Abnormal Neurologic Findings in Patients With Sickle Cell Disease Without a History of Major Neurologic Events

Abstract

Background and Objectives

Patients with sickle cell disease (SCD) are prone to symptomatic neurologic complications. Previous studies reported accrual of neural injury starting at early age, even without having symptomatic neurologic events. The aim of this study was to assess the prevalence and risk factors of abnormal neurologic findings in patients with SCD with no history of major symptomatic neurologic events.

Methods

Our study extracted patients diagnosed with SCD from the Cooperative Study of Sickle Cell Disease. Patients who underwent a neurologic evaluation were included in our analysis. Patients with previous documented major symptomatic neurologic events were excluded. We compared patients with SCD with abnormal neurologic findings with those without in terms of clinical and laboratory parameters using multivariate binary logistic regression.

Results

A total of 3,573 patients with SCD were included (median age = 11 [IQR = 19] years, male = 1719 [48.1%]). 519 (14.5%) patients had at least one abnormal neurologic finding. The most common findings in descending order were abnormal reflexes, gait abnormalities, cerebellar dysfunction, language deficits, nystagmus, abnormal muscle tone and strength, Romberg sign, Horner syndrome, and intellectual impairment. History of eye disease (odds ratio [OR] = 2.76, 95% confidence interval [CI] = 1.63–4.68) and history of osteomyelitis (OR = 2.55, 95% CI 1.34–4.84) were the strongest predictors of abnormal neurologic findings, followed by smoking (OR = 1.59, 95% CI 1.08–2.33), aseptic necrosis (OR = 1.57, 95% CI 1.06–2.33), hand-foot syndrome (OR = 1.48, 95% CI 1.04–2.12), and male sex (OR = 1.42, 95% CI 1.01–2.02).

Discussion

Neurologic deficits are relatively common in patients with SCD, even without documented major neurologic insults. They range from peripheral and ophthalmic deficits to central and cognitive disabilities. Patients with SCD should have early regular neurologic evaluations and risk factor modification, particularly actively promoting smoking cessation.

TAKE-HOME POINTS

• → Neurologic deficits are common in SCD, even without major neurologic insults.

• → Eye disease and osteomyelitis strongly predict abnormal neurologic findings.

• → Smoking cessation promotion is important in preventing neurologic deficits.

Introduction

Sickle cell disease (SCD) is the most common form of hereditary hemolytic anemia and is most prevalent in the Middle East, sub-Saharan Africa, and India.^1,2 Patients with SCD are prone to a wide range of early neurologic complications such as ischemic strokes, transient ischemic attack, seizures, posterior reversible encephalopathy syndrome, and cerebral venous sinus thrombosis.^3,4 Even in the absence of such events, silent cerebral infarcts (SCI) are a common cause of neurologic decline and found in nearly 22% of patients with SCD.⁵ Although usually clinically covert, SCI are associated with a high risk of recurrence of future vascular events.⁶ SCI are also associated with poor school and cognitive functioning, subtle visual-motor speed deficits, and higher neurologic soft sign (NSS; sensory integration and motor coordination impairment) scores.^7-9 Moreover, evidence suggests that children with SCD with normal MRIs may still experience neurocognitive impairments.¹⁰ Most of the previous studies have assessed the effect of seizures, meningitis, and strokes on the neurologic function in SCD; however, reports are scarce regarding effects of SCD on the nervous system in patients without such major events.^11-15

Given its high prevalence, transcranial Doppler ultrasound is used in screening protocols to assess the future risk of stroke to guide decisions to pursue chronic transfusion therapy.¹⁶ Unfortunately, this is not always feasible in middle-income and low-income countries, where SCD is the most prevalent.^17,18 Even in the lack of accessible chronic transfusion therapy, other interventions can be implemented especially pertaining to stroke prevention, school accommodations, and surveillance for developmental delay and cognitive impairments.¹⁹ As such, it would be extremely useful to use basic clinical parameters to predict future neurologic decline.

Other than NSS scores, clinical parameters such as lower hemoglobin concentration, higher systolic blood pressure, and male sex have shown to be associated with SCI in patients with SCD.^5,20 Studies exploring the importance of physical examination, a very basic element of patient assessment, in predicting neurologic decline, are very limited. Studies exploring that do not specify whether patients have had a history of neurologic event or whether their neurologic involvement was silent. In addition, they showed that NSS may reliably identify the presence of minor MRI lesions and identify the population at risk of developing strokes.^7,21

Therefore, this study aimed to explore the prevalence of abnormal neurologic findings in patients with SCD with no prior documented history of major neurologic events. We further attempt to identify clinical parameters that predict such abnormal findings. This has the potential to assist in identifying those at risk, especially in low-income countries, where more expensive screening methods are not widely available.

