Diabetic Patients With Rosacea Increase the Risks of Diabetic Macular Edema, Dry Eye Disease, Glaucoma, and Cataract : The Asia-Pacific Journal of Ophthalmology

Secondary Logo

Journal Logo

Original Study

Diabetic Patients With Rosacea Increase the Risks of Diabetic Macular Edema, Dry Eye Disease, Glaucoma, and Cataract

Wang, Fang-Ying MD*,†,‡; Kang, Eugene Yu-Chuan MD‡,§,∥; Liu, Chun-Hao MD‡,¶; Ng, Chau Yee MD†,‡; Shao, Shih-Chieh MS#; Lai, Edward Chia-Cheng PhD**; Wu, Wei-Chi PhD‡,§; Huang, Yi-You MD*; Chen, Kuan-Jen MD‡,§; Lai, Chi-Chun MD‡,††; Hwang, Yih-Shiou PhD‡,§,‡‡,§§

Author Information
Asia-Pacific Journal of Ophthalmology: November/December 2022 - Volume 11 - Issue 6 - p 505-513
doi: 10.1097/APO.0000000000000571



Rosacea is a chronic inflammatory disease with variable presentations, but the lesions are mainly found over the central face; accordingly, it is classified into 4 subtypes: erythematotelangiectatic rosacea, papulopustular rosacea, phymatous rosacea, and ocular rosacea.1 Rosacea has been associated with several systemic comorbidities, such as gastrointestinal, cardiovascular, neurological, psychiatric, metabolic, and autoimmune diseases,2–5 probably due to its complicated pathophysiology involving genetic factors, environmental stimulation, microbiome colonization or infection, disturbed neurovascular responses, and immunological dysregulation, which gives rise to localized and even systemic recurrent inflammation.2,3,5 Rosacea is not only a cutaneous disease but is more akin to a systemic inflammatory disease and may be related to the gut-skin axis.6,7 However, because of the characteristic relapse-and-remission course, long-term treatment of rosacea may be required. These medications possibly play a role in systemic comorbidities.6 Ocular rosacea is defined as eyelid and ocular surface inflammation, typically with symptoms of foreign body sensation, photophobia, tearing, conjunctival hyperemia, meibomian gland inflammation, and blurred vision.8,9 However, due to its systemic inflammatory characteristic of rosacea, its association with eyes is likely to extend considerably beyond mere ocular rosacea.

Diabetes mellitus (DM) is a leading chronic disease worldwide, and its prevalence continues to increase.10 The systemic involvement of DM leads to several macrovascular and microvascular complications,11 causing a great economic burden.12 In ophthalmology, DM is associated with several ocular diseases, including cataract, glaucoma, diabetic retinopathy (DR), and diabetic macular edema (DME).13,14 Considering the pathophysiology in diabetes, this metabolic disease is suggested to have a systemic inflammatory component.15 Inflammation in response to hyperglycemia can accelerate atherosclerosis and diabetic vascular complications.16,17 Inflammation can also deteriorate glucose control and lead to hyperglycemia, generating a vicious cycle in diabetes.18 In addition to the angiogenesis, inflammation also affects the pathogenesis of diabetic eye diseases.19–21

The role of inflammation in diabetic eye diseases such as DR has been highlighted.22 In addition, anti-inflammatory drugs are being assessed for the management of diabetic eye diseases.23 Because rosacea involves systemic inflammation, which may indirectly affect diabetes control, it may be associated with ocular complications of diabetes. However, no study has yet evaluated this association. In this study, we investigated the association between rosacea and eye diseases in patients with diabetes.


Data Source

This retrospective cohort study used claims data from the National Health Insurance (NHI) Research Database (NHIRD) released by the Ministry of Health and Welfare in Taiwan. Disease diagnoses and procedures were coded according to the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM), and Taiwan NHI reimbursement code. Detailed information regarding NHI and NHIRD has been described previously.24–26 The study was approved by the Chang Gung Memorial Hospital Institutional Review Board (no. 201900967B0) and adhered to the tenets of the Declaration of Helsinki. Informed consent was waived because the data were all deidentified, and no personal information could be disclosed.

Inclusion of the Study Patients

Patients with a diagnosis code of DM and the use of any hypoglycemic agents between January 1, 1997, and December 31, 2013, were included. The presence of rosacea (ICD-9-CM: 695.3) was ascertained using ≥2 outpatient diagnoses in dermatology.27,28 For patients diagnosed as having rosacea before DM, the index date was the date of DM diagnosis. For patients diagnosed as having rosacea after DM, the index date was the date of rosacea diagnosis.29 For each patient in the rosacea group, we included 20 age-matched and sex-matched patients in the nonrosacea group.29 Matching was performed using an age range of ±2 years. We excluded patients with missing demographics, history of retinal disorders (DR, retinal vascular occlusion, separation of retinal layers, degeneration of retina, chorioretinal inflammations, and other retinal disorders), history of receiving vitreoretinal intervention (intravitreal injection, laser, scleral buckling, and vitrectomy), and diagnosis with similar symptoms to rosacea (seborrheic dermatitis and cutaneous lupus erythematosus) (Fig. 1).

Flowchart of the study enrollment and exclusion.


