Article Text

Download PDFPDF

Risk of corneal ulcer in patients with end-stage renal disease: a retrospective large-scale cohort study
  1. Ren-Long Jan1,2,
  2. Ming-Cheng Tai3,
  3. Shih-Feng Weng4,
  4. Chun Chang5,
  5. Jhi-Joung Wang6,7,
  6. Yuh-Shin Chang2,8
  1. 1 Department of Pediatrics, Chi Mei Medical Center, Liouying, Tainan, Taiwan
  2. 2 Graduate Institute of Medical Science, College of Health Science, Chang Jung Christian University, Tainan, Taiwan
  3. 3 Department of Ophthalmology, Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan
  4. 4 Department of Healthcare Administration and Medical Informatics, Kaohsiung Medical University, Kaohsiung, Taiwan
  5. 5 Department of Education, University of Taipei, Taipei, Taiwan
  6. 6 Department of Medical Research, Chi Mei Medical Center, Tainan, Taiwan
  7. 7 Department of Anesthesiology, Chi Mei Medical Center, Tainan, Taiwan
  8. 8 Department of Ophthalmology, Chi Mei Medical Center, Tainan, Taiwan
  1. Correspondence to Dr Yuh-Shin Chang, Department of Ophthalmology, Chi Mei Medical Center, 901 Zhonghua Rd, Yongkang District, Tainan 71004, Taiwan; yuhshinchang{at}


Background/Aims To investigate the risk of corneal ulcer in patients with end-stage renal disease (ESRD).

Methods This retrospective, nationwide, matched cohort study included 92 967 patients with ESRD recruited between 2000 and 2009 from the Taiwan National Health Insurance Research Database. The same number of age-matched and sex-matched patients without ESRD were selected from the Taiwan Longitudinal Health Insurance Database, 2000 as the control group. Data for each patient were collected from the index date until December 2011. Corneal ulcer incidence rate and risk were compared between the groups. A Cox proportional hazards regression was used to calculate the HR for a corneal ulcer after adjustment for potential confounders. The cumulative corneal ulcer incidence rate was calculated using Kaplan-Meier analysis.

Results In total, 660 patients with ESRD and 591 controls showed a corneal ulcer during follow-up; thus, the corneal ulcer incidence rate in patients with ESRD was 1.54 times (95% CI 1.38–1.72) that in the control patients. After adjustment for potential confounders, including diabetes mellitus and HIV disease, patients with ESRD were 1.17 times (95% CI 1.03 to 1.33) more likely to develop a corneal ulcer in the cohort for the total sample. Among patients with diabetes mellitus, the corneal ulcer incidence rate was significantly higher in the ESRD group, and diabetes mellitus significantly increased corneal ulcer risk even after adjustment for other confounders in the cohort.

Conclusion ESRD increases the risk of a corneal ulcer, particularly in patients with ESRD with diabetes mellitus. Regular ocular examinations are suggested for patients with ESRD.

  • corneal ulcer
  • end-stage renal disease
  • Taiwan Longitudinal Health Insurance Database

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.


End-stage renal disease (ESRD) is a major public health problem worldwide because of its high prevalence, poor prognosis and high healthcare utilisation.1 2 The rising ESRD prevalence burdens global healthcare resources and impacts patients and their families. Compared with other countries, Taiwan has a remarkably high incidence rate and prevalence of ESRD.3–5 There are several possible explanations for the high incidence and prevalence of ESRD in Taiwan, such as free coverage for dialysis therapy without copayment, provided by the launch of the National Health Insurance (NHI), increased survival rate and overall life expectancy, owing to the better healthcare system and low transplantation and mortality rates in patients undergoing dialysis.3

Corneal ulcers are a leading cause of visual impairment and blindness worldwide and are caused by infection from pathogens including bacteria, viruses and fungi.6–8 Patients with a corneal ulcer often present with redness, tearing, pus formation, severe pain, photophobia or blurred visual acuity. Corneal ulcers are commonly associated with dry eye, which typically arises from an unbalanced tear film related to multifactorial conditions, topical ophthalmic corticosteroid use and inflammation. Furthermore, wearing contact lenses and ocular surface injury increase corneal ulcer risk.6–8

Dry eye symptoms and a compromised ocular surface are more common among patients with ESRD. Several reports demonstrated reduced basal tear secretion and tear film stability in patients with ESRD.9–11 Dry eye symptoms and a compromised ocular surface are important predisposing factors for a corneal ulcer.7 8 Additionally, ocular symptoms including itching, burning, irritation and a foreign body sensation are well known and common complications in patients with ESRD.12 13 A higher risk for increased ocular problems necessitates that patients with ESRD use topical steroid or lubricant eye drops in the long term. Meanwhile, several reports considered long-term use of steroid or lubricant eye drops a leading cause of corneal ulceration.6–8 Recently, many studies showed that toll-like receptor activation may play roles in inflammation processes resulting in ESRD development.14 15 Toll-like receptors are expressed in renal cells and on the ocular surface; activation involves both innate immunity and various inflammatory ocular surface conditions.16 Therefore, it is clinically relevant to determine whether ESRD is a predictor of a corneal ulcer. However, no previous studies have discussed the association between ESRD and corneal ulcers. Therefore, we designed a cohort study using a nationwide population-based dataset to evaluate the association between ESRD and corneal ulcers in Taiwan. In our cohort study, all the corneal ulcers included were infected corneal ulcers, because we excluded the sterile ulcers.

