1. Patient demographic characteristics and clinical features
A total of 29 eyes from 26 patients (14 men and 12 women) were included. Twenty-one patients were unilateral and 3 cases were bilateral (two cases had simultaneous onset in both eyes, and one case had a 5-week interval between both eyes). The mean age of onset was 55.7±9.8 years, except one patient with HSV infection, whose age was 16 years. The average time from onset to presentation was 16.56±13.27 days. The median follow-up time was 8.25 (3.38-17.25) months. A total of 11.6% (3/26) had a history of herpesvirus infection elsewhere in the body, and 11.6% (3/26) had immunosuppression. One patient showed CRAO at the time of onset. The median duration of oral antiviral therapy was 15 (9-35.25) weeks. Twenty-two eyes (75.8%) received at least one intravitreal injection of ganciclovir. (Table 1)
2. Visual prognosis and risk factors
The mean visual acuity at initial examination was 1.36±0.81logMAR. 1.36±0.81 eyes had initial visual acuity ≥1 logMAR, 8 eyes (27.6%) had initial visual acuity between 1 logMAR and 0.5 logMAR, and 4 eyes (13.8%) had initial visual acuity ≤0.5 logMAR. The mean visual acuity at last follow-up was 1.27 ±0.98 logMAR, and 15 (51.7%) had a last visual acuity ≥1 logMAR in the last eye. Four eyes (13.8%) had a last visual acuity between 1 logMAR and 0.5 logMAR, and 10 eyes (34.5%) had a last visual acuity≤0.5 logMAR. Comparing visual acuity at the last follow-up with that at the first visit, 16 (55.2%) eyes improved, and 13 (44.8%) did not improve. Logistic regression analysis revealed that risk factors for failure to improve vision after treatment included retinal detachment, necrotizing retinitis involving the posterior pole, and arteritis involving the large retinal arteries. Age and intravitreal injection of antiviral drugs were not associated with failure to improve vision after treatment.
3. Extent of retinal necrosis and visual prognosis
The partitioning of retinal necrosis at initial diagnosis was as follows: 31% (9/29) of eyes were in Zone 1, 41% (12/29) of eyes were in Zone 2, and 28% (8/29) of eyes were in Zone 3. Independent t tests also revealed that eyes with necrosis in Zone 2 and Zone 3 had a better visual prognosis than did those with necrosis in Zone 1 (P=0.014 and P=0.025, respectively) (Figure 2). After treatment, the mean time to regression of retinal necrosis was 32.05 ±13.75 days, except for one patient with CMV who had not completely resolved at the end of follow-up. The mean resolution time of the VZV-infected patients was 27.6±14.2 days, while the complete resolution time of the unidentified pathogen was 37±4.06 days. There was only one case of CMV infection with complete resolution, and the duration was 37 days. The time to complete resolution in one patient with HSV was 62 days. The statistical analysis showed that the time to resolution of necrosis was significantly correlated with visual prognosis (P=0.021).
At initial diagnosis, 17 of the 29 affected eyes showed vasculitis involving the large retinal arteries, and the remaining 12 eyes were limited to peripheral arterioles or had no obvious arteritis. Analysis revealed that vasculitis involving large retinal arteries at the time of first diagnosis was associated with poor visual prognosis in the affected eye (P<0.01) (Figure 2).
4. Retinal detachment and risk factors
Patients who developed retinal detachment had a worse visual prognosis (P<0.01) (Figure 2); therefore, we further analyzed the risk factors for retinal detachment. A total of 10 (34.5%) eyes exhibited retinal detachment; 1 eye developed retinal detachment at presentation, and 9 eyes developed retinal detachment during treatment. The median time from onset to retinal detachment was 36 (21-77.5) days. The following factors were used to analyze the risk factors for retinal detachment: age at onset, more than 14 days from onset to treatment, vasculitis involving the main artery around the optic disc, and extent of retinal necrosis. The results showed that only arteritis involving large retinal arteries was a risk factor for retinal detachment (OR =12.38; P=0.029). Another factor associated with retinal detachment is the number of viral DNA copies in the aqueous humor. Our results showed a significant correlation between viral load and retinal detachment (P=0.028).
Retinal detachment occurred in 27.2% (6/22) of patients receiving intravitreal antivirals and 50% (3/6) of untreated patients. Due to the small number of cases, the effect of intravitreal antiviral injection on retinal detachment cannot be statistically significant. However, there was a statistically significant difference in the occurrence of retinal detachment between patients who did and did not intravitreal antiviral therapy.
5. Viral DNA copy number and visual prognosis
Aqueous humor samples were collected from 20 out of 29 eyes, and PCR confirmed that 85% (17/20) of the eyes were VZV-infected, 10% (2/20) were CMV, and 5% (1/20) were HSV. The average viral copy number in VZV-infected patients was 5.73 ± 0.77 log10 copies/ml. Spearman correlation analysis also revealed that the VZV viral load in the aqueous humor at initial presentation was significantly associated with visual prognosis (P=0.013) (Figure 3A) and the extent of retinal necrosis. In patients with posterior pole retinal necrosis, the number of viral DNA copies was significantly greater than that in patients with peripheral retinal necrosis (P=0.03) (Figure 3B). Further analysis also revealed a significant correlation between viral load and retinal detachment (P=0.019) (Figure 3C).
NLR and visual prognosis
To study whether an elevated NLR correlates with visual acuity prognosis, we collected the NLR data of all patients at their first visit before starting antiviral treatment. Due to the relatively small number of patients with CMV and HSV infections, we excluded these data and analyzed the correlation between the NLR and visual prognosis in patients (12 patients in total) with VZV infection. The results showed that the NLR of VZV-infected patients at first presentation was significantly correlated with the prognosis of visual acuity (P=0.033) (Figure 4). Next, we further analyzed the relationship between the NLR and risk factors affecting visual prognosis, such as retinal detachment, the retinal necrosis area, and arteritis involving large retinal arteries. The mean NLR in the retinal detachment group was 3.6 (3.3, 5.4), whereas it was 2.14 (2.11, 2.75) in the nonretinal detachment group. There was a significant difference between them (P=0.003), which suggested that a higher NLR may be a risk factor for retinal detachment. However, the NLR was not correlated with the retinal necrosis area or arteritis involving large retinal arteries. (Figure 4)