Prospective, National Clinical and Epidemiologic Study on Imported Childhood Malaria in the United Kingdom and the Republic of Ireland

Abstract

Background: 

Current knowledge of clinical features of imported childhood malaria is largely limited to small, retrospective, and/or single-center case series. This prospective, population-based study describes the epidemiology and clinical features of imported childhood malaria in children <16 years in the United Kingdom and Republic of Ireland.

Methods: 

Active prospective national surveillance with clinical data collection was performed between January 1, 2006 and January 31, 2007 through the British Pediatric Surveillance Unit and capture-recapture analysis using cases reported independently to respective national surveillance centers.

Results: 

There were 290 cases, including 186 reported through the British Pediatric Surveillance Unit with clinical details. Capture-recapture analysis estimated the burden of imported childhood malaria to be 2.8/100,000 per year for the United Kingdom and 4.6/100,000 per year for Ireland. Black-African children born in the United Kingdom and Ireland and traveling to West Africa during school holidays without antimalarial prophylaxis accounted for the majority of cases. Thirty of 117 children (26%) who had traveled to a malaria-endemic country had previously been diagnosed with malaria, reflecting missed opportunities to educate families on malaria prevention. A third of children (46/148) with Plasmodium falciparum malaria fulfilled World Health Organization criteria for severe or potentially complicated malaria, although only 11/46 (24%) required intensive care. The choice of antimalarials varied considerably among hospitals and within the same hospital. However, recrudescence occurred in only 1 child and none died.

Conclusions: 

Interventions to prevent imported childhood malaria should focus on Black-African families traveling to West Africa, while pediatricians should be offered clearer guidance on the treatment of childhood malaria.

The World Health Organization estimates that around 250 million individuals develop malaria every year and almost a million die of the infection, particularly children <5 years old.¹ The burden of disease is greatest in Sub-Saharan Africa, where the predominant Plasmodium falciparum species causes most of the severe complications and deaths from malaria.² In Europe, endemic malaria has been eliminated, while rates of malaria in parts of Africa are dropping,³ but cases imported from malaria-endemic countries have continued to increase in the past 4 decades as air travel becomes more accessible, acceptable, and affordable.^4,5 In the United Kingdom, around 2000 imported malaria cases are reported to the national Malaria Reference Laboratory (MRL) annually, of which 10% to 15% occur in children <15 years.^6,7 In the Republic of Ireland, imported malaria was rare before 2003, with fewer than 20 annual notifications to the Health Protection Surveillance Centre,⁸ but this has recently increased, with almost a hundred cases reported in 2006.^9,10

Enhanced surveillance for malaria in the United Kingdom and Ireland is co-ordinated by national reference centers, but it is not known how accurately their estimates reflect the true burden of disease in children. In addition, because their data are primarily based on laboratory notifications, knowledge of clinical features of imported childhood malaria is largely limited to small, retrospective, single-center case series. The extensive literature on childhood malaria in Africa is of limited relevance to nonendemic countries, where children of all ages are likely to be nonimmune to malaria. National guidelines exist for management of imported childhood malaria,¹¹ but it is unclear if they are followed or whether the recommended therapy is effective. The objectives of this study were to estimate the burden of imported childhood malaria in the United Kingdom and Ireland; to describe clinical and laboratory features, management, complications, and outcome; and, to identify risk factors for and complications of severe malaria in a nonendemic setting.

METHODS

The British Pediatric Surveillance Unit (BPSU) of the Royal College of Pediatrics and Child Health is an active surveillance program for monitoring rarer pediatric conditions in the United Kingdom and Ireland.¹² Consultant pediatricians are requested to report cases to the BPSU by completing an orange report card sent to them every month. The BPSU malaria study began on January 1, 2006 for 13 months and included all children <16 years diagnosed with malaria through microscopic examination of blood smears and/or malaria antigen detection using commercially available assays. Pediatricians notifying malaria cases to the BPSU were sent a questionnaire requesting details of previous history of malaria, travel history, antimalarial prophylaxis, presenting features, laboratory investigations, concurrent infections, complications, management and outcome.

