Predicting retinal tears in posterior vitreous detachment

  • Published on

  • View

  • Download


<ul><li><p>Predicting retinal tears in posterior vitreous detachmentKelly D. Schweitzer, MD*, Amaka A. Eneh, MD*, Jonathan Hurst, BS, MD*, Mark D. Bona, MD*,Karim J. Rahim, MSc*, Sanjay Sharma, BSc, MD, FRCS, MS (Epid), MBA*</p><p>ABSTRACT RSUM</p><p>Objective: The purpose of this study is to determine whether patients with acute posterior vitreous detachment (PVD) who developdelayed retinal tears within the first 6 weeks after initial presentation have predictive characteristics.</p><p>Design: Prospective cohort study.Participants: All patients presenting to the Hotel Dieu Hospital Emergency Eye Clinic between September 2008 and July 2009</p><p>diagnosed with acute PVD were offered enrollment.Methods: At the initial visit, patients were given the previously validated Queens University Posterior Vitreous Detachment Patient Diary to</p><p>record their daily symptoms for 6 weeks. Two or 6 weeks later, patients were reexamined in detail, and their diaries were collected andanalyzed. Exact logistic regression was used to establish characteristics predictive of delayed retinal tears.</p><p>Results: In our study population of 99 patients, 2 developed delayed retinal tears. One had retinal hemorrhages and the other had acloud-like floater at initial presentation. Vitreal or retinal hemorrhage, large number of floaters at initial presentation, and high floaterfrequency at initial presentation indicated a high risk of delayed retinal tear formation, yielding a median unbiased estimated odds ratioof 36.18 with p value 0.009. No other presenting risk factors or symptomatology followed daily over the first 6 weeks after acute PVDwere predictive of delayed retinal tear formation.</p><p>Conclusions: PVD patients with retinal or vitreal hemorrhage, a significant number of floaters or a cloud like appearance to the floaters,or high floater frequency are at higher risk of developing delayed retinal tears.</p><p>Objet : Dterminer si les patients atteints dun dcollement postrieur aigu du vitr (DPV), qui dveloppent une dchirure de la rtine retardement dans les six premires semaines suivant la prsentation, ont des caractristiques prdictives.</p><p>Nature : tude prospective de cohorte.Participants : Tous les patients qui staient prsents la clinique durgence oculaire de lHpital Htel-Dieu dans les mois de</p><p>septembre 2008 juillet 2009 et avaient reu un diagnostic de DPV aigu, ont t invits participer.Mthodes : la premire visite, les patients recevaient de lUniversit Queens le Carnet de suivi du patient sur le dcollement</p><p>postrieur du vitr pour y noter quotidiennement leurs symptmes pendant six semaines. Aprs deux ou six semaines, les patientssubissaient un nouvel examen approfondi et lon recueillait et analysait les carnets. La rgression logistique exacte servait tablirles caractristiques prdictives du retard des dchirures de la rtine.</p><p>Rsultats : Parmi les 99 patients de notre tude, 2 avaient dvelopp une dchirure rtinienne retardement; lun avait une hmorragiertinienne et lautre, une apparence de corps flottants nuageux, la premire prsentation. Lhmorragie du vitr ou de la rtine ainsique le grand nombre et la frquence des corps flottants la premire prsentation indiquaient un risque lev de formation desdchirures rtiniennes retardement, donnant une probabilit moyenne estime 36,18 avec une valeur p de 0,009. Aucun autrefacteur de risque ou symptme apparent dans les suivis quotidiens des 6 premires semaines suivant le DPV aigu na prdit laformation des dchirures rtiniennes retardement.</p><p>Conclusions : Les patients ayant un DPV avec hmorragie de la rtine ou du vitr, un nombre important de corps flottants ou de corpsflottants ayant lapparence de nuages ou encore une forte frquence de corps flottants courent davantage de risque de dchiruresde la rtine retardement.Posterior vitreous detachments (PVD) are a commonage-related condition among patients aged 45 years orolder. The pathophysiology of PVD involves the separa-tion of the vitreous cortex from the internal limiting lam-ina of the retina.1 This typically causes the patient to seeflashing lights and floaters at varying amounts of severityand frequency.2 The main complications associated withPVDs are retinal detachments or breaks either at initialpresentation or at a later date.2,3 Therefore, all patientspresenting with flashes or floaters are typically examinedusing slit lamp biomicroscopy, indirect ophthalmoscopy,and scleral indentation both at initial presentation andduring a 6-week follow-up visit. PVDs have a high inci-dence in the general population and they necessitate mul-</p><p>From the *Department of Ophthalmology, Queens University, Kingston,Ont.</p><p>Originally received Feb. 26, 2011 Final revision Jun. 24, 2011 AcceptedJul. 