Table 2.

A summary of the putative influences on cognitive outcomes in patients with gliomas.

Factor		Evidence/Key Points
Premorbid and patient-related factors	Performance/functional status	Some limited evidence to indicate that this may influence the risk of postoperative cognitive dysfunction (32).
	Comorbidities	Hypothyroidism has been associated with impaired MMSE scores (77). Other medical comorbidities are likely to also be relevant.
	Baseline cognitive status	Likely to be an important predictor of cognitive outcomes (71, 79, 80).
	Genetics	Polymorphisms in COMT, BDNF, and DRD2 genes may be associated with cognitive performance in specific domains (147). IDH mutations may also be associated with improved cognitive outcomes (7). Polymorphisms in inflammation, DNA repair, and metabolism pathways may also be associated with cognitive function (148). APOE status has been shown to be associated with cognitive outcomes in some (126, 131) but not all (150) studies.
	Educational attainment/cognitive reserve	Cognitive reserve is most commonly evaluated via proxy, typically years of formal education, with mostly no effect on cognitive outcomes observed (28, 39, 72, 171), but there are exceptions (71) and it may be relevant for HGG but not LGG (42). Higher education levels may predict responsiveness to cognitive rehabilitation (172). NART IQ, a measure of premorbid IQ, is another proxy measure of cognitive reserve and has been positively associated with cognitive outcomes in some (73, 171) but not all (72) studies.
	Age/brain reserve	Age is the commonest proxy measure for brain reserve, which may (28, 71–73) or may not (38, 61, 74) influence cognitive outcomes. Evidence for and against a role for age in influencing cognition comes from prospective studies. Younger age may predict responsiveness to cognitive rehabilitation (172).
	Gender	Limited evidence. One study indicated improved language performance in females immediately after surgery, but scores were comparable to males one year after surgery (78).
	Functional anatomy	Changes to the white matter tracts and functional networks of the brain (81) and arcuate fasciculus lateralization (82) may influence cognitive function, particularly language outcomes.
	Fatigue	No strong evidence that it influences cognitive outcomes specifically in brain tumour patients, but fatigue is common in patients with brain tumours (162) and is a common side effect of several anti-epileptic drugs (163). Further studies are required.
	Mood disorders	There is some evidence for an association between mood state in cognitive outcomes (8, 26, 64, 74, 169), but whether mood changes influence cognitive outcomes or vice versa remains to be fully elucidated.
Tumour-related factors	Tumour location	Conflicting results, with evidence for (23–35) and against (3, 36–40) a role in cognitive outcomes, but on balance it is most likely important. Likely to depend on the cognitive domain evaluated (41).
	Tumour laterality	Many studies suggest right-sided tumours are associated with a lower risk of cognitive impairment (15, 33, 43–49), sometimes irrespective of the exact location of the tumour within the hemisphere (50), but this may depend on the cognitive modality under assessment (51–55). Other studies disagree with the relevance of laterality (36, 56, 57), but overall tumour laterality is likely to influence cognitive outcomes.
	Tumour grade	Evidence for (4, 32, 42, 54, 62–67) and against (15, 40, 50, 60, 68, 69) a role of tumour grade in influencing cognitive outcomes; others indicate it may be important for some but not all cognitive domains (41). Overall, there is more evidence supporting a role of tumour grade in cognitive outcomes than evidence against a link.
	Tumour volume	Some evidence for a role of tumour volume on cognitive outcomes (4, 54, 60), but there is more evidence against such a role (3, 15, 36, 39, 42, 47, 52, 61). It is likely that this variable has complex interactions with other putative influences on cognitive outcomes, including extent of tumour resection and surgical approach.
	Tumour biology	Molecular and histopathological profiles may (7, 52, 54, 70) or may not (41, 60) influence cognitive outcomes. The conflicting findings may result from differences in the specific tumour characteristics studied and differences in study design.
	Associated clinical features, including: Epilepsy/seizures Cerebral oedema Hydrocephalus	Epilepsy may influence cognitive outcomes but evidence suggests the relationship is driven primarily by the use of anti-epileptic drugs (AEDs; see below) (152). The limited evidence supporting improved cognition in patients receiving corticosteroids are likely to be due to the resolution of cerebral oedema (49) (see below). Hydrocephalus is known to be associated with cognitive dysfunction but the specific relationship in patients with gliomas is poorly studied to date.
