Clinical Reasoning: A 50-Year-Old Man With Intracerebral Hemorrhage and Tortuous Retinal Arterioles

A 50-year-old White man presented with an acute-onset (nonthunderclap) dull right-sided retro-ocular headache. His initial blood pressure was 190/110 mm Hg, with no fever or neck stiffness. Neurologic examination identified left sided facial palsy (sparing the forehead), brachiocrural hemiparesis (4/5), and hemiataxia. Funduscopic examination showed no papilledema. Extraocular movements, visual fields, and sensation were normal. He had no known history of arterial hypertension, diabetes, renal, or liver disease. He reported no history of head trauma, smoking, or substance abuse. He used no medications but had not seen his primary care physician in the past 10 years. A head CT revealed an acute intracerebral hemorrhage (ICH) in the right lentiform nucleus.

Intracerebral hemorrhage classifications differentiate cerebral small vessel disease (CSVD) from other secondary etiologies of spontaneous (nontraumatic) ICH.¹ Imaging signs common to all CSVD etiologies are white matter hyperintensities, recent small subcortical infarcts, lacunes, cerebral microbleeds, and dilated perivascular spaces.² Deep perforator (arteriolosclerosis-related) arteriopathy and cerebral amyloid angiopathy (CAA) are the 2 most common causes of CSVD.¹ Other secondary etiologies of spontaneous ICH include macrovascular lesions (arteriovenous malformation, cavernoma, dural arteriovenous fistula, and aneurysm), tumors, hematologic disorders (e.g., severe thrombocytopenia), cerebral venous thrombosis, hemorrhagic infarction, substance abuse, and rare vasculopathies (e.g., moyamoya, vasculitis, and reversible cerebral vasoconstriction syndrome).¹

A head CT-angiogram did not demonstrate a macrovascular lesion. Platelets and coagulation parameters were normal. A brain MRI identified deep and periventricular white matter hyperintensities (Figure, A), predominantly deep cerebral microbleeds, and a previous ICH in the left lentiform nucleus (Figure, B). Diffusion-weighted imaging was negative for an acute brain infarct. Gadolinium-injected sequences showed no parenchymal or meningeal enhancement. The brain MRI findings suggested that the ICH was most likely secondary to CSVD.

Deep perforator arteriopathy is the most frequent cause of CSVD. Risk factors for deep perforator arteriopathy include age, arterial hypertension, diabetes, smoking, excess adiposity, and alcohol overuse. In deep perforator arteriopathy, imaging signs of CSVD are predominantly found in deep brain regions (basal ganglia, thalamus, and pons). In a patient aged 50 years or older with spontaneous ICH, a probable diagnosis of sporadic CAA can be made with 2 strictly lobar hemorrhagic lesions or 1 strictly lobar hemorrhagic lesion and 1 white matter feature (severely dilated perivascular spaces in the centrum semiovale or white matter hyperintensities in a multispot pattern).³ In a patient with CSVD, acute cortical subarachnoid hemorrhage and chronic superficial siderosis are highly specific for CAA.⁴ Rarely, CSVD is secondary to a monogenic disease, radiotherapy, small vessel cerebral vasculitis, or iatrogenic CAA.⁵

In the absence of previous medical follow-up, it was unknown whether the patient's elevated blood pressure was reactive to the acute ICH or secondary to chronic arterial hypertension. His creatinine level was elevated at 280 µmol/L (normal: 50–110). Both kidneys were small on echography (9.6 and 9.3 cm, normal: 10–14 cm), suggesting chronic renal disease. In addition, the patient had 2 renal cysts. A transthoracic echocardiogram identified left ventricular concentric hypertrophy. His glycated hemoglobin was 5.3% (normal: 4.0–5.6). Fundus examination showed arteriolar tortuosity involving several blood vessels (Figure, C and D).

Considering the elevated blood pressure and extracerebral anomalies compatible with hypertension-related target organ damage, the most likely diagnosis was deep perforator arteriopathy in the setting of previously unknown arterial hypertension. However, given the severity of CSVD, the patient's young age (≤50 years),⁶ and the retinal arteriolar tortuosity, the possibility of a monogenic cause of CSVD was explored.

Well-established monogenic causes of CSVD are summarized in Table 1. On further questioning, the patient had no family history of stroke, early-onset dementia, migraine with aura, or unexplained CSVD. He was born to nonconsanguineous parents. He had no history of motor or mood disturbance, cognitive impairment, or migraine with aura to support the diagnosis of CADASIL. His white matter hyperintensities spared the anterior temporal lobes and external capsules, which decreased the likelihood of (but did not exclude) CADASIL.

