Intracranial hemorrhage and Covid-19: A retrospective analysis of 1675 hospitalized Covid-19 Brazilian patients

Abstract

Objective

The present study seeks to overcome the lack of data on Covid-19 associated intracranial hemorrhage (ICH) in Brazil.

Methods

This is a retrospective, single-center case series of consecutive patients. It is a subanalysis of a larger study still in progress, which covers all neurological manifestations that occurred in patients admitted between March 1st, 2020 and June 1st, 2022, with active SARS-CoV-2 infection confirmed by polymerase chain reaction test. All patients with non-traumatic ICH were included.

Results

A total of 1675 patients were evaluated: 917 (54.75 %) had one or more neurological symptoms and 19 had non-traumatic ICH, comprising an incidence of 1.13 %. All patients had one or more risk factors for ICH. The presence of neurological manifestations before the ICH and ICU admission showed a statistically significant relationship with the occurrence of ICH (X² = 6.734, p = 0.0095; OR = 4.47; CI = 1.3–15.4; and FET = 9.13; p = <0.001; OR = 9.15; CI = 3.27–25.5 respectively).

Conclusion

Our findings were largely congruent with the world literature. We believe that the assessment of risk factors can accurately predict the subgroup of patients at increased risk of ICH, but further studies are needed to confirm these hypotheses.

1. Introduction

Although the global pandemic has reached its third anniversary, Covid-19 still poses a threat to healthcare systems and much remains to be clarified about its manifestations and complications. SARS-CoV-2 is primarily a respiratory virus, but it is capable of affecting several other organs and systems. The nervous system can be affected in a myriad of ways, with manifestations ranging from typical symptoms of upper airway infection such as headache and anosmia, to autoimmune diseases and neurovascular syndromes, which have gained prominence in research due to their uniquely high prevalence, morbidity and mortality.¹

The first report of intracranial hemorrhage (ICH) in Covid-19 patients occurred in the first months of the pandemic. Since then, the studies carried out have differed considerably regarding its incidence and mortality,2, 3, 4 which may be caused by demographic factors and conflicting medical interventions among healthcare services around the world. Developed countries have more organized and available healthcare services, which can modify the incidence, mortality and risk factors of ICH when compared to developing countries, where the lack of medications, ventilators and intensive care unit (ICU) beds were common.⁵ Furthermore, data on Covid-19 neurological manifestations lacks large studies involving the Brazilian population.

The present study seeks to overcome the lack of Brazilian data regarding the intracranial hemorrhagic manifestations of Covid-19, identifying its incidence, mortality and relationship with other neurological manifestations of the disease. Additionally, we sought to assess the presence of other possible causes and risk factors for ICH in these patients.

2. Materials and methods

This is a single-center, retrospective and observational case series of consecutive patients that took place at the Hospital de Clínicas of Passo Fundo, Brazil. It is a subanalysis of another study still in progress, which covers all neurological manifestations presented by patients admitted between March 1st, 2020 and June 1st, 2022, with active SARS-CoV-2 infection confirmed by polymerase chain reaction test, which was approved by the Ethics and Research Committee of Faculdade Meridional - IMED and is registered under CAAE 60913222.8.0000.5319 (final opinion 5,646,360). The Informed Consent Form was waived by the Research Ethics Committee, as this study did not interfere in patient management and performed an anonymous treatment and analysis of the data, without individual identification of the participants and, consequently, without addition of risks or harm to their well-being.

Medical records were evaluated in search of neurological manifestations during and after hospitalization. All patients with imaging-confirmed non-traumatic ICH were included in this case series. The management of Covid-19 was carried out according to the local protocol, and treatment of the neurological condition followed the standard indications.

