Clinical features of retinopathy after cardiopulmonary resuscitation

Our study showed that retinopathy, including retinal hemorrhage and cotton wool spots, could occur after CPR in adults, and when retinopathy occurred, the CPR time was longer than in other cases. There were no other factors that had a significant influence on the occurrence of retinopathy after CPR, including underlying disease, cause of arrest, laboratory tests and others. Retinopathy cases presented with retinal hemorrhages with or without cotton wool spots, and the main feature of retinal hemorrhages was intraretinal hemorrhage. In some cases, RNFL hemorrhage was observed. These retinal hemorrhages were observed in the mid-periphery in all cases and were distributed to the peripapillary or posterior pole in 80% of the cases. All retinal hemorrhages and cotton wool spots disappeared over time without deterioration.

Retinal hemorrhage in children receiving CPR has been documented in several studies, with an incidence rate of 11–27% [5‐9]. However, according to results of these studies, when abusive head trauma was excluded, retinal hemorrhage associated with CPR was confirmed in 1 out of 43 children by Odom et al. [5], 1 out of 22 children by Pham et al. [6], and 1 out of 38 children by Binenbaum et al. [7], and the average prevalence was approximately 3.3%. Retinal hemorrhage associated with CPR in adults has been reported in only a few cases [13, 14], and its prevalence is unknown. As our study was retrospective and investigated patients who visited the ophthalmic clinic with various symptoms, it was not possible to confirm the exact prevalence. However, as 10 of the 47 (21%) patients had medical records and 10 of the 24 (42%) patients met the inclusion criteria and had no medical history of trauma, the rate of retinal hemorrhage associated with CPR may be higher than that in children.

The patterns of retinal hemorrhage associated with CPR were superficial intraretinal peripapillary hemorrhage in a report by Binenbaum et al. [7] and scattered peripapillary hemorrhage in pediatric cases according to Pham et al. [6] In adults, retinal hemorrhages present as Valsalva retinopathy in the form of peripapillary retinal hemorrhage, Terson syndrome-like appearance, and vitreous hemorrhage [13, 19, 20]. Most of the retinal hemorrhages in our study were intraretinal hemorrhages, similar to those observed in pediatric studies. However, cotton wool spots were found in one-third of the eyes with retinopathy, which was different from the results of other studies. In contrast, in other studies of pediatric and adult cases showing hemorrhages around the optic disc, the hemorrhage was mainly in the mid-periphery, and 60% of the hemorrhages were around the optic disc in our study.

The cause of retinal hemorrhage after CPR remains unclear [21, 22], but several hypotheses have been proposed. In relatively early studies, it was thought that increased thoracic pressure due to chest compression impeded intraocular venous drainage by increasing intracranial pressure through the jugular venous pressure and the paravertebral venous plexus [9, 21, 23]. As normal jugular vein valves prevent the retrograde flow of increased thoracic pressure, this hypothesis can only be applied in cases of jugular valve damage [24]. The pressures of the aorta and jugular vein bulb during CPR are 61 mmHg and 29 mmHg at peak compression and 23 mmHg and 15 mmHg at relaxation, respectively, which are lower than normal [25]. For this reason, besides increased thoracic pressure, hypoxia, Valsalva retinopathy, and Purtscher’s retinopathy have been suggested as possible mechanisms [9, 19, 21, 26]. Binenbaum et al. [7] argued that hypoxia could be more severe in out-of-hospital cardiac arrest; however, because retinal hemorrhage does not appear more frequently in out-of-hospital cardiac arrest, hypoxia is not a significant factor related to retinal hemorrhage. Valsalva retinopathy can occur when there is damage to the jugular vein valve; therefore, it is difficult to identify this as a major mechanism related to the development of retinal hemorrhage. In our study, the hemorrhage pattern did not show the characteristic findings of Valsalva, such as pre-retinal hemorrhage and sub-internal limiting membrane hemorrhage [27].