Hemorrhagic presentations of monogenic CSVD are more common in COL4A1/2-related disorders and hereditary CAA. Intracerebral hemorrhage can occur in CADASIL, CARASIL, CARASAL, and Fabry disease but is less common than small subcortical infarcts. In the patient, hereditary CAA was unlikely due to the predominantly deep signs of CSVD. On the contrary, in COL4A1/2-related disorders, ICH and cerebral microbleeds predominantly involve deep brain regions, particularly the basal ganglia.⁷ The pattern of CSVD in COL4A1/2-related disorders can mimic deep perforator arteriopathy. Retinal arteriolar tortuosity, while not always present, is an important finding that strongly suggests the diagnosis of COL4A1/2-related disorders.⁸

A targeted molecular screening of all known CSVD genes was performed and identified a heterozygous variant of unknown significance affecting a splice site in intron 16 of the COL4A1 gene (NM_001845:c.903 + 1G > A).

Criteria that can support the pathogenicity of a variant include its nature (e.g., null variant in a gene where loss of function leads to disease, de novo variant in a patient without family history of the disease), increased prevalence of the variant in affected individuals compared with controls, cosegregation of the variant in multiple affected family members, in vitro or ex vivo functional studies that show a damaging effect of the variant on the gene or its product, and computational evidence (in silico algorithms).⁹

The COL4A1 variant identified in the patient was absent from variant databases. However, the involvement of a splice site increased the likelihood that the variant was pathogenic. Because COL4A1 is expressed in skin fibroblasts, the effect of the variant was evaluated with functional studies performed on a skin biopsy sample. The latter confirmed that the variant leads to the skipping of exon 16, resulting in the loss of 5 glycine residues in the triple helix domain of the COL4A1 protein. This result confirmed the pathogenic nature of the c.903 + 1G > A variant and the diagnosis of a COL4A1-related disorder.¹⁰

COL4A1/2-related disorders are an AD monogenic disease in which heterozygous pathogenic variants of COL4A1 or COL4A2 genes disrupt the triple helix formation of type IV collagen. Most pathogenic variants affect glycine residues in type IV collagen.⁸ The systemic manifestations of COL4A1/2-related disorders are related to the presence of type IV collagen in several organs.

Classical phenotypes in COL4A1/2-related disorders are CSVD-related ICH,¹¹ familial porencephaly (fluid-filled cavities in the brain caused by antenatal ICH), hereditary angiopathy with nephropathy (glomerulopathy, cysts, hematuria, renal insufficiency, aneurysms, and muscle cramps syndrome),¹² and eye disease (isolated arteriolar tortuosity, anterior segment anomaly [Axenfeld-Rieger] with congenital iris defects, corneal opaqueness, and glaucoma). Other manifestations include liver cysts, Raynaud phenomenon, and supraventricular cardiac arrythmias. The expressivity of the disease is extremely variable, with overlap between phenotypes. In the patient, the disease presented with CSVD-related ICH, nephropathy with cysts, and retinal arteriolar tortuosity. After the diagnosis, the patient reported a history of unexplained muscle cramps during childhood. The nephropathy may have caused arterial hypertension, which in turn may have contributed to the CSVD in addition to the COL4A1 pathogenic variant.

Two other diseases related to pathogenic variants in COL4A1 and COL4A2 present with CSVD and usually lead to small subcortical infarcts rather than ICH. In pontine autosomal dominant microangiopathy and leukoencephalopathy (PADMAL), pathogenic variants in the 3′ UTR region of COL4A1 lead to upregulation of COL4A1, with infarcts that predominantly affect the pons.¹³ Duplications of COL4A1 and COL4A2 also lead to an upregulation of gene expression and present with ischemic rather than hemorrhagic CSVD.¹⁴

Diagnosing monogenic causes of CSVD is important for management. Owing to reports of ICH after minor head trauma in patients with COL4A1/2-related disorders, activities at risk for head trauma should be avoided.^10,11 Strict control of blood pressure is necessary to potentially slow the progression of CSVD and decrease the risk of stroke. Patients should be screened for extraneurologic manifestations of the disease. If patients have comorbidities that require the use of antithrombotic drugs, the latter may increase the risk of ICH and should only be prescribed after careful weighing of their benefits and risks. Based on expert opinion, intravenous thrombolysis is not recommended in acute ischemic stroke related to COL4A1/2-related disorders.¹⁰ In fetuses that inherit a COL4A1 or COL4A2 pathogenic variant, a cesarean delivery may be considered to avoid the risk of head trauma during vaginal delivery and secondary ICH.^10,11 Genetic counseling is necessary, and screening of family members should be discussed.¹⁰

This report emphasizes the need to consider monogenic causes of CSVD in patients with suggestive features, even if signs of chronic arterial hypertension are present. COL4A1/A2-related disorders can mimic deep perforator arteriopathy and should be considered in patients with deep ICH, CSVD, and retinal arteriolar tortuosity. Collaboration between neurologists, neuro-ophthalmologists, neuroradiologists, and geneticists is key to diagnosing and managing monogenic causes of CSVD.