The Malaria Reference Laboratory (MRL) enhanced surveillance was used as a comparison data source for UK cases in a capture-recapture analysis. The MRL identifies cases from routine reports submitted by UK hospital laboratories to the Health Protection Agency and from positive malaria slides sent to the MRL for species verification. The MRL also receives notifications of malarial deaths from the Office for National Statistics. Cases in the Republic of Ireland are reported by local hospitals to the Heath Protection Surveillance Centre (HPSC).^9,10

The BPSU study did not request patient names but used birth-date, gender, postcode, hospital, diagnosis date, and Plasmodium species to cross-link with the national databases for capture-recapture analysis. The capture-recapture methodology is often applied to epidemiologic studies to estimate the true number of cases when 2 or more independent data sources are available for the same population.^13,14

All data were entered into Microsoft Excel 2003 and analyzed using Stata 9. Estimated cases using capture-recapture analysis and 95% confidence intervals (95% CI) were calculated according to Howitz et al.^13,14 Mid-2006 pediatric population estimates were obtained from the Office for National Statistics, General Register Office for Scotland, Northern Ireland Statistics and Research Agency and Central Statistics Office, Republic of Ireland. Clinical and laboratory features were compared between travelers from the United Kingdom/Ireland and visitors form endemic countries because the former are less likely to be semi-immune to malaria and may, therefore, present with more severe features. Continuous, normally distributed variables are given as means with 95% CI and compared using the Student t test. Proportions are compared by χ² or Fisher exact test, as appropriate. Age and time periods are given as medians and interquartile ranges (IQR) and compared using the Mann-Whitney U test. Risk factors for severe malaria were assessed using logistic regression incorporating variables with a P value of <0.20 in the univariate analysis and then excluding the least significant variable until only those risk factors with P < 0.05 remained. This study was approved by the Leicestershire, Northamptonshire, and Rutland Multicenter Research Ethics Committee (Reference: 05/Q2502/120).

RESULTS

The BPSU orange card return rate from 30,059 cards sent in 2006 was 93.7% (monthly range: 91.5%–93.8%).¹² During the surveillance period, 214 case reports were received, of which 28 were duplicates (n = 26) or reported in error (n = 2). Of the remaining 186 cases, 155 (83%) were from England, 3 (2%) from Scotland, 2 (1%) from Wales, 26 (14%) from Republic of Ireland, and none from Northern Ireland. Fourteen case questionnaires (12 from England and 2 from Ireland) were not completed (8%).

Over the same surveillance period, 226 cases were reported to the MRL, of which 108 of the 148 UK cases (73%) with completed questionnaires had also been reported to MRL. It was not possible to link the 12 BPSU cases without questionnaires to MRL cases using the limited information on BPSU report cards because >100 MRL cases had not been reported to BPSU. However, assuming the same proportion of BPSU cases (73%) would have been reported to MRL, then 9 of the 12 BPSU cases without questionnaires would have been reported to both surveillance systems and the remaining 3 cases to BPSU only. Thus, of the 226 cases reported to MRL, an estimated 109 cases would not have been reported to BPSU (Table 1).

TABLE 1: Number of Cases and Incidence of Imported Childhood Malaria in the United Kingdom and Republic of Ireland

Capture-recapture analysis estimated that there were 309 (95% CI, 289–329) cases of pediatric malaria in the United Kingdom (burden of infection, 2.8/100,000 per year; 95% CI, 2.6–3.0 per 100,000). The sensitivity of BPSU was 52% (160/309 cases), 73% (226/309 cases) for MRL, and 87% (269/309 cases) for the 2 combined. In Ireland, because of fewer cases, the 2 BPSU cases with incomplete questionnaires could be linked to the HPSC dataset. Capture-recapture analysis using 18 cases identified through both data sources, 8 through BPSU and 12 through HPSC, estimated a total of 43 (95% CI, 26–50) cases (burden of infection, 4.6/100,000 per year; 95% CI, 2.8–5.4 per 100,000) (Table 1). The sensitivity of BPSU in Ireland was 60% (26/43 cases), 70% (30/43 cases) for HPSC and 91% (39/43 cases) for the 2 combined.

