Clinical manifestations

Presentation and course

Many of these collections are entirely asymptomatic and are noted incidentally on neuroimaging studies. Some may be associated with a large or rapidly-growing head. A retrospective and population-based study of 176 children with benign external hydrocephalus determined that most infants were born with normal head circumference that rapidly increased to abnormally high values before 6 months of age (106). Still others may be accompanied by signs of increased intracranial pressure, including irritability, lethargy, vomiting, split cranial sutures, tense and bulging anterior fontanel, and dilated scalp veins. If they cause mass effect, abnormalities may be noted on neurologic examination, or the infant may present with seizures. It has only been with the advent of CT, and then later MRI, that accurate delineation of these collections could be done, making interpretation of pre-imaging data difficult at best.

Prognosis and complications

Most published case series of children with benign enlargement of subarachnoid spaces of infancy (subarachnomegaly) report that these children have relatively normal development without treatment. Although some older reports reported poor outcome, including risk for hydrocephalus, and advocated shunting (93; 70), these data are based on cases where quality imaging was not available, and most likely these children had subdural hematomas rather than benign subarachnomegaly. Ment and colleagues reported 17 of 18 infants had no evidence of neurologic abnormality, and the 1 child that showed abnormalities was preterm and had suffered multiple other CNS insults (62). They termed the condition “benign enlargement of the subarachnoid spaces.” Furthermore, they demonstrated that the enlargement spontaneously improved on follow-up imaging, as previously reported (75). Alvarez and colleagues reported that most children with this condition had normal development, although more than half had a history of gross motor and language delay; however, all of the delays were transient and resolved by 15 to 18 months of age (06). Furthermore, they found a family history of macrocephaly in 88% of these infants, suggesting that benign familial macrocephaly and benign enlargement of the subarachnoid spaces were related. They also expressed concerns that reports of surgery for this condition were being published and, given the benign and self-limiting nature, advocated deferring treatment unless there were other risk factors for hydrocephalus. Reports of additional infants with benign enlargement of the subarachnoid spaces demonstrated that by age 2 years the excess fluid resolved, mild developmental delay was common but self-limited, family history of macrocephaly was common, and children often remained macrocephalic (81; 08; 37; 68; 34; 77; 87; 47). A study that looked at long-term outcome showed overall good outcome with these patients, with a slight increase in attention deficits with antidotal histories of learning disabilities and behavior problems, but intelligence was within normal range (66). Halevy and colleagues looked at infants with a mean age of 9 months with this condition using the Mullen Scales of Early Learning, and the only difference from controls was hypotonia (transient in 9/20 and persistent in 2/20) (36). A review from Khosroshahi and Nikkhah further suggests that children with benign enlargement of the subarachnoid space ultimately have normal final developmental status (44).

As mentioned above, benign enlargement of the subarachnoid spaces (subarachnomegaly) has been reported to be associated with a higher risk for subdural hematoma after minor head trauma or spontaneously (43; 51; 73; 76; 80; 38; 99; 94). Of note, a report demonstrated that children with benign enlargement of subarachnoid spaces generally had normal to large head sizes at birth, presented with macrocephaly, and had good outcome. A second group of children in the same paper who had enlarged subarachnoid spaces and chronic subdural hematomas also did well, suggesting that even if these children have a higher risk for subdurals with minor trauma, most will self-resolve without treatment (38). In fact, a study by Greiner and colleagues showed that review of 168 imaging studies done for macrocephaly demonstrated 108 with enlarged subarachnoid spaces and 6 with subdural hematoma; however, the authors caution that nonaccidental trauma should still be considered in these children (35). In contrast, a retroactive study of 110 children showed no association between the subdural hematoma and benign enlarged arachnoid spaces, challenging the common hypothesis. The study did, however, find that susceptibility is increased for subarachnoid/subpial hematomas (31).