Covariates were demographics (age and sex), proxy variables for DM severity (duration of diabetes, diabetic neuropathy, and foot ulcers), number of retinal examinations in the previous year, type of DM, comorbidities (dyslipidemia, hypertension, ischemic heart disease, chronic kidney disease, and stroke), as well as Charlson Comorbidity Index score,30 antidiabetic medications for proxy variables for glucose control, antihypertensive medications, and other medications (antiplatelet, statins, and fibrates). Comorbidities and diseases were identified as having any inpatient diagnosis or 2 outpatient diagnoses in the previous year before the index date. Medications were identified according to the prescriptions within the 3 months before the index date in either the outpatient or pharmacy claims data.


The primary outcomes were new-onset ocular diseases (newly diagnosed DR, DME, glaucoma with medical treatment, dry eye disease, age-related macular degeneration, retinal vascular occlusion, optic neuritis, ischemic optic neuropathy, retinal detachment, and uveitis) and ocular interventions (retinal laser, intravitreal injection, vitrectomy, cataract surgery, and glaucoma interventions). Glaucoma with medical treatment was defined as the diagnosis of glaucoma combining any intraocular pressure–lowering agent used at the same visit. Glaucoma interventions included iridotomy, trabeculoplasty, cyclophotocoagulation, trabeculotomy, trabeculectomy, and shunting surgery. The occurrence of ocular diseases was defined as having ≥2 new-onset outpatient diagnoses or any inpatient diagnosis at ophthalmology. The ocular intervention was identified using Taiwan reimbursement codes.31,32

On the basis of previous relevant publications, we also defined secondary positive control outcomes as 3 disease groups: systemic inflammatory diseases (psoriasis, ankylosing spondylitis, Behcet syndrome, rheumatoid arthritis, and inflammatory bowel diseases), gastrointestinal diseases (peptic ulcer, irritable bowel syndrome, and esophageal reflux), and psychiatric diseases (anxiety, major depression disorder, bipolar disorder, and schizophrenia). The occurrence of the aforementioned diseases was defined as having at least 2 new-onset outpatient diagnoses or any inpatient diagnosis at a relevant department, such as dermatology or rheumatology for psoriasis, and rheumatology or gastroenterology for other systemic inflammatory and gastrointestinal diseases. The ICD-9-CM codes used to define diseases (including covariates and outcomes) in this study are shown in Supplementary Digital Content Table 1 (Supplemental Digital Content 1, https://links.lww.com/APJO/A186).


Propensity score matching (PSM) was conducted to balance the distribution of covariates between groups. The propensity score was the predicted probability of being in the rosacea group given certain values of covariates by the multivariable logistic regression without considering interaction effects. The variables used to calculate propensity scores were all of the covariates, in which the follow-up duration was replaced with the index date (Table 1). The matching was processed using a greedy nearest neighbor algorithm with a caliper of 0.2 times the SD of the logit of the propensity score, with random matching order and without replacement. The balance of the baseline characteristics between groups was checked using the absolute value of standardized difference between the groups, where a value <0.1 was considered a negligible difference.

