As far as I know it has been demonstrated that single nucletdide polymorphism (SNP) of CYP3A4, CYP3A5 and ABCB1 genes determine significant changes in the pharmacokinetic profile of tacrolimus (and cyclosporine as well). Although it seems that the final effect is mediated by complex intaractions among the SNPs the exact nature of these interactions is not yet completely understood. Probably stratification for SNPs may prove useful in modelling real-life data of tacrolimus, provided this information is available.

Kind regards

https://www.ncbi.nlm.nih.gov/pubmed/20170205

https://www.ncbi.nlm.nih.gov/pubmed/26543771

Your example presents a continous decrease of serum concentrations. This might be an argument against an assumed hidden administration. In the time interval from 12 to 60h post dose the values decrease for about 50% giving a rough estimate of of T1/2 of about 50h.

M.Sardina already pointed on the importance of polymorphisms. Of course this includes ethno-PK differences. Choi et al 2007 investigated healthy Korean subjects and found in the CYP3A5*3 genetic polymorphism subgroup prolonged elimination half lifes. In the 3 patients with diplotype TTT-TTT the range of T1/2 values was 21.0 to 46.9h.

It is not a surprise that you observe these plasma concentration profiles in some patients, you do not have to assume un-recorded drug administration.

Kind regards

If you use the 17-->8.7 values, it looks like a T1/2 of ~50 hrs, but the earlier values do not seem to fit into a simple PK model.  My first thought was is there some redistribution from tissue storage sites back into plasma, and are the posted values actually are a combination of residual drug from the initial dose plus redistributed drug?  That might explain the apparent lack of linearity among the interim levels from 17 to 8.7.

Another possible explanation might involve polymorphism in metabolism and saturation of 1-2 enzymes by a metabolite or feed back mechanism that decreased interim metabolism and produced non-linear data.  Just a thought, my colleagues.

The patients condition have an input in Tacrolimus concentration in clinical practice if you have patients with dehydration then the drug concentration is high. 

You might see Aasberg A, Midtvedt K, van Guilder M, Storset E, Bremer S, Bergan S, Jelliffe R, Hartmann A, and Neely M: Inclusion of CYP3A5 genotyping in a nonparametric population model improves dosing of tacrolimus early after transplantation. Transplant International on behalf of Steunstichting ESOT 2013; 26: 1198 – 1207. Very best regards, Roger Jelliffe

You might also see Aasberg A, Midtvedt K, van Guilder M, Storset E, Bremer S, Bergan S, Jelliffe R, Hartmann A, and Neely M: Inclusion of CYP3A5 genotyping in a nonparametric population model improves dosing of tacrolimus early after transplantation. Transplant International on behalf of Steunstichting ESOT 2013; 26: 1198 – 1207

It should also pay attention in addition to dehydration, which is already mentioned, when his value and also with which medication is given!

For example reaction with a hazard can be assumed dace tacrolimus exhibit an interaction with drugs that are related to cyclosporine!

The level of tacrolimus increase for example imidazole, antifungals,,,, macrolides danazol, omerprazol, bromocriptine, cortisone, dapsone, ethinyl estradiol, gestodene, lignocaine,, nicardipine, nifedipine, verapamil, quinidine!

The values of tacrolimus may reduce; rifampicin, theophylline, phenobarbital, karbamzepin,  isoniazid!

About this an interaction should be taken into accounts because they influence directly the level of drug!

.

 Of course it is understood that the monitoring of the drug in the blood and the desirable value of Cyclosporin Cyclosporin is from 100-300 ng / ml and for tacrolimus is considered that the effective serum concentrations of the drug on, o,5-3 ng / ml !

For TDM, it is important to obtain a blood sample just prior to the next dose in order to document the Cmin. especially for drugs with a low therapeutic index and narrow therapeutic serum concentration, like tacrolimus.  If you know Tmax, Cmax also can be obtained subsequently.  Many clinicians take random blood samples that have no meaning.  More emphasis must be placed on teaching therapeutics to medical students and practicing docs w/o adequate training in PK.

I reviewed a case many years ago where a pt was taking tacrolimus as an outpatient and did not shake the bottle before taking it.  The formulation had "separated" while standing and the pt essentially took only vehicle and no active drug.  This led to a GvH reaction and rejection of the transplanted kidney and the pt's death.

Homogeneity of a formulation and proper formulation development and pt instructions are critical to delivering the correct dose of liquid formulations.  Back in 1937, sulfanilamide was dissolved in diethylene glycol b/c of solubility issues.  Thirty to 50 pts developed kidney failure, and in 1938 the Food Drug and Cosmetic Act was enacted in the US to ensure the safety of medications sold to the public.  It seems as if we are always reacting (to a crisis) rather than being pro-active and preventing one!

 It is always a good idea to get back to "fundamentals."

Dear All:

     I had replied to this but do not see it, so I will try again. First, it is not enough simply to get a trough drug sample before the next dose. One does much  better with Bayesian analysis, making a model of drug behavior in each individual patient. Optimal sampling strategies such as D-optimal or multiple model optimal design are now available in clinical software for this. Other samples taken at other times do have meaning and can also be analyzed. Also, interacting multiple model (IMM) sequential Bayesian analysis permits optimal tracking of drug behavior in acutely ill unstable patients (dehydrated or receiving dialysis, etc.,, for example) with high intra-individual variability, developing maximally precise regimens for the near future, using multiple model (MM) dosage design and nonparametric (NP) models of drug behavior which hit desired target goals (individually selected for each patient based on sensitivity and tolerance) rather than slavish acceptance of general guidelines. These methods are  discussed in a recent book edited by Dr. Michael Neely and myself, entitled : “Individualized Drug Therapy for Patients – Basic Techniques, Relevant Software, and Clinical Applications”. Imprint: Academic Press. Release Date: November 15, 2016. Print Book ISBN : 9780128033487. Pages: 432. More information at http://store.elsevier.com/product.jsp?isbn=9780128033487&pagename=search  

Well, Roger, it depends on what you are trying to determine, your training, and your facility with computers.  If one is a pharmacokineticist like yourself with all the tools, sure your approach will generate more data, but for the clinician w/o advanced training who is trying to determine if the drug is in the therpeutic window, then trough levels know where you are and how much latitude you have in dosing.

