American Heart Association

Monthly Archives: August 2013

Stroke Quality Metrics as Basis for Value Based Reimbursement and Purchasing

Waimei Tai, MD

As previously whispered and predicted, CMS recently announced that they are piloting evaluation of new claims-based measures that address 30-day Stroke readmissions and mortality starting August 12, 2013.

The plan of course, is not only to improve the quality of care we provide our stroke patients, but also to potentially adjust payment for performance, as the other CMS core measures have done for MI, heart failure, pneumonia, and surgical outcomes.

Here’s what the Stroke Measures will look like:

STK-1 VTE Prophy
STK-2 Discharged on Antithrombitc Therapy
STK-3 Anticoagulation for Afib and Aflutter
STK-4 Thrombolytic Therapy
STK-5 Antithrombotic Therapy By End of Hospital Day 2
STK-6 Discharged on Statin
STK-8 Stroke Education
 STK-10 Assessed for Rehab

Does this look familiar? Why yes, they are incredibly similiar to what many hospitals already do for  TJC (formerly JCAHO) has metrics for those that are designated primary stroke centers or comprehensive stroke centers. This designation is voluntary for hospitals, but since many counties do EMS diversion, this designation is important for the hospitals to capture and serve those patients.

Similarly, the American Heart Association has a voluntary participation program (like they do for heart attack and CHF) for Get with the Guidelines in Stroke with process metrics.

Anyways, so what’s all this fuss about?

I guess even though CMS claims they will case-mix adjust these metrics, I’m a bit hesitant to think that meeting these metrics will ensure patients will get the best stroke care. I’m all about decreasing variation and providing best value care (hence I’m a CERC fellow) but I’m not sure if these metrics really reflect differences in quality.

A few weeks ago I had a patient with a hypercoagulable state secondary to metastatic cancer who we discharged home on anticoagulation as the clinical science dictates. I got a reminder email from our quality auditor asking why she didn’t get a statin at discharge (as our Comprehensive Stroke Center designation would have us monitor for). Well, I should have done a better job documenting that her abnormal liver function tests from her metastatic cancer would have precluded use of a statin. And no evidence for statin use exists in this type of patient. But I don’t think CMS will read that deeply. Without my documentation correction, we would have been dinged for that metric. 

HIAT2 Score Predicts Poor Responders to Endovascular Therapy

Matthew Edwardson, MD

Sarraj A, Albright K, Barreto AD, Boehme AK, Sitton CW, Choi J, et al. Optimizing Prediction Scores for Poor Outcome After Intra-ArterialTherapy in Anterior Circulation Acute Ischemic Stroke. Stroke. 2013

Results of recent clinical trials suggest no benefit from endovascular therapy over IV-TPA alone. The search continues to identify subgroups of patients who may respond better to endovascular therapy, thereby justifying the use of this costly but potentially beneficial intervention. To this end, Sarraj and colleagues developed the HIAT2 score to identify patients who would respond poorly to endovascular therapy regardless of recanalization rate.

The authors retrospectively selected 163 stroke patients at UT Houston with ICA or MCA occlusion who underwent intervention with IA-TPA and/or mechanical clot removal. Using logistical regression, they identified variables that predicted poor outcome to come up with the HIAT2 score: age (≤59=0, 60-79=2, ≥80 years=4), glucose (<150=0, ≥150 mg/dl=1), NIHSS (≤10=0, 11-20=1, ≥21=2) and ASPECTS (8-10=0, ≤7=3). More than 80% of patients with a HIAT2 score ≥ 5 had a poor outcome (mRS 4-6). The HIAT2 outperformed all prior scores predicting outcome following endovascular therapy.

By adding neuroimaging findings in the form of the ASPECTS to the prediction score, the HIAT2 improves our ability to identify those who will respond poorly to endovascular therapy. The HIAT2 validates what most of us were already doing in clinical practice, that is, taking patient age, exam severity, and poor prognostic signs on neuroimaging into account when selecting patients for further intervention. Considering the number of points allotted for patients with age ≥ 80, the HIAT2 also validates the practice of many centers not to consider patients ≥ 80 for endovascular therapy. Whether the HIAT2 score will change clinical management in the future remains uncertain. The HIAT2 holds promise to aid in patient selection for future clinical trials so that we can identify patients who might benefit from endovascular therapy.

SWI improves detection of cerebral microbleeds regardless of reviewer expertise.