Data were extracted according to a protocol created by the authors, and included age, sex, date of symptoms onset, hospitalization and discharge/death, previous comorbidities and risk factors for ICH, clinical information such as severity of the respiratory condition and use of anticoagulation, complications, procedures, ICU admission, laboratory data, neuroimaging, specific neurological treatment, outcome and follow-up data. Neuroimaging was reviewed by experienced Neurosurgeons. Descriptive variables were grouped. Continuous variables were summarized as mean, standard deviation (SD) and range. Categorical variables were summarized in totals and percentages. All comparisons were performed using the chi-square test (X²) or Fisher's exact test (FET), as applicable, and confidence intervals (CI) were adjusted to 95 %. All p values were considered significant at the p = 0.05 level and were not adjusted for multiple comparisons.

3. Results

A total of 1675 patients had their medical records analyzed. Of these, 917 (54.75 %) had one or more neurological manifestations and 19 had non-traumatic ICH, comprising an overall incidence of 1.13 %. Eighteen patients developed ICH during hospitalization and 1 patient developed it after 2 months of hospital discharge. Most patients were male (13/19), and the mean age was 53 years (SD: 14.2; range: 26–76), with 3 patients younger than 32 years.

Among patients with ICH, 4 suffered an ischemic stroke (IS) and 1 suffered a cerebral venous sinus thrombosis before the hemorrhage, and ICH occurred in a corresponding location in 4 of these, characterizing a hemorrhagic conversion of the ischemic events and not a spontaneous hemorrhage. However, one of the patients with IS hemorrhagic conversion also developed a subdural hemorrhage without history of trauma, and was included in the ICH without hemorrhagic conversion group for incidence estimates (Fig. 1 ). Subtracting the 3 cases of pure hemorrhagic conversion from the analysis we have a total of 16 ICH, an incidence of 0.95 % among all patients with Covid-19 admitted to our center (Table 1, Table 2 ).

Fig. 1.

All images are axial CT-scans that belong to patient 15. A and B) show a bilateral collection of subacute and acute blood in the posterior fossa subdural space; the red arrows signal very small areas of hemorrhagic conversion in the ischemic nervous tissue. C) shows intraventricular blood. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 1.

Individual characteristics of patients with Covid-19 associated ICH.

