Perioperative central retinal artery occlusion after prone posterior cervical spine surgery treated with delayed hyperbaric oxygen therapy: a case report and targeted literature review Authors: Xinyu Tang, Yucheng Hao, Yang Yu, Kun Lin, Xianhui Kang Published: 2026 ABSTRACT BackgroundPerioperative central retinal artery occlusion (CRAO) is a rare but potentially devastating cause of visual loss after non-ocular surgery. Prone spinal and posterior cervical procedures may increase ocular risk through local orbital compression, impaired venous drainage and reduced ocular perfusion. Management is particularly difficult after surgery, when diagnosis may be delayed, conventional reperfusion strategies may be impractical or relatively contraindicated, and the role of delayed hyperbaric oxygen therapy (HBOT) remains uncertain.Case presentationA 53-year-old man with atlantoaxial subluxation underwent posterior atlantoaxial reduction and internal fixation under general anaesthesia in the prone position. Intra-operative haemodynamics remained stable; however, after the patient was returned to the supine position, a transient pressure mark and erythema were noted over the right orbit. After extubation, the patient reported blurred vision in the right eye and had a reduced right pupillary light reflex. Ophthalmological assessment showed hand-motion visual acuity, posterior pole retinal oedema and a cherry-red spot at the macula. Multimodal imaging showed findings consistent with CRAO, with fundus fluorescein angiography providing additional supportive evidence. Thrombolysis was considered relatively contraindicated in the early postoperative setting, and endovascular intervention was judged unlikely to provide meaningful benefit. As visual function remained severely impaired after initial conservative measures, HBOT was started approximately 52 h after the first postoperative visual complaint and delivered once daily at 2.0 ATA for 14 sessions. Best-corrected visual acuity later improved to 20/25 in the affected eye, with interval structural improvement on retinal imaging, although severe visual field loss persisted with only slight improvement.ConclusionsThis case highlights perioperative CRAO as a rare but serious complication of prone posterior cervical spine surgery. Early recognition and prevention of ocular compression remain essential. In selected postoperative cases where standard reperfusion options are limited, delayed HBOT may still be worth considering, although its therapeutic role requires further investigation.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12871-026-04043-z. Introduction Perioperative visual loss is an uncommon but potentially devastating complication of non-ocular surgery [1]. Among its recognised causes, central retinal artery occlusion (CRAO) is particularly serious because visual loss is often sudden, profound and difficult to reverse [2]. During prone spinal and posterior cervical procedures, local orbital compression and impaired ocular perfusion may contribute to retinal ischaemia, with systemic vascular risk factors such as diabetes mellitus potentially increasing susceptibility [1, 3]. CRAO is generally managed as an ophthalmic emergency, but treatment remains difficult. The therapeutic window is narrow, the evidence base is heterogeneous and visual recovery is often incomplete [2, 4–6]. These difficulties are amplified after surgery, when diagnosis may be delayed, thrombolysis may be relatively contraindicated and the role of delayed hyperbaric oxygen therapy (HBOT) remains uncertain [1, 2, 5, 6]. We report a case of perioperative CRAO following prone posterior cervical surgery in which clinically meaningful visual recovery occurred despite delayed initiation of HBOT. We discuss the likely perioperative mechanism, summarise clinically relevant evidence on HBOT timing, and consider the implications for similar postoperative cases in which standard reperfusion options are limited. Case presentation A 53-year-old man was admitted on 10 October 2025 with a long-standing history of chronic neck pain and headache. He was diagnosed with atlantoaxial subluxation. He had no documented history of hypertension, diabetes mellitus, dyslipidaemia, atrial fibrillation or smoking, and no known pre-existing ophthalmic disease or preoperative visual symptoms. Cervical computed tomography (CT) demonstrated an os odontoideum, and dynamic imaging showed marked C1–C2 instability with an atlanto-dental interval of 12 mm. Cervical magnetic resonance imaging (MRI) revealed hypertrophic soft tissue posterior to the C1–C2 level, thinning of the spinal cord at the corresponding level and marked posterior intramedullary signal change. Computed tomography angiography (CTA) of the carotid and vertebral arteries showed vascular anomalies: the right vertebral artery partially transitioned into the right posterior inferior cerebellar