Topical corticosteroids preparations are commonly used in the treatment of ocular inflammation. For anterior segment inflammation, topical steroids treatment is usually preferred over systemic steroids owing to its safety, convenience, and low risk of systemic side effects. Although serum drug levels are apparently low in patients receiving topical corticosteroids [1], there may be a change in levels of plasma cortisol and adrenocorticotropic hormone. [2-6] Fukushima et al. [7,8] showed in two different publications an early effect on serum glucose levels, within hours, after subconjunctival injection of dexamethasone in diabetic and non diabetic rats and an early transient increase in blood glucose levels on the day of surgery in diabetic patients. Our previous study [9] evaluated the effect of dexamethasone eye drops on blood glucose profile following cataract operations and demonstrated that postoperative dexamethasone eye drops have a greater effect on the blood glucose profile of diabetic compared to non diabetic patients.

The aim of the present study is to examine changes in serum glucose concentrations and in haemoglobin A1c (HbA1C) in diabetic patients treated with repeated topical applications of steroidal eye drops.

All community pharmacies in use by the Clalit Health Services health maintenance organization (HMO) in Israel are computerized and report to a central repository. All prescriptions for ocular topical steroids filled by diabetic members (ICD-9 code: 250.0) of the "central district" of the HMO between January 1^st, 2001 and July 31^st, 2006, were included (32,454 prescriptions filled by 8995 patients). These included Dexamethasone phosphate 0.1%, Fluorometholone 0.1%, Hydrocortisone acetate 1.5%, Prednisolone acetate 1%.

This HMO dispenses steroidal eye drops with nominal and almost equal co-payment, which ensures that all prescriptions were documented and that drug selection was not influenced by financial aspects. Since there is co-payment we assume that most of the medications that were filled were indeed used by the patients. The collection of data for the study was done anonymously and in conformity with all country laws and the declaration of Helsinki. The study was approved by the Institutional Review Board (Barzilai Medical Center).

To exclude patients receiving steroidal eyedrops treatment for a short period of time, we included only those patients that filled at least 3 consecutive prescriptions at least once every two months. Other exclusion criteria included patients receiving systemic corticosteroids or other drugs that could affect blood glucose (tricyclic antidepressants, diuretics, epinephrine, estrogens, lithium, phenytoin, and salicylates) at the time of the study, a systemic disease known to influence blood glucose levels (i.e. acromegaly, chronic renal failure, cushing syndrome, hyperthyroidism, pancreatic cancer, pancreatitis). Cessation of treatment was defined as not receiving the medications for at least 3 months or more. Based on these criteria, we included 2697 diabetic patients using altogether 18,692 ocular topical steroidal prescriptions.

To evaluate glycemic control, all fasting blood glucose tests and HbA1c laboratory tests performed by diabetics during the study period (2 weeks before till 8 weeks after the beginning of steroid drops treatment) in the district, were documented (225,304 tests).

Of the 8995 diabetic patients followed between January 1^st, 2001 and July 31^st, 2006, 2697 (30%) were treated with ocular topical steroidal medications (filled at least three prescriptions). Of those, 1360 (50.4%) had laboratory data for their glycemic control (fasting blood glucose levels and HbA1c) measured around the period of topical steroidal treatment (2 weeks before till 8 weeks after the beginning of steroid drops treatment). These 1360 patients were included in this study and considered our study group, 589 males and 771 females with mean age of 67.8±9.1 years. After starting topical steroids, only laboratory data acquired while still being on topical steroids were included.

We investigated the effect of ocular topical steroidal medications on glycemic control among diabetic patients.

Statistical analysis

Student's t-test was used for continuous variables and x² test for proportions (SPSS ver.12 (SPSS Inc. Chicago, IL, USA)). Multiple unpaired t-tests were performed, testing each time point against the baseline time point for both random blood glucose measurement and HbA1c measurement. To account for multiple testing, the p values were corrected according to Bonferroni correction. P-value less than 5% was considered statistically significant.

Before starting topical steroids treatment, the baseline fasting glucose level was 145.8 ± 2.1 (standard error of the mean=SEM) mg/dl and HbA1c 7.6 ± 0.1 %.

Figure 1 shows the change in fasting blood glucose levels and HbA1c levels over time while being on topical steroidal treatment. We noted an increase in blood glucose levels, peaking to a mean value of 157.4 ± 6.3 mg/dl on the 3^rd week on topical steroids (p= 0.05, t-test) and an increase in HbA1c to 8.2 ± 0.3% on the 7^th week under topical steroids treatment (p= 0.03, t-test).