Gillian Gordon Perue, MBBS, DM
Cerebral Microbleeds (CMBs) have become increasingly important because clinically they may represent an early imaging biomarker for asymptomatic patients at high risk of intracranial hemorrhage and/or dementia. Therefore, early detection of CMBs may provide an opportunity to intervene and slow the progression of commonly associated cerebrovascular diseases such as hypertensive vasculopathy and cerebral amyloid angiopathy.

In an elegantly designed study recently published in Stroke, Cheng and colleagues evaluate the detection and quantification of cerebral microbleeds on GRE vs. SWI among three blinded independent rates of varying levels of expertise (research expert, practicing radiologist and a 4thyear radiology resident). The authors made an effort to ensure standardization and minimize bias by training each rater on reading GRE sequences prior to the start of the study and having each rater review the studies more than 2 weeks apart in random order without clinical information. 

The study included 31 participants in a 3T MRI based study on biomarkers for cerebral amyloid angiopathy (CAA). Among these patients, 9 had CAA and 22 were healthy non- stroke controls. For each rater, the use of SWI improved the detection of CMBs when compared to detection by GRE: Rater 1 (13%, p 0.25), Rater 2 (30%, p=0.04), and Rater 3 (184%, p0.008), Kappa 0.57-0.74.

Interestingly, among the control group 9 total CMBs were found on GRE vs. 19 on SWI. This raises the question; how much is too much when it comes to microbleeds?

That is to say, “Can the presence of a few microbleeds be considered normal?” “And if so, what is the normal range?” As our imaging techniques advance and become more accessible for clinical use, the more relevant these questions become. I invite you to read the article and join the discussion on cerebral microbleeds.  

A population-based study of disability & institutionalisation after TIA and stroke: results from the Oxford Vascular Study

Peter Hannon, MD

Luengo-Fernandez R, Paul NLM, Gray AM, Pendlebury ST, Bull LM, Welch SJV, et al.Population-Based Study of Disability and Institutionalization AfterTransient Ischemic Attack and Stroke10-YeaResults of the OxfordVascular Study. Stroke. 2013

Luengo-Fernandez and colleagues utilize a large UK population-based study, the Oxford Vascular Study, to investigate the impact of TIA and stroke on disability and rate of institutionalization over 5 years. Seven hundred and forty eight index strokes and 440 TIA cases were identified from 2002-2007, with patients followed up at 1, 6, 12 and 60 months. The mean age of study participants was 74 years old.  

Disability among stroke patients jumped from 21% to 43% after their stroke, and among TIA patients climbed from a pre-morbid rate of 14% to 23% at 5 years. 47% of the stroke patients and 27% of the TIA patients died during the study. Eleven percent of stroke patients and 9% of TIA patients required long-term nursing or residential care. Significant predictors for disability included age, event severity, pre-morbid disability, subsequent vascular events, being widowed or single, diabetes and atrial fibrillation.  

Interestingly, the authors note that this study mirrored results from a study of a similar population in the early 1980’s regarding disability at 1 and 12 months after stroke, suggesting that even with advances in stroke care, there is still significant work to be done.  As you would expect, early disability after a stroke translated to a worse outcome, making it clear once again that prevention and early intervention are paramount in stroke care. The gradual increase in disability and need for long-term nursing care after TIA were notable, and it would have been very interesting to see how this compared to an age-matched healthy population control.  Overall, this study reaffirms the high cost to patient and community surrounding TIA and stroke, and specifically makes a case for stroke prevention, early intervention and special attention to follow-up care after any cerebrovascular event.

Perfusion-Diffusion Mismatch on MRI Can Predict Progression in TIA or Minor Stroke

Waimei Tai, MD

Asdaghi N, Hill MD, Coulter JI, Butcher KS, Modi J, Qazi A, et al. Perfusion MR Predicts Outcome in High-Risk Transient Ischemic Attack/Minor Stroke: A Derivation–Validation Study. Stroke. 2013

Asdaghi et. al published an interesting article recently looking at MRI performed within 24 hours of an acute ischemic event to see if it would predict progression of clinical symptoms. Using a threshold cut off of 10ml in perfusion/diffusion mismatch on the baseline MRI (performed within 24 hours of initial onset of symptoms) they demonstrated that subjects with that imaging criteria were at greater risk for both radiologic and clinical progression if infarct.

This group chose to use threshold of Tmax >4 ml as their definition of perfusion delay. This is similar to DEFUSE1 threshold (the DEFUSE authors later concluded that threshold was too sensitive and tended to overcall penumbra, and changed the threshold to Tmax >6 in the DEFUSE2 study). The use of 
Tmax >4 ml is one possible reason why Asdaghi et al. found final infarct growth of 3.5ml in the validation cohort when they had predicted >10 ml of at risk tissue on the initial acute stroke MRI. 

