Article Text

Download PDFPDF

Prevalence of refractive errors in the European adult population: the Gutenberg Health Study (GHS)
  1. Christian Wolfram1,
  2. René Höhn1,
  3. Ulrike Kottler1,
  4. Philipp Wild2,
  5. Maria Blettner3,
  6. Jens Bühren4,
  7. Norbert Pfeiffer1,
  8. Alireza Mirshahi1
  1. 1Department of Ophthalmology, University Medical Center Mainz, Mainz, Germany
  2. 2Department of Internal Medicine, University Medical Center Mainz, Mainz, Germany
  3. 3University Medical Center Mainz, Institute of Medical Biostatistics, Epidemiology and Informatics (IMBEI), Mainz, Germany
  4. 4Department of Ophthalmology, University Hospital Frankfurt, Frankfurt, Germany
  1. Correspondence to Dr Christian Wolfram, University Medical Center, Department of Ophthalmology Mainz, Langenbeckstr. 1, Mainz D-55131, Germany; christian.wolfram{at}


Objective To study the distribution of refractive errors among adults of European descent.

Design Population-based eye study in Germany with15 010 participants aged 35–74 years.

Methods The study participants underwent a detailed ophthalmic examination according to a standardised protocol. Refractive error was determined by an automatic refraction device (Humphrey HARK 599) without cycloplegia. Definitions for the analysis were myopia <−0.5 dioptres (D), hyperopia >+0.5 D, astigmatism >0.5 cylinder D and anisometropia >1.0 D difference in the spherical equivalent between the eyes. Exclusion criterion was previous cataract or refractive surgery.

Results 13 959 subjects were eligible. Refractive errors ranged from −21.5 to +13.88 D. Myopia was present in 35.1% of this study sample, hyperopia in 31.8%, astigmatism in 32.3% and anisometropia in 13.5%. The prevalence of myopia decreased, while the prevalence of hyperopia, astigmatism and anisometropia increased with age. 3.5% of the study sample had no refractive correction for their ametropia.

Conclusions Refractive errors affect the majority of the population. The Gutenberg Health Study sample contains more myopes than other study cohorts in adult populations. Our findings do not support the hypothesis of a generally lower prevalence of myopia among adults in Europe as compared with East Asia.

  • Epidemiology
  • Optics and Refraction
  • Public health

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.


Refractive errors are very common and affect the majority of the population. Due to the fact that most refractive errors can be corrected easily and usually have no direct pathological implications, the epidemiology of refractive errors has not been studied in depth, particularly not for the European population.

Population-based surveys in other regions of the world demonstrated that the prevalence of myopia tends to decrease with age, whereas the prevalence of hyperopia reveals the opposite tendency.1–6 Reports of possible gender differences in the prevalence of refractive errors have been controversial. Women have been found to be more hyperopic1 or more myopic.7 Other studies failed to confirm any gender differences.4 ,8 ,9

Prevalence rates of refractive errors have also been reported to differ by race, with Caucasians presenting higher prevalence rates for myopia and hyperopia than black or Hispanic subjects.3 A higher prevalence of myopia among Asians has been discussed,10 but recent epidemiological studies do not support this hypothesis.4 ,7

Both genetic and environmental factors have been found to play a role in the aetiology of refractive errors.11 Although there is evidence that refractive errors among siblings correlate closely,12 the genetic background of myopia remains unclear, as it may involve a complex interaction of multiple genes.13 Environmental associations have also been described between myopia and urban environment as well as with higher education.6

Possible ethnic, genetic and environmental differences make it difficult to project findings from other areas in the world and to apply them to the situation in Europe. This study is the first population-based cross-sectional study in Germany to assess the prevalence and distribution of refractive errors in the general adult population in Germany.


The Gutenberg Health Study (GHS) is a population-based, prospective, observational cohort study in the Rhine-Main region, located in the State of Rhineland-Palatine in midwestern Germany. It has a total of 15 010 participants aged 35–74 years who were evaluated after random sampling from the regional registration office. Aside from the investigation of ophthalmological conditions and eye diseases, the study also focuses on cardiovascular and metabolic diseases, cancer, diseases of the immune system and mental diseases. A detailed description of the GHS and the study methods have been published by Wild et al.14

Inclusion criteria were sufficient knowledge of the German language to understand the study documents, and the physical and psychological capabilities to travel to the study centre and cooperate during the investigations. Exclusion criteria for the analysis of refractive errors were previous cataract or refractive surgery or other reasons that may have distorted the refraction of the eye (such as a history of perforating corneal injury). Cases with information on only one eye were excluded as well in order to allow comparability of the refraction between both eyes.