artery, while the left vertebral artery originated directly from the aortic arch. On 11 October 2025, the patient underwent posterior atlantoaxial reduction and internal fixation under general anaesthesia. The procedure was performed in the prone position, with the head secured in mild flexion using a Jackson table system, cranial pins connected to skull traction, and intra-operative neurophysiological monitoring. The pins were used to apply traction rather than to suspend the head free of facial support; the face was supported by a prone headrest during the procedure. The operation lasted approximately 3.5 h, and no intra-operative technical surgical complication was documented. Intra-operative fluid management included 3500 mL of crystalloid, 150 mL of 20% human albumin and 480 mL of cell-salvaged autologous blood. Estimated blood loss was 1000 mL, and urine output was 650 mL. The lowest measured intra-operative haematocrit and haemoglobin were 42.5% and 13.9 g/dL, respectively. Haemodynamic parameters remained stable, with systolic blood pressure ranging from 110 to 140 mmHg, diastolic blood pressure from 60 to 80 mmHg and heart rate from 60 to 70 beats per minute. After the patient was returned to the supine position at the end of surgery, a transient pressure mark and erythema were noted over the right orbit, extending to the zygomatic arch; these findings subsided spontaneously within 10 min. After extubation, the patient complained of blurred vision in the right eye. Physical examination showed a reduced pupillary light reflex on the right side. Supportive treatment was started on the day of surgery, including supplemental oxygen, ocular massage and topical intra-ocular pressure-lowering therapy with carteolol hydrochloride and brinzolamide eye drops. On postoperative day 1, comprehensive ophthalmological assessment was performed, followed by multidisciplinary consultation. Visual acuity in the right eye was limited to hand motion. The cornea was clear, the anterior chamber was clear and of normal depth, and the pupil was round with a sluggish light reflex. Intra-ocular pressure was considered normal by digital palpation, as tonometry was not available in the immediate postoperative setting. Fundoscopy showed mild optic disc pallor with relatively well-defined margins, retinal oedema at the posterior pole and a cherry-red spot at the macula. Multimodal imaging showed macular perfusion abnormality and inner retinal hyperreflectivity, consistent with acute retinal ischaemia and supporting the diagnosis of acute CRAO (Fig. 1). Fundus fluorescein angiography performed on 14 October 2025 showed delayed retinal vascular filling in the right eye. Fluorescein dye was rapidly injected through antecubital venous access. When interpreted together with the fundus appearance and OCT findings, the angiographic findings further supported the diagnosis of CRAO (Supplementary Figure S1). Fig. 1Baseline multimodal ophthalmic findings of the right eye on 12 October 2025. a Ultra-widefield colour fundus photograph of the right eye showing posterior pole retinal whitening with a cherry-red spot at the fovea (arrow). b Optical coherence tomography angiography of the right eye showing macular perfusion abnormality. c Macular optical coherence tomography of the right eye showing increased reflectivity and thickening of the inner retinal layers, consistent with acute retinal ischaemic change Baseline multimodal ophthalmic findings of the right eye on 12 October 2025. a Ultra-widefield colour fundus photograph of the right eye showing posterior pole retinal whitening with a cherry-red spot at the fovea (arrow). b Optical coherence tomography angiography of the right eye showing macular perfusion abnormality. c Macular optical coherence tomography of the right eye showing increased reflectivity and thickening of the inner retinal layers, consistent with acute retinal ischaemic change Given the patient’s early postoperative status following cervical spine surgery, thrombolytic therapy was considered relatively contraindicated, and endovascular intervention was judged unlikely to provide meaningful clinical benefit. After multidisciplinary discussion on postoperative day 1, conservative treatment was continued. Intermittent ocular massage and intra-ocular pressure-lowering therapy were maintained, together with vasoactive treatment using intravenous alprostadil. Additional supportive medication included citicoline and oral mecobalamin. Despite these initial measures, visual function remained severely impaired, with no meaningful early improvement. HBOT was therefore started on postoperative day 2, approximately 52 h after the first postoperative visual complaint, after logistical arrangements had been completed. Treatment was delivered once daily at 2.0 ATA with a 65-min plateau time per session, for a total of 14 sessions. This regimen reflected the local hyperbaric