Purpose To describe novel underlying associations of classic acute macular neuroretinopathy (AMN).
Methods Multimodal imaging case series evaluating patients with classic AMN lesions and previously unreported underlying aetiologies.
Results Six patients were included (five women, one man, mean age 30±7 years). Mean distance best corrected visual acuity at initial presentation was 0.21±0.3 logMAR (mean Snellen acuity: 20/30, range 20/15–20/100) and at last follow-up visit 0.09±0.17 logMAR (Snellen acuity: 20/20, range 20/15–20/60). All cases but one had bilateral lesions and showed typical parafoveal hyporeflective lesions on infrared imaging, which corresponded to the hyper-reflectivity in the Henle's layer with attenuation of the external limiting membrane, the ellipsoid zone and interdigitation zone. Underlying diseases included thrombocytopenia and anaemia associated with dengue fever, acute lymphoblastic leukaemia, chronic kidney disease and ulcerative colitis, while Valsalva-like manoeuvre was found to be a potential trigger. Other novel associations included the use of lisdexamphetamine.
Conclusions Classic AMN may be associated with leukaemia, dengue fever, ulcerative colitis and chronic kidney disease, probably as a result of chorioretinal hypoxia in the setting of thrombocytopenia and anaemia. Adrenergic agonists such as lisdexamphetamine may also contribute to the manifestation of AMN.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Acute macular neuroretinopathy (AMN) is an increasingly diagnosed disorder typically affecting young healthy people. Based on biomicroscopic examination, early studies localised AMN lesions to superficial retinal layers.1–3 Using spectral domain optical coherence tomography (SD-OCT) classic AMN lesions can be assigned to the outer retina, presenting as hyper-reflectivity of the outer plexiform layer (OPL), Henle's layer and outer nuclear layer (ONL), which ultimately lead to localised disruption of the interdigitation zone (IZ), ellipsoid zone (EZ) and the external limiting membrane (ELM). Later in the disease course, subsequent thinning of the ONL ensues.2 ,3 AMN lesions are clearly visible on near-infrared (NIR) reflectance and red-free imaging, manifesting as characteristic hyporeflective parafoveal wedge shaped or teardrop shaped areas.2 ,3 Short wavelength (488 nm) fundus autofluorescence (FAF) and NIR (787 nm) FAF can also identify subtle changes correlating with these typical lesions.2 ,3
The original paper by Bos and Deutman described AMN in healthy young women taking oral contraceptives.1 Since then numerous aetiologies have been reported,2 ,4 ,5 including caffeine intake, epinephrine/ephedrine therapy, viral and flu-like illnesses, diabetic retinopathy, migraines, trauma, hypotension and anaphylactic shock, postpartum hypotension, rhinosurgery, contrast media administration and even antithymocyte globulin infusion.4 The pathogenesis remains unknown, although a vascular theory has been proposed, based on the vasoactive nature of the precipitating factors as well as the anatomical localisation of pathology.2
The aim of this study is to describe previously unreported aetiologies of classic AMN and broaden the spectrum of clinical presentation.
Patient selection and setting
This retrospective observational case series included six patients from three tertiary referring institutions: Feinberg School of Medicine at Northwestern University Chicago, USA, Retina clinic, University of Sao Paulo, Brazil, and Retina Vitreous Associates Medical Group, Los Angeles, USA. The study adhered to the tenets of the Declaration of Helsinki.
Patients were included if they were diagnosed with classic AMN with previously unreported aetiologies. AMN was diagnosed according to the following clinical criteria: parafoveal grey dark wedge shaped lesions with NIR imaging and hyper-reflectivity of the OPL/ONL junction with associated attenuation of the underlying EZ and IZ on SD-OCT at baseline along with subsequent thinning of the ONL during the follow-up. Data collection included age, sex, medications, medical history, best corrected visual acuity (BCVA), and clinical findings with slit-lamp examination and ophthalmoscopy. Image analysis included combinations of the following imaging modalities, whenever available: colour fundus and red-free photography, SD-OCT, NIR reflectance, blue (488 nm) FAF, NIR (787 nm) FAF, fluorescein angiography, indocyanine angiography (Heidelberg Engineering, Germany or Topcon, USA) and microperimetry (MP-1 microperimeter, Nidek Technologies, Italy).
