- Email: [email protected]
Laser flare photometry and complications of chronic uveitis in children
Laser Flare Photometry and Complications of Chronic Uveitis in Children JANET L. DAVIS, MD, LEONARDO M. DACANAY, MD, GARY N. HOLLAND, MD, AUDINA M. BERROCAL, MD, MICHAEL J. GIESE, OD, PHD, AND WILLIAM J. FEUER, MS
● PURPOSE: To investigate possible relationships between laser flare photometry values, complications of uveitis, and outcomes in children with chronic uveitis. ● DESIGN: Retrospective chart review. ● METHODS: We evaluated patients with active, noninfectious anterior, intermediate, or panuveitis who were 16 years of age or younger and who had laser flare photometry at one of two academic institutions. Complications enumerated at baseline and during follow-up were compared with laser flare photometry values and to anterior chamber cell counts. ● RESULTS: At least one laser flare photometry value (“flare”), defined as baseline measurement, was available for 59 patients (41 girls, 18 boys; mean age, 10.3 ⴞ 3.5 years); 38 of these patients had at least one additional measurement during follow-up (median 11 months). Complications of uveitis were present in 35 patients (59%) at baseline. There was a positive association between increased laser flare photometry values and complications at baseline (any complication [P ⴝ .007], posterior synechiae [P ⴝ .003]). The development of complications during follow-up was associated with the presence of complications at baseline (P ⴝ .018). A subgroup of patients with low flare at baseline had no complications during follow-up regardless of treatment given.
Accepted for publication March 12, 2003. InternetAdvance publication at ajo.com March 28, 2003. From the Bascom Palmer Eye Institute and Department of Ophthalmology (J.L.D., A.M.B., W.J.F.), University of Miami School of Medicine, Miami, Florida, and the Ocular Inflammatory Disease Center (L.M.D., G.N.H., M.J.G.), Jules Stein Eye Institute and Department of Ophthalmology, David Geffen School of Medicine at UCLA, University of California, Los Angeles, Los Angeles, California. This study was supported in part by Research to Prevent Blindness, Inc, New York, New York (Drs. Davis, Holland), the Skirball Foundation, Los Angeles, California (Dr. Holland), the Richard B. Shapiro Uveitis and Glaucoma Fund, Jules Stein Eye Institute (Dr. Holland), the Emily Plumb Estate Trust Fund, Jules Stein Eye Institute (Dr. Holland), and the David May II Endowed Professorship (Dr. Holland). Doctor Holland is a recipient of a Research to Prevent Blindness Physician-Scientist Award. Inquiries to Janet L. Davis, MD, Bascom Palmer Eye Institute, University of Miami School of Medicine, 900 NW 17th St, Miami, FL 33136; fax: (305) 326-6417; e-mail: [email protected] 0002-9394/03/$30.00 doi:10.1016/S0002-9394(03)00315-5
©
2003 BY
● CONCLUSIONS:
There is a positive relationship between laser flare photometry values and the prevalence of complications of uveitis in children. Laser flare photometry provides a novel way to monitor children with uveitis. Future study will be needed to determine whether values have prognostic importance and whether a treatment strategy that minimizes flare results in fewer uveitic complications. (Am J Ophthalmol 2003;135: 763–771. © 2003 by Elsevier Inc. All rights reserved.)
C
HRONIC UVEITIS IN CHILDREN HAS HISTORICALLY
been associated with a high rate of complications.1–3 Ocular factors predictive of these complications, and the best means to monitor children with uveitis is not well established. Laser flare photometry provides an objective and quantitative, noninvasive, in vivo measurement that correlates with aqueous humor protein levels in the anterior chamber of the eye. Its full role in the assessment of uveitis remains undefined, but a previous study involving a heterogeneous group of patients with uveitis showed associations between laser flare photometry values and the prevalence of some complications of uveitis, including posterior synechiae, cataracts, and macular edema.4 In contrast, there were no associations between these complications and the number of anterior chamber cells by clinical assessment. These observations suggest that laser flare photometry may be a useful supplement to the traditional strategy of monitoring cells during follow-up of patients with uveitis. We investigated the association between laser flare photometry values and the prevalence of uveitic complications, both at initial measurement and during follow-up, in a population of children with chronic uveitis.
METHODS LASER FLARE PHOTOMETRY MEASUREMENTS WERE AT-
tempted routinely for all children with uveitis as a part of standard clinical evaluations at the two participating academic centers. Laser flare photometry readings could not be obtained for all children, or at every visit for
ELSEVIER INC. ALL
RIGHTS RESERVED.
763
periocular injection of corticosteroid; or topical prednisolone acetate 1% at a frequency of more than four times daily for longer than 4 consecutive weeks at any time during follow-up. Conservative therapy was all other treatments, including observation. Athough the more intensive therapy group encompassed a broad range of treatments, these arbitrary categories reflect our treatment practices, and patients could be placed easily into one of the two groups on the basis of available information. A more specific evaluation of treatment was not possible because of the retrospective nature of the study. One of our major goals was to investigate the relationship between flare and complications in patients treated conservatively; this goal was appropriately served by these cateogories. Other definitions were consistent with those used in our previous studies of laser flare photometry.4,8 “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. Patients were identified as having “glaucoma/ocular hypertension” if they had a previous diagnosis of glaucoma/ocular hypertension; had undergone prior glaucoma/ocular hypertension surgery; were using intraocular pressure-lowering medications chronically; had intraocular pressure measurements greater than 24 mm Hg at any visit; or had visual field or optic disk changes typical of those seen with glaucomatous optic nerve damage; and had no history of corticosteroid-induced elevation of intraocular pressure. Patients were considered to have “cataracts” if records identified them as having 2⫹ or more nuclear or cortical opacity or any posterior subcapsular lens opacities. “Macular edema” was identified either clinically or by fluorescein angiography. “Hypotony” was defined as an intraocular pressure of less than 5 mm Hg. Patients were considered to have “vision loss” if best-corrected visual acuity dropped 2 or more Snellen lines on follow-up examination. Only one patient, who had light perception vision, was judged to have vision loss at baseline. For purposes of categorizing patients, “low flare” was defined as a laser flare photometry value of less than 20 photon units/msec, and “high flare” was defined as a laser flare photometry value of 20 photon units/msec or greater, in accordance with definitions of high and low flare in a previous publication.8 Laser flare photometry was performed by technicians using the Kowa FM-500 (Kowa Company, Ltd., Electronics and Optics Division, Tokyo, Japan). Technicians were unaware of the treatment or the presence or absence of complications (other than posterior synechiae that could be observed during measurement). Measurements were performed using standard procedures described by the Kowa Laser-Cell Flare Photometry Medical Advisory Board.4,8 Briefly, measurements were repeated until seven acceptable readings (difference between two background measurements less than 15%) had been taken for the study eye; the lowest and highest readings were deleted, and the machine then calculated the mean laser flare photometry
children included in this study, because of age, lack of cooperation, or other factors unrelated to their diagnoses. We retrospectively reviewed the records of all children for whom laser flare photometry values had been obtained at least once and who had active, noninfectious anterior uveitis, intermediate uveitis, or panuveitis. A child was defined as a patient whose age was 16 years or less at baseline examination. The first examination at which laser flare photometry was performed was selected as the baseline visit. Only one eye of each patient was studied; in cases with bilateral uveitis, the right eye was selected as the study eye, unless the right eye was not evaluable, in which case the left eye was used. Information about each patient was obtained from a review of medical records. The following demographic and historical information was obtained for each patient: age at baseline examination, sex, and past ocular surgery. The following information was obtained for each study eye at the baseline examination: best-corrected visual acuity, intraocular pressure, clinical assessment of anterior chamber cells, laser flare photometry value, and the presence or absence of six complications associated with uveitis: posterior synechiae, cataract, glaucoma or ocular hypertension, hypotony, macular edema, and vision loss from any cause. The inclusion of posterior synechiae in this list of complications is justified because they are a measure of severe disease and their presence predicts poorer outcomes.2,5 Clinical assessments had all been performed by one of two authors (J.L.D., G.N.H.). Anterior chamber cells were scored clinically by the method of Hogan, Kimura, and Thygeson.6 Uveitis was classified as anterior, intermediate, or pan-uveitis according to International Uveitis Study Group criteria.7 Cause, systemic disease associations, or ocular syndrome were identified for each case, if known. Patients with at least one follow-up clinical examination were evaluated for the development of complications after baseline. Records were reviewed for each clinical examination until development of a new complication was noted or until last follow-up clinical examination for patients who did not develop complications during follow-up. The following information was recorded for study eyes at each follow-up examination: duration since baseline examination, best-corrected visual acuity, intraocular pressure, clinical assessment of anterior chamber cells, laser flare photometry value, if performed, and the presence or absence of the six complications described above. Baseline and follow-up laser flare photometry values were compared if follow-up laser flare photometry was obtained within the 6 months before or after the development of a new complication or the 6 months before the last clinical examination of patients without new complications. For purposes of this study, “more intensive therapy” was defined as the use of oral prednisone, immunosuppressive drugs, immunomodulatory biologics, or a combination of these agents for longer than 6 consecutive weeks; a 764
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
JUNE 2003
value and standard deviation of the remaining five readings. The mean value for each patient was used for comparisons. Photometry readings are expressed as photon units / msec. This study was approved by the Institutional Review Boards at the University of Miami and the University of California, Los Angeles. There is a linear relationship between laser flare photometer readings and total protein content in aqueous humor.9 For statistical analysis, either the laser flare reading or its logarithm was used. The raw data expressed as photon units/msec were considered inappropriate for some statistical analysis because assumptions of normality and variance homogeneity were not met. Transformation of data is recommended in these circumstances10; a logarithmic conversion suited the laser flare photometry data best, and was used in analyses. The primary outcome was defined as prevalence of complications for patients grouped by different levels of flare. The secondary outcome was defined as the development of new or additional complications during follow-up in patients with different levels of flare who were receiving conservative or more intensive therapy. The associations between prevalence of complications at baseline and patient and ocular characteristics were evaluated with the Fisher exact test, the 2 test, or logistic regression as appropriate. Kaplan-Meier survival analysis with the log-rank test and Cox proportional hazards survival regression were used to study incidence of new complications in relation to baseline characteristics, treatment, and follow-up laser flare photometry values. The use of some additional analyses are documented in the tables. A P value of less than .05 was considered to be statistically significant.