TABLE 1 - Characteristics of Patients With Diabetes With and Without Rosacea Before and After PSM
Before PSM After PSM
Variables Rosacea (n=4191) Nonrosacea (n=92,154) STD Rosacea (n=4096) Nonrosacea (n=16,384) STD
Male 1622 (38.7) 36,940 (40.1) −0.03 1596 (39.0) 6365 (38.8) <0.01
Age (y) 56.5±13.2 57.5±13.3 −0.07 56.6±13.2 56.5±13.3 0.01
Age (y)
 ≤40 407 (9.7) 8118 (8.8) 0.03 396 (9.7) 1621 (9.9) −0.01
 41–54 1516 (36.2) 31,334 (34.0) 0.05 1474 (36.0) 5930 (36.2) <0.01
 55–64 1132 (27.0) 25,180 (27.3) −0.01 1107 (27.0) 4372 (26.7) 0.01
 ≥65 1136 (27.1) 27,522 (29.9) −0.06 1119 (27.3) 4461 (27.2) <0.01
Diabetes severity
 Diabetic duration
  Rosacea before DM 1579 (37.7) 10,786 (11.7) 0.63 1484 (36.2) 5899 (36.0) <0.01
  <5 y 1337 (31.9) 50,588 (54.9) −0.48 1337 (32.6) 5305 (32.4) 0.01
  ≥5 y 1275 (30.4) 30,780 (33.4) −0.06 1275 (31.1) 5180 (31.6) −0.01
 Diabetic neuropathy 255 (6.1) 5983 (6.5) −0.02 254 (6.2) 1057 (6.5) −0.01
 Diabetic foot ulcer 58 (1.4) 1982 (2.2) −0.06 58 (1.4) 241 (1.5) <0.01
Retinal examinations in the previous 1 y 0.27±0.66 0.24±0.64 0.04 0.26±0.64 0.25±0.71 0.02
Type of diabetes
 Type 1 331 (7.9) 6223 (6.8) 0.04 323 (7.9) 1303 (8.0) <0.01
 Type 2 3860 (92.1) 85,931 (93.2) −0.04 3773 (92.1) 15,081 (92.0) <0.01
 Dyslipidemia 1488 (35.5) 33,742 (36.6) −0.02 1427 (34.8) 5907 (36.1) −0.03
 Hypertension 2047 (48.8) 47,368 (51.4) −0.05 1986 (48.5) 7944 (48.5) <0.01
 Ischemic heart disease 481 (11.5) 10,959 (11.9) −0.01 466 (11.4) 1887 (11.5) <0.01
 Chronic kidney disease 283 (6.8) 8152 (8.8) −0.08 280 (6.8) 1082 (6.6) 0.01
 Stroke 218 (5.2) 6657 (7.2) −0.08 216 (5.3) 879 (5.4) <0.01
Charlson Comorbidity Index score 1.45±1.30 1.60±1.28 −0.11 1.42±1.26 1.41±1.41 0.01
Antidiabetic medication
 Sulfonylurea 2585 (61.7) 61,752 (67.0) −0.11 2558 (62.5) 10,089 (61.6) 0.02
 Metformin 3011 (71.8) 64,960 (70.5) 0.03 2930 (71.5) 11,777 (71.9) −0.01
 Alpha-glucosidase inhibitors 410 (9.8) 8849 (9.6) 0.01 398 (9.7) 1591 (9.7) <0.01
 Thiazolidinediones 369 (8.8) 8637 (9.4) −0.02 368 (9.0) 1502 (9.2) −0.01
 Meglitinides 233 (5.56) 5194 (5.64) <0.01 227 (5.5) 881 (5.4) 0.01
 DPP4i 303 (7.23) 6615 (7.18) <0.01 296 (7.2) 1199 (7.3) <0.01
 Insulin 343 (8.2) 8830 (9.6) −0.05 339 (8.3) 1390 (8.5) −0.01
Antihypertensive medication
 Calcium channel blockers 1213 (28.9) 28,937 (31.4) −0.05 1193 (29.1) 4677 (28.5) 0.01
 Beta-blockers 1136 (27.1) 23,614 (25.6) 0.03 1102 (26.9) 4404 (26.9) <0.01
 ACEis/ARBs 1564 (37.3) 35,717 (38.8) −0.03 1531 (37.4) 6176 (37.7) −0.01
 Thiazide 186 (4.4) 4102 (4.5) <0.01 178 (4.3) 703 (4.3) <0.01
 Alpha-blockers 148 (3.5) 3275 (3.6) <0.01 142 (3.5) 568 (3.5) <0.01
Other medication
 Antiplatelet 773 (18.4) 19,305 (20.9) −0.06 761 (18.6) 3058 (18.7) <0.01
 Statins 1239 (29.6) 25,526 (27.7) 0.04 1194 (29.2) 4921 (30.0) −0.02
 Fibrates 431 (10.3) 9796 (10.6) −0.01 423 (10.3) 1662 (10.1) 0.01
Follow-up duration (y) 5.2±3.6 5.0±3.5 0.07 5.2±3.6 5.1±3.6 0.02
Data are presented as frequency (%) or mean±SD.
ACEi indicates angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; DDP4i, dipeptidyl peptidase-4 inhibitor; DM, diabetes mellitus; PSM, propensity score matching; STD, standardized difference.

The risk of time-to-event outcome between groups was compared using the Fine and Gray subdistribution hazard model, which considered death during follow-up a competing risk.33 The study group (rosacea vs nonrosacea) was the only explanatory variable in the survival analysis. A 2-sided P value <0.05 was considered statistically significant. All statistical analyses were performed using SAS, version 9.4 (SAS Institute), including the procedure of “psmatch” for PSM and “phreg” for survival analyses and a macro of “%cif” for generating cumulative incidence function under the Fine and Gray subdistribution hazard method for drawing plots.


Inclusion and Exclusion of the Study Patients

From January 1, 1997, to December 31, 2013, 2,099,303 patients were diagnosed as having DM under treatment. Among them, 5459 were diagnosed as having rosacea in the dermatology department. Next, 20 times the number of patients with DM without rosacea were identified. After application of the exclusion criteria and 1:4 PSM, 4096 and 16,384 patients (total: 20,480) were in the rosacea and nonrosacea groups, respectively (Fig. 1).

Patient Characteristics

Before PSM, no substantial differences were noted in most covariates, except the duration of diabetes, Charlson Comorbidity Index score, and use of sulfonylurea. After PSM, no variable exhibited a substantial between-group difference (all absolute standardized difference values <0.1) (Table 1). The overall study population comprised ∼39% of men (mean age: 56 y old), consisting of 92.1% of type 2 and 7.9% of type 1 DM. The most common comorbidity was hypertension (48.5%), followed by dyslipidemia (35.8%), ischemic heart disease (11.5%), and chronic kidney disease (6.7%). The most frequently used hypoglycemic agent was metformin (71.8%), followed by sulfonylurea (61.8%), and 8.4% of the patients received insulin therapy. In the rosacea group, 36.2% of the patients had rosacea before the DM diagnosis, whereas 32.6% and 31.1% of the patients had a DM duration of <5 years and ≥5 years at the time of the rosacea diagnosis.

