- Even with the composite, 46% had divergent symptom-sign trajectories over 9 months — discordance isn't just a single-snapshot phenomenon, it persists over time
- TBUT aligned best with symptom changes over time (47.5%) — suggesting tear film instability is the sign patients are most sensitive to
- Schirmer had the highest sign-dominant rate (39.2%) — tear deficiency is frequently 'under-reported' by patients (declining tear volume doesn't always register subjectively)
- In multivariate analysis, antidepressant use associated with temporal concordance (OR 1.70, 95% CI 1.08-2.67, p=0.02) — may reflect more consistent symptom reporting in this group, or effects of comorbid mental health
- Female + higher bodily pain associated with symptom-dominant discordance in the corneal staining model (female OR 2.11, p=0.01; bodily pain per point OR 1.03, p=0.03)
9. Who falls into which phenotype?
Integrating findings across the 3 methods, the typical profiles are (statistical tendency, not individual prediction):
| Phenotype | Typical profile | Possible mechanism |
|---|---|---|
| Symptom-dominant | • Younger (<60) • Female • High overall bodily pain scores • High OSDI but normal exam • High self-reported discomfort • Often coexists with other chronic pain (fibromyalgia, chronic headache) |
Neurosensory amplification (corneal nerves or central pain processing) |
| Sign-dominant | • Older (>65) • Sjögren's / autoimmune disease • Marked corneal staining, low Schirmer • Self-reports 'fine' or 'no problem' • Paradoxically higher 'relief from treatment' |
Corneal hypoesthesia (chronic inflammation / autoimmune reduces nerve density) |
| Concordant | • Symptoms and signs both mild OR both severe • On antidepressants (temporal method) • Subjective + objective responses move in parallel |
Classic DED (tear film instability + surface inflammation) |
10. Clinical meaning — why this matters
① For patients: your experience is real
If you're symptom-dominant, the most important message is — your discomfort isn't malingering or 'oversensitivity'. About 38% of DREAM's 535 patients fit this phenotype; it's common. You may have been told 'the exam looks fine' or 'just use artificial tears' — but this reflects current diagnostic limits, not the absence of your symptoms. Discuss with your ophthalmologist: (1) evaluation for neuropathic ocular pain; (2) coexisting meibomian gland dysfunction; (3) advanced options like autologous serum drops.
If you're sign-dominant, the key message is — 'no symptoms' ≠ 'no problem'. About 39% of the cohort fit this; many had Sjögren's or other autoimmune disease. The cornea may be damaged even when you don't feel it; by the time pain returns, ulceration or irreversible vision loss may have occurred. Adhere to drops and follow-up — don't stop or delay because 'I feel fine.'
② For clinicians: 3 phenotypes → 3 lines of thinking
💡 Authors' suggested clinical use
- Rank-based (cross-sectional): for 'which phenotype is this patient?' — useful for triage on initial referral
- Discomfort-based: for 'how much is this patient's suffering affecting QoL?' — drives treatment intensity and prioritization
- Temporal: for 'is the treatment working? time to change?' — use M0→M3→M6 trajectory at follow-up
The 3 methods are complementary, not competing — together they give a fuller picture. But this framework is not yet a TFOS DEWS II standard — guideline-based stepwise therapy remains the clinical backbone.
11. Taiwan NHI dry-eye coverage (April 2026)
Taiwan's NHI covers two main classes for DED: 14.5 artificial tears and 14.9.3 Cyclosporine eye drops (e.g., Restasis). Current rules below.
① 14.5 Artificial tears (effective July 1, 2024)
| Item | Content |
|---|---|
| Initial criteria (any one) | 1. Schirmer < 5 mm in ≥1 eye, OR 2. TBUT < 5 sec, OR 3. DED-related corneal pathology (including exposure keratopathy) |
| Documentation | Chart documentation with relevant notes, photos, or video |
| Quantity limit | 15 mL bottle: 1 per month ≤ 10 mL bottle: 4 per 3 months |
② 14.9.3 Cyclosporine eye drops (e.g., Restasis)
| Item | Content |
|---|---|
| Eligibility | DED severity Level 3 or above |
| Required criteria (ALL needed) | 1. Schirmer without anesthesia < 5 mm/5min, strip required 2. TBUT ≤ 5 sec 3. External color photo + fluorescein-staining photo (severe corneal PEE, conjunctival hyperemia/staining, filamentary keratitis, ulcer, symblepharon, keratinization) 4. Failed prior topical anti-inflammatory, punctal plugs, or artificial tears |
| Review process | Pre-authorization required; re-reviewed every 6 months |
| Continuation | (1) Symptomatic improvement (dryness, photophobia, discharge) + (2) Any one improved: Schirmer / TBUT / corneal staining / conjunctival hyperemia/staining |
③ How does this relate to DREAM's discordance findings?
