Please know that any use or reproduction of content ... - Blog Staff

(1)

Please know that any use or reproduction

of content must systematically and clearly

state the following copyright:

(2)

M A J O R A R T I C L E

H5N1 Avian In ﬂ uenza in Children

Ahmet Faik Oner,¹Nazim Dogan,²Viktor Gasimov,³Wiku Adisasmito,⁴Richard Coker,⁵Paul K. S. Chan,⁷Nelson Lee,⁷ Owen Tsang,⁸Wanna Hanshaoworakul,⁹Mukhtiar Zaman,¹⁰Ebun Bamgboye,¹¹Anna Swenson,¹²Stephen Toovey,⁶ and Nancy A. Dreyer¹²

1Yuzuncu Yil University, Van, and²Ataturk University Medical School, Erzurum, Turkey;³Azerbaijan Ministry of Health, Baku;⁴University of Indonesia, Depok;⁵London School of Hygiene and Tropical Medicine, and⁶Royal Free and University College Medical School, Department of Infection and Immunity, Academic Centre for Travel Medicine and Vaccines, London, United Kingdom;⁷Faculty of Medicine, The Chinese University of Hong Kong, and⁸Hospital Authority Infectious Disease Centre at Princess Margaret Hospital, Hong Kong, SAR;⁹Ministry of Public Health, Nonthaburi, Thailand;

10Khyber Teaching Hospital, Peshawar, Pakistan;¹¹St Nicholas Hospital, Lagos, Nigeria; and¹²Outcome Sciences, Inc., Cambridge, Massachusetts

Background. Avian inﬂuenza continues to pose a threat to humans and maintains the potential for greater transmissibility. Understanding the clinical presentation and prognosis in children will help guide effective diagnosis and treatment.

Methods. A global patient registry was created to enable systematic collection of clinical, exposure, treatment, and outcomes data on conﬁrmed cases of H5N1. Bivariate and multivariate statistical tools were used to describe clinical presentation and evaluate factors prognostic of survival.

Results. Data were available from 13 countries on 193 children <18 years who were conﬁrmed as having been infected with H5N1; 35.2% of cases were from Egypt. The case fatality rate (CFR) for children was 48.7%, with Egypt having a very low pediatric CFR. Overall, children aged≤5 years had the lowest CFR and were brought to hospitals more quickly and treated sooner than older children. Children who presented for medical care with a complaint of rhinorrhea had a 76% reduction in the likelihood of death compared with those who presented without rhinorrhea, even after statistical adjustment for age, having been infected in Egypt, and oseltamivir treatment (P= .02). Delayed initiation of treatment with oseltamivir increases the likelihood of death, with an overall 75% increase in the adjusted odds ratio for death for each day of delay.

Conclusions. The presence of rhinorrhea appears to indicate a better prognosis for children with H5N1, with most patients surviving regardless of age, country, or treatment. For individuals treated with oseltamivir, early initiation of treatment substantially enhances the chance of survival.

Although much of the attention on influenza has di- minished, especially since the relatively mild pandemic of H1N1 swineflu, H5N1 avian influenza continues to occur, with human cases continuing to be reported in 2011 [1,2]. A large and ineradicable avian reservoir for this infection means that it may reemerge as an impor- tant threat to human health in many countries [3], with a number of clades of possibly differing virulence circulating in different regions [4]. This article describes the clinical presentation of H5N1 avian influenza, including

identification of prognostic factors and treatment effectiveness, for children infected with laboratory-confirmed influenza A (H5N1).

METHODS

This investigation utilized the global avian influenza registry, which includes 391 cases of laboratory- confirmed influenza A (H5N1). Using standard defi- nitions and data collection procedures, information was gathered from medical records; clinical and field investigations, including government sources; and published case reports. Information about presenting symptoms, treatments, and survival was sought. The registry methods are described in full elsewhere [5].

Cases were recorded as having occurred from 1997 through 2010. The 18 earliest cases were from the ini- tial 1997 outbreak, before World Health Organization

Received 8 November 2011; accepted 23 February 2012; electronically published 15 March 2012.

