OBJECTIVE. Iron-deficiency anemia occurs with a peak prevalence of 4% to 8% in children between 1 and 3 years of age. Case reports have suggested an association between iron-deficiency anemia in healthy children and ischemic stroke. Our objective was to investigate whether iron-deficiency anemia is associated with stroke in young children.
METHODS. A case-control study was conducted of case patients who were selected from the stroke registry at the Hospital for Sick Children (Toronto, Ontario, Canada) and control subjects selected from a database of healthy children who were prospectively enrolled in an outpatient setting. Children were aged 12 to 38 months and were previously healthy with no identifiable risk factors for stroke. Age, gender, mean corpuscular volume, platelet count, and hemoglobin and ferritin levels were collected. Iron-deficiency anemia was defined as a hemoglobin level of <110 g/L, mean corpuscular volume <73 fL, and serum ferritin level <12 μg/L. Stroke was defined according to clinical and radiologic criteria.
RESULTS. Case (n = 15) and control (n = 143) subjects were similar with respect to median age and percentage of boys. Case patients had a lower median hemoglobin level and mean corpuscular volume and a higher median platelet count. Iron-deficiency anemia was significantly more common among case patients (8 [53%] of 15) than control subjects (13 [9%] of 143).
CONCLUSIONS. Previously healthy children with stroke were 10 times more likely to have iron-deficiency anemia than healthy children without stroke. Furthermore, children with iron-deficiency anemia accounted for more than half of all stroke cases in children without an underlying medical illness, which suggests that iron-deficiency anemia is a significant risk factor for stroke in otherwise healthy young children. Primary prevention and early identification of iron-deficiency anemia must remain a priority.
Childhood ischemic stroke includes both arterial ischemic stroke (AIS) and sinovenous thrombosis (SVT) and occurs at a frequency of 2 to 3 per 100000 children each year in North America.1 Although childhood stroke is rare, it can lead to devastating long-term consequences and occasionally to death. A prospective cohort study of 163 children who survived ischemic stroke demonstrated that >40% had persistent moderate to severe neurologic deficits.2 In recent years, the case fatality rate has been estimated at 10% to 20%.3–6
Several risk factors for childhood ischemic stroke have been consistently reported in large cohort studies of children with stroke. These include sickle cell disease, cardiac structural lesions, chronic systemic disease, cerebral arterial disease, coagulation disorders, head trauma, and subacute varicella zoster infection.1,3,6–8
Several case reports and 1 case series of 6 children suggested an association between iron-deficiency anemia (IDA) in healthy children and ischemic stroke.9–13 One large cohort study found anemia to be present in 24% of 115 previously healthy children with AIS.8 The strength and the magnitude of this association have not been previously reported.
IDA is a relatively common disorder in developed countries, such as Canada, the United Kingdom, and the United States, occurring with a peak prevalence of 4% to 8% in children between 12 and 36 months of age.14–16 The health implications of IDA are thought to be cognitive and motor developmental delay, which may not be improved with iron therapy.17,18
The aim of this study was to investigate whether IDA is a risk factor for childhood ischemic stroke. We conducted a case-control study to compare the prevalence of IDA in otherwise healthy young children at the time of stroke with the prevalence of IDA in age-matched healthy children.
The sampling frame for the study was previously healthy children between 12 and 38 months of age, because this is a group with a high prevalence of nutritional IDA. A case-control study design was used, with selection of case patients and control subjects from children who were enrolled prospectively in 2 databases.
Case patients were selected from the Canadian Pediatric Ischemic Stroke Registry. Children who presented to the Hospital for Sick Children (HSC) between January 1992 and December 2004 with AIS or SVT were considered for eligibility. AIS was defined as a focal neurologic deficit of acute onset and a computed tomography scan or MRI scan of the brain showing a lesion characteristic of a focal arterial infarct in a vascular territory consistent with the neurologic presentation. SVT was defined as headache, altered level of consciousness, seizure, or focal neurologic deficit with venography (computed tomography or MRI) showing occlusion or focal reduction of flow within cerebral venous sinuses or cerebral veins. A manual review of the hospital charts of each child who met these criteria was undertaken to assess eligibility for the study. Only children who did not have sickle cell disease, chronic systemic disease (excluding IDA), malignancy, cerebral arterial disease, meningitis, encephalitis, head trauma, recent surgery, heart disease, or indwelling catheters were included as case patients. All children who were enrolled in the stroke registry were screened for prothrombotic factors, which included activity assays for antithrombin, protein C, protein S, and the lupus anticoagulant, as well as immunologic assays for anticardiolipin antibody and molecular assays for the presence of factor V Leiden and the G20210A mutation in the prothrombin gene. Children with any of these factors identified after the stroke were not excluded; case patients with acute minor infections were also not excluded.