Goetting and Sowa and Weedn et al. [21, 22] reported that the cause of retinal hemorrhage observed in their study could be Purtscher’s retinopathy associated with chest compression. At that time, an increase in intrathoracic pressure due to chest compression was known to be the cause of Purtscher’s retinopathy; therefore, the mechanism was similar to previous theories. Purtscher’s retinopathy is an occlusive microvasculopathy that develops after trauma and presents with retinal findings, including cotton wool spots, retinal hemorrhages, and Purtscher’s flecken [28, 29]. Fundus findings in this study showed lesions similar to those of Purtscher’s retinopathy, such as superficial to intraretinal hemorrhages and cotton wool spots, although there were no Purtscher’s flecken. In a systematic review, Purtscher’s flecken was present in only approximately 60% of patients, and in the study by Agrawal and McKibbin, Purtscher’s flecken was found in 3 out of 10 patients who received chest compression [28, 29]. For this reason, although there was no Purtscher’s flecken, it is possible that the retinopathy in our study was caused by the same mechanism as that of Purtscher’s retinopathy caused by chest compression during CPR.

It has recently been accepted that the pathogenesis of Purtscher’s retinopathy is due to microembolization followed by arteriolar precapillary occlusion [28]. After manual CPR, pulmonary bone marrow embolism and fat embolism were found in 26.7% and 20% of patients, respectively, and pulmonary bone marrow embolism was correlated with CPR time [30]. Because the fat embolus is relatively small and easily deformable, it can move to the retina through the foramen ovale (present in 20–25% of adults), lung capillaries, or arteriovenous shunts around the lungs [31]. Occlusion of the precapillary arteriole by an intermediate-sized embolus causes Purtscher’s flecken. However, in the case of a small fat embolus, obstruction of the retinal capillary can cause cotton wool spots or retinal hemorrhage [32]. For this reason, it seems that the appearance of Purtscher’s retinopathy was caused by fat emboli derived from the rip or sternum caused by chest compression in patients who received a long CPR time in our study. Agrawal and McKibbin [32] reported the disappearance of cotton wool spots, retinal hemorrhage, and Purtscher flecken at 6 months in Purtscher’s retinopathy, whereas in our study, retinal hemorrhage and cotton wool spots disappeared at 5.6 months, showing similar results.

In our study, 9 of 10 patients in the retinopathy group had diabetes mellitus, which was higher than the 27–31% reported in other studies (binomial distribution test, p = 0.002) [33, 34]. As fundus findings, including microaneurysm, hard exudate, retinal swelling, venous beading, and neovascularization in diabetic retinopathy, except dot-blot hemorrhage and cotton wool spot, were not observed and abnormal findings gradually disappeared during follow-up, it is assumed that the possibility of diabetic retinopathy is low. In addition, in 4 of the 10 patients with fundus findings within 12 months before CPR, diabetic retinopathy was not found before CPR.

This study has several limitations. This was a retrospective, small-case study. In terms of patient selection, there was a selection bias because most patients were referred for treatment because of complaints of ophthalmic discomfort. In addition, the underlying diseases of patients and those that cause CPR were diverse. In terms of patient examination, the initial condition of the patients was not stable, and there were limitations in performing fundus examination in all patients. To establish differential diagnosis from other retinal diseases, including diabetic, hypertensive retinopathy, anemic retinopathy, and others and to assess the retinal anatomic and vascular status, ophthalmic examinations, such as OCT, OCT angiography, and fundus angiography, were required. However, these had not been performed in several patients. Patients with systemic conditions that could commonly cause retinal hemorrhages were excluded, cases with retinopathy due to systemic complication were excluded by additional examination and medical record check, and laboratory tests were checked to identify potential confounding factors that could cause retinal hemorrhage. Additionally, in all included retinopathy patients, the findings of retinopathy had improved over time, suggesting that the effects of other retinal diseases were minimal.

In conclusion, as the success rate of CPR increases and its practice becomes more common [35, 36], it is becoming more likely that fundus examinations will be performed in patients who have received CPR. If a retinal hemorrhage is found in these patients, it is necessary to differentiate it from other retinal diseases. In cases of intraretinal hemorrhage with or without RNFL hemorrhage or cotton wool spots after CPR, retinopathy associated with CPR can be suspected if there are no characteristic findings of other retinal diseases. Retinopathy appears when CPR is prolonged and disappears over time.