Epidemiology of primary glaucoma: prevalence, incidence, and blinding effects

Introduction

The main objectives of an epidemiological study are to describe the frequency of a disease in a given population and to identify risk factors that may be associated with the disease. As with many other

Corresponding author. Tel.: +39 0672596144; Fax: +39 062026232; E-mail: cedrone@uniroma2.it

diseases, the biggest methodological problem encountered in studies that aim to investigate the prevalence and incidence of glaucoma is the definition of the disease.

Prevalence of glaucoma

Since the 1920s, numerous studies have been conducted to determine the prevalence of glaucoma. In

4

many cases, however, the participants were simply persons with high intraocular pressure (IOP) because in those days glaucoma was regarded as synonymous with ocular hypertension (HIOP). After the introduction of gonioscopy (Barkan, 1938), two types of glaucoma were distinguished: primary open-angle glaucoma (POAG) and primary angle-closure glaucoma (PACG). Most of the prevalence studies cited above focused on POAG, which — according to the initial findings — affected 2% of individuals over 40 years of age. This figure reflected the number of subjects with HIOP (i.e., IOPW20 mmHg, a figure that represents the statistical mean of values observed in numerous population studies increased by a sum equal to twice the standard deviation of the mean), which is not always associated with optic neuropathy, a hallmark of glaucoma. It failed to include persons who did have glaucomatous changes in the optic disc and in the visual fields despite the fact that their IOP was statistically normal.

Since 1972, most of the figures cited above have been considered inaccurate representations of the prevalence of POAG because the studies that generated them failed to respect certain methodological requisites (Khan, 1972): (1) the study population must be well defined, and no subgroup should be systematically excluded from the investigation; (2) the criteria used to define persons affected by the disease must be clearly described;

(3)the number of subjects in the population who are eligible for enrollment must be specified; and

(4)the participation rate, i.e., the percentage of subjects actually examined must be reported (if possible, for each sex and age group).

The first study that was unanimously acknowledged to be well designed and appropriately conducted was the 1963 Ferndale Glaucoma Survey (Hollows and Graham, 1966), which examined 92% of the eligible population of a town in South Wales. POAG was diagnosed in subjects who presented an IOP of W 20 mmHg, cupping of the optic disc, and visual-field defects. This study introduced the concept of routine visual-field testing with the Friedman analyzer (even though this testing was done on only one out of three patients). This approach led to the identification of certain cases of low-tension

glaucoma (LTG), half of which were subsequently reclassified as POAG based on IOPs observed at later visits. In addition, the study revealed that the majority of cases of glaucoma are primary and of the open-angle type.

So far, 55 other population-based surveys have been conducted in the populations of an entire town, village, region, or nation or in those selected with a clearly defined random or clustered sampling procedure. The racial or ethnic groups examined in these studies can be broadly grouped into the following categories: Asian (Alsbirk, 1973; Arkell et al., 1987; Hu, 1989; Shiose et al., 1991; Rauf et al., 1994; Foster et al., 1996, 2000; Jacob et al., 1998; Dandona et al., 2000a, b; Metheetrairut et al., 2002; Bourne et al., 2003; Ramakrishnan et al., 2003; Iwase et al., 2004; Rahman et al., 2004; Yamamoto et al., 2005; Raychaudhuri et al., 2005; Vijaya et al., 2005, 2006, 2008a, b; He et al., 2006; Casson et al., 2007); Black (Wallace and Lovell, 1969; Mason et al., 1989; Tielsch et al., 1991; Wormald et al., 1994; Leske et al., 1994; Buhrmann et al., 2000; Rotchford and Johnson, 2002; Ekwerekwu and Umeh, 2002; Rotchford et al., 2003; Ntim-Amponsah et al., 2004; Friedman et al., 2006); Hispanic (Quigley et al., 2001; Anton et al., 2004; Varma et al., 2004); Mixed (Salmon et al., 1993; Sakata et al., 2007); White (Hollows and Graham, 1966; Bankes et al., 1968; Leibowitz et al., 1980; Bengtsson, 1981; Martinez et al., 1982; Gibson et al., 1985; Tielsch et al., 1991; Ringvold et al., 1991; Klein et al., 1992; Coffey et al., 1993; Dielemans et al., 1994; Leske et al., 1994; Hirvela et al., 1994; Giuffre´et al., 1995; Ekstrom, 1996; Mitchell et al., 1996; Cedrone et al., 1997; Bonomi et al., 1998; Reidy et al., 1998; Wensor et al., 1998; Kozobolis et al., 2000; Friedman et al., 2006; Sakata et al., 2007; Topouzis et al., 2007). Prevalence rates of POAG and PACG are reported in Tables 1–3.

These studies varied widely in terms of the eye examination methods and case definitions used, in particular the criteria adopted for defining glaucomatous nerve damage. With the advent of automated perimetry, the definitions of glaucoma used in epidemiologic studies improved. With manual kinetic field testing, the outcome of the examination and the reliability of findings were

dependent on the psychological/physical conditions of the patient and of the examiner. The automated examination allowed objective evaluation of the reliability of the findings, which are generally rated as good when false positive or false negative rates are r33% and loss of fixation rates are o20%. Furthermore, it was more sensitive and specific than manual perimetry and thus capable of detecting early visual-field changes that were missed with older methods.