Rosacea and Ocular Outcomes

The results of the primary ocular outcomes are presented in Table 2. During the mean follow-up period of 5 years, patients with rosacea had significantly higher risks in DME [subdistribution hazard ratio (SHR): 1.31, 95% confidence interval (CI): 1.05–1.63], glaucoma with medical treatment (SHR: 1.11, 95% CI: 1.01–1.21), and dry eye disease (SHR: 1.55, 95% CI: 1.38–1.75) than those without rosacea (Fig. 2A–C). Among the surgical interventions, the risk of cataract surgery (SHR: 1.13, 95% CI: 1.02–1.25) was significantly higher in the rosacea group than in the nonrosacea group (Fig. 2D). No significant association was found in DR, age-related macular degeneration, retinal vascular occlusion, optic neuritis, ischemic optic neuropathy, retinal detachment, uveitis, or other ocular interventions (retinal laser, intravitreal injection, vitrectomy, and glaucoma interventions).

TABLE 2 - Primary Ocular Outcomes in Patients With Diabetes With and Without Rosacea
Before PSM After PSM
Outcome Rosacea (n=4191) Nonrosacea (n=92,154) Rosacea (n=4096) Nonrosacea (n=16,384) SHR (95% CI) P
Ocular diseases
 Diabetic retinopathy 417 (9.95) 9104 (9.88) 412 (10.1) 1639 (10.0) 1.02 (0.92–1.13) 0.733
 Diabetic macular edema 105 (2.5) 1820 (2.0) 103 (2.5) 319 (2.0) 1.31 (1.05–1.63) 0.017
 Dry eye disease 394 (9.4) 5448 (5.9) 382 (9.3) 1022 (6.2) 1.55 (1.38–1.75) <0.001
 Glaucoma with medical treatment 614 (14.7) 11,978 (13.0) 592 (14.5) 2187 (13.3) 1.11 (1.01–1.21) 0.027
 Age-related macular degeneration 70 (1.7) 1297 (1.4) 70 (1.7) 245 (1.5) 1.15 (0.89–1.50) 0.290
 Retinal vascular occlusion 19 (0.45) 382 (0.41) 19 (0.46) 78 (0.48) 0.98 (0.59–1.62) 0.931
 Ischemic optic neuropathy 3 (0.07) 45 (0.05) 3 (0.07) 8 (0.05) 1.51 (0.40–5.69) 0.542
 Optic neuritis 2 (0.05) 60 (0.07) 2 (0.05) 13 (0.08) 0.62 (0.14–2.74) 0.527
 Uveitis 52 (1.24) 933 (1.01) 52 (1.27) 170 (1.04) 1.003 (0.75–1.35) 0.983
 Retinal detachment 11 (0.26) 254 (0.28) 9 (0.22) 52 (0.32) 0.70 (0.34–1.41) 0.314
Surgical intervention
 Retinal laser 177 (4.2) 4353 (4.7) 172 (4.2) 803 (4.9) 0.86 (0.73–1.01) 0.066
 Intravitreal injection 30 (0.72) 809 (0.88) 29 (0.71) 141 (0.86) 0.83 (0.56–1.23) 0.356
 Vitrectomy 41 (0.98) 986 (1.07) 39 (0.95) 189 (1.15) 0.83 (0.59–1.18) 0.297
 Cataract surgery 478 (11.4) 9462 (10.3) 467 (11.4) 1694 (10.3) 1.13 (1.02–1.25) 0.024
 Glaucoma with surgical interventions 46 (1.1) 905 (1.0) 44 (1.1) 190 (1.2) 0.93 (0.67–1.29) 0.671
PSM indicates propensity score matching; SHR, subdistribution hazard ratio.
Data are presented as frequency (%).

Time to cumulative incidence in diabetic macular edema (A), glaucoma (B), dry eye disease (C), and cataract surgery (D) in the nonrosacea and rosacea groups.

Rosacea and Systemic Inflammatory/Gastrointestinal/Psychiatric Diseases

The results of the secondary outcomes are presented in Table 3. Among the systemic inflammatory diseases, patients with rosacea had a significantly greater risk of psoriasis than those without rosacea (SHR: 2.99, 95% CI: 2.19–4.09). By contrast, no significant between-group difference was observed in other outcomes. Among the 3 gastrointestinal diseases, a higher risk of irritable bowel syndrome was noted in the rosacea group (SHR: 1.31, 95% CI: 1.13–1.53). Among the 4 psychiatric diseases, rosacea was associated with a greater risk of anxiety (SHR: 1.37, 95% CI: 1.15–1.64) and depression (SHR: 1.34, 95% CI: 1.05–1.70). By contrast, the risks of bipolar disorder and schizophrenia were not different between the 2 groups.