Patient	Age, sex	Comorbidities and risk fators for ICH	Covid-19 severity	ICU at ICH diagnosis	Neurological manifestations prior to ICH	ICH subtype	Clinical complications	ICH consequences	ICH Management	Observations
1	62, M	Therapeutic anticoagulation (UFH), obesity	Severe	Yes	No	SAH	Bacterial pneumonia, dialytic AKI	Diffuse cerebral edema	Conservative	Started therapeutic UFH on the same day as ICH. ICH ictus 1 h after dialysis session.
2	32, M	Prophylactic anticoagulation (UFH)	Severe	Yes	Yes	IPH, IVH, SAH	Bacterial pneumonia, intestinal tract infection, dialytic AKI, hepatic disfunction, hematuria	Subfalcine and uncal herniation, perilesional edema	Conservative	ICH ictus during dialysis session.
3	68, M	Therapeutic anticoagulation (UFH), HTN, dyslipidemia	Severe	Yes	Yes	IPH, IVH, SAH	Bacterial pneumonia, dialytic AKI, hematuria, septic shock	Diffuse cerebral edema	Conservative	ICH secondary to CVST. Started therapeutic UFH a few hours before the ICH. Hematuria onset concomitant with ICH.
4	46, M	Therapeutic anticoagulation (UFH), HTN, DM	Severe	Yes	Yes	IPH	CVST, DVT, PE, hematuria, bacterial pneumonia, non-dialytic AKI	Subfalcine and uncal herniation, hemispheric edema	Decompressive craniectomy	None.
5	61, F	Therapeutic anticoagulation (Fondaparinux), HTN, DM, obesity, previous VTE, HF	Severe	Yes	Yes	IPH, IVH, SAH	Dialytic AKI, UTI	Diffuse cerebral edema, Subfalcine herniation	Conservative	ICH diagnosed in a patient with slow awakening.
6	47, M	Prophylactic anticoagulation (Fondaparinux), HTN, obesity, previous IS	Severe	Yes	Yes	IPH	Bleeding in different sites, dialytic AKI, bacterial pneumonia, pancreatitis, septic shock	Diffuse posterior fossa edema, obstructive hydrocephalus, brainstem compression	EVD	ICH ictus 5 h after dialysis session. Hemorrhagic diathesis concomitant with ICH.
7	62, M	Prophylactic anticoagulation (enoxaparin)	Non-Severe	No	Yes	IPH, IVH, SAH	Enterorrhagia, bacterial pneumonia, meningitis	None	Conservative	Patient with previous hydrocephalus and EVD placement.
8	76, F	Prophylactic anticoagulation (Fondaparinux), HTN, asthma	Severe	Yes	Yes	IPH	Spontaneous breast ecchymosis	None	Conservative	None.
9	63, F	Prophylactic anticoagulation (UFH), HTN, DM	Non-Severe	No	No	IPH, SAH	None	None	Conservative	Covid-19 patient referred to our center due to ICH.
10	47, M	Prophylactic anticoagulation (Fondaparinux), obesity	Severe	Yes	No	IPH, SAH	Dialytic AKI	Diffuse cerebral edema, subfalcine herniation	Conservative	ICH ictus 5 h after dialysis session.
11	63, M	Therapeutic anticoagulation (UFH), asthma	Severe	Yes	Yes	IPH	Dialytic AKI, 2 PE, bacterial pneumonia, UTI, hematuria, pneumothorax CRA with ROSC	None	Conservative	Started therapeutic UFH 1 h before the ICH.
12	61, M	Therapeutic anticoagulation (Enoxaparin), HTN, DM, obesity, atrial fibrilation	Severe	Yes	Yes	IPH	Central venous line infection, bacterial pneumonia, UTI, dialytic AKI, acute pulmonary edema, pressure ulcer, septic shock	None	Conservative	ICH ictus 4 h after dialysis session.
13	41, M	Prophylactic anticoagulation (Enoxaparin), obesity	Severe	Yes	Yes	IPH, IVH, SAH	Dialytic AKI, central venous line infection, septic shock	Diffuse cerebral edema	Conservative	None.
14	30, F	Prophylactic anticoagulation (UFH), obesity	Severe	Yes	Yes	IPH	Dialytic AKI, oral bleeding, bacterial pneumonia, pneumothorax, 2 CRA with ROSC	Diffuse cerebral edema	Conservative	ICH ictus 2 h after dialysis session. Oral bleeding concomitant with ICH.
15	49, M	Prophylactic anticoagulation (Warfarin), prior anticoagulation (Warfarin), previous IS, Eisenmenger syndrome	Severe	No	Yes	Subacute SDH, IPH, IVH	Septic shock, IS	Posterior fossa edema, brainstem compression	EVD	Only patient to present ICH after hospital discharge. Hemorrhagic conversion of IS + subacute HSD due to probable coumarin intoxication.
16	26, M	Prophylactic anticoagulation (UFH), prior anticoagulation (Warfarin)	Severe	No	Yes	SAH	IS, pancreatitis, CRA with ROSC	Diffuse cerebral edema	Conservative	ICH ictus 3 h after dialysis session.
17	52, F	Therapeutic anticoagulation (UFH), obesity	Severe	Yes	Yes	IPH, SAH	Dialytic AKI, DVT, PE, nasal bleeding	Subfalcine and uncal herniation, perilesional edema	Conservative	None.
18	53, M	Prophylactic anticoagulation (Warfarin), prior anticoagulation (Warfarin), HTN, DM, COPD, dyslipidemia, previous AMI, CCS, HF	Non-Severe	No	Yes	IPH, SAH	Bacterial pneumonia, IS	Perilesional edema (IS)	Conservative	Hemorrhagic conversion of IS.
19	74, F	Therapeutic anticoagulation (Fondaparinux), HTN, DM	Severe	Yes	Yes	IPH	Dialytic AKI, IS	Perilesional edema (IS)	Conservative	Hemorrhagic conversion of IS.
Abbreviations: AKI = Acute kidney injury, AMI = Acute myocardial infarction, CCS = Chronic coronary syndrome, COPD = Chronic obstructive pulmonar disease, CRA = Cardiorespiratory arrest, CVST = Cerebral venous sinus thrombosis, DM = Diabetes mellitus, DVT = Deep vein thrombosis, EVD = Extraventricular drain, F = Female, HF = Heart failure, HTN = Hypertension, ICH = Intracranial hemorrhage, IS = Ischemic stroke, IPH = Intraparemchymal hemorrhage, IVH = Intraventricular Hemorrhage, M = Male, PE = Pulmonary embolism, ROSC = Return of spontaneous circulation, SAH = Subarachnoid hemorrhage, SDH = Subdural hemorrhage, UTI = Urinary tract infection, VTE = Venous thromboembolism