Control subjects were selected from a database of children who were prospectively enrolled in an observational study of iron deficiency.19 Children were recruited from a university-affiliated community-based pediatric practice in Toronto between November 2002 and March 2004 while attending a well-child visit. Exclusion criteria were acute febrile illness at enrollment (which may falsely increase the serum ferritin level, which was the primary outcome measure in the study), history of illness or medication associated with iron deficiency or anemia, previous diagnosis of iron depletion or anemia, current use of iron supplements, and breast milk or formula as primary source of milk at or within 8 weeks of enrollment (which provides more available iron). For the purposes of this study, children were included when blood work had been obtained.
For both case patients and control subjects, the following variables were collected: age, gender, mean corpuscular volume (MCV), platelet count, and hemoglobin and serum ferritin levels. For case patients, hematologic parameters that were obtained before the initial administration of intravenous fluids or intravenous medications either at HSC or at the child's referring hospital were abstracted from hospital charts. For control subjects, all variables were obtained concurrently with enrollment in the observational study. IDA was defined as serum ferritin level of <12 μg/L, MCV of <73 fL, and hemoglobin level of <110 g/L.20–23 Thrombocytosis was defined as a platelet count of >450 × 109/L.24
Baseline characteristics of the 2 groups were compared using the Pearson χ2 test and the Mann-Whitney U test for gender and age, respectively. Hemoglobin level, MCV, and platelet count were compared as continuous variables by using the Mann-Whitney U test. The association between stroke and IDA and thrombocytosis were calculated using odds ratios (ORs) and 95% confidence intervals (CIs). Hemoglobin levels, MCV, and platelet count were compared as continuous variables by using the Mann-Whitney U test. Significance was defined as P ≤ .05. Binomial logistic regression was used to examine interaction between variables. Data analysis was performed by using SPSS 13 (SPSS Inc, Chicago, IL). Data collection for the stroke registry and the community cohort study was approved by the HSC Research Ethics Board, and informed parental consent was obtained.
Fifty-six previously healthy children who were 12 to 38 months of age presented to HSC with stroke between January 1992 and December 2004. Fifteen (27%) of these children were generally healthy before stroke and were included as case patients (Table 1). Reasons for exclusion were heart disease (n = 15), meningitis (n = 5), malignancy (n = 5), head trauma (n = 5), cerebral arterial disease (n = 3), nephrotic syndrome (n = 2), mastoiditis (n = 2), and other (n = 4). Initial hemoglobin level, MCV, and platelet count were available for all case patients. Serum ferritin level was available for all cases with hemoglobin levels of <110 g/L and MCV of <73 fL.
A total of 332 healthy children who attended a well-child visit at the community practice met full eligibility criteria, and 143 of these had blood work available and were included as control subjects. Data for platelet counts were missing for 10 control subjects.
There were no significant differences in median age (24 vs 21 months) or percentage of boys (73% vs 65%) for case patients as compared with control subjects. Mean and median hemoglobin levels were significantly lower in case patients than in control subjects (88 and 92 g/L vs 116 and 120 g/L, respectively; P < .01). Mean and median MCVs were also significantly lower in case patients than in control subjects (65 and 57 fL vs 76 and 77 fL, respectively; P < .01). The mean and median platelet counts were significantly higher in case patients than in control subjects (458 and 415 × 109/L vs 358 and 351 × 109/L, respectively; P = .05). IDA was significantly more common among case patients (8 of 15 [53%]) than control subjects (13 of 143 [9%]; OR: 12; 95% CI: 4–37). Thrombocytosis was also more common among case patients (7 of 15 [47%]) than control subjects (21 of 133 [16%]; OR: 5; 95% CI: 2–14). Using logistic regression to control for platelet count, the association between IDA and stroke remained significant (adjusted OR: 10; 95% CI: 3–33). There was no statistically significant interaction between IDA and thrombocytosis.
