Letter to Editor RE “Diagnosing dry-eye: Which tests are most accurate?”

We read with interest the recent paper by Eric Papas concerning the appropriate testing for diagnosing dry eye [[1]Papas E. Diagnosing dry-eye: Which tests are most accurate? Contact Lens and Anterior Eye XXXXX.Google Scholar], and were concerned by the comment that “Confronted with these insights, clinicians might be forgiven for concluding that it would be better (and cheaper) to toss a coin instead of following the current TFOS DEWS II recommendations.” which could be taken out of context, even though the article goes on to state “This would be unfair however, since the guidelines actually specify the diagnostic hurdle as being the presence of “symptoms and at least one positive result of the markers of homeostasis”.[[2]Wolffsohn J.S. Arita R. Chalmers R. Djalilian A. Dogru M. Dumbleton K. et al.TFOS DEWS II Diagnostic Methodology Report.Ocular Surface. 2017; 15: 539-574Crossref PubMed Scopus (1117) Google Scholar]. We certainly agree with the author that diagnosis is central to the role of any clinician and consensus is critical to the patient (for clarity and consistency between clinicians), to the clinician (for consistency with fellow eye care professionals and to inform the treatment approach) and to regulators (for accurate prevalence estimation and allocation of resources). However, merely relying on sensitivities and specificities, which is the modelling approach undertaken in this manuscript, is fundamentally flawed, as outlined in the Tear Film and Ocular Surface Society (TFOS) Diagnostic report (section 5)[[2]Wolffsohn J.S. Arita R. Chalmers R. Djalilian A. Dogru M. Dumbleton K. et al.TFOS DEWS II Diagnostic Methodology Report.Ocular Surface. 2017; 15: 539-574Crossref PubMed Scopus (1117) Google Scholar]. That sensitivity and specificity are not appropriate for diagnostic reasoning has also been identified by authors outside of the ophthalmic field [3Llewelyn H. Sensitivity and specificity are not appropriate for diagnostic reasoning.BMJ. 2017; 358j4071Google Scholar, 4Moons K.G. van Es G.A. Deckers J.W. Habbema J.D. Grobbee D.E. Limitations of sensitivity, specificity, likelihood ratio, and bayes' theorem in assessing diagnostic probabilities: a clinical example.Epidemiology. 1997; 8: 12-17Crossref PubMed Scopus (157) Google Scholar]. This is primarily because there is no ‘gold standard’ test/techniue to compare the diagnosis against. The criteria by which the ‘disease’ group is chosen will lead to spectrum and selection bias (excluding individuals that do not fit the ‘healthy’ or ‘disease’ criteria set and recruiting a ‘disease’ group with more severe disease will lead to artificially raised sensitivity and specificity) and selection bias (when efficacy of metrics that were used in the selection and differentiation of subjects are directly compared to a novel test that was not used as part of the inclusion criteria) [[2]Wolffsohn J.S. Arita R. Chalmers R. Djalilian A. Dogru M. Dumbleton K. et al.TFOS DEWS II Diagnostic Methodology Report.Ocular Surface. 2017; 15: 539-574Crossref PubMed Scopus (1117) Google Scholar]. This will lead to much of the variability evident in the author’s table 2 [[1]Papas E. Diagnosing dry-eye: Which tests are most accurate? Contact Lens and Anterior Eye XXXXX.Google Scholar], highlighting the problem when there is no consensus around the definition/diagnosis of a disease. Moreover, the high heterogeneity in the methodology and reference standards of individual diagnostic studies included in the modelling may significantly compromise the clinical utility and applicability of the trends highlighted by the current study, which would therefore warrant judicious interpretation. Some signs are also found to occur during later and more severe stages of the disease, such as ocular surface staining [2Wolffsohn J.S. Arita R. Chalmers R. Djalilian A. Dogru M. Dumbleton K. et al.TFOS DEWS II Diagnostic Methodology Report.Ocular Surface. 2017; 15: 539-574Crossref PubMed Scopus (1117) Google Scholar, 5Wang M.T.M. Muntz A. Lim J. Kim J.S. Lacerda L. Arora A. et al.Ageing and the natural history of dry eye disease: A prospective registry-based cross-sectional study.The Ocular Surface. 