Gonzales CA, Ladas JG, Davis JL, Feuer WJ, Holland GN. Relationships Between Laser Flare Photometry Values and Complications of Uveitis. Arch Ophthalmol. 2001;119(12):1763-1769. doi:10.1001/archopht.119.12.1763
To determine whether relationships exist between elevated laser flare photometry values and common abnormalities and complications associated with uveitis.
We retrospectively studied all patients with uveitis on whom laser flare photometry measurements ("flare") were obtained (N = 111) at 2 academic medical centers. The first laser flare photometry values obtained for each patient were compared with the presence or absence of the following abnormalities or complications associated with uveitis: keratic precipitates, posterior synechiae, cataract, macular edema, optic disc edema, and glaucoma. In bilateral cases, the eye with the higher flare was used in primary analyses.
Flare was significantly higher in patients with posterior synechiae(P<.001) and in those with macular edema (P = .02) than in patients with uveitis who did not have these complications. Flare was significantly higher in patients with prior cataract surgery or cataract at the study visit than in those without cataracts(P = .001). There was no significant difference in flare between patients with and without keratic precipitates, optic disc edema, or glaucoma. No relationships were found between abnormalities or complications and the level of inflammatory cells or flare as determined by clinical assessment. We also identified an inverse relationship between flare and visual acuity that was not completely explained by the presence of complications in a stepwise regression model.
Although causal relationships were not established, associations between flare and some complications of uveitis suggest that aqueous humor protein may be an important factor in the development of these problems. Consequently, laser flare photometry could play a role in predicting outcomes or monitoring therapy for patients with uveitis.
SECONDARY complications such as cataracts, cystoid macular edema, or glaucoma can cause substantial visual morbidity in patients with uveitis. In contrast, the infiltration of inflammatory cells into uveal tissue probably does not directly affect vision.1 Although the mechanisms for development of sight-threatening complications are incompletely understood, alterations of the blood-aqueous humor and blood-retinal barriers, specifically the entry of serum proteins and inflammatory cells into the fluid-filled compartments of the eye, may cause complications by changing the internal ocular milieu.2
In keeping with these concepts, the evaluation and treatment of patients with uveitis has traditionally been based on inflammatory cell numbers in the aqueous and vitreous humors, a measure of disease severity that is clinically determined with slitlamp biomicroscopy.3 Clinical determination of aqueous humor protein levels ("flare") has been considered less important. This belief may have evolved from the greater difficulty in clinically identifying changes in protein levels. Whereas cells can be counted, the quantification of aqueous humor protein levels by slitlamp biomicroscopy is more subjective, based on the perceived intensity of light reflected off protein molecules within the aqueous humor (Tyndall effect).4
Laser flare photometry provides a more precise way to identify changes in aqueous humor protein. The Kowa FM-500 Flare Meter (Kowa Co Ltd, Electronics and Optics Division, Tokyo, Japan) projects a diode laser beam into the anterior chamber and measures the amount of light scattered by protein molecules in the aqueous humor.5 Laser flare photometry values are correlated with other objective measurements of aqueous humor protein concentration.6 We sought to determine whether relationships could be identified between laser flare photometry values and the presence of several abnormalities and complications associated with uveitis.
Laser flare photometry measurements are routinely performed on patients with uveitis at the 2 participating academic centers during their clinical evaluations. We retrospectively reviewed the records of all patients with uveitis at both centers for whom laser flare photometry values had been obtained. The first examination at which laser flare photometry was performed was selected as the study visit. In patients with bilateral uveitis, the eye with the higher laser flare photometry value was selected as the study eye. A secondary analysis was performed using right eyes for all bilateral cases.
Laser flare photometry was performed with the Kowa FM-500 using standard techniques as described by the Kowa Laser-Cell Flare Photometry Medical Advisory Board. Measurements were repeated until 7 acceptable readings (difference between 2 background measurements <15%) had been obtained from the study eye; the lowest and highest readings were deleted, and the machine then calculated the mean laser flare photometry value and SD of the remaining 5 measurements. The mean value for each patient was used for comparison. Laser flare photometry values are expressed as photon units per millisecond.