medicine service’s routine practice for retinal ischaemic disorders, informed by published literature, hyperbaric medicine consensus recommendations and institutional experience, with the aim of increasing dissolved oxygen delivery to the ischaemic retina while maintaining tolerability in the early postoperative period. Published HBOT protocols for CRAO vary in treatment pressure, session duration, treatment frequency and total number of sessions, and no universally accepted regimen has been established [7–11]. The total number of sessions was determined according to the planned treatment course and clinical response. Compound anisodine hydrobromide was also administered by subcutaneous injection adjacent to the right superficial temporal artery as part of adjunctive ophthalmological treatment. During follow-up after HBOT and adjunctive treatment, the patient demonstrated marked improvement in central visual acuity. On 29 October 2025, best-corrected visual acuity was 20/25 in the right eye and 20/20 in the left eye. Examination of the right eye showed a clear cornea, a quiet anterior chamber, a round pupil with a positive relative afferent pupillary defect, a well-defined optic disc margin and an attached retina within the observable field. Serial retinal imaging showed interval structural improvement during follow-up, as illustrated by the serial fundus photographs and OCT images in Fig. 2. The follow-up OCT images on 18 and 29 October 2025 provide the most direct structural comparison (Fig. 2b, d). Visual field testing showed persistent severe field loss with slight interval improvement (Supplementary Figure S2). The patient was discharged in stable condition on 30 October 2025 after otherwise uneventful postoperative recovery. Fig. 2Serial multimodal retinal imaging during follow-up after hyperbaric oxygen therapy. a Ultra-widefield colour fundus photograph of the right eye obtained on 18 October 2025. b Macular optical coherence tomography (OCT) B-scan of the right eye obtained on 18 October 2025. c Colour fundus photograph of the right eye obtained on 29 October 2025. d Macular OCT B-scan with flow overlay of the right eye obtained on 29 October 2025. Because follow-up imaging was performed according to clinical availability at each visit, the imaging modalities were not fully identical across time points. Despite this, the serial examinations demonstrated interval structural improvement during follow-up after hyperbaric oxygen therapy Serial multimodal retinal imaging during follow-up after hyperbaric oxygen therapy. a Ultra-widefield colour fundus photograph of the right eye obtained on 18 October 2025. b Macular optical coherence tomography (OCT) B-scan of the right eye obtained on 18 October 2025. c Colour fundus photograph of the right eye obtained on 29 October 2025. d Macular OCT B-scan with flow overlay of the right eye obtained on 29 October 2025. Because follow-up imaging was performed according to clinical availability at each visit, the imaging modalities were not fully identical across time points. Despite this, the serial examinations demonstrated interval structural improvement during follow-up after hyperbaric oxygen therapy Discussion This case is worth discussing for two reasons that are directly relevant to perioperative practice. First, it occurred in a recognised high-risk perioperative setting—prone posterior cervical surgery—where perioperative visual loss remains uncommon but potentially devastating [1]. Second, meaningful visual recovery occurred despite HBOT being initiated more than 48 h after symptom onset, at a point when conventional reperfusion strategies would generally be considered impractical because of the elapsed time and potentially contraindicated in the early postoperative setting [2]. CRAO is an ophthalmic emergency, and early recognition remains central to any chance of visual salvage [2]. Postoperative cases, however, are different. Diagnosis may be delayed, thrombolysis is often not feasible, and treatment decisions are commonly made under considerable uncertainty [1, 2]. Against this background, the value of the present case lies not in suggesting that delayed treatment can be relied upon, but in illustrating a clinically important grey zone: recovery may still occur in selected postoperative patients even when the most favourable window has been missed. The key question, therefore, is how such recovery should be interpreted. On balance, a multifactorial mechanism appears most plausible, with external orbital compression likely to have been an important contributor [1, 12]. Although cranial pins were used for skull traction, the head was not suspended free of facial support; therefore, direct or peri-orbital contact with the prone headrest remained possible. The transient pressure mark and erythema over the right orbit immediately after repositioning support that possibility. Although these