Arachnoid cysts are developmental anomalies and are the most common intracranial cysts. They can be associated with macrocephaly and hydrocephalus in infants and can also cause mass effect, leading to seizures and abnormal findings on neurologic examination. Although most cysts are static in size, they may change in size (increase or decrease) and may rupture spontaneously or after minor trauma; rupture can lead to subdural hematoma and require surgical intervention (71; 17; 22; 105). A study by Mattei and colleagues demonstrated that children with enlarged subarachnoid spaces during infancy are more likely later to develop arachnoid cysts, suggesting a related process of abnormal arachnoid development and CSF absorption (57).

Outcome for subdural hematomas depends on etiology and age at presentation, which relates to the extent of associated cerebral damage. Traumatic acute subdural hematomas have the worst outcome because there is often diffuse damage to the developing brain; many of these children will develop spasticity, seizures, and developmental delay (89; 03). Older patients tend to suffer more focal damage, although they also have less ability to recover from brain injury. Chronic subdural hematomas secondary to shunt placement usually have good outcome (89; 32). Subdural effusions secondary to meningitis have been reported not to affect outcome (86; 29), although changes in etiology over the past 2 decades has resulted in more cases requiring surgical intervention and worse outcome (98). Subdural empyemas require prompt diagnosis and treatment to achieve good outcome (96; 54).

Clinical vignette

This full-term infant boy had a normal physical examination. His head circumference was 37 cm, which is approximately 75th percentile. His father had a large head, as did a number of paternal relatives. At 4 months of age, his head circumference had grown to 44 cm (about 90th percentile).

By 6.5 months, his head circumference was 48 cm (well above 99th percentile).

The anterior fontanel was slightly enlarged but soft, and there was no splaying of the sutures. He was normal developmentally except for mild gross motor delay (sat without support at 8 months, and pulled to stand at 13 months). No treatment was considered necessary.

During follow-up, he remained neurologically normal. Head growth slowed and paralleled the growth curve. There was never any evidence of elevated intracranial pressure or abnormalities on neurologic examination. At age 3 years, his development was completely normal and head circumference was 52.5 cm (about 95th percentile). He was diagnosed with “benign enlargement of the subarachnoid spaces of infancy.”

Biological basis

Etiology and pathogenesis

The most common cause of excessive head growth in infants is hydrocephalus. However, it has long been recognized that another group of children have abnormally large heads and frontal fluid collections on imaging and follow relatively benign courses. Review of the literature is confusing because many different terms have been used to describe these children (102; 89; 72; 32; 02; 30); parenthesis in the list of terms below indicate that the name has been used either with or without those words as part of the name:

(Benign) external hydrocephalus
Benign extraaxial fluid collections of infancy
Benign expansion of the subarachnoid spaces
Idiopathic external hydrocephalus
(Benign) extracerebral (fluid) collections (of infancy)
(Cryptic or benign) pericerebral fluid collections of infancy
Primitive megalencephaly
Benign obstructive hydrocephalus
(Benign) subdural effusions
(Benign) subdural hematomas
(Benign) subdural fluid collections of infancy
Subdural fluid collections
Pseudohydrocephalus megalencephaly
Extracerebral intracranial fluid collections of childhood
Benign enlargement of subarachnoid space(s) (of infancy)
Benign communicating hydrocephalus
Benign subdurals of infancy
Extraventricular hydrocephalus
Subarachnomegaly

It is now known that the extracerebral fluid is CSF and resides in the arachnoid space. These children do not have subdural fluid, nor do they have hydrocephalus, and these terms should no longer be used when describing these patients. It has been advocated to use the terms “benign enlargement of the subarachnoid spaces of infancy” or “subarachnomegaly.” The diagnostic criteria include: macrocephaly (greater than 98th percentile), enlargement of the subarachnoid space in the frontal or frontoparietal regions symmetrically, no to mild ventriculomegaly, and normal development or perhaps minor gross motor delay. There is often a family history of macrocephaly, and this condition is likely part of the spectrum of benign familial macrocephaly (72; 32; 30; 14).

