Statistics from Altmetric.com
New incidence and prevalence studies required
Weber designed the first applanation tonometer in 1867, Goldmann constructed the modern applanation tonometer in 1954, but only 27 years ago Ehlers et al1 presented their study that showed the correlation between central corneal thickness (CCT) and intraocular pressure (IOP) applanation tonometric measurement. The number of publications about central corneal thickness measurements has risen exponentially over the past few years,2 but do we use this information? The implementation of this inverse correlation between CCT and IOP on the glaucoma/ocular hypertension (OHT) patients has a great impact on our knowledge, understanding, classification, and treatment of glaucoma. This means that many OHT patients should be reclassified as normal; from 30% as Argus3 found or even up to 65% according to Herndon et al.4 Of the normal tension glaucoma (NTG) patients a similar proportion can be reclassified; 31% according to Copt et al.5 But this is only the beginning. All the data concerning prevalence of OHT, NTG, and primary open angle glaucoma (POAG) must be revised as well as all the data concerning natural history and typical morphological changes of OHT and NTG. One of the most important consequences should be that corneal pachymetry is a must when conducting a study with OHT or NTG patients, otherwise the population study is false and the results are not valid. Nevertheless, most publications dealing with OHT and NTG patients do not put these criteria to use, and we continue to gather inaccurate information. The Ocular Hypertension Treatment Study6 has to readjust its eligibility assessment—include CCT pachymetry—otherwise the results of this important study will be doubted.
Visual field examination and interpretation is one of the pillars of ophthalmological diagnoses. Bengtsson et al7 introduced in 1997 the Swedish interactive threshold algorithm (SITA), a static perimetry operating system/software package. SITA standard presents a reduced number of stimuli by an average of 25% and test time reduction up to 50%.7–9 SITA has greater patient acceptance than full threshold strategy.10,11 The mean sensitivity is about 1 dB higher than in a full threshold program and the age dependent decrease of mean sensitivity is 25% smaller.12,13 The intersubject variance is 30% smaller.9,13 In glaucoma patients the SITA standard shows a larger number of significantly depressed points.14 Perimetry is still the gold standard for most ophthalmologists for glaucoma monitoring. Should we be satisfied with a blurred presentation of visual field instead of a clear and sharp presentation only because of greater patient acceptance? All the deviations of sensitivity in comparison with full threshold strategy are the result of the algorithm behind SITA, as well as the lower intersubject variance that is flattened by an algorithm, and not because of shorter examination time. It can be deduced, if we look at a similar strategy, that tendency oriented perimetry (TOP), has similar drawbacks.15,16
What we are measuring is the ability of a method to detect the patients who were classified as having glaucoma with another method . . . but not the overall ability to detect the pathology
In the Hippocratic aphorisms the term glaucoma is first mentioned as a diagnosis. But, still there is no agreement over the definition of glaucoma or what are the criteria to diagnose glaucoma. A review of 182 articles about POAG published in 1980, 1985, 1990, and 1995 in the American Journal of Ophthalmology, Ophthalmology, and Archives of Ophthalmology showed that only 66% included a definition of POAG.17 Among these, there are various combinations of optic disc visual field criteria and IOP to define POAG. In the population based Rotterdam Study, different commonly used criteria for the diagnosis of POAG resulted in prevalence figures ranging from 0.1% to 1.2%.18 Millions of people all over the world are treated for glaucoma—in the United States more than one million. Bohn et al19 have shown that in almost half (44.7%) of 544 patients who were treated for POAG, the only parameter used for diagnosis was IOP. As long as there are no accepted criteria for glaucoma diagnosis the correct time for first medical intervention is unclear and studies are difficult to compare. The same problem arises as we try to compare various screening methods and calculate their sensitivity and specificity. Furthermore, no method can detect all glaucoma cases.20–23 The sensitivity and specificity is falsified, as some people who have glaucoma and who are detected with one method will not be detected by another method and vice versa. Actually, what we are measuring is the ability of a method to detect the patients who were classified as having glaucoma with another method, which is inaccurate in itself, but not the overall ability to detect the pathology. To minimise those problems a classification and definition system should be agreed upon.
A final remark, over the years it has been suggested that the new diagnostic tools enable us to detect glaucoma much earlier. Is it not time to implement these methods and to do new population based incidence and prevalence studies of glaucoma with the new diagnostic devices? I would expect a shift to the left of the age distribution. Glaucoma is ready for an update.
Note in Proof
New incidence and prevalence studies required