A recent paper by Thabit describes a curious finding . The authors measured total and free (i.e., nonprotein-bound) eravacycline levels at ascending doses in a mouse model. They found strikingly small increases in free drug levels when titrating up total doses.
The effect was rather dramatic: an increase in protein binding from 12% to 93% (yes, ninety-three percent!) was observed. Obviously, higher doses of eravacycline resulted in proportionally higher degrees of protein binding, eventually reaching a plateau. In other words, an atypical non-linear exposure curve was seen.
This is a problematic feature; after all, one would like to achieve adequate free drug levels to penetrate tissues as well as reach and kill microbes that are not in the blood stream. Remember, it is free drug concentration and free AUC/MIC ratio that are key determinants for tetracycline efficacy.
Free drug concentration and free AUC/MIC ratio are key determinants for tetracycline efficacy
As a matter of fact, blood levels of tetracyclines are quite low even with IV dosing, and rhe same is true for eravacycline. PK data from a Phase 1 study conducted by Connors  confirms a rather flat concentration-time curve (see below):
Table: Eravacycline PK – adapted from Ref 
With a peak fCmax of only 0.185 mg/L, blood and tissue levels are clearly well below the MIC90 values of many pathogens, which for the most part fall into the 0.25-1.0 mg/L range . This makes clinicians uneasy who do not fully subscribe to the PK/PD mantra, which for tetracyclines tells us that Cmax is not the best predictor of efficacy and fAUC/MIC is all that matters.
Eravacycline’s favorable fAUC/MIC ratio is driven primarily by a long T½ in PK/PD calculations, so we hope that the infecting pathogens ‘understand’ this concept and ‘behave’ accordingly. The drug also has a rather large VD , which provides significant intracellular levels and may boost efficacy.
This is all well and good, but intracellular levels matter when the action is also intracellular (like for Neisseriae, Chlamydophila, Legionellae and a few others) but not for most other pathogens, including those causing pneumonia which reside in the extracellular space.
What is gnawing on us is that PK/PD analysis was wrong regarding the dose selection for the IGNITE2 (cUTI) study. This is the infamous Phase 3 study that failed to show non-inferiority to levofloxacin, the comparator drug. Such mishaps are not supposed to happen in this enlightened age. But – to be fair – nothing is ever just black or white, and dose selection still is a bit of an art form.
Could it be that PK/PD models favored higher doses, but the company still chose to take a calculated risk with a lower but per chance suboptimal cUTI dose?
We do hope that this was not the case here, but stranger things have happened before
Now back to protein binding. With a drug like eravacycline, the described effect on protein binding could have a negative impact on efficacy. Doses of 1 mg/kg q12h – infused over 1 hour – were used in IGNITE1, the pivotal cIAI trial and found to be efficacious. However, a slightly lower dose employed in the cUTI trial did not make the grade. This all points to a very delicate balancing act as far as dosing is concerned.
We would not go so far as to blame the recently failed cUTI trial on this phenomenon alone, although it may have been a contributing factor. Tetraphase certainly has not provided a satisfactory explanation for this.
This paradoxical effect on protein binding, so different from other antibiotic drug classes, has already been described in tigecycline and other tetracyclines, but it is an oddity, nonetheless ,. Are there other drugs that behave similarly? In our admittedly cursory search we came across none. This phenomenon may deserve further scrutiny.
Time for us to check if the PK profile of omadacycline has similarly atypical protein binding characteristics. Well, according to a recent poster, omadacycline behaves like a different animal: Villano et al. report concentration-independent protein binding of approx. 21%. This translates into 79% of unbound drug, much different from other tetracyclines which are usually highly protein-bound .
 A Thabit. Eravacycline Pharmacokinetics and Challenges in Deﬁning Humanized Exposure In Vivo. AAC 2016; 60: 5072
 G Zhanel. Review of Eravacycline, a Novel Fluorocycline Antibacterial Agent. Drugs 2016; 76:567
 K Connors. Phase I, open-label, safety and pharmacokinetic study to assess bronchopulmonary disposition of intravenous eravacycline in healthy men and women. AAC 2014;58:2113
 R Singh. Plasma Protein binding of eravacycline in mouse, rat, rabbit, cynomolgus monkey, African green monkey and human using microdialysis [abstract no. A-015 plus poster]. 53rd ICAAC; 201
 S Villano. In-vitro Protein Binding with Omadacycline, a First in Class Aminomethylcycline Antibiotic. ASM Microbe Poster 518, 2016