We have a list of drugs that we tend to use most and we favor continuous when possible and drug with the largest volume of distribution. I'll find the list and attach it with my next post.

I recommend the excellent pharmakokinetic tool provided on the following website: http://www.jjpreisenberger.de.

We use the algorithym and tables in these two articles.

Freebairn R, Cohen J, Lipman J Prescription of Antimicrobial Agents in Patients Undergoing Continuous Renal Replacement Therapy In: Rello J., Kollef M. (et al.) (Eds.) Infectious Diseases in Critical Care 2007, ISBN: 978-3-540-34405-, 2007

Choi G, Gomersall CD, Tian Q, Joynt GM, Freebairn RC, Lipman J. Principles of antibacterial dosing in continuous renal replacement therapy. Crit Care Med. 2009 Jul;37(7):2268-82

Another older but classic paper is this one:

I recommend the articles:

Heintz BH, Matzke GR, Dager WE. Antimicrobial dosing concepts and recommendations for critically ill adult patients receivint continuous renal replacement therapy or intermittent hemodialysis. Pharmacotherapy 2009;29(5):562-77.

Mushatt DM, Mihm LB, Dreisbach AW, Simon EE. Antibiotic dosing in slow extended daily dialysis. CID 2009;49:433-7.

Carcelero E, Soy D. Antibiotic dose adjustment in the treatment of MRSA infections in patients with acute renal failure undergoing continuous renal replacement therapies. Enferm Infecc Microbiol Clin 2012;30(5):249-56.

In 2006, we conducted with our pharmacy department a six months prospective study to determine when we should order Cmin of aminoglycosides in case of septic shock or severe sespis; we used at this time conventional loading dose i.e 5mg/kg for Genta or Tobra and 15 mg/kg for Amikacin. We demonstrated that Cmin determinations were required only in case of severe renal deficiency defined by an estimated clearance < 60 ml/min. This study as been reported in ESICM in Vienna in 2009 ( eposter 477), but the main remarks were the absence of Cmax ( 1 hour post loading dose).

According to the literature, we used , for at least four years, larger loading doses: 7.5 or 8mg/kg for Genta or Tobra and 20 or 25 mg/kg for Amikacin (dose depending of the physician on duty!) and monitored the Cmin 20 to 24 hours after administration. In case of CRRT and a Cmin < 1mg/l, we redose at 5mg/kg for G/T and 15mg/kg for AMK if Cmin < 5mg/l and then wait for a new dosage on day 2. If Cmin level are higher , we wait six hours for a new Cmin determination.

For time-dependent antibiotics, we use only continuous infusion in case of severe sepsis secondary to pseudomonas aeruginosa. For Piperacillin/taxobactam we order 4.5 g every 8 hours, and 2g every two times a day.

Hallo,

we have shown (Seyler et al, Crit Care 2011) that Trotman recommendations did not result in adequate beta-lactams concentrations if you want to target difficult-to-treat pathogens, such as Pseudomonas or ESBL.

We are preparing another paper on our experience with "unadapted" drug regimens - in other words, we give the same drug regimen as for normal renal function for the first 48 hours, then we reduce regimens to those proposed by Trotman et al. to avoid overdosage.

Importantly, this strategy makes sense if you target gram-negative pathogens with higher thresholds of susceptibility to beta-lactams or if you suspect Pseudomonas aeruginosa infections, while "recommended" regimens will be absolutely adequate if you deal with very sensitive strains.

For amino glycosides, I agree on the concept of giving "high" dose, provided you keep Cmin below the threshold of potential toxicity - in some cases, high-flow RRT will allow to give high drug regimens and to induce rapid elimination - this would result in almost daily drug administration, with potentially increased antimicrobial activity.

For Vancomycin, we use only continuous infusion - data on a new drug regimen, prospectively validated in 32 patients, have been submitted to JAC: 35 mg/kg loading dose over 4-hr followed by 14 mg/kg over 24 hours. For intermittent infusion, article suggest 750 to 1000 mg q12h, but this will be dependent on the intensity of RRT.

The use of algorithm demands the knoledge of pharmacokinetic data for particular antibiotics. Unfortunately these data were obtained under very individual clinical situations. More , these data are being changed during CRRT therapy . In conclusion , there are no accurate methods for solving this problem

I perfectly agree with the Erasmus team, the main remaining problem is the Beta-lactam administration which is not yet evidence based. I take into consideration that JL Vincent and colleagues propose to use normal dose during the first 48 hours and then reduce doses to avoid toxicity as proposed in the literature.

We once derived a general predictive rule based on pharmacokinetic data.

http://ndt.oxfordjournals.org/content/11/8/1582.abstract

However, not being a nehrologist I cannot tell wether there are more accurate methods available now.

We have a clinical pharmacist on our daily rounds and he using the algorithim in his little book.

Stephen, you are very right. In several of the above mentioned references it is not easy to find corrections for dose of dialysis or hemofiltration. It will of course matter for the drug removal wether we have a filtration of 2000 versus 4000 ml/h, and also if replacement fluid is given pre or post filter. In the former we have reduced drug clearance because of dilution of the blood. It is really strange that these considerations is so easily "forgot" in guidelines. Only in some of the references (se from Ross Freebairn above) this problem is partially covered. A large problem however is the lack of clinical data in this area.