TABLE 3 - Other Secondary Outcomes
Before PSM After PSM
Outcome Rosacea (n=4191) Nonrosacea (n=92,154) Rosacea (n=4096) Nonrosacea (n=16,384) SHR (95% CI) P
Systemic inflammatory disease
 Psoriasis 68 (1.62) 563 (0.61) 67 (1.64) 91 (0.56) 2.99 (2.19–4.09) <0.001
 Rheumatoid arthritis 63 (1.5) 1155 (1.3) 54 (1.32) 211 (1.29) 1.03 (0.76–1.39) 0.854
 Inflammatory bowel disease 36 (0.86) 807 (0.88) 32 (0.78) 150 (0.92) 0.86 (0.59–1.25) 0.424
 Ankylosing spondylitis 20 (0.48) 267 (0.29) 19 (0.46) 47 (0.29) 1.63 (0.95–2.77) 0.075
 Behçet syndrome 1 (0.02) 28 (0.03) 1 (0.02) 6 (0.04) 0.67 (0.08–5.56) 0.711
 Sarcoidosis 1 (0.02) 13 (0.01) 1 (0.02) 4 (0.02) 1.002 (0.11–8.97) 0.998
Gastrointestinal diseases
 Irritable bowel syndrome 223 (5.3) 3280 (3.6) 215 (5.2) 664 (4.1) 1.31 (1.13–1.53) 0.001
 Esophageal reflux 134 (3.2) 2263 (2.5) 129 (3.1) 445 (2.7) 1.18 (0.96–1.43) 0.111
Helicobacter pylori 48 (1.15) 815 (0.88) 45 (1.10) 146 (0.89) 1.25 (0.89–1.73) 0.195
Psychiatric diseases
 Anxiety 175 (4.2) 2443 (2.7) 165 (4.0) 487 (3.0) 1.37 (1.15–1.64) 0.001
 Depression 96 (2.3) 1367 (1.5) 88 (2.1) 266 (1.6) 1.34 (1.05–1.70) 0.018
 Bipolar 18 (0.43) 299 (0.32) 18 (0.44) 50 (0.31) 1.45 (0.84–2.48) 0.180
 Schizophrenia 5 (0.12) 130 (0.14) 5 (0.12) 18 (0.11) 1.12 (0.42–3.01) 0.824
PSM indicates propensity score matching; SHR, subdistribution hazard ratio.
Data are presented as frequency (%).


The pathogenic mechanisms of both rosacea and diabetes include systemic inflammatory components. Although the link between rosacea and diabetes remains poorly understood, increased oxidative stress and systemic inflammation have been implicated in the occurrence of rosacea, insulin resistance, and diabetes complications. However, the association between rosacea and eye diseases among diabetic patients has not previously been investigated. We performed a population-based study to investigate the association between rosacea and eye diseases among diabetic patients.

Rosacea may more closely resemble a systemic chronic inflammatory disease34 than a localized facial disease, associated with many systemic complications. The pathogenesis of rosacea is complicated. Triggering factors activate toll-like receptor 2, increase the active form of cathelicidin peptide LL-37, and induce leukocyte chemotaxis and inflammatory events, which activate nuclear factorκB and promote angiogenesis.35 Vascular and inflammatory changes present as clinical rosacea symptoms, such as erythema, papules, and pustules. In addition, oxidative stress has recently been recognized as a key mechanism underlying insulin resistance and diabetes complications.36,37 Angiogenesis, oxidative stress, and inflammation play critical roles in the pathogenesis of both rosacea and diabetic eye diseases.19–21

In this study, we observed that rosacea was positively associated with DME but not with DR. DR and DME are the leading causes of vision-threatening diseases among patients with diabetes,37,38 and the pathophysiology of both diseases involves complex microvascular changes due to inflammation and angiogenesis21,39 that cannot be explained by any single pathway. Disease development is widely accepted as resulting from the combined effects of angiogenesis and inflammation.20 Systemic inflammation may not be correlated with local retinal inflammation in DR,40,41 despite the elevation of systemic inflammatory biomarkers in DR,41 similar to DME42–44 and other disorders associated with macular edema, such as uveitis,44 retinal vascular occlusion,45 and retinal dystrophy.46 We hypothesize that systemic inflammation may play a more critical role in DME than DR. However, determining detailed associations between systemic inflammation and the biomolecular pathophysiology of DME requires further study.

Rosacea may increase the risk of blepharitis, leading to an imbalance in tear formation, which is associated with dry eye disease.47,48 However, rosacea is not associated with an increased risk of corneal damage in dry eye disease.49 Diabetes is a known risk factor for glaucoma and cataract development.50,51 In our study, we identified rosacea as an additional risk factor for glaucoma requiring medication and cataracts requiring surgery among patients with diabetes. An increase in systemic inflammation may contribute to the development of cataracts52 and glaucoma.53 Glaucoma medication has been reported to be associated with blepharitis and dry eye disease.54 However, glaucoma and cataract development are also associated with steroid use,55 which is often used to treat dry eye disease.56 Short-term systemic steroids are also prescribed to treat rosacea and may require long-term administration due to the relapsing course of rosacea. Further investigations examining patients with matched steroid prescriptions remain necessary to clarify the possible confounding effects of steroid use on symptom development.