Demographic characteristics and multimodal evaluation
Six patients (five women, one man) were clinically diagnosed with classic AMN and included in this case series. Demographic characteristics, systemic treatments administered prior to baseline, medical history and associated systemic aetiologies can be found in table 1. Mean age was 30±7 years. Mean distance BCVA at presentation was 0.21±0.3 logMAR (mean Snellen acuity: 20/30, range 20/15–20/100) and at last follow-up visit 0.09±0.17 logMAR (mean 20/20, range 20/15–20/60) (table 1). Mean follow-up duration was 2±0.7 months.
Exemplary case reports
A healthy 27-year-old woman from Brazil experienced bilateral central vision loss, 10 days after the onset of headaches and flu-like symptoms. She was not taking any medications. The tourniquet test was positive and laboratory workup revealed low leucocyte count and thrombocytopenia of 96 000/μL (normal range 140 000–440 000/μL). Serological investigation was negative for syphilis, tuberculosis, malaria, HIV, cytomegalovirus and sarcoidosis but positive for dengue fever. At presentation, BCVA was 20/20 right eye (OD) and 20/40 left eye (OS). Baseline clinical and multimodal imaging findings are described in figure 1. At the 9-week follow-up BCVA was 20/20 OD and 20/25 OS with notable fading of the hyporeflective NIR lesions, incomplete restoration of the hyper-reflective bands on SD-OCT and reduction of the number of absolute scotomata on microperimetry (figure 1E, F). Images of a second dengue case (case 2) are shown in figure 2.
A 31-year-old woman presented with a 6-week history of decreased and distorted vision, OD>OS. History revealed decreased appetite, night sweats, epistaxis, petechiae and enlarged lymph nodes. She was diagnosed with BCR/abl negative acute lymphoblastic B cell leukaemia without central nervous system involvement. Chemotherapy consisting of vincristine and daunorubicin was started 5 days after diagnosis. Further, she received intrathecal methotrexate, dexamethasone, acyclovir, bactrim, fluconazole and moxifloxacin for neutropenic fever. Ophthalmoscopic examination demonstrated intraretinal haemorrhages and Roth spots in the posterior pole both eyes (OU), attributed to concurrent thrombocytopenia of 6000/μL and anaemia (Hb 4.2 g/dL, normal range 11.7–16 g/dL). The patient was initially seen in our practice during her follow-up, at day 23 of course 1 of chemotherapy. Platelet count had improved to 35 000/µL. Vision was 20/25 OU. Ophthalmoscopy revealed resolving intraretinal haemorrhages. Fluorescein angiography showed no evidence of ischaemia, non-perfusion or leakage. While both eyes showed multiple classic, mainly parafoveal, hyporeflective dark greyish lesions, two of these lesions OS corresponded to areas of subtle hyperautofluorescence with blue (488 nm) FAF and hypoautofluorescence with (787 nm) NIR FAF, which is atypical for AMN (figure 3A). Corresponding SD-OCT changes are described in figure 3B, C. Follow-up revealed resolution of intraretinal haemorrhage and hyporeflectivity on NIR while the OCT hyper-reflective bands recovered partially. The hyperautofluorescence on 488 nm FAF and hypoautofluorescence on NIR-FAF OS showed reduced intensity at follow-up, however the corresponding localised disruption/loss of the IZ persisted.
A 28-year-old Caucasian woman with myopia (−2.75D) with a history of end-stage idiopathic renal disease of unknown aetiology, who underwent kidney transplant (2004). The patient was recently admitted for graft chronic rejection and fibrosis, and was seen 2 weeks after discharge from the hospital for viral pneumonia and dehydration. During hospitalisation she was found to have leucopenia (3100/µL, normal range 3600–10 700/µL), anaemia (Hb 6 g/dL), hyponatraemia (124 mEq/L, normal range 135–145 mEq/L) and thrombocytopenia (113 000/µL), which resolved during admission. During hospitalisation, she complained of blurred vision, which began almost immediately following an episode of vomiting. When the patient presented 2 weeks later in our clinic at Northwestern University, BCVA was 20/70 OD and 20/40 OS. Slit-lamp examination showed bilateral resolving subconjunctival haemorrhages. Indirect ophthalmoscopy revealed central retinal pigment epithelium (RPE) mottling OU and intraretinal haemorrhage, OS. NIR imaging revealed large hyporeflective lesions OU. OCT revealed corresponding attenuation of EZ, IZ and ELM with subtle hyper-reflectivity of the Henle's layer (figure 4A). Two months later, the AMN lesions had improved with fading of the hyporeflectivity on NIR and partial reconstitution of the hyper-reflective bands on SD-OCT. BCVA had recovered to 20/20, OD and 20/30, OS.
