FluidSim Capabilities

What is FluidSim for and what does it do?

 In contrast to blood, which is continually flowing and mixing, the fluids in the inner ear do not flow and are unstirred.  So, when drug is applied, it doesn’t spread uniformly throughout the fluids. Instead, the drug slowly distributes along the fluid-filled spaces, a process that can take days. The rate of spread is so slow that if the drug leaks appreciably to the blood will never reach distant parts of the ear.


Therefore, for any drug being applied to the ear with a given methodology, we need to know how that drug distributes with distance and time. Distribution depends strongly on the physical properties of the drug. One of the primary tenets of pharmacotherapy is that the drug must reach the target tissue at a therapeutic concentration. In many cases this doesn’t happen for local drug therapies of the ear. Commonly the drug doesn’t reach the apical parts of the ear which are necessary to hear speech frequencies. For drug therapies of the cochlea, we therefore need to determine whether drug is reaching the targeted regions.

 Situations where FluidSim can contribute:

1) We are giving a drug systemically that is potentially ototoxic at concentrations above 1 ng/mL. How do we design safety and toxicology studies at minimum cost?

Before you start animal experiments, use FluidSim Simulations to help identify potential safety concerns and toxicity for the ear. Analysis can use the agent’s chemical characteristics to predict whether the therapeutic reaches the inner ear and in what concentration. Simulations provide a cost-effective way to design experiments with a minimum number of time points. They also help interpret blood and perilymph measurements, focusing on times and locations when drug concentrations exceed toxic levels, thereby saving animals and reducing costs.

 As data are collected, simulations are refined until the drug time courses in blood and perilymph are established. These data help guide the design of the toxicology screening study. A well-designed, efficient study helps minimize the costs involved.


 2) Will my small molecule therapy that works well in mice will work equally well in humans?

 No, not necessarily. The mouse cochlea is short (~ 5 mm) and the human is much longer (~30 mm). This means that drugs that distribute well and treat all frequency regions of the mouse cochlea do not necessarily distribute completely along the human cochlea, thereby leaving some regions untreated. Figure 3 shows dexamethasone (Dex) distribution along scala tympani of the mouse, guinea pig and human calculated by FluidSim for a sustained, 24 hour application.

Figure 3: Dexamethasone gradients along ST calculated by FluidSim for a sustained application of dexamethasone (such as by dexamethasone suspension in gel) for 24 hours. The elimination half-times of dexamethasone were 40 min for ST and 87 min for SV. The “glitches” near the basal part of ST are the result of perilymph-CSF exchange through the cochlear aqueduct. In each case the “flattening” of the curve towards the apex is due to exchange with scala vestibuli across the helicotrema.

 In the mouse, Dex is distributed along the entire length of the cochlea, with concentration at the apex just 2.5 x lower than at the base. In the guinea pig, the gradient is larger and is 35 x lower at the apex relative to the base. The gradient in the human is even larger, 136,000 x lower at the apex than the base, in part because a steady state has not been fully reached at 24 hours. Nevertheless, it is clear that apical regions of the human cochlea are treated with much lower levels of drug than basal regions.

It is also important to recognize that this is not an extreme example. How far drugs distribute along ST depends on how rapidly the substance is eliminated to blood. The half-time for Dex is estimated to be 40 min. The elimination rate of valproate, used in Frequency Therapeutics FX-322 formulation, was measured to be 19 min (McLean et al. 2021). The elimination rate of triamcinolone-acetonide was measured to be even faster, with a half time of 12 min (Salt et al. 2019). None of these drugs are suitable for treating the speech regions of the human cochlea. Figure 4 shows a figure from a corporate presentation by Frequency Therapeutics recognizing the limited spread of their therapeutic along the cochlea. The probably contributed to the failure of clinical trials with the therapy and the subsequent demise of the research program.

In conclusion, many drugs are not suitable for therapy of the speech regions of the human cochlea. Calculations with FluidSim can be instrumental in deciding whether a specific drug will distribute well, or not.

 Figure 4: Calculated influence of Frequency Therapeutics FX-322 formulation along the human cochlea (figure taken from one of their public corporate presentations). It is notable that most of the speech frequency regions (all frequencies below 6 kHz) were NOT treated by their formulation, because of the limited distribution of drug along the cochlea.


