CT Perfusion for Identification of Patients at Risk for Delayed Cerebral Ischemia during the Acute Phase after Aneurysmal Subarachnoid Hemorrhage: A Meta-analysis
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.271235
Source of Support: None, Conflict of Interest: None
Keywords: CT perfusion, delayed cerebral ischemia, meta-analysis, subarachnoid hemorrhageKey Messages: The analysis has investigated the correlation between CTP defcits in the acute phase and impending DCI after aneurysmal SAH, revealing that CTP allows early and effcient detection of patients at risk for DCI. Although aSAH has a higher morbidity and mortality rate, CTP may allow close monitoring and preemptive therapy for improvement of the prognosis in these patients.
Delayed cerebral ischemia (DCI) is a devastating complication after aneurysmal subarachnoid hemorrhage (aSAH). It typically occurs 4–14 days after aneurysm rupture and is associated with significant morbidity and mortality in patients who survive the initial hemorrhage.,, The definition of DCI is variable because it has long been described as a new onset of clinical deterioration not attributable to other known causes., In addition, recently cerebral infarction excluding other etiologies has been strongly recommended as a primary outcome measure to define DCI by expert consensus, and the clinical deterioration has been recommended only as a secondary outcome measure due to suspected lower expert agreement rates., However, the definition of DCI may limit the initiation of preemptive therapy once clinical deterioration and even cerebral infarction have occurred, and therefore, a technique that can detect DCI before poor outcomes is urgently needed, especially in patients in comas or under sedation. As a noninvasive and rapid medical imaging technology, computed tomography perfusion (CTP) has been frequently used to diagnose the occurence of DCI following aSAH.,,, However, the clinical role of CTP in the prediction of impending DCI remains unclear.,,, Therefore, we aimed to perform a meta-analysis to investigate the role of CTP in the identification of patients at risk for DCI during the acute phase (<4 days) after the initial bleeding.
Information sources and search
Relevant articles were searched by a systematical review of the electronic databases of PubMed, EMBASE, and Cochrane databases (last search July 2016). All studies evaluating the predictive value of CTP for DCI following aneurysm rupture were identified. The terms used for the electronic search included “subarachnoid haemorrhage” or “subarachnoid hemorrhage,” “DSA” or “DS angiography” or “digital subtraction angiography,” and “CTP” or “PCT” or “CT perfusion” or “perfusion CT” or “computed tomography perfusion” or “perfusion computed tomography.” We also identified relevant studies from the reference lists of all selected original and review articles. Two authors (H.G.S. and J.P.M.) performed this search, and manuscripts written in languages other than English or Chinese were excluded.
Study selection and inclusion criteria
After the literature search, search results were preliminarily screened via title and abstract information by an author (H.G.S.) to exclude apparently irrelevant articles. Then two additional authors (J.P.M. and Y.L.) independently reviewed the full text of the shortlisted articles and identified the articles that met the criteria for inclusion in our meta-analysis. Any disagreements were resolved by consensus. Inclusion criteria were as follows: (1) all patients with aSAH, defined by CT or magnetic resonance (MR) imaging or lumbar puncture; (2) CTP scan performed in the acute phase (<4 days after aSAH) for the prediction of DCI; (3) perfusion measured quantitatively, semi-quantitatively (comparing quantitative perfusion in one hemisphere to the contralateral side), or qualitatively (visual evaluation for side-to-side perfusion asymmetry on different color maps)., (4) definition for DCI included clinical deterioration (new neurological deficits or worsening on Glasgow Coma Scale and not explained by other causes), cerebral infarction (new infarction on follow-up CT or MR imaging), or both. Exclusion criteria were as follows: (1) conference abstracts; (2) same or overlapping patient populations reported in other articles (we only included the article with the largest population); (3) article could not provide adequate data to meet the need of our analysis.
CTP parameters included cerebral blood volume (CBV), cerebral blood flow (CBF), mean transit time (MTT), and time to peak (TTP). In all enrolled articles, we collected the best parameter or the definition of abnormal CTP scan result and chose the data, which were then correlated with the greatest overall predictive value for DCI for rebuilding of the fourfold table. We assessed CTP as a single test object and not by its individual components. In case these data were not provided, we calculated from additional information or contacted the corresponding author by e-mail to request for the data.
Besides the term “DCI,” there are many other terms used in the literature. The following terms were also considered as DCI: delayed ischemic neurological deficits, delayed neurological deficit, secondary cerebral ischemia, symptomatic vasospasm, secondary cerebral infarction, delayed cerebral infarction, and vasospasm-associated infarction.,, Because all terms were comparable with the DCI definition that was recently recommended by expert consensus, articles with these terms were involved in our analysis.
The Quality Assessment of Diagnostic Accuracy Studies (QUADAS) tool,, which is widely used to assess the quality of included studies, was implemented in our analysis. This tool assesses bias according to 14 items and each item can be answered as “yes,” “no,” or “unclear.” The item information of included articles was extracted.
Besides the above-mentioned information, extracted information also included characteristics of patients and studies, as well as the CTP performance. Two authors (H.G.S. and J.P.M.) independently performed this information extraction and any disagreements were resolved through discussion [Table 1] and [Table 2].