RESULTS WE EVALUATED 59 CHILDREN WITH UVEITIS. DEMOGRAPHIC
and ophthalmic characteristics are shown in Table 1. Of 41 children with a diagnosis of anterior uveitis, 17 (29% of all patients) had arthritis (juvenile rheumatoid arthritis [JRA] or juvenile idiopathic arthritis [JIA]). Bilateral uveitis was present in 52 children (88%). In one patient with bilateral uveitis attributed to sympathetic ophthalmia, laser flare photometry values from the left (sympathizing) eye were used, as the right (inciting) eye was not evaluable because of posttraumatic changes. A mild anterior chamber cellular reaction (ⱕ 1⫹ cells) was present in 59% of patients. The median laser flare photometry value for all patients (23.0 photon units/msec [range, 0.8 – 675 photon units/msec]) fell within our definition of high flare. Overall, 58% of eyes had at least one complication at baseline: cataracts (37%), posterior synechiae (37%) glaucoma/ocular hypertension (12%), and macular edema (5%). VOL. 135, NO. 6
LASER FLARE PHOTOMETRY
AND
There were several differences between the baseline characteristics of patients at the two centers (data not shown), but the ranges for all factors from both centers overlapped. At site 1, there were more boys (P ⫽ .072) and a higher proportion of patients with a diagnosis of JRA (P ⫽ .071, comparing prevalence of all diagnoses). Several other differences were consistent with a higher proportion of patients having severe disease at site 1; patients at site 1 had higher median laser flare photometry values (P ⫽ .03), lower visual acuities (P ⫽ .004), and a higher prevalence of complications (P ⫽ .05). In contrast, there was a higher proportion of patients at site 2 who had been assigned baseline clinical scores of 2⫹ or 3⫹ for anterior chamber cells (P ⫽ .002, 2 test). There was a statistically significant positive correlation between log laser flare photometry value and the presence of complications at baseline (P ⫽ .001, Table 2). When the large number of eyes with posterior synechiae were excluded, the relationship remained, although it was less strong (P ⫽ .017). Because the primary site of inflammation may influence rate of complications, we also investigated the association between flare and the prevalence of complications for the subgroup of patients with anterior uveitis only (n ⫽ 41). Results were similar to those for all patients; those with anterior uveitis and high flare had a significantly higher prevalence of any complication (P ⫽ .031) and of posterior synechiae (P ⫽ .008). Anterior chamber cell counts did not correlate with baseline complications (P ⫽ .54, exact 2 test, Table 2). Because there were generally higher clinical scores for baseline cells at site 2 than at site 1, we compared clinical scores for cells to prevalence of complications for each site individually, and no statistical associations were identified. Table 3 provides follow-up data that were available for 45 patients (mean duration of follow-up 15 ⫾ 14 months, median 11, range, 1– 64). Of these patients, 33 had laser flare photometry values at follow-up. We found no difference in the proportion of patients who had follow-up examinations between those with complications at baseline (28 of 35 children [80.0%]) and those without complications at baseline (17 of 24 children [70.8%], P ⫽ .54, Fisher exact test). Also, there was no difference in the proportion of patients who had follow-up examinations between those with low flare at baseline (20 of 26 children [76.9%]) and those with high flare at baseline (25 of 33 children [75.8%], P ⫽ 1.00, Fisher exact test). The frequencies of individual complications during follow-up are shown in Table 3 as descriptive data. The cumulative percentage of eyes with new complications rose over the period of follow-up (Table 4). Stated rates for individual complications do not account for patients lost to follow-up; thus, true rates may be higher or lower than shown in Table 4. Cataract and glaucoma/ocular hypertension made up the majority of new complications. There were significantly more new complications in eyes with complications at baseline than in those without complicaCHRONIC UVEITIS COMPLICATIONS
765
TABLE 1. Patient Demographic and Ocular Characteristics at Initial Examination
Number of patients Sex Female (patients [percentage]) Mean age (years ⫾ SD) Category of uveitis (patients [percentage]) Anterior uveitis Idiopathic iridocyclitis JRA-associated Intermediate uveitis Panuveitis Laterality (patients [percentage]) Bilateral Median visual acuity (range) Laser flare photometry value Median (photon units/msec, [range]) Mean log value ⫾ SD Anterior chamber cells ⬍1⫹ 1⫹ 2⫹ 3⫹ Complications (eyes) Any Posterior synechiae Cataract Glaucoma/ocular hypertension Hypotony Macular edema Vision loss
All Patients
Patients With Low Flare*
Patients With High Flare†
59
26
33
41 (70%) 10.3 ⫾ 3.5
19 (73%) 10.7 ⫾ 3.6
22 (67%) 9.9 ⫾ 3.4
41 (70%) 24 (41%) 17 (29%) 9 (15%) 9 (15%)
17 (66%) 10 (38%) 7 (27%) 3 (12%) 6 (23%)
24 (72%) 14 (42%) 10 (30%) 6 (18%) 3 (9%)
52 (88%) 20/20 (20/15–HM)
24 (92%) 20/20 (20/15–6/200)
28 (85%) 20/30 (20/15–HM)
23.0 (0.8–675) 1.43 ⫾ 0.67
7.0 (0.8–18.8) 0.82 ⫾ 0.30
70.9 (21.7–675) 1.90 ⫾ 0.46
11 (19%) 24 (41%) 13 (22%) 11 (19%)
3 (12%) 14 (44%) 6 (23%) 3 (12%)
8 (24%) 10 (30%) 7 (21%) 8 (24%)
35 (59%) 22 (37%) 22 (37%) 7 (12%) 0 (0%) 3 (5%) 1 (2%)
10 (38%) 4 (15%) 6 (23%) 4 (15%) 0 0 0
25 (76%) 18 (55%) 16 (49%) 3 (9%) 0 3 (9%) 1 (3%)
P Value‡
0.78§ 0.39¶ 0.51㛳
0.45§ 0.003** ⬍0.001* ⬍0.001¶ 0.94**
0.007§ 0.003§ 0.060§ 0.69§ 0.25§ 1.00§
HM ⫽ hand movement; JRA ⫽ juvenile rheumatoid arthritis; SD ⫽ standard deviation. *Low flare defined as laser flare photometry values ⬍20 photon units/msec. † High flare defined as laser flare photometry values ⱖ20 photon units/msec. ‡ Comparison of patients with low flare versus those with high flare. § Fisher exact test. ¶ Two sample t test. 㛳 Exact 2. **Mann-Whitney test.
tions at baseline (P ⫽ .018, Kaplan-Meier method with log-rank test, Table 4). Only one of 17 eyes without complications at baseline developed a new complication during follow-up; the patient was a 7-year-old girl with bilateral idiopathic iridocyclitis who had high flare at baseline and developed glaucoma/ocular hypertension at 4 months of follow-up. Laser flare photometry values at baseline were statistically associated with the development of new complications during follow-up, but the association was not strong: a 2-log larger baseline laser flare photometry value was associated with a 1.35 higher risk of new complications (P ⫽ .047, Cox regression). Baseline complications and baseline flare were not independent predictors of new complications, however, based on a Cox proportional hazards 766
AMERICAN JOURNAL
survival regression. Baseline laser flare photometry values in patients with anterior uveitis were weakly associated with development of new complications at follow-up (P ⫽ .048). A relationship between laser flare photometry values at follow-up and new complications could not be identified after accounting for complications at baseline. Also, there was no statistical difference between incidence of new complications during follow-up and the intensity of treatment (P ⫽ .59, Kaplan-Meier method with log-rank test). For the 33 eyes with both baseline and follow-up laser flare photometry values, mean values decreased during follow-up; mean baseline values were 76.3 ⫾ 145.9 photon units/msec and mean last follow-up values were 51.0 ⫾ 76.0 photon units/msec, a drop of ⫺25.4 ⫾ 95.2 photon units/msec from baseline to last follow-up. This result was OF
OPHTHALMOLOGY
JUNE 2003
P Value
complications in patients with low flare cannot be attributed to inadequate follow-up. Likewise, the single, new complication that occurred during follow-up was seen at 4 months in a patient with high flare, whereas those without complications at baseline who had low flare (none of whom had new complications) were followed up for a mean of 21.3 months (range, 3– 64).