We noted that rosacea increased the risk of psoriasis in patients diagnosed as having diabetes. No study has evaluated the significant association between rosacea and psoriasis. As these 2 diseases were both mainly presented with skin rashes, they may have similar skin presentations on the face, possibly misleading the diagnosis. However, rosacea and psoriasis were both associated with chronic inflammatory conditions. For example, studies indicate that both tumor necrosis factor-alpha and interleukin-17 play pivotal roles and treatment targets in rosacea and psoriasis, indicating the common property of systemic inflammation in these 2 diseases.57,58 Rosacea and psoriasis also share some common predisposing factors or comorbidities, such as microbiotic environment dysfunction, metabolic cardiovascular, and psychiatric complications.59

Rosacea is associated with various gastrointestinal diseases.6 Although a previous meta-analysis indicated an increased risk of inflammatory bowel disease in patients with rosacea,3 our study with a diabetic population did not find this association. However, we demonstrated an increased incidence of irritable bowel syndrome in patients with diabetes diagnosed as having rosacea. This agrees with a previous Danish nationwide population-based cohort study.60 Rosacea and irritable bowel syndrome have some common factors,6 such as small bowel bacterial overgrowth and systemic inflammation, which is supported by elevated peripheral tumor necrosis factors. The exact pathogenic link, however, remains unclear.

In psychiatric diseases, we observed that patients with rosacea were at a significantly higher risk for anxiety and depression but not for schizophrenia or bipolar disorder. This is consistent with a previous Danish nationwide cohort study.61 The association between inflammation and anxiety or depressive disorders has been well studied. For example, a UK study demonstrated a higher risk for anxiety and depression among patients with an inflammatory disorder.62 Increased inflammatory markers are also found in patients with anxiety and depression.63 The effect of inflammation on critical brain structures that are involved in anxiety or depression, such as the amygdala or reward circuit, has been reported in recent imaging studies.64 Moreover, patients with rosacea may also suffer from psychosocial distress, which may explain the increased risk of anxiety or depression.65 Patients with anxiety or depression may have different medical-seeking behavior from the general population, thus leading to more diagnosis of rosacea. However, the association between anxiety or depression and rosacea requires further investigation.

This study has some limitations. First, the NHIRD did not include laboratory examinations. We could not analyze serum glucose and glycated hemoglobin levels, although we matched the patients based on their medication use for diabetes. Furthermore, the database did not have examination reports such as optical coherence tomography, which has been used for identifying central involvement and different inflammation-associated characteristics in DME.66 Moreover, the dataset lacks information on anterior chamber angle morphology and intraocular pressure, which is a main evaluation of steroid-responder and glaucoma.67 The retrospective design precluded controlling for the use of steroids, which may affect the development of cataract and glaucoma. The dataset did not include sufficient information to permit the classification of rosacea subtypes, preventing analyses examining differences among rosacea subtypes. Some of the outcomes had small case numbers (eg, retinal detachment), making the identification of significant differences challenging. In addition, we did not adjust for multiple testing in this study, and the conclusions derived from the study should be interpreted cautiously, as multiple comparison adjustments can affect significance. Finally, the retrospective database study could only demonstrate possible association of the diseases, although the big data in real-world application could provide important information in disease association and help with further investigation.68 The mechanism, cause, and effect of the outcomes require prospective and experimental studies. To the best of our knowledge, this is the first study to analyze the association between rosacea and ocular diseases. Rosacea was positively associated with ocular disease development, including DME, glaucoma, and dry eye disease. Patients with rosacea also had an increased risk of receiving cataract surgery. However, no significant associations were found between rosacea and ocular vascular diseases, including DR, age-related macular degeneration, retinal vascular occlusion, ischemic optic neuropathy, or ocular inflammatory diseases (such as uveitis and optic neuritis). The frequencies of retinal laser, intravitreal injection, and glaucoma interventions were comparable between groups with and without rosacea.


The authors thank Alfred Hsing-Fen Lin, Ben Yu-Lin Chou, and Jane Yan-Jen Shiu for their assistance with the statistical analysis during the completion of the manuscript.


