Clinical and Diagnostic Findings of Acute Macular Neuroretinopathy and Paracentral Acute Middle Maculopathy in the COVID-19 Era Free

Growing evidence suggests an association between the infection from the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) and eye disorders [1‒6]. Early in the pandemic, the transmission of SARS-CoV-2 through the ocular surface was suggested [7], and more recent findings confirmed this hypothesis [8]. Although ocular involvement may be only of the surface and clinically mild in most of the cases, severe manifestations and imaging alterations of the internal ocular structures have been described [9‒12]. Among them, individual cases of rare retinal disorders in patients previously affected by the novel coronavirus disease (COVID-19) or concomitantly infected by SARS-CoV-2 have been reported [13‒29].

Acute macular neuroretinopathy (AMN) [30] and paracentral acute middle maculopathy (PAMM) [31] are retinal disorders associated with retinal vascular alterations, but their exact pathogenesis remains unclear. Scotomas and best-corrected visual acuity (BCVA) reduction are the main symptoms of both these conditions, and they are frequently persistent. Both the AMN and PAMM diagnoses are largely based on optical coherence tomography (OCT) imaging, which demonstrates structural alterations of the retinal layers, which can be long lasting [32].

The aim of this review was to collect and analyze the clinical presentation and diagnostic features of AMN and PAMM associated with COVID-19 – considering also when they are coincident – and to compare our results with previous reports regarding these retinal disorders, in order to recognize possible specific characteristics.

A literature search was performed by consulting PubMed, Scopus, and Embase. The following terms were searched: “(COVID-19 OR SARS-CoV-2 OR coronavirus) AND ([acute macular neuroretinopathy] OR [paracentral acute middle maculopathy]).” Inclusion criteria were as follows: (1) publication date from January 31, 2020 to January 31, 2022; (2) English language; (3) original research or case report; (4) free full-text availability. One-hundred and fifty-six records were identified, and after removing duplicates, three investigators independently performed the screening of the sixty-three studies selected (Fig. 1).

After title and abstract evaluation, thirty-four studies presented a different main topic, and eight additional works were not original reports; of the remaining twenty-one eligible studies, two focused on another main topic and in another two the full text were not available. Therefore, seventeen original reports were included, reflecting 22 patients.

Patients’ demographic and clinical characteristics were collected (Table 1). Among the 22 subjects, the female gender was predominant (68%) and the mean (range) age was pair to 38 ± 17 years (21–75). Medical history identified comorbidities in about half of these patients (55%). Two subjects were affected by type 2 diabetes [18, 19], and one presented arterial hypertension and received a kidney transplant [15]. Previous ophthalmic disorders were detected in just 1 patient, who presented unilateral vision loss (VL) due to post-traumatic glaucoma [20]. At the time of the ophthalmic evaluation, 1 patient was hospitalized for femur fracture complicated with pulmonary embolism [24], and another was affected by severe COVID-19 acute respiratory distress syndrome requiring corticosteroid treatment [14]. One patient with a negative medical history was receiving corticosteroid and low-molecular-weight heparin for elevated D-dimer levels [26], and 3 patients without relevant comorbidities were assuming oral estrogen/progesterone [16, 17].

Clinical presentations of COVID-19 and ocular manifestations are displayed in Table 2. Almost all patients had a diagnosis of COVID-19 confirmed with nasopharyngeal swab (18/22) polymerase chain reaction or with serology (3/22). In one case, the laboratory diagnostic test performed was not specified. Asymptomatic or mild symptomatic presentations of the infection were the most common (3/19 and 13/19, respectively), and only 1 patient reported ocular manifestations (itchy eyes with subconjunctival hemorrhage) in association to mild COVID-19. Pneumonia was diagnosed in one case, 1 patient presented an acute episode of pulmonary embolism, and one isolated case of severe acute respiratory distress syndrome was reported.

Data on the timing of ocular symptom onset were available for 20 patients. Thirteen patients presented ocular symptoms after the infection, with a mean (range) delay of 26 days (7–120), while in seven cases the symptoms were concomitant with COVID-19. The most common ocular manifestation was the acute onset of a scotoma (17/21), in few cases associated with BCVA reduction (3/17), dyschromatopsia (2/17), and photopsia (1/17). Isolated BCVA reduction or VL were other reported symptoms (3/21 and 1/21, respectively). The ophthalmic examination showed BCVA reduction and VL in all patients complaining these symptoms (7/7), while BCVA was more frequently normal in patients presenting scotomas without subjective BCVA alterations (10/14).