4/18 (22%) 16/26 (62%) 15/15 (100%)
0.001†
DISCUSSION
8/11 (73%) 12/24 (50%) 9/13 (69%) 6/11 (55%)
0.54§
TABLE 2. Prevalence of Complications at Baseline Examination in Patients Grouped by Initial Laser Flare Photometry Value and by Initial Clinical Anterior Chamber Cell Count Prevalence of Complications*
Patient Groups
By laser flare photometry value (photon units/msec) ⬍10 10–100 ⬎100 By clinical cell count‡ ⬍1⫹ 1⫹ 2⫹ 3⫹
IN THIS RETROSPECTIVE STUDY, THERE WAS AN ASSOCIA-
*Number of patients with any complication at baseline examination/number of patients with specified laser flare photometry values or clinical cell count. † Significance analyzed with logistic regression in which log laser flare photometry value was evaluated as a continuous independent variable and presence of any complication was used as the dependent variable. ‡ As defined by Hogan and associates.6 § Exact 2 test.
influenced by some individual large decreases. Change was correlated with baseline laser flare photometry values; greater baseline values tended to decrease more during follow-up (P ⫽ .007; Figure 1) In contrast, the median change was an increase of 0.5 photon units/ms (range ⫺359.2 to ⫹ 169.1 photon units/msec). Change was not correlated with follow-up duration. Among the points shown in Figure 1, the largest increase in flare during follow-up was in a patient treated conservatively, whereas the largest drop in flare occurred in a patient treated more intensively. Overall, however, we could not identify a statistical association between change in flare and intensity of treatment. Table 5 shows patients grouped by baseline flare (low vs high) and by medical therapy (conservative vs more intensive). Among 25 children with high flare at baseline, complications developed in 11 (44%) during follow-up (four of seven [57%] who were treated conservatively and seven of 18 [39%] who were treated more intensively; P ⫽ .66, Fisher exact test). In contrast, complications developed in only two (10%) of 20 children with low flare at baseline (P ⫽ .02 when compared with all patients with high flare at baseline, Fisher exact test); both were treated intensively. The duration of follow-up was longer for those patients with low flare in whom complications did not develop (median, 13 months; range, 1– 64) than for those with high flare in whom complications did develop (median, 8 months; range, 3–32); thus, the lower rate of VOL. 135, NO. 6
LASER FLARE PHOTOMETRY
AND
tion between higher laser flare photometry values and the prevalence of complications of uveitis at baseline, similar to the findings of Gonzales and associates.4 In contrast to that study, the current investigation contained a more homogeneous group that included just children with uveitis. A cause and effect relationship between flare and complications was not established, however. Uveitis is known to alter the blood-aqueous barrier in the anterior segment, changing the protein composition of the aqueous humor.11 Elevated intraocular protein may be a cause of complications, or it may simply be a marker for other pathologic processes in the uveitic eye that lead to complications.8,12 Our results are consistent with the high rates of complications that have been reported in children with uveitis by others.1–3 Because not every patient had follow-up data, it is possible that the true rates of complications are less than those we report, as it is more likely that patients with complications would return for follow-up than would those without complications. Nevertheless, follow-up did not differ statistically between children with and without complications at baseline or between children grouped on the basis of low versus high flare at baseline. Previous studies have shown that the presence of posterior synechiae at initial examination is predictive of poorer outcomes.2,5 In this study, both higher baseline laser flare photometry values and presence of complications at baseline were associated with the subsequent development of new complications on univariate analyses. Furthermore, patients with baseline complications continued to develop additional complications during follow-up, even with more intensive therapy. Flare was not identified as an independent predictor of complications during follow-up, but intervals were relatively brief and patients were receiving medical treatment. In addition, when complications are preexisting, new events of the same type cannot occur. Because of the strong association between flare and baseline complications, it is possible that flare, measured early in the course of disease before development of complications, would predict their development. A longitudinal study with consistent follow-up would be necessary to demonstrate that. The intensity of treatment was not independently associated with the development of new complications. Small patient numbers and selection of CHRONIC UVEITIS COMPLICATIONS
767
TABLE 3. Data From Follow-up Examinations
Number of eyes with follow-up data Duration of follow-up Mean (months ⫾ SD) Median (months [range]) Treatment during follow-up interval Conservative therapy㛳 More intensive therapy** Log laser flare photometry value†† Number of eyes evaluated Mean increase (⫾ SD) Median visual acuity at last follow-up examination (range) Eyes with new complications Any complication Posterior synechiae Cataracts Glaucoma/ocular hypertension Hypotony Macular edema Vision loss
All Patients
Patients With Low Flare*
Patients With High Flare†
45
20
25
15 ⫾ 14 11 [1,64]
17 ⫾ 17 13 [1,64]
13 ⫾ 10 8 [1,32]
P Value‡
0.52§ 0.35¶
16 (36%) 29 (64%)
9 (45%) 11 (55%)
7 (28%) 18 (72%)
33 0.03 ⫾ 0.28‡‡ 20/20 [20/15–CF]¶¶
16 0.00 ⫾ .18 20/20 [20/15–6/200]
17 0.06 ⫾ 0.35 20/30 [20/20–CF]
13 0 5 5 0 1 2
2 0 1 0 0 1 0
11 0 4 5 0 0 2
0.54§§ 0.003§
CF ⫽ counting fingers; SD ⫽ standard deviation. *Low flare defined as laser flare photometry values ⬍20 photon units/msec. † High flare defined as laser flare photometry values ⱖ20 photon units/msec. ‡ Comparison of patients with low flare versus those with high flare. § Mann-Whitney test. ¶ Fisher exact test. 㛳 Conservative therapy was all treatments other than aggressive therapy, as defined below, including observation. **More intensive therapy was defined as the use of oral prednisone, immunosuppressive drugs, immunomodulatory biologics, or a combination of these agents for longer than 6 consecutive weeks; a periocular injection of corticosteroid; or topical prednisolone acetate 1% at a frequency of more than 4 times daily for longer than 4 consecutive weeks. †† Measurement closest to last examination, if made within 6 months prior to examination. ‡‡ P ⫽ .48, paired t test; when compared with baseline value. §§ Two sample t test. ¶¶ P ⫽ .52, Wilcoxon paired rank test, when compared with baseline value.
therapy based on clinical perception of uveitic activity (based in part on flare levels) may have obscured these relationships. Among patients with high flare at baseline who were treated conservatively, those in whom complications did not develop were followed up for shorter durations (median, 3 months; range, 2–11) than those in whom complications did develop (median, 7.5 months; range, 3–28); thus, it is possible that patients with high flare are at increased risk for eventual complications if they are treated conservatively. Consensus is developing among uveitis specialists that intensive treatment will improve outcomes for children with uveitis. Our numbers were too small to show that complications are reduced by more intensive therapy, but our study does show that intensive therapy alone does not 768
AMERICAN JOURNAL
prevent complications in all cases, especially in children with preexisting complications. In view of the potential adverse effects of systemic immunosuppressive therapy, it is important to identify factors that predict an increased risk of uveitis-associated complications, and thus a need for aggressive therapy. There may be a subgroup of patients with chronic uveitis and low flare in whom complications do not develop and who retain excellent vision over many years despite a persistent anterior chamber cellular reaction, even in the absence of intensive therapy. Of concern is whether the use of immunosuppressive drugs in such children poses a greater risk than the persistent low-grade cellular reaction. Even with intensive treatment, some children with low flare in this series did develop complications, indicating OF
OPHTHALMOLOGY
JUNE 2003
TABLE 4. Incidence of New Complications in Patients Grouped by Interval to Last Examination Prevalence of Complications* Follow-up Interval
12 months 24 months 36 months§
Patients Grouped by Baseline Flare† All Patients
22% (⫾7%) 27% (⫾8%) 66% (⫾16%)
Patients Grouped by Baseline Complications‡
Low
High
Absent
Present
6% (⫾6%) 6% (⫾6%) 30% (⫾21%)
34% (⫾10%) 43% (⫾12%) 100%
7% (⫾7%) 7% (⫾7%) 7% (⫾7%)
30% (⫾9%) 38% (⫾11%) 100%
*Cumulative percentages (⫾ standard errors). † The difference in cumulative proportion without new complications during follow-up is significantly different between eyes with low flare (⬍20 photon units/msec) and high flare (ⱖ 20 photon units/msec) at baseline (P ⫽ .021, Kaplan-Meier method and log rank test). ‡ The difference in cumulative proportion without new complications during follow-up is significantly different between eyes with complications at baseline and those without complications at baseline (P ⫽ .018, Kaplan-Meier method and log rank test). § Few patients were followed up longer than 30 months, limiting the reliability of findings.