Epidemiologic Features (n = 290)

Epidemiologic information was available for 172 BPSU cases with complete questionnaires and 118 MRL/HPSC cases not identified through BPSU. The median age at diagnosis was 8.0 years (IQR, 4.6–12.4 years). Imported malaria cases peaked in the summer (146 cases [50%] between July and October), with a smaller peak in January-February, and P. falciparum accounted for 247 of 290 (85%) cases. Of 262 cases where information was available, 235 (90%) were acquired in Africa, mainly West Africa (206 cases, 79%), particularly Nigeria (145 cases, 55%), whereas the rest (27 cases, 10%) were from the Indian Subcontinent. P. falciparum acquired in West Africa was responsible for all cases in Ireland and Scotland. In England, 157 of 258 (61%) cases were diagnosed in London and 90% (142/157) had P. falciparum malaria, while over half of all Plasmodium vivax cases (13/24, 54%) were diagnosed in 3 English regions—West Midlands, Yorkshire, and the North West.

When the 172 BPSU cases with completed questionnaires were compared with the 118 MRL/HPSC cases that were not reported to the BPSU, there was no difference in age at diagnosis, month of diagnosis, country of infection, Plasmodium species or region of diagnosis, suggesting that the BPSU cases were representative. Most hospitals reported only 1 case (48/79 hospitals, 61%), 14 (18%) reported 2, 8 (10%) reported 3 and 9 (11%) reported ≥4 cases. Almost all children (155 cases, 90%) were Black-African and none was Caucasian. None was known to have an underlying immune problem or HIV at the time of infection. The median duration of illness at presentation was 4 days (IQR, 2–6 days) and 34% (59 cases) had a past history of malaria (Table 2). The clinical (Table, Supplemental Digital Content 1, https://links.lww.com/INF/A318) and laboratory features (Table, Supplemental Digital Content 2, https://links.lww.com/INF/A319) of malaria among travelers and visitors are available online.

TABLE 2: Epidemiologic Features and Outcome of Children With Imported Malaria

Malaria Among Travelers (n = 117, 68%)

Most children (95/117 cases, 81%) had been born in the United Kingdom/Ireland, while others had emigrated mainly from West (13/22) and East (4/22) Africa. Almost two-thirds (117 cases, 61%) had not taken any antimalarial prophylaxis and only 9 (8%) had taken appropriate prophylaxis, although in 20 others (17%), the antimalarial regimen taken was not specified. Seventy-nine children (67.5%) had been visiting relatives, while 13 (11%) had been on holiday. Interestingly, 12 of the children with nonsevere P. falciparum malaria had been sent to West Africa to attend boarding school; none had taken antimalarial prophylaxis.

Malaria Among Visitors (n = 55, 32%)

Most visitors were from West Africa (47 cases, 85%). Reasons for visiting included new immigrants (28 cases, 52%), holiday (24 cases, 44%), foreign student (2 cases, 3%), and visiting relatives (1 case, 2%). Visiting children were older (median 9.6 years [IQR, 5.6–12.8 years] vs. 7.3 [4.2–12.1]; P = 0.011), more likely to have had malaria previously (29/55 [53%] vs. 30/117 [26%], χ² = 12.2, P < 0.0001) and developed symptoms later than travelers (median 14 days [IQR, 7–18 days] vs. 7 [4–14]; P < 0.0001). They were also less likely to present with fever (50/55 [91%] vs. 116/117 [99%], χ² = 7.54, P = 0.006) or have severe thrombocytopenia (platelet counts <50 × 10⁹/l; 1/55 [2%] vs. 19/117 [16%], χ² = 7.57, P = 0.006) (Tables 1 and 2, online only).

Plasmodium falciparum (n = 148) and Severe Malaria (n = 46)

Among P. falciparum malaria cases, the median parasite count was 2% (IQR, 0.2%–4%; range, 0.1%–21%). Forty-six children (31%) fulfilled WHO criteria for severe or potentially complicated malaria,² including hyper-parasitemia >5% (n = 30, 20%), convulsions (n = 8, 5%), jaundice (n = 8, 5%), coma (n = 4, 3%), hypoglycemia (blood glucose <2.2 mmol/L, n = 2, 1%), and 1 case each (1%) of respiratory acidosis, spontaneous bleeding and metabolic acidosis. None had severe anemia (<5 g/dL) or renal impairment. Severe malaria was associated with age <5 years (22/46 [48%] vs. 25/102 [25%], χ² = 8.0, P = 0.005), thrombocytopenia (<150 × 10⁹/l) (34/46 [74%] vs. 50/102 [49%], χ² = 8.0, P = 0.005), and diagnosis outside London (31/46 [67%] vs. 39/102 [38%], χ² = 10.8, P = 0.001). In a logistic regression model, thrombocytopenia (OR = 3.9, 95% CI = 1.6–9.2, P = 0.002), age <5 years (OR = 2.9, 95% CI = 1.3–6.8, P = 0.01), and diagnosis outside London (OR = 2.8, 95% CI = 1.3–6.1, P = 0.01) remained independently associated with severe malaria. The adjusted odds ratio increased to 8.1 (95% CI, 2.5–26.7; P = 0.001) when thrombocytopenia was replaced with severe thrombocytopenia in the model. Eleven children (6% of all cases, 7% of P. falciparum cases and 24% of severe P. falciparum malaria cases) were admitted to a pediatric intensive care unit for coma (n = 5), convulsions (n = 3), and 1 case each of hyper-parasitemia, circulatory shock requiring inotropes and cardiac monitoring (Table 2).