Transient enlargement of the subarachnoid space is also common immediately after shunt insertion in the hydrocephalic infant (78). However, in the presence of an open fontanel, such enlargement is usually mild and transient. Various chromosomal and metabolic disorders may also result in subdural effusion or enlargement of the subarachnoid space (72). CSF may enter the subdural space through an arachnoid tear caused by rapid ventricular decompression during shunt placement (103) and also may occur after trauma (32). Bacterial meningitis, especially with H influenza, can result in subdural effusions in infants (86; 29). If these effusions become infected, or if pathogens from a nonmeningitic source reach the subdural space, subdural empyema ensues (98). Although any malignancy may cause a subdural effusion, they are rare and usually seen only in histiocytoses (108). Acute subdural hematoma, then evolving into chronic subdural hematoma, follows head trauma (89). Subdural hygroma (CSF mixes with blood) most commonly forms after opening of the arachnoid, such as during placement of a ventriculoperitoneal shunt (32). Arachnoid and related extracerebral cysts are developmental in origin (22).

The width of the normal subarachnoid space varies considerably during infancy (52) and is especially wide earliest in life (50). It is entirely possible that many infants diagnosed with subarachnoid fluid collections may just be normal variants (45). The size of the subarachnoid space decreases sharply around the gestational age of 32 weeks and remains at a median of 2 mm until birth; however, it increases again after birth, reaching a peak at around 7 months, and then again decreases (100). This reduction in subarachnoid space in the fetus is associated with rapid myelination of the brain, suggesting that alterations in brain water content may be related. Normal values for arachnoid space are less than 4 to 8 mm in the first year of life, and less than 5 mm in the second year (77; 38). The most rapid period of head growth is also during the first year of life; thus, enlargement of the subarachnoid space likely drives increases in head size. Width of the subarachnoid space has been shown to be associated with growth of the overlying skull in both term (50) and preterm infants, especially during periods of rapid head growth (04). Focal widening of the subarachnoid space in infants with craniosynostosis has been reported, hypothesized to be due to increased CSF pressure at the sites of skull fusion causing preferential redistribution of CSF into more expandable skull areas (20). Expanded subarachnoid spaces have also been reported as more common in infants with occipital plagiocephaly, and it was speculated that dilation of the subarachnoid space was the primary event that secondarily drove skull deformation (83). Bateman and colleagues demonstrated in 3 infants by magnetic resonance venography that collapsible sinuses and venous outflow obstruction could be related to arachnoid fluid accumulation (11).

As mentioned above, reports on prevalence, treatment, and outcome of infants with extracerebral fluid collections are difficult to reconcile, given that many different terms were used to describe the same fluid collections, or alternatively different collections were lumped together under the same name. For example, lumping infants with subdural collections and hydrocephalus with the more benign subarachnomegaly was common; in large part this was due to difficulties discerning the various fluid collections by ultrasound and CT. Clearly, infants with enlarged subarachnoid spaces (“subarachnomegaly” or “benign enlargement of the subarachnoid spaces of infancy”) represent a different population; thus, the old terms of “benign subdural collections of infancy” (16) and “idiopathic external hydrocephalus” (06) should be abandoned. Proposed mechanisms for enlargement of the subarachnoid space include reduced CSF uptake due to immaturity of the arachnoid granulations and disproportionate growth of the skull faster than the brain (38; 32). Another study gives a different hypothesis based on the finding that subarachnoid space fluid diffusivity is reduced in patients with enlarged subarachnoid spaces of infancy. Whitehead and colleagues suggest there is an unbalanced osmotic/hydrostatic pressure in the frontotemporal subarachnoid spaces that is leading to the development of symptoms (101). Sainz and colleagues hypothesized that venous hypertension resulting in increased intracranial pressure, acceleration of head growth, and resulting enlargement of the cortical subarachnoid is a possible source for enlargement of the subarachnoid spaces (82). There is also a genetic component as it is much more prevalent in infants whose parents have large heads (06; 30; 44). Benign familial macrocephaly is thought to be inherited in an autosomal dominant fashion and is diagnosed by macrocephaly in multiple family members without a primary cause, such as hydrocephalus, or genetic or metabolic causes, such as achondroplasia, fragile X, and others (10; 26). Reports of enlarged arachnoid spaces in twins and triplets support the autosomal dominant inheritance of this condition (23; 19; 33).

