Femmes enceintes et grippe A : vaccins et Tamiflu à gogo et à l’aveugle, selon des articles du Lancet

Dans le numéro du 8 août de la revue The Lancet, on peut lire les conclusions d’une étude financée par le Center for Disease Control and Prevention des Etats-Unis (CDC) : Denise J Jamieson et al. « H1N1 2009 influenza virus infection during pregnancy in the USA » (Etats-Unis: Gripe A H1N1 pendant la grossesse ; Lancet. 2009 Aug 8;374(9688):451-458).

Selon les auteurs, les femmes enceintes pourraient être plus à risque de faire des complications (pneumonie, détresse respiratoire…) suite à une grippe A. Le taux d’admission à l’hôpital serait plus élevé que dans la population générale. Jamieson et al. recommandent un traitement systématique et précoce par antiviraux (oseltamivir – Tamiflu, zanamivir – Relenza), dès le début des symptômes, considérant que, même si on ne sait pas grand-chose des effets de ces médicaments sur le fœtus, les bénéfices devraient l’emporter sur les risques… Donc à l’aveugle. Un commentaire par Punam Mangtani et al. paru dans le même numéro du Lancet conforte ces thèses et insiste aussi sur la vaccination des femmes enceintes, qui leur paraît nécessaire. Vaccination qui devrait comprendre le vaccin contre la grippe saisonnière, et ce quel que soit le trimestre de la grossesse.

Est-ce bien raisonnable ? Il m’a paru intéressant d’entrer dans les détails, pour que chacun puisse juger par lui-même et se poser la question de la solidité des preuves scientifiques sur lesquelles se basent de telles recommandations, surtout dans un cas aussi sensible que celui de vaccins et de médicaments administrés pendant la grossesse.

D’autant que les conclusions de Jamieson et al. – et même leur estimation des pourcentages – se basent sur un très faible nombre de cas : 31 cas confirmés et 3 cas probables de grippe A chez des femmes enceintes, recensés entre le 15 avril et le 18 mai, et dont 32% ont été hospitalisées. Sur les 45 décès signalés au CDC entre le 15 avril et le 16 juin, 6 étaient des femmes enceintes dont la santé était relativement bonne avant la contamination par une grippe A qui s’est compliquée par des pneumonies et des insuffisances respiratoires. Ces femmes ont toutes été traitées par des antiviraux, commencés 6 à 15 jours après le début des symptômes. Les auteurs considèrent non pas que le Tamiflu et/ou le Relenza ont été inefficaces, mais que le traitement a été trop tardif. Sur ces 6 femmes enceintes, 5 avaient des fœtus viables, et une césarienne d’urgence a été pratiquée. Les nourrissons ne présentent aucun signe de grippe, et, mis à part celui né à 27 semaines, ils ont rapidement pu quitter l’hôpital.

Les auteurs sont soit des employés du CDC soit des départements santé et/ou santé publique de plusieurs Etats américains ou de structures fédérales impliquées dans la santé publique; ils forment le « Novel Influenza A (H1N1) Pregnancy Working Group ». Aucun conflit d’intérêt n’est déclaré.

Un commentaire signé par trois auteurs européens – Punam Mangtani (épidémiologiste britannique), Tippi K Mak (chargé de pharmacovigilance en Suisse) et Dina Pfeifer (du département vaccins et immunisations de l’OMS), est paru dans le même numéro du Lancet sous le titre « Pandemic H1N1 infection in pregnant women in the USA« .

Leur commentaire rappelle certes les limitations et biais multiples de l’étude de Jamieson et al., mais insiste néanmoins sur le fait que les femmes enceintes doivent être considérées comme une population à risque et traitées en conséquence, notamment par la vaccination – qui serait bénéfique aussi aux nouveau-nés – et par les médicaments antiviraux, même lorsque le traitement est tardif. Mangtani et al. considèrent aussi qu’une vaccination des femmes enceintes contre la grippe saisonnière est souhaitable, quel que soit le trimestre de la grossesse.