FIGURE 1. This scattergram includes 33 patients who had both baseline and follow-up laser flare photometry measurements. Points above the diagonal indicate patients whose laser flare photometry values increased, whereas those below the line indicate patients whose values decreased. The decrease in the overall mean is due to the influence of a few points with high baseline values.
that children with uveitis require close follow-up, regardless of findings or treatment. Because intensive corticosteroid treatment itself can cause cataract and ocular hypertension, additional long-term follow-up of many paVOL. 135, NO. 6
LASER FLARE PHOTOMETRY
AND
tients will be required to determine the management strategy that provides the greatest benefits with the least risk. Decreases in laser flare photometry values from baseline to final follow-up occurred in several patients, most notaCHRONIC UVEITIS COMPLICATIONS
769
770 TABLE 5. Development of Complications During Follow-up in Patients Grouped by Baseline Laser Flare Photometry Value and Intensity of Treatment All Patients
Conservative
AMERICAN JOURNAL OF
OPHTHALMOLOGY
Findings at baseline Number of eyes Laser flare photometry value (photon units/msec) Mean ⫾ SD Range Findings at follow-up New complications Number of patients Duration of follow-up (months) Median Range Mean laser flare photometry value (photon units/ msec ⫾ SD) at follow-up (number of patients evaluated) No complications at follow-up Number of patients Duration of follow-up (months) Median Range Mean laser flare photometry value (photon units/msec ⫾ SD) at follow-up (number of patients evaluated)
‡
Patients Without Complications at Baseline High Flare†
Low Flare* More Intensive
9
11
8.0 ⫾ 5.2 3.6–18.8
9.0 ⫾ 4.8 0.8–16.5
0 —
§
Conservative
‡
7
High Flare†
Low Flare*
More Intensive
§
Conservative
‡
More Intensive
§
Conservative‡
More Intensive§
18
7
5
1
4
131.6 ⫾ 136.2 28.3–360.6
153.5 ⫾ 174.1 21.7–675.4
6.9 ⫾ 5.4 3.6–18.8
8.0 ⫾ 3.4 3.6–13.1
32.5 32.5
34.9 ⫾ 24.0 22.6–70.9
2 (18%)
4 (57%)
7 (39%)
0
0
0
1 (25%)
7.5 3–28 116.0 ⫾ 89.9 (4)
8 3–32 133.9 ⫾ 147.0 (4)
—
—
—
—
16 6–26 9.9 (1)
—
—
—
4 — 73.8 (1)
9 (100%)
9 (82%)
3 (43%)
11 (61%)
7 (100%)
5 (100%)
1 (100%)
3 (75%)
13 3–64 6.8 ⫾ 2.4 (7)
13 1–26 9.2 ⫾ 4.3 (8)
3 2–11 141.5 (1)
13 1–30 51.3 ⫾ 44.5 (8)
22 3–64 6.5 ⫾ 2.5 (6)
13 9–26 8.3 ⫾ 4.6 (5)
3 — (0)
6 3–14 26.3 ⫾ 11.8 (3)
*Low flare defined as laser flare photometry values ⬍ 20 photon units/ms. High flare defined as laser flare photometry values ⱖ 20 photon units/ms. ‡ Conservative therapy was all treatments other than aggressive therapy, as defined below, including observation. § More intensive therapy was defined as the use of oral prednisone, immunosuppressive drugs, immunomodulatory agents, or a combination of these agents for longer than 6 consecutive weeks; a periocular injection of corticosteroid; or topical prednisolone acetate 1% at a frequency of more than 4 times daily for longer than 4 consecutive weeks. †
JUNE 2003
bly those with higher baseline flares. Contrary to the assumptions of some clinicians, it may be possible to reduce aqueous humor protein concentrations with treatment in some patients. Current concepts regarding the persistence of elevated flare in patients with uveitis may relate to the inability to appreciate small changes in flare clinically; small differences in aqueous humor protein content can be identified with laser flare photometry, however.4 It is unclear whether reduction of flare would alter the course of uveitis or reduce the risk of complications, but it is reasonable to speculate that therapy to minimize flare may be an effective strategy to improve outcomes for pediatric patients with uveitis. Clinical scores for anterior chamber cells were not associated with either complications at baseline or with new or additional complications during follow-up. This observation may reflect a true lack of association between the number of anterior chamber cells and complications, or it may reflect an inability of clinicians to quantitate cells with sufficient precision to identify a true relationship. In either case, clinical cell counts alone would appear not to be sufficient to assess the severity of uveitis. Precise measurement of flare provides an alternative means of assessment. Our results support the use of laser flare photometry to monitor uveitis in children. Although our small numbers and lack of controls limit our ability to draw firm conclusions about the observed associations, we believe they are sufficiently convincing to support continued investigation of this technique as a method for monitoring children early in the course of uveitis and potentially for predicting their risk of complications so that therapy of the proper intensity can be prescribed. ACKNOWLEDGMENTS
Kowa Company, Ltd, Electronics and Optics Division, Tokyo, Japan, provided Dr Holland with an FM-500 Flare Meter for use in research activities during a portion of the period in which data used in this article were being
VOL. 135, NO. 6
LASER FLARE PHOTOMETRY
AND
collected. Doctor Holland was an unpaid member of the Kowa Laser Cell-Flare Photometry Medical Advisory Board from 1991 to 1995. The authors have no other interest in the products or techniques described in this report or in competing techniques.
REFERENCES 1. Edelsten C, Reddy MA, Stanford MR, Graham EM. Visual loss associated with pediatric uveitis in English primary and referral centers. Am J Ophthalmol 2003;135:757–762. 2. Wolf MD, Lichter PR, Ragsdale CG. Prognostic factors in the uveitis of juvenile rheumatoid arthritis. Ophthalmology 1987;94:1242–1248. 3. Tugal-Tutkun I, Havrlikova K, Power WJ, Foster CS. Changing patterns in uveitis of childhood. Ophthalmology 1996;103:375–383. 4. Gonzales CA, Ladas JG, Davis JL, et al. Relationships between laser flare photometry values and complications of uveitis. Arch Ophthalmol 2001;119:1763–1769. 5. Edelsten C, Lee V, Bentley CR, et al. An evaluation of baseline risk factors predicting severity in juvenile idiopathic arthritis associated uveitis and other chronic anterior uveitis in early childhood. Br J Ophthalmol 2002;86:51–56. 6. Hogan MJ, Kimura SJ. Signs and symptoms of uveitis, I. Anterior uveitis. Am J Ophthalmol 1959;47:155–170. 7. Bloch-Michel E, Nussenblatt RB. International Uveitis Study Group recommendations for the evaluation of intraocular inflammatory disease. Am J Ophthalmol 1987;103:234 – 235. 8. Ladas JG, Yu F, Loo R, et al. Relationship between aqueous humor protein level and outflow facility in patients with uveitis. Invest Ophthalmol Vis Sci 2001;42:2584 –2588. 9. Shah SM, Spalton DJ, Taylor JC. Correlations between laser flare measurements and anterior chamber protein concentrations. Invest Ophthalmol Vis Sci 1992;33:2878 –2884. 10. Fleiss JL. The design and analysis of clinical experiments. New York, Wiley & Sons, 1986, 59 –68. 11. Zirm M. Proteins in aqueous humor. Adv Ophthal 1980;40: 100 –172. 12. Nguyen QD, Humphrey RL, Dunn JP, Humayun MS. Elevated vitreous concentration of monoclonal immunoglobulin manifesting as schlieren in juvenile rheumatoid arthritisassociated uveitis. Arch Ophthalmol 2001;119:293–296.
