Table 1.
Qualitative analysis of systematic review for intra-arterial thrombolysis treatment for central retinal artery occlusion.
| Type of study | No. of IAT subjects (total subjects) | Average time to thrombolysis in hours | Agent and dose | Study design features and major study limitations (bias) | Results | |
|---|---|---|---|---|---|---|
| Mercier et al. (14) | Retrospective Case Series | 14 (14) | 8.0 | t-PA | No control group, not included in quantitative meta analysis | 6/14 (42.9%) had significant visual improvement correlating to improvement of ≥0.3 on the log MAR scale |
| Mean: 35 ± 13 mg | ||||||
| Ahn et al. (15) | Retrospective cohort study with control group | 57 (101) | 22.7 | Urokinase | All patients (IAT and controls) met the same predetermined inclusion/exclusion criteria and were treated concurrently, which limited selection bias | Overall, 24/57 (42.1%) in the IAT group showed significant final visual improvement (improved log MAR ≥ 0.3) vs. 11/44 (25%) in the ST group (P = 0.09b; Ahn et al., Table 3)c |
| Predetermined subgroup analysis for severity of CRAO (CRAO stages) | Patients with final visual outcome 20/200 (legally blind threshold) or better: 11/57 (19.3%) in IAT group vs. 2/44 (4.5%) in ST group (P = 0.026) | |||||
| Time from symptom onset to treatment significantly different between groups representing a potential selection bias that may favor IAT therapy | Of incomplete stage CRAO patients, 11/13 (84.6%) in the IAT group showed significant final visual improvement (improved log MAR ≥ 0.3) vs. 3/13 (23.1%) in the ST group (P = 0.002) | |||||
| Maximum: 500 | The incomplete CRAO subgroup did not have significantly different time to treatment differences resulting in a relative decrease in selection bias within this subgroup | |||||
| Schumacher et al. (16) | Prospective randomized controlled trial | 42 (82) | 12.8 | Maximum: 500 | Prospective randomized controlled trial significantly reduces risk of selection bias. Planned enrollment 200 subjects (100 in each treatment group) Inclusion criteria included all non-arteritic CRAO patients regardless of CRAO stage (incomplete, subtotal, total) resulting in a heterogenous patient population. Relative number of incomplete stage CRAO is not known between groups and, therefore, it is possible there were different relative severity levels of CRAO between groups. This possibility could have influenced results. Randomization should reduce this risk, but performing stratified randomization with CRAO stage as one of the stratum may have reduced this allocation risk |
A prespecified interim efficacy analysis of the available data on the first 70 enrolled subjects was performed. The probability to detect a significant difference between the two groups upon study completion (200 subjects enrolled) was calculated based on the data of the 70 patients while assuming that the remaining 130 patients would resemble the initial 70 patients. They calculated an 8.1% probability of a statistical difference in favor of the ST group and a 0.1% probability of a statistical difference in favor of IAT therapy. As a result, the DSMB recommended halting enrollment early. A total of 84 patients had been enrolled at this point with data available on 82 patients for final intention to treat analysis |
| Overall, 24/42 (57.1%) in the IAT group showed significant visual improvement at 1 month (improved log MAR ≥ 0.3) vs. 24/40 (60%) in the ST group (P = 0.83)c | ||||||
| Time to treatment was on average 2 h longer in the IAT group compared to the ST group, which could have resulted in a treatment bias unfavorable to the IAT group | Mean log MAR improvement of the IAT group was 0.447 (SD ± 0.545) and of the ST group was 0.443 (SD ± 0.549) | |||||
| Zhang et al. (17) | Retrospective case series | 49 (49) | <6 | Urokinase | No control group, not included in quantitative meta analysis | At 6 months, 18/49 (36.7%) patients had regained VA to > 0.6 |
| 35/49 (71%) had recanalization after IAT. Those with recanalization received greater recovery of visual acuity with a VA of 0.6 ± 0.04 in recanalized patients compared to 0.002 ± 0.0012 in non-recanalized patients (P < 0.05) | ||||||
| 200,000 IU to 1 million IU | Retrospective analysis of 49 consecutive patients treated by IAT within 6 h of presentation | 16/21 (76.2%) had significant visual improvement within IAT group vs. 7/21 (33.3%) within ST group (P = 0.018) as defined by one line improvement on Snellen chart or one VA category improvement when VA is worse than 20/400c | ||||
| Mean: 626,000 | ||||||
| Aldrich et al. (18) | Retrospective cohort study with control group | 21 (42) | 9.3 | t-PA | Consecutive CRAO patients enrolled but no consistent preddetermined inclusion/exclusion criteria, thereby representing risk for selection bias | |
| Maximum: 20 mg | Time from symptom onset to treatment significantly different between groups representing a potential selection bias that may favor IAT therapy | 7/21 (33.3%) had significant visual improvement within IAT group vs. 1/21 (4.8%) within ST group (P = 0018) as defined by 3-lines or more improvement on Snellen chart | ||||
| Multiple 3 mg aliquots delivered until VA improvement | Initial VA better for IAT group vs. ST group (Aldrich et al., Table 2) though not statistically different (P = 0.31) may still present a bias toward better outcomes in IAT group | |||||