Correspondence: Dr Nancy A. Dreyer, Outcome Sciences, 201 Broadway, Cam- bridge, MA 02139 ([email protected]).

Clinical Infectious Diseases 2012;55(1):26–32

[email protected].

DOI: 10.1093/cid/cis295

by guest on March 18, 2013http://cid.oxfordjournals.org/Downloaded from

(3)

(WHO)–certified laboratory confirmation was available. Of the remaining 373 cases, 358 (96%) had laboratory confirmation from a WHO-accredited laboratory and 15 (4%) were confirmed by a local laboratory.

Nearly half of the cases (193 of 391) were <18 years at the time of diagnosis. Pediatric cases were recorded from 11 countries: Azerbaijan (5), Bangladesh (1), Cambodia (3), China (6), Egypt (68), Hong Kong (11), Indonesia (59), Laos (1), Thailand (13), Turkey (12), and Vietnam (14). Pediatric cases are further categorized as aged 0–5 years (n = 91, including a single patient aged <1 year), 6–11 years (n = 46), and 12–17 years (n = 56).

Statistical Methods

Differences in categorical variables by age group were exam- ined usingχ²or Fisher exact tests. Differences in continuous variables by age group were compared using the nonpara- metric Wilcoxon rank sum and Kruskal-Wallis tests because the data were not normally distributed. A P value <.05 was considered statistically significant. A Bonferroni adjustment was used to account for multiple comparisons between the various age groups. The overall αlevel of .05 was divided by the number of comparison to give theαlevel considered statistically significant for multiple comparisons. Relative risks and associated 95% confidence intervals (CIs) are also presented. A multivariate logistic regression approach was used to examine the odds of death for individuals with and without rhinorrhea while controlling for age, country (Egypt versus others), and oseltamivir treatment [6,7]. The small number of cases did not allow for addition of potential confounders other than age, country, and oseltamivir treatment. Two

models were used: one included all cases with information recorded regarding the presence or absence of rhinorrhea (n = 100), and the other included only individuals who had information about both the presence or absence of rhinorrhea and who were treated with oseltamivir (n = 44).

RESULTS

Table 1 shows the distribution of cases by age group and country and the corresponding case fatality rates (CFRs) for all age groups. The CFR for pediatric cases was 48.7% in con- trast to the overall CFR of 57.5%, with substantial variability by age and country. Young children were more likely to survive than older children and adults, with children aged ≤5 years showing a markedly lower CFR (28%) than older children (P< .01 for all comparisons). The CFR for those aged 6–11 years was also lower than that of those aged 12–17 years (P= .003).

Children aged ≤5 years were brought for medical attention, hospitalized, and treated with antiviral therapy earlier than older children, with a median of 3 days from symptom onset to start of treatment in the 0–5-year age group versus 7 days for older groups (P= .01; Table2). The time from symptom onset to antiviral treatment was similar for those aged 6–11 years and those aged 12–17 years, despite the much higher mortality rate in the 12–17-year age group. The median of 9 days from symptom onset to death was virtually identical for all ages.

There were some differences by age group in symptoms reported atﬁrst presentation for medical care (Table3). Young children (≤5 years) reported rhinorrhea more frequently, and

Table 1. Case Fatality Rate by Country and Age Group

Country

Pediatric Cases

All Ages 0–5 Years 6–11 Years 12–17 Years All Pediatric Cases

Azerbaijan NA 0/1 (0.0%) 2/4 (50.0%) 2/5 (40.0%) 6/9 (66.7%)

Bangladesh 0/1 (0.0%) NA NA 0/1 (0.0%) 0/1 (0.0%)

Cambodia 1/1 (100.0%) 1/1 (100.0%) 1/1 (100.0%) 3/3 (100.0%) 6/6 (100.0%)

China NA 2/6 (33.3%) NA 2/6 (33.3%) 16/28 (57.1%)

Egypt 2/47 (4.3%) 2/11 (18.2%) 5/10 (50.0%) 9/68 (13.2%) 34/106 (32.1%)

