Macular Hole Formation Secondary to Bacterial Septic Embolism Demonstrated by Serial Spectral-domain Optical Coherence Tomography Imaging

Abstract

Purpose: To describe a case of macular hole (MH) formation secondary to bacterial septic embolism and demonstrate the changing thereof using serial spectral-domain optical coherence tomography (SD-OCT) images.

Methods: A single case report.

Results: A 69-year-old woman who had been diagnosed with Klebsiella pneumonia septicemia was referred for decreased vision in her left eye. Impending MH was detected by SD-OCT and retinitis with multiple retinal hemorrhages by funduscopy. One week after the initial visit, a full-thickness MH had formed, and 13 weeks after the initial visit, we were able to close the MH following vitrectomy and gas tamponade. The MH was formed by retinal thinning due to outer retinal disruption without vitreomacular traction.

Conclusion: Metastatic retinal septic embolism and consequent outer retinal disruption can cause MH. Ophthalmologists must be aware of the possibility of MH in cases of retinal metastatic bacterial septic embolism, especially if the lesion infiltrates the fovea.

• bacterial septic embolism
• macular hole
• optical coherence tomography
• spectral domain
• visual acuity

Macular hole (MH) formation secondary to bacterial septic embolism is extremely uncommon; to our knowledge, only 1 case has been reported.1 We experienced a case of MH formation secondary to bacterial septic embolism and were able to observe and monitor the changes in the lesion for 13 weeks before vitrectomy. The changing of MH in this case can be demonstrated using serial spectral-domain optical coherence tomography (SD-OCT) images.

Case

A 69-year-old woman was referred for decreased vision in her left eye. She had previously been diagnosed with Klebsiella pneumonia septicemia secondary to a hepatic abscess.

At the initial visit, her best-corrected visual acuity (BCVA) was 20/28 in her right eye and 20/133 in her left eye. Slit-lamp examination revealed mild inflammatory response in the anterior chamber and anterior vitreous humor in her left eye and no inflammation in her right eye. Funduscopy showed multiple intraretinal hemorrhages in the fovea, and SD-OCT (Spectralis OCT, Heidelberg Engineering, Heidelberg, Germany) showed outer retinal disruption, subretinal punctate spots, and impending MH (Figure 1, A2, A3) in her left eye. Funduscopy and SD-OCT of her right eye revealed no abnormal findings. We recommended that the patient undergo pars plana vitrectomy (PPV) and remain in a prone position for 2 weeks afterward. However, as the patient was unable to assume this posture due to recent cervical spinal surgery, the vitrectomy was postponed, and we monitored her condition at regular intervals with SD-OCT.

Figure 1.  Sequential fundus, foveal horizontal, vertical section, surface, and outer retinal images obtained by fundus photography and spectral-domain optical coherence tomography (SD-OCT) of the left eye. (A1–A5) At the initial visit, multiple hemorrhagic infiltrations of the fovea with impending macular hole (MH) formation and an irregular margin of the outer retinal layer defect (arrow) and subretinal punctate spots (arrowhead) were observed. The outer retinal layers, including the photoreceptor layers, were disrupted, and foveal detachment was evident (the best-corrected visual acuity (BCVA) was 20/133). The measured thickness of the foveal retinal tissue was 54 µm (A2). (B1–B5) Images obtained 1 week later show slight enlargement of the outer retinal retina defect (arrow) and initiation of full thickness MH formation (arrowhead). (C1–C5) Four weeks after the initial visit, the macular hemorrhages had almost disappeared (C1). The MH (arrowhead) had expanded, and intraretinal pseudocysts had formed (C2, C3). (D1–D5) Images obtained 7 weeks after the initial visit show enlargement and additional swelling of the edges of the MH. (E2–E5) Seven weeks after surgery, the MH was closed, the intraretinal pseudocysts (arrowhead) had resolved, and the outer retinal defect had partially recovered (E3, arrow). (F1–F5) Seven months after the surgery, a closed MH and remnant foveal photoreceptor disruption were observed (the BCVA was 20/67). The measured thickness of the foveal retinal tissue was 88 µm (F2).

One week after the initial visit, the intraretinal hemorrhage had decreased slightly (Figure 1, B1), and SD-OCT showed a full-thickness MH and increased disruption of the inner segment–outer segment junction (Figure 1, B2, B3). Five weeks after the initial visit, the intraretinal hemorrhage was much decreased (Figure 1, C1), and SD-OCT showed that the edge of the hole was swollen and intraretinal pseudocysts had formed (Figure 1, C2, C3). Seven weeks after the initial visit, SD-OCT showed that the hole had enlarged (Figure 1, D2, D3), and we again recommended PPV. Thirteen weeks after the initial visit, the preoperative BCVA in the patient's left eye was 20/200. Cataract surgery, PPV, internal limiting membrane peeling by the inverted flap technique, and 14% perfluoropropane gas tamponade were performed on the left eye. Complete posterior vitreous detachment (PVD) was observed intraoperatively. Seven weeks after the operation, the patient's BCVA in her left eye was 20/133, and SD-OCT showed closure of the MH and resolution of the intraretinal pseudocysts (Figure 1, E2, E3). Seven months after the operation, the patient's BCVA in her left eye was 20/67. SD-OCT showed a closed MH with persistent disruption of the photoreceptor layer (Figure 1, F2, F3). During the course of the disease, she was treated with intravenous cefotaxime and metronidazole.

Discussion

We report the first serial SD-OCT images showing the changes in retinal anatomy during MH formation secondary to bacterial septic embolism. In our case, the intraretinal hemorrhage was observed in the fovea. This feature is not common in bacterial septic embolism patients.1 Most idiopathic MHs show incomplete PVD or significant vitreofoveal traction, which is considered to be the main mechanism underlying idiopathic MH.2–4

Our case differs from previously reported cases of idiopathic MH in 2 respects. First, serial SD-OCT images and intraoperative findings showed neither vitreofoveal traction nor tractional posterior vitreous membrane, suggesting another mechanism of MH formation. OCT imaging has been reported to be able to accurately visualize the posterior vitreous status and showed 100% agreement with actual posterior hyaloid status.5,6 These findings indicate that the complete PVD preceded MH formation.

Second, outer retinal layer disruption was observed during the initial stages of MH formation (Figure 1, A1–A5). The presence of punctate spots or inflammatory debris in the subretinal space on the initial SD-OCT image (Figure 1, A3) is an indication that retinal inflammation was the cause of the outer retinal layer disruption.7 These findings of outer retinal disruption were not reported in the original Gass classification. After MH formation, the edge of the hole becomes progressively elevated by a cuff of subretinal fluid; this is accompanied by thickening of the neurosensory retina, possibly due to retinal hydration by recruitment of vitreous fluid via the hole, which causes further elevation of the hole edges as previously.8,9

Theoretically, metastatic emboli might move through the intraretinal capillary plexus, which is located in the inner retinal layers, and thus the inner retinal layers would be involved initially. In our case, however, retinal inflammation was mostly confined to the outer retina, which might be due to the advanced stage of retinitis where the inflammation in the inner retina had propagated and covered the entire thickness of the retina. Differences between the inner and outer retina in terms of vulnerability to inflammation and ability to recover are also possible. Further studies are required to reveal the mechanism responsible for this phenomenon.

Bacterial septic embolism-related MH is rare, but our case report indicates that its clinical impact is important. Therefore, ophthalmologists should be aware of these lesions, especially if they infiltrate the fovea.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.