Medical records provided information about each patient included in the study. The following data were obtained for each eye at the study visit: best-corrected visual acuity, intraocular pressure, clinical grades of anterior chamber cells and flare, laser flare photometry values, and the presence or absence of 6 abnormalities or complications (keratic precipitates, posterior synechiae, cataract, macular edema, optic disc edema, and glaucoma). Although keratic precipitates are not necessarily vision-limiting complications of uveitis, we chose to investigate this abnormality because of its potential clinical relevance. Similar deposits in other tissues, such as the trabecular meshwork, might lead to functional problems. Also, certain types of keratic precipitates (eg, large granulomatous deposits) can probably damage the corneal endothelium. Because of the retrospective nature of our data, we were unable to identify the type of keratic precipitates present in most cases. The following demographic and historical information was obtained for each patient: age, sex, past ocular surgery, the characteristics of disease onset (sudden or indolent), and the duration of ocular disease. Uveitis was classified as anterior, intermediate, posterior, or panuveitis according to International Uveitis Study Group criteria.7 Cause, systemic disease associations, and ocular syndrome were identified for each case if known. All clinical assessments of cells and flare were performed by 1 of 2 observers(J.L.D. and G.N.H.) using standard definitions.4
Posterior synechiae were defined as adhesions between the iris and lens capsule or large pigment clumps on the anterior lens capsule consistent with past adhesions. For purposes of this study, patients were identified as having glaucoma if they had undergone prior glaucoma surgery, were using intraocular pressure–lowering eye drops for a long-term period, or had visual field or optic disc changes typical of those seen with glaucomatous optic nerve damage in the absence of a history of corticosteroid-induced elevation of intraocular pressure. Patients were considered to have cataracts if records identified them as having 2+ or more opacities of any kind (nuclear sclerotic, cortical, or posterior subcapsular) or any degree of posterior subcapsular lens opacity. Macular and optic disc edema were identified either clinically or by fluorescein angiography.
This study adhered to institutional review board policies at the University of Miami (Miami, Fla) and the University of California, Los Angeles.
The 2-sample t test, 1-way analysis of variance followed by least significant difference multiple comparisons, and logistic regression were used to determine whether laser flare photometry values differed between patients with and without each of the 6 uveitic abnormalities or complications. The raw data, expressed as photon units per millisecond, were considered inappropriate for statistical analysis because assumptions of normality and variance homogeneity were not met. Transformation of data is recommended in these circumstances8; a logarithmic conversion suited the laser flare photometry data best and was used in the analyses. Potential relationships between clinical grades of cells or flare and each of the 6 abnormalities or complications of uveitis were analyzed using the Mann-Whitney test. The Kendall τ rank correlation was used to analyze the relationships between laser flare photometry values and (1) intraocular pressure and (2) clinical grades of flare. The relationship between laser flare photometry values and visual acuity was analyzed with multiple linear regression.
We evaluated 111 patients with uveitis. There were 66 women and 45 men with a mean age of 42.5 years (range, 5-85 years). The characteristics of uveitis in these patients are summarized in Table 1 and Table 2. We did not find any significant relationships between laser flare photometry values and onset of disease, duration of disease, or anatomic classification.
Measures of inflammation (laser flare photometry value expressed as mean photon units per millisecond and as mean log of laser flare photometry value, clinical grade of flare, and clinical grade of cell) in patients with and without abnormalities and complications associated with uveitis are listed in Table 3. The presence of keratic precipitates was the most frequent of these findings (57 patients; 52%), followed by cystoid macular edema (43 patients; 41%), presence of cataract or history of cataract extraction (43 patients; 41%), and posterior synechiae (42 patients; 40%). Optic disc edema was less common (13 patients; 13%). The denominator used to calculate percentages varied for each abnormality or complication because the clinical record was incomplete for some patients; a patient was not included in a given analysis if there was no mention of either the presence or absence of the abnormality or complication being considered.