Table 2.

Grouped characteristics of patients with Covid-19 associated ICH.

CHARACTERISTICS	TOTAL (%)
Age (in years)	Mean: 53,3 (SD: 14,2; range 26–76)
-Over 53	9 (47,4%)
-Under 53	10 (52,6%)
-Under 32	3 (15,8%)
Sex
-Male	13 (68,5%)
-Female	6 (31,5%)
Comorbidities and risk factorsfor ICH	17 (89,5%)
HTN	9 (47,4%)
DM	6 (31,5%)
Obesity	8 (42,1%)
Asthma	2 (10,5%)
COPD	1 (5,25 %)
Dyslipidemia	2 (10,5%)
Atrial fibrilation	1 (5,25 %)
Past major cardiovascular event	3 (15,8%)
Past VTE	1 (5,25 %)
HF	3 (15,8%)
CCS	1 (5,25 %)
Prior anticoagulation	3 (15,8%)
Covid-19 Hospitalization
Covid-19 severity
-Severe	16 (84,2%)
-Non-severe	3 (15,8%)
Anticoagulation during hospitalization	19 (100 %)
-Prophylactic	11 (57,9%)
-Therapeutic	8 (42,1%)
-Medication used
-Unfractionated heparin	9 (47,4%)
-Low molecular weight heparin (Enoxaparin ou Fondaparinux)	8 (42,1%)
-Warfarin	2 (10,5%)
Clinical complications	19 (100 %)
-IS	4 (21 %)
-CVST	1 (5,25 %)
-VTE	3 (15,8%)
-Bleeding at different sites	9 (47,4%)
-Septic shock	5 (26,3%)
-Overlaping bacterial infections	12 (63 %)
-AKI	13 (68,5%)
-Dialytic AKI	12 (63 %)
-CRA with ROSC	3 (15,8%)
-Pneumothorax	3 (15,8%)
-Pancreatitis	2 (10,5%)
-Critical illness polyneuropathy	2 (10,5%)
ICU at the ICH occurence	14 (73,7%)
Other neurological manifestations prior to ICH	16 (84,2%)
ICH characteristics
Days between the onset of Covid-19 symptoms and ICH	Mean: 22 (SD: 10,2; range 2–50)
ICH subtype
-IPH	17 (89,5%)*
-SAH	11 (57,9%)¬
-IVH	6 (31,5%)
-SDH	1 (5,25 %)
Hemorrhagic conversion of IS	3 (15,8%)§
Hemorrhage secondary to CVST	1 (5,25 %)
Diagnostic suspicion
-Pupillary reflex examination	14 (73,7%)
-Other	5 (26,3%)
ICH complications
-Severe cerebral edema	11 (57,9%)
-Cerebral herniation	7 (36,8%)
-Hydrocephalus	1 (5,25 %)
Management
-Conservative	16 (84,2%)
-EVD	2 (10,5%)
-Decompressive craniectomy	1 (5,25 %)
Outcome
Death	19 (100 %)
Abbreviations: AKI = Acute kidney injury, CCS = Chronic coronary syndrome, COPD = Chronic obstructive pulmonar disease, CRA = Cardiorespiratory arrest, CVST = Cerebral venous sinus thrombosis, DM = Diabetes mellitus, EVD = Extraventricular drain, HF = Heart failure, HTN = Hypertension, ICH = Intracranial hemorrhage, IS = Ischemic stroke, IPH = Intraparemchymal hemorrhage, IVH = Intraventricular Hemorrhage, ROSC = Return of spontaneous circulation, SAH = Subarachnoid hemorrhage, SD = Stadard deviation, SDH = Subdural hemorrhage, UTI = Urinary tract infection, VTE = Venous thromboembolism * Seven of the 17 IPH were isolated, and 10 were associated with other ICH subtypes. ¬ Two of the 11 SAH were isolated and 9 were associated with other ICH subtypes. § Among these patients, patient 15 also had a subacute SDH, which was not a consequence of the IS hemorrhagic conversion.