Accumulation of fluid in the subarachnoid space is common because it is a real space; however, the subdural space in actually a misnomer because it is only a potential space, and fluid collection into this space is always abnormal (55). Acute subdural hematomas are almost always due to trauma and derive from bleeding from disrupted bridging veins. Subdurals can be composed of blood or may be a mixture containing CSF if the arachnoid membrane is torn. This blood coagulates and clots and is slowly resorbed; this evolving mixture of resorption products has been called various names in the literature (eg, subdural hematoma, chronic subdural, subdural hygroma, subdural effusion, subdural hemorrhage, and hygroma), but in principle they are all variants of chronic subdural hematoma. Acutely, the protein content is high, but over time it progresses to look like CSF on MRI. If the fluid becomes purulent during this process, it is then referred to as subdural empyema (32; 14).

It is worth noting that enlargement of the subarachnoid space has been reported to be associated with a higher risk for subdural hematoma after minor head trauma, postulated to be due to increased venous stretching (43; 51; 73; 76; 80; 38; 35). However, it has also been argued that enlargement of the space reduces relative brain and skull displacement and, thus, is actually protective (79). Subdural hematoma has also been suggested to occur spontaneously in the setting of subarachnomegaly, macrocrania, or severe dehydration (99). A theoretical basis for acute subdural hematomas after trauma is stretching of traversing cortical veins and their known vulnerability in the subdural space (104; 73). Alternatively, those occurring spontaneously have been proposed as secondary to the bleeding from the inner dural plexus, which is prominent during infancy and then undergoes regression (55). Also, glutaric aciduria type I can present with bilateral subdural hematomas and macrocephaly. This condition is usually symptomatic, in contrast to the asymptomatic presentation of benign enlargement of subarachnoid spaces (13; 72; 14).

Arachnoid cysts are developmental anomalies whereby CSF is trapped between layers of the arachnoid membrane. They are the most common intracranial cysts. In younger children (younger than 2 years old) they can be associated with macrocephaly and hydrocephalus. Arachnoid cysts can also cause mass effect, leading to seizures and abnormal findings on neurologic examination. Although most cysts are static in size, they may change in size (increase or decrease) and may rupture spontaneously or after minor trauma (17; 22; 105). Furthermore, a common etiology between arachnoid cysts and enlargement of the subarachnoid space has been proposed (57).

Epidemiology

Macrocephaly without hydrocephalus was found in 1% of children referred to a developmental clinic (69). A Swedish study found 0.5% of the population had macrocephaly (74). A study showed that the prevalence of macrocephaly can vary by age--only 67% of fetuses diagnosed with macrocephaly had enlarged heads at birth, and in later childhood only 35% still had large heads (12). Among pediatric patients subjected to MRI due to macrocephaly in the literature, between 58% to 75% of children imaged also met criteria for benign external subarachnoid spaces (94).

“Subarachnomegaly” or “benign enlargement of subarachnoid spaces of infancy” is usually diagnosed in the first year of life, usually due to a combination of rapid head growth and large head; although these children often have moderately large heads at birth, they usually become macrocephalic between 6 and 12 months of age. At that point, the rate of head growth normalizes and parallels the normal curve. Although the excess fluid usually resolves, most continue to have large heads as adults, especially if it runs in the family. It is slightly more common in males (30). A retrospective study of 88 children with macrocephaly demonstrated that 77.5% of them had subarachnomegaly (61). It is also more common in infants born prematurely (38). A more recent study showed that macrocephaly in Korean children was more common at birth, and although the etiologies were more diverse, it was a benign finding that did not require intervention (41).

Arachnoid cysts are the most common intracranial masses, representing about 1% of all intracranial masses. They can be found in about 0.5% of the population. They are often asymptomatic and are found on routine imaging for other reasons. They can be identified even on intrauterine imaging. In childhood, they are more common in boys, but in adulthood, they are equally prevalent in men and women (17; 22; 105). As mentioned above, there may be a relationship between arachnoid cysts and enlargement of the subarachnoid space (57).

