In patients with cerebral infarction, after 1 year of follow-up, the cumulative recurrence rate was 23.29% [3]. Nguyen Van Thanh surveyed 203 patients with cerebral ischemic stroke, after 3 months of follow-up, the recurrence rate was up to 19.2% [11]. And Bui Chau Tue surveyed 307 patients with cerebral infarction, after 6 months of follow-up, the recurrence rate was up to 20.54% [10].
Thus, according to our study as well as most other authors, patients with cerebral infarction due to internal carotid artery occlusion have a lower recurrence rate than patients with cerebral infarction in general. This can be explained by two arguments: first, internal carotid artery occlusion causes greater brain damage than usual, leaving less brain parenchyma, so it is less likely to recur; second, the internal carotid artery has been blocked so there is no more blood flow, so if the brain still has collateral blood flow, the possibility of recurrence depends less on the damage of the blocked artery but on the condition of other arteries (blood clots in the internal carotid artery only cause recurrent infarction due to debris from the distal part, or if from the proximal part, it must go through the collaterals of the external carotid artery). This was confirmed in the study of Thanvi B et al [128] with a 5-year recurrence rate of 14% in the internal carotid artery occlusion group compared to 40% in the internal carotid artery stenosis group (similar mortality: 29 vs 32%).
4.4 ANALYSIS OF PROGNOSIC FACTORS
Maybe you are interested!
-
Outcomes of Functional Recovery, Mortality, and Recurrence in Patients with Cerebral Infarction and Internal Carotid Artery Occlusion -
Survey of prognostic factors of cerebral infarction due to internal carotid artery occlusion - 2 -
Early outcome assessment of laparoscopic radical prostatectomy - 1 -
Research on accounting approach and market approach in predicting default of Vietnamese enterprises - 23 -
Teaching Objectives and Learning Outcome Test:
4.4.1. Prediction of functional recovery outcome
Three models were developed with different sets of variables that are predictive of functional outcome in patients with internal carotid artery occlusion. The models differed only in imaging variables, with model 1 using the status of the M1 segment of the middle cerebral artery, which can be obtained by noninvasive angiography (CTA or MRA) even when the patient presents very early and is a candidate for revascularization. This has two implications: first, it is feasible, since the trend and actual protocols of many hospitals include angiography in the hyperacute phase, right at the beginning of brain imaging. Second, because the variables in this model can be collected very early, they can be used to rapidly predict the outcome of
patients, providing additional basis for the treating physician to decide whether or not to use aggressive treatment measures such as venous recanalization or mechanical recanalization.
From model 1, a formula for calculating the outcome prediction score was established (formula (2), page 79), and preliminary evaluation on the study sample itself gave quite high prognostic values, with sensitivity and specificity of 77.2% and 83.3%, respectively, and an overall predictive accuracy of 80.3%. This predictive value of course needs to be verified with another patient sample in another study before it can be used clinically.
The other two models are more accurate, but are slower, since it takes time for brain lesion images to appear (on CT) and stabilize.
Regarding the prognostic factors, it can be seen that their hazard ratios are relatively stable across models, that is, stable when adjusting for different sets of variables. Blood glucose is the least influential factor, with an HR of approximately 1 and a 95% confidence interval containing the value 1 in it.
Hypertension was a favorable predictor of outcome, with a hazard ratio of approximately 0.75, meaning that people with hypertension had a 25% lower relative risk of a poor outcome than people without hypertension. However, the 95% confidence interval for the HR of this factor was quite wide and covered the value of 1, meaning that the predictive power and statistical significance were not high. There is no reasonable explanation for this inverse association, possibly because hypertension in the acute phase helps to restore some cerebral perfusion, but the blood pressure values at admission were not statistically significantly associated with outcome. Therefore, the sample size of this study was probably not large enough to demonstrate a true association.
Coronary artery disease is a factor with a very high risk ratio, from 6 to more than 7, meaning that people with coronary artery disease have a risk of a bad outcome 6-7 times higher than people without this factor, but the confidence interval is very wide, and also contains a value of 1, so it needs to be evaluated in a larger sample size to narrow the confidence interval and increase the predictive significance, thereby also finding a more accurate risk ratio. To explain the relationship between coronary artery disease and functional outcomes, it can be understood that any comorbidity
may adversely affect patient outcomes. Furthermore, coronary artery disease is an important disease, and its mechanism is related to atherosclerosis, and it has the same risk factors as cerebral infarction, and patients with both cerebral infarction and coronary artery disease have polyvascular disease, the ability to recover will be poor and recurrence will increase.
Age has a fairly good predictive significance, with each increase in age increasing the risk of a poor outcome by about 8 to 9% (HR 1.083 to 1.096), with a narrow 95% confidence interval that does not contain the value 1, indicating good predictive significance. To more easily see the effect of age, the HR can be calculated for each 10-year increase, in which case the new HR will be equal to the exponential HR
10. As a result, for every 10-year increase in age, the risk of a poor outcome increases by 2.2 to 2.5 times (1.083^10 = 2.22; 1.096^10 = 2.5). The role of age in predicting patient outcomes is easy to explain. Generally, older people have poorer general health, more brain atrophy, and poorer recovery and compensation capabilities of nerve cells. Therefore, it is reasonable that our study found that age is a valuable factor in predicting patients' functional recovery outcomes.