Ils reconnaissent qu’il n’y a pas de données solides quant au rapport bénéfices – risques, particulièrement pour le premier trimestre de la grossesse, ce qui ne les empêche pas de dire que le vaccin ne semble pas avoir posé de problèmes. Il faudra mettre en place un système de recueil des effets secondaires, disent-ils.

© Pharmacritique pour le texte en français

Voici un extrait du commentaire de Mangtani et al. (suivi d’un plus long extrait de l’article de Jamieson). Chacun se fera une idée en lisant sur quoi se basent les conclusions et préconisation de ces auteurs.

“Although the true level of increased risk remains uncertain, the US analysis clearly indicates a need to provide particularly close medical care to pregnant women who might have pandemic H1N1 viral infection. These findings also indicate the importance of preventive measures including immunisation, which is also likely to benefit the offspring,⁶ and of the need for timely intervention with neuraminidase inhibitors in confirmed or strongly suspected cases in pregnant women. Although much less ideal or effective, there is some suggestion that delayed use of oseltamivir (>48 h since symptom onset) might still reduce maternal mortality, as seen in non-pregnant individuals infected with other influenza viruses.^{[7] and [8]} Use and safety data in pregnancy are scarce with neuraminidase inhibitors,⁹ thus any new information about the safety and effectiveness of drug or drug combinations, dose, or duration of therapy in pregnancy should be well communicated.

Seasonal influenza vaccination in all trimesters of pregnancy has been recommended for several years in some countries, including the USA. Although safety data are scarce, particularly in the first trimester, the available vaccine safety data have not raised concerns.¹⁰ Recommendations for H1N1 influenza vaccination in pregnancy have been recently issued by the Strategic Advisory Group of Experts on Immunization, an independent WHO advisory group¹¹ and the US Advisory Committee on Immunization Practices.¹² »

Voici la dernière partie de l’article de Jamieson et al.: « H1N1 2009 influenza virus infection during pregnancy in the USA » 

“This study summarises the cases of pregnant women with pandemic H1N1 virus infection in the USA and shows that this virus can cause serious illness in healthy pregnant women. Before the present outbreak, published work of influenza virus of swine origin in pregnant women was limited to a single case in 1988: a 32-year-old previously healthy pregnant woman at 36 weeks’ gestation was infected with a swine influenza virus, contracted through exposure to pigs, and later died of complications related to primary viral pneumonia.^{[11] and [12]}

On the basis of our investigation, pregnant women seem to be at increased risk for complications from pandemic H1N1 virus infection, with a higher estimated rate of hospital admission than in the general population. Although the decision to admit a pregnant woman is complex and might include considerations beyond simply the severity of disease, that a high proportion (>10%) of influenza-related deaths in the USA have been in pregnant women is concerning. In the previous influenza pandemics of 1918 and 1957, mortality seemed to be higher in pregnant women than in non-pregnant populations, although appropriate comparison groups were often not available. In a series of 1350 pregnant women reported during the 1918 pandemic,⁸ about 50% developed pneumonia and o
f these women, more than half died (overall case fatality rate 27%), with the highest mortality in the third trimester.

During the pandemic of 1957, 50% of deaths due to Asian influenza in Minnesota among women of reproductive age occurred in pregnant women.⁷ The groups of women selected for inclusion in reports from the 1918 and 1957 pandemics might have been biased towards more severe cases, and in 1918, the diagnosis of influenza was based on the clinical syndrome alone because human influenza viruses were not identified until 1933. Thus, comparison of severity of disease in pregnant women in previous pandemics compared with the present outbreak, particularly in view of the few documented cases so far, is difficult. However, that all six deaths reported during the present outbreak were in relatively healthy pregnant women is noteworthy.