Hong Kong 1/8 (12.5%) 0/1 (0.0%) 1/2 (50.0%) 2/11 (18.2%) 6/18 (33.3%)

Indonesia 16/21 (76.2%) 11/14 (78.6%) 22/24 (91.7%) 49/59 (83.1%) 107/124 (86.3%)

Laos NA NA 1/1 (100.0%) 1/1 (100.0%) 1/1 (100.0%)

Nigeria NA NA NA NA 1/1 (100.0%)

Pakistan NA NA NA NA 1/4 (25.0%)

Thailand 2/4 (50.0%) 5/6 (83.3%) 3/3 (100.0%) 10/13 (76.9%) 17/25 (68.0%)

Turkey 0/6 (0.0%) 1/3 (33.3%) 3/3 (100.0%) 4/12 (33.3%) 4/13 (30.8%)

Vietnam 3/3 (100.0%) 2/3 (66.7%) 7/8 (87.5%) 12/14 (85.7%) 26/55 (47.3%)

Total 25/91 (27.5%) 24/46 (52.2%) 45/56 (80.4%) 94/193 (48.7%) 225/391 (57.5%)

Abbreviation: NA, not applicable.

(4)

headache and myalgia less frequently, than older children and adults. Headache was a much more frequent complaint on presentation for those aged 12–17 years than for all other age groups (P< .01). Bleeding gums were reported more frequently for those aged 6–17, though these ﬁndings are based on especially small numbers.

With regard to whether any particular signs, symptoms, or tests carry prognostic value for survival from avian inﬂuenza during theﬁrst 24 hours of hospital admission, children who died were more likely to have had decreased leukocyte, lymphocyte, and platelet counts and to have had elevated alanine aminotransferase (ALT), aspartate aminotransferase (AST), Table 2. Time Course of Illness From Onset of Symptoms to Antiviral Treatment and Death, by Age Group (N = 391)

Age (years) Case Fatality Rate

Median Days From Symptom Onset to:

Presentation for

Medical Care Hospitalization^a

Antiviral Treatment^a

(n = 242 Treated) Death (n = 225 Fatal) No.^b Median (min–max) No.^b Median (min–max) No.^b Median (min–max) No.^b Median (min-max)

0–5 25/91 (27.5%) 31 0 (0–20) 85 3 (0–20) 33 3 (0–11) 24 9 (3–17)

6–11 24/46 (52.2%) 28 2 (0–9) 44 4.5 (0–25) 20 7 (2–21) 22 10.5 (2–32)

12–17 45/56 (80.4%) 34 1 (0–10) 55 5.0 (0–13) 24 7 (1–14) 44 9 (3–31)

All ages 225/391 (57.5%) 193 1 (0–20) 374 5.0 (0–25) 170 6 (0–23) 215 9 (2–32)

aMedians are significantly different between age groups.

bTotal no. differs according to completeness of timing data.

Table 3. Frequency of First Symptoms Reported at Presentation for Medical Care, by Age

Symptom No.

Age (years)

PValue^a

0–5 6–11 12–17 ≥18

Fever 303 70/73 (95.9%) 34/37 (91.9%) 43/47 (91.5%) 138/146 (94.5%) .71

Unexplained respiratory illness^b 275 48/56 (85.7%) 30/34 (88.2%) 38/47 (80.9%) 99/138 (71.7%) .06

Tachypnea 182 9/28 (32.1%) 9/27 (33.3%) 15/37 (40.5%) 41/90 (45.6%) .51

Abnormal breath sounds 113 6/21 (28.6%) 5/18 (27.8%) 6/19 (31.6%) 10/55 (18.2%) .57

Sore throat/pharyngitis 169 13/29 (44.8%) 14/28 (50.0%) 19/34 (55.9%) 28/78 (35.9%) .22

Cyanosis 99 2/19 (10.5%) 1/15 (6.7%) 1/18 (5.6%) 3/47 (6.4%) .93

Excessive sputum production 132 2/20 (10.0%) 3/22 (13.6%) 2/27 (7.4%) 15/63 (23.8%) .19