The NIHSS score was strongly predictive of outcome in all three models, with narrow 95% confidence intervals that did not include the value 1. This demonstrates that the predictive value of the NIHSS score is very strong, and remains significant even in the presence of variables that represent the severity of brain damage on imaging. Specifically, for each additional NIHSS point, the risk of poor outcome increased by 11% to 16% (HR 1.113 to 1.156) compared to baseline.
The NIHSS score is quite understandable in predicting functional outcome, because the initial clinical severity is also a factor reflecting the severity of brain damage, and the greater the degree of brain damage, the poorer the ability to recover later. A similar explanation can be made when talking about two other variables assessing the degree of brain damage, that is, the severity of stroke, which are the cerebral infarction lesion on imaging assessed by region and by the ASPECTS scale.
Although the NIHSS score is only a clinical assessment, it has great value in expressing the severity of the patient, and has a role that is not inferior to, and also supplements the role of imaging in assessing stroke severity to predict functional recovery outcomes. This once again explains why the NIHSS score has now become indispensable in the practice of stroke diagnosis and treatment in the world, as well as indispensable in research and clinical trials on stroke patients.
The characteristics of cerebral infarction lesions on imaging were found to be predictive of the risk of poor outcome, with each increase in ASPECTS score resulting in a 30% reduction in the risk of poor outcome compared to the previous baseline score (HR 0.73), with a narrow 95% confidence interval that did not include the value 1. If assessed by the extent of cerebral infarction lesions on imaging (lesion segmentation), the large-area group had a 38-fold increased risk of poor outcome, the large-area group had a 4.7-fold increased risk, and the border-area group had a more than 2-fold increased risk, compared to the small-area group with shallow or deep brain lesions. The confidence intervals of the HRs at these levels were also relatively wide, but only one comparison had a confidence interval containing the value 1.
In general, imaging is still considered more accurate and objective than clinical assessment of stroke severity. However, CT scan images in the acute phase often do not clearly show the limits and size of the lesion. MRI provides better images but is not a commonly performed examination in the acute phase, moreover, the lesions on diffusion MRI often tend to show slightly over the actual size of the lesion later. Therefore, models 2 and 3 also have good meaning when combining both clinical and imaging to predict outcome.
Ipsilateral M1 segment status, after adjustment for other variables in model 1, was predictive of adverse outcome at a moderately significant level, with the risk of adverse outcome nearly doubling in those with severe-occlusive M1 stenosis compared with those with normal or mild M1 stenosis (HR 1.706). The 95% confidence interval was not large, but contained a value of 1, and therefore warrants further investigation.
The status of the ipsilateral middle cerebral artery M1 segment is a variable that reflects the combination of two factors, one is the M1 segment itself, and the second is the status of collateral perfusion. If there is no collateral, then whether the M1 segment is actually normal or not, the image will still show a loss of signal. Meanwhile, if the collateral is good, the perfusion to the M1 segment is good, then the stenosis or occlusion of this artery segment plays a decisive role. In fact, the collateral for the anterior cerebral artery is often good, specifically in 88.4% of the cases in our study, so the collateral for the M1 segment of the middle cerebral artery is almost representative of the status of collateral blood supply when the ipsilateral internal carotid artery is occluded. This is a variable that can collect data very early after the onset of stroke, possibly by CTA or MRA, so it has practical significance in helping to predict the outcome early if no intervention is performed, as a basis for deciding on intervention.
Comparison of functional outcome predictions with literature
Comparing studies in the literature, we did not find any studies that evaluated the factors predicting functional recovery outcomes in patients similar to our study. There was only one retrospective study by Matsubara [94] in 2013, on 16 patients with internal carotid artery occlusion with severe acute cerebral infarction and acute endovascular intervention, which noted atrial fibrillation as a risk factor when all 4 patients with atrial fibrillation had poor outcomes, while 40% of patients with atherosclerosis had good outcomes. This study was too small and did not have a similar design for comparison.
Expanding on prognostic studies on patients with ischemic stroke in general, we recorded the following results.
Similar to our study, in the general population of cerebral infarction, the initial stroke severity expressed by the NIHSS score has been confirmed by many studies to have predictive value for functional recovery. Specifically, Weimar's study in 2004 [136], on a large sample size of 1079 patients and tested on 1307 patients, recorded that NIHSS at admission had an OR of 1.313 in predicting functional outcome after 100 days. Dhamoon's study in 2009 [50] on 525 patients
Stroke severity as expressed by the NIHSS score was also noted as a prognostic factor.
predict functional outcome at both 6 months and 5 years.
Nguyen Ba Thang's study [17] on 149 patients recorded that the NIHSS score had predictive value for functional outcome at 2 months with OR 0.741. Truong Van Son in his 2010 study [15] on 243 patients also recorded that only the NIHSS score had predictive value for functional outcome after 30 days (OR = 2.902; for NIHSS
>6 compared to NIHSS 0-6). Phan Van Mung's study in 2009 [6] on 71 patients through multivariate analysis recorded that an NIHSS score of 16 or higher had an OR of 18.95 compared to a score lower than this level in predicting poor outcomes (BI<60 or death) after 3 months.