Methods

Study Design and Population Sample

This retrospective sample of patients with SCD was derived from the multicenter Cooperative Study of Sickle Cell Disease (CSSCD), which was conducted from 1977 to 1995. The methods for CSSCD have been previously reported.^22,23 Access to data was obtained from the Biologic Specimen and Data Repository Information Coordinating Center.²⁴ Patients with SCD were enrolled in 3 phases, and this study was limited to those enrolled in phase 1 (n = 4,085), which spanned from 1978 to 1988 (recruitment ended in 1981, except for infants younger than 6 months, who were still enrolled until 1988). Of those enrolled during phase 1, only those who underwent a neurologic evaluation and had no prior major neurologic events (cerebrovascular accidents, seizures, and meningitis) were included in the analysis (n = 3,573). We compared clinical and laboratory parameters in patients with abnormal neurologic findings with those of patients without such findings. At study entry, medical history, physical examination, and laboratory tests were obtained. Patients were then annually (4 annual visits) assessed for the development of acute events, and laboratory tests were collected (e.g., hemoglobin, red blood cells, white blood cells, platelets, reticulocytes, hemoglobin F, albumin, and lactate dehydrogenase). The mean value of these laboratory variables was calculated from the initial visit to the fourth annual visit when available. The hematology staff members performed the neurologic physical examination. Those who were discovered to have significant abnormalities were assessed by a neurologist for confirmation of findings. All patients underwent a neurologic evaluation on entry or at their first annual visit. The possible predictors of abnormal neurologic findings assessed were age, sex, race, hemoglobin genotype, the presence of non-neurologic complications of SCD (e.g., acute chest syndrome, pneumonia, eye disease, and average number of pain crises), smoking history, family history of seizures or stroke, related laboratory tests, and previous transfusions or operations, such as cholecystectomy, splenectomy, hip replacement, and laser eye therapy. Previous eye disease includes glaucoma, cataract, uveitis, and retinopathy.²⁵

Outcomes

The primary outcome was to determine the prevalence of abnormal neurologic findings in patients with SCD who have no known previous neurologic events. Secondary outcomes included assessing potential predictors of such abnormal neurologic findings and evaluating the implications of the results on clinical practice.

Standard Protocol Approvals, Registrations, and Patient Consents

Approval to access the CSSCD study data was obtained from the BIOLINCC (Biologic Specimen and Data Repository Information Coordinating Center) website (biolincc.nhlbi.nih.gov/home/), which is an open access data repository. Ethical approval was obtained from the Research Ethics Committee, School of Medicine, Jordan University of Science and Technology.

Statistical Analysis

Statistical analyses were performed using SPSS (version 25, IBM corporation, Armonk, NY). To assess the difference in clinical parameters between patients with normal and abnormal neurologic examination, the Mann-Whitney U test was used for continuous variables and the chi-square test for categorical variables. Data for continuous variables were recorded as median (interquartile range [IQR]). Differences in clinical parameters that were statistically significant on univariate analysis were further assessed using a multivariate binary logistic regression model. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated for each variable. Significance was set at p value <0.05.

Data Availability

Datasets used and/or analyzed in this study can be accessed from the BIOLINCC (Biologic Specimen and Data Repository Information Coordinating Center) website: biolincc.nhlbi.nih.gov/studies/csscd/.

Results

A total of 3,573 patients with SCD met the inclusion criteria and were included in our study (median age = 11 [IQR = 19] years, male = 1719 [48.1%]). Of them, 519 (14.5%) patients had at least 1 abnormal neurologic finding on examination (382 [10.7%] with only 1 abnormal finding and 137 [3.8%] with at least 2). The total number of patients assessed for each element of the full neurologic examination is listed in Table 1. The most commonly observed abnormal neurologic finding was an abnormal deep tendon reflex (n = 238/2,829 [8.4%]; knee (n = 187/2,822 [6.6%]), ankle (n = 142/2,821 [5.0%]), biceps (n = 119/2,822 [4.2%]), triceps (n = 115/2,817 [4.1%]), ulnar (n = 110/2,809 [3.9%]), radial (n = 110/2,816 [3.9%]), and plantar (n = 20/2,818 [0.7%]). Abnormal gait was observed in 91/2,810 (3.2%) patients (3 (0.1%) were unable to walk, 11 (0.4%) had broad-based ataxia, and 77 (2.7%) had a limp) while abnormal cerebellar function was found in 117/2,734 patients (4.3%). Other less frequent abnormal findings included nystagmus 62/2,812 (2.2%), abnormal muscle tone and strength 59/2,831 (2.1%), Romberg sign 27/2,776 (1.0%), Horner syndrome 29/2,813 (1.0%), language deficits 23/938 (2.5%), and intellectual impairment 10/2,037 (0.5%). The number of patients who had at least 1 abnormal deep tendon reflex were 238 of 2,829 (8.4%). Among them, abnormal reflexes were predominantly bilateral (ranging from 81.2% to 89.5%), except for the plantar reflex, which showed a bilateral occurrence in 60% of cases. The study revealed a predominance of absent-hypoactive reflexes. For upper extremity reflexes, including biceps, triceps, ulnar, and radial reflexes, the percentages of absent-hypoactive responses ranged between 64.7% and 66.9%. Similarly, lower extremity reflexes, such as knee, ankle, and plantar reflexes, displayed substantial absent-hypoactive patterns, ranging from 70.5% to 85.0% (Table 2). The ratio of bilateral to unilateral involvement for absent-hypoactive reflexes was notably higher than for hyperactive reflexes for all the included sites (except for plantar reflexes): biceps (24.6 vs 3.1), triceps (24.6 vs 3.2), ulnar (17.2 vs 2.7), radial (23.0 vs 3.2), knee (5.0 vs 3.1), ankle (6.0 vs 3.5), and plantar (1.4 vs 2.0) reflexes. This indicates that hyporeflexia is associated more with bilateral involvement and hyperreflexia is associated with both bilateral and unilateral involvement.