28, 2011Correspondence to Amaka Eneh, Hotel Dieu Hospital, Department ofOphthalmology, 166 Brock Street, Kingston, ON K7L 5G2; 7aae1@</p><p></p><p>CAtiple eye examinations. Thus, the management of PVDsconsumes significant amounts of resources.4</p><p>There is considerable interest in identifying symptomsthat can predict the later development of retinal detach-ments or breaks so that follow-up visits can be reserved forpatients with serious risk factors. A recent study by vanOverdam et al.5 suggested that among patients with iso-lated PVDs, only those with more than 10 floaters, a cur-tain or cloud, vitreous hemorrhages, or retinal hemor-rhages at initial presentation need to be rescheduled for afollow-up visit. All other patients can be instructed to re-turn if the number of floaters increases.</p><p>The purpose of this study is to determine if our patientswho present with acute PVD and later develop retinal tears</p><p>Can J Ophthalmol 2011;46:4814850008-4182/11/$-see front matter 2011 Canadian Ophthalmological Society.Published by Elsevier Inc. All rights reserved.doi:10.1016/j.jcjo.2011.09.010N J OPHTHALMOLVOL. 46, NO. 6, DECEMBER 2011 481</p><p></p></li><li><p>shaers</p><p>Predicting retinal tears in PVDSchweitzer et al.display these predictive characteristics, and to look for ad-ditional predicting factors.</p><p>METHODS</p><p>Study populationAll patients presenting to the Hotel Dieu Hospital Ur-</p><p>gent Eye clinic diagnosed with acute PVD were offeredenrollment in the study. The Urgent Eye clinic is a tertiarycare centre accepting urgent referrals from ophthalmolo-gists, optometrists, and medical doctors in Kingston andsurrounding areas of southeastern Ontario. Exclusion cri-teria of the study included retinal breaks or detachments atinitial presentation, prior diagnosis of ocular disease, pa-tients under the age of 40 years or who had symptoms for1 month, or an identified history of blunt trauma to theeye. Patients with any history of ocular surgery except un-complicated phacoemulsification cataract extraction within the bag IOL insertion without posterior capsular rup-ture or vitreous loss were also excluded.</p><p>All patients were asked about risk factors for retinal tearsand underwent a full ophthalmologic examination includ-ing slitlamp biomicroscopy, indirect ophthalmoscopy, andscleral indentation. Goldmann 3-mirror examination wasused as required based on clinical examination. All patients</p><p>Fig. 1Instructions sheet outlining the procedure for recordUniversity Posterior Vitreous Detachment Patient Diary. FloaterA, B, and C represent 1-3 floaters, 3-10 floaters, and &gt;10 floatwere seen by a Queens University ophthalmology resident</p><p>482 CAN J OPHTHALMOLVOL. 46, NO. 6, DECEMBER 2011as well as an attending staff physician, vitreoretinal fellow,or vitreoretinal surgeon.</p><p>On completion of the examination and a thorough ex-planation of the nature of the diagnosis, enrollment andconsent was obtained. Each patient then received aQueens University Posterior Vitreous Detachment Pa-tient Diary. The diary included an instructions sheet(Fig. 1). The bulk of the diary centered around the mainsymptoms of PVD including the frequency and quality ofphotopsias and floaters. There was also room for any othersymptoms the patient wished to record (Fig. 2). At theinitial presentation, patient demographic information, vi-sual acuity, and risk factors for retinal tears were recordedin the diary. Risk factors included subjective vision loss,family history of retinal detachment, myopia 6D, pseu-dophakia, lattice degeneration, vitreous hemorrhage, reti-nal hemorrhage, and tobacco dust. Symptoms from thefirst day were recorded by the examining physician withthe patients input to ensure proper understanding. Foursymptoms were recorded and they included floater fre-quency, floater group (A through D), flash frequency, andflash intensity (1-10). The images of the various floatergroups (Fig. 2) were adapted from van Overdam et al.4</p><p>Groups A, B, and C represent 1-3 floaters, 3-10 floaters,and 10 floaters respectively. Group D represents a cur-</p><p>symptoms of posterior vitreous detachment in the Queenspes are represented in the lower left corner of the page. Groupsrespectively. Group D represents a curtain or cloud.ingtain or cloud (Fig. 1 and Fig. 2).4</p></li><li><p>pagd.</p><p>Predicting retinal tears in PVDSchweitzer et al.All patients were examined at 6 weeks. If additional riskfactors such as retinal or vitreal hemorrhage, or pigmentvisualized in the anterior vitreous were present on the ini-tial exam, patients had another exam 2 weeks after initialpresentation. A full examination was repeated at each visit.At every visit, the patient was reminded to return to clinicimmediately if they noticed increasing photopsias, an in-creased number of floaters, or a visual field defect. All dia-ries were collected at the 6-week visit.</p><p>Study design was approved by the Queens UniversityResearch Ethics Review Board.</p><p>Data analysisLongitudinal data from the diary was recorded on Mi-</p><p>crosoft Excel (Microsoft, Seattle, Wash) and statistical analysiswas done using version 9.2 of the SAS System for Windows(SAS Institute, Cary, NC). Retinal or vitreous hemorrhage,and multiple floaters at presentation (group C or D) weretested with exact logistic regression.5-7 In addition, any longi-tudinal changes in the 4 symptoms recorded in the diary thatmay predict a retinal tear were explored.