Granted, because this study was based on a registry, we don’t know if the perfusion mismatch found on the initial scan may have led clinicians to change their management for the patient, and thus, potentially salvaging some of that penumbral region.

How does this change management? Well, for one, if a patient has a perfusion delay (at what Tmax threshold is still controversial) then that patient likely warrants close clinical monitoring. Such a patient may need recanalization therapy if a vessel occlusion is present or possibly hypertensive therapy if no occlusion or a distal occlusion is found. Since the initial scans were performed at 24 hours after onset of symptoms, this may suggest that using radiologic criteria for reperfusion therapy makes more sense than an absolute time window.

This study confirms other research that suggests perfusion mismatch portends a higher risk of radiologic and clinical progression.

Could styloid bone be a cause of cervical carotid artery dissection?

Nandakumar Nagaraja, MD

Renard D, Azakri S, Arquizan C, Swinnen B, Labauge P, Thijs V. Styloid and Hyoid Bone Proximity Is a Risk Factor for Cervical Carotid Artery Dissection. Stroke. 2013


Cervical carotid artery dissection (CAD) as a cause of stroke is commonly encountered in patients with history of neck trauma, chiropractor manipulation of neck and connective tissue disorder such as fibromuscular dysplasia. Some patients with cervical carotid artery dissection may have only history of projectile vomiting or violent coughing prior to the event and in a few patients a definite precipitating factor may not be identified. Renard and colleagues evaluated the role of anatomic factors such as the proximity of styloid and hyoid bone to the Internal Carotid Artery (ICA) in patients with CAD. 

The authors performed a retrospective review of 88 CAD patients, 88 age and sex matched controls without dissection and 32 vertebral artery dissection patients. One rater blinded to clinical data and radiological reports evaluated the axial CTA images of the neck of these patients and measured the distance from the center of the ICA to the styloid and hyoid bone. They found that regardless of the side of the dissection patients with CAD had shorter styloid-ICA and hyoid-ICA distance compared to age and sex matched non dissection control group and vertebral artery dissection group. Among patients with CAD the styloid-ICA distance was significantly shorter on the side of dissection (ipsilateral) compared to the non dissection (contralateral) side. Similar comparison was not significant for the hyloid-ICA distance. 

The findings of this study shed light on the possibility of external mechanical compression by bony structures on ICA as a potential cause of CAD. If these results are confirmed with larger studies then in the future styloid bone surgery may be a potential treatment option for patients with CAD to prevent further events. 

If the workup for CAD reveals no identifiable precipitating cause, I would probably measure the styloid-ICA distance just for the fun of doing it though this would not change the clinical management. The mean ipsilateral styloid-ICA distance was 53.9 mm in patients with CAD in this study. Would you consider measuring the styloid ICA distance in similar patient population?

Gender disparities in IV rt-PA use

There is an observed gender disparity in thrombolytic use for treatment of acute ischemic stroke. Studies have shown women are less likely to receive IV rt-PA than men for reasons that are not entirely clear. In this interesting analysis from the Netherlands, investigators of the PRACTISE study looked at sex differences in the use of IV rt-PA. 

A total of 5515 patients were included in the study. Women were on average four years older than men. Fewer women presented to the emergency room within the 4 hour window (27% vs. 33%). In fact, onset to door time was on average 27 minutes longer in female patients. The authors found however, once patients do get to the hospital within the narrow treatment window women were treated just as often as men with IV rt-PA. Of the female patients, 41% were age 80 and above, whereas only 22% of the male patient population were in this age group. This suggests that within this cohort, more women are experiencing their first stroke at an older age compared to men because they are on average living longer than men.
In a meta-analysis of 18 studies published in 2009, the gender disparity in the use of IV rt-PA was clearly demonstrated. 

Women with acute stroke were consistently less likely to receive thrombolysis compared to men. If the difference truly lies in the timeliness of presentation, then the exact cause for this could be multi-factorial. It is be more difficult to tease out the factors contributing to the delay depending on the local conditions and population.

Concomitant headache influences long-term prognosis after acute cerebral ischemia of non-cardioembolic origin

Tareq Kass-Hout, MD

Maino A,  Algra A, Koudstaal PJ, van Zwet EW, Ferrari MD, Wermer MJH. Concomitant Headache Influences Long-Term Prognosis After AcuteCerebral Ischemia of Noncardioembolic Origin. Stroke. 2013.