The participants’ examinations took place at the study centre at the University Medical Center in Mainz between April 2007 and April 2012. Written informed consent was obtained from each participant prior to any examination, according to the tenets of the Declaration of Helsinki.

All participants underwent extensive medical examinations as well as a comprehensive ophthalmological examination. The refraction was determined automatically by an automatic refractor (Humphrey HARK 599). Cycloplegic medication was not used. Refractive errors were measured in spherical and cylindrical dioptres (D), where for cylindrical power a negative sign convention was used.

Our definitions for refractive errors were myopia <−0.5 spherical D, hyperopia >+0.5 spherical D, anisometropia >1.0 spherical D difference between eyes and astigmatism >0.5 cylinder D, as these are the most commonly used definitions for refractive errors in the literature. We then calculated the prevalence rates of different definitions of myopia (<−1.0/≤−6.0/≤−9.0 D) and hyperopia (>+3.0 D).

Participants were also asked about their use of glasses or contact lenses. The prevalence of uncorrected refractive errors was then assessed by the presence of myopia or hyperopia (according to the autorefractive result) in both eyes and the information that no refractive correction was used.

Prevalence rates of the respective refractive errors were assessed by computing absolute and relative frequency of refractive errors in the whole cohort and in subgroups after stratifying for age and gender. 95% CIs were calculated for all the relative frequencies of refractive errors in the population. Differences in frequencies of findings between men and women were assessed by logistic regression adjusting for age. Testing for age trends was performed by logistic regression adjusting for sex. Analyses were done using PASW Statistics V.20 and SAS V.9.2.


Of the entire study sample, 692 subjects had undergone cataract surgery and 74 subjects had had refractive surgery and were excluded from the analysis. An additional 64 subjects were excluded due to a history of corneal injuries, corneal transplantation or other conditions that may have affected the refraction. Data on refractive errors were missing in 237 subjects, including those whose refractive data were available for one eye only. Data were missing mainly for technical reasons or unwillingness to participate in the ophthalmic examination; thus, data of 13 959 study participants were included in the analysis.

The mean value of refractive errors in the study population was −0.401 D for right eyes and −0.395 D for left eyes. We observed no statistically significant difference between left and right eyes. For the purpose of easier illustration, the following findings therefore describe right eyes only.

Figure 1 shows the distribution of refractive errors within the study population. The range of refractive errors was −21.50 to +13.88 spherical D. The graph does not follow a Gaussian distribution as it is skewed towards myopia (v=−1.457).

Figure 1

Distribution of refractive errors in the Gutenberg Health Study.

The overall prevalences of refractive errors in the GHS are shown in table 1.

Table 1

Prevalence rates of refractive errors in the Gutenberg Health Study by sex and decades of age

The prevalence of myopia defined as <−1.0 D was 26.2% and for higher myopia (<−3.0, ≤−6.0 and ≤−9.0 D) 11.8%, 3.5% and 1.0%, respectively. Hyperopia of >3.0 D was prevalent in 4.5%.

Myopic errors were more common in male than female study participants, whereas hyperopia was observed more often in women (myopia p=0.021, hyperopia p=0.001). A more significant gender difference appeared regarding astigmatism, namely a higher prevalence in men (p=<0.0001).

We noted major differences in the prevalence of refractive errors among the age groups. Younger age groups showed a higher prevalence of myopia, whereas the older study participants displayed a higher prevalence of hyperopia. Over the age of 70 years, the percentage of hyperopes was again slightly lower (and percentage of myopes slightly higher) than in those aged 65–69 years, indicating that the potential hyperopic shift terminates at this age. The majority of the population in all age groups was affected by some refractive error; beyond the age of 60 years emmetropia was nearly as common as myopia (figure 2).

Figure 2

Myopia, hyperopia and emmetropia in the Gutenberg Health Study.

Uncorrected errors in subjects with binocular myopia or hyperopia and who did not use any correction (glasses or contact lenses) were observed in 3.5% (N=283) of the overall population. Uncorrected myopic errors were more common (85.1% of all uncorrected errors) than uncorrected hyperopic errors.


The GHS is the first population-based study of refractive errors in adults in Germany. There have only been a few epidemiological studies that have addressed the prevalence of refractive errors among adults in Europe, and their cohorts were small.8 ,15 To our knowledge, our study is the largest epidemiological study of refractive errors worldwide, and we are the first to provide reliable epidemiological data on the normal adult population in Germany.