Table 1.

Frequency of Abnormal Neurologic Findings Among Patients With SCD Without Previous Neurologic Insults

Abnormal neurologic finding	N (%)	Total examined
Number of abnormal neurologic findings		3,573
None	3054 (85.5)
One	382 (10.7)
Two or more	137 (3.8)
Extraocular movement	14 (0.5)	2,764
Facial power, sensation, and reflexes	19 (0.7)	2,787
Palatal reflex	6 (0.2)	2,775
Tongue movements	5 (0.2)	2,774
Muscle tone and strength	59 (2.1)	2,831
Muscle tone at the right side	28 (1.0)	2,828
Muscle tone at the left side	29 (1.0)	2,826
Muscle power at the right side	27 (1.0)	2,828
Muscle power at the left side	33 (1.2)	2,828
Body reflexes	238 (8.4)	2,829
Biceps	119 (4.2)	2,822
Triceps	115 (4.1)	2,817
Ulnar	110 (3.9)	2,809
Radial	110 (3.9)	2,816
Knee	187 (6.6)	2,822
Ankle	142 (5.0)	2,821
Plantar	20 (0.7)	2,818
Involuntary movement	9 (0.3)	2,828
Nystagmus	62 (2.2)	2,812
Romberg sign	27 (1.0)	2,776
Cerebellar function	117 (4.3)	2,734
Horner syndrome	29 (1.0)	2,813
Body sensation	18 (0.6)	2,775
Tactile	10 (0.4)	2,769
Position	7 (0.3)	2,731
Vibration	10 (0.4)	2,710
Pain present	11 (0.4)	2,778
Language deficits	23 (2.5)	938
Intellectual function	10 (0.5)	2,037
Gait	91 (3.2)	2,810
Unable to walk	3 (0.1)
Broad-based ataxia	11 (0.4)
Limp	77 (2.7)

Table 2.

Deep Tendon Reflexes Laterality Pattern and Grading Among Patients With Abnormal Reflexes

Reflex	Laterality		Absent-hypoactivea	Hyperactiveb	Mixedc	Total
Biceps			77 (64.7%)	41 (34.4%)	1 (0.9%)	119
	Unilateral	13 (10.9%)	3 (2.5%)	10 (8.4%)	—
	Bilateral	106 (89.0%)	74 (62.1%)	31 (26.0%)	1 (0.9%)
Triceps			77 (66.9%)	38 (33.0%)	—	115
	Unilateral	12 (10.4%)	3 (2.6%)	9 (7.8%)	—
	Bilateral	103 (89.5%)	74 (64.3%)	29 (25.2%)	—
Ulnar			73 (66.3%)	37 (33.6%)	—	110
	Unilateral	14 (12.7%)	4 (3.6%)	10 (9.0%)	—
	Bilateral	96 (87.2%)	69 (62.7%)	27 (24.5%)	—
Radial			72 (65.4%)	38 (34.5%)	—	110
	Unilateral	12 (10.9%)	3 (2.7%)	9 (8.1%)	—
	Bilateral	98 (89.1%)	69 (62.7%)	29 (26.3%)	—
Knee			132 (70.5%)	54 (28.8%)	1 (0.5%)	187
	Unilateral	35 (18.7%)	22 (11.7%)	13 (6.9%)	—
	Bilateral	152 (81.2%)	110 (58.8%)	41 (21.9%)	1 (0.5%)
Ankle			105 (73.9%)	36 (25.3%)	1 (0.7%)	142
	Unilateral	23 (16.1%)	15 (10.5%)	8 (5.6%)	—
	Bilateral	119 (83.8%)	90 (63.3%)	28 (19.7%)	1 (0.7%)
Plantar			17 (85.0%)	3 (15.0%)	—	20
	Unilateral	8 (40.0%)	7 (35.0%)	1 (0.5%)	—
	Bilateral	12 (60.0%)	10 (50.0%)	2 (10.0%)	—

^aAbnormal reflex for plantar reflex.

^bClonus reflex for plantar reflex.

^cAbsent-hypoactive on one side and hyperactive on the other side.

