This page is an English translation of the Korean original. If there is any discrepancy, the Korean version prevails.
| Item | Details |
|---|---|
| A. Classification codes | ICD-10: I64 (Stroke not specified as haemorrhage or infarction) | Special Copayment Reduction: V191 (Special Copayment Reduction for cerebrovascular disease: 5% copayment for up to 30 days from the registration date) |
| B. Synonyms | Undetermined stroke, cryptogenic stroke, unspecified stroke |
| C. Adult onset | Can occur at any age but is diagnosed mainly in older adults; this classification code applies when imaging has not been performed or hemorrhage/infarction cannot be differentiated on initial examination. |
| 1. History | In the pre-CT era, ischemic and hemorrhagic stroke could not be differentiated during life, and I64 is a residual category for coding this diagnostic uncertainty. |
| 2. Causes | Either an ischemic or hemorrhagic mechanism is possible; applies when classification is impossible due to imaging not being performed, rapid death, uncertainty in image interpretation, etc. |
| 3. Genetics | Depending on the underlying etiology, the genetic background of cerebral infarction or intracerebral hemorrhage applies in the same way; in cryptogenic stroke, genetic testing for inherited thrombophilia and vascular disease is important. |
| 4. General epidemiology | Accounts for about 5~10% of all stroke diagnoses, with a higher proportion in regions with low CT access or in cases of early emergency death. |
| 5. Pathophysiology | Classified as I64 due to diagnostic uncertainty; the actual pathophysiology may be either an ischemic or hemorrhagic mechanism, or a mixed type. |
| 6. Symptoms | The same sudden neurological deficits as in ischemic or hemorrhagic stroke (hemiparesis, speech disturbance, decreased consciousness, etc.) appear. |
| 7. Signs | Neurological signs common to stroke, such as motor, sensory, language, and visual field deficits, appear, and efforts to differentiate by signs specific to hemorrhage or infarction are needed. |
| 8. Screening | Management of stroke risk factors and ensuring early access to CT/MRI are the key strategies for minimizing diagnostic uncertainty. |
| 9. Diagnosis | Perform non-contrast head CT or MRI as quickly as possible to differentiate hemorrhage/infarction, and reclassify to a specific code (I60~I63). |
| 10. Treatment | Before imaging, general supportive care for stroke is provided; after differentiation, treatment is switched to specific therapy appropriate to the stroke type. |
| 11. Prognosis | Prognosis is difficult to predict due to diagnostic uncertainty, and treatment delay negatively affects prognosis. |
| 12. Prevention | Strengthening early imaging diagnosis systems, managing stroke risk factors, and prompt hospital arrival through education on stroke symptoms are key to minimizing use of I64 and to prevention. |
| 13. References | Sacco RL, et al. An updated definition of stroke for the 21st century: a statement for healthcare professionals. Stroke. 2013 … and 9 other papers |
ICD-10 code I64 (stroke not specified as haemorrhage or infarction) is a residual classification code used in stroke diagnosis when ischemic and hemorrhagic stroke cannot be distinguished or differentiating tests have not been performed. Historically, in the pre-CT era (before the 1970s), differentiation of ischemic cerebral infarction from hemorrhagic stroke during life was virtually impossible, and all strokes of this period could only be diagnosed in a category corresponding to the current I64. Until the 19th century, stroke was described simply as a single category, 'apoplexy,' and only the indirect method of finding bloody CSF on lumbar puncture could be used for diagnosing hemorrhagic stroke during life.
Since the introduction of CT in the early 1970s, differentiation of stroke types has advanced rapidly, and the frequency of clinical use of the I64 code has steadily decreased. However, the I64 code is still applied in developing countries with limited medical resources, in settings with limited CT access such as local clinics or emergency departments, in cases of death immediately after onset, or in cases where imaging findings are ambiguous and hemorrhage and infarction are difficult to differentiate with certainty. The concept of cryptogenic stroke is sometimes used interchangeably with I64, but strictly speaking, cryptogenic stroke means a stroke confirmed as ischemic but of unclear cause (cryptogenic in the TOAST classification) and should be classified as I63. In principle, I64 is used only when the stroke type itself—ischemic or hemorrhagic—is unknown.
In international statistical comparison studies, the proportion of I64 use is sometimes used as an indicator of the health care system and CT access. In high-income countries with high CT access, the proportion of I64 use is 2~5% or less of all strokes, whereas some reports from regions with limited medical resources show it reaching more than 30%. This is an important public health indicator reflecting gaps in diagnostic systems. The World Health Organization (WHO) and national stroke societies propose reduction in the proportion of I64 use as an indicator of health care quality improvement and recommend expanding CT access and strengthening stroke care systems for this purpose.