All patients had one or more risk factors for intracranial hemorrhage. The most frequent were anticoagulation (100 %), dialysis (12/19–63 %) and arterial hypertension (9/19–47.4 %). Regarding the respiratory condition, 16 (84.2 %) patients had a severe Covid-19 (defined as the need for MV) and 14 (73.7 %) were admitted to the ICU before the hemorrhage. During admission, only 4 patients had elevated in prothrombin or activated plasma thromboplastin times. At the time of ICH, important alterations in coagulation studies were present in 7 of the 15 patients with such data available. Other common findings at the time of ICH were blood gas analysis alterations (100 %), especially hypercapnia (10/19 – 52.6 %), normocytic normochromic anemia (13/19–68.4 %) and worsening kidney function (acute kidney injury in 13 patients). Thrombocytopenia occurred in 3 patients only.

Intracranial hemorrhage was more common in patients who had other neurological deficits before the ictus compared to those who did not (1.74 % vs 0.39 %) and a significant relationship was seen between the phenomena (X² = 6.734,p = 0.0095; OR = 4.47; CI = 1.3–15.4). In addition, the risk of ICH in this subgroup increases with ICU admission (3.4 % vs 0.7 %; X² = 9.16,p = 0.002; OR = 4.9; CI = 1.57–15 0.4), but is not greater than the overall risk of ICU admission alone (3.48 % vs 0.39 %; FET = 9.13; p=<0.001; OR = 9.15; CI = 3.27–25.5). This is probably due to the fact that the vast majority of patients admitted to the ICU had some type of neurological manifestation (87.8 %). Patient sex was non-significant (incidence of ICH in men 0.77 % vs 0.35 %; X² = 0.56,p = 0.45; OR = 1.45; CI = 0.54–3.83) (Table 3 ).

Table 3.

ICH incidence and relationship with other findings.

	Total (%)	With ICH(%)	p-value
Patients hospitalized with Covid-19	1.675 (100 %)	19 (1,13 %)	–
- Male	1.005 (60 %)	13 (1,29 %)	0,45
- Female	670 (40 %)	6 (0,89 %)
- With other neurological manifestations prior to ICH	917 (54,74 %)	16 (1,74 %)	0,0095
- Without other neurological manifestations prior to ICH	758 (45,26 %)	3 (0,39 %)
- ICU admission prior to ICH	402 (24 %)	14 (3,48 %)	<0,001
- Without ICU admission prior to ICH	1.273 (76 %)	5 (0,39 %)
- With neurological manifestations + ICU admission prior to ICH	353 (21,07 %)	12 (3,4%)	0,002
- Without neurological manifestations + ICU admission prior to ICH	564 (33,73 %)	4 (0,7%)
Abbreviations: ICH = Intracranial hemorrhage, ICU = Intensive Care Unit

It is noteworthy that all patients had one or more clinical complication during hospitalization. Acute kidney injury was the most common, occurring in 13 (68.4 %) (dialysis was performed in 12), and superimposed bacterial infections, usually multiple, occurred in 12 (63 %) cases. Meningitis occurred only 1 time, in a patient with an extraventricular drain (EVD). Another frequent complication was spontaneous bleeding in other sites. These occurred in 9 patients, all anticoagulated and 6 undergoing dialysis. In 3 patients, bleeding started shortly before the diagnosis of ICH. Interestingly, half of pancreatitis cases in hospitalized Covid-19 patients occurred in patients who subsequently developed ICH (2/4–50 %). Individual and grouped demographic and clinical information are summarized in Table 1, Table 2.

