Should the Moyamoya Syndrome “Puff of Smoke” trigger cerebral
revascularization surgery in children with sickle cell disease?
The moya-moya syndrome describes a tangled cloud of small fragile
collateral blood vessels, resulting from the ischemic brain’s desperate
release of vascular growth signals when it does not receive adequate
blood flow. In this issue of Pediatric Blood and Cancer, theStroke in Sickle Cell Revascularization Surgery Retrospective
Study (SiSCRS) [1] reports retrospective data from 15 medical
centers to examine the benefit of cerebral revascularization surgery
(CRS) in children with sickle cell disease (SCD) who also have Moyamoya
Syndrome (SCD-MMS). Ischemic stroke is one of the most devastating
complications of SCD, leading to motor, sensory, and cognitive deficits.
Ischemia can also lead to “silent cerebral infarcts” detected only by
magnetic resonance imaging of the brain (MRI). In sickle cell anemia
without any intervention, the incidence of stroke can be 20% and the
incidence of silent cerebral infarct is as high as 39% [2].
Therapeutic progress has greatly reduced the cerebrovascular
complications of SCD. Transcranial doppler ultrasound screens for
high-risk patterns of blood flow [3], and MRI screens for silent
cerebral infarcts. Strategies to improve oxygen transport to the brain
include correcting anemia, diluting sickle erythrocytes, and reducing
hemolysis. Both chronic transfusions and hydroxyurea are used to reduce
stroke risk, and hematopoietic stem cell transplant can halt sickle cell
pathophysiology [2].
However, therapies are less effective in reducing the risk of ischemic
stroke for the subset of SCD with the most severe cerebrovascular
abnormalities: SCD-MMS. Current American Society of Hematology
guidelines “suggests evaluation for revascularization surgery in
addition to regular blood transfusion” for treating SCD-MMS combined
with a history of stroke or transient ischemic attack but categorizes
this as a “conditional recommendation based on very low certainty in
the evidence about effects” [2]. The SiSCRS group aimed to reduce
the uncertainty surrounding revascularization surgery as a treatment
using retrospective data from 15 medical centers with expertise in CRS
for SCD-MMS.
SiSCRS comparison of Surgery and Conventional treatment indicate
moderate support (OR=0.27) that CRS is associated with the post-surgical
reduction of cerebrovascular events (CVEs). As further evidence of the
benefit of CRS, the authors present the result that significantly fewer
CVEs were found post-treatment in the Surgery group compared to
pre-surgery. However, ”there were no statistically significant
differences when comparing rates of designated ischemic events (ischemic
strokes or TIAs) by group (Table 2).”
A major strength of the study is that 15 centers pooled a larger
retrospective sample of SCD-MMS than previous single-institution studies
[1]. However, even with the 78 children with CRS (Surgery group) and
63 children in the non-surgery (Conservative group), the differences
between the two groups had mixed statistical significance in
multivariate analyses. This was especially notable for those with a
history of prior cerebrovascular events (CVEs).
The analyses carefully considered age and length of risk period as
potential confounders but, as in any retrospective analysis, potential
unmeasured confounders could not be accounted for. The sample size per
medical center made it impossible to determine site effects.
Indications for CRS were not described, but referral to CRS could be
triggered when a stroke occurs despite chronic transfusion therapy
[4]. However, at least two lines of evidence suggest that, if the
CRS was done at the time that recurrences would diminish anyway, then
the benefit of CRS could be confounded with the benefit of time.
(1) Stroke rates in non-transfused SCD children decline with age,
notably between the 6-9yo compared to the 10-19yo range [5]. The
“age of treatment start” was 11.0 + 4.7 for the Surgery group
and 6.8 + 4.0 for the Conservative Group. The graphs showing the
decreased rate of CVEs with age in Figures S15- S18 resemble some of the
data from natural history. Statistical models used in the article show
an inverse correlation between age and logit-transformed data for CVEs,
but the scatter of the data is very large.
(2) The highest stroke recurrence rates were in the first 3 years after
the initial stroke for non-transfused children [6].
Additional potential confounders could be hidden by major changes in the
national standard of care that were introduced during the period of data
collection from 1990 to 2017: screening using transcranial doppler
examination [3], hydroxyurea for pediatric SCD [7], and
heterogeneous adoption of other interventions like automated
erythrocytapheresis for stroke prevention [8]. Aspirin was used more
widely in the Surgery group than the Conservative group, so that the
protective effect of aspirin against CVE might be another confounding
variable.
The adverse events reported within 30 days of surgery included 2
ischemic strokes and 2 TIAs among the 78 children with CRS. There were
no perioperative deaths and no permanent neurologic sequelae, indicating
a low risk compared to other neurosurgical procedures.
SiSCRS should be praised for providing a solid foundation for a larger
multicenter controlled study. Their data estimate effect size and
present sample sizes to make power calculations possible. They identify
confounders like age, follow-up duration, prior CVE, and aspirin usage;
a future study might need to design stratification by age. They
highlight that future multivariate analysis could examine the role of
CVE type and the time-dependent relationship between CVE history and
subsequent CVE occurrence. Whether randomized assignment to CRS vs no
CRS has equipoise is not clear. Asymptomatic MR screening for silent
infarcts might provide a cohort with early detection of SCD-MMS.
Functional data with Modified Rankin Scale could be collected across all
patients.
Thus, SiSCRS is an important retrospective analysis that suggests a
likely retrospective benefit of CRS for SCD-MMS, but must be interpreted
with caution. The benefit of CRS attains statistical significance only
in some of the comparisons. The rate of stroke in pediatric SCD
decreases with age, creating a bias in favor of the older group (i.e.,
the Surgery Group) to have fewer CVEs. Which children with SCD-MMS
benefit from CRS, and when to refer, must be demonstrated with future
studies that build upon the foundation provided by SiSCRS data.
REFERENCES
1. Robert, A., et al., Cerebral Revascularization Surgery Reduces
Cerebrovascular Events in Children with Sickle Cell Disease and Moyamoya
Syndrome Results of the Stroke in Sickle Cell Revascularization Surgery
(SiSCRS) Retrospective Study. Pediatric Blood and Cancer, in press.