CHRONIC UVEITIS COMPLICATIONS
771
● PURPOSE: To investigate possible relationships between laser flare photometry values, complications of uveitis, and outcomes in children with chronic uveitis. ● DESIGN: Retrospective chart review. ● METHODS: We evaluated patients with active, noninfectious anterior, intermediate, or panuveitis who were 16 years of age or younger and who had laser flare photometry at one of two academic institutions. Complications enumerated at baseline and during follow-up were compared with laser flare photometry values and to anterior chamber cell counts. ● RESULTS: At least one laser flare photometry value (“flare”), defined as baseline measurement, was available for 59 patients (41 girls, 18 boys; mean age, 10.3 ⴞ 3.5 years); 38 of these patients had at least one additional measurement during follow-up (median 11 months). Complications of uveitis were present in 35 patients (59%) at baseline. There was a positive association between increased laser flare photometry values and complications at baseline (any complication [P ⴝ .007], posterior synechiae [P ⴝ .003]). The development of complications during follow-up was associated with the presence of complications at baseline (P ⴝ .018). A subgroup of patients with low flare at baseline had no complications during follow-up regardless of treatment given.
Accepted for publication March 12, 2003. InternetAdvance publication at ajo.com March 28, 2003. From the Bascom Palmer Eye Institute and Department of Ophthalmology (J.L.D., A.M.B., W.J.F.), University of Miami School of Medicine, Miami, Florida, and the Ocular Inflammatory Disease Center (L.M.D., G.N.H., M.J.G.), Jules Stein Eye Institute and Department of Ophthalmology, David Geffen School of Medicine at UCLA, University of California, Los Angeles, Los Angeles, California. This study was supported in part by Research to Prevent Blindness, Inc, New York, New York (Drs. Davis, Holland), the Skirball Foundation, Los Angeles, California (Dr. Holland), the Richard B. Shapiro Uveitis and Glaucoma Fund, Jules Stein Eye Institute (Dr. Holland), the Emily Plumb Estate Trust Fund, Jules Stein Eye Institute (Dr. Holland), and the David May II Endowed Professorship (Dr. Holland). Doctor Holland is a recipient of a Research to Prevent Blindness Physician-Scientist Award. Inquiries to Janet L. Davis, MD, Bascom Palmer Eye Institute, University of Miami School of Medicine, 900 NW 17th St, Miami, FL 33136; fax: (305) 326-6417; e-mail: [email protected] 0002-9394/03/$30.00 doi:10.1016/S0002-9394(03)00315-5
©
2003 BY
● CONCLUSIONS:
There is a positive relationship between laser flare photometry values and the prevalence of complications of uveitis in children. Laser flare photometry provides a novel way to monitor children with uveitis. Future study will be needed to determine whether values have prognostic importance and whether a treatment strategy that minimizes flare results in fewer uveitic complications. (Am J Ophthalmol 2003;135: 763–771. © 2003 by Elsevier Inc. All rights reserved.)
C
HRONIC UVEITIS IN CHILDREN HAS HISTORICALLY
been associated with a high rate of complications.1–3 Ocular factors predictive of these complications, and the best means to monitor children with uveitis is not well established. Laser flare photometry provides an objective and quantitative, noninvasive, in vivo measurement that correlates with aqueous humor protein levels in the anterior chamber of the eye. Its full role in the assessment of uveitis remains undefined, but a previous study involving a heterogeneous group of patients with uveitis showed associations between laser flare photometry values and the prevalence of some complications of uveitis, including posterior synechiae, cataracts, and macular edema.4 In contrast, there were no associations between these complications and the number of anterior chamber cells by clinical assessment. These observations suggest that laser flare photometry may be a useful supplement to the traditional strategy of monitoring cells during follow-up of patients with uveitis. We investigated the association between laser flare photometry values and the prevalence of uveitic complications, both at initial measurement and during follow-up, in a population of children with chronic uveitis.
METHODS LASER FLARE PHOTOMETRY MEASUREMENTS WERE AT-
tempted routinely for all children with uveitis as a part of standard clinical evaluations at the two participating academic centers. Laser flare photometry readings could not be obtained for all children, or at every visit for
ELSEVIER INC. ALL
RIGHTS RESERVED.
763
periocular injection of corticosteroid; or topical prednisolone acetate 1% at a frequency of more than four times daily for longer than 4 consecutive weeks at any time during follow-up. Conservative therapy was all other treatments, including observation. Athough the more intensive therapy group encompassed a broad range of treatments, these arbitrary categories reflect our treatment practices, and patients could be placed easily into one of the two groups on the basis of available information. A more specific evaluation of treatment was not possible because of the retrospective nature of the study. One of our major goals was to investigate the relationship between flare and complications in patients treated conservatively; this goal was appropriately served by these cateogories. Other definitions were consistent with those used in our previous studies of laser flare photometry.4,8 “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. Patients were identified as having “glaucoma/ocular hypertension” if they had a previous diagnosis of glaucoma/ocular hypertension; had undergone prior glaucoma/ocular hypertension surgery; were using intraocular pressure-lowering medications chronically; had intraocular pressure measurements greater than 24 mm Hg at any visit; or had visual field or optic disk changes typical of those seen with glaucomatous optic nerve damage; and had no history of corticosteroid-induced elevation of intraocular pressure. Patients were considered to have “cataracts” if records identified them as having 2⫹ or more nuclear or cortical opacity or any posterior subcapsular lens opacities. “Macular edema” was identified either clinically or by fluorescein angiography. “Hypotony” was defined as an intraocular pressure of less than 5 mm Hg. Patients were considered to have “vision loss” if best-corrected visual acuity dropped 2 or more Snellen lines on follow-up examination. Only one patient, who had light perception vision, was judged to have vision loss at baseline. For purposes of categorizing patients, “low flare” was defined as a laser flare photometry value of less than 20 photon units/msec, and “high flare” was defined as a laser flare photometry value of 20 photon units/msec or greater, in accordance with definitions of high and low flare in a previous publication.8 Laser flare photometry was performed by technicians using the Kowa FM-500 (Kowa Company, Ltd., Electronics and Optics Division, Tokyo, Japan). Technicians were unaware of the treatment or the presence or absence of complications (other than posterior synechiae that could be observed during measurement). Measurements were performed using standard procedures described by the Kowa Laser-Cell Flare Photometry Medical Advisory Board.4,8 Briefly, measurements were repeated until seven acceptable readings (difference between two background measurements less than 15%) had been taken for the study eye; the lowest and highest readings were deleted, and the machine then calculated the mean laser flare photometry
children included in this study, because of age, lack of cooperation, or other factors unrelated to their diagnoses. We retrospectively reviewed the records of all children for whom laser flare photometry values had been obtained at least once and who had active, noninfectious anterior uveitis, intermediate uveitis, or panuveitis. A child was defined as a patient whose age was 16 years or less at baseline examination. The first examination at which laser flare photometry was performed was selected as the baseline visit. Only one eye of each patient was studied; in cases with bilateral uveitis, the right eye was selected as the study eye, unless the right eye was not evaluable, in which case the left eye was used. Information about each patient was obtained from a review of medical records. The following demographic and historical information was obtained for each patient: age at baseline examination, sex, and past ocular surgery. The following information was obtained for each study eye at the baseline examination: best-corrected visual acuity, intraocular pressure, clinical assessment of anterior chamber cells, laser flare photometry value, and the presence or absence of six complications associated with uveitis: posterior synechiae, cataract, glaucoma or ocular hypertension, hypotony, macular edema, and vision loss from any cause. The inclusion of posterior synechiae in this list of complications is justified because they are a measure of severe disease and their presence predicts poorer outcomes.2,5 Clinical assessments had all been performed by one of two authors (J.L.D., G.N.H.). Anterior chamber cells were scored clinically by the method of Hogan, Kimura, and Thygeson.6 Uveitis was classified as anterior, intermediate, or pan-uveitis according to International Uveitis Study Group criteria.7 Cause, systemic disease associations, or ocular syndrome were identified for each case, if known. Patients with at least one follow-up clinical examination were evaluated for the development of complications after baseline. Records were reviewed for each clinical examination until development of a new complication was noted or until last follow-up clinical examination for patients who did not develop complications during follow-up. The following information was recorded for study eyes at each follow-up examination: duration since baseline examination, best-corrected visual acuity, intraocular pressure, clinical assessment of anterior chamber cells, laser flare photometry value, if performed, and the presence or absence of the six complications described above. Baseline and follow-up laser flare photometry values were compared if follow-up laser flare photometry was obtained within the 6 months before or after the development of a new complication or the 6 months before the last clinical examination of patients without new complications. For purposes of this study, “more intensive therapy” was defined as the use of oral prednisone, immunosuppressive drugs, immunomodulatory biologics, or a combination of these agents for longer than 6 consecutive weeks; a 764
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
JUNE 2003
value and standard deviation of the remaining five readings. The mean value for each patient was used for comparisons. Photometry readings are expressed as photon units / msec. This study was approved by the Institutional Review Boards at the University of Miami and the University of California, Los Angeles. There is a linear relationship between laser flare photometer readings and total protein content in aqueous humor.9 For statistical analysis, either the laser flare reading or its logarithm was used. The raw data expressed as photon units/msec were considered inappropriate for some statistical analysis because assumptions of normality and variance homogeneity were not met. Transformation of data is recommended in these circumstances10; a logarithmic conversion suited the laser flare photometry data best, and was used in analyses. The primary outcome was defined as prevalence of complications for patients grouped by different levels of flare. The secondary outcome was defined as the development of new or additional complications during follow-up in patients with different levels of flare who were receiving conservative or more intensive therapy. The associations between prevalence of complications at baseline and patient and ocular characteristics were evaluated with the Fisher exact test, the 2 test, or logistic regression as appropriate. Kaplan-Meier survival analysis with the log-rank test and Cox proportional hazards survival regression were used to study incidence of new complications in relation to baseline characteristics, treatment, and follow-up laser flare photometry values. The use of some additional analyses are documented in the tables. A P value of less than .05 was considered to be statistically significant.