Pregnant women are also at increased risk for pregnancy complications during seasonal influenza epidemics. In one study, pregnant women were more likely to be admitted for a cardiopulmonary event during influenza season than were those who were post partum (a group considered similar to pregnant women demographically and by health status). The risk varied by weeks of gestation, with odds ratios of 1·06 (95% CI 0·68–1·67) during weeks 1–7, 2·52 (1·74–3·65) during weeks 21–26, and 4·67 (3·42–6·39) during weeks 37–42.

A study from Nova Scotia⁵ showed that women were more likely to be admitted to the hospital for respiratory illness during influenza season during pregnancy than during the year before pregnancy. In pregnant women with no comorbidities (defined as pre-existing diabetes, pulmonary disease, heart disease, renal disease, and anaemia), rate ratios for hospital admissions by trimester of pregnancy compared with the year before pregnancy were 1·7 (1·0–2·8) for first, 2·1 (1·3–3·3) for second, and 5·1 (3·6–7·3) for third trimesters. The effects of pregnancy were greatest in women with one or more comorbidities, with rate ratios of 2·9 (1·5–5·4), 3·4 (1·9–6·0), and 7·9 (5·0–12·5), respectively. In our investigation, a fifth of women admitted to hospital had a history of mild asthma, but only one reported using any medications for asthma.

The present circulating pandemic H1N1 virus is sensitive to the neuraminidase inhibitors oseltamivir and zanamivir. In randomised controlled clinical trials,¹³ these drugs have reduced the severity of seasonal influenza if started within 48 h of illness onset. Although data suggest that oseltamivir can reduce mortality in admitted patients even when started more than 48 h after illness onset, CDC recommendations for pregnant patients are that antiviral drugs be started as soon as possible after the onset of influenza symptoms. The benefit is expected to be greatest if started within 48 h of onset. However, many pregnant women in our series were not treated with either of these drugs at the time of their presentation with influenza-like illness. Furthermore, none of those who died were treated within 48 h of illness onset.

We did not record a delay in diagnosis of influenza in pregnant women. However, in many cases, there seem to have been delays in initiation of antiviral therapy. Health-care providers might have been reluctant to treat patients with antiviral drugs because they were pregnant, or before laboratory confirmation of disease, or the pregnant woman might have been reluctant to take an antiviral drug. As with most drugs,^{[14] and [15]} information about the safety and effectiveness of these anti-influenza drugs during pregnancy is scarce.^{[16], [17] and [18]} In view of the expected effects of pandemic H1N1 influenza virus on the pregnant woman, the benefits of treatment with these drugs are likely to outweigh potential risks to the fetus.^{[19] and [20]}

Current CDC guidance suggests that antiviral treatment with oseltamivir or zanamivir is recommended for groups at high risk for influenza complications from infection with H1N1 virus, including pregnant women (oseltamivir is preferred). However, some women were diagnosed before the availability of these guidelines. Communication messages aimed at pregnant women and their health-care providers should include information about the benefits and risks of anti-influenza drugs, and about the increased risk of influenza complications in pregnant women. Additionally, improved understanding of the effects of influenza during pregnancy, both seasonal influenza and novel strains, and of anti-influenza drugs used for treatment is crucial; additional data could be obtained through establishment of a national pregnancy registry.

Since most women in this series are still pregnant, little is known about the effects of the pandemic H1N1 virus on the fetus. Although no infections have been reported in infants born to women with H1N1 virus infection, these infants might have had more subtle effects from maternal pandemic H1N1 virus infection. Furthermore, the effects of seasonal influenza on the fetus are not well understood. Although viraemia seems to be rare in seasonal influenza²¹ and placental transmission seems to occur infrequently,²² highly pathogenic strains of influenza virus, such as avian influenza A (H5N1), might be transmitted across the placenta. This was shown in a pregnant woman infected with H5N1, with viral genomic sequences identified in the placental cytotrophoblasts and in the fetal respiratory tract.^{[23] and} [24] Y Shu, H Yu and D Li, Lethal avian influenza A (H5N1) infection in a pregnant woman in Anhui Province, China, N Engl J Med 354 (2006), pp. 1421–1422. ^[24] Additionally, even in the absence of placental transmission, the fetus could be adversely affected by influenza or its effects. For example, fever, which often accompanies influenza, has been associated with an increased risk for neural tube defects when occurring in the first trimester²⁵ and with other adverse neonatal or developmental outcomes, when occurring later in pregnancy;
^{[26], [27] and [28]} thus, treatment of fever with acetaminophen is recommended.²⁰ Seasonal influenza has been associated with a small increased risk of birth defects in some studies, although others have not reported this association.²⁹

