Early oseltamivir therapy improves the outcome in critically ill patients with influenza: a propensity analysis.

Dear Editor,

Influenza affects between two and three million people worldwide each year, with complications responsible for a significant number of excess hospitalizations in intensive care units (ICUs) [1]. Since the newsworthy 2009 A(H1N1) pandemic (pdm), publications about influenza in ICUs remain scarce, with crucial outstanding issues on prognosis factors including the timing of antiviral treatments [2-5]. Here, we present a multicenter prospective study of critically ill influenza-infected patients aimed to identify prognosis factors associated with death.

This study was conducted from December 2008 to April 2013 in the 12 polyvalent ICUs from the Lyon catchment area (France). All adult patients admitted with microbiologically confirmed influenza infection were included. Following univariate comparisons, the independent contribution of patients’ characteristics to in-hospital mortality was analyzed by backward stepwise multivariate analysis in a logistic regression model. Propensity score-matching was further used to compare similar patient populations receiving oseltamivir within or after 2 days of the onset of symptoms.

Over the study period, 201 patients were enrolled with the following main reasons for ICU admission: respiratory distress (n = 174, 87%), shock (n = 13, 6.5%), and neurological failure (n = 7, 3.5%). Infections were mostly caused by type A influenza virus (n = 171, 85%), with A(H1N1)pdm encountered in half of cases (n = 100, 50%). Other characteristics of the patients are presented in Table 1. Only 40 (20%) patients received antiviral treatment by oseltamivir within 2 days of the onset of symptoms. Day-28, ICU, and in-hospital mortalities were 18% (37/201), 21% (43/201), and 26% (53/201), respectively. Survivors were more likely to have received early oseltamivir therapy, with a significant trend in proportions of in-hospital mortality with increasing time from onset of symptoms to initiation of treatment (p = 0.01). Oseltamivir administration within 2 days of the onset of symptoms appeared as the sole independent determinant of a favorable outcome (OR 0.26; 95% CI 0.08–0.79, p = 0.02). After propensity score-matched analysis, oseltamivir therapy within 2 days of the onset of symptoms was associated with reduced in-hospital mortality (Supplementary Table 1).

Table 1

Patients’ characteristics according to outcome

	Total n = 201	Survivors n = 148	Non-survivors n = 53	Univariate analysis p	Multivariate analysis OR (95% CI)
Clinical features
Agea (years)	63 ± 16	62 ± 15	66 ± 17	0.088	1.03 (1.00–1.05)*
Male sex	107 (53)	81 (55)	26 (49)	0.477
Influenza vaccination	25 (12)	19 (13)	6 (11)	0.773
BMIa (kg m−2)	27 ± 7	28 ± 7	26 ± 7	0.168
No comorbidity	19 (9)	14 (9)	5 (9)	0.996
SAPS II	44 ± 17	40 ± 15	54 ± 17	< 0.001
Clinical course
Symptom duration before ICU admission (days)	5.1 ± 5.0	5.1 ± 4.5	5.1 ± 6.0	0.931
Influenza diagnosis before ICU admission	20 (10)	14 (9.5)	6 (11)	0.698
Days from admission to influenza diagnosisa	2.3 ± 3.4	2.0 ± 2.7	2.9 ± 4.9	0.112
Microbiological results
Viral subtype				0.707
Influenza A(H1N1)pdma	100 (50)	73 (49)	27 (51)	–
Other type A	71 (35)	54 (37)	17 (32)	–
Type B	30 (15)	21 (14)	9 (17)	–
Bacterial co-infection at admission	48 (24)	38 (26)	10 (19)	0.319
VAP in the evolutiona	51 (25)	31 (21)	20 (38)	0.016	1.91 (0.89–4.10)
Initial organ failures
Type
Respiratory	109 (54)	74 (50)	35 (66)	0.044
Cardiovascular	86 (43)	58 (39)	28 (53)	0.003
Neurological	50 (25)	30 (20)	20 (38)	0.011
Hematological	24 (12)	8 (5.4)	16 (30)	< 0.001
Renal	21 (10)	10 (6.8)	11 (21)	0.004
Hepatic	2 (1.0)	1 (0.7)	1 (1.9)	0.459
Number of failed organs	1.5 ± 1.2	1.2 ± 1.1	2.1 ± 1.2	< 0.001
SOFA scorea	6.6 ± 3.9	5.9 ± 3.8	8.7 ± 3.7	< 0.001	1.19 (1.08–1.30)*
Organ support during ICU stay
Mechanical ventilation	179 (89)	126 (85)	53 (100)	0.003
ARDS	113 (56)	73 (49)	40 (75)	0.001
NIV only	36 (18)	31 (22)	5 (9.4)	0.061
Days	14 ± 20	12 ± 17	19 ± 28	0.084
ECMOa	8 (4.0)	4 (2.7)	4 (7.5)	0.211
Catecholamines (days)	4.7 ± 8.8	3.3 ± 6.8	8.2 ± 12.4	0.005
RRT (days)	3.9 ± 12.0	2.7 ± 8.8	7.3 ± 17.9	0.073
Other treatments
Prone positiona	38 (19)	23 (15)	15 (28)	0.041
Neuromuscular blockade	67 (33)	45 (30)	22 (42)	0.141
Nitric oxide	19 (9.5)	11 (7.4)	8 (15)	0.101
Oseltamivir	146 (73)	109 (74)	37 (70)	0.591
Administration before ICU admission	17 (8.5)	13 (8.7)	4 (7.5)	0.999
Dose (mg/day)	201 ± 84	197 ± 82	215 ± 98	0.257
≤ 2 days after onset of symptomsa	40 (20)	36 (24)	4 (7.5)	0.009	0.26 (0.08–0.79)*
Days from onset of symptoms to initiation	4.9 ± 3.5	4.6 ± 3.6	5.7 ± 3.1	0.086
Steroidsa	91 (45)	63 (43)	28 (53)	0.288