3) Is my potential drug candidate suitable for local therapy of the human ear?

 A lot of information is gained by calculating the physical properties of the drug, WLOGP (lipid partition coefficient) and TPSA (topological polar surface area). They can be calculated for any molecule on the SwissADME site (http://http://www.swissadme.ch). When WLOGP is plotted against TPSA it forms the so-called “egg plot”; Daina et al. 2017, providing a valuable indication of how the molecule will behave, as in Figure 5.

Figure 5: Left: As molecules become larger, more polar and less lipid-soluble (lower right of plot) they pass less easily through biological membranes. Molecules within the elliptical statistical boundaries pass through the blood brain barrier (yellow) or through the gut epithelium (white) respectively (Daina & Zoete, 2016; Daina et al. 2017). Right: Some example molecules relevant to the ear. Those in red are poorly suited to therapy of the cochlea, while those in green are retained in perilymph sufficiently to allow distribution towards apical cochlear regions. The red/green boundary is for illustrative purposes only and is not statistically based on data.

 As shown in the right panel of Figure 5, molecules within or close to the “egg yolk” (in the red shaded area) commonly pass through biological membranes and are rapidly eliminated from perilymph. When applied intratympanically, they are generally NOT SUITABLE for treating speech regions of the cochlea.

 A more detailed analysis of the molecule can be provided by analysis with FluidSim. If WLOGP and TPSA values for the molecule are entered on the “molecules” page FluidSim will then calculate the likely distribution of that molecule in the ear. FluidSim calculates rates of elimination from ST and SV are based on correlations established by measurement of elimination rates for 13 molecules in the Salt lab over the past 10 years (Figure 6). The fitted lines from these plots are used by FluidSim to predict the ST elimination rate, SV elimination rate and RWM permeability for each molecule, based on the “Egg Vector” derived from WLOGP and TPSA values. Middle ear elimination, by default is set to 49.6 min, which is the average of the 7 molecules for which middle ear elimination was measured (note that middle ear elimination can only be measured when drug is applied as a solution and cannot be measured when drug is applied as a suspension).


Figure 6: Left: Correlations between the rate of elimination from ST, rate of elimination from SV, RW membrane permeability and Middle Ear Elimination versus the distance the molecule lies from the center of the “egg yolk” on the egg plot. These plots are based on elimination and RW entry measurements for 13 molecules (some published; others proprietary). Elimination from the middle ear is relatively uniform across a wide range of molecule properties and is regarded as non-specific in nature (probably mediated by the lymphatic system).

Calculations by FluidSim for the specific delivery protocol used with the molecule allow an initial estimate of drug distribution to be made for the chosen species. Such calculations should be later supplemented with PK measurements for the molecule in an animal model.

 4) Is my potential drug candidate suitable for systemic therapy of the human ear?

 The “egg plot” analysis described above (the plot of WLOGP against TPSA) can be helpful when considering therapy of the ear by a drug given systemically. Molecules within the yellow ellipse (egg yolk) will typically pass through the blood-brain barrier (BBB) and enter the brain. The blood-labyrinth barrier (BLB) is comparable with the BBB, but may be somewhat less “tight”, allowing a greater range of molecules to enter as shown by the green ellipse in Figure 7.

It was shown in mice given fluorescein systemically that initial samples collected from the posterior semi-circular (predominantly perilymph) had higher fluorescein concentrations than later samples (predominantly CSF). This may indicate the BLB is more permeable, or may be explained by other mechanisms (perilymph volume smaller relative to blood flow, lower volume turnover of perilymph relative to CSF, etc). Whatever the mechanism, drug concentrations in perilymph may be higher than those in CSF, allowing “less permeable” molecules to reach therapeutic concentration.

Figure 7: The green ellipse represents the approximate range of molecular properties making the molecule useful for systemic therapy of the inner ear. The ellipse is arbitrarily larger than the brain entry ellipse and is not statistically based on measured data (since virtually no appropriate data are available).