The severity of brain damage on imaging has also been noted by some studies to have predictive value for functional recovery outcomes after ischemic stroke. Specifically, the ASPECTS score with a cut-off point of 7 has predictive value for 3-month outcomes with an OR of 43.1 in the study by Le Tu Phuong Thao in 2009 [19]. Phan Van Mung's study also in 2009 [6] noted that ASPECTS has predictive value for poor outcomes but only in univariate correlation; when analyzed in multivariate terms, this variable did not surpass the NIHSS score in predicting functional outcomes after three months. Nguyen Ba Thang's study in 2006 [17] also noted that the size of brain lesions on CT scan, assessed by lesion segmentation or by ASPECTS, has predictive value for functional outcomes after two months. Cao Minh Chau's study [1] used infarct size on CT to predict functional recovery after 3 months, while Nguyen Thi Hung's study [2] noted that brain damage in the entire middle cerebral artery perfusion area or borderline cerebral infarction was a predictor of poor functional outcome.
Age of onset is also a predictor of outcome that has been noted in many studies. One can mention the study by Dhamoon in 2009 [50] in which age had an OR of 0.76 and 0.95 in predicting good functional outcome at 6 months and 5 years after cerebral infarction, respectively. Similarly, the study by Weimar in 2004
[136] also noted that age had predictive value for poor outcome at 100 days with an OR of 1.051.
Domestic studies have mostly recorded the same, with older age predicting worse outcomes. Typically, Nguyen Ba Thang's study [17] recorded that age had an odds ratio of 0.928 for predicting good functional outcomes after 2 months; Phan Van Mung recorded that age ≥65 compared to younger age had an odds ratio of 6.69 for poor outcomes after 3 months. In addition, Cao Minh Chau's study [1] recorded that people over 65 had lower independent outcomes than people aged 45 to 64; Pham Van Phu's study [14] recorded that older people had a lower rate of independent outcomes than younger people.
4.4.2. Prediction of fatal outcome
Mortality assessed here includes death directly due to the initial stroke, death due to recurrent stroke, death due to subsequent cardiovascular events, death due to other causes, and death of unknown cause. Of the three models obtained, the first model includes factors that can be collected very early after onset, which is significant for early prediction; the other two models predict later because they have to wait for brain damage to be clearly seen on imaging, mainly CT scan.
The role of initial stroke severity in predicting outcome is not difficult to explain, in particular, it is related not only to the patient's ability to recover, as in the analysis of functional recovery outcomes above, but also to the patient's ability to survive, both through the acute phase and long-term survival, which depends on the possibility of later complications that increase mortality. In addition, people with severe stroke are more likely to have general effects on the function of other organ systems, especially the cardiovascular system, and are also more likely to have traumatic falls, which are also related to death from other causes. This explains why our three models clearly recorded the role of variables reflecting stroke severity, including NIHSS score at admission, cerebral infarction lesion imaging assessed by lesion segmentation or by ASPECTS score.
For admission NIHSS scores, all three models showed that increasing points increased the risk of death, with each additional point increasing the risk of death by 10–11% (HR 1.097–1.115). The 95% confidence intervals for the HRs in these models were narrow and did not include the value 1, indicating good predictive significance. In other words, for an HR of 2, the gap in scores would be 7 (1.097^7 = 1.91; 1.115^7 = 2.14), meaning that a 7-point increase in NIHSS score would double the risk of death.
The risk of death was increased by approximately 4% (HR = 0.959) for each point decrease in the ASPECTS score. According to the stratification, only large infarctions increased the risk of death by nearly 2.1 times compared to small infarctions, while large infarctions of the middle cerebral artery and border infarctions were less lethal, but the 95% confidence interval of the hazard ratio always covered the value of 1, so the predictive significance was not high. This low contribution of imaging in predicting death can be explained by the large role of the NIHSS score in the model.
Reviewing the literature, although there are no studies predicting mortality in patients with internal carotid artery occlusion, many studies predicting the outcome of ischemic stroke in general have noted the relationship between stroke severity and mortality. The NIHSS score is the most commonly recorded factor related to this outcome. Specifically, a study by Wiemar C [136] found that the NIHSS score was significant in predicting 100-day mortality with an OR of 1.168. A study by Cao Phi Phong and Phan Dang Loc [13] found that NIHSS <9 was significant in predicting mortality at discharge with an OR of 12.6. A study by Le Tu Phuong Thao [19] also found that an NIHSS score of 16 points or more was significant in predicting mortality within 3 months with an OR of 12.8.
Saposnik G's study [118] in 2011 with a large sample size (n=12262) assessed severity using the CNS score (Canadian Neurological Scale), a scale with similar significance to the NIHSS, and noted that the CNS score was predictive of mortality at both 6 months and 5 years, with the odds ratio of the mild to severe score groups ranging from