Data are number (%) or mean ± standard deviation, as appropriate

OR odds ratio, CI confidence interval, BMI Body Mass Index, ICU intensive care units, VAP Ventilator-Associated Pneumonia, ARDS Acute Respiratory Distress Syndrome, SOFA Sepsis-Related Organ Failure Assessment, NIV non–invasive ventilation, ECMO extra-corporeal membrane oxygenation, RRT renal replacement therapy

*p < 0.05

aVariables included in the backward stepwise logistic regression

The present study emphasizes the delay in oseltamivir administration as a major risk factor for in-hospital mortality. Influenza is a public health problem that, each year, causes both severe illness and deaths in high-risk populations [1]. With the exception of the 2009 pandemic, data on critical illnesses attributable to influenza are scarce. Thus, we designed the present study to provide current information on influenza disease in ICUs in the real situation of a specific territory. Concerning the severity of the patients’ illness or the risk factors of death, our cohort is in agreement with previous studies on influenza-associated critical illness mainly drawn from 2009 pandemic studies [2, 3]. Nevertheless, our study does not confirm the negative impact on patients’ outcomes of the influenza A(H1N1)pdm virus subtype. Importantly, as confirmed by propensity analysis, oseltamivir administration was associated with better outcomes when administrated within 2 days of the onset of symptoms. This is the key message of our work, consistent with recent studies in non-severe forms of the disease, including meta-analysis of randomized clinical trials [4]. Indeed, the recent literature highlights the efficacy of oseltamivir to reduce the duration of symptoms, respiratory tract complications, and hospital admittance [4, 5]. Our results extend to the ICU setting the relationship between the delay of oseltamivir administration and the effectiveness of the treatment in patients with either A(H1N1)pdm or other influenza virus subtypes.

In conclusion, this real-life study emphasizes oseltamivir efficacy on in-hospital outcome when administrated within 2 days of the onset of symptoms. Even if physicians’ awareness of the influenza disease has been undeniably enhanced since the last pandemic, many efforts are still required to improve influenza-infected patient management in ICUs, including early oseltamivir administration.

Below is the link to the electronic supplementary material.

Supplementary material 1 (DOCX 14 kb)

Supplementary material 2 (DOCX 13 kb)