 FluidSim can readily calculate entry into perilymph from the blood and can follow a changing blood time course. There is very little data available from which suitable parameters can be derived. The Salt lab fitted FluidSim simulations to measurements of perilymph concentration resulting from systemic fluorescein or systemic FITC-labeled dextran (4000 FW). The entry half-times were 5180 min, and 14,000 min respectively, which are far slower than the corresponding elimination times for these molecules, demonstrating the asymmetry of the BLB.

 While the egg-plot provides a rough guide to which molecules will be suitable for systemic therapy of the ear, FluidSim cannot predict entry parameters as there is insufficient data to base a prediction on. Nevertheless, with a limited amount of PK data where blood and perilymph concentrations were measured suitable parameters can be extracted by FluidSim and used to simulate the entire perilymph time course resulting from the therapy.


 5) We want to inject drug solution into the posterior semi-circular canal of the mouse. How much volume must be injected to fill the entire perilymph space with drug solution? How much difference does it make if the injection pipette isn’t properly sealed and we have a little fluid leakage at the injection site?


FluidSim excels at comparing injection conditions, such as different injection sites, different outlet locations, different injection rates and different volumes injected for different animal species.

The example in Figure 8 shows FluidSim calculations for posterior SCC injections at 0.1 uL/min in a mouse. The top left panel shows the distribution with distance through the mouse ear at different time points during the injection. At 2 min, the PSCC and vestibule are filled but scala tympani (distances up to ~4mm) only increases after 5 – 10 min. Loading the ear is substantially accomplished by a 10min injection, unless the concentration at the base of ST is critical. Drug is retained in the ear well after the injection stops.

Panels C and D of the Figure show the result when there is a slight leakage at the injection site, due to inadequate sealing. With a small leak, drug never reaches the cochlear apex with a 10 min injection and does not reach the basal half of ST even with the full 30 min injection. In addition, drug is rapidly washed out of perilymph as soon as the injection ceases (panel D).


Figure 8: Calculated drug distribution (A) and time courses (B) for a 30 min, 0.1 uL/min injection from a pipette sealed into the posterior SCC of the mouse. Perilymph is well-loaded with drug in about 10 min and drug is retained well in the ear after the injection is turned off. C and D: Similar calculations for the situation when the pipette is incompletely sealed and is allowing a small fluid leak of just 0.1 uL/min. (Without sealing, leakage can be >10x this rate in the mouse). Perilymph concentrations in the cochlea are substantially lower, especially in scala tympani (i.e. distances up to ~4 mm) and drug in the vestibule is rapidly washed away when injection is turned off. (adapted from Ohlemiller et al., 2021)

 FluidSim simulations therefore provide a rationale for how much drug volume needs to be injected to load specific regions of the inner ear.

 Mouse perilymph can be loaded more rapidly by injecting at a higher rate.


This figure shows the distribution of drug within the mouse ear for injection at 0.5 uL/min. At this rate the ear becomes well loaded within 3-4 minutes.

 Drug distribution depends on factors such as the rate of drug elimination from perilymph to blood, how well the pipette is sealed, and the injection rate. It also depends on the PK properties of the drug being used.

Ideally the FluidSim calculation should be set up to evaluate the specific conditions for the drug formulation being applied.

 6) The drug we are injecting into the SCC of the mouse is potentially toxic to the brain. How much of the injected drug will be pushed into CSF through the cochlear aqueduct during the injection?

 As scala tympani concentration rises during the injection, an increasing amount of drug will flow across the cochlear aqueduct into CSF, because  the aqueduct is providing the outlet for volume flow. This figure shows the total drug amounts in the cochlea and in CSF as calculated by FluidSim.


Amounts are calculated for an 8 min injection at 0.5 uL/min.

 As soon as the cochlea is loaded with drug (orange curve, 2-3 mins), the amount being driven into CSF starts rising. The longer the injection, more drug will be driven into CSF.

 This show why it is important to limit the injection rate and volume, if delivery into CSF can cause potential problems.

 FluidSim can calculate the amounts involved for you specific delivery conditions and taking into account the PK properties of the drug you are using.



Submit a question related to your own study interests. If an analysis with FluidSim would be helpful in your situation, we will add the example here. Email your question / example to This email address is being protected from spambots. You need JavaScript enabled to view it..

Alec N Salt, PhD

March 27, 2024



Toxicity Assessments for Hearing and Balance

 Before you start animal experiments measuring concentrations in the ear and their influence on hearing – use FluidSim Simulations.