Rhinorrhea 173 24/45 (53.3%)^c,d 6/24 (25.0%) 5/31 (16.1%)^e 8/73 (11.0%)^e <.01

Diarrhea 180 7/30 (23.3%) 5/26 (19.2%) 8/33 (24.2%) 20/91 (22.0%) .97

Abdominal pain 136 5/26 (19.2%) 5/19 (26.3%) 8/21 (38.1%) 13/70 (18.6%) .28

Vomiting 158 9/30 (30.0%) 7/22 (31.8%) 8/26 (30.8%) 11/80 (13.8%) .08

Headache 151 5/27 (18.5%)^c 4/17 (23.5%)^c 18/26 (69.2%)^d,e,f 28/81 (34.6%)^c <.01

Fatigue or malaise 121 4/21 (19.1%) 4/15 (26.7%) 7/24 (29.2%) 21/61 (34.4%) .60

Myalgia 135 2/23 (8.7%)^d 6/21 (28.6%) 4/23 (17.4%) 28/68 (41.2%)^e .01

Neurologic involvement 106 1/21 (4.8%) 2/18 (11.1%) 0/16 (0%) 3/51 (5.9%) .57

Psychiatric 91 0/19 (0%) 2/14 (14.3%) 1/14 (7.1%) 4/44 (9.1%) .46

Bleeding gums and/or nose 133 0/23 (0%) 2/23 (8.7%) 3/25 (12.0%)^d 0/62 (0.0%)^c .02

Enlarged liver 93 0/18 (0%) 1/15 (6.7%) 1/16 (6.3%) 1/44 (2.3%) .62

Conjunctivitis 134 1/23 (4.4%) 1/22 (4.6%) 0/23 (0%) 1/66 (1.5%) .64

aOverallPvalue. Comparisons between age groups are considered significant atP<.01. Reduction inαlevel is due to application of a Bonferroni correction for multiple comparisons.

bUnexplained respiratory illness is defined as including cough, shortness of breath, or difficulty breathing.

cSignificantly different from children aged 12–17.

dSignificantly different from those aged≥18.

eSignificantly different from children aged 0–5.

fSignificantly different from children aged 6–11.

(5)

creatinine, and hematocrit values; children who survived were more likely to have had lower hemoglobin levels (Table 4).

Creatine kinase and lactate dehydrogenase levels at presentation for medical care did not show statistically significant relationships with likelihood of survival, but small numbers counsel caution in interpretation. Examination of the many clinical signs and symptoms reported at presentation for medical care (Table 5)demonstrate that the presence of rhinorrhea is associated with a decreased risk of death, especially for children aged ≤5 years (relative risk of death for children with rhinorrhea, 0.13; 95% CI, .03–.53). Also, the nonspecific symptom characterized as “unexplained respiratory illness with cough, shortness of breath, or difficulty breathing” appears to carry some prognostic value, showing a decreased

risk of death, particularly in children ≤5 years. Similarly, 4 other symptoms (diarrhea, headache, fatigue or malaise, and myalgia) also showed some weak but consistent evidence that they may be associated with a better prognosis in children.

Other symptoms, including fever, excessive sputum production, sore throat, vomiting, and tachypnea, did not show any consistent relation to the likelihood of death.

Although small numbers preclude examination of speciﬁc age groups, we investigated whether the improved survival among children who presented with rhinorrhea might have been due to their having received medical attention sooner or having been treated more quickly than children who did not present with these symptoms. For children who presented with rhinorrhea, the time to presentation for medical care was Table 4. Survival by Median Laboratory Values Within 24 Hours of Hospital Admission

Laboratory Parameter No.

