Preclinical Investigation

In 2004, Eckenhoff et al. published the finding that the inhaled anesthetic isoflurane enhances peptide pathophysiology (oligomerization of amyloid-beta) associated with Alzheimer’s disease (AD).^[1] This investigation began after observations of postoperative cognitive dysfunction after surgery and anesthesia.^[2] Since that time, several investigators have examined the cellular and molecular links between anesthetic agents and neurodegeneration. Early preclinical research began with in vitro cellular studies examining the pathophysiologic architecture of anesthetics and neurotoxicity, and investigation has now expanded to cognitive testing in animals. Various intravenous and inhalational agents have been studied, and the reader is referred to the specific references provided for further information.

Isoflurane. Isoflurane is a volatile anesthetic agent that has been used clinically for many decades. Given the common and long-standing use of isoflurane, much of the anesthetic neurotoxicity research started with this agent. As mentioned above, Eckenhoff et al. first demonstrated enhanced amyloid-beta (Aβ) oligomerization with isoflurane in vitro with pheochromocytoma cells.^[1] In addition to increased Aβ production and oligomerization, further in vitro investigation has revealed that isoflurane induces loss of cytosolic calcium regulation,^[3] mitochondrial damage,^[4] and metabolic stress on the endoplasmic reticulum (ER).^[5] Each of these pathways has been demonstrated to result in cellular apoptosis. As research has progressed, preclinical investigation of isoflurane has also involved cognitive testing in animal models. Isoflurane exposure has indeed been associated with behavioral changes, spatial memory deficits, and learning impairment in a rodent model.^[6] In one study with aged rats, no cognitive deficits were detected 4 months after isoflurane exposure, suggesting that isoflurane-induced cognitive impairment may be a transient phenomenon.^[7] A review of pertinent literature is available in Table 7.1.

Sevoflurane. Sevoflurane is another halogenated volatile anesthetic agent also commonly used in clinical practice. Sevoflurane neurotoxicity studies have demonstrated similar results compared to those involving isoflurane. For example, sevoflurane has also been shown to promulgate Aβ pathophysiology, induce caspase activation, and increase neuroapoptosis with both in vitro and in vivo investigation.^[8] Additionally, neonatal mouse sevoflurane exposure has also been linked to cognitive impairment and abnormal social behaviors.^[9] Interestingly, however, environmental enrichment has been shown to reverse sevoflurane-induced memory impairment in young mice.^[10] This study implies that learning environment may enhance synaptic plasticity and mitigate anesthetic-induced neurotoxicity. Certainly, further investigation is warranted. Studies representative of the current literature on sevoflurane neurotoxicity are outlined in Table 7.2.

Desflurane. Desflurane, the newest volatile anesthetic agent available, has demonstrated clinical value due to its fast emergence profile from anesthesia. It follows that desflurane may leave a different neurophysiologic “signature” on the brain compared to other inhalational agents, such as isoflurane. As such, laboratory studies have directly compared isoflurane and desflurane, and there is evidence that desflurane may have a less neurotoxic profile compared to isoflurane.^[4] When recently compared directly, isoflurane – but not desflurane – was shown to induce cytotoxicity and learning impairment in mice.^[4] Alternatively, in young mice, desflurane has recently been shown to induce more neuroapoptosis compared to isoflurane and sevoflurane,^[11] and other laboratories have demonstrated similar neurotoxic profiles in desflurane compared to the other volatile agents.^[12] Confounding reasons for these discrepancies are unclear, though they may relate to different strains of mice used in experiments and differences in experimental design (including dose of volatile anesthetic administered). Literature representative of desflurane neurotoxicity research is summarized in Table 7.3.

Nitrous oxide. Nitrous oxide (N₂O) is a unique anesthetic agent in many regards. Unlike the halogenated ether anesthetics above, N₂O acts as an N-methyl-D-aspartate (NMDA) antagonist and has effects at other various ligand- and voltage-gated channels.^[13] Furthermore, given the relatively low potency of N₂O, it is used adjunctively rather than primarily as a maintenance anesthetic.

Initial concerns regarding N₂O centered on the ability of N₂O to inhibit vitamin B12 and methionine synthase activity.^[14] Though there are indeed case reports of patients experiencing neurologic deficits with N₂O exposure, this has been in the setting of either chronic exposure or extremely low preexisting levels of vitamin B12.^{[15, 16]} Another finding has related to the formation of reversible neurotoxic vacuole formation in adult rats exposed to N₂O.^[17] The significance of these vacuoles is unclear, though cell death only ensued with prolonged N₂O exposure. Furthermore, both vacuole formation and cell death were blocked with concomitant benzodiazepine exposure, indicating a possible excitotoxic-mediated injury. Regardless, both concurrent benzodiazepine administration (as is commonly given perioperatively) and limiting N₂O exposure seem to impede this pathology from occurring. Lastly, the combination of N₂O and isoflurane has been shown to increase apoptosis and Aβ levels in vitro.^[18] Of note, these findings were not appreciated when either N₂O or isoflurane were administered alone. This suggests a potentially synergistic interaction between these agents, though in vitro confirmation is warranted.