</p><p>RESULTS</p><p>Over the course of the study, 2 patients offered enrollmentdeclined participation due to an inability to keep the 6-week</p><p>Fig. 2Daily recording pages for PVD symptoms. These pagsymptoms for 6 weeks after the initial visit. The floater groupsFloater shapes are represented in the upper right corner of the&gt;10 floaters respectively. Group D represents a curtain or cloufollow-up visit. A total of 102 patients were given a Queens</p><p>CAUniversity Posterior Vitreous Detachment Patient Diary.Three patients were lost to follow-up. Nine patients weregiven 2-week follow-up appointments. In total, we analyzed99 diaries. Of these, 2 people developed retinal tears within 6weeks. One of these patients had both vitreous and retinalhemorrhages and had been seen at a 2-week follow-up ap-pointment. The other patient reported a curtain or cloud pat-tern of floaters at presentation (group D floaters).4</p><p>In our study, the first patient presenting with a delayedretinal tear had group D floaters but also presented withthe highest floater frequency possible, and did not indicateany change in floater frequency. In terms of flashes, thispatient only reported flashes on a total of six days duringthe 6-week study period. The intensity ranged from 2 of 10to 5 of 10 and the flash frequency, when flashes occurred,was once every 18 h. The second patient presenting with adelayed retinal tear had retinal hemorrhages at presenta-tion. This patient had group A floaters throughout thestudy period and demonstrated no changes in flash fre-quency or intensity. However, this patient did report adecrease in floater frequency from once every minute toonce every hour on day 3. On day 28, there was anotherdecrease in floater frequency to once every 6 h. Regardlessof the decrease noted in floater frequency, this patientwould still have been scheduled for follow-up because he/</p><p>were duplicated to provide enough daily records to recordpted from van Overdam et al4 are in the top right hand corner.e. Groups A, B, and C represent 1-3 floaters, 3-10 floaters, andesadashe had a retinal hemorrhage at presentation.</p><p>N J OPHTHALMOLVOL. 46, NO. 6, DECEMBER 2011 483</p></li><li><p>Predicting retinal tears in PVDSchweitzer et al.In total, 15 of 99 patients in our study fulfilled thecriteria for scheduling a follow-up visit, namely, vitreal orretinal hemorrhage and group C or D floaters. Becausethese criteria captured our 2 patients who developed de-layed retinal tears, they displayed 100% sensitivity for ourstudy population. The specificity, positive predictive value,and likelihood ratios were 84.5%, 11.8%, and 6.45 respec-tively (Table 1). The median unbiased estimated odds ratio(OR) was 12.244 (p 0.0561) (Table 2). In addition, ourdata suggested that stronger predictors can be explored byconsidering the floater frequency at presentation. As bothpatients presenting had the highest floater frequency, weattempted models that included high floater frequency atpresentation. We found that when using hemorrhage,group C or D floaters, and the highest floater frequency atpresentation, the estimated the OR was 36.18 (p 0.009)(Table 2). The sensitivity remained 100%, and the speci-ficity increased to 94.8%. The likelihood ratio increased to19.2, and the positive predictive value increased to 40.0%(Table 1). Again, we used exact logistic regression to com-pensate for the small sample size.</p><p>We did not detect any changes in floater frequency,floater type, flash frequency, or flash intensity over the6-week period that were predictive of delayed retinal tears.Nor were the other risk factors recorded at presentationpredictive of delayed retinal tears.</p><p>DISCUSSION</p><p>It has been suggested that patients with floater group Cor D or vitreal or retinal hemorrhages at presentationshould be scheduled for reexamination.5 Patients not</p><p>Table 1Sensitivity, specificity, positive likelihood ratios, and pre-examination for delayed retinal tears among acute PVD patienincluding 1 additional criterion (high flash frequency at presenta</p><p>Predictor Sensitivity (%)</p><p>Hemorrhage or floaters in group C or D 100Hemorrhage or floaters in group C or D combined with</p><p>initial floater frequency of once every 6 h or higher 100Hemorrhage or floaters in group C or D combined with</p><p>initial floater frequency of once every hour or higher 100Hemorrhage or floaters in group C or D and initial floater</p><p>frequency once every minute 100</p><p>Table 2Estimated OR of the criteria for reexamination for delaOverdam et al.5 and the effects of including 1 additional criterio</p><p>Predictor</p><p>Hemorrhage or floaters in group C or DHemorrhage or floaters in group C or D combined with initial floater frequency o</p><p>6 h or higherHemorrhage or floaters in group C or D combined with initial floater frequency o</p><p>hour or higherHemorrhage or floaters in group C or D and initial floater frequency once everyNote: OR, odds ratio.</p><p>484 CAN J OPHTHALMOLVOL. 46, NO. 6, DECEMBER 2011within the above group should...</p></li></ul>


View more >