A 30-year-old Caucasian woman with recurrent ulcerative colitis was hospitalised with nausea, abdominal cramping and vomiting. Her medications included oral contraceptives and mesalamine enemas. She was treated with intravenous fluids for dehydration and the laboratory workup revealed anaemia. According to the patient, she was given a dose of intravenous corticosteroids, and within a day she noticed vision loss OU associated with multiple small scotomata. One week after the acute scotomata onset she presented to us. BCVA was 20/100, OD, 20/80, OS. Indirect ophthalmoscopy revealed a flame-shaped haemorrhage and cotton wool spots along the superotemporal arcade OD. Hyporeflective central lesions were visible on NIR and SD-OCT showed corresponding disruption of the IZ and EZ (figure 4B). Follow-up examination, 3 weeks later, revealed improvement of BCVA to 20/60, OD and 20/30 OS, with incomplete recovery of the EZ and IZ.
The cases presented here highlight thrombocytopenia and anaemia as potential factors contributing to the onset of AMN. Leucocytosis, increased production of leucoblasts and hypovolaemia leading to hyperviscosity may be contributing factors as well. These factors were probably the main cause of classic AMN in our patients suffering from dengue fever (case 1 and case 2), acute lymphoblastic B cell leukaemia (case 3), chronic kidney insufficiency (case 4) and ulcerative colitis (case 5). Other novel associations included the use of lisdexamphetamine (case 6) for the treatment of attention-deficit/hyperactivity disorder (ADHD).
Localisation of the level of ischaemia in AMN has been elusive. Paracentral distribution of AMN lesions has been a consistent finding indicating that the retinal capillary plexus may be implicated in the aetiology of this disorder. The presence of outer retinal involvement including initial ONL and OPL hyper-reflectivity, followed by subsequent thinning of the ONL along with photoreceptor disruption suggest ischaemia at the deep capillary plexus (DCP). In contrast, paracentral acute middle maculopathy (PAMM), a recently coined term to describe hyper-reflectivity at the inner nuclear layer level, likely results from ischaemia at the level of the intermediate capillary plexus (ICP). It is therefore our hypothesis that ischaemia of the ICP results in PAMM as a result of middle retinal involvement while ischaemia of the DCP alone results in AMN with outer retinal involvement, with varying degrees of overlap depending on the underlying pathogenesis and overall health of the retinal vascular system. We propose that each of the unique systemic aetiologies presented in this report have contributed to DCP ischaemia and the development of AMN. The location of the OPL within the watershed zone and the high oxygen requirement due to the high synaptic activity in this layer predispose for initial ischaemic injury.
Reports of ocular manifestations in dengue fever are rare and include intraretinal haemorrhage, cotton wool spots, tortuosity of vessels, macular oedema and exudation, retinal capillary ischaemia as well as RPE changes and choroidal effusion syndrome,6–9 whereas ocular findings such as intraretinal haemorrhage, Roth spots, cotton wool spots, retinal vein occlusions and non-arteritic anterior ischaemic optic neuropathy in the context of leukaemia have been extensively described.10 ,11 The appearance of classic AMN lesions in these disease entities has never been described so far. Marked thrombocytopenia has been associated with ocular findings in dengue fever, whereas anaemia, thrombocytopenia and hyperviscosity have been associated with ocular findings in leukaemia.6 ,9 Increased capillary permeability, endothelial dysfunction and haemorrhagic diathesis, platelet destruction, occlusion of precapillary arterioles by immune complex deposition as well as consumptive coagulopathy have been proposed as a cause of the associated retinal findings in these entities and may also be the leading cause of focal ischaemia of the deep capillary network in our patients.6 In particular, profound anaemia and thrombocytopenia may cause retinal capillary ischaemia due to lack of oxygen supply, or due to endothelial injury at the superficial (cotton wool spot) and/or intermediate and deep levels, leading to manifestation of PAMM or AMN, respectively.10 ,12 Hyperviscosity due to accumulation and infiltration of leucoblasts (case 3), due to leucocytosis, as well as hypovolaemia (case 4) or dehydration (case 5) leading to venous stasis and hypoxia are further precipitating factors for the onset of AMN.11 Classic AMN was previously described in one patient with enteritis as well as in three women after delivery (two undergoing caesarean delivery and one undergoing spontaneous vaginal delivery).4 ,5 It is very likely that in these cases hypovolaemia and hyperviscosity were contributing factors to AMN. Forty-six per cent of the reported cases mention flu-like symptoms/fever as an associated factor of the onset AMN.5 So one may argue that the viral infection was the main triggering factor for AMN in case 1 and case 2 rather than the change in blood count. However an alternative hypothesis could be that viral febrile illness and hyperviscosity due to sweating, reduced water intake and leucocytosis may have contributed to focal ischaemia in DCP in these previously reported cases.