RESULTS WE EVALUATED 59 CHILDREN WITH UVEITIS. DEMOGRAPHIC
and ophthalmic characteristics are shown in Table 1. Of 41 children with a diagnosis of anterior uveitis, 17 (29% of all patients) had arthritis (juvenile rheumatoid arthritis [JRA] or juvenile idiopathic arthritis [JIA]). Bilateral uveitis was present in 52 children (88%). In one patient with bilateral uveitis attributed to sympathetic ophthalmia, laser flare photometry values from the left (sympathizing) eye were used, as the right (inciting) eye was not evaluable because of posttraumatic changes. A mild anterior chamber cellular reaction (ⱕ 1⫹ cells) was present in 59% of patients. The median laser flare photometry value for all patients (23.0 photon units/msec [range, 0.8 – 675 photon units/msec]) fell within our definition of high flare. Overall, 58% of eyes had at least one complication at baseline: cataracts (37%), posterior synechiae (37%) glaucoma/ocular hypertension (12%), and macular edema (5%). VOL. 135, NO. 6
LASER FLARE PHOTOMETRY
AND
There were several differences between the baseline characteristics of patients at the two centers (data not shown), but the ranges for all factors from both centers overlapped. At site 1, there were more boys (P ⫽ .072) and a higher proportion of patients with a diagnosis of JRA (P ⫽ .071, comparing prevalence of all diagnoses). Several other differences were consistent with a higher proportion of patients having severe disease at site 1; patients at site 1 had higher median laser flare photometry values (P ⫽ .03), lower visual acuities (P ⫽ .004), and a higher prevalence of complications (P ⫽ .05). In contrast, there was a higher proportion of patients at site 2 who had been assigned baseline clinical scores of 2⫹ or 3⫹ for anterior chamber cells (P ⫽ .002, 2 test). There was a statistically significant positive correlation between log laser flare photometry value and the presence of complications at baseline (P ⫽ .001, Table 2). When the large number of eyes with posterior synechiae were excluded, the relationship remained, although it was less strong (P ⫽ .017). Because the primary site of inflammation may influence rate of complications, we also investigated the association between flare and the prevalence of complications for the subgroup of patients with anterior uveitis only (n ⫽ 41). Results were similar to those for all patients; those with anterior uveitis and high flare had a significantly higher prevalence of any complication (P ⫽ .031) and of posterior synechiae (P ⫽ .008). Anterior chamber cell counts did not correlate with baseline complications (P ⫽ .54, exact 2 test, Table 2). Because there were generally higher clinical scores for baseline cells at site 2 than at site 1, we compared clinical scores for cells to prevalence of complications for each site individually, and no statistical associations were identified. Table 3 provides follow-up data that were available for 45 patients (mean duration of follow-up 15 ⫾ 14 months, median 11, range, 1– 64). Of these patients, 33 had laser flare photometry values at follow-up. We found no difference in the proportion of patients who had follow-up examinations between those with complications at baseline (28 of 35 children [80.0%]) and those without complications at baseline (17 of 24 children [70.8%], P ⫽ .54, Fisher exact test). Also, there was no difference in the proportion of patients who had follow-up examinations between those with low flare at baseline (20 of 26 children [76.9%]) and those with high flare at baseline (25 of 33 children [75.8%], P ⫽ 1.00, Fisher exact test). The frequencies of individual complications during follow-up are shown in Table 3 as descriptive data. The cumulative percentage of eyes with new complications rose over the period of follow-up (Table 4). Stated rates for individual complications do not account for patients lost to follow-up; thus, true rates may be higher or lower than shown in Table 4. Cataract and glaucoma/ocular hypertension made up the majority of new complications. There were significantly more new complications in eyes with complications at baseline than in those without complicaCHRONIC UVEITIS COMPLICATIONS
765
TABLE 1. Patient Demographic and Ocular Characteristics at Initial Examination
Number of patients Sex Female (patients [percentage]) Mean age (years ⫾ SD) Category of uveitis (patients [percentage]) Anterior uveitis Idiopathic iridocyclitis JRA-associated Intermediate uveitis Panuveitis Laterality (patients [percentage]) Bilateral Median visual acuity (range) Laser flare photometry value Median (photon units/msec, [range]) Mean log value ⫾ SD Anterior chamber cells ⬍1⫹ 1⫹ 2⫹ 3⫹ Complications (eyes) Any Posterior synechiae Cataract Glaucoma/ocular hypertension Hypotony Macular edema Vision loss
All Patients
Patients With Low Flare*
Patients With High Flare†
59
26
33
41 (70%) 10.3 ⫾ 3.5
19 (73%) 10.7 ⫾ 3.6
22 (67%) 9.9 ⫾ 3.4
41 (70%) 24 (41%) 17 (29%) 9 (15%) 9 (15%)
17 (66%) 10 (38%) 7 (27%) 3 (12%) 6 (23%)
24 (72%) 14 (42%) 10 (30%) 6 (18%) 3 (9%)
52 (88%) 20/20 (20/15–HM)
24 (92%) 20/20 (20/15–6/200)
28 (85%) 20/30 (20/15–HM)
23.0 (0.8–675) 1.43 ⫾ 0.67
7.0 (0.8–18.8) 0.82 ⫾ 0.30
70.9 (21.7–675) 1.90 ⫾ 0.46
11 (19%) 24 (41%) 13 (22%) 11 (19%)
3 (12%) 14 (44%) 6 (23%) 3 (12%)
8 (24%) 10 (30%) 7 (21%) 8 (24%)
35 (59%) 22 (37%) 22 (37%) 7 (12%) 0 (0%) 3 (5%) 1 (2%)
10 (38%) 4 (15%) 6 (23%) 4 (15%) 0 0 0
25 (76%) 18 (55%) 16 (49%) 3 (9%) 0 3 (9%) 1 (3%)
P Value‡
0.78§ 0.39¶ 0.51㛳
0.45§ 0.003** ⬍0.001* ⬍0.001¶ 0.94**
0.007§ 0.003§ 0.060§ 0.69§ 0.25§ 1.00§
HM ⫽ hand movement; JRA ⫽ juvenile rheumatoid arthritis; SD ⫽ standard deviation. *Low flare defined as laser flare photometry values ⬍20 photon units/msec. † High flare defined as laser flare photometry values ⱖ20 photon units/msec. ‡ Comparison of patients with low flare versus those with high flare. § Fisher exact test. ¶ Two sample t test. 㛳 Exact 2. **Mann-Whitney test.
tions at baseline (P ⫽ .018, Kaplan-Meier method with log-rank test, Table 4). Only one of 17 eyes without complications at baseline developed a new complication during follow-up; the patient was a 7-year-old girl with bilateral idiopathic iridocyclitis who had high flare at baseline and developed glaucoma/ocular hypertension at 4 months of follow-up. Laser flare photometry values at baseline were statistically associated with the development of new complications during follow-up, but the association was not strong: a 2-log larger baseline laser flare photometry value was associated with a 1.35 higher risk of new complications (P ⫽ .047, Cox regression). Baseline complications and baseline flare were not independent predictors of new complications, however, based on a Cox proportional hazards 766
AMERICAN JOURNAL
survival regression. Baseline laser flare photometry values in patients with anterior uveitis were weakly associated with development of new complications at follow-up (P ⫽ .048). A relationship between laser flare photometry values at follow-up and new complications could not be identified after accounting for complications at baseline. Also, there was no statistical difference between incidence of new complications during follow-up and the intensity of treatment (P ⫽ .59, Kaplan-Meier method with log-rank test). For the 33 eyes with both baseline and follow-up laser flare photometry values, mean values decreased during follow-up; mean baseline values were 76.3 ⫾ 145.9 photon units/msec and mean last follow-up values were 51.0 ⫾ 76.0 photon units/msec, a drop of ⫺25.4 ⫾ 95.2 photon units/msec from baseline to last follow-up. This result was OF
OPHTHALMOLOGY
JUNE 2003
P Value
complications in patients with low flare cannot be attributed to inadequate follow-up. Likewise, the single, new complication that occurred during follow-up was seen at 4 months in a patient with high flare, whereas those without complications at baseline who had low flare (none of whom had new complications) were followed up for a mean of 21.3 months (range, 3– 64).