Blood cultures were positive in 2 visitors and 1 traveler (Table, Supplemental Digital Content 1, https://links.lww.com/INF/A318) and included Salmonella typhi (2 cases) and Acinetobacter baumannii (1 case). Three other children with convulsions had a lumbar puncture but did not have meningitis.

Treatment, Complications, and Outcome

Thirty-eight children with P. falciparum malaria (26%) were treated with quinine and pyrimethamine/sulfadoxine (Fansidar), while 45 children (30%) were treated with quinine alone and 25 (11%) received oral atovaquone-proguanil (Malarone). The remaining children with P. falciparum malaria were treated with a range of different antimalarial combinations (Table, Supplemental Digital Content 3, https://links.lww.com/INF/A320). The median duration of in-patient stay was 2 days (IQR, 1–4 days; range, 0–9 days) (Table 2). Twenty-eight children (16.3%) were sent home from the Emergency Department with oral antimalarials. One UK-born child (0.6%) with nonsevere P. falciparum malaria who had presented with vomiting and been given oral atovaquone-proguanil returned with 10% parasitemia after treatment completion (Table 2). She was successfully retreated with 7 days of intravenous quinine. No childhood malaria deaths were identified.

DISCUSSION

We have reported the largest national prospective study on imported malaria in children. The burden of malaria in children is higher than that reported in official statistics. Black-African children born in the United Kingdom/Ireland and traveling to West Africa during school holidays without antimalarial prophylaxis remain the main group at risk for imported malaria, although the contribution of visitors diagnosed with malaria has increased.¹⁵ A substantial proportion had severe malaria, some requiring intensive care. Treatment guidelines were followed haphazardly, and many cases had had malaria previously, implying missed opportunities for health education during their previous hospital admission.

The assumptions on which the capture-recapture methodology is based were largely fulfilled in our study. We studied a relatively closed population, the 2 sources were independent (BPSU relies on clinical reporting by pediatricians, while national reference centers receive laboratory reports), only cases fulfilling strict case definitions were included and almost all cases were uniquely identifiable between data sources, although we had to make some assumptions regarding 12 UK cases without a completed BPSU questionnaire.^13,14 Although both data sources are hospital-based, our estimates are likely to reflect the burden of laboratory-proven disease in the population because general practitioners use hospital laboratories for blood tests to diagnose malaria,¹⁵ and are recommended to refer all children with P. falciparum malaria (which accounted for 85% of all cases) to the hospital for further management.¹¹ The level of under-reporting by national reference centers was close to the 25% we had predicted before the study and compares with 20% in Finland, 55% in France, and 59% in the Netherlands.^5,16 In the United Kingdom, the current level of under-reporting (27%) is better than the 41% estimated in 1990.¹⁷

The number of childhood malaria cases reported to MRL has declined from 416 cases in 1996 to 226 in 2006, although the proportion acquired in Africa (88% vs. 90%) and cases due to P. falciparum (82% vs. 85%) have not changed since 1999–2003.⁷ The decrease in nonfalciparum cases since the early 1990s cannot be explained by changes in travel pattern to the Indian subcontinent or Sub-Saharan Africa, which have both continued to increase during the past 10 years,¹⁸ or by differences in population demographics, since 3.6% of the UK population are of Indian, Pakistani, or Bangladeshi origin, while only 0.8% are Black African.¹⁹ Possible reasons include increased travel to urban cities where malaria is declining and increased use of antimalarial prophylaxis among travelers, although we have previously reported lower prophylaxis uptake among children traveling to non-African countries (5% vs. 44%, P < 0.001).⁷ The lower number of cases reported from Northern Ireland may in part be a result of under-reporting—in a recent study, only 29% of adult and pediatric malaria cases in Northern Ireland had been reported to the MRL.²⁰

