DETAILED DESCRIPTION OF THE INVENTION

[0028] Reference is made to FIG. 1, which shows a preferred embodiment of the invention. The drug to be delivered to the animal has been stored in reservoir 10. Reservoir 10 is a pre-pressurizing container that stores the drug at a desired mechanical pressure during the filling process. Typically, reservoir 10 has an inner liner for contacting the stored drug in a drug stability-preserving manner. The inner liner may be made from polymers acceptable for being in contact with pharmaceutical solutions for long periods of time or other materials having the required properties as are known in the art. The inner liner is typically in physical contact with an outer layer used as the structural element of reservoir 10. The outer layer may be made from metals such as steel, aluminum, or similar metal capable of preserving an internal pressure for several years. Alternatively, the structural layer may be made from a variety of polymers such as polyolefins including specifically particle filled polymers that are capable of storing the drug at the desired mechanical pressure for several years.

[0029] At time of use, flow tube 11 is connected to reservoir 10 in such a manner that the drug fluid in reservoir 10 may flow through flow tube 11 to the flow outlet (microneedle array) and into the animal to be treated. The microneedles (not shown) may be made of stainless steel as is conventional, or may be made of polysilicon or silicon, or may be preferably made of shape memory allow because of the very high elastic limit of these materials. Flow tube 11 is made of a flexible, drug compatible polymer such as Vialon™ that may be alternately pinched shut using pinch bar 14 and pinch stop 12 to stop the flow of the drug, or may be reopened to flow by withdrawing pinch bar 14 from pinch stop 12.

[0030] In order to regulate the rate of delivery of the drug into an animal such that the proper liquid volume, and hence the proper quantity of drug substance, is delivered at the proper time over the period when the drug delivery system is used, flow of the drug liquid down flow tube 11 must be regulated. This is accomplished with a calculated time sequence of pinching and unpinching of flow tube 11 using pinch bar 14 and pinch stop 12, thereby regulating the flow of the drug liquid through flow tube 11 as shown in FIG. 2. The actual rate of drug delivery is then the volumetric flow rate of the drug when flow tube 11 is open multiplied by the fraction of time flow is permitted during a cycle.

[0031] The calculated time sequence to accomplish the desired schedule of drug delivery has been stored in microprocessor 17. It includes a basic time interval or cycle that is continuously repeated. This calculated time sequence is illustrated in FIG. 2. The curves in FIG. 2 show the status of flow tube 11, open or closed, as a function of time.

[0032] Curve B in FIG. 2 shows one cycle, the shortest repeated increment in the calculated time sequence. The cycle begins at time t0 and concludes at time tc. Curve A in FIG. 2 illustrates the opening and closing of flow tube 11. When the cycle begins, pinch bar 14 has pinched flow tube 11 closed against pinch stop 12. At time t1, pinch bar 11 is moved away from pinch stop 12 by means of cam operating motor 15 and cam 18, opening flow tube 11. Later, at time t2, pinch bar 12 is moved back against pinch stop 14, closing flow tube 11. This pinching and unpinching action of flow tube 11 is controlled by microprocessor 17. At time t1, microprocessor 17 sends a signal to cam operating motor 15 to rotate cam 18 180 degrees. This rotation of cam 18 causes pinch bar 14 to move away from pinch stop 12, opening flow tube 11, and initiating flow of the drug liquid through flow tube 11. At time t2, microprocessor 17 sends another signal to cam operating motor 15, causing cam 18 to rotate another 180 degrees. This rotation of cam 18 causes pinch bar 14 to move against pinch stop 12, closing flow tube 11, and stopping flow of the drug liquid. Other methods of pinching and unpinching the flow tube as are known in the art, such as with a solenoid, or valve, may be used.

[0033] Curve C in FIG. 2 shows several cycles of the calculated time sequence during a period of time when the programmed drug delivery rate is constant. In this case, the time interval that flow is permitted each cycle (t1-t2) is the same. If a higher drug delivery rate is required, the fraction of time each cycle that flow tube 11 is open is increased. Similarly, if a lower drug delivery rate is required, the fraction of time flow tube 11 is open is decreased.

[0034] In an ideal world, factors affecting the actual rate of liquid flow down the flow tube such as the pressure in the reservoir, the viscosity of the drug fluid, and the inside diameter of the flow tube are constant. Thus the actual time that flow is permitted during the cycle would be directly calculated from the designed flow rate and the desired delivery rate as in equation (1) below. In an ideal world, this invention would not be needed.

[0035] In the real world, though, the reservoir pressure changes as the reservoir empties of the drug liquid. The drug liquid viscosity changes as the temperature of the liquid changes. The flow tube inside diameter also changes as temperature changes, and, if use of the system requires that the flow tube be changed, the inside diameter of the new tube is almost certainly different than that of the replaced tube. These and other changes can be sufficiently large to cause significant changes in the drug delivery rate. This is especially important in the case of insulin for the treatment of diabetes. It can also be true for Factor VIII for treating hemophiliacs, heparin for treating clotting disorders, theophyllin for treating asthma, morphine for treating severe pain, and other drugs that have a very low therapeutic index.