Treatment-related factors	Timing of treatment in disease course	Superior cognitive outcomes have been observed in those with LGG treated with radiotherapy or surgery later in the course of their disease compared to those treated at the time of diagnosis (88, 137, 140, 173).
	Surgery	Overall effects of surgery Large variation in findings. Several patterns observed, including: - a deterioration in cognition following surgery often with partial or full recovery or long-term improvement in the ensuing months (3, 24, 33, 41, 47, 48, 64, 71, 78, 86–100) - an improvement in cognition (37, 52, 101–103) - no effect of surgery at all (28, 37, 68, 104–106) - deterioration in specific cognitive domains but no overall cognitive impairment at the group level (108) - a mixed picture: some patients experiencing cognitive improvement and others in the same study who experience cognitive decline or no change (39, 58, 60, 76, 110–113) - or improvements in some cognitive domains and deterioration (2) and/or no change (46) in others Overall, surgery influences cognition negatively initially followed by a recovery over several months in most cases. Surgical approach Limited specific evidence to indicate that this influences cognitive outcomes apart from use of awake craniotomies/intraoperative brain mapping having a positive (93, 118, 119) or negative (57) influence, although there is likely overlap with evidence for tumour location, which is a crucial influence on surgical approach undertaken. Extent of tumour resection A number of studies indicate that higher extent of tumour resection does not negatively influence cognitive outcomes (3, 4, 36, 38, 47, 49, 50, 52, 60), and some evidence it may positively influence cognitive outcomes (116). Complications during or following surgery Evidence to suggest the development of infarcts are associated with worse cognitive outcomes following surgery (93, 114, 115).
	Chemotherapy	Most evidence indicates no effect on cognition (3, 68, 123, 124), but may have a negative (18) or a positive effect (122). Likely to be influenced by several variables including the specific chemotherapy regime chosen and timing of administration.
	Radiotherapy	One of the factors most strongly associated with adverse cognition in patients with brain tumour (32, 39, 129–138), with neuroimaging correlates of the cognitive effects identified (131, 132, 135, 137), but not all studies have found such a relationship (3, 28, 68, 139–141), and in some cases mixed results (61, 142–144) and improvements in cognition following radiotherapy (122) have been noted. Effects of radiotherapy on cognitive outcomes are likely to relate to the specific anatomical structures that have been irradiated and their laterality (136, 138). Studies comparing combinations or radiotherapy and/or chemotherapy treatments suggest no difference in cognitive outcomes according to oncological treatment administered (122, 146), but follow-up in these studies was short.
	Anti-epileptic drugs	AEDs are a well-recognised cause of cognitive decline (36, 46, 49, 131, 133, 153, 154), particularly the older anti-epileptic medications (155), and use of multiple AEDs in a single patient increases the risk further (36). Limited data suggest no influence of seizures or anti-epileptic medications on cognitive outcomes in patients with brain tumours (61, 63, 157), but these studies include limitations of small sample size (157) and lack of comprehensive cognitive assessment (use of MMSE alone) (61). Overall, most evidence supports its role in cognition.
	Steroids	Well-recognised cognitive effects, particularly in studies of patients with systemic conditions requiring steroid treatment, but also in patients with gliomas (161). Some evidence against a role of steroid use in determining cognitive outcomes (63). Evidence of improved cognition with steroid use (49) is likely secondary to resolution of cerebral oedema. Role is likely to be influenced by the presence and extent of cerebral oedema, and the effects of that on cognition.
Other	Specific cognitive tests administered	Many studies identified in the systematic review found impaired performance in some but not all cognitive tests administered to patients. Furthermore, evidence indicates that the rate of recovery of cognitive function after surgery varies by cognitive domain, with language, attention, and executive functions being the slowest domains to recover (93). This highlights the importance of administering comprehensive test batteries to patients to ensure all cognitive changes are captured.
	Timing of test administration	Length of time since diagnosis has been shown in some studies to affect cognitive outcomes in patients with gliomas (125), but not in others (74). This may be in part due to the timing in relation to treatments administered.