Simulations provide a cost-effective way to design experiments with a minimum number of time points. They also help interpret blood and perilymph measurements, saving animals and reducing costs

Turner Scientific’s FluidSim program is the most sophisticated inner ear simulator presently available.

 As soon as you know the time course of the potentially toxic agent in blood, FluidSim can predict the time course for perilymph of the ear. The best time points for perilymph measurements can then be selected. FluidSim can be trained to replicate the amount of drug reaching the ear (by fitting to measured data), providing a complete time course and distribution of the agent within the ear.

 The inclusion of FluidSim analysis into the project allows toxicology experiments to be well-designed and efficient, minimizing the costs involved.


 The flow chart above shows how FluidSim simulations can guide the toxicity evaluation process. FluidSim allows the experimental design at every step to be optimized, minimizing the animal numbers and costs involved.

 FluidSim can be downloaded from the Turner Scientific website, allowing you to perform the calculations yourself.


  • The experienced staff at Turner Scientific can help guide you, providing FluidSim simulations and advice at each step of the process.


Contact Turner Scientific at This email address is being protected from spambots. You need JavaScript enabled to view it.


The Function of the Inner Ear

In 1861 Prosper Meniere fundamentally changed our understanding of the ear and was ostracized for it.

 This summary is based on the content of rare otolaryngology textbooks available in Dr. Richard Chole’s collection at Washington University and of other vintage texts available through Google Books Advanced Search. During the 19th century there were major advances in understanding of how the ear worked. Meniere's true contribution is often overlooked.

A modern understanding of the inner ear was summarized (Cummings, Otolaryngology Head and Neck Surgery, 5th Edition, 2010) as:  “The two important functions of the ear are hearing and balance. The portion of the inner ear that deals with hearing is the cochlea and the portion of the inner ear that deals with balance is collectively known as the vestibular organs (semicircular canals, utricle and saccule).”

But it wasn’t always this way.

In 1734, the consensus view was that the ear was entirely the “Hearing Organ”. Under this view, different parts of the ear performed different analyses of the sound.

The Semi-circular canals coded the directionality of sound and provided enhancement or attenuation of auditory stimuli

The Saccule and Utricle coded the hearing of noises

The Cochlea coded the hearing of tones


From Thomson 1734: (images of the original texts)The inner ear was believed to be air-filled. This was because cerebrospinal fluid pressure rapidly becomes negative after death. Disturbance of the stapes when you are trying to look inside the ear allows air to be rapidly drawn into the labyrinth as the perilymph is sucked into the cranium through the cochlear aqueduct. By the time you could look inside, it was air-filled.

Sound was thought to be transmitted to all the organs inside the ear. The entire labyrinth was regarded as “the immediate organ of hearing”.

  Thomson made the argument that there is no spiral cochlea in birds or fish, so the semicircular canals are part of the “immediate organ of hearing”. Furthermore the semicircular canals act in a way that the “impression of sounds increases and fortifies itself” as it passes through the ear.

 At this point, just consider what was seen when the stapes was removed and you looked through into the vestibule The below figure shows a 3D reconstruction of the human ear with the stapes, surrounding bone and vestibular endolymphatic structures removed. The inside of the structure is shown silver. Without the soft tissues of the saccular and utricular maculae you would see depressions in the bone underlying the structures when the stapes was removed. These were regarded as mechanical structures to amplify/focus the sound and were accordingly given the terms “fovea hemisherus” and “fovea semielliptica” (analogous to the fovea of the eye). With the stapes applying sounds directly above these structures it seemed logical that they were involved in the amplification of the sound.


 Now let us consider the "distasteful" work of Flourens  1817, as described in an 1884 textbook by Burnett. Flourens was the first to suggest that disruption of the semicircular canals caused “peculiar disturbances of equilibrium of the body”

However, Flourens experiments were performed in conscious, un-anesthetized animals and involved drilling holes into the ear or brain and in some cases pulling on the nerves. Damage to one SCC was observed to cause the head to snap in a specific direction.  Even then, they were regarded as unethical and were largely ignored by the medical and scientific communities.