[0036] For a given use of this drug delivery device, such as the delivery of insulin to treat diabetes, a certain drug delivery rate will be desired at a particular time. Based on the design of the system for this application, that is, the design pressure in the reservoir, the design viscosity of the drug solution, the length and inside diameter of the flow tube, and the separation of heating block and heat sensor, the time required for the drug beneath the heating block to travel to the heat sensor can be calculated. In actual use, though, the time measured will usually be slightly different than this nominal design time, and in some cases will be significantly different. If sufficiently different, the user can be alerted to the problem and replace the drug reservoir component. If the differences are small, as they will be in most cases, the differences can be compensated for. If the measured time is shorter than the nominal time, this means that the rate of drug flow is higher. The fraction of the cycle that the flow tube is open can then be shortened by an appropriate amount such that the desired delivery is achieved. Similarly, if the measured time is longer than the nominal time, this means that the rate of drug flow is lower. The fraction of the cycle that the tube is open may then be lengthened by an appropriate amount such that the desired delivery is achieved.

[0037] To reduce or eliminate these real world variations, a metering means is incorporated into the drug delivery system, also shown in FIG. 1. As mentioned above, at the beginning of a cycle, flow tube 11 is closed. In a first preferred embodiment, a small pulse of electrical power from microprocessor 17 is sent to heating block 13, causing the temperature of the small amount of drug liquid beneath heating block 13 to be raised a small amount (this small amount of heat will not raise the temperature enough to alter the potency of the drug). Flow tube 11 is then unpinched, allowing the drug liquid to flow along flow tube 11. The small amount of warmer drug liquid quickly moves to heat sensor block 16 where it is sensed. A signal is sent from heat sensor 16 to microprocessor 17 where the time that the warmer liquid passed through heat sensor 16 is recorded. In this way, the time required for the fluid to flow the distance from the heating block 13 to heat sensor block 16 is measured. This time will be defined as TM.

[0038] In a second preferred embodiment, as shown in FIG. 3, the heating block is replaced with laser 45 focused on the flow channel 11 by lens 44 that will heat the liquid in the flow channel at location 48 through absorption of the laser radiation by the liquid. Preferred lasers are infrared lasers operating at wavelengths where the liquid is highly absorbing, or UV lasers where the liquid may also be absorbing. The flow tube 11 may be made of glass or of any engineering polymer such as polycarbonate which is transparent at the wavelength of laser 45.

[0039] The heated liquid moves downstream to location 42 where it passes through the sensing block, now shown to be comprised of laser 46, lens 47, lens 49 and detector 41. Preferred wavelengths for sensing laser 46 are visible or near infrared since very inexpensive, fast and sensitive detectors for use as detector 41 are available in these wavelength regions. The illumination from laser 46 is focused onto the flow channel at location 42 by lens 47. Because the heated liquid has a lower density than surrounding liquid due to its higher temperature, its refractive index is lower than the surrounding liquid. Because the refractive index is lower, this heated increment of liquid acts like a lens, redirecting some of the illumination from laser 46 focused at location 42.

[0040] In the absence of this heated increment of liquid, most of the illumination from laser 46 that is focused onto location 42 by lens 47 is refocused onto detector 41 by lens 49. When the heated increment of liquid passes through location 42, the illumination that is redirected by the lensing effect of the heated increment of liquid is not focused onto detector 42 by lens 49. Thus the intensity of the illumination reaching detector 41 is reduced, providing a signal that indicates the presence of the heated increment of liquid.

[0041] The thermal time of flight TM can be determined by recording the time that the laser heats the increment of liquid and the time that the detector detects the heated increment of liquid and calculating the time difference between these two times. Further, the velocity of the liquid in the flow channel can be determined by measuring the physical separation of location 48 where laser 45 heats the increment of liquid and location 42 where the passing of the heated increment of liquid is detected, and dividing by the “thermal time of flight”.

[0042] A third preferred embodiment of the heating means and sensing means is shown in FIG. 4. Grating 53 has been added to the heating laser optical system of FIG. 3. In FIG. 4, laser 55 now heats two increments of liquid at region 58 of the flow channel. Radiation from laser 55 is now focused onto flow tube 11 through lens 54 and grating 53. Because the periodicity of grating 53 can be precisely defined, the focal length of lens 54 is precisely known, and the wavelength of laser 55 is precisely known, the separation of the two heated increments of liquid in flow tube 11 at location 58 is precisely known. As these two heated increments of fluid move along flow tube 11, they pass, in sequence, location 60 where the pass through the focused illumination of laser 56. In the same way as described in preferred embodiment two, each of these heated increments of heated liquid in turn causes a portion of the illumination from laser 56 focused at location 60 by lens 57 to not be captured by detector 51. Hence the passing of each of these heated increments is detected by detector 51, and the time at which each of these heated increments passes location 60 can be recorded. The “thermal time of flight”, TM, is the time difference between these two recorded times. Since the spatial separation between the heated increments is known, the velocity of the liquid can be calculated as the spatial separation divided by the “Thermal time of flight”.