The finding that a quarter of traveling children had previously had malaria is concerning because it suggests that these children continue to travel without appropriate precautions and reflects missed opportunities by clinicians to educate families on malaria prevention. The BPSU study also identified children sent to malaria-endemic countries for schooling as a new and particularly high-risk group because of their prolonged stay in a malaria-endemic country.²¹ Although UK guidelines currently recommend that long-term travelers should also take antimalarial prophylaxis, data on safety and efficacy of long-term antimalarial use in children is limited, and adherence as well as cost are likely to be problematic.²¹

The clinical presentations in this prospective study were similar to retrospective studies in nonendemic countries.²² In particular, vomiting was common and should prompt clinicians to treat with intravenous antimalarials. The finding that a third of P. falciparum cases fulfilled WHO criteria for severe or potentially complicated malaria may be an overestimate, because nonsevere cases might be less likely to be captured by BPSU methodology. However, the higher proportion with hyperparasitemia compared with 5% to 10% in previous studies,²² and the higher proportion of severe malaria cases outside London where malaria is less common, both suggest a delay in diagnosis²² which, together with young age and nonimmunity to malaria, have previously been associated with more severe disease.²³ Thrombocytopenia was also associated with severe malaria, which has been reported in some pediatric malaria studies in endemic settings.^24,25 In our study, almost all children with life-threatening P. falciparum malaria had neurologic involvement (coma, convulsions or reduced consciousness) rather than severe anemia or respiratory distress. The predominance of neurologic presentations of severe malaria is more typical of settings outside Africa.²⁶

Unlike previous retrospective, single-center case series,²² the BPSU study is national and found that, of the hospitals reporting a case of malaria, >90% reported 3 or fewer cases over the 13-month surveillance period. This finding suggests that most hospitals in the United Kingdom/Ireland have limited experience in diagnosing and treating childhood malaria. Of particular concern was the wide variation in (and often inadequate) treatment. One explanation might be publications dissuading UK pediatricians from using oral quinine in children because of its bitter taste,²⁷ which contradict current UK guidance.¹¹ Another concerning finding was the increasing trend towards outpatient treatment of seemingly uncomplicated P. falciparum malaria with oral atovaquone-proguanil (Malarone). Although this antimalarial combination is safe and effective if taken correctly, vomiting (up to 44%) is such a common side-effect that it may not be suitable for outpatient treatment in children.^28,29 In addition, assessing the severity of malaria in children is difficult as they can frequently deteriorate despite adequate treatment, and assumptions that children from African families are partially immune and can, therefore, be treated as outpatients are not based on any evidence.

Although it is reassuring that there were no long-term complications or deaths caused by malaria, this study has raised several areas of concern. Almost a third of children with imported malaria were <5 years, and half of them presented with severe malaria. Children, especially those traveling to visit the countries from which their families migrated previously, continue to travel to malaria-endemic countries without appropriate prophylaxis. General practitioners caring for families who originally migrated from Africa or south Asia need to target health messages on malaria to them. Additionally, malaria remains a relatively rare diagnosis in most British and Irish hospitals. Therefore, the diagnosis may potentially be delayed, the severity of infection not appreciated and/or appropriate management not initiated. This may particularly be true for hospitals outside London where children were more than twice as likely to present with severe malaria in our study. It is, therefore, imperative that national guidelines are precise in their recommendations and smaller hospitals have access to specialist pediatric infectious disease units for advice and assistance in managing complicated cases.

ACKNOWLEDGMENTS

The authors acknowledge BPSU, supported by the Department of Health, for facilitating the data collection and the reporting clinicians, particularly those who completed the questionnaires. The authors thank all members of the RCPCH who reported cases of imported malaria and completed the questionnaires, Valerie Smith and Marie Blaze for providing the MRL data, and Patricia Garvey and Paul McKeown at the Health Protection Surveillance Centre (HPSC) in the Republic of Ireland for linking the Irish cases identified through the BPSU with HPSC notifications.