3.1. ICH characteristics and outcome

The diagnosis of ICH was performed with head CT scan in all patients. On average, bleeding occurred about 22 days after the onset of Covid-19 symptoms (SD: 10.2; range: 2–50). Patient 15 was the only one to have an ICH after hospital discharge; he arrived at our service with a posterior fossa IS with hemorrhagic transformation, in addition to a subacute subdural hematoma (SDH) with no history of recent trauma (Fig. 1).

In total, 17 intraparenchymal (IPH), 11 subarachnoid (SAH), 6 intraventricular hemorrhages (IVH) and only 1 subacute SDH were detected. Ten patients had more than one type of hemorrhage. Isolated IPH occurred 7 times and isolated SAH was seen only 2 times (Fig. 2 ). Among the 17 IPH, 9 presented multiple foci of bleeding, and 9 had a volume >30 ml. ICH caused brain herniation in 7 patients (36.8 %), severe edema in 11 (57.9 %) and hydrocephalus in 1 (5.25 %) (Fig. 3, Fig. 4 ). As most patients were sedated at the time of diagnosis, the suspicion almost always arose through pupillary reflex examination.

Fig. 2.

All images are axial CT-scans belonging to patient 1. The red arrows signal bilateral SAH foci in the cerebral sulci (A and C) and in the Sylvian cistern (B). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3.

All images are axial CT-scans. A and B belong to patient 2. A) shows a large IPH with intraventricular rupture causing a significant midline shift; the red arrow points to an SAH foci. B) shows the cranial extension of the hematoma; the red arrows also signal cortical SAH foci. Fig. 3 C through F belong to patient 3. C) demonstrates a significant hematoma in the pons; the red arrow signals a small IPH foci. D) goes on to show a more cranial image, with diffuse hemorrhagic foci. Most are small, with the exception of the one located in the right gangliocapsular region. E) demonstrates more IPH foci, and also shows the presence of blood in the right ventricle. F) shows more IPH foci; the red arrows signal small IPH foci. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 4.

All images are axial CT-scans. A) belongs to patient 8. We visualize a medium-sized, unifocal IPH in the right frontal lobe. B) belongs to patient 17. A large intraparenchymal hemorrhage can be seen in the left hemisphere, causing a significant midline shift; the red arrows signal cortical SAH foci. C and D) belong to patient 10. They show a bifocal IPH, located in the left temporo-occipital region and in the left frontal lobe, respectively; the red arrows signal cortical SAH foci. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Most patients were managed conservatively, with only 1 undergoing decompressive craniectomy and 2 undergoing EVD placement due to hydrocephalus. Interestingly, patient 7 had non-communicating hydrocephalus on hospital admission, and underwent EVD placement prior to intracranial bleeding.

All patients died due to the neurological and/or respiratory condition.

4. Discussion

4.1. Covid-19 and ICH: Causal relationship or coincidence?

The big question that practitioners seek to answer is: Does Covid-19 have a causal relationship or is it coincidentally associated with ICH?

The incidence of ICH in the general population is 24.6/100,000 person-years, a figure about 45 times lower than the 1.13 % found in our series and lower than all estimates found in the literature, that range from 0.1 to 4.4 %.3, 4 ^{, 6} Furthermore, the ICH rate in Covid-19 hospitalized patients is also higher than in patients hospitalized with respiratory infections caused by other viruses such as influenza (0.7 % vs 0.2 %; adjusted risk ratio = 2.85; CI = 1.35–6.03).⁷ However, no hemorrhages were observed in patients without classic risk factors in our study, which is once again consistent with the world literature.6, 8 From very early on, this finding led researchers to question the role of Covid-19 in ICH pathophysiology.