1. Wilkin J, Dahl M, Detmar M, et al. Standard classification of rosacea: report of the National Rosacea Society Expert Committee on the Classification and Staging of Rosacea. J Am Acad Dermatol. 2002;46:584–587.
2. Holmes AD, Spoendlin J, Chien AL, et al. Evidence-based update on rosacea comorbidities and their common physiologic pathways. J Am Acad Dermatol. 2018;78:156–166.
3. Wang FY, Chi CC. Association of rosacea with inflammatory bowel disease: a MOOSE-compliant meta-analysis. Medicine. 2019;98:e16448.
4. Egeberg A, Hansen PR, Gislason GH, et al. Clustering of autoimmune diseases in patients with rosacea. J Am Acad Dermatol. 2016;74:667–672.
5. Haber R, El Gemayel M. Comorbidities in rosacea: a systematic review and update. J Am Acad Dermatol. 2018;78:786–792.
6. Wang FY, Chi CC. Rosacea, germs, and bowels: a review on gastrointestinal comorbidities and gut-skin axis of rosacea. Adv Ther. 2021;38:1415–1424.
7. De Pessemier B, Grine L, Debaere M, et al. Gut-skin axis: current knowledge of the interrelationship between microbial dysbiosis and skin conditions. Microorganisms. 2021;9:353.
8. Vieira AC, Mannis MJ. Ocular rosacea: common and commonly missed. J Am Acad Dermatol. 2013;69:S36–S41.
9. Stone DU, Chodosh J. Ocular rosacea: an update on pathogenesis and therapy. Curr Opin Ophthalmol. 2004;15:499–502.
10. International Diabetes Federation. IDF Diabetes Atlas, 8th ed. 2017. http://www.diabetesatlas.org/. Accessed January 13, 2022.
11. Beckman JA, Creager MA. Vascular complications of diabetes. Circ Res. 2016;118:1771–1785.
12. American Diabetes Association. Economic costs of diabetes in the US in 2012. Diabetes Care. 2013;36:1033–1046.
13. Sayin N, Kara N, Pekel G. Ocular complications of diabetes mellitus. World J Diabetes. 2015;6:92–108.
14. Chhablani J, Wong K, Tan GS, et al. Diabetic macular edema management in Asian population: expert panel consensus guidelines. Asia Pac J Ophthalmol (Phila). 2020;9:426–434.
15. Wellen KE, Hotamisligil GS. Inflammation, stress, and diabetes. J Clin Invest. 2005;115:1111–1119.
16. Lontchi-Yimagou E, Sobngwi E, Matsha TE, et al. Diabetes mellitus and inflammation. Curr Diab Rep. 2013;13:435–444.
17. Biondi-Zoccai GG, Abbate A, Liuzzo G, et al. Atherothrombosis, inflammation, and diabetes. J Am Coll Cardiol. 2003;41:1071–1077.
18. Collier B, Dossett LA, May AK, et al. Glucose control and the inflammatory response. Nutr Clin Pract. 2008;23:3–15.
19. Rübsam A, Parikh S, Fort PE. Role of inflammation in diabetic retinopathy. Int J Mol Sci. 2018;19:942.
20. Capitão M, Soares R. Angiogenesis and inflammation crosstalk in diabetic retinopathy. J Cell Biochem. 2016;117:2443–2453.
21. Romero-Aroca P, Baget-Bernaldiz M, Pareja-Rios A, et al. Diabetic macular edema pathophysiology: vasogenic versus inflammatory. J Diabetes Res. 2016;2016:2156273.
22. Abcouwer SF, Antonetti DA. A role for systemic inflammation in diabetic retinopathy. Invest Ophthalmol Vis Sci. 2013;54:2384.
23. Mesquida M, Drawnel F, Fauser S. The role of inflammation in diabetic eye disease. Semin Immunopathol. 2019;41:427–445.
24. Hsieh CY, Su CC, Shao SC, et al. Taiwan’s National Health Insurance Research Database: past and future. Clin Epidemiol. 2019;11:349–358.
25. Hsing AW, Ioannidis JP. Nationwide population science: lessons from the Taiwan National Health Insurance Research Database. JAMA Intern Med. 2015;175:1527–1529.
26. Lin LY, Warren-Gash C, Smeeth L, et al. Data resource profile: the National Health Insurance Research Database (NHIRD). Epidemiol Health. 2018;40:e2018062.
27. Wu CY, Chang YT, Juan CK, et al. Risk of inflammatory bowel disease in patients with rosacea: results from a nationwide cohort study in Taiwan. J Am Acad Dermatol. 2017;76:911–917.
28. Hua TC, Chung PI, Chen YJ, et al. Cardiovascular comorbidities in patients with rosacea: a nationwide case-control study from Taiwan. J Am Acad Dermatol. 2015;73:249–254.
29. Zhou Z, Rahme E, Abrahamowicz M, et al. Survival bias associated with time-to-treatment initiation in drug effectiveness evaluation: a comparison of methods. Am J Epidemiol. 2005;162:1016–1023.
30. D’Hoore W, Bouckaert A, Tilquin C. Practical considerations on the use of the Charlson Comorbidity Index with administrative data bases. J Clin Epidemiol. 1996;49:1429–1433.
31. Kang EY, Kang C, Wu WC, et al. Association between add-on dipeptidyl peptidase-4 inhibitor therapy and diabetic retinopathy progression. J Clin Med. 2021;10:2871.
32. Liu CH, Kang EY, Lin YH, et al. Association of ocular diseases with schizophrenia, bipolar disorder, and major depressive disorder: a retrospective case-control, population-based study. BMC Psychiatry. 2020;20:486.
33. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94:496–509.
34. Wollina U. Is rosacea a systemic disease? Clin Dermatol. 2019;37:629–635.
35. Holmes AD, Steinhoff M. Integrative concepts of rosacea pathophysiology, clinical presentation and new therapeutics. Exp Dermatol. 2017;26:659–667.
36. Hurrle S, Hsu WH. The etiology of oxidative stress in insulin resistance. Biomed J. 2017;40:257–262.