Maino et al. conducted a study, recently published online in stroke, to better evaluate the pathophysiology of acute cerebral ischemia that is associated with headaches. This study shed a light on a key question, as it is still unknown whether concomitant headache reflects a partly different pathogenesis and thus may influence long-term prognosis after Transient Ischemic Attack (TIA) or minor ischemic stroke of non-cardioembolic origin.

Patients in this retrospective study were recruited from LiLAC (Life Long After Cerebral ischemia) cohort. Participants were grouped based on the presence or absence of headache at presentation. Of 2473 participants, 420 (17%) experienced headache during the acute event. Participants with headache were at lower risk of any first vascular event or vascular death (HR 0.83; 95% CI 0.71-0.97 and adjusted HR 0.73; 95% CI 0.61- 0.87, respectively). The risk of any cardiac or cerebral event, however, did not differ between the two groups (HR 0.88; 95% CI 0.67-1.14 and 0.97; 95% CI 0.76-1.24).

In short, this study suggests that patients who experienced TIA or minor ischemic stroke in association with headaches have a better vascular prognosis than those without concomitant headache. This may, at least partly, reflect a different pathogenesis.

Stroke, depression, and mortality among very old people

Vasileios-Arsenios Lioutas, MD

Carl Hornsten, Hugo Lövheim, and Yngve Gustafson. The Association Between Stroke, Depression, and 5-Year Mortality Among Very Old People. Stroke. 2013

The association between stroke, depression, and five-year mortality among very old peopleIn this population-based study, Drs Hornsten et al examine the effect of depression on mortality among stroke survivors of very old age. Depression is being increasingly recognized as a common long-term effect in stroke survivors.

In this study, depression seems to be associated with increased five-year mortality in very old (defined as ≥85 years of age) patients. Stroke patients without depression or depressed patients without stroke were found to have increased chances of survival compared to stroke patients with ongoing depression. Is it notable that although the effect appears robust in univariate analysis, adjusting for cognitive decline and functional dependence significantly weakened the association, underlining the complex interplay between those factors.

The results are interesting, but the limitations of the study do not allow for robust conclusions: The sample of stroke survivors with ongoing depression was relatively small. Additionally, depression was not diagnosed by medical record review or antidepressant use but rather by using the Geriatric Depression Scale, therefore essentially excluding a number of successfully treated patients and not allowing answer to the question whether depression per se or untreated depression only are associated with increased mortality. Further investigation of the effect of depression and its treatment in this unique but increasing subpopulation of survivors is necessary.

Hemicraniectomy for malignant stroke: New analysis from HAMLET

Claude Nguyen, MD

Geurts M, van der Worp HB, Kappelle LJ, Amelink GJ, Algra A, and Hofmeijer J. Surgical Decompression for Space-Occupying Cerebral Infarction:Outcomes at 3 Years in the Randomized HAMLET Trial. Stroke. 2013

Collectively, several signature trials assessing the effectiveness of hemicraniectomy for malignant edema after stroke have indicated that the procedure reduces mortality, although survivors tend to have moderate-to-severe disability.  To see if there was any change in this trend, Geurts et al. reviewed data at the three-year mark for patients from the Hemicraniectomy After Middle cerebral artery infarction with Life-threatening Edema Trial (HAMLET).  Unfortunately, the group did not find much improvement in the functional outcomes of both surgical and non-surgical groups at three years.  They also found no difference in the quality of life measures and depression scores between groups. 

This study does not add much to the conundrum that hemicraniectomy may not help to improve quality of life despite reducing mortality.  Although one would have hoped that such patients receiving hemicraniectomy would continue to recover over time, we know that recovery tends to plateau, and any improvement after six months, especially in the case of a large MCA stroke, would be negligible.  We should continue to keep in mind that the inclusion criteria were conservative, with no one included above the age of 60 years.  With our population living well beyond this cutoff, we will see more and more patients with malignant strokes who are much older than the patients in these trials, and therefore need revised trials to see whether hemicraniectomy should be used. 

Personally, this study has not altered my philosophy when treating these patients.  I remember seeing a patient for follow-up who was independent before her stroke, who had received a hemicraniectomy and was now a modified Rankin Scale of 5.  Despite other practitioners questioning the true benefit of surgery, the patient later reported that she was glad to have lived despite her severe disability.  As we know, evidence based medicine does not always translate to personalized medicine; in the end, we must help the patient and their family make the best decision possible.