Intravenous anesthetics. Propofol is one of the most commonly used anesthetics used for both induction and maintenance of anesthesia. Research has demonstrated that drugs with GABAergic activity – like propofol – can induce neuroapoptosis in the developing rodent brain.^[19] This has indeed been shown in vitro with immature rat and mouse neurons, where propofol-induced calcium dysregulation and cytoskeleton destabilization, respectively, were shown to induce apoptosis.^{[20, 21]} Propofol-induced neuroapoptosis has since been demonstrated in vivo as well using an infant mouse model.^[22] Clinical implications for these findings are unclear, research to assess for anesthetic neurotoxicity in children is ongoing.^[23] On the other end of the age spectrum, propofol has been shown to induce tau protein phosphorylation, a central component of neurofibrillary pathology.^[24] Neurofibrillary tangles (NFT) are indeed believed to play a role in AD pathophysiology.^[25] Further studies may be informative, as propofol may directly modulate this pathophysiologic substrate of AD. A review of pertinent literature is available in Table 7.4.

Like propofol, ketamine is frequently used in the field of anesthesiology, as it serves as both an anesthetic and potent analgesic. From a neurotoxicity standpoint, ketamine has been shown to induce apoptotic neurodegeneration in various animal models.^{[19, 26]} Longer exposure periods may lead to a higher degree of neurodegeneration and subsequent cognitive impairment.^[26] The mechanism by which this neuroapoptosis occurs may be via NMDA receptor up-regulation,^[27] and with this up-regulation, there may be a greater degree of glutamate-induced excitotoxicity upon binding NMDA.

Clinical Investigation

Despite the wealth of laboratory data available on anesthetic neurotoxicity, clinical research has lagged behind preclinical investigation. There are likely a multitude of reasons for this chasm, with both patient- and experimental-specific challenges that make designing these studies difficult. For example, consenting patients with any preexisting cognitive deficits may prove challenging. Long-term follow-up may also be problematic, both in terms of patient availability for follow-up (as these patients may have a higher comorbidity burden) and in selection of appropriate cognitive testing batteries. The latter issue alludes to challenges with experimental design, where, until very recently, there has been no universally agreed-upon definition of perioperative neurocognitive disorder (PND; formerly postoperative cognitive dysfunction, POCD) or consensus as to which cognitive testing tools should be used for postoperative cognitive evaluation. Furthermore, other factors such as preoperative functional status and nonanesthetic pharmacologic influence may affect perioperative cognition.^[28] Sufficiently isolating any effects on cognition due to anesthetic agents or technique is a difficult task. Nonetheless, given the mounting laboratory evidence of anesthetic neurotoxicity, the time is ripe for clinical investigation.

Recently, a perioperative method for analyzing biomarkers of AD in vitro via human cerebrospinal fluid (CSF) has been developed.^{[29, 30]} Interestingly, a decreased CSF Aβ/tau protein ratio has been found to be a possible harbinger of AD,^[31] and lower preoperative CSF Aβ/tau protein ratios may correlate with changes in postoperative cognitive function.^[30] Interestingly, desflurane and isoflurane may be associated with different CSF characteristics of Aβ and tau protein.^[29] It should be noted, however, that these were relatively small single-center studies, and the exact clinical relevance of these CSF findings are not yet completely understood. Regardless, these findings have highlighted interesting hypotheses, and follow-up investigation is warranted.

Additionally, Zhang et al. piloted a clinical study examining the effects of three different anesthetic protocols on postoperative cognitive function.^[32] Though all 45 patients received a spinal anesthetic, one cohort of patients was also administered isoflurane anesthesia, a second cohort of patients was administered desflurane anesthesia, and the third cohort received spinal anesthesia alone. In this pilot study, 27% of patients experienced postoperative cognitive decline in the isoflurane group, as compared to 0% in the other two groups. This was, however, a pilot study, and the conclusions that can be drawn are limited. Nonetheless, a subsequent large-scale, well-designed, and adequately powered trial would be informative.

Overall, much work remains to be done in the realm of clinical neurotoxicity research. Biomarkers associated with AD may hold implications for predicting PND, though further investigation is needed. As isoflurane and desflurane are associated with differences in CSF AD biomarker characteristics and potential postoperative cognitive trajectory in preliminary studies, further research is needed.