In upper extremity reflexes, bilateral involvement predominantly featured absent-hypoactive patterns than hyperactive patterns (biceps: 62.1% vs 26.0%, triceps: 64.3% vs 25.2%, ulnar: 62.7% vs 24.5%, radial: 62.7% vs 26.3%), while unilateral cases leaned toward hyperactive patterns than absent-hypoactive patterns (biceps: 8.4% vs 2.5%, triceps: 7.8% vs 2.6%, ulnar: 9.0% vs 3.6%, radial: 8.1% vs 2.7%). Conversely, in lower extremity reflexes, both bilateral and unilateral instances were predominantly marked by absent-hypoactive responses, with hyperactive responses being less frequent.

MRI results were available only for 92 patients. Thus, the data were not sufficient to be entered into the regression model. Approximately 22 of the 92 patients had abnormal brain MRI (brain ischemia or atrophy), 9 (41%) of them had abnormal neurologic findings, and 13 (59%) of them had normal neurologic examination. On the contrary, 70 patients had normal brain MRI: 21 (30%) of them had abnormal neurologic findings and only 49 (70%) of them had normal neurologic examination.

Univariate analysis (Tables 3 and 4) revealed that in comparison with patients with SCD with normal neurologic examination, those with abnormal findings had a significantly higher number of male patients (51.8% vs 47.0%, p = 0.050), were older at study entry (median = 17.0 (IQR = 21.0) years vs 14.0 (IQR = 16.0) years, p < 0.001), had higher smoking rates (47.7% vs 33.5%, p < 0.001), were more likely to be overweight/obese with body mass index ≥ 25 kg/m² (12.4% vs 7.7%, p < 0.001), had higher systolic blood pressure (median = 108.3 (IQR = 14.8) mm Hg vs 105.8 (IQR = 14.0) mm Hg, p < 0.001), had higher diastolic blood pressure (median = 65.6 (IQR = 12.0) mm Hg vs 64.7 (IQR = 11.0) mm Hg, p = 0.006), and had more painful crises (57.9% vs 47.4%, p < 0.001). Moreover, they were more likely to have a history of one of the following non-neurologic conditions: aseptic necrosis (25.3% vs 12.0%, p < 0.001), eye disease (12.1% vs 6.7%, p < 0.001), osteomyelitis (10.4% vs 5.1%, p < 0.001), hand-foot syndrome (41.2% vs 33.7%, p = 0.004), renal insufficiency (4.4% vs 2.7%, = 0.050), or lung infarction (5.6% vs 3.6%, = 0.050). They were also more likely to have had blood transfusion (65.4% vs 53.7%, p < 0.001). Laboratory data revealed that patients with abnormal neurologic examination had higher reticulocyte count (p = 0.025) and lower hemoglobin F (Hgb F, p = 0.005) and albumin (p < 0.001) levels.

Table 3.

Differences Among Categorical Variables Between Patients With Normal and Patients With Abnormal Neurologic Examination

Variable	Normal N (%)	Abnormal N (%)	p Value
Sex (female)	1237 (53.0)	248 (48.2)	0.050
Sex (male)	1097 (47.0)	266 (51.8)	0.050
Hemoglobin SS and SB variants	1402 (60.4)	306 (59.8)	0.789
Race (African American)	2275 (97.9)	502 (98.0)	0.870
BMI ≥25 kg/m2	157 (7.7)	75 (12.4)	<0.001
G6PD deficiency	110 (9.5)	29 (11.8)	0.266
Pneumonia vaccination	1149 (49.2)	240 (46.7)	0.298
Smoker (includes cigarettes, cigars, and pipes)	589 (33.5)	183 (47.7)	<0.001
Previous history of
Eye disease	140 (6.7)	54 (12.1)	<0.001
Hematuria	264 (12.6)	66 (14.4)	0.302
Nephrotic syndrome	16 (3.5)	45 (2.2)	0.080
Renal insufficiency	56 (2.7)	20 (4.4)	0.050
Acute priapism	121 (12.1)	37 (14.9)	0.234
Hearing loss	125 (6.0)	25 (5.5)	0.712
Heart disease	254 (12.2)	61 (13.6)	0.399
Aseptic necrosis	280 (12.0)	130 (25.3)	<0.001
Hand-foot syndrome	660 (33.7)	170 (41.2)	0.004
Osteomyelitis	107 (5.1)	47 (10.4)	<0.001
Spleen sequestration	227 (11.2)	53 (12.0)	0.607
Pneumonia	1115 (53.7)	256 (56.9)	0.223
Lung infarction	75 (3.6)	25 (5.6)	0.050
Liver sequestration	162 (7.8)	45 (10.1)	0.115
Painful crisis	992 (47.4)	358 (57.9)	<0.001
Blood transfusion	1124 (53.7)	297 (65.4)	<0.001
Family history of
Stroke	12 (0.6)	6 (1.3)	0.088
Seizures	19 (0.9)	6 (1.3)	0.423

Abbreviations: BMI = body mass index; G6PD = glucose-6-phosphate dehydrogenase.