Materials and methods


Taiwan launched a single-payer NHI scheme on 1 March 1995, which provides extensive medical care coverage for all residents. As of 2007, the programme enrolled >98% of the total Taiwanese population of 22.96 million. Data were obtained from the Taiwan National Health Insurance Research Database (NHIRD), which supplies enciphered identification numbers, the patients’ sexes and birth, admission and discharge dates. The NHIRD also includes the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnoses and procedure codes, prescription details and NHI-covered costs. A public database was used for the analysis; therefore, the requirements for ethical approval and informed consent were waived by the Institutional Review Board of Chi-Mei Medical Center, Tainan, Taiwan.

Study design

A new-onset ESRD group and a matched non-ESRD (control) group were involved in our retrospective, nationwide, matched cohort study.

Study participants

We recruited patients and controls from 2000 to 2009. In total, 92 967 patients with ESRD who started their first dialysis treatment after 31 December 2000 and received a code 585 catastrophic illness certificate between 1 January 2000 and 31 December 2009 were included. We excluded patients with missing data and of unknown sex. Patients diagnosed with a corneal ulcer (ICD-9-CM code 370.0, without marginal corneal ulcer (ICD-9-CM code 370.01) and Mooren’s ulcer (ICD-9-CM code 370.07)) prior to ESRD were also excluded.

For each patient with ESRD, one control without ESRD was randomly selected from the Longitudinal Health Insurance Database 2000, an NHIRD subset that contains full claim data for one million beneficiaries (4.34% of the total population) systematic and randomly selected in 2000. We matched the 92 967 controls by sex, age and index date. The index date for patients with ESRD was the first dialysis date, and the index date for the matched control was created by matching the ESRD subject’s index date. Moreover, we excluded controls diagnosed with a corneal ulcer before the index date. We followed each patient until the end of 2011 or until censoring because of death to determine the corneal ulcer incidence rate.

To determine all patients who developed a corneal ulcer, every patient was tracked from his or her index outpatient visit or hospitalisation through December 2011. Only patients who were evaluated by an ophthalmologist and diagnosed to have corneal ulcer were included. We recorded demographic data (eg, age and sex) and information on comorbidities including diabetes mellitus (ICD-9-CM code 250) and HIV disease (ICD-9-CM code 042 and V08). The inclusion criteria for diabetes mellitus and HIV disease were documentation of the condition at least once in an inpatient setting or ≥3 times in an ambulatory setting within 1 year prior to the initial ESRD diagnosis on the dialysis medical service date.

Statistical analysis

SAS V.9.4 for Windows (SAS Institute, Cary, North Carolina, USA) was used for analysis. Pearson’s Χ2 test was used to compare the demographic characteristics and comorbid disorders between the ESRD and control groups. The incidence rate was calculated as the number of corneal ulcer cases identified during follow-up divided by the total number of person-years (PY) for each group by age, sex and select comorbidities. The incidence rate ratio (IRR), which compares the risk of developing a corneal ulcer between the ESRD and control groups, was calculated using Pearson’s Χ2 test. A Cox proportional hazard regression analysis was performed to calculate the adjusted HR for developing a corneal ulcer. Cumulative incidence rates for corneal ulcers in ESRD were evaluated with a Kaplan–Meier analysis, and differences in cumulative incidence rate curves were analysed using the log-rank test. Furthermore, we subdivided patients into three age subgroups for further analysis: <50 years, 50–64 years and ≥65 years. Data are presented as the mean ±SD, and 95% CIs are provided when applicable. Statistical significance was defined as p<0.05.


Demographic data

Between 2000 and 2009, 92 967 patients with ESRD and 92 967 controls were recruited after excluding ineligible subjects. Table 1 provides demographic data and data on evaluated comorbidities for the patients with ESRD and age-matched and sex-matched controls. The mean age of the patients with ESRD and control patients was 62.23 (SD, 14.64) years. Of the 92 967 patients with ESRD, 46 039 (49.52%) were male and 46 928 (50.48%) were female, with 18 015 (19.38%) aged <50 years, 29 723 (31.97%) aged 50–64 years and 45 229 (48.65%) aged ≥65 years. Patients with ESRD exhibited a significantly higher prevalence of diabetes mellitus than the controls. The mean follow-up periods for the patients with ESRD and control patients were 4.70 (SD, 3.26) and 6.49 (SD, 2.96) years, respectively.