Concerns about influenza’s effects on the fetus were raised during previous pandemics. In the pandemic of 1918, high rates of pregnancy loss and preterm delivery were reported,^{[8] and [9]} and during the pandemic of 1957–58, possible increases in CNS defects and other adverse outcomes were shown.^{[30], [31], [32] and [33]} Follow-up of outcomes of pregnancy to women infected with H1N1 virus is needed to improve understanding of the possible effects of this novel virus.

Once available, vaccination will be an essential component of the public health response to this influenza, and US guidelines place pregnant women in a high-priority group for receipt of pandemic influenza vaccine.²⁰ However, in one study³⁴, pregnant women had the lowest vaccine coverage level (14·4% in 2004) of all adult population groups recommended to receive influenza vaccination,³⁴ despite trivalent inactivated vaccine being recommended by the US Advisory Committee on Immunization Practices and the American College of Obstetricians and Gynecologists during any trimester of pregnancy.^{13 and 35} The low level of use of influenza vaccine in pregnant women³⁴ is disconcerting and has important implications for future pandemic vaccination planning.

In addition to the protection provided to mothers, influenza vaccination also seems to provide benefit to infants; in a randomised study in Bangladesh,³⁶ inactivated influenza vaccine reduced proven influenza illness by 63% in infants aged 6 months or younger. The reasons for the low level of influenza vaccination coverage in the USA are not well understood, although concerns about vaccine safety are often cited by mothers as barriers to vaccination.³⁷ Information about safety of influenza vaccine during pregnancy is scarce, but available data show no evidence of adverse effects on women or their infants.³⁸ Knowledge gaps in women and their health-care providers must be addressed³⁹ to improve vaccination coverage for pregnant women during interpandemic periods and during a pandemic.²⁰

Our investigation has several limitations. Ascertainment of women infected with pandemic H1N1 influenza virus was dependent on surveillance and laboratory testing methods used by state public health authorities during the outbreak. These methods varied by state and by the timing during the outbreak. For example, because of limitations on laboratory capacity and the likelihood that results of testing would be unlikely to change clinical care, laboratory testing for the virus varied as the outbreak progressed: early in the outbreak, most people with influenza-like illness who presented for medical care were tested, but later in the outbreak, testing was focused on more severe cases (eg, those admitted).

Furthermore, data from a survey of obstetricians and gynaecologists in 2004,⁴⁰ suggested that pregnant women might be less likely to be tested than were those who were not pregnant. Of the obstetricians and gynaecologists who had seen pregnant women whom they suspected of having influenza, 84% reported that they rarely or never tested pregnant women for influenza. This report also summarises information that was collected by the states and reported to the CDC. In view of the substantial stress on the public health system during this outbreak, states might not have reported all cases and might have been most likely to report severe cases.

Another limitation is that health-care providers might be more likely to admit a pregnant woman than a non-pregnant person with similar findings, which could lead to an exaggerated admission rate in pregnant women. Last, the estimated proportion of all pandemic H1N1 deaths in pregnant women is an unstable estimate, in view of the small number of deaths reported so far. If we increased the reporting period by 1 week, the proportion of pandemic H1N1 influenza deaths in pregnant women would be 8% (seven of 87) instead of 13% (six of 45).

Findings from this study will be crucial to inform public health planning for pregnant women, both for this virus and for other novel pathogens. Crucially, health-care providers have to realise that pregnant women are at increased risk for severe disease and complications from pandemic H1N1 influenza virus infection, and should start treatment with anti-influenza drugs promptly.”