Purchase this article with an account.
J. C. Pastor, I. Fernández, E. Rodríguez de la Rúa, J. M. Ruiz Moreno, J. Elizalde, C. V. Rui Martinho, M. Alfaiate, M. S. Figueroa, S. Velilla, Retina 1 Project; Retina 1 Project. Validation of a Predictive Risk Formula for Developing Proliferative Vitreoretinopathy (PVR) in Patients With Retinal Detachment (RD). Invest. Ophthalmol. Vis. Sci. 2007;48(13):5709.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To validate a predictive formula (Pastor JC et al. Curr Eye Res 2005 30:147-153) estimating the risk of development of PVR in patients with RD from new data collected in 13 centers in Spain and Portugal.
A multicenter prospective study is going on, collecting 83 pre, intra and postoperative variables (follow up 3 months). For this work data from 546 consecutive patients with RD admitted for surgery has been used to validate a formula previously developed using 335 patients collected from 5 of the participating groups 5 years ago (G0). Five groups have been considered. Global sample (GS), new patients from the 5 centers which defined the formula (G1) and patients from 3 centers with different PVR rates: G2 low (1.69%); G3 medium (8.6%); and G4 high (21%). Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) have been calculated.
Vitrectomy, more non-visible retinal breaks and re-operations were more frequently found in GS than in G0. PVR incidence (10.8%) in GS was similar to the obtained in G0 (9.9%), but sensitivity [49.2% (CI 95%:36.8%-61.6%)] and specificity [46% (CI 95%:41.6%-50.4%)] were significantly lower than previous published (sensitivity 67%, specificity 81%).In G1 (n: 143) PPV was 12.9% (CI 95%: 5.1%-28.9%) and NPV 94.6% (CI 95%: 88.8%-97.5%).In G2 (n: 60), PPV was 50% (CI 95%: 9.5%-90.5%) and NPV 98.3% (CI 95%: 90.9%-99.7%). In G3 (n: 93) PPV was 66.7% (CI 95%: 30%-90.3%) and NPV 95.4% (CI 95%: 88.8%-98.2%). In G4 (n: 101) PPV was 22.1% (CI 95%: 13.8%-33.3%) and NPV 81.3% (CI 95%: 64.7%-91.1%).
Predictive formulas for PVR based on clinical risk factors have a limited value because they are greatly influenced by characteristics of RD and its surgical management which have been modified during the last 5 years. This emphasizes the role of large multicenter, prospective studies such as Retina 1, providing an updated information on a regular basis.
This PDF is available to Subscribers Only