Intensive Care Med (2016) 42:946–948
DOI 10.1007/s00134-016-4267-x
LETTER
Extracorporeal membrane oxygenation
(ECMO) does not impact on amikacin
pharmacokinetics: a case–control study
E. Gélisse1,2, M. Neuville2, E. de Montmollin3, L. Bouadma2,4, B. Mourvillier2, J. F. Timsit2,4 and R. Sonneville2,5*
© 2016 Springer-Verlag Berlin Heidelberg and ESICM
Dear Editor,
More than half of adult patients receiving extracorporeal
membrane oxygenation (ECMO) require an antibiotic
therapy during their stay in ICU [1]. Although amikacin
is one of the most used aminoglycosides for the treat-
ment of severe infections, data on its pharmacokinetics in
patients receiving ECMO support are scarce. We report
on the distribution of peak serum concentration (Cmax) of
amikacin and on factors associated with insufficient Cmax
in a population of critically ill patients receiving ECMO.
We conducted an observational single-centre study
of patients admitted to a general ICU. Inclusion criteria
were (1) need for ECMO support and (2) a suspected
Gram-negative infection requiring loading dose of ami-
kacin. Amikacin was routinely administered intrave-
nously at dose of 25  mg/kg of total body weight over a
30-min infusion time, as described previously [2]. Cmax
was measured 30  min after the end of infusion and
trough serum concentration 24  h after the end of infu-
sion (Cmin). Patients from the ECMO group (ECMO)
were matched with 50 critically ill patients without
ECMO support (controls). The matching procedure was
performed for characteristics that had been previously
identified as independent predictors for insufficient Cmax,
namely (1) body mass index, (2) cirrhosis and (3) 24-h
fluid balance [2]. This study was approved by the appro-
priate ethics committee.
The ECMO group consisted of 50 episodes of amika-
cin initiation under ECMO (venoarterial ECMO, n = 43
*Correspondence: romain.sonneville@aphp.fr
2
Department of Intensive Care Medicine and Infectious Diseases, Hôpital
Bichat‑Claude‑Bernard, Assistance Publique Hôpitaux de Paris AP-HP, 46
rue Henri Huchard, 75018 Paris, France
Full author information is available at the end of the article
and venovenous ECMO, n  =  7) in 46 patients. The
median time between ECMO cannulation and Cmax was
4  days (interquartile range, 1.0–7.5  days). On the day of
Cmax, 100 % of ECMO patients required mechanical ven-
tilation, 96  % vasopressors and 44  % renal replacement
therapy. Baseline characteristics did not differ between
the two groups, except for age and reason for ICU admis-
sion (Table 1). There was no significant difference in Cmax
between ECMO and controls [71.7 (58.9–79.7)  mg/l in
ECMO vs 68.4 (53.0–81.0)  mg/l in controls, p  =  0.36].
The proportions of insufficient [Cmax  <60  mg/l: 13/50
(26  %) in ECMO vs 17/50 (34  %) in controls], adequate
[60  mg/l  <  Cmax  <  80  mg/l: 25/50 (50  %) in ECMO vs
18/50 (36 %) in controls] and excessive [Cmax > 80 mg/l:
12/50 (24 %) in ECMO vs 15/50 (30 %) in controls] ami-
kacin peak concentrations were similar in the two groups.
Cmin was determined in 43/50 (86 %) cases in the ECMO
group. The proportion of measurements above the
toxic threshold of 5  mg/l was similar in the two groups
[28/43 (65 %) in ECMO versus 30/50 (60 %) in controls,
p = 0.67].
The results of our study suggest that ECMO does not
significantly impact on peak and trough amikacin levels
in plasma. Although ECMO patients differed from non-
ECMO patients for some parameters (including age,
reason for ICU admission, time between ICU admis-
sion and Cmax measurement, and SOFA scores), none of
them were shown to be independently associated with
insufficient Cmax in a previous study [2]. Using a 25 mg/
kg standardized amikacin loading dose, we observed an
insufficient Cmax in 25  % of ECMO cases, stressing the
need for even higher doses and systematic therapeutic
drug monitoring in those patients [3–5].
 Page 947 of 948