Taiwan NHI's criteria are predominantly objective-sign-based (Schirmer, TBUT, staining photos). For DREAM's 'symptom-dominant' phenotype, this means: even with severe subjective symptoms, if objective exam doesn't meet criteria, NHI won't cover artificial tears or Cyclosporine.
This isn't NHI being 'unreasonable' — objective-criteria-based coverage is a reasonable sustainability mechanism. But for patients, the practical implication is: symptom-dominant patients may more often need self-pay options (OTC artificial tears, self-pay Restasis, autologous serum). Recommendations: (1) get a thorough exam with documentation; (2) discuss your treatment ladder with your doctor; (3) ask your doctor about cost ranges for self-pay options.
* NHI coverage rules from the National Health Insurance Administration (NHIA) 'Drug Coverage Regulations' April 2026, Chapter 14 Ophthalmology. Actual application per NHIA notice.
12. Limitations and caveats
⚠️ Honest read of the evidence
The study contributes meaningfully (large sample, multi-center, 3 methods, 12-month follow-up), but its proper place in the evidence hierarchy:
- Single secondary analysis: DREAM tested omega-3; discordance is post-hoc; needs replication in independent cohorts
- No control arm: DREAM compared omega-3 vs placebo, both 'treated' — can't isolate treatment effects on discordance
- Excluded post-refractive-surgery patients: a major symptom-dominant clinical group (post-LASIK DED) — findings don't generalize
- Only moderate-severe DED with both symptoms and signs: excludes mild DED and pure neuropathic ocular pain (severe symptoms, zero signs)
- No corneal nerve assessment: no confocal microscopy or esthesiometry — neurosensory inferences are indirect, not directly demonstrated
- Each method rests on assumptions; no gold standard: each definition has its own assumptions, none is THE truth
- Linear normalization assumption: the 5-sign composite assumes linearity and equal weighting, which may not hold in all settings
- Author conflicts of interest: lead co-author Asbell discloses ties (consulting/speaker/research) with 12+ ophthalmology-related companies (Abbvie, Amgen, Azura, CS Pharmaceuticals, Glia, Harrow, Horizon, Iollyx, LinkBiologic, Premark, Regeneron, Santen, Senju, Trefoil, Vindico). Other authors: no disclosures. Disclosed COIs ≠ biased conclusions, but readers should weigh this
- Not yet in TFOS DEWS II: this 3-method framework is currently a 'research perspective' — guideline-based stepwise therapy remains the clinical backbone
13. Common myths
14. Advanced: for readers who want more depth
This section uses more technical language and suits medical students, residents, optometrists, and readers interested in methodology. General readers can stop at the previous section.
① Neuropathic ocular pain (NOP) and central sensitization
- NOP concept formalized by Rosenthal et al. (Ocul Surf 2009) 'Corneal pain without stain' — persistent corneal pain without corresponding staining or objective signs
- Mechanism hypotheses: (a) peripheral sensitization — corneal nerves over-react after inflammation/injury; (b) central sensitization — altered trigeminal nucleus or higher-order pain processing; (c) descending modulation failure
- Diagnostic tools: (a) in vivo confocal microscopy — quantifies corneal nerve density/length, dendritic cell density; (b) esthesiometry (Cochet-Bonnet or non-contact) — quantifies mechanical/thermal threshold
- Treatment ladder (local→systemic): preservative-free tears → topical cyclosporine/lifitegrast → autologous serum → scleral lens → neuromodulators (gabapentin, pregabalin, SNRI like duloxetine) → CBT → advanced nerve blocks. Requires multidisciplinary care (ophthalmology, pain, psychiatry)
② Chronic Overlapping Pain Conditions (COPCs)
Levitt et al. (Mol Pain 2017) proposed DED may belong to the COPC family — chronic pain syndromes sharing central sensitization, including fibromyalgia, chronic headache, TMD, chronic pelvic pain, IBS. COPC patients often have multiple chronic pains, elevated bodily pain scores (SF-36), female predominance — overlapping closely with DREAM's symptom-dominant profile. Clinically, a symptom-dominant DED patient with other chronic pain conditions should be evaluated as possibly within COPC, shifting treatment toward systemic pain modulation rather than purely local therapy.