All patients underwent multimodal ophthalmic assessment (Table 3), which was heterogeneous between different studies. When performed, near-infrared reflectance always detected hypo-reflective lesions (12/12), and fundoscopy showed abnormalities in the majority of cases (12/17). On the other hand, several patients had no alterations on indocyanine green angiography (4/5), fundus autofluorescence (6/9), and fluorescein angiography (7/13).

As displayed in Table 4, all 22 patients underwent OCT imaging, which was abnormal in all cases; similarly, OCT angiography (OCTA) frequently identified vascular alterations (9/11). After OCT imaging, AMN was diagnosed in 13 patients and 2 subjects had both AMN and PAMM. A diagnosis of isolated PAMM was made in 5 patients, while in 2 cases it was associated with other retinal vascular disorders.

AMN patients often had hyper-reflectivity (HR) of the outer plexiform layer (OPL) (11/13), of the outer nuclear layer (ONL) (6/13), and ellipsoid or interdigitation zones (EZ and IZ, respectively) disruption (10/13). In most cases, the presence of HR and EZ/IZ abnormalities resulted combined (9/13). When performed, OCTA identified attenuation of the deep capillary plexus (DCP) of the deep vascular complex (7/9).

The most common OCT finding in PAMM was an alteration of the inner nuclear layer (INL) (3/5), associated with other areas of HR of inner plexiform layer and OPL (3/5), while no signs of EZ/IZ disruption were detected (0/5). In one case, the OCT alterations suggestive for PAMM were not specified. When performed, OCTA showed the attenuation of both the DCP and the superficial capillary plexus (SCP) of the superficial vascular complex (1/1). In two more cases, PAMM was associated with cilioretinal artery occlusion (1/2) and with central retinal artery occlusion (1/2). In the first case, the inner retinal layers had HR and increased thickness, while in the second one HR of the INL and increased thickness of the ganglion cell layer and of the nerve fiber layer were detected.

One of the two patients with both AMN and PAMM showed several areas of HR, including OPL, ONL, IN, and the Henle’s fiber layer, associated with EZ/IZ disruption and to DCP attenuation at OCTA, while the second patient presented HR of the inner retinal surface and signs of disruption of the outer retinal layers.

Two patients received a specific treatment for central retinal artery occlusion and cilioretinal artery occlusion, while no therapy was prescribed in all the remaining cases (20/22). Among patients presenting follow-up data (15/22; 65 ± 48 days, range 35–180), a full recovery was observed in just 1 case, a partial recovery was identified in eight cases, while 6 cases showed no clinical or imaging sign of improvement.

AMN and PAMM are retinal vascular diseases with a still unclear pathogenesis. With the advent of modern retinal multimodal imaging, in particular, OCTA, it is now evident that they are distinct disorders with some common features. A recent study showed these similarities and differences between AMN and PAMM developing in the same eye: a common pathogenic pathway has been hypnotized that may result from impairment of the venous outflow channels of the DCP, with a difference in the depth of this impairment in the DCP between AMN and PAMM [33]. Several environmental triggers as well as other ocular and systemic diseases have been reported as associated with AMN [30] and PAMM [31]. The case reports of patients developing AMN and PAMM after or in concomitance with SARS-CoV-2 infection are growing, and a possible association must be considered.

AMN is a retinal disorder mainly associated with flu-like illness by several viral agents, use of oral contraceptives, and administration of catecholamines, antecedent trauma, and systemic shock [30]. All these different associations could be unified in the pathophysiology if a microvascular damage or hypoperfusion pathophysiology is considered [34]. The attenuation of DCP signal at OCTA increases the likelihood of this etiology [35].