The question was, "During TDM, Is it possible that the concentration of Tacrolimus keep being high over a period of several days without no drug administration?"  That can be answered with Cmin and Cmax data.  That is what is done in bedisde PK.  Multiple samples means multiple analyses and more costs.  Sometimes it is not necessary to determine  "multiple model (IMM) sequential Bayesian analysis" in typical pts., but is useful in acutely ill unstable patients.  The method used and the number and timing of sampling should be designed to answer the question, not generate data that may not relate to the question.

Reminds me of the old definition of a pharmacologist, "Someone who shoots a drug into a patient and gets a paper out of it."

Well, David:

     Is that what you think of me, that I do stuff to get generate unrelated data and push papers out of it? That is something I have never done and never will! Cmax plus Cmin are already more than you had mentioned before. MM optimal approaches get the most informationeven from from optimally timed single or paired samples, often better than Cmax and Cmin, Sloppy approaches  mean poor care. Good software is now free to those who are interested - perhaps you. We try to plan things in advance for each patient to get the most information from the fewest samples, with mnimal cost, and to hit desired targets with maximum precision. This, i would suggest, is what good bedside pharmacokinetics is about and clearly NOT to generate useless data or push out papers. You might look at the recent book edited by Michael Neely and myself - : “Individualized Drug Therapy for Patients – Basic Techniques, Relevant Software, and Clinical Applications”. Imprint: Academic Press. Release Date: November 15, 2016. Print Book ISBN : 9780128033487. Pages: 432. More information at

http://store.elsevier.com/product.jsp?isbn=9780128033487&pagename=search

Good bedside pharmacokinetics is what this book is about. You might find it useful. The tools are available and free to all.  The software, for laptop computers, is free. Costs will be minimized, and patient care (therapeutic precision) will be optimized. Interested in improving the current state of bedside PK?

Very best regards,

Roger Jelliffe

No, Roger, that was not meant as a criticism of you or your work, just an observation.  I'm surprised you personalized it.

One of my objectives is to teach clinicians how to get data they can use in pt management with the least amount of effort and for the lowest price.  That's all.  There is a great elegance to simpliicity.  Cmin and Cmax are still simple and useful parameters to obtain with little effort and low cost, and can supply important therapeutic information.

The joke about the pharmacologist was just that, a joke.  Come on, lighten up a little.  Being a scientist no longer means sitting in a dark room like Edison waiting for the light to go on.  Today time is short and just about everyone has their hand out for $$, I'm just trying to suggest pathways that make it possible to improve therapeutics w/o going back to school for a PhD in pharmacokinetics.  Framing the upper and lower limits of the therapeutic window is the starting point, IMHO.

A few years ago, I spoke at the International Opioid Symposium and heard several MDs discuss taking ramdom blood levels to determine opioid levels.  They did not know when the last dose was given nor the amount of the dose, but they got a "blood level."  My comments were directed at the lesser trained, not a man with your expertise.  Also, I confess to not being what I call a "gadget person."   You may love putting data into a computer, but I do not.  Just trying to present some functional approaches to TDM that will provide more significant data than random sampling.

Strong CYP3A4 inhibitors may extend half-life enough one may only need to dose tacrolimus once weekly to obtain therapeutic drug levels. Ritonavir is one of those agents.

The TAC half-life is about 43h which is elevated but can be seen in clinical practice. Your concentrations decrease with time so this could be consistent with a drug drug interaction with a CYP3A4/5 inhibitor (macrolides, azoles, protease inhibitors...) and or an hepatic failure in this patients (do you have access to the hepatic enzymes?). In my view, the T1/2 is quite long for a CYP3A5*3 but in our center we have limited experience with homozygotous *3/*3 because they are very rare. So, this could be also considered whether the rate of *3 is important in your country. In brief, nothing so surprising.

Thank you for all the insightful and valuable discussions. About the patient in my database I cannot check the pharmacogenetics nor the co-administered drug that may help elucidate the reason behind this. No further information about this patient was available...

The only information I have is an increase of GGT (from 86 to 800) and PALC (from 60 to 109), AST decrease from 1711 to 51 and ALT decrease from 934 to 103 (I don't even have the unit). It's hepatic transplantation so dysfunction temporally of the transplanted liver can also be the cause.

Later in my database I observe 1 patient with relatively high concentration (19.2, 30, 27.4, 22.9, 21, 17.6, 13.4, 15.3, 13.4, 13.2, 11.2 8.4, 7.8) during a period of 12 days without any dose recorded. It was a small baby of 7 months and 6.3 kg, with hepatic transplantation. This time, I am not so surprised about that any more. I guess on some patient, the graft hepatic just needs sometimes to function stably.

Great admire for all the modellers that found a way to interpret the data!

if in the toxic range, first order kinetics may change to zero order kinetics, or indeed drug/drug interactions should be considered