Table 4.

Difference Among Continuous Variables Between Patients With Normal and Patients With Abnormal Neurologic Examination

Variable	Normal Median (IQR)	Abnormal Median (IQR)	p Value
SBP (mm Hg)	105.8 (14.0)	108.3 (14.8)	<0.001
DBP (mm Hg)	64.7 (11.0)	65.6 (12.0)	0.006
Age at entry visit (y)	14.0 (16.0)	17.0 (21.0)	< 0.001
Hemoglobin (g/dL)	9.0 (2.7)	9.0 (2.5)	0.827
RBC (*1012/L)	3.0 (1.4)	3.0 (1.3)	0.999
WBC (*109/L)	10.7 (4.6)	10.9 (4.2)	0.431
Platelets (*109/L)	387.9 (184.4)	378.6 (194.3)	0.608
Reticulocytes (%)	9.2 (8.4)	9.9 (9.2)	0.025
Hemoglobin F (%)	4.7 (6.6)	3.7 (5.6)	0.005
Albumin (g/dL)	4.4 (0.4)	4.2 (0.5)	0.001
LDH (mg/dL)	374.8 (302.6)	360.0 (255.8)	0.246

Abbreviations: DBP = diastolic blood pressure; IQR = interquartile range; LDH = lactate dehydrogenase; RBC = red blood cells; SBP = systolic blood pressure; WBC = white blood cells.

Using logistic regression, we then explored whether the aforementioned significant variables were predictors of abnormal neurologic examination (Table 5). The logistic regression model was statistically significant, χ² (17) = 84.68, p < 0.001, explained 12.5% (Nagelkerke R2) of the variance, and correctly classified 82.6% of cases. Sensitivity was 9.9%, specificity was 98.6%. History of eye disease (OR = 2.76, p < 0.001, 95% CI 1.63–4.68) and history of osteomyelitis (OR = 2.55, p = 0.004, 95% CI 1.34–4.84) were the strongest predictors of abnormal neurologic findings, followed by smoking (OR = 1.59, p = 0.018, 95% CI 1.08–2.33), aseptic necrosis (OR = 1.57, p = 0.024, 95% CI 1.06–2.33), hand-foot syndrome (OR = 1.48, p = 0.029, 95% CI 1.04–2.12), and male sex (OR = 1.42, p = 0.050, 95% CI 1.01–2.02).

Table 5.

Binary Logistic Regression Predicting the Likelihood of Having an Abnormal Neurologic Examination Among Patients With SCD Who Did Not Experience a Previous Major Neurologic Event

Variable	Odds ratio	95% CI	p Value
Sex (male)	1.42	1.01–2.02	0.050
Age at entry visit (y)	1.01	0.99–1.03	0.179
Smoking	1.59	1.08–2.33	0.018
BMI ≥25 kg/m2	1.36	0.80–2.32	0.254
SBP	0.99	0.97–1.02	0.881
DBP	0.99	0.96–1.02	0.647
Previous history of
Eye disease	2.76	1.63–4.68	<0.001
Aseptic necrosis	1.57	1.06–2.33	0.024
Osteomyelitis	2.55	1.34–4.84	0.004
Hand-foot syndrome	1.48	1.04–2.12	0.029
Renal insufficiency	1.54	0.66–3.59	0.310
Blood transfusion	1.17	0.79–1.74	0.427
Lung infarction	1.49	0.70–3.16	0.293
Painful crisis	1.05	0.75–1.49	0.743
Blood and liver tests
Reticulocytes (%)	1.01	0.97–1.03	0.753
Hemoglobin F (%)	0.97	0.94–1.01	0.276
Albumin <3.5 (g/dL)	1.05	0.66–1.68	0.818

Abbreviations: BMI = body mass index; CI = confidence interval; DBP = diastolic blood pressure; LDH = lactate dehydrogenase; SBP = systolic blood pressure.

Discussion

By virtue of their disease process, patients with SCD commonly experience overt neurologic events such stroke or seizures or covert events such as SCI, which despite being silent, carry tremendous risk for developing future stroke and neurologic decline. Thus, identifying clinical parameters that predict neurocognitive deficits, especially in those with asymptomatic involvement, may be particularly helpful. We first assessed the prevalence of abnormal neurologic physical findings, representative of neurologic involvement, in patients with SCD with no history of neurologic event and found that 14.5% had at least 1 abnormal finding, most commonly an abnormal deep tendon reflex. We then assessed which clinical parameters significantly predicted an abnormal neurologic examination in patients with SCD and found that history of eye disease, osteomyelitis, smoking, aseptic necrosis, hand-foot syndrome, and male sex were all significant predictors. Closely monitoring patients with SCD with such attributes may help prevent neurologic decline, especially in places where early screening may not be feasible.