Prevention

Subdural hematomas are most commonly secondary to nonaccidental trauma; thus, they are preventable through education. Subdurals can also occur after placement of shunts for hydrocephalus, and improved operative technique may lead to reduction in postoperative complications. A survey estimated the annual incidence of subdural hematomas among British and Irish infants aged 0 to 1 year at 24.1 per 100,000; most resulted from inflicted head injury (39). Subarachnomegaly and arachnoid cysts have no identifiable cause.

Differential diagnosis

Confusing conditions

Extracerebral collections may initially look like cerebral atrophy, but neuroimaging can easily distinguish them (62; 43; 58; 32). Infants with cerebral atrophy demonstrate: globally increased fluid around the brain with widening of the sylvian fissures; normal to small head size with normal to low rate of growth; normal- to large-sized ventricles; abnormal neurologic examination or development (16; 62; 06; 38; 30). Other causes of enlarged subarachnoid spaces include genetic conditions: mucopolysaccharidoses types I (Hurler), II (Hunter), and III (Sanfilippo), achondroplasia, Sotos syndrome, and glutaric aciduria type I (72; 14; 42). However, history, physical examination, and imaging findings will usually be abnormal. Diagnostic criteria that make “benign enlargement of the subarachnoid spaces of infancy” or benign “subarachnomegaly” more likely are: family history of macrocephaly; normal development (through 18 to 24 months of age); no neurologic abnormalities on examination; no neurologic disease; symmetric bifrontal enlargement of the subarachnoid spaces, widened interhemispheric fissure, normal to mildly dilated ventricles on imaging, and resolution of these imaging findings by 2 years of age (72; 32; 30). A study demonstrated that diffusion tensor imaging can distinguish between children with benign enlargement of subarachnoid spaces and controls (88).

Prior to the advent of MRI, CT and ultrasound were used to distinguish the various fluid collections. In retrospect, much of the data published before the ready availability of MRI are difficult to interpret, given that it was not always possible to accurately define the location of the extraaxial collection; this is why so many terms were used to define the same patient population, including many that incorrectly described it as hydrocephalus or subdural fluid. Certainly, the differential for extracerebral fluid collections includes subdural hematomas and chronic subdural as well as arachnoid cysts. Furthermore, mega cistern magna can sometimes be confused with a posterior fossa arachnoid cyst. MRI has dramatically improved the ability to accurately distinguish these various collections.

Diagnostic workup

Neuroimaging is the key to diagnosing the type and size of an extracerebral collection. However, a careful developmental examination and history are essential, as are serial head circumference measurements.

Ultrasound (85; 92) or CT (32) will localize the collection and show its size with great sensitivity. The first step to consider is brain sonography via anterior fontanel, as it is fast, safe, and inexpensive. Improvements in technique and equipment have allowed for increased accuracy in evaluation of ventricular size, extracerebral fluid collection, and other structural abnormalities (44).

In general, CT and ultrasound are far less accurate than MRI in determining whether the subdural or subarachnoid space is involved (102; 09; 63; 05; 32; 48). Intravenous contrast will enhance cortical veins and show them coursing through a dilated subarachnoid space; veins are flattened against the cortical surface by subdural effusions (49).

MRIs show the same phenomena without the need for contrast injection, demonstrating the flow-void sign; vessels within the fluid space are seen with enlargement of the subarachnoid space (59; 09; 49; 05). Color Doppler ultrasound can also show the transiting cortical veins and may be useful in the neonate unable to undergo neuroimaging (21; 107).

In benign enlargement of the subarachnoid spaces of infancy, the fluid is isodense with CSF on MRI (102; 09). The fluid is most prominent in the frontal, parietal, and interhemispheric regions; ventricles are normal to slightly enlarged (08; 102; 09; 77). The upper limit of normal for the frontal subarachnoid space varies but many suggest greater than 6 to 8 mm in the first year of life and greater than 4 mm in the second as abnormal (77; 05; 38; 30).