37. Giacco F, Brownlee M. Oxidative stress and diabetic complications. Circ Res. 2010;107:1058–1070.
38. Cheung N, Mitchell P, Wong TY. Diabetic retinopathy. Lancet. 2010;376:124–136.
39. Bandello F, Battaglia Parodi M, Lanzetta P, et al. Diabetic macular edema. Dev Ophthalmol. 2017;58:102–138.
40. Lechner J, O’Leary OE, Stitt AW. The pathology associated with diabetic retinopathy. Vision Res. 2017;139:7–14.
41. Tang J, Kern TS. Inflammation in diabetic retinopathy. Prog Retin Eye Res. 2011;30:343–358.
42. Vujosevic S, Simó R. Local and systemic inflammatory biomarkers of diabetic retinopathy: an integrative approach. Invest Ophthalmol Vis Sci. 2017;58:BIO68–BIO75.
43. Figueras-Roca M, Molins B, Sala-Puigdollers A, et al. Peripheral blood metabolic and inflammatory factors as biomarkers to ocular findings in diabetic macular edema. PLoS ONE. 2017;12:e0173865.
44. Ilhan C, Citirik M, Uzel MM, et al. The usefulness of systemic inflammatory markers as diagnostic indicators of the pathogenesis of diabetic macular edema. Arq Bras Oftalmol. 2020;83:299–304.
45. Figueras-Roca M, Matas J, Llorens V, et al. Systemic contribution of inflammatory mediators to the severity of diabetic and uveitic macular edema. Graefes Arch Clin Exp Ophthalmol. 2021;259:2695–2705.
46. Deobhakta A, Chang LK. Inflammation in retinal vein occlusion. Int J Inflam. 2013;2013:438412.
47. Verhagen F, Kuiper J, Nierkens S, et al. Systemic inflammatory immune signatures in a patient with CRB1 linked retinal dystrophy. Expert Rev Clin Immunol. 2016;12:1359–1362.
48. Tong L, Lim L, Tan D, et al. Corneal Subspecialty Workgroup, College of Ophthalmology, Academy of Medicine, Singapore. Assessment and management of dry eye disease and meibomian gland dysfunction: providing a Singapore framework. Asia Pac J Ophthalmol (Phila). 2021;10:530–541.
49. Vora GK, Gupta PK. Intense pulsed light therapy for the treatment of evaporative dry eye disease. Curr Opin Ophthalmol. 2015;26:314–318.
50. Hung N, Kang EY, Lee TW, et al. The risks of corneal surface damage in aqueous-deficient dry eye disease: a 17-year population-based study in Taiwan. Am J Ophthalmol. 2021;227:231–239.
51. Mitchell P, Smith W, Chey T, et al. Open-angle glaucoma and diabetes: the Blue Mountains eye study, Australia. Ophthalmology. 1997;104:712–718.
52. Drinkwater JJ, Davis WA, Davis TME. A systematic review of risk factors for cataract in type 2 diabetes. Diabetes Metab Res Rev. 2019;35:e3073.
53. Ang MJ, Afshari NA. Cataract and systemic disease: a review. Clin Exp Ophthalmol. 2021;49:118–127.
54. Nijm LM, De Benito-Llopis L, Rossi GC, et al. Understanding the dual dilemma of dry eye and glaucoma: an international review. Asia Pac J Ophthalmol (Phila). 2020;9:481–490.
55. Karahan M, Kilic D, Guven S. Systemic inflammation in both open-angle and angle-closure glaucoma: role of platelet-to-lymphocyte ratio. Bratisl Lek Listy. 2021;122:45–48.
56. Renfro L, Snow JS. Ocular effects of topical and systemic steroids. Dermatol Clin. 1992;10:505–512.
57. Amir Ali A, Vender R, Vender R. The role of IL-17 in papulopustular rosacea and future directions. J Cutan Med Surg. 2019;23:635–641.
58. Sand FL, Thomsen SF. Off-label use of TNF-alpha inhibitors in a dermatological university department: retrospective evaluation of 118 patients. Dermatol Ther. 2015;28:158–165.
59. Duman N, Ersoy Evans S, Atakan N. Rosacea and cardiovascular risk factors: a case control study. J Eur Acad Dermatol Venereol. 2014;28:1165–1169.
60. Egeberg A, Weinstock LB, Thyssen EP, et al. Rosacea and gastrointestinal disorders: a population-based cohort study. Br J Dermatol. 2017;176:100–106.
61. Egeberg A, Hansen PR, Gislason GH, et al. Patients with rosacea have increased risk of depression and anxiety disorders: a Danish nationwide cohort study. Dermatology. 2016;232:208–213.
62. Dregan A, Matcham F, Harber-Aschan L, et al. Common mental disorders within chronic inflammatory disorders: a primary care database prospective investigation. Ann Rheum Dis. 2019;78:688–695.
63. Lee CH, Giuliani F. The role of inflammation in depression and fatigue. Front Immunol. 2019;10:1696.
64. Felger JC. Imaging the role of inflammation in mood and anxiety-related disorders. Curr Neuropharmacol. 2018;16:533–558.
65. Heisig M, Reich A. Psychosocial aspects of rosacea with a focus on anxiety and depression. Clin Cosmet Investig Dermatol. 2018;11:103–107.
66. Chung YR, Kim YH, Ha SJ, et al. Role of inflammation in classification of diabetic macular edema by optical coherence tomography. J Diabetes Res. 2019;2019:8164250.
67. Phulke S, Kaushik S, Kaur S, et al. Steroid-induced glaucoma: an avoidable irreversible blindness. J Curr Glaucoma Pract. 2017;11:67–72.
68. Cheng CY, Soh ZD, Majithia S, et al. Big data in ophthalmology. Asia Pac J Ophthalmol (Phila). 2020;9:291–298.

cataract; diabetes; diabetic macular edema; dry eye disease; glaucoma; mood disorder; psoriasis; rosacea

Supplemental Digital Content

Copyright © 2022 Asia-Pacific Academy of Ophthalmology. Published by Wolters Kluwer Health, Inc. on behalf of the Asia-Pacific Academy of Ophthalmology.