In the Cyclopedia of Anatomy and Physiology 1839, edited by Todd, it is stated that the vestibule is part of the hearing mechanism 

Todd further dismisses the work of Flourens 

 An anatomic textbook of the same year (1839) by Wistar and Horner shows a mid-modiolar section of the human cochlea in which a membrane divides the spiral tube into two compartments. It should be appreciated that histology of the ear was in its infancy and the fibrous basilar membrane was the main structure that was observed within. At this time the sensory organ and sensory cells were not visible (subsequently reported by Corti in 1851) as was the membrane forming a boundary between the endolymph space and scala vestibuli (reported by Reissner in 1854).

From the text It is apparent that both the vestibular stuctures and the cochlea are still believed to be involved only in hearing.

This was a period of rapid discovery and in 1844 a text by Cruveilhier identifies the membranous labyrinth containing endolymph (attributed to  the French anatomist Breschet) within an outer compartment containing perilymph. The text emphasizes that there is no air in the labyrinth even though some anatomists (Ribes) still hold that view. The entire labyrinth remains the “organ of hearing”

 In 1861, just prior to his death in 1862, Prosper Menière published his paper attributing the dizziness of a patient to a pathology of the semicircular canals. From the above texts, this observation was presented to a medical and scientific community that believed the inner ear was only involved in hearing. Meniere's perspective was not well accepted.

From a translation by Politzer, 1902 

Although Meniere's conclusions were not initially accepted, some began to consider his observations. In 1869 Tröltsch presents Menière’s cases, with conjecture that “the seat of the disease” was the semicircular canals. Elsewhere in the text he also compares  Menière’s observations with those of Flourens.

 In 1882, Winslow shows some acceptance of the concept but suggests that vestibular structures “must be considered auditory in function, as well as special organs of the new sense of equilibrium”.

 By 1899 more detailed anatomical descriptions of the cochlea had made it to the textbooks, with the organ of Corti, Reissner’s membrane, inner hair cells, outer hair cells and stria vascularis all identified.

 One of the most reputable textbooks of the period, for which English translations are available, was assembled by Adam Politzer. In 1861 Politzer had spent a research fellowship in Paris, which involved visits to the  Institute for Deaf-Mutes where Menière was the “physician in residence”. He was well aware of Menière’s concepts. There were two editions of Politzer’s  texts, one in 1883 and a second in 1902. 

In the introductory paragraphs of both versions concerning the anatomical division of the ear, the labyrinth is included in the “sound-perceiving apparatus” with no mention of the involvement in equilibrium and balance. 

In the section relating to the physiology of the ear the 1883 version stated that the function of the vestibular ototliths was to damp the sound. By 1902, the vestibular apparatus was recognized, including the involvement in head motions and equilibrium. Furthermore the pioneering work of Flourens was now recognized.
Concerning Meniere’s disease, the 1883 edition presented the disease as a brief subsection on the topic of haemorrhages into the labyrinth. In contrast the 1902 version included a detailed, comprehensive 14 page discussion of Menière’s disease. This publication played an important role in bringing Menière’s work to the attention of “mainstream” ENT physicians of the period By 1915 most textbooks has incorporated the new thinking, that the vestibular system played a role in the equilibrium of the body.

But there were hold-outs. In “The Medical Record” in 1908, George Gould presented an 8 page diatribe against the entire concept of Menière’s disease and the “false and snapped-up theory that the labyrinth was the organ of equilibrium".


Prosper Meniere’s contribution to the field was much greater than characterizing the medical condition now known as Meniere’s disease. The case presented in the 1861 article was important because it was a clear description of a balance disorder resulting from pathology of a semi-circular canal, presented by a reputable physician. It suggested the semicircular canals were primarily involved in balance and not in hearing, strongly in contrast to the prevailing scientific consensus. It took over 40 years for the truth to become widely accepted by the scientific and medical communities. Challenging the concensus is often a difficult battle.

But – the story doesn’t end there. Remember the 2010 Cummings textbook in which it was stated “The portion of the inner ear that deals with hearing is the cochlea and the portion of the inner ear that deals with balance is collectively known as the vestibular organs (semicircular canals, utricle and saccule).” This documents the current consensus.