When compared with major population-based studies on refractive errors among adults in the world, the GHS population is more myopic than most other study samples, which all used the same definitions for myopia and hyperopia (table 2).

Table 2

Prevalence rates of refractive errors among adults in population-based eye studies

The only study we identified with a higher prevalence of myopia is a Japanese study,4 which differs significantly from other Asian studies.6 ,7 ,16 A higher prevalence of myopia was reported in both Asian children17 and in Asian adult populations compared with Caucasian populations.1 ,5 ,8 As our study's proportion of myopic study participants was also very high (35.1%), we cannot confirm the hypothesis of a generally lower prevalence of myopia in European adults. Nor did we observe a major difference in the prevalence of higher myopia when comparing other studies’ findings in populations of different ethnicities. Higher myopia (−6 D) was present in 3.3% of the GHS population, which is lower than in the Japanese Tajimi study (5.5%)4 and higher than in the Beijing study (2.6%).6 Compared with the estimated US prevalence of higher myopia (defined there as ≤−5 D)3 of 4.5%, we noted a slightly higher prevalence of 5.6% in the GHS. The high prevalence of myopia in our study can be partly attributed to the fact that our study had lower age limits than most other studies, when considering the dynamics of refractive changes between different decades of age. However, other studies from China,16 India,18 Bangladesh19 and Barbados20 mostly had lower age limits and demonstrated even lower prevalences of myopia, which may also have to do with the fact that these study populations came from rural areas. As the environment of the GHS population in the Rhine-Main area in Germany is generally more urban, we assume that there may be a correlation between urbanity and higher prevalence of myopia, as demonstrated in the Beijing Eye Study.6

Our study reveals major differences in the prevalence of hyperopia and myopia in different age groups. All cross-sectional studies concur that hyperopia is more prevalent in older age groups. This hyperopic shift has also been found in the longitudinal perspective.21 A second—myopic—shift has been documented for the age group around 70 years of age1 ,4 due to the myopic effect of nuclear lens opacities and cataracts. This myopic effect is also present in our study as the prevalences in the oldest two age groups in our study indicate.

The association between refractive errors and gender has not been significant in most studies. In our study, myopia was more common among men and hyperopia among women, even though at relatively low levels of significance. A more pronounced gender difference was observed for the prevalence of astigmatism which was significantly higher in men. This has not been described before.

Uncorrected refractive errors are often addressed as major global public health problems.22 Even in industrial countries, the prevalence of uncorrected refractive errors has been considered a major reason for avoidable visual impairment. Estimations reveal that uncorrected refractive errors account for more than 75% of low vision cases.23 Our study, however, shows a different result, namely that uncorrected refractive errors are not very common in the GHS cohort, which is consistent with findings from the EPIC study in Norfolk (GB)24 that described only 1.9% of a normal adult population as having insufficient correction. As a limitation in our investigation, we must mention that we did not measure the subjects’ current refractive correction. A greater difference between the objective refraction and the current correction may have resulted in a higher prevalence of inadequate refractive correction.

Another potential limitation is that we did not employ cycloplegia in our study. As Krantz et al25 demonstrated in their study of more than 2500 adult participants that the difference in the refractive result in the spherical equivalent with or without cycloplegia was only 0.29 D, we may have overlooked some individuals with latent hyperopia, especially in the younger study participants; nevertheless, an additional cycloplegic examination would not have revealed major differences in our cohort.

As in any population-based study, our study population may not be representative of the general population. Due to the fact that the participation in the study took more than 5 h, we assume that the working population in the lower decades of age is slightly under-represented in our study as compared with the standard population for the study region. We therefore expect that the prevalence data for myopia in our study population are probably to some minor degree lower than in the general population, whereas the prevalence of hyperopia is probably higher.

In summary, the distribution of refractive errors in various populations reveals some differences and many similarities between different population-based studies. Even though ethnic differences for the prevalence of myopia were described among children,11 ,17 we believe that global differences may be less pronounced among adults. This hypothesis needs further investigation in two ways: first by improving the comparability of studies in particular through adjustment of prevalence data to standard populations and second by extending the knowledge about internal (eg, genetic) and external (eg, environmental and educational) influences on refractive errors.



  • Contributors The eight authors are justifiably credited with authorship, according to the authorship criteria. In detail: CW, NP and AM—conception, design, analysis and interpretation of data, drafting of the manuscript, final approval given; RH—acquisition of data, analysis and interpretation of data, final approval given; UK and PW —acquisition of data, critical revision of manuscript, final approval given; MB and JB—critical revision of manuscript, final approval given.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval Ethics committee of the University Medical Center Mainz.

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