The infection sustained by SARS-CoV-2 can cause multiorgan microvascular changes, including endothelial cell swelling and damage (endotheliitis), microscopic blood clots (microthrombosis), capillary congestion, and a damage to pericytes, which are fundamental to capillary integrity and barrier function, tissue repair (angiogenesis), and scar formation [36]. In a recent autopsy study, SARS-CoV-2 ribonucleic acid was detected in the retina of 3 out of 14 patients deceased for COVID-19 [37]. Our review reported the diagnosis of AMN in 15 patients with confirmed infection of SARS-CoV-2. In nine out of these 15 patients, OCTA was performed, and in seven of them, alterations of DCP signal were observed. Thus, it is reasonable to assume that SARS-CoV-2 infection could be associated with the development of AMN through pathophysiological mechanisms similar to the ones assumed for AMN following flu-like illnesses. The microvascular pathophysiology with a subsequent hypoperfusion could explain the damages to the outer retinal layers observed in this retinal disease. Moreover, the possible link between the infection and the development of AMN is corroborated by the low prevalence of other known risk factors for AMN (i.e., the use of oral contraceptives in the majority) in the patients identified in this review. Another observation supporting this association is the onset time of AMN symptoms. In fact, in most patients, the symptoms are usually reported 1 or 2 weeks after the SARS-CoV-2 infection. However, this is true with the exception of 2 patients, in whom reported symptoms were after a longer period from the infection (30 and 120 days).

In line with previous evidence about AMN in patients without SARS-CoV-2 infection [30, 34], our review noted a predominance of the female gender (10 of 13 patients). In our sample, the mean (range) age was 37 ± 18 years (21–75), while in the previous two main reviews on the topic it was 29 years [30] and 27 years [34]. Three women were taking oral contraceptives, and the use of this medication has been reported in several patients with AMN [30, 34].

In our review, the clinical presentation of AMN in patients with COVID-19 was mainly characterized by acute onset of central or paracentral scotoma that was present in 11 out of 13 patients, followed by BCVA reduction (3 of 13 patients). These symptoms are in accordance with those reported in AMN not associated with COVID-19 [30, 34]. Among the patients we identified, there was a single case of photopsia [22], a symptom not before reported in the previous literature.

Concerning the OCT findings in COVID-19 AMN patients, the majority of patients had HR of the OPL and disruption of EZ/IZ, while only half of them showed HR of the ONL. In most cases, the OPL and ONL HR and EZ/IZ disruption were combined. All these features are common in non-COVID AMN patients [30]. However, contrary to previous findings in AMN not associated with COVID-19 [30], in these cases the OPL was more frequently involved than the ONL. When performed, OCTA showed attenuation of the signal of the DCP. The same OCTA finding is well described in the literature for AMN in non-COVID patients [30, 34] and is coherent with the supposed vascular pathophysiology of the disease. In a recent study, Cabral et al. [38] reported an interesting OCTA analysis of the DCP in non-COVID AMN patients: this plexus is always involved in this disease, but with two different patterns. A central DCP-draining venule was present in one pattern, and in the other there were central DCP flow defects within the lesion. In the studies considered in our review, this sub-analysis was not performed, but it would be useful to better understand the real pathophysiology of the disease, adding more clear evidence on the reduction on venous outflow in the DCP.

In PAMM pathogenesis, several associations have been proposed with ocular disorders, especially occlusions of the retinal vein or artery [39, 40], hypertensive and diabetic retinopathies [41], congenital glaucoma [42], ocular surgery [43, 44], and local retrobulbar anesthesia [45]. The most common systemic diseases associated with PAMM are intracranial hypertension [46], meningitis [47], vascular surgery [48], pregnancy [49], and specific drugs assumption [50, 51]. As observed in AMN, in some patients, PAMM was linked to a recent viral flu-like disease [52].

A recent study described a retinal vascular model in which an oxygen gradient between the superficial vascular complex and deep vascular complex exposes the middle retina to a high risk of ischemia [53]. OCTA findings in PAMM patients show a reduction in blood flow in DCP [31], suggesting this blood flow model as potentially associated with PAMM development. Even if all the exact pathophysiological mechanisms of PAMM are still unclear [54, 55], the microvascular alterations caused by SARS-CoV-2 infection and the presence of viral ribonucleic acid in the retina [37] suggest a possible association between COVID-19 and the development of PAMM. As observed in AMN, this association is reported in those with concomitant or recent diagnosis of COVID-19 (14–35 days prior) in all these patients.

Most of the patients identified in this review were female and only one was a male, while in non-COVID PAMM no difference of prevalence between the two genders has been observed [31]. The mean (range) age was 40 ± 15 years (26–65), while in the literature it ranges from 49 to 53 years [31].