However, the saccule has been shown to respond to acoustic stimulation and is the basis of the clinical VEMP response used to test saccular function. Similarly, SCC, saccule and utricle receptors have been shown to respond to acoustic clicks at above 60 dB SL


It is therefore likely that the entire ear responds to air-borne acoustic stimulation, especially for loud sounds. Meniere was correct in associating the vestibular organs with balance but this does not exclude the possibility that vestibular organs could serve a dual purpose.

Texts in which it is assumed that all vestibular function is related to balance and all cochlear function is related to hearing may be oversimplifying the physiology of the ear. Current consensus may be oversimplified.

We finish with a timeline of some of the major contributors to our understanding of the inner ear. This period provided the foundation for what we know as the field of Otology today. The first otology journal, Archiv für Ohrenheilkunde (Archive of Otology) was published in Halle, Germany in 1864. The editors were Anton von Tröltsch, Hermann Schwartze (in Halle) and Adam Politzer in Wien. The journal subsequently became the European Archives of Oto-rhino-laryngology.


Miscellaneous Science-Related Notes

1) Scientific Consensus

 The media use the term "scientific consensus" to justify a certain point of view and to add support that it is "correct". The problem is that history has shown that the consensus view is often incorrect.

To quote Aaron Kheriaty, a fellow at the Ethics and Public Policy Center:

"Science is an ongoing search for truth & such truth has little to do with consensus. Every major scientific advance involves challenges to a consensus. Those who defend scientific consensus rather than specific experimental findings are not defending science but partisanship." 

Most of the discoveries that I made in my career studying the inner ear went against the prevailing scientific consensus. Some were initially dismissed, but over time all have now become generally accepted. If you look back at history, many major breakthroughs also went against the consensus of the time. (Galileo anyone? Also see our history page: the ear was initially thought to be concerned only with hearing and not with balance). The consensus view (otherwise known as the current dogma) is often incorrect. Today, discussion of consensus is a method used to shut down those with opposing views. It is mostly used by narrow-minded people who don't understand science and who have difficulty reconciling multiple viewpoints. The concept that "all scientists agree" only occurs when the scientists who disagree are censored, are prevented from publishing their work by reviewers holding the majority view, or who choose to stay silent (perhaps to keep their grant funding). This is destroying science.

The reluctance to consider results that disagree with the current dogma led Max Planck to state

"A scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die and a new generation grows up that is familiar with it."

One can only hope that those in the censorship-industrial-complex pushing false, supposedly consensus views hurry up and get on with their part. 

Dissenting views should be a normal part of any discourse

Artificial Intelligence (AI) will only make the situation worse. AI is generally based on the programmer's selection of "socially-acceptable" knowledge. 

As our society increasingly silences and ignores dissenting perspectives, "unexpected" outcomes (i.e. outcomes contrary to the consensus view) will occur more frequently. Critical thinking and skepticism are essential for good science.

There is a common misconception that drug solutions applied intratympanically (into the middle ear cavity) “stay around for a while”, giving time for the applied drug to spread into the inner ear. Some even think that delivery in this manner is relatively efficient, allowing most of the applied drug to enter inner ear perilymph over time.

Unfortunately, it’s not true. The middle ear delivery process is neither efficient or simple.

The middle ear is specialized to maintain its gas-filled state and has powerful homeostatic processes to remove fluids and foreign substances. The cells of its ventral surface are ciliated i.e. have long motile cilia which can push fluid along their surface. Especially when the patient or animal is conscious and upright, the cilia propel fluid volume towards the Eustachian tube. Each time the subject swallows, the tensor tympani muscle attached to the tympanic membrane contracts, pulling the tympanic membrane inwards, pressurizing the middle ear space and driving fluid down the Eustachian tube. As the tensor tympani relaxes, air from the pharynx is drawn back into the middle ear, replacing the expelled fluid with air. Over time the applied drug formulation is driven out of the middle ear to the pharynx and from there it is swallowed.

Even when the subject is anesthetized and supine, with no volume losses down the Eustachian tube, there are additional highly active processes that remove foreign substances.