Given that in high-income countries nowadays virtually all patients with SCD are routinely monitored with brain MRI and TCD, and this level of monitoring may not be readily accessible in low–middle-income countries due to various constraints. The strength of this study is that it summarizes the neurologic examination findings in asymptomatic patients with SCD with neurologic involvement and other clinical findings predictive of neurologic involvement. This is particularly useful in low–middle-income countries, where such patients should be referred to tertiary centers for imaging and further testing, whenever feasible.

The most common abnormal neurologic finding in our SCD cohort was an abnormal deep tendon reflex, mostly absent-hypoactive and most commonly in the lower limbs (namely knee and ankle). Hyperreflexia indicates CNS involvement and most likely results from SCD-related effects on cerebral circulation. On the contrary, hyporeflexia is rather unique because it suggests SCD-related peripheral nerve involvement.²⁶ This is further supported by the finding that most of the hyporeflexia involvement was bilateral in nature. Evidence from animal studies demonstrates diffuse sciatic nerve myelin instability, which could be attributed to increased permeability of the blood-nerve barrier leading to the accumulation of demyelinating factors.²⁷ Although most studies explore central neurologic complications of SCD,²⁸ there is evidence to suggest that SCD creates a neurotoxic environment for peripheral nerves.²⁷ Moreover, the etiology of abnormal gait in SCD may be central due to cerebellar infarcts, leading to manifestations of broad-based ataxia and additional cerebellar signs, or due to peripheral neuropathy. Further studies exploring the peripheral neurologic complications of SCD may improve our understanding of disability patterns and the nature of pain in this population.

Silent cerebral infarcts are an extremely common cause of unexplained neurologic decline in SCD. Studies have shown that, by their 6th birthday, 27% of SCD will have silent cerebral infarcts.⁵ This number will increase to 37% and 39% by the 14th and 18th birthdays, respectively.⁴ Apart from the direct effect of sickling and occlusion on brain vessels, the chronic hypoxic environment might distort the cerebral circulation in a unique way. Sickling episodes and the subsequent hypoxia will induce endothelial hyperplasia and nonatherosclerotic vascular stenosis. Ultimately, new small collateral vessels resembling “puff of smoke” will form or what is called Moyamoya syndrome.^29,30 All in all, the aforementioned structural changes will result in a gradual chronic neurocognitive decline even without punctuating acute neurologic events.³¹

The MRI findings in our study showed that 30% (21/70) of patients who had normal brain MRI were found to have abnormal neurologic findings. This was consistent with the findings from other studies, which suggested that children with SCD with normal MRIs may still experience neurocognitive impairments.¹⁰ It is worth noting that some of this 30% may be due to peripheral processes, such as peripheral neuropathy. Therefore, to better understand the underlying mechanisms contributing to these neurologic findings, further investigations into the specific etiology of the abnormalities observed in this subgroup of patients are warranted. Thus, it is important to screen patients with SCD for neurologic involvement using regular neurologic examination and transcranial Doppler ultrasound in addition to MRI to facilitate early detection and intervention to prevent subsequent neurologic decline.¹⁶ Relying solely on MRI may not be sufficient to detect all potential neurologic impairments because certain peripheral factors can contribute to the observed abnormalities.

Our multivariable model demonstrated that history of eye disease, osteomyelitis, smoking, aseptic necrosis, hand-foot syndrome, and male sex were all significant predictors of abnormal neurologic findings. These predictors reflect a general scheme of vasculopathy, endothelial dysfunction, and microinfarcts and is consistent with previous studies.^4,32

Osteomyelitis, hand-foot syndrome, and aseptic necrosis episodes reflect poor disease control and result in more sickling and hemolysis, which releases more inflammatory mediators and vascular growth factors, commonly resulting in diffuse vasculopathy.³² The strong association between these bone pathologies and abnormal neurologic findings may also be suggestive of fat embolus syndrome, wherein bone marrow necrosis causes nontraumatic shedding of multiple fat emboli, which travel throughout the venous system, resulting in various organ-specific manifestations.³³ Should the emboli reach the arterial system, a range of cerebral manifestations may occur, from mild confusion or anxiety to overt focal symptoms or even encephalopathy.^34-36 Moreover, in those with painful episodes, subtle neurologic symptoms may be overlooked in the context of the overall clinical presentation and thus often go unnoticed for years, allowing them to progress. In addition, certain patients with SCD may experience chronic or neuropathic pain, necessitating the use of potent pain medications, such as opioids, which might obscure the presence of underlying neurologic abnormalities.³⁷ Thus, our findings suggest that patients recovering from osteomyelitis or aseptic necrosis should be evaluated by a neurologist to document any neurologic deficits, in the hope of curbing further decline.