findings do not establish causation, they make a purely spontaneous vascular event less convincing as the sole explanation. The operative context also supports a pressure-related component. Although there was no evidence of severe perioperative anaemia, the intra-operative blood loss may have narrowed the margin of safety in the presence of any pressure-related reduction in ocular perfusion. Importantly, stable systemic haemodynamics do not exclude local ocular hypoperfusion when direct orbital pressure or impaired venous drainage is present [1, 12]. Prone positioning, head position and venous congestion can all affect intra-ocular pressure and ocular perfusion during anaesthesia [1, 12]. Similar perioperative cases after prone cervical or spinal procedures have been reported, including cases accompanied by periorbital swelling or ophthalmoplegia [3, 13]. The mechanism proposed here therefore remains inferential, but it is clinically plausible because it is grounded in the operative setting, the immediate postoperative orbital findings and similar patterns described in previous perioperative reports [3, 13]. Current evidence still favours early HBOT in CRAO, particularly when treatment can be started within the first 24 h [4, 8, 14, 15]. In retrospective cohort and case-series data, shorter onset-to-HBOT intervals have generally been associated with greater visual improvement [4, 8, 9, 14, 15]. In particular, Kim et al. reported the greatest visual benefit when HBOT was initiated within 9 h, and St Peter et al. found a stronger improvement signal among patients treated within 8 h [4, 14]. Gur et al. similarly found that shorter time to HBOT initiation was independently associated with greater short-term BCVA improvement, while the initial HBOT protocol also appeared to influence early outcome [9]. Other recent data support consideration of HBOT within 24 h, although the observed benefit appears to be influenced by patient selection, treatment completion and protocol heterogeneity [8, 9, 11, 15]. In addition to timing, HBOT protocols for CRAO remain heterogeneous. Reported regimens differ in treatment pressure, session duration, treatment frequency and total number of sessions, and current evidence does not define a single optimal protocol. The present regimen should therefore be interpreted as a pragmatic local protocol applied in an early postoperative patient, rather than as a universally accepted CRAO regimen. Nevertheless, this evidence base has important limitations. Study designs vary, treatment protocols differ, and most published clinical data remain retrospective or observational [4–6, 8, 9, 14–16]. Recent systematic reviews have therefore reached cautious rather than definitive conclusions. Rosignoli et al. did not show a clear improvement in final visual outcome, whereas Elfil et al. reported potential BCVA benefit and a favourable safety profile but emphasised heterogeneity and the need for prospective, standardised studies [5, 6]. The most defensible interpretation is therefore that current evidence supports early treatment, while the limits of reversibility remain uncertain. Recent and clinically relevant evidence is summarised in Table 1 [4–6, 8, 9, 11, 14–21]. Most contemporary studies still favour treatment within 24 h, whereas a smaller body of comparative and case-based evidence suggests that visual recovery may still occur beyond the conventional window in selected patients [17, 18]. Table 1Evidence relevant to HBOT timing, protocol heterogeneity and outcome interpretation in retinal artery occlusionAuthor, yearDesignPopulationTreatment timingKey finding and relevance Section A. Evidence supporting early HBOT, especially within 24 h Kim, 2025 [4]Retrospective single-centre cohortCRAO treated with HBOTMedian 9.5 h; ≤9 h vs. > 9 hEarlier HBOT was associated with greater visual improvement, strongest within 9 h. Supports shorter onset-to-treatment intervals. Thoongsuwan, 2025 [8]Retrospective cohortAcute CRAO; HBOT vs. non-HBOTWithin 24 hHBOT was associated with better visual and structural outcomes. Supports early HBOT while illustrating protocol and outcome heterogeneity. St Peter, 2023 [14]Retrospective case seriesNon-arteritic CRAO treated with HBOT< 24 h; subgroup < 8 hVisual acuity improved overall, with a stronger signal within 8 h. Supports a graded timing effect. Gur, 2026 [9]Retrospective historically controlled cohortAcute non-inflammatory CRAO; different HBOT protocolsMedian 6 h; 75% within 10 h; 5% beyond 24 hHigher initial pressure and shorter onset-to-HBOT latency were associated with greater short-term BCVA improvement. Supports timing sensitivity and protocol heterogeneity. Orwig, 2025 [15]Retrospective cohortNon-arteritic CRAO treated with HBOTWithin 24 hVisual improvement was reported among patients completing treatment. Supports early HBOT and the importance of treatment completion. Kim, 2020 [19]Comparative