Three case patients with stroke were identified with previously described prothrombotic risk factors. All had subacute varicella infection, 1 of whom was also heterozygous for the factor V Leiden mutation, and none had IDA. Six case patients with stroke were noted to have acute minor infection (acute otitis media, pneumonia, acute gastroenteritis, or upper respiratory tract infection), 4 with IDA and 2 without IDA. Five case patients with stroke and IDA had SVT, 2 had AIS, and 1 had both. All 7 case patients with stroke and without IDA had AIS (Table 1).
The results of this case-control study suggest that previously healthy children who develop vasoocclusive stroke are 10 times more likely to have IDA than healthy children who do not develop stroke. Furthermore, children with IDA accounted for more than half of stroke cases in children without an underlying medical illness, suggesting that IDA is a significant risk factor for stroke in otherwise healthy children.
Of note, SVT was more closely associated with IDA than was AIS. Although a previously reported case series of 6 children with IDA and stroke showed equal numbers of children with AIS and SVT,13 our data may indicate that IDA is a greater risk factor for SVT than for AIS.
The strength and magnitude of several other risk factors for stroke have been identified. As reported in the published literature, these include activated protein C resistance (OR: 9.4; 95% CI: 2.0–44.6), elevated lipoprotein(a) (OR: 7.2; 95% CI: 3.8–13.8), antiphospholipid antibodies (OR: 6.08; 95% CI: 1.5–24.3), prothrombin gene variant (OR: 4.7; 95% CI: 1.4–14.6), factor V Leiden mutation (OR: 4.3; 95% CI: 2.8–6.5), subacute varicella infection (estimated OR: 3), and methylenetetrahydrofolate reductase (MTHFR) mutation (OR: 2.64; 95% CI: 1.53–4.50).8,25,26 IDA, with an OR of 10 (95% CI: 3–33) for stroke, is at least as strong a risk factor as these other risk factors. Furthermore, IDA was more common than other prothrombotic risk factors in this population of children with stroke (8 of 15 [53%] vs 3 of 15 [20%]). We were unable to assess whether IDA is an independent or additive risk factor.
Three mechanisms to explain an association between IDA and childhood ischemic stroke have been suggested: a hypercoagulable state directly related to iron deficiency and/or anemia; thrombocytosis secondary to IDA; and anemic hypoxia, whereby a mismatch between oxygen supply and end-artery oxygen demand leads to ischemia and infarction.13 Our data support an association between thrombocytosis and stroke. Healthy children who develop stroke may be 5 times more likely to have thrombocytosis than children who do not develop stroke; however, no statistically significant interaction between IDA and thrombocytosis was found. IDA seems to be an independent risk factor for stroke.
This study has several limitations. There may have been differences between case patients and control subjects other than stroke. First, children were excluded as control subjects when they had a history of breast milk or formula intake in the previous 8 weeks or an acute febrile illness but were not excluded as case patients for these reasons. Breast milk and iron-fortified formula are considered to be protective against IDA, and acute illness may falsely increase the serum ferritin level; therefore, it is more likely that these biases would lead to an underestimation of the magnitude of the association between IDA and stroke. Data on nutrition practices of case patients were not available, and we were not able to explore causal mechanisms for the development of IDA. Second, because of the rarity of childhood ischemic stroke, the time frame for enrollment of case patients was necessarily longer than that for control subjects (January 1992 through December 2004 versus November 2002 through March 2004). A change in the frequency of IDA between 1992 and 2002 could introduce bias; however, we believe that this is not likely because the published rate of iron deficiency and IDA among toddlers in Canada and the United States has not appreciably changed during this interval.14,20,27 Finally, the results from this study at most pertain to children in the age range explored, that is, 12 to 38 months, and the proportion of young children who have IDA and are at risk for stroke cannot be determined by this study.
Although IDA is thought to be an indolent disease with insidious long-term consequences, it may hold a 10-fold increased risk for acute stroke in well toddlers and may account for half of all strokes in otherwise well children of this age group. This research supports additional development of strategies aimed at the primary prevention and early detection of IDA in young children.