2020 Oct; 18: 736-741Crossref PubMed Scopus (31) Google Scholar], and may therefore be associated with higher diagnostic specificity. Thus, the author recommending “that this criterion be specified in diagnostic guidelines for dry eye disease” [[1]Papas E. Diagnosing dry-eye: Which tests are most accurate? Contact Lens and Anterior Eye XXXXX.Google Scholar] is biased toward patients with longstanding disease. The paper [[1]Papas E. Diagnosing dry-eye: Which tests are most accurate? Contact Lens and Anterior Eye XXXXX.Google Scholar] refers multiple times to a ‘correct’ diagnosis, but this is impossible to define unless unified criteria have been applied, which is not the case across the range of studies reviewed. In addition, the parameters modelled will vary depending on how the tests are performed. In regard to ocular surface staining, specific examples might include fluorescein volume [[6]Mooi J.K. Wang M.T.M. Lim J. Müller A. Craig J.P. Minimising instilled volume reduces the impact of fluorescein on clinical measurements of tear film stability.Contact Lens Anterior Eye. 2017 Jun; 40: 170-174Abstract Full Text Full Text PDF PubMed Scopus (41) Google Scholar] and instillation location for assessing fluorescein breakup time and corneal staining, as well as the illuminating light spectrum and observational cut-off filter [[7]Peterson R.C. Wolffsohn J.S. Fowler C.W. Optimization of Anterior Eye Fluorescein Viewing.Am J Ophthalmol. 2006; 142: 572-575.e2Abstract Full Text Full Text PDF PubMed Scopus (37) Google Scholar]; also whether lissamine green or fluorescein are used for conjunctival staining [[8]Eom Y. Lee J.-S. Keun Lee H. Myung Kim H. Suk Song J. Comparison of conjunctival staining between lissamine green and yellow filtered fluorescein sodium.Canadian J Ophthalmol. 2015; 50: 273-277Abstract Full Text Full Text PDF PubMed Scopus (27) Google Scholar]. The modelling in this paper confirms that single tests will be less accurate [[1]Papas E. Diagnosing dry-eye: Which tests are most accurate? Contact Lens and Anterior Eye XXXXX.Google Scholar], which is unsurprising. As the disease diagnosis must align with its definition (and the TFOS DEWS II definition includes both signs and symptoms [[9]Craig J.P. Nichols K.K. Akpek E.K. Caffery B. Dua H.S. Joo C.-K. et al.TFOS DEWS II Definition and Classification Report.Ocular Surface. 2017; 15: 276-283Crossref PubMed Scopus (1751) Google Scholar]), the TFOS DEWS II diagnostic criteria requires at least two predefined criteria from a limited range of options to be met for a diagnosis to be made [[2]Wolffsohn J.S. Arita R. Chalmers R. Djalilian A. Dogru M. Dumbleton K. et al.TFOS DEWS II Diagnostic Methodology Report.Ocular Surface. 2017; 15: 539-574Crossref PubMed Scopus (1117) Google Scholar]. Adding multiple tests (performed consistently) will improve the sensitivity and specificity in making a diagnosis, but at the risk of fewer clinicians having the time, expertise and equipment to make that diagnosis.[[2]Wolffsohn J.S. Arita R. Chalmers R. Djalilian A. Dogru M. Dumbleton K. et al.TFOS DEWS II Diagnostic Methodology Report.Ocular Surface. 2017; 15: 539-574Crossref PubMed Scopus (1117) Google Scholar] As signs and symptoms are acknowledged not to be strongly correlated in dry eye disease [[2]Wolffsohn J.S. Arita R. Chalmers R. Djalilian A. Dogru M. Dumbleton K. et al.TFOS DEWS II Diagnostic Methodology Report.Ocular Surface. 2017; 15: 539-574Crossref PubMed Scopus (1117) Google Scholar], it will also exclude a large number of people with dry eye, such that the highly sensitive and specific diagnosis will not, in fact, be ‘correct’! [[2]Wolffsohn J.S. Arita R. Chalmers R. Djalilian A. Dogru M. Dumbleton K. et al.TFOS DEWS II Diagnostic Methodology Report.Ocular Surface. 2017; 15: 539-574Crossref PubMed Scopus (1117) Google Scholar]. Hence, clinicians would absolutely NOT “be forgiven for concluding that it would be better (and cheaper) to toss a coin”[[1]Papas E. Diagnosing dry-eye: Which tests are most accurate? Contact Lens and Anterior Eye XXXXX.Google Scholar] and should still follow the well-established and carefully selected TFOS DEWS II diagnostic recommendation [[2]Wolffsohn J.S. Arita R. Chalmers R. Djalilian A. Dogru M. Dumbleton K. et al.TFOS DEWS II Diagnostic Methodology Report.Ocular Surface. 2017; 15: 539-574Crossref PubMed Scopus (1117) Google Scholar] until such time as improved consensus criteria are developed. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.