Table 1

Demographic characteristics and comorbid disorder comparisons between the ESRD and control groups

Corneal ulcer incidence rate

During the follow-up period, 1251 (1251/185 934, 0.67%) patients developed corneal ulcers; the proportion was significantly higher in patients with ESRD (660/92 967, 0.71%) than in controls (591/92 967, 0.64%; table 2). Additionally, there were significant differences in corneal ulcer incidence rate (ESRD, 15.10/10 000 PY; control, 9.81/10 000 PY) and IRR (1.54, 95% CI  1.38 to 1.72, p<0.001; table 2) between the groups.

Table 2

Risk of corneal ulcer in the ESRD and control groups

Regarding age, patients with ESRD aged <50 years exhibited the highest corneal ulcer incidence rate (18.66/10 000 PY), followed by those aged 50–64 years (17.03/1000 PY) and those ≥65 years (10.98/1000 PY). IRR values were significantly higher in the three ESRD age groups than among age-matched controls (table 2). Particularly, the incidence rate was 1.78 times higher among patients with ESRD aged <50 years than in age-matched controls (IRR, 1.78; 95% CI 1.43 to 2.21; p<0.001).

Corneal ulcer incidence rates were 16.11/10 000 PY for male patients with ESRD and 9.41/10 000 PY for male controls (IRR, 1.71; 95% CI 1.46 to 2.01; p<0.001). A significant difference was also observed between female patients with ESRD and female controls (IRR, 1.39; 95% CI 1.19 to 1.62; p<0.001; table 2).

Corneal ulcer incidence rates among diabetes mellitus was 20.93/10 000 PY and the IRR for corneal ulcer associated with diabetes mellitus indicated a significantly greater risk in patients with ESRD compared with control patients (IRR=1.96 (95% CI=1.50–2.56)).

Table 3 provides crude and adjusted corneal ulcer HRs during the follow-up period. After adjusting for age, sex and select comorbidities, ESRD remained an independent risk factor for a corneal ulcer (adjusted HR, 1.17; 95% CI 1.03 to 1.33); other significant risk factors in both groups included age <50 years (adjusted HRs (95% CI), 0.81 (0.70 to 0.93) for patients with ESRD aged 50–64 years and 0.63 (0.55 to 0.73) for patients with ESRD aged ≥65 years, p<0.05, table 3) and diabetes mellitus (adjusted HR, 1.82; 95% CI 1.59 to 2.07; p<0.05). Sex was not an independent risk factor for a corneal ulcer. Kaplan–Meier analyses revealed higher corneal ulcer cumulative incidence rates among patients with ESRD than controls; the log-rank test findings were also significant (p<0.0001; figure 1).

Table 3

Crude and adjusted HRs from Cox proportional hazard regressions and 95% CIs for corneal ulcer during the follow-up period in the study cohort

Figure 1

Cumulative incidence rate of corneal ulcer in patients with ESRD and controls during the follow-up period. ESRD, end-stage renal disease.


To our knowledge, our study is the first large-scale population-based study to evaluate the relationship between ESRD and a subsequent corneal ulcer. We analysed 92 967 patients with ESRD and 92 967 age-matched and sex-matched controls to explore the association between ESRD and corneal ulcers. The results show a significantly increased corneal ulcer risk in patients with ESRD compared with controls and indicate that ESRD was an independent risk factor for a corneal ulcer in the total cohort after accounting for age, sex, diabetes mellitus and HIV disease.

Many studies demonstrated that dry eye symptoms and a compromised ocular surface are common in patients with ESRD9–11; reduced lacrimation and lacrimal function impairment are frequent among these patients.10 A reduction in basal tear secretion might be postulated based on squamous metaplasia of the conjunctival epithelium and transformation of secretory conjunctival epithelium into nonsecretory keratinised epithelium.17 Besides tear secretion, tear film stability and tear quality were affected markedly in patients with ESRD.11 18 Once dry eye symptoms and a compromised ocular surface develop in patients with ESRD, the ocular surface may become vulnerable to various trauma and sources of infection. Therefore, a corneal ulcer is a frequently encountered complication of a dry-eye–related compromised ocular surface. In fact, dry eye symptoms and a compromised ocular surface resulting from xerophthalmia were important and significant risk factors for a corneal ulcer in many studies.7 8