One of the strongest predictors of abnormal neurologic examination in our cohort of patients with SCD was history of eye disease. Previous reports have shown that patients with SCD have a heightened risk of neuro-ophthalmic complications involving the retina, visual tract, and even the cortex.^25,38 The development of the visual pathway and other cortical functions, such as movement perception, reading and writing, are closely linked.³⁹ Retinal neurons are subject to the same SCD-related insults that affect the CNS, and given the reciprocal interaction between the retina and the rest of the visual pathway, retinal insults may have negative neurotropic effects on other visual pathway components.^39,40As such, the eye can be viewed as a window to the CNS, and examining the fundus and retina may provide an affordable and practical means of evaluating and stratifying patients with SCD according to their risk of subsequent neurologic decline.

Our study reported abnormal neurologic findings in nearly half (47.7%) of smokers, which was significantly higher than those in nonsmokers. Previous studies have linked smoking to increased frequency of acute chest syndrome episodes and community-acquired pneumonia in patients with SCD.⁴¹ Children exposed to tobacco smoke were found to have had more sickle cell crises requiring hospitalization than unexposed children.⁴² This could be related to the fact that exposure to tobacco products increases vaso-occlusive disease through increasing vascular inflammation, platelet aggregation, and expression of endothelial adhesion molecules.⁴³ An association between smoking and neurologic manifestations could thus be explained by the fact that neurologic findings in patients with SCD are mostly the result of vaso-occlusive events, which are more likely to occur in smokers. It is also possible that an acute neurologic insult may have occurred because of hypoxia in the setting of an acute chest syndrome or chest infection. A direct effect of smoking on the CNS through the generation of free radicals and inflammatory mediators in the general population was also previously suggested.⁴⁴ More studies are yet to be conducted to observe the presence and absence of abnormal neurologic manifestations in patients with SCD exposed to smoking. Considering that it is a modifiable risk factor, efforts should be directed at educating the general population and patients with SCD to avoid both firsthand and secondhand smoking.

We found that there was a significantly increased risk of abnormal neurologic findings in male patients. This was consistent with one study that reported an increased risk of silent cerebral infarct in male patients with sickle cell anemia.²⁰ Many factors may have contributed to these observed effects because male patients seem to have higher overall blood pressure readings than female patients. Although our study found no association between higher blood pressure readings and reported abnormal neurologic manifestations, this remains debatable because other studies have recognized the association.^20,45 Smoking status may also be a contributor to the observed effects because previous studies have suggested that men use tobacco products at a higher rate than women.⁴⁶ On the contrary, one study found no significant difference in the prevalence of CVA between male and female patients with SCD.⁴⁷

Fetal hemoglobin inhibits polymerization of deoxy sickle hemoglobin (HbS), and this is the rationale behind therapies such as hydroxyurea.⁴⁸ We found that a lower percentage of fetal hemoglobin (HbF) was not predictive of abnormal neurologic findings. Previous study findings have been inconsistent; some similarly demonstrate that HbF does not associate with silent infarct risk in patients with SCD with no history of CVA,^20,49,50 while others suggest that HbF may protect against silent white matter changes (WMC), decreasing the likelihood of WMC being present and their severity. It may therefore be beneficial to increase HbF levels in adult patients with WMC who have no history of CVA.^51,52 Some data suggest that the clinical relevance of HbF in predicting SCD severity is questionable, considering that it is the distribution of HbF among F cells that matters; only when the total HbF concentration is near 30% is it possible for the number of protected cells—those with sufficient concentrations to inhibit HbS polymerization—to approach 70%.⁵³ In addition, higher HbF levels are associated with reduced rates of acute painful episodes, leg ulcers, osteonecrosis, acute chest syndrome, and reduced disease severity. However, it had no clear association with silent cerebral infarction, among others.⁵⁴ Thus, HbF is not capable of uniformly modulating all complications of sickle cell disease, which is likely related to the variable pathophysiologic processes underlying these complications.

Several studies have demonstrated an association between elevated relative systolic blood pressure and SCI in patients with SCD.²⁰ Patients with SCD have lower blood pressure than the general population, which may be explained partly by the compensatory systemic vasodilatation resulting from chronic anemia. Elevated systolic blood pressure thus may offset this process.⁵⁵ In our SCD cohort, we found that systolic blood pressure was higher among patients with abnormal physical examination; however, the difference was not statistically significant. This can be due to the exclusion of patients with more severe disease, such as those with overt stroke or seizures, which might have limited our ability to detect clinically significant associations.

Given the high risk of strokes in patients with SCD, early screening is implemented to help prevent them.¹⁹ Noninvasive transcranial Doppler ultrasound is the primary method by which patients are risk stratified.⁵⁶ Based on transcranial Doppler velocity (TDC) measures, patients may be placed the patient may be considered for transfusion therapy or other treatment for prevention of stroke.^21,57 This has been shown to considerably reduce risk of stroke.^58,59 Despite its utility, however, it is not without limitations. TDC measures are operator dependent, can vary from time to time and most importantly, access is not readily available in lower income countries. This is specially relevant given that the highest prevalence of SCD is in countries of Africa.¹⁷ Moreover, it is ineffective in predicting silent cerebral infarcts, for which MRI is sought instead. Even if chronic transfusion therapy was not feasible in these countries, special considerations and interventions can be implemented that can help improve the neurologic morbidity. We recommend conducting regular neurologic examinations to patients with SCD, especially in resource-poor countries where screening transcranial Doppler or MRI is not readily available⁶⁰; in addition, particular consideration with more frequent neurologic evaluations should be given to those who have risk factors of developing abnormal neurologic findings. Thus, future screening guidelines should rely not only on neuroimaging but should take into consideration the neurologic history, examination findings, and potential risk factors.