Pediatric Cases (Age <18 years)

PValue

Fatal Cases Nonfatal Cases

Leukocyte count: median count, per mm³ 61 2800 (400–18 300) 5100 (2000–15 900) <.01 Lymphocyte count: median count, per mm³ 21 800 (250–1700) 2028 (1040–4256) <.01 Platelet count: median count, per mm³ 52 123 000 (35 000–314 000) 188 000 (122 500–528 000) <.01

Creatine kinase: median count, U/L 9 1430 (52–3429) 297 (82–1396) .54

Alanine aminotransferase: median, U/L 31 60 (8–8750) 22 (11–299) .03

Aspartate aminotransferase: median, U/L 31 263 (20–3230) 53 (16–107) <.01

Lactate dehydrogenase: median, U/L 11 1606 (604–4032) 1518 (420–4478) .93

Creatinine: median,μmol/L 22 0.70 (0.16–1.04) 0.38 (0.20–0.53) .02

Urea nitrogen: median, mg/dL 16 21 (10–58) 15 (7–22) .19

Hemoglobin 47 13 (10–37) 11 (10–14) .04

Hematocrit 35 37 (4–51) 33 (27–38) .01

Table 5. Relative Risk of Death on Presentation for Medical Care, by Symptom and Age Group

Symptom No.

Age (years)

0–5 6–11 12–17 ≥18

Fever 303 0.81 (.16, 4.22) 0.88 (.38, 2.06) 1.72 (.64, 4.62) 0.95 (.63, 1.43)

Unexplained respiratory illness with cough, shortness of breath, or difficulty breathing

275 0.47 (.27, .82) 2.80 (.50, 15.53) 0.84 (.73, 0.97) 0.87 (.72, 1.06)

Tachypnea 182 0.53 (.20, 1.41) 1.56 (.87, 2.78) 1.29 (1.03, 1.62) 0.92 (.73, 1.16)

Abnormal breath sounds 113 0.83 (.34, 2.05) 1.73 (.83, 3.61) 1.18 (.94, 1.49) 1.16 (.80, 1.68) Sore throat/pharyngitis 169 0.49 (.20, 1.21) 1.25 (.71, 2.19) 0.68 (.47, 0.98) 0.89 (.66, 1.20) Excessive sputum production 132 0.82 (.20, 3.43) 1.06 (.44, 2.52) 1.14 (.98, 1.31) 0.80 (.51, 1.25)

Rhinorrhea 173 0.13 (.03, .53) 0.75 (.32, 1.78) 0.68 (.33, 1.41) 0.49 (.20, 1.20)

Diarrhea 180 0.88 (.43, 1.78) 0.60 (.20, 1.83) 1.20 (1.00, 1.41) 1.25 (1.02, 1.54)

Abdominal pain 136 1.29 (.74, 2.24) 1.05 (.45, 2.45) 1.14 (.77, 1.69) 1.22 (.99, 1.52)

Vomiting 158 0.67 (.30, 1.47) 1.19 (.64, 2.22) 1.05 (.75, 1.47) 1.09 (.80, 1.48)

Headache 151 0.68 (.22, 2.09) 0.72 (.25, 2.05) 0.89 (.62, 1.27) 1.15 (.90, 1.46)

Fatigue or malaise 121 0.43 (.07, 2.43) 0.46 (.08, 2.72) 0.81 (.49, 1.33) 0.98 (.76, 1.26)

Myalgia 135 NA 0.50 (.15, 1.64) 0.59 (.22, 1.61) 0.80 (.59, 1.10)

All data are presented as relative risk of death (95% confidence interval).

Abbreviation: NA, not applicable.

(6)

not very different (median, 1 versus 3 days) for children who survived and children who died (n = 9 cases with information on time from symptom onset to presentation for medical care;

P= .45). However, children with rhinorrhea who survived were treated more quickly with antiviral therapy (median, 2.5 days from symptom onset to start of antiviral treatment) than those who died (median, 10 days; n = 17 cases with information on time from symptom onset to treatment with anti- virals; P< .01). Using multivariate modeling, children who presented with rhinorrhea had a 76% reduction in the risk of death (odds ratio [OR], 0.24; 95% CI, .08–.77) when simul- taneously controlling for oseltamivir treatment, country, and age group (Table6). Looking only at children prescribed oseltamivir, however, the survival beneﬁts associated with rhinorrhea were still remarkable but did not achieve statistical signiﬁcance (OR, 0.38; 95% CI, .03–5.55). Looking at time to initiation of treatment with oseltamivir, there was an increased odds of death for each day of delay (OR, 1.75; 95% CI, 1.17–2.61) when controlling for age, rhinorrhea, and country (Egypt).