The Valsalva-like manoeuvre during emesis and colitis (case 4 and case 5) increases intraocular and intravascular pressures and may further exacerbate hypovolaemia related, anaemia related and thrombocytopenia related capillary ischaemia. Previous reports on patients suffering from aplastic anaemia (and therefore pancytopenia) emphasised that Valsalva-like manoeuvres elevate the intrathoracic and intraocular pressures and, therefore, increase the incidence of retinal findings.13 Similarly, previous AMN reports indicated that Valsalva-like manoeuvres may trigger the onset of AMN.4
Although the shape of retinal lesions in cases 4 and 5 were somewhat atypical with central foveal involvement, the authors believe that based on their onset, appearance on multimodal imaging and their course, these lesions were consistent with classic AMN. We propose that foveal involvement in these cases may be suggestive of generalised chorioretinal hypoxia (anaemia leading to overall decreased oxygen content in blood in anaemia) and a circumferential involvement of the entire parafoveal DCP, which exacerbated the ischaemic insult to the outer retina.
Interestingly, in case 3 as well as case 5, the AMN lesions were found in conjunction with superficial intraretinal haemorrhages. This phenomenon has already been reported in a case of AMN associated with flu-like symptoms, after epinephrine injection during rhinosurgery and in another case where no underlying aetiology could be identified.3 ,4 ,13 ,14 Another report described AMN lesions resembling deep intraretinal haemorrhage and the authors assumed that co-localisation of haemorrhage in the OPL and classic AMN lesions may be possible.15 Interestingly one case (case 3) presented with superficial haemorrhage as well as with characteristic AMN lesions, which corresponded to hyperautofluorescence using the 488 nm FAF filter (and hypoautofluorescence at 787 nm). This is not characteristic for AMN, as most previous reports have described either no visible FAF changes, subtle hypoautofluorescent changes on 488 nm FAF or subtle hyperautofluorescent changes on 787 nm FAF.2 ,16 However one previous report detected AMN lesions presenting with subtle hyperautofluorescence on 488 nm FAF similar to our case.17 Although this finding is not characteristic for AMN, it can be explained by the reduction of filtering of blue light by decreased visual pigment density in outer segments, seen on OCT as loss/disruption of the IZ.18 ,19 Alternatively, as suggested by Lalezary et al,15 this lesion may represent co-localisation of the deeper retinal haemorrhage and AMN.
Lisdexamphetamine dimesylate (case 6), the first long-acting prodrug for the treatment of ADHD is an adrenergic agonist and sympathomimetic that leads to an increase of catecholamines in the extracellular space and has vasoactive effects similar to epinephrine, ephedrine and norepinephrine, drugs that have previously been associated with AMN.1 ,2 ,4 ,5 The increase of adrenergic transmitters, be it due to adrenergic drugs or due to trauma-induced catecholamine release, leading to retinal capillary network vasoconstriction has been previously linked to the onset of AMN.2 ,3 ,20–22 So far, about seven patients with AMN in the setting of sympathomimetic agents have been reported.5 ,21 The migraines and use of oral contraceptives (case 6), both factors previously associated with AMN, may have had additive effects contributing to the onset of AMN.4 ,23 The number of risk factors may determine extent and severity of ischaemia, and thus the size of AMN lesions in this series, similar to previously published cases with multiple associated risk factors that had large, rather unusual centre-involving AMN lesions.15 ,17
To summarise, this report presents several previously unrecognised potential aetiologies of classic AMN. Our findings further support the association between sympathomimetic drugs and development of AMN, but also add thrombocytopenia and anaemia, hyperviscosity due to leucocytosis, leucoblastocystis, hypovolaemia and dehydration as new risk factors. This report is important to alert clinicians of the potential for development of AMN changes with these disease entities and that blood count studies may be indicated in cases of AMN without apparent underlying cause.
Contributors Study design: AAF, LMJ, MRM; data collection: MRM, LMJ, ECS, GCdA, DDE, DS, AAF; data interpretation: MRM, AAF, LMJ, DS; writing the draft: MRM; critical review and approval: MRM, LMJ, ECS, GCdA, DDE, DS, AAF.
Funding This work was supported in part by an unrestricted grant from the Macula Foundation, New York City and by Research to Prevent Blindness, New York.
Competing interests None declared.
Ethics approval Northwestern University Institutional Review Board.
Provenance and peer review Not commissioned; externally peer reviewed.