Table 1  Patients’ characteristics

Variable ECMO n = 50 episodes Controls n = 50 episodes p value

Characteristics at ICU admission

 Number of patients 46 50
 Age (years) 61 (43–68) 63.5 (54–72) 0.03
 Cirrhosis 2 (4) 4 (8) 0.68
Reasons for ICU admission <0.001
 Acute respiratory failure 7 (15) 13 (26)
 Severe sepsis/septic shock 5 (11) 12 (24)
 Cardiogenic shock 21 (46) 4 (8)
 Cardiac surgery 10 (22) 11 (22)
 Cardiac arrest 5 (11) 3 (6)
 Other 2 (4) 7 (14)
 BMI (kg/m2) 26.5 (24.2–29.4) 26.3 (24.5–30.9) 0.50
 SAPS 2 51 (42.5–73.3) 53.5 (41–73.3) 0.98
Characteristics at time of Cmax
 SOFA score 12 (10–14) 9 (6–11) <0.001
 24-h fluid balance (ml) 100 (−925 to 1860) 100 (−1550 to 1019) 0.52
 Creatinine (µmol/l) (in patients without CRRT) 97 (79–146) 77 (63–124) 0.37
 CRRT 22 (44) 25 (50) 0.69
Suspected source of infection 0.05
 VAP 30 (60) 41 (82)
 Bacteraemia 4 (8) 1 (2)
 Other 16 (32) 8 (16)
Pharmacokinetic parameters
 Weight at time of Cmax (kg) 88.5 (73–100) 82 (70–93) 0.37
 Amikacin dose (mg) 2250 (1825–2500) 2000 (1713–2300) 0.17
 Amikacin regimen (mg/kg) 25.0 (25.0–25.6) 24.9 (24.5–25.5) 0.10
 Cmax, mg/L 71.7 (58.9–79.7) 68.4 (53.0–81.0) 0.36
 Cmax < 60 mg/l 13 (26) 17 (34) 0.51
 Cmax > 80 mg/l 12 (24) 15 (30) 0.65
 Amikacin AUC (mg h/ml) 973 (799–1193) 921 (663–1203) 0.55
 Cmin (mg/l) 8.5 (3.0–15.4) 9.6 (2.5–16.9) 0.45
Data are median (IQR) or n (%). Differences between groups were assessed with the chi-squared test, Fisher’s exact test, Student’s t test or Mann–Whitney U test, as
appropriate
IQR interquartile range, BMI body mass index, SAPS Simplified Acute Physiology Score, SOFA Sequential Organ Failure Assessment, ECMO extracorporeal membrane
oxygenation, GFR glomerular filtration rate, VAP ventilator associated pneumonia, CRRT continuous renal replacement therapy, Cmax amikacin peak concentration, Cmin
amikacin trough concentration, AUC area under the curve

Author details Accepted: 8 February 2016

1
 Department of Intensive Care Medicine, CHU Reims, Reims, France. Published online: 24 March 2016
2
 Department of Intensive Care Medicine and Infectious Diseases, Hôpital
Bichat‑Claude‑Bernard, Assistance Publique Hôpitaux de Paris AP-HP, 46 rue
Henri Huchard, 75018 Paris, France. 3 Department of Intensive Care Medicine,
CH Saint Denis Delafontaine, Saint‑Denis, France. 4 IAME U1137, Decision Sci-
ence in Infectious Diseases, INSERM, Université Paris Diderot, Sorbonne Paris References
Cité, Paris, France. 5 LVTS U1148, INSERM, Université Paris Diderot, Sorbonne 1. Schmidt M, Brechot N, Hariri S, Guiguet M, Luyt CE, Makri R, Leprince P,
Paris Cité, Paris, France. Trouillet JL, Pavie A, Chastre J, Combes A (2012) Nosocomial infections in
adult cardiogenic shock patients supported by venoarterial extracorpor-
Compliance with ethical standards eal membrane oxygenation. Clin Infect Dis 55:1633–1641
2. de Montmollin E, Bouadma L, Gault N, Mourvillier B, Mariotte E, Chemam
Conflicts of interest  S, Massias L, Papy E, Tubach F, Wolff M, Sonneville R (2014) Predictors of
On behalf of all authors, the corresponding author states that there is no insufficient amikacin peak concentration in critically ill patients receiving
conflict of interest. a 25 mg/kg total body weight regimen. Intensive Care Med 40:998–1005
 Page 948 of 948

3. De Rosa FG, Roberts JA (2014) Amikacin dosing in the ICU: we now know 5. Taccone FS, Laterre PF, Spapen H, Dugernier T, Delattre I, Layeux B, De
more, but still not enough. Intensive Care Med 40:1033–1035 Backer D, Wittebole X, Wallemacq P, Vincent JL, Jacobs F (2010) Revisiting
4. Roberts JA, Taccone FS, Lipman J (2015) Understanding PK/PD. Intensive the loading dose of amikacin for patients with severe sepsis and septic
Care Med. doi:10.1007/s00134-015-4032-6 shock. Crit Care 14:R53