4/18 (22%) 16/26 (62%) 15/15 (100%)
0.001†
DISCUSSION
8/11 (73%) 12/24 (50%) 9/13 (69%) 6/11 (55%)
0.54§
TABLE 2. Prevalence of Complications at Baseline Examination in Patients Grouped by Initial Laser Flare Photometry Value and by Initial Clinical Anterior Chamber Cell Count Prevalence of Complications*
Patient Groups
By laser flare photometry value (photon units/msec) ⬍10 10–100 ⬎100 By clinical cell count‡ ⬍1⫹ 1⫹ 2⫹ 3⫹
IN THIS RETROSPECTIVE STUDY, THERE WAS AN ASSOCIA-
*Number of patients with any complication at baseline examination/number of patients with specified laser flare photometry values or clinical cell count. † Significance analyzed with logistic regression in which log laser flare photometry value was evaluated as a continuous independent variable and presence of any complication was used as the dependent variable. ‡ As defined by Hogan and associates.6 § Exact 2 test.
influenced by some individual large decreases. Change was correlated with baseline laser flare photometry values; greater baseline values tended to decrease more during follow-up (P ⫽ .007; Figure 1) In contrast, the median change was an increase of 0.5 photon units/ms (range ⫺359.2 to ⫹ 169.1 photon units/msec). Change was not correlated with follow-up duration. Among the points shown in Figure 1, the largest increase in flare during follow-up was in a patient treated conservatively, whereas the largest drop in flare occurred in a patient treated more intensively. Overall, however, we could not identify a statistical association between change in flare and intensity of treatment. Table 5 shows patients grouped by baseline flare (low vs high) and by medical therapy (conservative vs more intensive). Among 25 children with high flare at baseline, complications developed in 11 (44%) during follow-up (four of seven [57%] who were treated conservatively and seven of 18 [39%] who were treated more intensively; P ⫽ .66, Fisher exact test). In contrast, complications developed in only two (10%) of 20 children with low flare at baseline (P ⫽ .02 when compared with all patients with high flare at baseline, Fisher exact test); both were treated intensively. The duration of follow-up was longer for those patients with low flare in whom complications did not develop (median, 13 months; range, 1– 64) than for those with high flare in whom complications did develop (median, 8 months; range, 3–32); thus, the lower rate of VOL. 135, NO. 6
LASER FLARE PHOTOMETRY
AND
tion between higher laser flare photometry values and the prevalence of complications of uveitis at baseline, similar to the findings of Gonzales and associates.4 In contrast to that study, the current investigation contained a more homogeneous group that included just children with uveitis. A cause and effect relationship between flare and complications was not established, however. Uveitis is known to alter the blood-aqueous barrier in the anterior segment, changing the protein composition of the aqueous humor.11 Elevated intraocular protein may be a cause of complications, or it may simply be a marker for other pathologic processes in the uveitic eye that lead to complications.8,12 Our results are consistent with the high rates of complications that have been reported in children with uveitis by others.1–3 Because not every patient had follow-up data, it is possible that the true rates of complications are less than those we report, as it is more likely that patients with complications would return for follow-up than would those without complications. Nevertheless, follow-up did not differ statistically between children with and without complications at baseline or between children grouped on the basis of low versus high flare at baseline. Previous studies have shown that the presence of posterior synechiae at initial examination is predictive of poorer outcomes.2,5 In this study, both higher baseline laser flare photometry values and presence of complications at baseline were associated with the subsequent development of new complications on univariate analyses. Furthermore, patients with baseline complications continued to develop additional complications during follow-up, even with more intensive therapy. Flare was not identified as an independent predictor of complications during follow-up, but intervals were relatively brief and patients were receiving medical treatment. In addition, when complications are preexisting, new events of the same type cannot occur. Because of the strong association between flare and baseline complications, it is possible that flare, measured early in the course of disease before development of complications, would predict their development. A longitudinal study with consistent follow-up would be necessary to demonstrate that. The intensity of treatment was not independently associated with the development of new complications. Small patient numbers and selection of CHRONIC UVEITIS COMPLICATIONS
767
TABLE 3. Data From Follow-up Examinations
Number of eyes with follow-up data Duration of follow-up Mean (months ⫾ SD) Median (months [range]) Treatment during follow-up interval Conservative therapy㛳 More intensive therapy** Log laser flare photometry value†† Number of eyes evaluated Mean increase (⫾ SD) Median visual acuity at last follow-up examination (range) Eyes with new complications Any complication Posterior synechiae Cataracts Glaucoma/ocular hypertension Hypotony Macular edema Vision loss
All Patients
Patients With Low Flare*
Patients With High Flare†
45
20
25
15 ⫾ 14 11 [1,64]
17 ⫾ 17 13 [1,64]
13 ⫾ 10 8 [1,32]
P Value‡
0.52§ 0.35¶
16 (36%) 29 (64%)
9 (45%) 11 (55%)
7 (28%) 18 (72%)
33 0.03 ⫾ 0.28‡‡ 20/20 [20/15–CF]¶¶
16 0.00 ⫾ .18 20/20 [20/15–6/200]
17 0.06 ⫾ 0.35 20/30 [20/20–CF]
13 0 5 5 0 1 2
2 0 1 0 0 1 0
11 0 4 5 0 0 2
0.54§§ 0.003§
CF ⫽ counting fingers; SD ⫽ standard deviation. *Low flare defined as laser flare photometry values ⬍20 photon units/msec. † High flare defined as laser flare photometry values ⱖ20 photon units/msec. ‡ Comparison of patients with low flare versus those with high flare. § Mann-Whitney test. ¶ Fisher exact test. 㛳 Conservative therapy was all treatments other than aggressive therapy, as defined below, including observation. **More intensive therapy was defined as the use of oral prednisone, immunosuppressive drugs, immunomodulatory biologics, or a combination of these agents for longer than 6 consecutive weeks; a periocular injection of corticosteroid; or topical prednisolone acetate 1% at a frequency of more than 4 times daily for longer than 4 consecutive weeks. †† Measurement closest to last examination, if made within 6 months prior to examination. ‡‡ P ⫽ .48, paired t test; when compared with baseline value. §§ Two sample t test. ¶¶ P ⫽ .52, Wilcoxon paired rank test, when compared with baseline value.
therapy based on clinical perception of uveitic activity (based in part on flare levels) may have obscured these relationships. Among patients with high flare at baseline who were treated conservatively, those in whom complications did not develop were followed up for shorter durations (median, 3 months; range, 2–11) than those in whom complications did develop (median, 7.5 months; range, 3–28); thus, it is possible that patients with high flare are at increased risk for eventual complications if they are treated conservatively. Consensus is developing among uveitis specialists that intensive treatment will improve outcomes for children with uveitis. Our numbers were too small to show that complications are reduced by more intensive therapy, but our study does show that intensive therapy alone does not 768
AMERICAN JOURNAL
prevent complications in all cases, especially in children with preexisting complications. In view of the potential adverse effects of systemic immunosuppressive therapy, it is important to identify factors that predict an increased risk of uveitis-associated complications, and thus a need for aggressive therapy. There may be a subgroup of patients with chronic uveitis and low flare in whom complications do not develop and who retain excellent vision over many years despite a persistent anterior chamber cellular reaction, even in the absence of intensive therapy. Of concern is whether the use of immunosuppressive drugs in such children poses a greater risk than the persistent low-grade cellular reaction. Even with intensive treatment, some children with low flare in this series did develop complications, indicating OF
OPHTHALMOLOGY
JUNE 2003
TABLE 4. Incidence of New Complications in Patients Grouped by Interval to Last Examination Prevalence of Complications* Follow-up Interval
12 months 24 months 36 months§
Patients Grouped by Baseline Flare† All Patients
22% (⫾7%) 27% (⫾8%) 66% (⫾16%)
Patients Grouped by Baseline Complications‡
Low
High
Absent
Present
6% (⫾6%) 6% (⫾6%) 30% (⫾21%)
34% (⫾10%) 43% (⫾12%) 100%
7% (⫾7%) 7% (⫾7%) 7% (⫾7%)
30% (⫾9%) 38% (⫾11%) 100%
*Cumulative percentages (⫾ standard errors). † The difference in cumulative proportion without new complications during follow-up is significantly different between eyes with low flare (⬍20 photon units/msec) and high flare (ⱖ 20 photon units/msec) at baseline (P ⫽ .021, Kaplan-Meier method and log rank test). ‡ The difference in cumulative proportion without new complications during follow-up is significantly different between eyes with complications at baseline and those without complications at baseline (P ⫽ .018, Kaplan-Meier method and log rank test). § Few patients were followed up longer than 30 months, limiting the reliability of findings.
