Methodological Quality
Of the six domains from AGREE II, the lowest score was in the applicability domain with a mean of 18.8% while clarity of presentation domain had the highest score with a mean of 62%. The lowest score was in the applicability domain (EURETINA guideline) and the highest was in two domains: scope and purpose (RCO), and clarity of presentation (SERV).
None of the CPG assessed reached the minimum suggested for a high-quality CPG (at least 3/6 domains > 60% including domain 3) or moderate quality CPG (≥ 3 domains score > 60%, except Domain 3)(19). Even when RCO CPG included a hematologist and PPP CPG scored the better methodological quality items (Domain 3: Rigour of development = 40.4%), all the CPG are considered “Low quality CPG” (≥ 2 domain score < 60% and domain 3 score < 50%)(19). Previous studies have assessed and gave some suggestion to improve PPP quality: “The inclusion of a diverse representation of clinical, scientific, and methodological experts, record the inclusion of patients or patient representatives in CPG development, and ensure that patients, patient organizations, and interested members of the public have an opportunity to review the CPG and describe how their comments were addressed” (27).
In item-9(Strengths/limitations of the body of evidence are clearly described), most of the guidelines had lower scores, that could be due to lack of high-quality evidence in ophthalmology, primarily in systematics reviews(28–30), however retina is the sub-specialty with more published systematic reviews, followed by cornea, glaucoma and cataract surgery. (28) Also, constant development of new and several types of anti-angiogenics can be a reason.
Item 10 (methods for formulating the recommendations) is the most important about clinical relevance, however only PPP and SERV reported a method for evidence and grading of recommendations (SIGN/GRADE for PPP and US Agency for Health Research and Quality for SERV). CEC reports its own level of evidence. In recent years, most of the developer organizations are including the GRADE framework(31). Not including how evidence was assessed nor how the recommendation was formulated could be a problem with consensus-based recommendations. However, even in this case, a complete description of the percent of agreement between members, will give transparency to the CPG(32). Transparency in a CPG is a very important matter, and includes several issues contained in AGREE assessment like: a complete description of the questions, retrieved evidence, who assessed the evidence (including conflict of interest), benefit/harms balance, judgments for or against recommending a specific intervention, and who peer-reviewed the draft guidance(33).
Including the date of update is a very important issue assessing CPGs, especially considering that recommendations become outdated after 3 years(34). RCO and PPP guidelines detailed the update date (2025 and 2024 respectively). Waiting more than 3 years to review a CPG is not recommended(34), so an efficient way is to implement systematic updating, with the possibility of performing partial updating(35). A special committee/working group focused on the disease, can lead to annual updates, as in other CPG development groups like hypertension(36). The RCO was the unique CPG that included criteria for monitoring and/or auditing. Having an accessible CPG with an adequate description of audit criteria, useful to audit the implementation of the CPG in the future as has been done in glaucoma(37, 38) and Herpes Simplex Keratitis(39).
On an overall Guideline Assessment, all guidelines except for PPP (Score 4) were scored 3 and recommended with modifications. This recommendation is based mainly on methodological assessment, but also on content recommendations. Even when AGREE II scores were low for all the guidelines, some of these guidelines are useful tools for the diagnosis and management of RVO. SERV and CEC have simple and useful algorithms.
Guideline clinical recommendation meta-synthesis
When assessing the evidence matrix (Guideline clinical recommendation meta-synthesis), risk factor reports are different between assessed guidelines. The most important risk factor for RVO is hypertension (meta-odds ratio of 2.82%), followed by advanced age (1.60 for every decade increase), heart attack history (2.23), stroke history (2.07), total cholesterol (1.32 for every mmol/L increase) and creatinine (1.04 for every ten-mmol/L) (1). Other authors report different OR for several risk factors: hypertension (OR = 1.10), Heart failure (OR = 1.30), Ischemic heart disease (OR = 1.37), Peripheral artery disease (OR = 1.89), and stroke (OR = 2.21) (2). Atrial fibrillation is also a risk factor for RVO(40), however other factors like non-O blood groups are known risk factors for thrombotic and cardiovascular disease, but not for RVO(41).
Other authors report that arterial hypertension, high-density lipoprotein level, hyperlipidemia, peripheral artery disease, stroke, anticoagulation, aspirin use, age-related macular disease, glaucoma, intraocular pressure, refractive errors, diabetes, liver disease, and renal disease are risk factors with an effect exceeding 10% (42). Some studies report that in patients under 50 years old, hypertension and hyperlipidemia were the most frequent CV risk factors for RVO, however congenital thrombophilic disorder was the third in frequency(43).
Diabetes is one of the risk factors still in doubt (Table Nº 4.). A meta-analysis showed that diabetes is associated with RVO (OR = 1.68), with CRVO (OR = 1.98), but not with BRVO (OR = 1.22, CI: 0.95–1.56) (44). Antiphospholipid syndrome and Thrombophilia tests recommendations vary between guidelines, especially with CEC. Based on guidelines, there’s not enough evidence to recommend anticoagulation or antiplatelet therapy, but new studies suggest that anticoagulation could lead to better VA improvement and fewer recurrency compared with antiplatelet therapy, considering the bleeding risk(45). Recommendations for systemic diagnosis in patients with RVO is an important outcome that must be issued by an internal medicine/hematology/primary care specialist(46, 47).