This study was supported by a grant-in-aid from Danone Institute of Canada (community cohort) and grants from the Heart and Stroke Foundation of Ontario and the Bloorview Children's Hospital Foundation (stroke registry). The Pediatric Outcomes Research Team is supported by a grant from the Hospital for Sick Children Foundation. Dr deVeber is the recipient of a Stroke Investigator Award from the Heart and Stroke Foundation of Ontario. The funding bodies had no role in design or conduct of the study; collection, management, analysis, or interpretation of the data; or preparation, review, or approval of the manuscript.
We thank Trenna Sutcliffe, MD, Sheila Jacobson, MD, Michael Peer, MD, and Shelley Westergard, RN, for recruitment of control subjects.
- Accepted May 25, 2007.
- Address correspondence to Patricia C. Parkin, MD, Division of Pediatric Medicine and the Pediatric Outcomes Research Team, Hospital for Sick Children, 555 University Ave, Toronto, Ontario, Canada M5G 1X8. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Lynch JK, Hirtz DG, deVeber G, Nelson KB. Report of the National Institute of Neurological Disorders and Stroke workshop on perinatal and childhood stroke. Pediatrics.2002;109 :116– 123
- ↵deVeber G, MacGregor D, Curtis R, Mayank S. Neurologic outcome in survivors of childhood arterial ischemic stroke and sinovenous thrombosis. J Child Neurol.2000;15 :316– 325
- ↵Lanthier S, Carmant L, David M, Larbrisseau A, de Veber G. Stroke in children: the coexistence of multiple risk factors predicts poor outcome. Neurology.2000;54 :371– 378
- Fullerton HJ, Elkins JS, Johnston SC. Pediatric stroke belt: geographic variation in stroke mortality in US children. Stroke.2004;35 :1570– 1573
- ↵Sebire G, Tabarki B, Saunders D, et al. Cerebral venous sinus thrombosis in children: risk factors, presentation, diagnosis and outcome. Brain.2005;128 :477– 489
- ↵Askalan R, Laughlin S, Mayank S, et al. Chickenpox and stroke in childhood: a study of frequency and causation. Stroke.2001;32 :1257– 1262
- ↵Belman AL, Roque CT, Ancona R, Anand AK, Davis RP. Cerebral venous thrombosis in a child with iron deficiency anemia and thrombocytosis. Stroke.1990;21 :488– 493
- Baptist EC, Castillo SF. Cow's milk-induced iron deficiency anemia as a cause of childhood stroke. Clin Pediatr (Phila).2002;41 :533– 541
- Saxena K, Ranalli M, Khan N, Blanchong C, Kahwash SB. Fatal stroke in a child with severe iron deficiency anemia and multiple hereditary risk factors for thrombosis. Clin Pediatr (Phila).2005;44 :175– 180
- Booth IW, Aukett MA. Iron deficiency anemia in infancy and early childhood. Arch Dis Child.1997;76 :549– 554
- ↵Grantham-McGregor S, Ani C. A review of studies on the effect of iron deficiency on cognitive development in children. J Nutr.2001;131 :659S– 666S
- ↵Martins S, Logan S, Gilbert R. Iron therapy for improving psychomotor development and cognitive function in children under the age of three with iron deficiency anemia. Cochrane Database Syst Rev.2001;(2) :CD001444 .
- ↵Lozoff B, De Andraca I, Castillo M, Smith JB, Walter T, Pino P. Behavioral and developmental effects of preventing iron-deficiency anemia in healthy full-term infants [published correction appears in Pediatrics. 2004;113:1853]. Pediatrics.2003;112 :846– 854
- Sherriff A, Edmond A, Hawkins N, Golding J. Hemoglobin and ferritin concentrations in children 12 and 18 months. Arch Dis Child.1999;80 :153– 157
- Expert Scientific Working Group. Summary of a report on assessment of the iron nutritional status of the United States population. Am J Clin Nutr.1985;42 :1318– 1330
- ↵Nowak-Gottl U, Strater R, Heinecke A, et al. Lipoprotein(a) and genetic polymorphisms of clotting factor V, prothrombin, and methylenetetrahydrofolate reductase are risk factors of spontaneous ischemic stroke in childhood. Blood.1999;94 :3678– 3682
- ↵Brotanek JM, Gosz J, Weitzman M, Flores G. Iron deficiency in early childhood in the United States: risk factors and racial/ethnic disparities. Pediatrics.2007;120 :568– 575
- Copyright © 2007 by the American Academy of Pediatrics