DISCUSSION

The clinical presentation of avian inﬂuenza in children differs in some meaningful ways from that in adults. Unlike some earlier reports that characterized H5N1 infections as carrying a higher mortality rate in children, the lower mortality rate in children aged ≤5 years in this large case series is quite strik- ing, especially because the survival beneﬁt is evident even when type of presenting symptom, antiviral treatment, time to treatment initiation, and country are taken into account [8,9].

Of note, in a small series from Vietnam (N = 36), Kawachi et al reported that children aged 6 years were at higher risk of fulminant disease with acute respiratory distress syndrome (ARDS) than younger children [10]. It might be that the less mature immune systems of younger children mount a response that is less harmful to their hosts.

The presence of rhinorrhea at presentation is more common in children aged≤5 years and appears to be associated with a markedly decreased risk of death in this age group.

This symptom appears to have prognostic value, even after ac- counting for the most obvious explanations for the increased survival seen in children with this presentation, such as having been seen sooner, or having received theirﬁrst dose of antiviral therapy early in the course of their illness, or coming from Egypt, which has a lower CFR than other countries [6, 7]. In this registry, 35% (68 of 193) of the patients <18 years of age were from Egypt, as were more than half (52%

[47 of 91]) of the patients aged≤5 years. However, the protec- tive effect of rhinorrhea was still evident in the 0–5-year age group when tested both by excluding Egyptian cases from analysis and by using statistical analyses to control simul- taneously the effects of country (see model 1 in Table6, the only model that had enough cases to permit inclusion of Egypt as an additional covariate). Thus, there remains an in- triguing difference in the frequency of rhinorrhea as a presenting symptom and its apparent prognostic value, which declines with increasing age. One might speculate that this represents primary inoculation of the virus into the upper rather than the lower airways or perhaps a less injurious pathway to immune activation [11]. Regardless of the expla- nation, there remained beneﬁt from early use of antiviral Table 6. Rhinorrhea and Odds Ratios for Death Among Children: 2 Models

No. Unadjusted OR (95% CI) No. Adjusted OR (95% CI) PValue All cases with information on rhinorrhea (model 1)

Rhinorrhea noted at presentation for medical care (yes/no)

100 0.11 (.04, .28) 100 0.24 (.08, .77) .02

Oseltamivir treatment (yes/no) 100 0.94 (.43, 2.09) 100 1.41 (.50, 3.96) .51

Age, years 100 100

0–5 0.10 (.03, .30) 0.15 (.04, .52) <.01

6–11 0.32 (.09, 1.14) 0.32 (.08, 1.24) .10

12–17 Ref Ref

Egypt (versus all other countries) 100 0.04 (.01, .35) 100 0.16 (.02, 1.78) .14

Oseltamivir-treated cases with information on rhinorrhea (model 2)

Rhinorrhea at presentation for medical care 56 0.15 (.04, .52) 44 0.38 (.03, 5.55) .48 Days from symptom onset to oseltamivir treatment

(delay in starting treatment)

44 1.85 (1.30, 2.64) 44 1.75 (1.17, 2.61) .01

Age, years 56 1.21 (1.05, 1.38) 44 1.38 (1.02, 1.85) .04

Abbreviations: CI, confidence interval; OR, odds ratio; Ref, reference group.

(7)

therapy. Asymptomatic and mild cases of H5N1 have been previously reported, and one might speculate that children presenting with rhinorrhea represent a subset of patients with generally milder illness [12].

The recording of other symptoms may not be as reliable as rhinorrhea, which can be observed by the clinician. For example, the paucity of myalgia as a presenting symptom in children aged 0–5 years may simply reﬂect the inability of young children or their parents to accurately report this symptom, rather than a true difference in symptom occurrence by age.

