LAUSR

Dear Editor, Descemet stripping automated endothelial keratoplasty (DSAEK) and Descemet membrane endothelial keratoplasty (DMEK) have become standard posterior corneal replacement techniques due to excellent visual outcomes and faster visual rehabilitation compared with traditional full-thickness transplants [1–3]. Graft adherence is achieved by injecting an air or gas bubble into the anterior chamber. However, graft detachments remain the most common complication of endothelial keratoplasties (EK) [3, 4]. Postoperative head positioning has been thought to help prevent graft dislocations by maximizing the tamponade effect of the air or gas bubble on the graft. However, patient adherence to positioning has not been well-studied. We assessed patient adherence to supine positioning after EK using a head positioning sensor. Patients undergoing either DMEK or DSAEK at Gavin Herbert Eye Institute, University of California, Irvine between December 2017 and May 2018 were included. This study was approved by the institution’s Institutional Review Board. The head positioning sensor uses the MetaWear C sensor board (MbientLab, San Francisco, CA) (Fig. 1a), which comprises a 3-axis accelerometer and gyroscope. The device is Bluetooth-enabled such that it connects to an iPhone (Apple, Cupertino, CA) via the MetaWear iOS application interface. The sensor is attached to an adjustable headband that is fitted around a patient’s head (Fig. 1b). The iOS application records head angle data in spherical angle coordinates and records a binary output of whether the head angle is in the correct position. This device was previously validated in a pilot study of healthy volunteers [5, 6]. A net deviation > 45° from the calibrated reference point in any axis or combination of axes was defined as being out of position for this report. This threshold was determined by simulating postoperative positioning in a healthy subject. We used 20% SF6 in DMEK and air in DSAEK. Postoperatively, patients were fitted with the device, which was then calibrated to set the supine head position as the reference point. The MetaWear iOS application collected data every 30 s for up to 24 h. Patients were instructed to wear the headband continuously until their postoperative day one visit and were given an iPod with the MetaWear iOS application. At the postoperative visit, the device was returned and the data was uploaded. All patients were asked whether they removed the headband at any point; patients who wore the device < 8 hours were excluded from the analysis. A total 36 patients (43 eyes) underwent EK for either Fuchs’ dystrophy or pseudophakic bullous keratopathy. Nine patients declined to participate; three patients were not asked to participate due to cognitive decline. Twenty-one patients (22 eyes; 11 DMEK, 11 DSAEK) tolerated the head sensor. Nine patients were excluded from the analysis due to user error or premature removal of the device. The mean duration of time patients wore the device was 12.5 (± 5.7) hours. The mean and median percentage time out of position was 31.0% (± 19.8%) and 30.8% (range, 7.2–72.6%), respectively. The mean percentage time out of position was statistically similar between the first 8 hours of wear compared with the entire recorded duration (p = 0.98). In regard to graft status, 86% of the eyes had completely attached grafts on postoperative day one. Three (14%) eyes (1 DMEK, 2 DSAEK) had a small detachment (< 20%); no eyes required rebubbling. Both DSAEK eyes had coexisting ocular morbidities and a history of multiple intraocular surgeries. The percentage time out of position for the three patients with partially detached grafts was 30.1%, 35.7%, and 42.1%. * Elizabeth Shen [email protected]