Lim & Hussl (1975) described active pinocytosis of macromolecules by the bounding epithelium, delivering substances to the lymphatic system for drainage. Such a mechanism is consistent with the rapid loss of drug from solution in the middle ear in a non-specific manner (all drugs are lost at a similarly high rate). Drugs lost to the lymphatics eventually enter the vasculature at a nearby lymph node. Some of the middle ear drug loss also occurs directly to the vasculature. 

The figure below shows the time course of loss for some drugs from the anesthetized guinea pig middle ear (i.e. with no fluid loss to the Eustachian tube). It was quantified by taking samples from the round window niche at the time of perilymph sampling. Loss is relatively non-substance-specific. Even large molecules, such as dextran, rapidly decline in concentration after local applications in the form of a solution. Such a non-specific process is consistent with loss by pinocytosis.


The rate of elimination from the middle ear has a major influence on perilymph drug levels reached. Shown below is a calculation for gentamicin entry into perilymph by FluidSim using parameters appropriate for gentamicin. The normal middle ear elimination (MEE) half time is about 75 min (from the studies above). In the calculations we have changed elimination half-times by a factor of 2, higher and lower, altering the rate of decline as a function of time. Persistence time in the middle ear has an enormous influence on perilymph concentration, with a higher, later peak when elimination occurs more slowly (blue curves). All other calculation parameters were unchanged across the 3 conditions. Differences in perilymph concentration, time course and distribution along the cochlea (not shown) are a direct result of the different middle ear kinetics.


 A longer persistence of drug concentration may be provided by delivering drug as an undissolved suspension (which dissolves slowly over time) or as a timed release formulation from a substrate (polymers, particles, etc).

 Delivering the drug in a gel formulation (Poloxamer 407, hyaluronate, etc) does NOT slow the rate of decline appreciably. Gels do not act as timed-release agents. They may slow volume loss to the Eustachian tube, but they do not prevent losses to the lymphatic and vascular systems. Delivering the drug in a gel does not solve the problem. Also note that although the gel may remain in the middle ear for a few days, this does not mean that drug is also present in the gel. The drug rapidly diffuses out leaving a no-drug gel behind. Kinetics of the drug and the gel are governed by completely different kinetic processes. The presence of gel in the middle ear gives no indication of the amount of drug present.

 Why Middle Ear Concentration Measurements are necessary.

 It is virtually impossible to interpret perilymph drug concentration measurements without knowledge of the middle ear concentration time course. As middle ear concentration often declines rapidly with time, this dominates the kinetics measured in perilymph. The situation may be simplified when drug suspensions or timed-release formulations are applied, providing more sustained drug levels over time.

 Especially when comparing across species, similar kinetics of the middle ear cannot be assumed. In one study where the kinetics of two drugs were compared in guinea pigs and humans, it was found that CHIR99021 was lost from the middle ear guinea pigs with a 56.4 min halftime while VPA was lost with a 48.6 min halftime (McLean et al.,.2021). These rates are close to those found for other drugs, as shown in the plots above. Comparable measurements with samples from the human middle ear found that the losses were even faster, with halftimes of 10.6 min (CHIR99021) and 8.9 min (VPA) respectively fitting the measurements. This is probably a consequence of the highly aerated mastoid in the human, with higher surface area to volume ratio than found in the simple, open bulla of the guinea pig. Nevertheless, in cases where kinetics of the middle ear are substantially different, the difference will certainly have a major influence on perilymph levels achieved.

 Unfortunately, the lack of awareness of how powerful the homeostatic processes of the middle ear really are has resulted in very few measurements of middle ear kinetics. They are likely to be affected by many factors, such as the drug volume applied, the orientation of the head, whether the subject is conscious or anesthetized, and the presence of other components in the applied formulation. There are no detailed studies in which the most important variables have been identified and quantified.

 Our ability to interpret perilymph kinetics requires a better knowledge and understanding of middle ear kinetics. That can only be achieved when middle ear measurements are performed as part of perilymph kinetic studies.


Lim DJ, Hussl B. Macromolecular transport by the middle ear and its lymphatic system. Acta Otolaryngol. 1975; 80(1-2):19-31.

 McLean WJ, et al. Improved Speech Intelligibility in Subjects with Stable Sensorineural Hearing Loss Following Intratympanic Dosing of FX-322 in a Phase 1b Study, Otol Neurotol. 2021 42(7):e849-e857.