In our review, 2 of the 7 patients presented PAMM associated with other vascular retinal concomitant diseases. The association with concomitant retinal vascular disease is possible and well described in the literature also in non-COVID-19 PAMM patients [39, 40].

Concerning clinical presentation, among COVID-19 PAMM patients, most had paracentral scotomas associated with BCVA reduction, only one presented with isolated BCVA reduction and only one had isolated paracentral scotoma. These symptoms found in COVID-19 PAMM patients correspond to the typical presentation of non-COVID PAMM [31].

At ophthalmic examination, in this review, fundoscopy was negative in 1 case, while small cotton wool spots were detected in 2 patients (one of them presenting with retinal hemorrhages) and an isolated hyperpigmented lesion was observed in 1 patient. These findings are peculiar and not common in non-COVID PAMM, in which the fundoscopy is usually normal or characterized in rare cases by deep slightly gray lesions [31]. It should be noted that cotton wool spots are seen in PAMM patient when this condition is associated with other retinal vascular diseases that cause an ischemia of the tissue [56]. These findings may not be specific for COVID-19 PAMM but could represent another sign of vascular damage in the pathophysiology of the disease.

All the five COVID-19 PAMM patients described in this review underwent OCT, demonstrating HR of INL in the absence of alterations of EZ and IZ. These findings are the same to those identified in non-COVID PAMM patients, which usually have an HR of the INL (even an involvement of OPL and inner plexiform layer is frequent) in the early stages of the disease and a subsequent permanent thinning of the same retinal layer in the chronic stages [31, 54].

OCTA findings were reported in only 1 out of 5 patients and showed the reduction in blood flow in the SCP and DCP, thus suggesting a microvascular damage. These findings are similar to those found in non-COVID PAMM patients. In fact, the DCP flow is frequently reduced [31], while the involvement of the SCP has been described in just few cases [54].

AMN and PAMM are retinal vascular diseases which have still to be clearly understood in their pathogenesis and in risk factors. Their clinical presentation is frequently not so evident, determining an underestimation of the prevalence of these diseases. The common pathogenetic feature of AMN and PAMM is the retinal microvascular alteration detectable with OCTA.

In this review, we reported a case series of AMN and PAMM in patients with a previous or concomitant infection from SARS-CoV-2. The microvascular changes in these cases are highlighted by the OCTA scans, which showed a reduction in the signal in SCP and DCP in PAMM patients and in DCP in AMN patients. Even if we are far from the determination of a direct link between COVID-19 and these retinal disorders, we could hypothesize that the vascular alterations associated with SARS-CoV-2 infection could be a possible risk factor for both AMN and PAMM. Moreover, a previous viral flu-like syndrome has been frequently reported before the development of AMN or PAMM in several previous reports. Furthermore, the majority of the patients we identified did not report other specific risk factors for the development of these retinal diseases.

Further research is needed to better assess the link between SARS-CoV-2 infection and AMN and PAMM, with a larger number of patients and a standard protocol of investigation, particularly with the multimodal retinal imaging, which is crucial for the investigation of AMN and PAMM. The study of microvasculature in other organs in patients with AMN or PAMM presenting in association with SARS-CoV-2 infection may also be helpful to better understand the pathogenesis of the disease.

An ethics statement is not applicable because this study is based exclusively on data already published in the literature.

E.P., R.A., D.S., I.A.I., N.P., F.G., G.G., M.S., A.S., S.D., C.A., and F.L. declare no conflict of interest on this topic. B.D.M. is supported to conduct COVID-19 neuroscience research by the UKRI/MRC (MR/V03605X/1); B.D.M. is also supported for additional neurological inflammation research due to viral infection by grants from the MRC/UKRI (MR/V007181//1), MRC (MR/T028750/1), and Wellcome (ISSF201902/3).

No external funding to support the production of this review.

Conceptualization: E.P. and F.L.; literature search and screening: E.P., S.D., and C.A.; original draft preparation: E.P. and F.L.; review and editing: E.P., R.A., D.S., I.A.I., N.P., F.G., G.G., M.S., A.S., B.D.M., S.D., C.A., and F.L. All the authors have read and approved the definitive version of the manuscript.

All data presented and analyzed in this review are available in each original article: references from 13 to 29. Further inquiries can be addressed to the corresponding author.