This study has several limitations. First, 98.3% of our sample are of African American ethnicity. However, this closely resembles the real world because nearly 93% of patients with SCD are of African descent.⁶¹ Second, although we excluded 3 major neurologic events that commonly occur in patients with SCD (meningitis, stroke, and seizures), we did not take into account other potential causes of abnormal neurologic findings such as genetic disorders, congenital abnormalities, perinatal hypoxia, trauma, or undocumented stroke/meningitis. Third, neurologic assessment is subjective because it is evaluator dependent. Due to the retrospective nature of data, there is lack of standardization of physical examination findings. Moreover, some of the details of physical examination findings were limited to what was described in the original study, such as the exact description of cerebellar signs. In addition, some of the physical examination findings could be attributed to defects outside of the neurologic system, such as gait problems, which could be due to musculoskeletal abnormalities. Fourth, we did not account for whether patients are on hydroxyurea or on prophylactic transfusions. In addition, our sample did not use a control group (people without SCD), which would have helped make us more confident about the origin of abnormal neurologic findings. We also had insufficient data on medications or brain imaging findings. Being on hydroxyurea, for example, would alter the baseline risk considerably and alter the clinical parameters assessed. In addition, specific medications might interact differently with the disease's pathophysiology, leading to varying neurologic manifestations. Moreover, it would have been helpful to see how abnormal neurologic examination findings correlated with lesions on imaging, which could help establish a stronger correlation between neurologic abnormalities and specific clinical symptoms and to determine the localization and extent of the disease burden and tailor specific interventions for these patients. Finally, a prospective study could provide further insights into the concepts raised in this study and potentially address the limitations more comprehensively. Nevertheless, our large multicenter sample with a relatively good documentation of neurologic findings and follow-up of various potential predictors provides a reliable description of this unique population.

Neurologic deficits are relatively common in patients with SCD, even without having a documented neurologic insult. These manifestations range from peripheral and ophthalmic deficits to the more prominent central and cognitive disabilities. The large variability of manifestations suggests a multidimensional neuropathology, involving chronic hypoxia, demyelination, and fat embolization, and not merely the old paradigm of vasculopathy and cerebrovascular events. Therefore, it seems that the current screening guidelines are inadequate in monitoring neurologic decline in patients with SCD. Patients with SCD should have regular neurologic evaluations, including comprehensive neurologic examination, retinal examination, and risk factor modification. The frequency of follow-ups should probably be tailored based on risk factors suggested by this study and other similar studies. Before implementing these screening protocols, long-term longitudinal studies are needed to evaluate the impact of these potential risk factors more clearly. Moreover, the feasibility and cost-effectiveness of these screening visits and potential protective interventions should be extensively studied.

Appendix. Authors

Name	Location	Contribution
Mohammed B. Nawaiseh, MD	Department of Ophthalmology, Jordanian Royal Medical Services, Amman, Jordan	Drafting/revision of the article for content, including medical writing for content; study concept or design; and analysis or interpretation of data
Ahmed M. Yassin, MD	Neurology, Jordan University of Science and Technology, Irbid, Jordan	Drafting/revision of the article for content, including medical writing for content; study concept or design
Mohammed Q. Al-Sabbagh, MD	Neurology, Kansas University Medical Center, Kansas City	Drafting/revision of the article for content, including medical writing for content; analysis or interpretation of data
Ahmad AlNawaiseh, MD	Internal Medicine, St. Elizabeth's Medical Center, Boston, MA	Drafting/revision of the article for content, including medical writing for content; analysis or interpretation of data
Hadil Zureigat, MD	Internal Medicine, Cleveland Clinic, OH	Drafting/revision of the article for content, including medical writing for content; analysis or interpretation of data
Dina Aljbour AlMajali, MD	Internal medicine, Henry Ford Health System, Detroit, MI	Drafting/revision of the article for content, including medical writing for content
Rund R. Haddadin, MD	Department of Internal Medicine, JCESOM, Marshall University, WV	Drafting/revision of the article for content, including medical writing for content; analysis or interpretation of data
Mohammad El-Ghanem, MD	Department of Clinical Sciences, College of Medicine, University of Houston, HCA Northwest Medical Center, TX	Drafting/revision of the article for content, including medical writing for content
Mohammad Abu-Rub, MD	Neurology, George Washington University, Washington, DC	Drafting/revision of the article for content, including medical writing for content

Study Funding

The authors report no targeted funding.

Disclosure

The authors report no relevant disclosures. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/cp.