The main confounding factor in most studies is parenteral anticoagulation, as it is used in the majority of hospitalized Covid-19 patients due to high rates of venous thromboembolism (VTE). The universal use of these drugs hinders the assessment of the role of Covid-19 in ICH pathogenesis. However, when we compare hospitalized patients under prophylactic anticoagulation with and without Covid-19, we notice that the incidence of ICH is higher in the former.

The IMPROVE study evaluated 10,866 patients hospitalized for medical reasons, 48.1 % of whom were receiving adequate prophylactic anticoagulation, with either unfractionated heparin (UFH) or light molecular weight heparin (LMWH), and only 10 ICH were registered.⁹ The PRIME study, which evaluated the safety and efficacy of enoxaparin (n = 477) compared to UFH (n = 482) for the prevention of VTE in medical patients, did not record ICH in any of the 959 patients included.¹⁰ Moreover, in the PRINCE study, which also compared the safety and efficacy of enoxaparin (n = 332) versus UFH (n = 333) for thromboprophylaxis, none of the 665 patients experienced intracranial bleeding.¹¹ Fondaparinux had its safety and efficacy for thromboprophylaxis evaluated in a randomized, double-blind, placebo-controlled trial that included 849 elderly patients. In that study, 425 patients were receiving fondaparinux daily, and no ICH was observed in either study group.¹² VTE prophylaxis data in medical ICU patients are scarce. However, in a study performed with 156 patients in a surgical ICU, comparing UFH (n = 75) to LMWH (n = 81), no ICH was observed.¹³

None of the studies on thromboprophylaxis demonstrated an incidence of ICH remotely close to the 1.13 % found in patients hospitalized for Covid-19. It is worth noting that even if we consider only the Covid-19 patients that developed ICH while under prophylactic anticoagulation, the 11 cases recorded generate an incidence higher than the observed in the thromboprophylaxis studies. Therefore, there is no doubt that Covid-19 increases the number of intracranial bleeds. It is also noteworthy that the incidence of ICH may be underestimated in several Covid-19 studies. Underdiagnosis of neurovascular events can be frequent in sedated patients, where neurological examination is impaired and the evaluation of the pupillary reflex is one of the few doors available for nervous system assessment. It is expected that difficulty in diagnosing ICH will be greater in underdeveloped countries and smaller centers, where imaging studies are not widely available. As an example, in our study we identified 4 patients with a sudden drop in the level of consciousness, non-photoreactive anisocoria and/or signs of lateralization who died before undergoing a head CT scan, either due to immediate unavailability of the device or hemodynamic instability and subsequent death.

The pathophysiology of hemorrhagic disease in Covid-19 has been unraveled over time, and today we have a decent, albeit incomplete, understanding of the subject. The literature points out that risk factors for ICH in patients with Covid-19 are similar to those of the general population.⁵ In addition, ICU admission, MV, extracorporeal membrane oxygenation and anticoagulation, treatments that have been shown to decrease Covid-19 mortality, increase the risk of intracranial bleeding2, 3, 4 6, 14 It is also known that SARS-CoV-2 invades the nervous system through angiotensinogen-converting enzyme 2 receptors, present in neural and vascular cells. Microscopy studies have demonstrated viral inclusions in vascular endothelial cells, which can cause mononuclear infiltration and consequent endothelitis/vasculitis, leading to deregulation of the blood–brain barrier and weakening the vascular wall. Hypercytokinemia is also believed to lead to unbridled activation of extracellular matrix metalloproteases, which degrade connective tissue and weaken the vascular wall. Furthermore, some reports indicate the presence of circulating antibodies against Platelet Factor 4 in some Covid-positive individuals, which leads to a systemic response similar to heparin-induced thrombocytopenia and to bleeding events. Finally, the hypercoagulable state caused by Covid-19 can lead to thrombotic microangiopathy of the vasa-vasorum, causing vessel ischemia and consequent weakness and rupture. It should also be pointed out that hypercapnia, which develops in patients with severe lung disease, increases cerebral blood flow and favors hemorrhage.4, 15, 16, 17