The actual cause of a stroke classified as I64 may be either an ischemic or hemorrhagic mechanism. However, because I64 is a code reflecting diagnostic uncertainty, it is more important to understand the causes of the situation in which diagnostic classification is impossible than the cause itself. The most common situation is imaging not performed: cases in which the patient dies before CT or MRI can be performed, imaging is impossible due to lack of medical infrastructure, or the patient or guardian refuses testing.
The second situation is uncertainty in interpreting imaging findings. Classification of stroke type is difficult when acute MRI or CT shows mixed findings, or when ischemic and hemorrhagic findings coexist, as in ischemic stroke with hemorrhagic transformation. In particular, mixed ischemic-hemorrhagic stroke should clinically be classified as ischemic stroke, but is sometimes miscoded as I64 when the hemorrhagic component is prominent. The third situation is death before hospital arrival or during emergency treatment, making further testing impossible. In such patients, confirmation of stroke type by autopsy is ideal but is rarely performed in practice.
Some rare stroke types do not show clear ischemic/hemorrhagic findings even on CT/MRI and are sometimes classified as I64. Examples include reversible cerebral vasoconstriction syndrome (RCVS), the early phase of venous sinus thrombosis, or acute ischemic stroke with a very small lesion in which the initial CT appears normal. In addition, some medical institutions use I64 as unknown stroke type even for ischemic stroke due to inadequate coding education, which emphasizes the importance of coding quality management.
The genetic background of stroke classified as I64 is the same as that of I63 (cerebral infarction) or I61 (intracerebral hemorrhage), respectively, according to the actual underlying etiology (ischemic or hemorrhagic). However, what is particularly clinically important in the context of I64 is the role of genetic testing suggesting inherited thrombophilia or vascular disease in cryptogenic stroke. In young stroke patients, if an underlying hereditary cause is not found, there is a risk of miscoding as I64, so systematic genetic testing is important.
Hereditary diseases to consider in cryptogenic stroke include CADASIL due to NOTCH3 gene variants, Fabry disease due to GLA gene variants, small vessel disease due to COL4A1/COL4A2 variants, inherited thrombophilias such as Factor V Leiden (F5 R506Q variant) and the prothrombin G20210A variant (F2), antiphospholipid syndrome (acquired but with genetic predisposition), sickle cell anemia (HBB variants), and MELAS due to mitochondrial DNA variants (m.3243A>G). These hereditary diseases are not detected without appropriate testing and sometimes remain as I64.
Genomic studies such as the 2021 STROKE-CHIP study have confirmed associations of cryptogenic stroke with susceptibility loci such as PITX2 (atrial fibrillation-related), ZFHX3, HDAC9, and ABO (blood group-related thrombotic risk), suggesting that some cryptogenic strokes are genetically well-defined subtypes of ischemic stroke. When there is familial clustering of stroke, panel genetic testing or whole exome sequencing (WES) can provide diagnostic value.
The frequency of use of the I64 code is strongly associated with CT/MRI access in the health care system. Worldwide, I64 is used in about 5~10% of all stroke diagnoses, but regional differences are large. In large stroke centers in high-income countries, the rate of I64 use is 2~3% or less, whereas in low- and middle-income countries with limited medical resources it reaches more than 30%. This reflects differences in CT access, maturity of emergency stroke care systems, and level of coding education. According to World Stroke Organization (WSO) data, the proportion of strokes diagnosed without CT/MRI remains high worldwide, affecting mortality and rehabilitation planning.
In Korea, the proportion of I64 use has been steadily decreasing with advances in stroke care and improved CT access. However, the I64 code is still used for patients who die after cardiopulmonary resuscitation in the emergency department, patients who die after out-of-hospital stroke, or cases in long-term care hospitals given only a clinical diagnosis of stroke without imaging. The proportion of I64 use tends to be relatively high among elderly patients, patients outside large cities, and socioeconomically vulnerable groups. Statistically, I64 patients have higher mortality than patients classified as I60~I63, because death before imaging or a severe condition often leads to I64 classification.
In international coding comparison studies, the proportion of I64 use is used as one of the performance indicators of health care systems. In certified stroke hospitals or institutions with stroke units, the proportion of I64 use is markedly lower, showing that appropriate emergency imaging and specialized care systems contribute to accurate stroke classification. Monitoring the I64 proportion in stroke registries is an important method for improving quality of care.
