UBC Faculty Research and Publications

What is the relationship between type 2 diabetes mellitus status and the neuroradiological correlates… Funnell, Clark; Doyle-Waters, Mary M; Yip, Samuel; Field, Thalia Jan 17, 2017

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-13643_2017_Article_410.pdf [ 505.45kB ]
JSON: 52383-1.0362027.json
JSON-LD: 52383-1.0362027-ld.json
RDF/XML (Pretty): 52383-1.0362027-rdf.xml
RDF/JSON: 52383-1.0362027-rdf.json
Turtle: 52383-1.0362027-turtle.txt
N-Triples: 52383-1.0362027-rdf-ntriples.txt
Original Record: 52383-1.0362027-source.json
Full Text

Full Text

PROTOCOL Open AccessWhat is the relationship between type 2diabetes mellitus status and theneuroradiological correlates of cerebralsmall vessel disease in adults? Protocol fora systematic reviewClark Funnell1, Mary M. Doyle-Waters2, Samuel Yip3 and Thalia Field1*AbstractBackground: Cerebral small vessel disease (CSVD) is a common cause of stroke, dementia, and functional decline.In recent years, neuroradiologic correlates of CSVD have been identified. These imaging findings, best characterizedon magnetic resonance imaging (MRI), include some combination of white matter hyperintensities, lacunes,cerebral microbleeds, enlarged perivascular spaces, and cerebral atrophy. Though some cohorts have reported thatparticipants with type 2 diabetes mellitus (T2DM), an important risk factor for CSVD, may have a distinctneuroradiologic phenotype, this relationship is not well-characterized. Adults with diabetes mellitus have a two- tothreefold higher incidence of ischemic stroke compared to controls and are an increasingly important populationgiven global trends of increasing diabetes prevalence. This study aims to determine if adults with CSVD and T2DMhave a distinct neuroradiologic phenotype.Methods: A systematic search of the literature will be conducted to find articles that report the MRI features ofCSVD in a cohort of participants including those with and without type 2 diabetes mellitus (T2DM). A number ofdatabases will be searched including MEDLINE, Embase, CINAHL, and Web of Science. Proceedings and abstractsfrom key conferences will also be reviewed and relevant journals hand searched for additional papers. Thereferences from selected papers will be scanned. Screening of potential articles, data extraction, and qualityappraisal will be performed in duplicate by independent reviewers. Odds ratios and 95% confidence intervals forthe presence versus absence of each neuroradiologic correlate of interest from each included study will becalculated. If sufficient homogeneity exists among studies, a meta-analysis will be performed for eachneuroradiologic correlate of CSVD. If heterogeneity of studies precludes data pooling, results will be presented innarrative form.Discussion: Determining whether a distinct neuroradiologic phenotype of CSVD exists in adults with T2DM willprovide insight into the underlying mechanisms of CSVD and guide future research on therapeutic targets.Systematic review registration: PROSPERO CRD42016046669Keywords: Systematic review, Cerebral small vessel disease, Stroke, Magnetic resonance imaging, Type 2 diabetesmellitus, White matter hyperintensities, Lacunes, Cerebral microbleeds, Perivascular spaces* Correspondence: thalia.field@ubc.ca1Department of Medicine, Division of Neurology, University of BritishColumbia, S169-2211 Wesbrook Mall, Vancouver, BC V6T 2B5, CanadaFull list of author information is available at the end of the article© The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Funnell et al. Systematic Reviews  (2017) 6:7 DOI 10.1186/s13643-017-0410-1BackgroundCerebral small vessel disease (CSVD) is a neurodegener-ative condition affecting the small blood vessels of thebrain. In contrast to cerebral large vessels, small vesselsare not visualized by contemporary imaging methodsand therefore cerebral small vessel disease is used to de-scribe the parenchyma lesions rather than the underlyingsmall vessel alterations [1]. The neuroimaging correlatesof CSVD include lacunar infarcts, white matter hyperin-tensities, enlarged perivascular spaces, microbleeds, andbrain atrophy [2]. CSVD is a neuroradiological diagnosisand the above findings may occur in adults with or with-out a history of clinically manifest stroke or dementia[3]. However, the presence and severity of CSVD neuro-imaging correlates are associated with risk factor burden,baseline cognition and function, and prognosis with re-spect to recurrent stroke and cognitive decline [4–6].CSVD is estimated to affect 250,000 people in Canadaalone and this number is anticipated to rise as the popu-lation ages. CSVD contributes to 20% of