FIGURE 1. This scattergram includes 33 patients who had both baseline and follow-up laser flare photometry measurements. Points above the diagonal indicate patients whose laser flare photometry values increased, whereas those below the line indicate patients whose values decreased. The decrease in the overall mean is due to the influence of a few points with high baseline values.
that children with uveitis require close follow-up, regardless of findings or treatment. Because intensive corticosteroid treatment itself can cause cataract and ocular hypertension, additional long-term follow-up of many paVOL. 135, NO. 6
LASER FLARE PHOTOMETRY
AND
tients will be required to determine the management strategy that provides the greatest benefits with the least risk. Decreases in laser flare photometry values from baseline to final follow-up occurred in several patients, most notaCHRONIC UVEITIS COMPLICATIONS
769
770 TABLE 5. Development of Complications During Follow-up in Patients Grouped by Baseline Laser Flare Photometry Value and Intensity of Treatment All Patients
Conservative
AMERICAN JOURNAL OF
OPHTHALMOLOGY
Findings at baseline Number of eyes Laser flare photometry value (photon units/msec) Mean ⫾ SD Range Findings at follow-up New complications Number of patients Duration of follow-up (months) Median Range Mean laser flare photometry value (photon units/ msec ⫾ SD) at follow-up (number of patients evaluated) No complications at follow-up Number of patients Duration of follow-up (months) Median Range Mean laser flare photometry value (photon units/msec ⫾ SD) at follow-up (number of patients evaluated)
‡
Patients Without Complications at Baseline High Flare†
Low Flare* More Intensive
9
11
8.0 ⫾ 5.2 3.6–18.8
9.0 ⫾ 4.8 0.8–16.5
0 —
§
Conservative
‡
7
High Flare†
Low Flare*
More Intensive
§
Conservative
‡
More Intensive
§
Conservative‡
More Intensive§
18
7
5
1
4
131.6 ⫾ 136.2 28.3–360.6
153.5 ⫾ 174.1 21.7–675.4
6.9 ⫾ 5.4 3.6–18.8
8.0 ⫾ 3.4 3.6–13.1
32.5 32.5
34.9 ⫾ 24.0 22.6–70.9
2 (18%)
4 (57%)
7 (39%)
0
0
0
1 (25%)
7.5 3–28 116.0 ⫾ 89.9 (4)
8 3–32 133.9 ⫾ 147.0 (4)
—
—
—
—
16 6–26 9.9 (1)
—
—
—
4 — 73.8 (1)
9 (100%)
9 (82%)
3 (43%)
11 (61%)
7 (100%)
5 (100%)
1 (100%)
3 (75%)
13 3–64 6.8 ⫾ 2.4 (7)
13 1–26 9.2 ⫾ 4.3 (8)
3 2–11 141.5 (1)
13 1–30 51.3 ⫾ 44.5 (8)
22 3–64 6.5 ⫾ 2.5 (6)
13 9–26 8.3 ⫾ 4.6 (5)
3 — (0)
6 3–14 26.3 ⫾ 11.8 (3)
*Low flare defined as laser flare photometry values ⬍ 20 photon units/ms. High flare defined as laser flare photometry values ⱖ 20 photon units/ms. ‡ Conservative therapy was all treatments other than aggressive therapy, as defined below, including observation. § More intensive therapy was defined as the use of oral prednisone, immunosuppressive drugs, immunomodulatory agents, or a combination of these agents for longer than 6 consecutive weeks; a periocular injection of corticosteroid; or topical prednisolone acetate 1% at a frequency of more than 4 times daily for longer than 4 consecutive weeks. †
JUNE 2003
bly those with higher baseline flares. Contrary to the assumptions of some clinicians, it may be possible to reduce aqueous humor protein concentrations with treatment in some patients. Current concepts regarding the persistence of elevated flare in patients with uveitis may relate to the inability to appreciate small changes in flare clinically; small differences in aqueous humor protein content can be identified with laser flare photometry, however.4 It is unclear whether reduction of flare would alter the course of uveitis or reduce the risk of complications, but it is reasonable to speculate that therapy to minimize flare may be an effective strategy to improve outcomes for pediatric patients with uveitis. Clinical scores for anterior chamber cells were not associated with either complications at baseline or with new or additional complications during follow-up. This observation may reflect a true lack of association between the number of anterior chamber cells and complications, or it may reflect an inability of clinicians to quantitate cells with sufficient precision to identify a true relationship. In either case, clinical cell counts alone would appear not to be sufficient to assess the severity of uveitis. Precise measurement of flare provides an alternative means of assessment. Our results support the use of laser flare photometry to monitor uveitis in children. Although our small numbers and lack of controls limit our ability to draw firm conclusions about the observed associations, we believe they are sufficiently convincing to support continued investigation of this technique as a method for monitoring children early in the course of uveitis and potentially for predicting their risk of complications so that therapy of the proper intensity can be prescribed. ACKNOWLEDGMENTS
Kowa Company, Ltd, Electronics and Optics Division, Tokyo, Japan, provided Dr Holland with an FM-500 Flare Meter for use in research activities during a portion of the period in which data used in this article were being
VOL. 135, NO. 6
LASER FLARE PHOTOMETRY
AND
collected. Doctor Holland was an unpaid member of the Kowa Laser Cell-Flare Photometry Medical Advisory Board from 1991 to 1995. The authors have no other interest in the products or techniques described in this report or in competing techniques.
REFERENCES 1. Edelsten C, Reddy MA, Stanford MR, Graham EM. Visual loss associated with pediatric uveitis in English primary and referral centers. Am J Ophthalmol 2003;135:757–762. 2. Wolf MD, Lichter PR, Ragsdale CG. Prognostic factors in the uveitis of juvenile rheumatoid arthritis. Ophthalmology 1987;94:1242–1248. 3. Tugal-Tutkun I, Havrlikova K, Power WJ, Foster CS. Changing patterns in uveitis of childhood. Ophthalmology 1996;103:375–383. 4. Gonzales CA, Ladas JG, Davis JL, et al. Relationships between laser flare photometry values and complications of uveitis. Arch Ophthalmol 2001;119:1763–1769. 5. Edelsten C, Lee V, Bentley CR, et al. An evaluation of baseline risk factors predicting severity in juvenile idiopathic arthritis associated uveitis and other chronic anterior uveitis in early childhood. Br J Ophthalmol 2002;86:51–56. 6. Hogan MJ, Kimura SJ. Signs and symptoms of uveitis, I. Anterior uveitis. Am J Ophthalmol 1959;47:155–170. 7. Bloch-Michel E, Nussenblatt RB. International Uveitis Study Group recommendations for the evaluation of intraocular inflammatory disease. Am J Ophthalmol 1987;103:234 – 235. 8. Ladas JG, Yu F, Loo R, et al. Relationship between aqueous humor protein level and outflow facility in patients with uveitis. Invest Ophthalmol Vis Sci 2001;42:2584 –2588. 9. Shah SM, Spalton DJ, Taylor JC. Correlations between laser flare measurements and anterior chamber protein concentrations. Invest Ophthalmol Vis Sci 1992;33:2878 –2884. 10. Fleiss JL. The design and analysis of clinical experiments. New York, Wiley & Sons, 1986, 59 –68. 11. Zirm M. Proteins in aqueous humor. Adv Ophthal 1980;40: 100 –172. 12. Nguyen QD, Humphrey RL, Dunn JP, Humayun MS. Elevated vitreous concentration of monoclonal immunoglobulin manifesting as schlieren in juvenile rheumatoid arthritisassociated uveitis. Arch Ophthalmol 2001;119:293–296.
CHRONIC UVEITIS COMPLICATIONS
771