observational studyCRAO; adjunctive HBOT vs. no HBOTWithin 24 hAdjunctive HBOT was associated with greater visual improvement. Provides supportive comparative evidence for early treatment. Section B. Evidence relevant to delayed or time-window-limited HBOT decisions Yang, 2025 [17]Comparative observational studyNon-arteritic CRAO; conservative treatment, intra-arterial fibrinolysis or HBOTBeyond conventional windowHBOT beyond the conventional window was associated with greater visual improvement than conservative treatment. Supports possible therapeutic value beyond the usual early window. Werling, 2026 [18]Case reportSingle CRAO caseBeyond 24 hVisual recovery occurred after HBOT beyond 24 h. Provides delayed-treatment case evidence. Hadanny, 2017 [20]Retrospective cohortNon-arteritic CRAO treated with HBOTVariable timing; interpreted with funduscopic findingsSuggested that reversibility may not depend on elapsed time alone. Supports residual-viability-based selection. Section C. Balancing and contextual evidence Rosignoli, 2022 [5]Retrospective study, systematic review and meta-analysisCRAO / RAO literatureVariable timing and protocolsDid not show clear improvement in final visual outcome. Provides balancing evidence against treating HBOT as established therapy. Elfil, 2025 [6]Systematic review and meta-analysisRAO; HBOT vs. controlVariable timing and protocolsReported potential BCVA benefit and favourable safety, but emphasised heterogeneity and need for prospective studies. Provides updated balancing evidence. Rozenberg, 2022 [16]Retrospective comparative analysisNon-arteritic CRAOVariable timing; HBOT added to standard careHBOT added to standard care was associated with better final visual outcome and appeared safe. Supportive, but less timing-specific. Celebi, 2021 [11]Narrative review / perspectiveCRAO considered for HBOTEmphasises patient selection and timingSummarises patient selection, timing and practical considerations. Provides contextual support for individualised decision-making. Au and Chong, 2025 [21]Retrospective prognostic cohortAcute CRAO treated with HBOTVariable; prognostic analysisThinner parafoveal thickness and shorter FFA delay were associated with better visual outcomes. Supports imaging-based outcome interpretation.This table summarises selected evidence relevant to treatment timing, protocol heterogeneity, patient selection and outcome interpretation in the present case. It is not an exhaustive systematic review. Section B includes studies directly addressing delayed treatment and evidence suggesting that retinal reversibility may not be determined by elapsed time aloneAbbreviations: BCVA Best-corrected visual acuity, CRAO Central retinal artery occlusion, HBOT Hyperbaric oxygen therapy, RAO Retinal artery occlusion Table 1Evidence relevant to HBOT timing, protocol heterogeneity and outcome interpretation in retinal artery occlusionAuthor, yearDesignPopulationTreatment timingKey finding and relevance Section A. Evidence supporting early HBOT, especially within 24 h Kim, 2025 [4]Retrospective single-centre cohortCRAO treated with HBOTMedian 9.5 h; ≤9 h vs. > 9 hEarlier HBOT was associated with greater visual improvement, strongest within 9 h. Supports shorter onset-to-treatment intervals. Thoongsuwan, 2025 [8]Retrospective cohortAcute CRAO; HBOT vs. non-HBOTWithin 24 hHBOT was associated with better visual and structural outcomes. Supports early HBOT while illustrating protocol and outcome heterogeneity. St Peter, 2023 [14]Retrospective case seriesNon-arteritic CRAO treated with HBOT< 24 h; subgroup < 8 hVisual acuity improved overall, with a stronger signal within 8 h. Supports a graded timing effect. Gur, 2026 [9]Retrospective historically controlled cohortAcute non-inflammatory CRAO; different HBOT protocolsMedian 6 h; 75% within 10 h; 5% beyond 24 hHigher initial pressure and shorter onset-to-HBOT latency were associated with greater short-term BCVA improvement. Supports timing sensitivity and protocol heterogeneity. Orwig, 2025 [15]Retrospective cohortNon-arteritic CRAO treated with HBOTWithin 24 hVisual improvement was reported among patients completing treatment. Supports early HBOT and the importance of treatment completion. Kim, 2020 [19]Comparative observational studyCRAO; adjunctive HBOT vs. no HBOTWithin 24 hAdjunctive HBOT was associated with greater visual improvement. Provides supportive comparative evidence for early treatment. Section B. Evidence relevant to delayed or time-window-limited HBOT decisions Yang, 2025 [17]Comparative observational studyNon-arteritic CRAO; conservative treatment, intra-arterial fibrinolysis or HBOTBeyond conventional windowHBOT beyond the conventional window was associated with greater visual improvement than conservative treatment. Supports possible therapeutic value beyond the usual early window. Werling, 2026 [18]Case reportSingle CRAO caseBeyond 24 hVisual