ischemic stroke[3, 4] and 45% of dementias and the current estimateddirect (medical expenses) and indirect (lost productivity)costs of stroke and dementia in Canada exceeds $30 bil-lion annually [6, 7]. There are few known effective ther-apies for CSVD. Treatment remains empiric and isdirected at controlling vascular risk factors includinghypertension, dyslipidemia, smoking, and hyperglycemia.Unlike large artery disease, these conventional risk fac-tors may explain only a small proportion of CSVD,highlighting the need for targeted therapies.It is not known whether there is a sequential progres-sion of neuroradiological findings of CSVD, where theappearance of certain imaging features consistently pre-cedes others [7]. Further, certain neuroimaging featuresof CSVD seem to co-aggregate more consistently thanothers [5, 8]. Insights regarding risk factors for theseseparate “phenotypes” of CSVD may yield novel insightswith respect to the mechanism of this disease and directtherapeutic strategies in the future.People with diabetes are an important populationgiven the increasing global burden of the disease (esti-mated at over 400 million) and the substantially elevatedincidence of stroke (two- to threefold higher) in peoplewith diabetes compared to controls without diabetes [9].In the recent Secondary Prevention of Small SubcorticalStrokes (SPS3) trial, which included participants with re-cent lacunar infarcts, those with a history of diabeteshad distinct neuroimaging characteristics on magneticresonance imaging (MRI) as compared with those with-out diabetes, with an increased odds of posterior circula-tion infarcts and a lower burden of microbleeds andenlarged perivascular spaces [10–12]. Whether thisphenotype is consistently observed in other cohorts isnot known.The most recent systematic review examining brainimaging findings in adults with diabetes excludes dis-cussion of findings such as microbleeds and enlargedperivascular spaces that have only more recently beenwidely described in larger contemporary cohorts inthe medical literature [13]. Furthermore, this reviewcombined findings from both MRI and computedtomography (CT) studies, which introduces substan-tial heterogeneity as CT has reduced sensitivity fordetecting findings of CSVD, in particular mild-to-moderate white matter hyperintensities, cerebralmicrobleeds, or enlarged perivascular spaces. Giventhat a distinct neuroimaging phenotype of CSVD mayunderlie distinct pathophysiological mechanisms andtherapeutic strategies in adults with T2DM, we willperform an updated systematic review of the literatureto compare MRI neuroimaging features of CSVD inadults with versus T2DM.AimTo determine whether a unique neuroradiologic pheno-type of CSVD is found in adults with T2DM. This willbe achieved by comparing the presence and severity ver-sus absence of features of cerebral small vessel disease(white matter hyperintensities, lacunar infarcts, cerebralmicrobleeds, and/or enlarged perivascular spaces) onMRI in adults with T2DM versus adults without T2DM.This protocol conforms to the PRISMA-P guidelines[14] [see Additional file 1].Inclusion and exclusion criteriaOur research question does not strictly align with thetraditional “Patient, Intervention, Comparison, Out-come” (PICO) format. For clarity, we will separately de-scribe the inclusion and exclusion criteria for studyparticipants, MRI scans, exposure (diabetes mellitus, asdescribed below), and outcomes separately.ParticipantsParticipantsStudies of interest are those which include adults aged18 or older receiving brain MRI for any reason. Studiesmust include a group of patients with and without dia-betes mellitus. Patients in either group with other riskfactors for CSVD (i.e., hypertension, dyslipidemia,cigarette smoking) shall not be excluded.MRIThe MRI scanner type and the sequences obtained mustbe stated in the paper or data supplement. If such infor-mation is not described, we will attempt to contact theauthors to obtain this information. If the informationcannot be obtained, the study would be excluded. Thesequences obtained must be deemed by the reviewers toFunnell et al. Systematic Reviews  (2017) 6:7 Page 2 of 7be acceptable for determination of the imaging featuresreported in the study (see Table 1).Acceptable MRI modalities for assessing CSVD fea-tures include T1 and T2/T2 fluid-attenuated inversionrecovery (FLAIR) imaging for white matter hyperintensi-ties, lacunes, and enlarged perivascular spaces and eitherT2*-gradient echo or susceptibility-weighted imaging(SWI) for cerebral microbleeds [1].ExposureThe exposure being considered is presence of diabetes.The method used to identify patients