Because I64 is a classification reflecting diagnostic uncertainty, the actual pathophysiology of strokes under this code remains unidentified. However, among cases classified as I64, the largest proportion are those presumed to be ischemic cerebral infarction based on clinical history and risk factors. Considering that most strokes are ischemic, 70~80% of cases with only a clinical diagnosis and no imaging are likely to actually be ischemic cerebral infarction.
The pathophysiology of cryptogenic ischemic stroke is highly heterogeneous. Recent studies using long-term ECG monitoring have confirmed that subclinical/paroxysmal atrial fibrillation is the underlying cause in a substantial proportion of cryptogenic strokes. In 3-year follow-up using an implantable cardiac monitor (ICM), atrial fibrillation was found in more than 30% of patients with cryptogenic stroke. In addition, patent foramen ovale (PFO) is recognized as an important mechanism causing cryptogenic stroke through paradoxical embolism, and PFO closure has been shown to be effective in preventing recurrence, particularly in cryptogenic stroke under age 60.
Hemodynamic ischemia occurs when cerebral perfusion decreases globally due to hypotension, reduced cardiac output, severe vascular stenosis, or the like, and watershed infarction is characteristic. Arterial dissection is an important cause of cryptogenic stroke in young patients and is confirmed as an intramural hematoma on MRI fat-saturated imaging (fat-sat T1). Thoroughly investigating these various mechanisms to minimize I64 classification and provide cause-specific tailored treatment is the goal of modern stroke care.
The clinical symptoms of stroke classified as I64 are the same as those of ischemic or hemorrhagic stroke in I60~I63. Sudden neurological deficit (acute neurological deficit) is the key clinical feature, and FAST symptoms (Face drooping, Arm weakness, Speech difficulty) are the most common. Typical presentations are sudden paralysis or weakness of the limbs on one side, sudden slurred speech or inability to speak, drooping or numbness of one side of the face, and sudden visual disturbance (loss of vision on one side or diplopia).
In hemorrhagic stroke, severe headache, nausea and vomiting, and decreased consciousness tend to be more prominent, whereas in ischemic stroke, headache tends to be relatively milder and focal neurological deficits corresponding to a specific vascular territory tend to be more prominent. However, differentiating ischemic from hemorrhagic stroke by clinical symptoms alone is unreliable, and this is the fundamental reason the I64 code exists. Historically, tools for differentiating stroke type using clinical indicators, such as 'bleeding scores' (Allen score, Siriraj stroke score), were developed, but their sensitivity and specificity are low and they cannot replace CT.
Sudden neurological deficit accompanied by seizure is more common particularly in hemorrhagic stroke (especially intracerebral hemorrhage and subarachnoid hemorrhage); in this case a hemorrhagic cause is clinically more suggested, but confirmation must still rely on imaging. Rapidly progressive decrease in consciousness suggests brain herniation due to massive hemorrhage or large infarction and requires immediate emergency treatment. In such situations, if the patient dies or is treated before imaging, classification as I64 is likely.
Physical examination findings in stroke patients classified as I64 are essentially the same as those of each stroke type in I60~I63. On neurological examination, signs common to stroke, such as motor deficits, sensory disturbance, speech disturbance, abnormal eye movements, and gait ataxia, appear, and attempts are made to differentiate the stroke type from these. However, as mentioned above, it is difficult to reliably differentiate ischemic from hemorrhagic stroke by clinical signs alone.
Blood pressure measurement provides important information; when systolic blood pressure is extremely high at 220 mmHg or more, the likelihood of hemorrhagic stroke is relatively higher, but reactive hypertension can also appear in ischemic stroke, so it is unreliable as a differentiating criterion. Blood glucose measurement is important for excluding stroke-like symptoms due to hypoglycemia; if symptoms improve after correction of hypoglycemia, it is likely not a stroke. If ECG shows atrial fibrillation or myocardial infarction, the likelihood of cardioembolic ischemic stroke increases.
The level of consciousness is assessed with the GCS (Glasgow Coma Scale), and the severity of neurological deficit is quantified with the NIHSS. Signs of meningeal irritation (nuchal rigidity, Kernig sign) suggest subarachnoid hemorrhage or infectious meningitis. Hypertensive retinopathy or Terson syndrome (subhyaloid hemorrhage) found on fundoscopy increases the likelihood of hypertensive stroke or subarachnoid hemorrhage, respectively. The stroke type can be estimated by synthesizing these clinical clues, but imaging is always required for confirmation.