recovery occurred after HBOT beyond 24 h. Provides delayed-treatment case evidence. Hadanny, 2017 [20]Retrospective cohortNon-arteritic CRAO treated with HBOTVariable timing; interpreted with funduscopic findingsSuggested that reversibility may not depend on elapsed time alone. Supports residual-viability-based selection. Section C. Balancing and contextual evidence Rosignoli, 2022 [5]Retrospective study, systematic review and meta-analysisCRAO / RAO literatureVariable timing and protocolsDid not show clear improvement in final visual outcome. Provides balancing evidence against treating HBOT as established therapy. Elfil, 2025 [6]Systematic review and meta-analysisRAO; HBOT vs. controlVariable timing and protocolsReported potential BCVA benefit and favourable safety, but emphasised heterogeneity and need for prospective studies. Provides updated balancing evidence. Rozenberg, 2022 [16]Retrospective comparative analysisNon-arteritic CRAOVariable timing; HBOT added to standard careHBOT added to standard care was associated with better final visual outcome and appeared safe. Supportive, but less timing-specific. Celebi, 2021 [11]Narrative review / perspectiveCRAO considered for HBOTEmphasises patient selection and timingSummarises patient selection, timing and practical considerations. Provides contextual support for individualised decision-making. Au and Chong, 2025 [21]Retrospective prognostic cohortAcute CRAO treated with HBOTVariable; prognostic analysisThinner parafoveal thickness and shorter FFA delay were associated with better visual outcomes. Supports imaging-based outcome interpretation.This table summarises selected evidence relevant to treatment timing, protocol heterogeneity, patient selection and outcome interpretation in the present case. It is not an exhaustive systematic review. Section B includes studies directly addressing delayed treatment and evidence suggesting that retinal reversibility may not be determined by elapsed time aloneAbbreviations: BCVA Best-corrected visual acuity, CRAO Central retinal artery occlusion, HBOT Hyperbaric oxygen therapy, RAO Retinal artery occlusion Evidence relevant to HBOT timing, protocol heterogeneity and outcome interpretation in retinal artery occlusion This table summarises selected evidence relevant to treatment timing, protocol heterogeneity, patient selection and outcome interpretation in the present case. It is not an exhaustive systematic review. Section B includes studies directly addressing delayed treatment and evidence suggesting that retinal reversibility may not be determined by elapsed time alone Abbreviations: BCVA Best-corrected visual acuity, CRAO Central retinal artery occlusion, HBOT Hyperbaric oxygen therapy, RAO Retinal artery occlusion HBOT may plausibly benefit selected patients with CRAO by increasing dissolved oxygen content in plasma and enhancing oxygen diffusion from the choroidal circulation to metabolically compromised inner retinal tissue while retinal arterial perfusion remains impaired [11]. This mechanism is most relevant when retinal ischaemia is incomplete, intermittent or not yet anatomically irreversible, rather than when infarction is fully established [11, 20]. In the present case, the initial fundus appearance and OCT findings, together with subsequent FFA findings, supported acute CRAO, but the subsequent improvement in central visual acuity and interval structural improvement on OCT suggest that at least part of the macular tissue may have remained viable despite delayed treatment. This interpretation is consistent with recent prognostic work in HBOT-treated CRAO, in which thinner parafoveal macular thickness in the 1–3 mm zone and shorter FFA perfusion delay were associated with better visual outcomes [21]. However, our case did not include formal quantitative OCT thickness analysis or standardised FFA transit-time measurement, so this comparison should be interpreted qualitatively. The persistent severe visual field loss despite good final central acuity also indicates that structural and central visual recovery did not translate into complete functional recovery. The present case is best understood within this grey zone. It does not argue against early treatment; rather, it reinforces why earlier treatment remains preferable whenever feasible. It does suggest that elapsed time alone may not fully capture reversibility in selected patients, particularly when the clinical context differs from typical spontaneous CRAO [11, 17, 18, 20]. This is most relevant when the clinical context points to pressure-related hypoperfusion rather than a purely embolic event. One possible explanation is that pressure-related hypoperfusion was incomplete, intermittent or partially relieved after repositioning, leaving residual tissue viability beyond the window usually assumed for a fixed arterial occlusion. This interpretation cannot be proven, but it is consistent