The Dalby Study (1977–1978) deserves mention because it was the first study in which all participants underwent visual-field testing with the automated perimeter developed and validated by Heiji and Krakau, which is the prototype of the Humphrey perimeter (Bengtsson, 1981). As a result, all cases of LTG in the population examined in this study should have been diagnosed. Despite its improved reliability, visual-field testing with an automated perimeter had its drawbacks. For one thing, it prolonged the eye examination by approximately 30 min. Therefore,

in many of the less recent studies, visual fields were not tested in all participants because it represented a true ‘‘bottle-neck’’ in the examination scheme. With the recently introduced program based on the Swedish Interactive Threshold Algorithms (SITA), examination times are considerably shorter than those associated with the program based on the Full Threshold Algorithm. In the assessment of the central 301 of the visual field, the SITA Standard and SITA Fast programs reduce test times by 50% and 66%, respectively. More recently, the introduction of perimetry based on the frequency doubling technology (FDT) has allowed increasingly rapid detection of even earlier changes in the visual field (less than 2 min per eye).

These new methods for visual-field testing have drastically reduced the duration of eye examinations in prevalence studies. This factor is particularly important in research settings because the reliability of prevalence figures is inversely proportional to the time required for examination of the total sample. The prevalence of a disease is

calculated by dividing the number of cases ascertained at a given time by the number of subjects in the population at that time. If case ascertainment is significantly prolonged — e.g., 2 years — the onset of some of the cases identified near the end of the study may have occurred long after the beginning of the study. In theory, with respect to the initiation of the study, these would represent incident rather than prevalent cases. Furthermore, by the end of the 2-year ascertainment period, glaucoma might have developed in some subjects who were examined in the early phases of the study and found to be healthy.

Visual-field testing of only some of the participants in a study can obviously lead to underestimations of the prevalence of glaucoma (Mason et al., 1989). However, in a substantial number of subjects who do not have glaucoma, the results of field tests meet the criteria for abnormality. Therefore, use of these results without an examination of the optic disc can lead to overestimation of glaucoma prevalence. A thorough examination of the optic disc is a fundamental part of the threetiered system of evidence (Foster et al., 2002), which was recently used to categorize glaucoma in almost all recent population-based prevalence surveys (Bourne et al., 2003; Rotchford et al., 2003; Iwase et al., 2004; Rahman et al., 2004; Raychaudhuri et al., 2005; Yamamoto et al., 2005; Vijaya et al., 2005, 2006, 2008a, b; He et al., 2006; Casson et al., 2007; Topouzis et al., 2007; Sakata et al., 2007). The prototype system was discussed by the ISGEO Glaucoma classification working group at the congress of the International Society for Geographical and Epidemiological Ophthalmology held in Leeuwenhorst, the Netherlands, in June 1998.

The rationale underlying this new classification system is that, although the level of IOP is one of the most consistent risk factors for the presence of glaucoma, the concept that statistically raised IOP is a defining characteristic for glaucoma has been almost universally discarded. This is based on several population-based studies that document the typical disc and field damage of glaucoma in people with a statistically normal IOP and, conversely, people with statistically elevated IOP and no evidence of optic neuropathy (Foster et al., 2002). The authors proposed to follow this current

7

convention except for category 3 diagnosis, as detailed below. The diagnosis of glaucoma is made according to the following three categories:

Category 1 (structural and functional evidence): Eyes with a cup-to-disc ratio (CDR) or CDR asymmetry W97.5th percentile for the normal population, or a neuroretinal rim width reduced to o0.1 CDR (between 11 and 1 o’clock or 5 and 7 o’clock) that also showed a definite visual-field defect consistent with glaucoma.

Category 2 (advanced structural damage with unproved field loss): If the subject could not satisfactorily complete visual-field testing but had a CDR or CDR asymmetry W99.5th percentile for the normal population, glaucoma was diagnosed solely on the structural evidence. In diagnosing category 1 or 2 glaucoma, there should be no alternative explanation for CDR findings (dysplastic disc or marked anisometropia) or the visual-field defect (retinal vascular disease, macular degeneration, or cerebrovascular disease).

Category 3 (optic disc not seen. Field test impossible): If it is not possible to examine the optic disc, glaucoma is diagnosed if: (A) The visual acuity is o3/60 and the IOP W99.5th percentile, or (B) The visual acuity is o3/60 and the eye shows evidence of glaucoma filtering surgery, or medical records were available confirming glaucomatous visual morbidity.

POAG is therefore optic nerve damage meeting any of the three categories of evidence below, in an eye which does not have evidence of angle closure on gonioscopy, and where there is no identifiable secondary cause. PACG is optic nerve damage meeting any of the three categories of evidence above, in an eye which has evidence of angle closure on gonioscopy, according to the following