The most effective strategy for minimizing use of I64 is to build medical infrastructure that allows access to CT or MRI immediately upon stroke onset. A stroke center system and emergency stroke protocol that allow head CT within 30 minutes of emergency arrival for all patients with suspected stroke are essential. For this, community education (recognition of FAST symptoms), stroke hotlines, and a rapid 119 transport system are important. In medical institutions without imaging equipment, access to initial CT can be improved by establishing a rapid transfer system and using teleradiology systems.
In terms of risk factor screening and management, preventing stroke itself is the fundamental strategy for not generating the I64 code. Regular screening and active treatment of blood pressure, blood glucose, cholesterol, and atrial fibrillation, together with education on smoking cessation and lifestyle improvement, should take place at the community level. In patients who have already had a stroke, patient education and follow-up systems should be strengthened so that faster hospital arrival and accurate diagnosis can occur upon recurrence.
Quality management of medical record coding also contributes to reducing the proportion of I64 use. Coding education for medical records staff and physicians should reduce cases miscoded as I64 despite available imaging results. The I64 proportion should be monitored regularly in stroke registries and national stroke statistics, and support for improving quality of care is needed for institutions with high proportions.
The most important diagnostic approach to avoid applying the I64 code is to perform non-contrast head CT immediately upon onset of stroke symptoms. CT can clearly exclude or confirm hemorrhagic stroke within about 15~20 minutes, playing a decisive role in determining the subsequent direction of treatment. When there is no hemorrhage on CT and acute ischemic change is suspected, or when CT is normal but clinical suspicion is strong, MRI DWI is performed to confirm acute ischemic infarction.
For differential diagnosis in the I64 situation, clinical scoring systems are used as adjuncts in addition to imaging. The Siriraj Stroke Score and Allen Score are scoring systems that differentiate hemorrhagic from ischemic stroke based on clinical features (level of consciousness, vomiting, headache, markers of atherosclerotic vascular disease), but their sensitivity and specificity are each around 70~80%, insufficient to replace CT. However, they can be used as adjunctive tools for treatment decisions in emergencies when CT is not immediately available.
It is important to reclassify the I64 code to the appropriate code among I60~I63 as soon as the stroke type is confirmed by imaging. In ischemic stroke classified as cryptogenic, systematic etiologic workup is performed, including long-term ECG monitoring (up to 30 days), transesophageal echocardiography (TEE), a thrombophilia genetic testing panel, vascular imaging (CTA, MRA, DSA), and autoantibody testing. Because percutaneous PFO closure is effective in preventing recurrence when PFO is found, PFO is screened with a bubble study on transthoracic echocardiography.
In patients classified as I64—that is, stroke patients in whom it has not yet been confirmed whether the stroke is ischemic or hemorrhagic—general supportive care is started first while imaging is performed promptly. Securing the airway, maintaining oxygen saturation, monitoring vital signs, measuring blood glucose and correcting hypoglycemia, and securing IV fluids are the immediate measures. Antiplatelet agents, anticoagulants, and thrombolytics must not be given before the stroke type is confirmed. Because applying thrombolysis intended for ischemic stroke to hemorrhagic stroke can cause fatal bleeding complications, initiation of reperfusion therapy must be withheld until the imaging results are confirmed.
Before the stroke type is confirmed, blood pressure is not actively lowered unless it is extremely high (systolic 220 mmHg or more). This is because excessive blood pressure lowering in ischemic stroke can damage the ischemic penumbra. If seizures occur, they are controlled with antiseizure medications, and if fever is present, antipyretic measures are taken. As soon as the stroke type is confirmed, treatment is switched to specific therapy for each stroke type (ischemic: thrombolysis, mechanical thrombectomy, antiplatelet agents; hemorrhagic: anticoagulation reversal, blood pressure control, surgical decompression).
Long-term treatment and rehabilitation proceed according to the treatment guidelines for each of I60~I63 after the stroke type is confirmed. In rare cases where the stroke type remains unconfirmed (not identifiable even at autopsy), treatment for preventing recurrence must be decided empirically, and because ischemic stroke is usually more likely, treatment centered on antiplatelet agents and risk factor correction is applied. Intensive rehabilitation by a multidisciplinary rehabilitation team (neurology, physical medicine and rehabilitation, occupational therapy, speech therapy, physical therapy) is essential for functional recovery.
The prognosis of stroke patients classified as I64 varies greatly with the actual underlying stroke type and severity, and diagnostic delay due to absence of imaging negatively affects prognosis. Cases in which the stroke type could not be confirmed because of no imaging results are usually associated with rapid death, severe condition, or lack of access to medical resources, so short-term mortality in the I64 group tends to be reported higher than in I60~I63.
