That DCH can be a life-saving treatment option for complete MCA stroke is no longer a matter of debate [3, 5, 9–11, 37–39]. However, there are only limited neuropsychological data on patients who underwent decompressive hemicraniectomy after complete MCA infarction. Only one previous study of 14 patients who underwent DCH focused on neuropsychological sequelae . In contrast to the promising data on survival for DCH, our findings on impairment of higher cortical functions after DCH were not at all encouraging: More than two-thirds of our patients met the criteria for dementia after the hemispheric stroke. In contrast to Leonhardt et al.  who found that mean visuo-spatial and visuo-constructive capabilities were nearly normal and self-assessed mood not at all impaired, none of the cognitive domain z values in our patient group fell within the normal range. The only results comparable to ours were for attention and non-verbal memory. The reason for the discrepancy between our findings and Leonhardt et al. is that they did not focus on memory and learning and - after observing the difficulties patients had with tests for attention and psychomotor speed -they just eliminated most of the tests for this domain.
The self-assessment of mood in our patient group revealed significant impairments. In contrast to Leonhardt , but congruent with other authors [9, 41], we found increased scores for the BDI, although the majority of the patients had received anti-depressive drugs. These results can be clearly explained by previous findings on right-sided brain pathology: Right-sided hemispheric lesions impair both verbal and non-verbal memory and learning [42, 43]. In patients who underwent right-sided surgical pallidotomy, one observed, in addition to learning deficits for verbal and non-verbal content, a deterioration of frontal executive and visuo-constructive functions . Patients with right-sided fronto-temporal lobe atrophy have problems with spatial memory  and are also prone to develop affective disorders such as depression  or problems with the emotional processing [47–50]. The results in our patients after massive stroke-induced destruction of the right hemisphere underline the importance of this side of the brain for cognition and mood.
In regard to cognitive outcome after DCH, our results may be too negative since 8 of our patients were older than 60 years, whereas other recent studies included only patients up to 60 years . Furthermore, the current guidelines of the European Stroke Organization (ESO) recommend a maximum cut-off age of 60 years for the DCH .
The higher age of our study population also explains the high proportion of patients with an unfavorable Rankin scale (75% in our study) compared to 43% in pooled data from the DESTINY, HAMLET and DECIMAL trials. Thus, the population analyzed here is certainly not comparable to the study groups in the above mentioned studies on DCH. The same is true for the variance of time from stroke onset to the neuropsychological tests. Therefore, our data should be interpreted cautiously with respect to these issues and might not be valid for a younger population examined within a narrower time interval after the treatment. Our findings are comparable to another study with a similar rate of unfavorable outcome (73.7%)  which also included older patients. Furthermore, our data on indicators of activities of daily living (NEADL, BI) may also not be representative of a younger population. Other studies which included older patients also found these indicators to be correlated negatively with age .
Another limitation of our investigative focus on cognitive deficits is that patients with speech-dominant hemispheric stroke were not included. Our findings for the depression questionnaires, therefore, are not transferable to patients with DCH over the speech-dominant hemisphere, as it is well known that mood disturbances appear more often in patients with right-sided stroke than in those with left-sided cerebral ischemia [43, 47]. Strategically unfavorable localizations of small right-sided strokes, i.e. in the frontal projections of the corpus callosum, can cause major depression . The same is true for cognitive functions; neuropsychological investigations of patients after neurosurgical interventional treatment of Parkinson's disease have shown the importance of the right brain for cognitive functions [44, 52].
In regard to both mood and cognition, our findings may significantly differ from patients with left-sided MCA infarctions and DCH. Since this study focused on cognitive functioning, reliable results were dependent on the patient's understanding of the test rules. The degree of comprehension is extremely difficult to quantify in partially aphasic patients, and there is no general rule as to how to correct the cognitive results for the individual type of aphasia. Therefore, we decided to exclude patients with speech-dominant hemispheric infarctions rather than misinterpreting their cognitive test results.
As in other studies on DCH, most of our patients were able to walk after DCH and had regained some independence, despite the persisting physical handicaps . Physical impairment is, however, just one element which affects human well-being; in our patient group, the quality of life was significantly lower than in healthy control patients, which matches the findings reported by other groups [4, 53, 54]. Not surprisingly, in our patients, hrQoL was significantly correlated with the Beck's depression inventory. This underlines the necessity to diagnose and to treat post-stroke depression early and effectively in such patients. Depression is often a cause of cognitive disturbances, and vice versa. Our data do not allow us to safely quantify the influence of depressive pseudo-dementia on the test results. However, since cognition and mood interact closely, it is very difficult or perhaps impossible to differentiate their mutual influences. The lack of a significant correlation between the BDI and cognitive functioning, however, indicates that depression was not the primary factor affecting cognitive performance in DCH patients. In addition to the lack of a significant correlation of cognitive scores with the BDI, our patients also showed abnormal results for verbal learning and visual constructive tests. These domains are usually spared in depressive pseudo-dementia [55, 56].
The presence of symptomatic epilepsy should have an influence on cognitive functioning as well as mood . Particularly right-sided brain pathology should predispose patients with epilepsy for depression . It is most likely that the sample size of our study population was too small to reproduce these effects.
All told, the findings demonstrate that the probable outcome after DCH of the non-speech dominant hemispheric infarction includes significant cognitive deficits along with a physical handicap and impaired mood, and consequently, a diminished quality of life. The prospect of living with depression combined with severe physical and cognitive handicaps may be inacceptable for many patients; for others, however, it might be the better option compared to 70-80% mortality with a conservative treatment .
A significant limitation of this study in regard to the discussion of post-hoc patient agreement is the lack of information on long-term cognitive deficits in patients who survived a conservative treatment. It is possible that neurological and cognitive sequelae in these patients are higher than in those who had been operated on due to the untreated swelling of the ischemic brain regions. To our knowledge no such data are available. The group of the operated patients who stated that if they had to decide again, they would not have opted for DCH did not differ significantly in their neuropsychological test results, and also did not display lower quality of life scores as compared to the patients who stated that they would again opt for DCH. However, the former patient group suffered from significantly higher scores for depressive mood (p = 0.039) than the latter. More than half of our patients were treated with anti-depressive drugs. It could be hypothesized that the patients whose post-hoc assessment of DCH was negative just needed a better anti-depressive treatment, especially since the agreement rates in our patient group were considerably lower than in other trials (e.g. ). Our data can neither confirm nor reject this hypothesis. The fact that there was no difference in BDI score between the patients who were administered anti-depressive drugs and those who were not might argue against this idea. However, all reported rates of retrospective agreement with DCH could be severely biased by the interview situation. The patients were interviewed by a person who was involved in their stroke treatment and thus in the life-saving process, which may have inhibited them from answering honestly, at the risk of appearing ungrateful.