All these changes act by facilitating bleeding, but the available evidence indicates that these effects alone are not capable of generating major bleeding events, and that ICH depends on a synergistic action between the vessel frailty caused by Covid-19 and known risk factors, especially anticoagulation. Corroborating the hypothesis that the direct effect caused by Covid-19 is decisive for the occurrence of ICH, in our study, bleeding was more frequent in patients who had other neurological deficits before ICH compared to those who did not (1.74 % vs 0.39 %; X² = 6.734,p = 0.0095; OR = 4.47; CI = 1.3–15.4), a fact that implies the presence and action of the virus in the central nervous system prior to hemorrhage.

Most cases of ICH in Covid-19 patients were IPH both in our series and in the literature,3, 4 ^, 6, 8 This may be explained by the fact that IPH is the most common type of ICH overall, but it may also indicate a SARS-CoV-2 tropism for small, intraparenchymal vessels in the brain.¹⁸

4.2. Additional findings and mortality

Mortality in ICH is higher when it is associated with Covid-19, ranging from 42 to 100 % in the literature.3, 4, 5, 6 ^{, 16} We believe that the 100 % mortality rate in our series is a consequence of both the severity of the clinical condition of most of our patients as well as the delay in ICH diagnosis, which hindered proper and timely medical and surgical management.

Meticulous analysis of our cases led to some interesting findings. Eight patients were diagnosed with hemorrhage within 5 h or less of a dialysis session, and 3 patients were diagnosed with ICH a few hours after starting therapeutic anticoagulation with UFH. Additionally, 3 patients had bleeding in other sites just before the diagnosis of ICH. These findings may indicate that the transition between anticoagulation regimens (dose increase), the post-dialysis period and situations of hemorrhagic diathesis can be characterized as high-risk moments for ICH, and that more frequent and rigorous neurological evaluation is warranted in these cases in order to reduce the delay in diagnosis and mortality.

It is important to emphasize that our study has certain limitations, in particular the impossibility of ruling out the effect of confounding factors in our statistical analyzes and the small number of cases with ICH, as well as its retrospective nature. Besides, the observed 100 % mortality rate prevented us from identifying prognostic factors in these patients.

5. Conclusion

This is the first large study on Covid-19 associated ICH in Brazil and findings were largely congruent with the world literature. The presence of neurological manifestations preceding hemorrhage proved to be a significant risk factor in our patients, as well as ICU admission. Furthermore, we speculate that there are some moments of greater risk for ICH occurrence, especially the first hours after dialysis sessions, anticoagulant dose increases and hemorrhagic diathesis. We believe that the joint assessment of risk factors and the moments of greatest risk for ICH can predict the occurrence of intracranial bleeding with decent accuracy, and that identifying high-risk patients will lead to a faster diagnosis and improvements in outcome. Nonetheless, further studies will be needed to confirm these hypotheses.

Ethics approval and consent to participate

The study was approved by the Ethics and Research Committee of Faculdade Meridional – IMED and is registered under CAAE: 60913222.8.0000.5319 (final opinion 5,646,360). The Informed Consent Form was waived by the Research Ethics Committee, as it was a non-interventionist retrospective study, which used only information from medical records, organizational information systems and imaging tests already performed, without changing or interfering in the management of the service research participants and, consequently, without adding risks or harm to their well-being, with anonymous treatment and analysis of the data, without individual identification of the participants.

Funding

The authors did not receive support from any organization for the submitted work.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Credit authorship contribution statement

Artur Eduardo Martio: Conceptualization, Methodology, Writing – review & editing. Pedro de Moraes Rêgo Soares: Conceptualization, Methodology. Octávio Ruschel Karam: Writing – review & editing. Wagner Lazaretto Padua: Methodology, Writing – review & editing. Luciano Bambini Manzato: Conceptualization, Writing – review & editing. Paulo Moacir Mesquita Filho: Methodology, Writing – review & editing.