with the observed clinical course. This may help explain why improvement was still observed despite the clearly suboptimal timing of HBOT [17, 18]. Seen in that light, the case should not be read as evidence that treatment windows are unimportant. Rather, it is a reminder that perioperative CRAO may be pathophysiologically heterogeneous, and that treatment decisions sometimes have to be made in a zone of uncertainty. In such selected postoperative patients, particularly when conventional reperfusion strategies are not feasible and the clinical context points to pressure-related hypoperfusion rather than a clearly fixed embolic event, delayed HBOT may still have therapeutic value beyond the conventional early window. This possibility deserves further clinical evaluation. For anaesthesiologists and perioperative teams, this case offers a practical framework rather than proof of a treatment effect: prevent avoidable ocular compression, recognise visual symptoms early, and escalate decision-making rapidly when CRAO is suspected. Prevention remains the first priority. During prone cervical or spinal surgery, the eyes and periorbital tissues should be checked after final positioning and reassessed whenever the head, table position or traction is adjusted, to ensure that there is no direct ocular or orbital pressure [1]. After surgery, any new visual complaint, pupillary abnormality, periorbital swelling or orbital pressure mark should be treated as a time-sensitive warning sign, prompting urgent ophthalmological assessment and multidisciplinary discussion [1, 2]. In the early postoperative period, conventional reperfusion strategies such as thrombolysis may be contraindicated, logistically unavailable or judged disproportionate to the expected benefit [2]. When standard reperfusion options are limited, delayed HBOT may be worth considering in carefully selected postoperative cases even after the most favourable window has passed [6, 17, 18, 20]. Nevertheless, this remains a single observation and should not be used to overstate what delayed treatment can achieve. Functional assessment was limited to 30 − 2 automated perimetry; 10 − 2 testing and manual kinetic perimetry were not performed, which limited more precise delineation of the residual central visual field. The mechanism is inferred rather than proven; spontaneous or delayed reperfusion cannot be excluded; and the contribution of HBOT cannot be separated from the natural course or from other concurrent measures. The case therefore supports prevention, vigilance, early escalation and multidisciplinary problem-solving, not therapeutic certainty. Conclusion This case highlights perioperative CRAO as a rare but serious complication of prone posterior cervical spine surgery. Prevention of ocular compression and early recognition of postoperative visual symptoms remain central to management. Although early treatment is preferred, delayed HBOT may still be considered in selected postoperative cases when pressure-related hypoperfusion is plausible and standard reperfusion options are unavailable. Supplementary Information Supplementary Material 1. Figure S1: Fundus fluorescein angiography of the right eye performed on 14 October 2025. a Infrared baseline image. b Early-phase fluorescein angiography image at 0:16.77. c Early-to-mid phase fluorescein angiography image at 0:20.78. d Late-phase fluorescein angiography image at 13:34.95. Serial fluorescein angiography showed delayed retinal vascular filling, which, in conjunction with the fundus and OCT findings, further supported the diagnosis of central retinal artery occlusion. Supplementary Material 1. Figure S1: Fundus fluorescein angiography of the right eye performed on 14 October 2025. a Infrared baseline image. b Early-phase fluorescein angiography image at 0:16.77. c Early-to-mid phase fluorescein angiography image at 0:20.78. d Late-phase fluorescein angiography image at 13:34.95. Serial fluorescein angiography showed delayed retinal vascular filling, which, in conjunction with the fundus and OCT findings, further supported the diagnosis of central retinal artery occlusion. Supplementary Material 2. Figure S2: Serial visual field examinations of the right eye during follow-up after hyperbaric oxygen therapy. a Visual field examination of the right eye performed on 19 October 2025. b Visual field examination of the right eye performed on 29 October 2025. These follow-up examinations showed persistent severe visual field loss with slight interval improvement. Supplementary Material 2. Figure S2: Serial visual field examinations of the right eye during follow-up after hyperbaric oxygen therapy. a Visual field examination of the right eye performed on 19 October 2025. b Visual field examination of the right eye performed on 29 October 2025. These follow-up examinations showed persistent severe visual field loss with slight interval improvement.