| Criteria for “significant visual improvement” in primary outcome easier to achieve compared to other studies (one line improvement on Snellen chart or one category improvement when VA worse than 20/400), although a post hoc analysis with a more restrictive 3-line improvement was also included | ||||||
| Pettersen et al. (19) | Retrospective case series | 8 (8) | 9.7 | t-PA | No control group, not included in quantitative meta analysis | 8 cases of CRAO treated with IAT were identified but only 6 had VA follow-up in clinic. Of the two who did not have clinic follow-up, neither had improved 24 h after IAT. Therefore, 6/8 patients had documented improvement after IAT |
| Range: 10—30 mg | 6/6 patients with VA follow-up in clinic at least 1 month after IAT had VA improvement. 3/6 by 1 Snellen line and 3/6 improved by 2 or more Snellen lines | |||||
| Arnold et al. (20) | Retrospective cohort study with control group | 37 (56) | 3.7 | Urokinase | No significant difference in time from symptom onset to presentation (treatment) between the two groups. All presented in less than 6 hours as part of inclusion criteria minimizing selection bias | 8/37 improved to a clinically significant ≥ 0.6 VA in the IAT group vs. 0/19 in ST group (P = 0.04)c |
| Mean: 677,000 U | 24 of 37 patients in the IAT group received anterior chamber paracentesis as one of the standard therapies while just 4 of 19 received this modality in the ST group (P = 0.004). This represents the only identified significant difference between groups (baseline characteristic or treatment), but does represent a potential source of treatment bias potentially favoring the IAT group | 28/37 had any amount of log MAR improvement in the IAT group vs. 10/19 in the ST group (P = 0.13; derived from Arnold et al., Figures 1 and 2) | ||||
| Butz et al. (21) | Retrospective case series | 22 (22) | 7.6 | Urokinase | No control group, not included in quantitative meta analysis | 7/22 marked improvement in VA, 2/22 slight improvement |
| Mean: 642,000 SD 300 K | ||||||
| t-PA Mean: 27 ± 8 mg | ||||||
| Schmidt et al. (22) | Retrospective cohort study with control group | 62 (178) | 10.8 | Urokinase: 1 ml/min | No significant difference in time from symptom onset to treatment between the two groups (when comparing median times; P = 0.5), minimizing the potential for time to treatment selection bias | 36/62 (58%) of patients in the IAT group demonstrated partial or distinct visual improvement vs. 34/116 (29.3%) in the ST group (P < 0.001)c |
| 200,000–1.3 million IU; Median: 1 million IU | Severity of CRAO Stage was statistically similar between groups. 10/62 (16.1%) patients were incomplete (best stage) vs. 29/116 (25%) in the ST group (P = 0.19). This would tend to favor visual improvement in the ST group and, therefore, this difference would not represent any bias in favor of the IAT group | In subgroup analysis (not pre-specified), only subtotal stage CRAO demonstrated significant visual improvement with IAT therapy. Partial or distinct improvement was seen in 24/47 (51%) in the IAT group vs. 15/83 (18.1%) in the ST group (P < 0.001) | ||||
| t-PA: 40–80 mg | Specific reasons why a given patient was treated with IAT vs. ST only were not given other than patients were allocated to the ST group when they had “contraindications” to IAT. This lack of specificity in the methods section raises a concern for selection bias | |||||
| Median: 50 mg | ||||||
| IAT stopped once VA improvement identified | ||||||
| Richard et al. (23) | Retrospective case series | 53 (53) | 14.0 | t-PA | No control group, not included in quantitative meta analysis | 35/53 improved at 3 months, no correlation with time |
| Maximum: 40 mg | ||||||
| Weber et al. (24) | Retrospective cohort study with control group | 17 (32) | 4.2 | Urokinase | Patients had similar initial VA between groups | 11/17 (64.7%) of patients had visual improvement in the IAT group compared to 5/15 (33.3%) in the ST group (P = 0.16)c |
| 100,000–900,000 IU | 6 of 17 patients (35.3%) in the IAT group received anterior chamber paracentesis as one of the standard therapies while just 1 of 15 (6.7%) received this modality in the ST group (P = 0.09). While not a statistically significant difference between groups, this could represent a potential source of treatment bias potentially favoring the IAT group | |||||
| Mean: 594,000 | There was no predetermined specific inclusion/exclusion criteria and not all patients were treated concurrently, thereby introducing the risk of selection bias and treatment bias between groups | 3/17 total recovery, 2/17 marked improvement, 6/17 slight improvement vs. 5/15 minimal improvement in ST | ||||
| Patients presented with similar times from symptom onset to treatment reducing time to treatment as a potential selection bias | ||||||
| Schumacher and Schmidt (25) | Retrospective case series | 35 (35) | 4–2.5 days | Urokinase | No control group, not included in quantitative meta analysis | 23/35 (66%) were either complete, marked, or definite but less marked improvement in VA. 10/35 (29%) total or marked improvement, 13/35 (37%) definite improvement |
| Maximum: 1.2 million Units | ||||||
| t-PA | Continuation of prior studies (26–28)a | |||||
| Maximum: 70 mg | ||||||
IAT, intra-arterial thrombolysis; t-PA, tissue plasminogen activator; LogMAR, logarithim of the minimum angle of resolution; CRAO, central retinal artery occlusion; ST, standard therapy; DSMB, data safety monitoring board; VA, visual acuity; IU, international units.
aContinuation of prior studies.
bCorrected P-value which differs from the study reported inaccurate P-value.
cData used for pooled meta-analysis.