with diabetes mustbe stated in the paper. To maximize sensitivity, multipledefinitions of diabetes status will be accepted; thosebased on fasting blood glucose, oral glucose tolerancetest, hemoglobin A1c, use of anti-diabetes medications,or physician/hospital records will be deemed acceptable.Studies which do not include a description of how pa-tients with diabetes were identified will be excluded.Specification of the duration of diabetes or a measure ofdisease severity (hemoglobin A1c, presence of end-organdamage, or history of ketoacidosis or hyperosmolar state[15]) will not be required for inclusion, though this in-formation will be extracted if available and will contrib-ute towards assessment of risk of bias. If specificallyidentified, participants with type 1 diabetes shall be ex-cluded and if diabetes type is not specified in the paper,the authors will be contacted for clarification.OutcomesPrimary outcomes include the presence of MRI featuresof CSVD (any of: white matter hyperintensities, lacunarinfarcts, cerebral microbleeds and/or enlarged perivascu-lar spaces). Secondary outcomes focus on the severity ofMRI features of CSVD as per qualitative rating scales.Studies must include a rating system for presence and/or severity of the included imaging feature(s) and mustreport the prevalence/severity of the feature(s) separatelyfor subjects with and without diabetes. Studies which donot report results separately for participants with versuswithout diabetes shall be excluded. Commonly used, val-idated rating systems (examples in Table 1) are pre-ferred, but novel rating schemes will be deemedacceptable if they are adequately described as deter-mined by consensus. Studies which use only a novel rat-ing scheme that is not adequately described (byconsensus) shall be excluded if attempts to contact theauthors do not yield adequate clarification. If studiescontain data regarding the presence and/or severity ofcerebral atrophy, we will extract these data as well. How-ever, studies reporting cortical atrophy alone are to beexcluded because this finding in and of itself is notunique to CSVD.Types of studiesWe will consider any studies that report MRI features ofCSVD (at least one of white matter hyperintensities, la-cunar infarcts, cerebral microbleeds, enlarged perivascu-lar spaces) in adults with and without T2DM. To beeligible for inclusion, a study must report MRI findingsfor groups of participants both with and without dia-betes mellitus. To maximize sensitivity, we will place nolimitations on whether the imaging rater/s is/are blindedto clinical information, whether there is a measure ofintra-rater/inter-rater reliability, or whether the level ofexperience or background (radiologist, neurologist, etc.)of the raters is specified though this information will beextracted if available and will contribute towardsTable 1 Description of neuroimaging features of cerebral small vessel disease on MRIInternational consensus definition [2] Common rating scales/definitionsExamples of featurescounted as “present” in reviewWhite matterhyperintensities(WMH)Signal abnormality of variable size in the white matter withhyperintensity on T2-weighted images without cavitation• Age-related white mat-ter changes (ARWMC)score• Fazekas score• Scheltens score• Quantitative volumetricmeasurements• ARWMC score of 5 orgreater (moderate–severe)• Fazekas periventricular = 3and/or deep ≥2• WMH volume ≥7.7 mL [3]Lacunar infarcts Round/ovoid subcortical fluid-filled cavity 3–15 mm in diameter in theterritory of one perforating arteriole≥1 lacuneCerebralmicrobleeds(CMB)Small (≤10 mm) areas of signal void with associated blooming onT2*-MRI or other sequences sensitive to susceptibility effects• Brain observermicrobleed score(BOMBS)• Microbleed anatomicalrating scale (MARS)≥1 microbleedEnlargedperivascularspaces (PVS)Fluid-filled spaces following the typical course of a vessel throughgrey or white matter, isointense to CSF on all sequences. Linear whenimages parallel to the vessel and round or ovoid when perpendicular.In constrast with lacunes, no T2-hyperintense rim on T2/FLAIR unlessthey traverse WMH• Edinburgh scoreFunnell et al. Systematic Reviews  (2017) 6:7 Page 3 of 7assessment of risk of bias. Studies will be observational;cohort, case control, and cross-sectional study designswill be eligible for inclusion. The anticipation is that thevast majority of these studies will be cross-sectional. Wewill only include studies reporting the presence versusabsence of at least one of these four MRI changes in thetotal number of participants with versus without dia-betes (2 × 2 table). These data are sufficient to calculateodds ratios [16].MethodsAdherence to preferred reporting standards for system-atic review is confirmed with the PRISMA-P checklist(Additional file 1).The study is registered with PROSPERO (CRD42016046669).Search method for identification of studiesScoping searches were undertaken to ensure it was ap-propriate to commence a systematic review. A conciseMEDLINE (Ovid) search resulting in 147 referenceswhich were reviewed by TS and CF. Records from the17 papers they selected and five known papers were ex-amined in MEDLINE (Ovid). The MeSH terms werereviewed and categorized which contributed to theconceptualization of the search and the construction ofthe draft search.The intention of the review is to retrieve studies thatinclude a population of adults with and without diabeteswho have had a brain MRI examining the presence ofCSVD. The most sensitive search would include thesearch concepts CSVD and MRI, but this results in anunmanageable number of results. Therefore, these twoconcepts will be combined with diabetes mellitus whichshould capture the majority of relevant papers. Theabove strategy may miss articles in which the studypopulation is not described in the abstract of somepapers or not all populations may be mentioned, there-fore these papers, which could be missed using the willbe captured through two other searches. Search two willinclude CSVD, MRI, and adverse effects. Search threewill include CSVD but only studies where this is themain focus of the paper which will be combined withMRI. The three searches will be combined using theBoolean operator OR and duplicates will be removed.Together, the three searches should create a sensitivesearch strategy to capture relevant studies for this review(see Fig. 1). The subject headings of studies meeting theinclusion criteria will be examined to ensure all relevantterms have been captured. If needed, additional searcheswill be undertaken.The initial search will be developed in MEDLINE andadapted for the following databases: Embase (Ovid),CINAHL (EBSCOhost) and Web of Science (ThomsonReuters). Results will be restricted to after 1985 as litera-ture before 1985 would be prior to the clinical use of MRI.A language restriction shall not be applied to the search. Ifthere are relevant non-English abstracts, attempts shall bemade to translate them wherever possible. An illustrationdescribing the search conceptualization and a draft searchare included as Fig. 1 and Additional file 2.In regard to grey literature, the proceedings of theInternational Stroke Conference, European StrokeConference, European Stroke Organization, WorldStroke Organization, the American Diabetes Associ-ation Scientific Sessions, the World Diabetes Con-gress, and the Annual General Meeting of theAmerican Society of Neuroradiology will be searched.General searches through PapersFirst (WorldCat),ProceedingsFirst (WorldCat), and Web of Science(Thomson Reuters) will also be undertaken. The web-sites of pertinent organizations will also be examinedfor papers and the names of researchers.Fig. 1 Schematic of search conceptsFunnell et al. Systematic Reviews  (2017) 6:7 Page 4 of 7Several approaches will be undertaken to increase ourretrieval of relevant articles. The journals Stroke, Inter-national Journal of Stroke, Journal of Stroke and Cere-brovascular Diseases, Lancet Neurology, Diabetes,Neuroradiology, and American Journal of Neuroradiol-ogy will be hand-searched to ensure studies have notbeen missed. These journals are considered to be of thehighest impact for the clinical subject of interest. Subjectexperts on CSVD will also be contacted to enquire aboutany studies felt to be applicable but not retrieved by oursearch strategy. Papers meeting the inclusion criteria willbe searched in the Web of Science (Thomson Reuters)and Elsevier ScienceDirect for articles citing these pa-pers. References from included papers will also bereviewed.Data collectionA record will be kept of all searches and search decisionsto ensure reproducibility. Search results will be exportedto a citation management program (EndNote ver. 7.0).Duplicates will be removed and retained separately. Theresulting references will be exported separately to thetwo reviewers for independent review using MS Excel.Selection of studiesTwo authors (CF, TF) will independently screen all titlesand abstracts identified through the literature searchesand will exclude all records clearly not meeting inclusioncriteria. Disagreements will be resolved by consensus.The selection process will be pilot tested to ensure ahigh degree of agreement between reviewers. Full text ofthe remaining studies will then be retrieved. The sametwo authors (CF, TF) will