Diagnosis or monitoring criteria were not included in evidence matrix. However, OCT markers are reported in some of the guidelines. Also, some studies report that “baseline central subfield thickness (CST) < 464 µm, absence of subretinal fluid, absence of hyperreflective foci (HF). intact ellipsoid zone (EZ) and external limiting membrane (ELM), absence of epiretinal membrane (ERM) and absence of macular ischemia on FFA, were associated with a better response to intravitreal treatment at 12 months(48). Mean CRT decreased from 554.3µm (603.1µm for CRVO and 496.7µm for BRVO) to 314.4µm after 2 years, and to CRT for CRVOs was 254.2µm (for CRVO) and to 147.8µm for BRVO after 5 years. A formal comparison between drugs was not performed, however available data at 5 years show that most patients were receiving mixed treatment(49).
Most of the guidelines show enough evidence to recommend anti-VEGF as first line treatment and intravitreal steroids as second line treatment. A recent meta-analysis reported that BVCA improved from baseline to 2 years, 3 years (with a decline for CRVO) and 5 years (for both BRVO and CRVO)(49) with anti-VGEF or Dexamethasone. Treat and extend is a viable regimen for the treatment of macular edema secondary to RVO(50). In some reports there’s no significant difference between treat and extend with monthly and PRN treatment, but treat and extend had significant increase of injection frequency compared with PRN. And fewer injections compared with monthly. Even when Bevacizumab is not approved for the treatment of macular edema in RVO, several studies show that in the real-world setting, Bevacizumab can be effective at improving vision in patients with ME secondary to RVO(51). There are no available clinical trials (head-to head trials) to clearly state the best treatment for RVO patients. Some economic studies show better cost-effectiveness for Ranibizumab compared with Aflibercept(52); however, network meta-analysis showed slightly better visual outcomes for Aflibercept compared with Ranibizumab(53), while others show no difference between available anti-VEGF in efficacy (54–56) or in IOP (57). Recently anti-VEGF available like Faricimab are being evaluated in RVO (BALATON and COMINO clinical trials)(58), or Bevacizumab in some available case reports(59).
Comparing new anti-VEGF with Dexamethasone, there's not enough evidence to conclude which of them is better (in terms of VA and adverse effects) for the treatment of macular edema(60), or if combination therapy has the better safety profile compared with steroids alone(61). Network meta-analysis showed that anti-VEGF (Aflibercept and Ranibizumab) were superior to Dexamethasone for the treatment of macular edema associated with CRVO(53, 62, 63). Other studies reported that Dexamethasone intravitreal has better efficacy than anti-VEGF in the treatment of macular edema associated with RVO, but the safety profile (cataract development or exacerbation and high intraocular pressure) is inferior to other anti-VEGF(64). Currently there are no published cost-analysis of Dexamethasone compared with anti-VEGF in RVO, however available studies in Diabetic macular edema report that when Bevacizumab is available, Dexamethasone increases economic costs(65), but in other contexts, eye care-related medical costs are lower for Dexamethasone compared with anti-VEGF(66), especially as an early switch when there is no response to anti-VEGF(67).
SERV recommends that Ranibizumab is the safer anti-VEGF in patients with an adverse cardiovascular event less than 3 months ago. Some recent studies, including MA, reported that anti-VEGF treatment for RVO doesn’t increase risk of CV events, hypertension, or arrythmias, and that comparing Aflibercept with Ranibizumab has no difference in cardiovascular events(68). A French nationwide cohort study reported that for retinal diseases indications, Ranibizumab and Aflibercept have the same risk for myocardial infarction, stroke, or all-cause death, with a small stroke risk increase (not significant) in diabetic patients for Aflibercept(69).
Laser photocoagulation is an issue where views differ. Two meta-analysis showed no difference between the combination of intravitreal injections and retinal laser photocoagulation compared with single intravitreal injections in the treatment of macular edema associated with RVO, however laser photocoagulation with intravitreal injections decrease number of intravitreal injections in patients with BRVO but not in patients with CRVO(70, 71).
Few guidelines included details about follow-up, a relevant issue, considering that most patients with RVO will continue to receive anti-VEGF 5 years after initial treatment(72). It may be related to lower visual and anatomical gain in the real-world with anti-VEGF treatment, because of not achieving frequency of injections compared with clinical trials results(73).
A limitation of this systematic review was a language filter just for Spanish or English, however, just one guideline in another language (German) was available.
Finally, results don’t show which CPG is better, but we can report that PPP has the higher score in the domain “Rigour of Development”, which can mean that it has a better methodological quality compared with the other CPG.
Most of the CPGs for the diagnosis and management of retinal vein occlusion have a low methodological quality evaluated according to the AGREE-II.
PPP has the highest score in the domain “Rigour of Development”, however, they do score well in the “clarity of presentation” domain.
Among the CPGs evaluated, there is no clear recommendation on the type of anti-VEGF therapy to choose.