When ischemic stroke is confirmed, prognosis becomes markedly poor if the optimal time window for thrombolysis or endovascular treatment (4.5 hours or 24 hours) is missed because of delayed imaging confirmation. This emphasizes how important prompt imaging is in the I64 classification situation. In hemorrhagic stroke too, delay of early measures such as anticoagulation reversal and blood pressure control leads to hematoma expansion and worse prognosis.
In ischemic stroke classified as cryptogenic, failure to identify the cause prevents establishment of a secondary prevention strategy and increases the risk of recurrence. When occult atrial fibrillation is found through long-term ECG monitoring, switching to anticoagulant treatment can greatly lower the risk of recurrence. In stroke caused by PFO, the recurrence rate decreases significantly after closure, so thorough etiologic workup and tailored treatment are key to improving long-term prognosis.
Minimizing stroke cases classified as I64 is itself an important prevention goal. For this, first, prevention of stroke occurrence is important. Active control of major risk factors such as hypertension, diabetes, dyslipidemia, atrial fibrillation, and smoking reduces stroke occurrence itself. Regular physical activity, smoking cessation, moderation of alcohol, and healthy eating habits contribute to reducing stroke risk. In patients with atrial fibrillation, appropriate anticoagulation is decisive in preventing cardioembolic cerebral infarction, and when treatment is inadequate, the risk of stroke and of I64 coding increases.
Second is education on early recognition of stroke symptoms and prompt hospital arrival. Through public health campaigns in the community publicizing FAST symptoms, people are educated to call 119 immediately when stroke symptoms occur and to be transported to a medical institution capable of imaging. Shortening the time from symptom onset to hospital arrival (onset-to-door time) can reduce situations of death before imaging or inability to treat. Focused education and community support systems are needed especially for vulnerable groups such as the elderly, those living alone, and residents of rural areas.
Third is strengthening medical infrastructure. Increasing CT access in community hospitals and expanding stroke care systems (stroke units, stroke hotlines, rapid transport protocols) nationwide contribute most directly to minimizing use of I64. Using telemedicine and remote image reading (telestroke) systems, a system is built that allows expert interpretation even in regions without imaging equipment. It is also important to reduce cases miscoded as I64 despite available imaging results through medical record coding education.
Sacco RL, Kasner SE, Broderick JP, et al. An updated definition of stroke for the 21st century: a statement for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2013;44(7):2064-2089. DOI: 10.1161/STR.0b013e318296aeca. PMID: 23652265.
Feigin VL, Norrving B, Mensah GA. Global burden of stroke. Circ Res. 2017;120(3):439-448. DOI: 10.1161/CIRCRESAHA.116.308413. PMID: 28154096.
Hart RG, Diener HC, Coutts SB, et al. Embolic strokes of undetermined source: the case for a new clinical construct. Lancet Neurol. 2014;13(4):429-438. DOI: 10.1016/S1474-4422(13)70310-7. PMID: 24646875.
Sanna T, Diener HC, Passman RS, et al. Cryptogenic stroke and underlying atrial fibrillation. N Engl J Med. 2014;370(26):2478-2486. DOI: 10.1056/NEJMoa1313600. PMID: 24963567.
Mas JL, Derumeaux G, Guillon B, et al. Patent foramen ovale closure or anticoagulation vs. antiplatelets after stroke. N Engl J Med. 2017;377(11):1011-1021. DOI: 10.1056/NEJMoa1705915. PMID: 28902593.
Allen CMC. Clinical diagnosis of the acute stroke syndrome. Q J Med. 1983;52(208):515-523. PMID: 6665246.
Donnan GA, Fisher M, Macleod M, Davis SM. Stroke. Lancet. 2008;371(9624):1612-1623. DOI: 10.1016/S0140-6736(08)60694-7. PMID: 18468545.
Molina CA, Alvarez-Sabin J. Recanalization and reperfusion therapies for acute ischemic stroke. Cerebrovasc Dis. 2009;27(Suppl 1):162-167. DOI: 10.1159/000200455. PMID: 19342842.
Sposato LA, Cipriano LE, Saposnik G, et al. Diagnosis of atrial fibrillation after stroke and transient ischaemic attack: a systematic review and meta-analysis. Lancet Neurol. 2015;14(4):377-387. DOI: 10.1016/S1474-4422(15)70027-X. PMID: 25748102.
Kim JY, Kang K, Kang J, et al. Executive Summary of Stroke Statistics in Korea 2018: a report from the Epidemiology Research Council of the Korean Stroke Society. J Stroke. 2019;21(1):42-59. DOI: 10.5853/jos.2018.03125. PMID: 30732444.