independently assess the pa-pers for fulfillment of inclusion criteria. In case of differ-ences of opinion regarding study inclusion, a thirdreview author (SY) will serve as arbiter. To avoid doublecounting, if multiple publications based on the same co-hort of participants are retrieved, only the study report-ing the largest sample size will be used. The reasons forexcluding papers for which the full text was retrievedwill be documented.Data extraction and managementA data extraction form will be used to collect detailsfrom the included studies. The form includes informa-tion on study design, patient population, and presence ofneuroimaging features of interest (see Additional file 3).Two review authors (CF and TF) will independently ex-tract the data. The data extraction form will be pilottested on several papers to ensure consistency and thatall relevant information is being captured. If necessary, astatistician will review the extraction of data to furtherensure quality and reliability. Authors will be contactedfor missing data.We will extract the MRI features of interest (whitematter hyperintensities, lacunar infarcts, cerebral micro-bleeds, and/or enlarged perivascular spaces, cerebral at-rophy) accounted for in the study as well as any ratingscales used. “Presence” versus “absence” of each featurewill be determined as per Table 1. For studies using analternate rating scale not included in Table 1, criteria for“presence” versus “absence” will be discussed betweenthe raters (CF, TF) and with an additional expert ifdeemed necessary. For each feature included, we will ex-tract 2 × 2 tables (presence versus absence of MRI find-ings in subjects with versus without diabetes) frompublications, or we will reconstruct these from data inthe publication. For studies where the data does not per-mit a 2 × 2 table construction, we will contact the studyauthors to ask for the relevant data before decidingwhether or not to exclude the study from the systematicreview. Presence of adjustment for relevant covariates(i.e., smoking status, age, hypertension, dyslipidemia)and the covariates themselves included in the analysiswill be recorded.Assessment of methodological qualityWe will assess the methodological quality of each studyusing the Quality Assessment Tool for ObservationalCohort and Cross-sectional Studies from the NationalHeart, Lung and Blood Institute [17]. This tool was de-signed to provide a framework to focus on the key con-cepts for establishing the internal validity of cohort andcross-sectional studies. Use of a more rigorous assess-ment tool is precluded by the primarily cross-sectionalnature of our data.Two authors (CF, TF) will assess methodologicalquality independently and will resolve all disagree-ments through discussion or with arbitration by athird author (SY).Data synthesis and statistical analysisWe anticipate that there may be significant heterogen-eity in the prevalence of MRI features of CSVD in sub-jects with versus without diabetes across studies. Thereare several factors that could contribute to such hetero-geneity. These factors include the following: differencesin demographic and clinical features (e.g., age, hyperten-sion, renal disease, smoking, duration and severity ofdiabetes) among study cohorts; differences in definitionsof diabetes; technical differences between MRI scannersand acquisition protocols (e.g., magnet strength, slicethickness, sequences); use of different radiological ratingscales; and differences in MRI rater skill and reliability.An I2 statistic will be calculated for the studies to be in-cluded in each proposed meta-analysis (i.e. for each neu-roradiologic correlate of interest) with values of 25, 50,Funnell et al. Systematic Reviews  (2017) 6:7 Page 5 of 7and 75% suggesting low, moderate, or high degrees ofheterogeneity, respectively [18].Scales which report a dichotomized (i.e., present orabsent) or categorical (i.e., absent, mild, moderate, se-vere) shall be harmonized for meta-analysis if deemedappropriate by our statistician. Other types of ratingscales shall not be included in a meta-analysis andthe data based on any such data scale would be pre-sented in narrative form.If significant heterogeneity between studies, as de-termined by consultation with our statistician, pre-vents meaningful pooling of the data, we will limitourselves to providing a narrative description of ob-served trends. This would include a summary of thedesign of each of the studies reporting the preva-lence of each CSVD feature and the odds ratio forpresence of that feature from each individual study.If some studies (at least two) may be pooled into ameta-analysis, then for each CSVD