With respect to laboratory values, our ﬁndings might possess some clinical utility, indicating severity to clinicians and suggesting the need for aggressive antiviral and supportive therapy. Our ﬁndings are consistent with others. Grose and Chokephaibulkit reported leukopenia and thrombocytopenia in Vietnamese and Thai children suffering from H5N1 infection [9]. Examining Vietnamese pediatric H5N1 cases with ARDS, Kawachi et al reported leukopenia and thrombocytopenia to be predictors of fulminant disease with ARDS [10].

Furuya et al, in their meta-analysis of published pediatric H5N1 case series, found thrombocytopenia and leukopenia to be significantly associated with mortality [13]. Thesefindings with respect to leukopenia were confirmed in pediatric registry patients who died; registry patients who suffered a fatal outcome also demonstrated a lower median thrombocyte count within the first 24 hours of admission, if not always a true thrombocytopenia. This might suggest that a falling or relatively low thrombocyte count could be a very early indication of severe disease and poor outcome. For a variety of reasons, including consumption, peripheral sequestration, and myelo- suppression, leukopenia and thrombocytopenia are not infre- quently seen in other very severe infections.

Furuya et al also reported that a raised peripheral blood AST level trended toward significant association with mortality [13]; that association was confirmed in this larger patient registry dataset. Further, our data also revealed a significant association between a raised ALT level and mortality in children. These findings most likely represent widespread cellular and, in particular, hepatocyte insult consequent upon the severe inflammatory processes that accompany advanced H5N1 infection [14].

The raised hematocrit seen in fatal cases could be due to a degree of hemoconcentration possibly associated with dehy- dration and perhaps with vascular injury occasioned by the severe inﬂammatory processes known to accompany H5N1 infection. The lower mean hemoglobin levels seen in survivors could accord with the hemoconcentration seen in fatal cases.

With respect to antiviral treatment, by far the most frequently used antiviral therapy in registry cases was oseltamivir (91% of all antiviral therapies), and these data continue to show the beneﬁts of early treatment with oseltamivir. We also

tested the generalizability of this ﬁnding by looking at data from the 2 countries with the most cases, Egypt and Indonesia, and comparing antiviral effectiveness data between them and the rest of the countries in the registry. Because country of infection (and illness) is a variable that likely reﬂects the relative virulence of the circulating strain as well the sophistication and accessibility of the local healthcare system, we controlled by country, rather than by viral strain or clade, even though the latter might well indicate relative virulence [15,16].

The median number of days from symptom onset to antiviral treatment was markedly shorter for surviving cases compared with fatal cases for all pediatric age groups, regardless of country. Thesefindings confirm once again the importance of early initiation of effective antiviral therapy in human H5N1 infection [5], specifically in the pediatric setting.

Overall, the data from this registry show that children aged

≤5 years are more likely to survive infection with H5N1; they come to medical attention more quickly than adults and receive antiviral treatment more quickly than their older counterparts. The results also support the prognostic value of some laboratoryﬁndings on presentation as well the value of antiviral therapy, especially when initiated early in the course of infection.

Many limitations to the analysis and interpretation of these data remain, some due to data that were not available or tests that were either not performed or not recorded, and others to the relatively small number of cases available for analyses.

However, it should be recognized that this represents the largest collection of aggregated clinical data on avian inﬂuenza in humans, all of whom have laboratory evidence conﬁrming infection with H5N1. These data may provide insights to clinicians treating pediatric patients as well as meaningful clues to immune response that can be harnessed for more effective treatment of this highly lethal disease.

Notes

Financial support. This work was supported by a contract to Outcome Sciences, Inc, from F. Hoffmann-La Roche. The sponsor pro- vided scientiﬁc collaboration and had rights to nonbinding review of manuscripts but did not have the right to decide whether papers should be submitted for publication, to choose authors, or to approve the wording of any manuscripts.

Potential conﬂicts of interest. W. A., M. Z., A. F. O., E. B., and N. D. received modest support to facilitate data collection and review. R. C. has received funding from F. Hoffmann-La Roche, the manu- facturer of oseltamivir. P. K. S. C. and N. L. received funding support from F. Hoffmann-La Roche for an investigator-initiated study. N. A. D. and A. S. are employed by Outcome Sciences, Inc, a private company that specializes in patient registries and that received funding from F. Hoffmann-La Roche to create and conduct the registry study. S. T. is a former employee and a paid consultant to F. Hoffmann-La Roche and has been reimbursed by a number of inﬂuenza vaccine manufacturers.

