Prediction of risk and incidence of dry eye in critical patients1

Objectives: to estimate the incidence of dry eye, to identify risk factors and to establish a risk prediction model for its development in adult patients admitted to the intensive care unit of a public hospital. Method: concurrent cohort, conducted between March and June, 2014, with 230 patients admitted to an intensive care unit. Data were analyzed by bivariate descriptive statistics, with multivariate survival analysis and Cox regression. Results: 53% out of 230 patients have developed dry eye, with onset mean time of 3.5 days. Independent variables that significantly and concurrently impacted the time for dry eye to occur were: O2 in room air, blinking more than five times per minute (lower risk factors) and presence of vascular disease (higher risk factor). Conclusion: dry eye is a common finding in patients admitted to adults intensive care units, and care for its prevention should be established.


Methods
This is a concurrent cohort study, conducted in an ICU for adult patients of a public teaching hospital in Belo Horizonte, Minas Gerais. Currently, in this hospital, 30 intensive care beds intended for adults are available to the community.
Sample size calculation was carried out using the infinite population formula, by conservative criterion, since the studied population was unknown. In the estimation of the sample we considered the infinite population, confidence degree of 95%, margin of error of 6.5%, and proportion of interest of 55.1% in the incidence of lesions in the punctate cornea (8)   In Table 2 we observed that the values of the Pearson correlation coefficient (p) are all close to zero.
In addition, we also observed that both the global test and the tests for each variable do not show evidence for rejecting the null hypothesis of proportional risks.

Discussion
In previous studies (11)(12)(13)(14)(15) , conducted with patients in outpatient units, dry eye prevalence ranged from 10.8% to 57.1%. However, these studies were conducted with a profile of patients different from the one of the patients who participated in this study. The only study (4) on dry eye in critical patients found in the literature is a randomized clinical trial, comparing interventions for the prevention of the problem. We highlight that such study was conducted with a small sample (18 patients), and in a reality different from Brazil's.
As the reduction in the amount or the quality of tears can be the beginning of major alterations in the ocular surface, the prevention of dry eye becomes essential for critically ill patients. This problem can be identified and prevented by the nurses, who have nursing interventions to assist in the reduction of ocular complications.
Studies (16)(17) suggest that alterations in the ocular surface are highly prevalent, especially in the early days of hospitalization. It is estimated that the mean for the development of corneal injury is between 24 hours and 8.9 days (1,(6)(7)(8)18) . In the study (4) that specifically addressed dry eye, the mean for the emergence of the problem was approximately three days, but this time was estimated after the intervention implementation.  We also highlight that, in this study, lagophthalmos (eyeball exposure) was identified as the main risk factor for changes in the ocular surface. The datum is corroborated by other studies (17,(21)(22) , in which the patients had lagophthalmos frequency ranging from 31% to 54%. The fact is also pointed out by a study (23) in which this risk factor showed statistical significance (p=0.001), corroborating the results of this research.
Although the study (4)  Thus, in a way, these patients were exposed to various factors associated with ocular surface changes, and the emergence of dry eye.
Regarding the function of failure, we found close temporal relation between dry eye and hospitalization in the ICU, i.e. the longer the hospitalization time, the greater the risk of developing outcome in critically ill patients.
The risk factors identified in the study are related to the medical condition and treatment of the patients.

Conclusion
From the results, we can verify that dry eye in patients admitted to adults ICUs is a common finding, exposed to a set of internal and external risk factors that can collaborate for the emergence of the problem.
After bivariate analysis and multivariate analysis adjustment step, among the demographic and clinical factors identified, those that remained as better predictors for the phenomenon studied were: O 2 in room air, blinking more than five times per minute and presence of vascular disease.
The early recognition of risk factors for dry eye and, consequently, the adoption of preventive measures will certainly reduce the probability of ocular surface changes in critically ill patients.
It is recommended that the investigation of the risk factors described in NANDA I and which could not be validated according to the profile of the sample studied, such as: lifestyle, history of allergies, contact lenses, environmental factors and place where the patient lives.
In addition, we need studies that allow establishing what is the best nursing care for the prevention of the problem, particularly regarding critically ill patients.
We believe that this study may contribute to reflect on the relevance of the dry eye problem in critical and non-critical patients, in addition to a greater awareness and appreciation of the importance of eye care in patients admitted to adults ICUs, being a fundamental aspect for higher quality nursing care.