Figure 1: Changes in fasting blood glucose levels and HbA1c levels while being on topical steroidal treatment. One can note an increase in blood glucose levels, peaking to a mean value of 157.4±6.3 mg/dl on the 3^rd week on topical steroids (p= 0.05, t-test) and an increase in HbA1c to 8.2 ±0.3% on the 7^th week under topical steroids treatment (p= 0.03, t-test). Asterisks signify p<0.05.

Figure 2 compares the distribution of blood glucose levels in our patients 0-2 weeks before the topical steroids treatment was initiated versus 2-3 weeks after the beginning of treatment. As can be seen, blood glucose levels of >160 mg/dl was documented in 31% of patients before starting the treatment versus 42% on topical steroidal treatment. (p=0.0007).

Figure 2: Distribution of blood glucose levels in the study diabetic patients, 0-2 weeks before topical steroidal treatment versus 2-3 weeks after the beginning of treatment. As can be seen, blood glucose levels of >160 mg/dl was documented in 31% of patients before starting the treatment versus 42% on topical steroidal treatment. (p=0.0007).

We have shown that topical corticosteroid eye drops may induce a significant increase in serum glucose concentration in diabetic patients, and may interfere with their glycemic control.

Maintenance of near-normal glycaemia through intensive diabetes treatment have been proven to delay or prevent microvascular complications. [10,11] The diabetes associations throughout the world have set blood glucose and haemoglobin (Hb) A1c targets to assist care providers and patients with goal setting for diabetes management and control [12,13].

Our study demonstrated the systemic effect of topically administered corticosteroid drops. These drops may be systemically absorbed via several routes, namely the ocular tissues, the nasolacrimal tract or the gastrointestinal tract, when swallowed. [14] Punctal occlusion for a few minutes after the application of the drops might decrease the systemic absorption of the drops. Fukushima et al. [7] demonstrated an increase in blood glucose profile in response to subconjunctival dexamethasone injection in both diabetic and non diabetic groups, in a rat model. Their subsequent study [8] on diabetic patients demonstrated that subconjunctival steroid injection, in the setting of cataract surgery, induced a transient but significant increase in blood glucose on the day of surgery. Although the topically administered steroids may be systemically absorbed, the serum concentration achieved is much less effective compared with sub conjunctial injection.

Feldman-Billard et al. [15] showed recently that periocular steroidal injections in patients with type 2 diabetes induced a marked hyperglycemic effect. Our recent study [9] showed a significant increase in fasting blood glucose levels from 170 ± 55.5 mg/dl to 229 ± 76.8 mg/dl in diabetic patients treated with steroidal eye drops following cataract surgery. (p=0.05) This change was not demonstrated in our two control groups of diabetic patients treated postoperatively with diclofenac drops and non diabetic patients treated with steroidal drops. This finding was the rationale for our present study. Another study, published by Kymionis et al. in 2007 [16], found that intensive application of topical corticosteroids drops for a short

Period of time (7 days) raised the blood glucose levels in patients with controlled diabetes mellitus (p=0.003), which returned to pretreatment levels after discontinuation of the eye drops.

In accordance to these previous studies, our present study found a significant rise in plasma glucose levels and HbA1c, peaking at 3 weeks and 7 weeks, respectively, following topical steroids treatment.

Other causes for hyper glycemia in our study subjects were excluded, namely acromegaly, chronic renal failure, cushing syndrome, hyperthyroidism, pancreatic cancer, pancreatitis, and use of medications known to affect plasma glucose.

HbA1c is produced by glycation of hemoglobin A. Its level reflects the mean blood glucose level over a period of one or two months before the blood sample is taken, and indicates the status of glycemic control. The significant change in HbA1c, demonstrated here, probably reflects a worse glycemic control in the diabetic patient under topical steroid treatment.

Our present study has several limitations. First, its retrospective nature and non randomized design limits the power of our findings. Second, one must view with caution the assumption that filling a prescription for a medication equates with use of the medication as prescribed. Third, it is possible that the condition for which the corticosteroid is required (e.g. uveitis) could itself elevate blood sugar via physical or psychological stress.

Thus, we recommend that a prospective randomized clinical trial be undertaken, on a large series of patients, to further investigate the findings of this study.

In conclusion, topical steroids can affect blood glucose levels. Therefore, we recommend that blood glucose levels be monitored in diabetic patients receiving long-term treatment with steroid eye drops.