Log laser flare photometry values were significantly higher in eyes with present or past posterior synechiae than in eyes without posterior synechiae(P<.001), significantly higher in eyes with cataract or a history of cataract extraction than in phakic eyes without cataract (P = .001), and significantly higher in eyes with macular edema than in eyes without macular edema (P = .02). With regard to cataracts, eyes with a history of cataract extraction had significantly higher values (mean ± SD, 1.67 ± 0.49; n = 28) than phakic eyes without cataracts (mean ± SD, 1.26 ± 0.48; n = 61; P = .001). After eyes with prior cataract surgery were excluded, eyes with cataracts were found to have higher values (mean ± SD, 1.51 ± 0.64; n = 15) than phakic eyes without cataracts (mean ± SD, 1.26 ± 0.48; n = 61), but the difference was not significant (P = .10). There were no significant differences in values between eyes that did and did not have the following abnormalities or complications: keratic precipitates, optic disc edema, and glaucoma.
Logistic regression demonstrated that a 1-log unit (10-fold) increase in laser flare photometry value was associated with a 2.5-times greater likelihood of posterior synechiae (P = .002). A 1-log unit increase in laser flare photometry value was associated with a 2.6-times greater likelihood of macular edema (P = .03).
An analysis of variance with a test of first-order polynomials identified a significant relationship between laser flare photometry values and clinical grades of flare (P<.001) (Table 4). Although a significant negative correlation was found between intraocular pressure and clinical grades of flare (Kendall nonparametric τ; P = .04), a negative relationship between intraocular pressure and laser flare photometry values was not statistically significant. A potential relationship between intraocular pressure and log laser flare photometry values was also sought after dividing study eyes into 3 groups based on intraocular pressure (≤10 mm Hg, 11-20 mm Hg, and >20 mm Hg). Post hoc testing revealed that eyes with an intraocular pressure of 10 mm Hg or less had significantly higher values (mean ± SD, 1.6 ± 0.6) than eyes with intraocular pressures between 11 and 20 mm Hg (mean ± SD, 1.3 ± 0.5; P = 02). Patients with an intraocular pressure greater than 20 mm Hg had intermediate values (mean ± SD, 1.5 ± 0.5) that were not significantly different than values for eyes in the other 2 groups.
There was a significant correlation between fewer cells and glaucoma(P = .03), but the number of patients with glaucoma was small (n = 18), and these cases seemed to be scattered among all of the clinical grades of cells without a discernible pattern. There were no other statistically significant associations between abnormalities or complications and clinical grades of either flare or cells.
As a secondary analysis, comparisons were repeated using data from the right eye of all patients with bilateral uveitis. Results were generally in close agreement with the primary analyses presented previously, which used data from the eye with the highest laser flare photometry value. This agreement indicates the robust nature of our results. The only discrepancies included the following: there was a weak association between laser flare photometry values and keratic precipitates (mean ± SD, 1.4 ± 0.5 for eyes with current or past keratic precipitates vs 1.2 ± 0.5 for eyes without keratic precipitates; P = .05) that was not identified in the primary analysis. A stronger association between laser flare photometry values and macular edema was noted (mean ± SD, 1.5 ± 0.5 for eyes with macular edema vs 1.2 ± 0.5 for eyes without macular edema; P = .001) than in the primary analysis.
There was a negative correlation between log laser flare photometry values and visual acuity (expressed as the log of the minimum angle of resolution; P<.001). Linear regression was used to determine whether this correlation could be explained by complications of uveitis. Using univariate analysis, reduced visual acuity was significantly related to log laser flare photometry values (P<.001), cataract (P<.001), and macular edema (P = .02). A weak relationship between reduced visual acuity and glaucoma was found (P = .07). No relationships were identified between reduced visual acuity and keratic precipitates (P = .50), posterior synechiae (P = .40), or optic disc edema(P = .60). When variables were allowed to enter the regression model in a stepwise fashion, only log laser flare photometry values(P<.001) and cataract (P= .01) were significantly associated with reduced visual acuity. With all complications included in the model regardless of their statistically significant associations with vision, log laser flare photometry values were significantly correlated with reduced visual acuity (P<.001).