MRI of most subdural effusions shows intensity differences between the contents and CSF in at least 1 imaging sequence; a thin inner layer of CSF intensity is common (102; 09; 65). CT can demonstrate 90% of subdural fluid collections whereas MRI is estimated to be 100% sensitive; furthermore, MRI can provide additional information about the consistency of the fluid and better distinction between the fluid collection and CSF (32). The diagnoses of subdural empyema, chronic subdural hematoma, subdural hygroma, and arachnoid cysts are dealt with in separate clinical summaries.

The reader is referred to several reviews regarding the imaging guidelines in macrocephaly, hydrocephalus, and increased intracranial pressure, and how to distinguish benign enlargement of subarachnoid spaces (97; 56).

Management

The most important management question is whether treatment is needed at all. In the case of benign enlargement of the subarachnoid spaces, treatment is rarely necessary other than referral for early intervention if any developmental delays are noted. In the past, surgical treatment was sometimes done, although it is unclear whether these children had subarachnoid or subdural fluid collections given the lack of consistency in terminology and imaging. Furthermore, much of the rationale to treat was based on concerns about poor outcome, but subsequent data have demonstrated that these children do well without any treatment (62; 08; 06; 37; 68; 34; 77; 87; 47; 38; 36). Some of the authors in these reports only mentioned treatment with diuretics, but there are no data that this is either necessary or beneficial. The rare infant with progressive external hydrocephalus responds to temporary subarachnoid drainage (Eidlitz-Markus et at 2003); however, these children will have obvious neurologic abnormalities, extremely abnormal rate of growth of head size, bulging fontanels, splayed sutures, and other signs of increased intracranial pressure, and their conditions will likely represent an early form of hydrocephalus rather than benign enlargement of the subarachnoid space. A study demonstrated that intracranial pressure monitoring may be useful in children with macrocephaly who have both enlarged ventricles and subarachnoid spaces to distinguish which patients would benefit from ventriculo-peritoneal shunting (84).

Arachnoid cysts also rarely require treatment--watchful waiting is the best option, and surgery is only indicated if the child becomes symptomatic (17; 22; 105). Arachnoid cysts are often found in patients with headaches or seizures, but cause and effect is difficult to prove, and treatment does not always show benefit.

In general, management strategies are similar in infants with subdural effusions (chronic subdural hematomas). No treatment is necessary unless there is evidence of increased intracranial pressure, extremely rapid head growth, or neurologic symptoms such as seizures or mass effect (89; 32). When treatment is needed for symptomatic subdural effusion, there are several options. Medical therapies are usually ineffective or have transient benefits at best. Surgical choices include subdural puncture with fluid drainage (repeated as needed), burhole drainage, and subdural shunt procedures (89; 32). Too rapid drainage by any route can cause dangerous hypotension (95). With fast overnight drainage, a superficial subarachnoid CSF space expansion may form (91). Needle aspiration and burhole drainage are associated with high recurrence and infection rates and are no longer considered standards of care (90; 32). Although complications are not rare, subdural shunting seems to yield the best results (53; 28; 89; 18; 32), although not all authorities agree (64). Even when chronic subdurals are bilateral, usually only a unilateral shunt is needed; furthermore, the patient is rarely shunt-dependent, and removal of the shunt at a later date should be considered (89; 46; 32).

Subdural effusions in hydrocephalic infants following shunt insertion represent a special category of cases, and can be very difficult to treat. If the subdural develops in the chronic situation, one effective treatment is to drain the subdural collections and change the valve pressure (01; 89). If an acute subdural hematoma develops, emergency craniotomy may be needed. Treatment of acute subdural hematomas of an etiology is an emergency and beyond the scope of this review.

Subdural effusions that develop after meningitis self-resolve, and usually no treatment is needed. Subdural empyemas are emergencies and require appropriate antibiotics and prompt surgical intervention (Vankatesh et al 2006; 98).

Special considerations

Anesthesia

Special precautions are needed in the presence of increased intracranial pressure. Inhalation agents can accumulate in the subdural space if it is opened during an operative procedure. Every attempt should be made to limit residual air in the subdural space at closure.