③ Why does Schirmer dominate the 'sign-dominant' picture?
In DREAM's temporal analysis, Schirmer had the highest sign-dominant rate (39.2%) and lowest concordance (32.7%). Possible reasons:
- Absolute changes in tear volume (e.g., 9→6 mm/5min) have limited subjective impact — patients chronically adapted to gradual decline 'don't feel' the change
- Schirmer has high test-retest variability (mm differences across repeats in same patient) — 'direction of change' has noisy signal
- Tear production reflects aqueous deficiency; subjective discomfort also depends on lipid/mucin layers, surface inflammation, and neurosensory factors — one measure can't capture the whole picture
④ DREAM in relation to the TFOS DEWS II treatment ladder
TFOS DEWS II (2017) treats DED in a stepwise ladder (Step 1–4): lifestyle + tears (1), topical anti-inflammatory + punctal plugs (2), autologous serum + therapeutic contact lens (3), surgical (4). The ladder is primarily symptom-severity-based with no formal symptom-dominant vs sign-dominant branching. DREAM's contribution hints that future guidelines may need phenotype-based triage, with different priority orders for the two discordance subgroups. Still an open research direction — not in guidelines yet.
15. Summary: what you can do
If you're symptom-dominant (feel terrible but exam normal): (1) your experience is real, not malingering; (2) discuss in depth with your ophthalmologist — beyond basic tears, evaluate for neuropathic pain, add anti-inflammatory, consider autologous serum; (3) review lifestyle (screen time, AC, blink frequency, contact lens use); (4) assess for other chronic pain (fibromyalgia, chronic headache) — if present, may need multidisciplinary care; (5) don't expect a 'cure' — this phenotype usually requires long-term management.
If you're sign-dominant (severe exam but feel fine): (1) 'no symptoms' ≠ 'no problem'; (2) adherent treatment and regular follow-up are critical; (3) if Schirmer is persistently low + dry mouth, discuss Sjögren's workup with ophthalmology + rheumatology; (4) corneal hypoesthesia can also relate to diabetes, post-LASIK, neuropathy — address relevant comorbidities.
What you can't skip: visual breaks (20-20-20), avoid direct AC/heater airflow, maintain indoor humidity, warm compresses (especially if coexisting MGD), adequate sleep, smoking cessation — these help every DED phenotype and are baseline TFOS DEWS II recommendations.
What not to do alone: don't adjust or stop Cyclosporine on your own; don't mix multiple eye drops (especially steroids) without supervision; don't decide 'I don't hurt so I'll stop'; don't skip Sjögren's evaluation if your doctor suggests it.
One final note on the 2026 DREAM secondary analysis: the study's main value isn't a new treatment but validating the long-observed phenomenon that subjective and objective severity can diverge widely, with a quantifiable, reproducible classification framework. For patients: your experience deserves to be taken seriously. For doctors: don't rely on a single metric. Hopefully this digest enables a deeper, more evidence-grounded conversation at your next visit.