feature we willcalculate odds ratios and 95% confidence intervals forthe presence of the features in subjects with versuswithout diabetes from the 2 × 2 tables extracted fromthe data collection forms. Given the heterogeneity ofthe populations studied, assumption of a fixed effectsize across populations would not be justified, thusanalyses would be performed using a random effectsmodel [19]. Given the dichotomized (presence or ab-sence) or categorical (severity measure) nature of ourdata of, meta-analysis will be performed using theMantel-Haenzel method [20].If there are sufficient data to allow such analyses(in principle from as few as a single high qualitystudy, but if possible by pooling data from multiplestudies), we will perform subgroup analyses for par-ticipants with renal disease and participants withhypertension. In addition, if sufficient data are avail-able, we shall perform subgroup analyses by age anddiabetes duration.For diagnostic studies, knowledge of the mechanismsthat may induce publication bias or empirical absence forits existence is sparse (in contrast to intervention studies),however, we will attempt to assess for publication biasusing Deeks’ test [21]. Funding sources and conflict ofinterest will be extracted from included studies.Statistical analysis will be performed using RevMansoftware [22].Quality of evidenceA summary of findings table will be created. We willuse the Grading of Recommendations, Assessment,Development and Evaluation (GRADE) approach toassess the quality of evidence for the primary out-comes of interest [23].DiscussionThe neuroradiologic correlates of CSVD have only beenfully described in the last decade. Much research still re-mains into how the risk factors for CSVD lead to theemergence of the particular pathologic and neuroradio-logic correlates. Patients with diabetes mellitus representa large portion of adults with CSVD. Although any ef-fects of elevated glucose may be “dose-dependent” (i.e.,dependent on severity and duration of diabetes), wehypothesize that even dichotomized categorization ofparticipants into those with versus without diabetes willreveal significant differences between these two groups,when adjusted for confounding risk factors such as ageor history of hypertension.Whether or not such a distinct phenotype is uncov-ered by this review, additional analyses examining otherCSVD risk factors (i.e., hypertension) and their relation-ship to a distinct phenotype would further clarify pos-sible mechanisms of CSVD and help to guide targetedtherapeutic strategies.Additional filesAdditional file 1: PRISMA-P checklist. (DOCX 32 kb)Additional file 2: Search strategy. (DOCX 14 kb)Additional file 3: Data collection form. (DOCX 49 kb)AbbreviationsCSVD: Cerebral small vessel disease; EMBASE: Excerpta Medica database;FLAIR: Fluid-attenuated inversion recovery; MeSH: Medical subject heading;MRI: Magnetic resonance imaging; SWI: Susceptibility-weighted imaging;T2DM: Type 2 diabetes mellitusAcknowledgementsThe authors thank Penny Brasher for reviewing the manuscript.FundingTF is supported by a Vancouver Coastal Health Research Institute MentoredClinician Scientist Award.Availability of data and materialsData will be available on request from the authors and will be stored on anencrypted, password-protected hard drive. Requests for previously unpub-lished data contributed from other authors will be forwarded to those con-tributors for their direct agreement.Authors’ contributionsCF, TF and MDW conceptualized the review. CF drafted the protocolmanuscript. TF and MDW edited the protocol and MDW provided the searchstrategy. SY edited the manuscript. All authors read and approved the finalmanuscript.Authors’ informationNot applicable.Competing interestsThe authors declare that they have no competing interests.Consent for publicationNot applicable.Funnell et al. Systematic Reviews  (2017) 6:7 Page 6 of 7Ethics approval and consent to participateData from studies included in the review will have been obtained withparticipant consent and ethics approval. If this is not stated in themanuscript the authors will be contacted for confirmation.Author details1Department of Medicine, Division of Neurology, University of BritishColumbia, S169-2211 Wesbrook Mall, Vancouver, BC V6T 2B5, Canada.2Centre for Clinical Epidemiology and Evaluation, Research Pavilion, 708A-828West 10th Avenue, Vancouver, BC V5Z 1M9, Canada. 3Department ofMedicine, Division of Neurology, University of British Columbia, 8278-2775Laurel St., Vancouver, BC V5Z 1M9, Canada.Received: 21 May 2016 Accepted: 6 January 2017References1. Pantoni L. Cerebral small vessel disease: from pathogenesis and clinicalcharacteristics to therapeutic challenges. Lancet Neurol. 