Another possible pathogenic association between corneal ulcers and ESRD is the increased frequency of long-standing eye drop use in patients with ESRD for symptomatic control of irritated, red eyes, which is a common and well-known ocular complication in patients with ESRD.12 13 Irritated, red eyes could originate from aetiologies such as an increase in serum calcium or phosphate concentration, calcium phosphate metabolism disturbances and subsequent deposition of calcium phosphate salts as microcrystalline hydroxyapatite in the peripheral corneal epithelium and band keratopathy development.12 13 Recently, several studies showed that long-term steroid or lubricant eye drop use was a well-known and significant risk factor for corneal ulcers.6–8 The use of topical steroids in a simple corneal abrasion has the potential to cause bacterial and fungal invasion, leading to an infected corneal ulcer.19

Recently, several reports suggested that the toll-like receptors widely distributed in the renal tissue trigger chronic inflammation leading to renal fibrosis during ESRD development.14 15 Multiple toll-like receptors are expressed by corneal and can recognise and respond to invading pathogens on the corneal surface.16 Unfortunately, toll-like receptor activation potentially contributes to exacerbations of various ocular surface inflammatory conditions and destruction of the ocular surface during infection.16 The toll-like receptor might be involved in the association between ESRD and corneal ulcer.

The incidence rates of corneal ulcer were 15.10/10 000 PY in the ESRD group and 9.81/10 000 PY in the control group; these were higher than the general incidence rates of corneal ulcer according to Jeng’s report (2.76/10 000 PY) and the Olmsted County study (0.53/10 000 PY).20 21 The corneal ulcer incidence rate was higher in younger patients with ESRD aged <50 years. We explain the highest corneal ulcer incidence rate in patients with ESRD aged <50 years and lowest incidence rate in those aged ≥65 years with the following proposed reasons. First, the higher incidence rate in younger patients with ESRD probably implies a stronger inflammation reaction related to toll-like receptor activation, which exacerbates ocular surface dryness and destroys the ocular surface. In addition, younger patients with ESRD are more likely to wear contact lenses or engage in rigorous physical activities, thus increasing the risk of ocular surface trauma and subsequent corneal infection. Contact lens wear is a well-known predisposing risk factor for microbial keratitis at young ages.22 23 However, we could not evaluate the effect of this confounder because of a lack of data on contact lens wear in our databank from the NHIRD.

Corneal ulcers are vision-threatening. We evaluated several comorbidities: diabetes mellitus and HIV disease. However, only diabetes mellitus was associated with significantly higher corneal ulcer incidence rates in patients with ESRD than in controls and a significant risk factor in both groups. This finding is consistent with several previous studies that found diabetes mellitus to be a major risk factor for a corneal ulcer.24 25 Although Jeng’s study reported that HIV disease was a possible risk factor for corneal ulcers,20 we could not evaluate the relationship between HIV disease and corneal ulcers, as patients with HIV did not have corneal ulcers (tables 2 and 3).

This nationwide and population-based study has several strengths, including an accurate risk appraisal and high statistical power because the dataset included a large sample of patients with ESRD. The same database has been published in several previous studies.13 Additionally, because patients with visual discomfort visit an ophthalmologist rather than a general practitioner in Taiwan, there are reduced chances of misdiagnoses and selection bias at referral centres. Furthermore, this cohort study monitored corneal ulcer incidence rates in ESRD and control groups with maximum longitudinal data of 10 years.

There are some limitations to our study. Because the cohort’s medical history can only be traced to 1996, we could not confirm whether the controls had a history of ESRD before January 1996. In addition, we could not perform a chart review to ensure that the diagnoses recorded in the NHIRD were accurate and could not exclude the possibility of miscoding. Finally, several important confounding factors including contact lens use, occupation, mud or plant exposure and mild ocular trauma could not be evaluated.

In summary, our study shows that corneal ulcer risk was significantly higher in patients with ESRD than controls, and ESRD remained an independent risk factor after adjusting for diabetes mellitus and HIV disease. Moreover, diabetes mellitus was an independent risk factor for a corneal ulcer in patients with ESRD after adjustment for other confounders. These results suggest that clinicians should educate patients with ESRD about corneal ulcers to ensure appropriate and prompt ophthalmology visits. Furthermore, it is suggested that patients with ESRD must undergo regular ocular examinations.



  • R-LJ and M-CT contributed equally.

  • Handling editor James Chodosh.

  • Contributors All authors conceived the study. RLJ, MCT, SFW, CC and YSC conducted the study. RLJ, MCT, SFW and YSC analysed the results. JJW provided materials. RLJ, MCT and YSC wrote the article. All authors reviewed the manuscript.

  • Competing interests None declared.

  • Patient consent Detail has been removed from this case description/these case descriptions to ensure anonymity. The editors and reviewers have seen the detailed information available and are satisfied that the information backs up the case the authors are making.

  • Ethics approval Chi Mei Medical Center.

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

Linked Articles

  • At a glance
    Keith Barton James Chodosh Jost B Jonas