We undertook this project as a pilot study to determine whether relationships exist between elevated laser flare photometry values and selected abnormalities and complications associated with uveitis. The cross-sectional design allows rapid identification of such associations but does not confirm causal relationships.
We purposefully studied a heterogeneous population of patients with a variety of disease categories and diagnoses. No relationship was found between laser flare photometry measurements and specific categories of disease or diagnoses. Thus, we found no evidence that the relationships between abnormalities or complications associated with uveitis and laser flare photometry values were indirect, attributable to higher rates of those abnormalities or complications with particular diseases that are characterized by high aqueous protein levels. Future studies should control for potential confounding factors such as duration of disease. Although duration of disease most likely affects complication rates, we found no relationship between duration and laser flare photometry values.
Laser flare photometry has been used to characterize intraocular inflammatory reactions in a variety of disorders and clinical situations.5,6,9- 15 There is a significant correlation between laser flare photometry values and aqueous humor protein concentration measured ab externo,6 and laser flare photometry is believed to be a more reliable technique for measuring elevated protein levels in the anterior chamber than clinical grading of flare. Although we found the correlation between laser flare photometry values and clinical grades of flare to be statistically significant in this study, laser flare photometry values varied considerably for each clinical grade of flare. This fact may explain why relationships between clinical grades of flare and abnormalities or complications associated with uveitis have not been shown in this study or previous ones.
We found significantly higher laser flare photometry values in patients with posterior synechiae, cataracts or a history of cataract extraction, and macular edema. We also found an inverse relationship between laser flare photometry values and visual acuity that could not be explained on the basis of associations between laser flare photometry values and vision-limiting complications. Similar relationships were not identified between clinical grades of cells and complications or visual acuity.
The basis of the relationships between changes in aqueous humor protein composition or concentration as reflected by elevated laser flare photometry values and abnormalities or complications of uveitis is unknown, but it is possible that the proteins play a causal role. Breakdown of the blood-aqueous barrier in patients with uveitis is known to change the protein composition of the aqueous humor.16 The presence of fibrin could facilitate posterior synechiae formation. It is also possible that reactive chemokines, which probably represent a portion of aqueous humor proteins in the inflamed eye, play a role in the development of complications such as macular edema by altering vascular integrity or other cellular functions. Finally, the possibility remains that aqueous humor proteins do not cause damage but are merely markers linked to other disease processes that occur in an inflamed eye.
Possible relationships between aqueous humor protein levels and secondary cataracts are difficult to interpret. A significant relationship was identified when eyes with cataracts and those that had previously undergone cataract extraction were grouped together and compared with phakic eyes without cataracts. When tested separately, higher laser flare photometry values were found to be significantly different only for that subset of eyes that had undergone previous cataract surgery. Whereas this discrepancy might be attributable to small sample sizes, it raises the possibility that the relationship is actually related to the prior procedures rather than to the development of cataracts. In contrast to this possibility is the fact that laser flare photometry values have been shown to return to baseline levels after cataract extraction in otherwise healthy eyes.17 It is not known whether the same is true for eyes with uveitis, however. We chose to be inclusive in our definition of cataract in this study (any posterior subcapsular cataract or substantial nuclear sclerotic or cortical changes), even though posterior subcapsular cataracts are probably more closely related to uveitis than other lens opacities. We could not accurately differentiate the specific location and grade of opacities for all cases in this retrospective study. It is possible that future prospectively designed studies might identify specific relationships between elevated laser flare photometry values and cataracts if posterior subcapsular opacities alone were investigated.
