Following the first transcatheter aortic valve implantation (TAVI) performed by Cribier et al. in 2002 , the procedure has evolved considerably, to the point where it is now on the verge of posing a viable treatment option among intermediate surgical risk patients with symptomatic severe aortic stenosis . In the early days of the evolution of TAVI, cerebrovascular events were considered to be the “iceberg” standing in the way of the “TAVI Titanic” . Recent randomized trials reflecting improved patient selection, increased operator experience and iterated devices have somewhat dissipated those fears, with a similar incidence of cerebrovascular events already demonstrated between TAVI and surgical aortic valve replacement . However, these clinically overt events seem to be the tip of the iceberg, with silent cerebral micro-embolization and its potential cognitive consequences lurking beneath the surface . The emergence of embolic protection devices (EPDs), along with a trend towards treating lower surgical risk patients, has shed further light on these specific issues, which are increasingly recognized as a surrogate burden of the ischaemic cerebral insult imparted during TAVI .

In the setting of surgical aortic valve replacement, a recent review reported rates of cognitive decline of 50–70% within 1 week of cardiac surgery, whereas 10–20% of patients exhibited persistent cognitive decline at 1 year . A few seminal studies have attempted to navigate the deep and dark waters of cognitive evaluation early post-TAVI, mainly demonstrating preserved, and even improved, cognition . However, most of these studies focused on global cognition or memory during a short-term follow-up without serial evaluations. Moreover, changes in cognitive function were assessed by performing formal statistical tests on the mean or median cognitive scores in the setting of small sample size, therefore implying the risk of type II error. In contrast, recent studies evaluating the use of EPD have suggested that rates of early cognitive decline, on the basis of the Montréal Cognitive Assessment (MoCA) measurement, may be as high as 50–72% at 2 days post-TAVI and 37–55% at discharge . Interestingly, the use of EPDs seemed to mitigate this deleterious effect on cognition. Nonetheless, these studies defined cognitive decline as a drop of ≥ 1 point in the MoCA score, thus failing to properly take into account the variability of changes across tests, which depends on the stability and reliability of the assessment tool . In other words, in the worst-case scenario, these findings only reflect random variability.

To avoid these pitfalls of cognitive evaluation, some studies have used specific methods to account for this inherent variability . In the first of its type, Ghanem et al. evaluated the cognitive trajectory of 111 TAVI recipients, 32 of whom were followed for up to 2 years, using the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) test with alternate forms to counterbalance practice effects. Post-TAVI cognitive decline was defined as a drop of > 1 standard deviation compared with a subject’s score pre-TAVI, which occurred in 9% of patients overall. Early cognitive decline (within 3 days of TAVI) was apparent in 5.4% of patients, persisting in 50% of them, whereas 3.6% of TAVI recipients suffered from late cognitive decline (≥ 3 months post-TAVI). Only age was associated with the occurrence of cognitive decline. Among 229 patients aged ≥ 70 years undergoing TAVI, Schoenenberger et al. , using the Mini-Mental State Examination (MMSE) and also defining cognitive decline with the use of baseline standard deviation, reached roughly similar conclusions. In their cohort, 29 patients (12.7%) demonstrated cognitive decline (≥ 3 points decrease in MMSE) at 6 months post-TAVI including eight patients exhibiting a major decrease (≥ 5 points), for whom review of their records identified an obvious medical cause. Of note, the authors failed to isolate multivariable predictors of cognitive decline. These two studies used tests that allowed an integrative measure of global cognition through measurements of several cognitive domains. However, the standard deviation method used in these studies may be prone to overclassifying cognitive decline compared with more stringent methods such as practice-corrected reliable change index (RCI) or regression-based methods . Besides, both the MMSE and RBANS lack sensitivity in the detection of mild cognitive impairment. In contrast, the MoCA, albeit less commonly performed than the MMSE, demonstrated a greater sensitivity for the detection of subtle cognitive changes, due to the inclusion of a more comprehensive evaluation of executive functions, which are predominantly impaired in vascular cognitive impairment .

Recently, we reported the cognitive trajectory of 51 TAVI recipients during a 1-year follow-up period using the MoCA and practice-corrected RCI to define cognitive changes . On the basis of the RCI of MoCA score, four patients (7.8%) presented with 30-day cognitive decline, which persisted at 1 year in one patient (2.0%). Overall, 11.8% of patients exhibited cognitive decline at 1-year post-TAVI. Using five specific tests for some complex executive functions, we demonstrated that a quarter of TAVI recipients experienced deterioration in at least one of these tests at 1-month post-TAVI, which was transient in 60% of these patients and sustained in 40% (10% of the global cohort). Interestingly, we failed to demonstrate a meaningful association between cognitive changes and subsequent quality of life or functional status. Overall, the available evidence suggests that mid-term cognitive decline affects approximately 10–15% of TAVI recipients. Early cognitive decline (i.e. the most likely to result from the procedure itself) occurs in 2–10% of patients when considering global cognition, may affect 25% of them when specific cognitive domains are evaluated, and seems to persist in 25–50% of these patients.

On the contrary, an early and sustained cognitive improvement post-TAVI has been demonstrated in a sizeable proportion of TAVI recipients , ranging from 8% to 38%, depending on the assessment tool, the timing of the evaluation and the methods used to define cognitive changes. This improvement seems more likely among patients who are cognitively impaired at baseline . Importantly, Schoenenberger et al. demonstrated a lower pre-TAVI aortic valve area among patients with post-procedure cognitive improvement. This supports the hypothesis that improvements in cardiac output and, consequently, in cerebral blood flow post-TAVR may reverse some of the baseline alterations.

Admittedly, the current literature leaves us with as many questions as answers, highlighting the tremendous complexity of cognitive evaluation using heterogeneous methods and definitions. That said, what should we expect from our scientific journey on the still largely unexplored and winding path of post-TAVI cognition? First and foremost, we urgently need to validate and harmonize a suitable battery of neurocognitive tests for TAVI candidates, particularly within the setting of treating lower surgical risk patients, especially given the growing interest surrounding neuroprotective strategies such as EPDs. In keeping with the previous point, stringent methods that take into account the variability of cognitive tests should be mandatory to define significant cognitive changes post-TAVI in future studies. To fairly evaluate the role of TAVI in cognitive evolution, we also need to unravel the specific cognitive trajectory of the elderly population with medically managed severe aortic stenosis, as this has not been specifically assessed so far. Future research should aim to precisely elucidate the underlying mechanisms of post-TAVI cognitive decline. Indeed, most studies so far have focused on the cerebral embolic insult as the leading cause of cognitive decline. Although silent cerebral infarcts could conceivably be a major cause of early cognitive decline, they may be transient and thus, their implication in sustained cognitive decline remains questionable. Moreover, only one study has demonstrated a relationship between cerebral embolism and the occurrence of cognitive decline following surgical aortic valve replacement . This finding has recently been reproduced in only one small study in the setting of TAVI, demonstrating a moderate but significant correlation between cognitive decline and the number and volume of new cerebral lesions on diffusion-weighted magnetic resonance imaging . Identifying predictors of cognitive changes and evaluating neuroprotective strategies in high-risk subgroups are other areas of major interest. Finally, of particular importance, is the impact of cognitive changes on TAVI recipients’ quality of life and functional status, which is largely unknown to date. There is still much to be discovered in the fascinating land of post-TAVI cognition and it is safe to say that this incredible journey is thus far from drawing to an end…