All authors have submitted the ICMJE Form for Disclosure of Potential Conﬂicts of Interest. Conﬂicts that the editors consider relevant to the content of the manuscript have been disclosed.

(8)

References

1. Jhung MA, Davidson H, McIntyre A, et al. Preliminary results of 2009 pandemic inﬂuenza surveillance in the United States using the Aggre- gate Hospitalization and Death Reporting Activity. Inﬂuenza Other Respi Viruses2011; 5:321–7.

2. World Health Organization. Cumulative number of conﬁrmed human cases of avian inﬂuenza A/H5N1 reported to WHO. 9 August2011.

Available at: http://www.who.int/inﬂuenza/human_animal_interface/

country/cases_table_2011_08_09/en/index.html. Accessed 27 September 2011.

3. Suarez DL. Avian inﬂuenza: our current understanding. Anim Health Res Rev2010; 11:19–33.

4. Neumann G, Chen H, Gao GF, Shu Y, Kawaoka Y. H5N1 inﬂuenza viruses: outbreaks and biological properties. Cell Res2010; 20:51–61.

Epub 3 November 2009. Review.

5. Adisasmito W, Chan PKS, Lee N, et al. Effectiveness of antiviral treatment in human inﬂuenza A(H5N1) infections: analysis of a global patient registry. J Infect Dis2010; 202:1154–60.

6. Dudley JP. Age-speciﬁc infection and death rates for human A(H5N1) avian inﬂuenza in Egypt. Euro Surveill2009; 14:19198.

7. Fasina FO, Ifende VI, Ajibade AA. Avian inﬂuenza A(H5N1) in humans: lessons from Egypt. Euro Surveill2010; 15:19473.

8. Areechokchai D, Jiraphongsa C, Laosiritaworn Y, et al. Investigation of avian inﬂuenza (H5N1) outbreak in humans—Thailand, 2004.

MMWR Morb Mortal Wkly Rep2006; 55 Suppl 1:3–6.

9. Grose C, Chokephaibulkit K. Avian inﬂuenza virus infection of children in Vietnam and Thailand. Pediatr Infect Dis J 2004; 23:

793–4.

10. Kawachi S, Luong ST, Shigematsu M, et al. Risk parameters of fulminant acute respiratory distress syndrome and avian inﬂuenza (H5N1) infection in Vietnamese children. J Infect Dis2009; 200:510–5.

11. Oslund KL, Baumgarth N. Inﬂuenza-induced innate immunity: regu- lators of viral replication, respiratory tract pathology and adaptive immunity. Future Virol2011; 6:951–62.

12. Zaman M, Ashraf S, Dreyer NA, Toovey S. Human infection with avian inﬂuenza virus, Pakistan, 2007. Emerg Infect Dis 2011; 17:

1056–9.

13. Furuya H, Kawachi S, Shigematsu M, Suzuki K, Watanabe T. Clinical factors associated with severity in hospitalized children infected with avian inﬂuenza (H5N1). Environ Health Prev Med2011; 16:64–8.

14. de Jong MD, Simmons CP, Thanh TT, et al. Fatal outcome of human inﬂuenza A (H5N1) is associated with high viral load and hypercyto- kinemia. Nat Med2006; 12:1203–7.

15. Le QM, Ito M, Muramoto Y, et al. Pathogenicity of highly pathogenic avian H5N1 inﬂuenza A viruses isolated from humans between 2003 and 2008 in northern Vietnam. J Gen Virol2010; 91(Pt 10):

2485–90.

16. Govorkova EA, Ilyushina NA, Boltz DA, Douglas A, Yilmaz N, Webster RG. Efﬁcacy of oseltamivir therapy in ferrets inoculated with different clades of H5N1 inﬂuenza virus. Antimicrob Agents Che- mother2007; 51:1414–24.