We were unable to draw any conclusions regarding possible relationships between elevated laser flare photometry values and glaucoma; if they exist, relationships between aqueous humor protein and glaucoma are undoubtedly complex. Many factors are involved in the development of elevated intraocular pressure and glaucomatous optic nerve damage, including health of the ciliary body and aqueous humor production, the status of the anterior chamber angle and trabecular meshwork and the rate of aqueous humor outflow, and the susceptibility of the optic nerve to injury. Both animal studies18 and clinical experience3 have shown that intraocular pressure can be elevated or decreased in relation to the stage of disease in many forms of uveitis. Our ability to define and identify glaucoma among patients was limited by the retrospective nature of the study. Nevertheless, it is interesting to speculate about the relevance of our observations. The data demonstrated significantly higher laser flare photometry values in eyes with low intraocular pressure (<10 mm Hg) when compared with patients with normal intraocular pressure (11-20 mm Hg). This relationship might reflect disease chronicity, in which damage to the ciliary body and decreased aqueous humor production has occurred. Patients with high intraocular pressure (>20 mm Hg) had higher laser flare photometry values than those with normal pressure(11-20 mm Hg), although a statistically significant difference was not confirmed. If the relationship is true, it might reflect decreased aqueous humor outflow. In laboratory studies, Epstein et al19 demonstrated a decreased ability to perfuse the trabecular meshwork with solutions of higher protein content. We did not control for the use of aqueous humor suppressants, which may have influenced our findings.14,20 Future prospective studies should evaluate the relationships between laser flare photometry values, intraocular pressure, and glaucoma by examining each component (aqueous humor production, outflow, and optic nerve damage) separately.
This study has several limitations. The accuracy of retrospective data cannot be confirmed, and information was incomplete for some patients. Nevertheless, the robust nature of our data and the strength of some observed associations suggest that there are true relationships between elevated laser flare photometry values and abnormalities or complications associated with uveitis, though the nature of such relationships is yet to be determined. In addition, technical limitations may have influenced the results. Increased light scatter from pupils that dilated poorly because of posterior synechiae might have led to elevated laser flare photometry values. Oshika and Kato12 have reported laser flare photometry values that were 20% higher before dilation. In contrast, Sawa et al5 found that differences before and after dilation were not statistically significant. They also reported statistically higher laser flare photometry values in elderly patients, which were presumed to be caused by light scatter from cataracts. Whereas pupillary size and light scatter from cataracts might affect laser flare photometry values, the reported magnitude of these effects is too small to explain our findings. Because of the cross-sectional study design, laser flare photometry measurements and clinical examinations were performed at nonstandardized points in the course of each patient's disease; the single value does not reflect past severity of inflammation or anterior chamber protein composition or concentration, both of which might have influenced the development of complications that persisted after inflammation subsided.
In conclusion, this retrospective study has demonstrated associations between elevated laser flare photometry values and posterior synechiae, history of cataract, and macular edema in patients with uveitis. Additional studies will be needed to define these observed relationships more precisely and to clarify whether there are also relationships between elevated laser flare photometry values and other clinical findings that could not be confirmed in this study. Also, the basis for the association between changes in aqueous humor protein and decreased vision needs to be determined. If aqueous humor protein plays a role in the pathogenesis of uveitic complications, it must be determined whether the protein concentration in the aqueous humor, protein composition, or both affect complication rates. It also remains to be determined whether laser flare photometry values are independent predictors of complications, whether aqueous humor protein levels can be lowered therapeutically, and if so, whether lowering aqueous humor protein levels will also lower the associated risk of complications. These issues can be best investigated in longitudinal studies. Although grading of clinical flare has been deemed unimportant in uveitis treatment, physicians may eventually need to reconsider a role for monitoring changes in aqueous humor protein and using this information in clinical decision making.
Accepted for publication August 22, 2001.
This study was supported in part by Research to Prevent Blindness, New York, NY (Drs Davis and Holland); the Skirball Foundation (Dr Holland), and the Richard B. Shapiro Uveitis and Glaucoma Fund, Jules Stein Eye Institute, the David May II Endowed Professorship (Dr Holland), the Adelaide Stein Miller Fellowship Fund (Dr Gonzales), and the Jules Stein Eye Institute Clinical Research Center, University of California, Los Angeles. Dr Holland is a recipient of a Research to Prevent Blindness Lew R. Wasserman Merit Award.
Corresponding author and reprints: Gary N. Holland, MD, Jules Stein Eye Institute, 100 Stein Plaza, UCLA, Los Angeles, CA 90095-7003.