📚 HsiaoEye Dry Eye Series
- 8 Dry Eye Myths — TFOS DEWS II guidelines, artificial tears, cyclosporine, Taiwan NHI
- Symptom-Sign Discordance — 2026 DREAM 535 patients, 77% discordant, neurosensory amplification (you are here)
References
- [本文主要來源] Matar K, Yu Y, Augello P, Ying GS, Asbell PA, Sayegh RR; DREAM Research Group. Defining Discordance Between Signs and Symptoms in Dry Eye Disease: Insights From the DRy Eye Assessment and Management (DREAM) Study. Am J Ophthalmol. 2026;286:41-49. doi:10.1016/j.ajo.2026.02.031
- [DREAM 試驗主結果] Asbell PA, Maguire MG, Pistilli M, et al. n-3 Fatty Acid Supplementation for the Treatment of Dry Eye Disease. N Engl J Med. 2018;378(18):1681-1690. doi:10.1056/NEJMoa1709691
- [DREAM 試驗設計] Asbell PA, Maguire MG, Peskin E, Bunya VY, Kuklinski EJ; DREAM Study Research Group. Dry Eye Assessment and Management (DREAM) study: Study design and baseline characteristics. Contemp Clin Trials. 2018;71:70-79. doi:10.1016/j.cct.2018.06.002
- [TFOS DEWS II 定義] Craig JP, Nichols KK, Akpek EK, et al. TFOS DEWS II Definition and Classification Report. Ocul Surf. 2017;15(3):276-283. doi:10.1016/j.jtos.2017.05.008
- [TFOS DEWS II 流行病學] Stapleton F, Alves M, Bunya VY, et al. TFOS DEWS II Epidemiology Report. Ocul Surf. 2017;15(3):334-365. doi:10.1016/j.jtos.2017.05.003
- [症狀-徵象不一致流行病學] Ong ES, Felix ER, Levitt RC, Feuer WJ, Sarantopoulos CD, Galor A. Epidemiology of discordance between symptoms and signs of dry eye. Br J Ophthalmol. 2018;102(5):674-679. doi:10.1136/bjophthalmol-2017-310633
- [Discordance 預測因子] Vehof J, Sillevis Smitt-Kamminga N, Nibourg SA, Hammond CJ. Predictors of discordance between symptoms and signs in dry eye disease. Ophthalmology. 2017;124(3):280-286. doi:10.1016/j.ophtha.2016.11.008
- [症狀徵象缺乏關聯原始觀察] Nichols KK, Nichols JJ, Mitchell GL. The lack of association between signs and symptoms in patients with dry eye disease. Cornea. 2004;23(8):762-770. doi:10.1097/01.ico.0000133997.07144.9e
- [神經病理性眼痛概念] Rosenthal P, Baran I, Jacobs DS. Corneal pain without stain: is it real? Ocul Surf. 2009;7(1):28-40. doi:10.1016/s1542-0124(12)70290-2
- [神經病理性眼痛系統回顧] Galor A, Levitt RC, Felix ER, Martin ER, Sarantopoulos CD. Neuropathic ocular pain: an important yet underevaluated feature of dry eye. Eye (Lond). 2015;29(3):301-312. doi:10.1038/eye.2014.263
- [乾眼症與慢性重疊疼痛狀況] Levitt AE, Galor A, Chowdhury AR, et al. Evidence That Dry Eye Represents a Chronic Overlapping Pain Condition. Mol Pain. 2017;13:1744806917729306. doi:10.1177/1744806917729306
- [Shtein 不一致乾眼症論文] Shtein RM, Harper DE, Pallazola V, et al. Discordant Dry Eye Disease (An American Ophthalmological Society Thesis). Trans Am Ophthalmol Soc. 2016;114:T4.
- [症狀-徵象相關性歷史研究] Sullivan BD, Crews LA, Messmer EM, et al. Correlations between commonly used objective signs and symptoms for the diagnosis of dry eye disease: clinical implications. Acta Ophthalmol. 2014;92(2):161-166. doi:10.1111/aos.12012
- [TFOS DEWS II 診斷方法] Wolffsohn JS, Arita R, Chalmers R, et al. TFOS DEWS II Diagnostic Methodology Report. Ocul Surf. 2017;15(3):539-574. doi:10.1016/j.jtos.2017.05.001
- [共聚焦顯微鏡與乾眼神經] Cox SM, Kheirkhah A, Aggarwal S, et al. Alterations in corneal nerves in different subtypes of dry eye disease: an in vivo confocal microscopy study. Ocul Surf. 2021;22:135-142. doi:10.1016/j.jtos.2021.08.004
- [性別差異] Lu PT, Lee CY, Sun CC. Sex differences and discordance between symptoms and signs of dry eye disease. Am J Ophthalmol. 2024;260:14-20. doi:10.1016/j.ajo.2023.10.008
- [台灣健保給付規定] 衛生福利部中央健康保險署. 《全民健康保險藥品給付規定》民國 115 年 4 月版，第 14 章「眼科用藥」14.5「人工淚液」、14.9.3「Cyclosporine 眼用製劑」、附表二十七「全民健康保險眼科含 cyclosporine 製劑事前審查申請書」。