2010;9:689–701.2. Wardlaw JM, Smith EE, Biessels GJ, Cordonnier C, Fazekas F, Frayne R,Lindley RI, O’Brien JT, Barkhof F, Benavente OR, et al. Neuroimagingstandards for research into small vessel disease and its contribution toageing and neurodegeneration. Lancet Neurol. 2013;12:822–38.3. Valdes Hernandez Mdel C, Morris Z, Dickie DA, Royle NA, Munoz Maniega S,Aribisala BS, Bastin ME, Deary IJ, Wardlaw JM. Close correlation betweenquantitative and qualitative assessments of white matter lesions.Neuroepidemiology. 2013;40:13–22.4. Staals J, Booth T, Morris Z, Bastin ME, Gow AJ, Corley J, Redmond P, StarrJM, Deary IJ, Wardlaw JM. Total MRI load of cerebral small vessel diseaseand cognitive ability in older people. Neurobiol Aging. 2015;36:2806–11.5. Staals J, Makin SD, Doubal FN, Dennis MS, Wardlaw JM. Stroke subtype,vascular risk factors, and total MRI brain small-vessel disease burden.Neurology. 2014;83:1228–34.6. Benavente OR, Pearce LA, Anderson D, Bazan C, Hart RG. Abstract 66: MRIpredictors of stroke recurrence in patients with recent lacunar stroke: theSPS3 trial. Stroke. 2014;45:A66.7. Aribisala BS, Wiseman S, Morris Z, Valdes-Hernandez MC, Royle NA, ManiegaSM, Gow AJ, Corley J, Bastin ME, Starr J, et al. Circulating inflammatorymarkers are associated with magnetic resonance imaging-visibleperivascular spaces but not directly with white matter hyperintensities.Stroke. 2014;45:605–7.8. Huijts M, Duits A, van Oostenbrugge RJ, Kroon AA, de Leeuw PW, Staals J.Accumulation of MRI markers of cerebral small vessel disease is associatedwith decreased cognitive function. A Study in first-ever lacunar stroke andhypertensive patients. Front Aging Neurosci. 2013;5:72.9. Umemura T, Kawamura T, Hotta N. Pathogenesis and neuroimaging ofcerebral large and small vessel disease in type 2 diabetes: a possible linkbetween cerebral and retinal microvascular abnormalities. J DiabetesInvestig. 2016. PMID 27239779.10. Shoamanesh A, Pearce LA, Bazan C, Catanese L, McClure LA, Marti-FabregasJ, Kase CS, Hart RG, Benavente OR. Cerebral microbleeds in 1278 lacunarstroke patients: the Secondary Prevention of Small Subcortical Strokes(SPS3) trial. Stroke. 2015;46:A213.11. Ohba H, Pearce L, Potter G, Benavente O. Abstract 151: enlargedperivascular spaces in lacunar stroke patients. The Secondary Prevention ofSmall Subcortical Stroked (SPS3) trial. Stroke. 2012;43:A151.12. Palacio S, McClure LA, Benavente OR, Bazan 3rd C, Pergola P, Hart RG.Lacunar strokes in patients with diabetes mellitus: risk factors, infarctlocation, and prognosis: the secondary prevention of small subcorticalstrokes study. Stroke. 2014;45:2689–94.13. van Harten B, de Leeuw FE, Weinstein HC, Scheltens P, Biessels GJ. Brainimaging in patients with diabetes: a systematic review. Diabetes Care. 2006;29:2539–48.14. Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M, Shekelle P,Stewart LA. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev. 2015;4:1.15. Young BA, Lin E, Von Korff M, Simon G, Ciechanowski P, Ludman EJ,Everson-Stewart S, Kinder L, Oliver M, Boyko EJ, Katon WJ. Diabetescomplications severity index and risk of mortality, hospitalization, andhealthcare utilization. Am J Manag Care. 2008;14:15–23.16. Szumilas M. Explaining odds ratios. J Can Acad Child Adolesc Psychiatry.2010;19:227–9.17. Quality Assessment Tool for observational cohort and cross-sectionalstudies. [https://www.nhlbi.nih.gov/health-pro/guidelines/in-develop/cardiovascular-risk-reduction/tools/cohort]. Accessed 15 Mar 2016.18. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency inmeta-analyses. BMJ. 2003;327:557–60.19. Hedges LV, Vevea JL. Fixed-and random-effects models in meta-analysis.Psychol Methods. 1998;3:486.20. Cochrane handbook for systematic reviews of interventions. [http://www.handbook.cochrane.org/]. Accessed 30 Aug 2016.21. van Enst WA, Ochodo E, Scholten RJ, Hooft L, Leeflang MM. Investigation ofpublication bias in meta-analyses of diagnostic test accuracy: a meta-epidemiological study. BMC Med Res Methodol. 2014;14:70.22. Review Manager (RevMan) [Computer program]. Version 5.3. Copenhagen:The Nordic Cochrane Centre, The Cochrane Collaboration; 2014. http://community.cochrane.org/tools/review-production-tools/revman-5/about.23. Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Ytter Y, Alonso-Coello P,Schunemann HJ. GRADE: an emerging consensus on rating quality ofevidence and strength of recommendations. BMJ. 2008;336:924–6.•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:Funnell et al. Systematic Reviews  (2017) 6:7 Page 7 of 7


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items