Pooled sensitivity, specificity, and odds ratios with 95% confidence intervals (CIs) were calculated, and the study heterogeneity was evaluated by I2 and Chi-square tests. In our analysis, the pooled odds ratio was calculated by a random-effects model rather than the less conservative fixed-effects model because there was wide variation in enrolled studies, such as different CTP protocols, postprocessing software, DCI definitions, and in the component of abnormal CTP result. Forest plots were drawn to display the variations of odds ratios from different studies, and the Beggar's funnel plot was plotted to detect publication bias. Under the right conditions, a summary receiver-operating characteristic (SROC) curve was plotted to assess the diagnostic value of CTP; the area under the SROC curve with its Q* point was also calculated. The above-mentioned statistical analyses were performed using Review manager 5.3 and Meta-Disc 1.4 software.
The search strategy yielded 183 articles. 161 articles were considered irrelevant for this analysis and were excluded on the basis of screening the titles and abstracts. 22 relevant articles were selected for full review and three articles,, were finally included in the analysis. 13 articles were not included because the reported data were insufficient for this analysis, and six articles were excluded because four were reduplicated studies and two were conference abstracts. Although one article reported that the CTP was done to detect impending DCI on day 3–5 after aSAH, early CTP was performed on day 3 in 39 of 43 patients (91%) and hence the article was included for analysis. The study selection is shown in [Figure 1].
The QUADAS tool was used to assess the methodological quality of the three included articles and revealed several potential biases. In all articles,,, CTP was performed before the time-window of DCI (4–14 days after aSAH). Therefore, it indicated the presence of disease-progression bias. In addition, two articles possibly presented reference standard review bias because they did not clearly state that the DCI was evaluated blinded to the results of the CTP scan.,
Three studies included 128 patients. DCI developed in 48 patients (37.5%). The meta-analysis revealed a pooled odds ratio of 32.15 (95% CI, 9.92–104.21) [Figure 2], suggesting that patients with a positive CTP test in the acute phase after aSAH were approximately 32 times as likely to develop DCI compared with those with a negative CTP test. The value of Chi-square was 0.18 (P = 0.92) and I2 was 0.0%, indicating that there was no statistically significant heterogeneity in these data [Figure 2]. The pooled sensitivity and specificity of CTP for detecting impending DCI after aSAH was 65% (95% CI: 0.49–0.78) and 91% (95% CI: 0.83–0.96). The SROC curve could be not drawn and the area under the SROC curve with its Q* point could not be calculated because of the small sample size. The Beggar's funnel plot is shown in [Figure 3].
In our previous meta-analysis, CTP was confirmed to have a high diagnostic value for the detection of DCI after aSAH, and the pooled sensitivity, specificity, and diagnostic odds ratio were 82%, 82%, and 20.96%, respectively. Mir et al. performed a meta-analysis for the evaluation of the utility of the usefulness of CTP in the determination of DCI following aSAH and demonstrated a pooled odds ratio of 23.14. Unfortunately, the subanalyses that CTP were performed in the acute phase and during the DCI time-window after aSAH were not done in both of the articles, which is a significant limitation. Cremers et al. collected mean values of CTP parameters to evaluate the values of prediction and diagnosis of CTP for DCI after aSAH, respectively. The study showed that CTP can be used to diagnose DCI in the DCI time-window (4–14 days after aSAH), but CTP cannot be used for predicting DCI on admission (<72 hours after aSAH). However, there were still limitations the article. First, studies that do not report the mean values of CTP parameters could not be enrolled for the meta-analysis. Some articles assessed perfusion deficits visually on the different color maps and showed an opposite result that DCI was significantly related to the positive CTP test result in the acute phase after aSAH., Second, although semi-quantitative measurements on admission reveal significant differences between DCI and non-DCI patients, the data could not be extracted for the meta-analysis due to a limited sample size.
We attempted to overcome some of the aforementioned limitations. In our eligibility criteria, CTP was performed only in the acute phase after aSAH and the data from all perfusion measures (quantitative, semi-quantitative, and visual assessments) were available for meta-analysis. Finally, our meta-analysis indicates that the patients with perfusion deficits detected by CTP scan in the acute phase after aSAH were approximately 32 times more likely to develop DCI compared with those patients without. This suggests that CTP allows for early and efficient identification of patients at risk for developing DCI.
It is noteworthy to mention that our analysis also revealed other limitations. First, there is a lack of consistency among the definitions for DCI. Second, CT perfusion maps are generated in different institutions by different CTP protocols and postprocessing software. Third, nonuniform definitions of an abnormal CTP test result are used. Fourth, different CTP parameters with various thresholds are used to evaluate CTP test results in quantitative analyses. Unfortunately, given such variability, subanalyses could not be performed due to the limited sample size. We will improve the analysis without delay when more relevant studies are reported in the future. Despite many differences among these studies, there was no statistically significant heterogeneity of odds-ratio in our meta-analysis. It proved that a strong association exists between CTP deficits during the acute phase and impending DCI after aSAH across some variations among CTP performance and DCI definition.
Our meta-analysis has shown a strong correlation between CTP deficits in the acute phase and impending DCI after aneurysmal aSAH, suggesting that CTP allows early and efficient detection of patients at risk for DCI. As a noninvasive and quick medical imaging technology, CTP can assist in the detection of abnormal brain perfusion before the occurrence of DCI, allowing for close monitoring and preemptive therapy to improve the prognosis of patients with aSAH.
Ethical standards and patient consent
We declare that this manuscript does not contain clinical studies or patient data.
This article is supported by the National Natural Science Foundation of China (No. 8140050475).
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2]