[0043] Yet another preferred embodiment of the heating and sensing means of this invention is shown in FIG. 5. As in the second preferred embodiment, shown in FIG. 3, an increment of liquid in flow tube 11 is heated by a laser. In FIG. 5, laser 75 is imaged onto an increment of the liquid flowing in flow tube 11 by lens 74. The wavelength of laser 75 is selected to efficiently heat the increment of liquid through absorption of the laser energy by the liquid. In this preferred embodiment, grating 73 has been introduced into the sensing optical path to cause two images of laser 76 to be created at region 80 along the flow tube by focusing lens 77. The heated increment of liquid first passes the first image of laser 76 at location 80, causing a portion of the illumination from laser 76 to be refracted out of the pupil of lens 79 and thereby not captured by detector 71. Further downstream, the heated increment of liquid passes the second image of laser 76 at location 80 causing a portion of the illumination from laser 76 to be refracted out of the pupil of lens 83 and thereby not captured by detector 82. In this embodiment, then, the “thermal time of flight” is the time interval between the detection of the passing of the heated liquid increment by detector 71 and the later detection of the same heated liquid increment by detector 82.

[0044] In each of these embodiments, the measurement of the “thermal time of flight,” TM, has been described. In the paragraphs below, the use of this “thermal time of flight” for near real time compensation of flow variation caused by changes in system parameters that cause variation in flow is described.

[0045] In general, the volume of fluid Q delivered from a pressurized reservoir to an outlet through a flow tube, where F is the flow through the tube in time T is dictated by the equation:

Q=FT (1)

[0046] In a specific device, such as the drug delivery system of this invention, that is designed to provide a nominal or designed flow rate, F0, the amount of fluid delivered, or discharge Q in time T can be calculated using:

Q=F0T (2)

[0047] If two points on the flow tube are selected, say a first point where a heater is placed to inject a small amount of heat into the increment of fluid at that point, and a second point where a heat sensor is placed to detect the presence of the heated increment of fluid when it passes, and the time that it takes the heated increment of fluid to flow from the first point to the second point is defined as T0, then the nominal discharge of fluid Q0 from the flow tube in time T0 can be calculated as:

Q0 =F0T0 (3)

[0048] Note that the physical significance of Q0 is the volume of fluid in the flow tube between the first and second points described above.

[0049] Further, if the device is designed with means to first stop fluid flow in the tube, for example, with the tube pinching means described above; to second energize the heating element to heat the increment of fluid beneath it; to third activate the fluid stop means to start fluid flow; to fourth monitor the heat sensor until it measures a fluid temperature rise indicative of the passage of the heated fluid increment; and fifth measure the time required for the heated fluid increment to flow from the heating element to the heat sensor, then, in the nominal or designed system, this time would be equal to T0.

[0050] Finally, if the desired flow discharge from the nominal or designed drug delivery system during a delivery cycle as described above is QD, then the time TD that the fluid is permitted to flow before being stopped by fluid stop means can be calculated as: 1

[0051] Thus it can be easily seen that for the nominal or designed system, the determination of the time TD required to permit flow to achieve the desired drug delivery rate, which, of course, can be varied at the discretion of the user or his physician at any time by simply changing QD, is easily calculated from designed system parameters.

[0052] In use, the actual parameters that govern flow through the flow tube, such as the viscosity of the fluid, or the pressure in the drug reservoir, which is known to decrease as fluid is removed from the reservoir, both of which are dependent upon the temperature of the fluid, will rarely be the same as the design parameters. More importantly, when the reservoir component of the system is interchanged with a new, filled reservoir, the inside diameter of the flow tube in the new reservoir component will not be the exactly the same as the inside diameter of the tube in the reservoir component being replaced. Since flow in the flow tube varies with the fourth power of flow tube inside diameter, this is an especially important and perhaps dominant determinant of the actual flow, and hence the actual delivery of drug. Thus it is critical to provide a drug delivery system that automatically compensates for the unknown and unpredictable variations in these parameters in order to insure accurate drug delivery.

[0053] The provision of such a drug delivery system is the main object of this invention. In the following paragraphs, the method by which the invention automatically compensates for these variations of use and manufacture is described.

[0054] Case 1. Pressure variations.

[0055] When the pressure in the reservoir is higher than nominal, the fluid in the flow tube will flow at a higher rate. Conversely, when the pressure in the reservoir is lower than nominal, the fluid will flow at a lower rate. When the fluid flows at a higher rate, the time required to permit flow will be shorter than the nominal time, and when the fluid flows at a lower rate, the time required to permit flow will be longer than the nominal time. The key will be to determine a new time, TD′ to permit fluid flow that provides for the desired drug delivery QD. Of course, it would be very easy in principle to include a pressure sensor on the reservoir component. However, since the reservoir component is intended to be disposable, this would add cost and complexity to a system component that needs to be as inexpensive as possible. The present invention avoids the need for such a pressure sensor. For ease of description, assume that the inside diameter of the flow tube is the nominal inside diameter such that the volume of fluid in the flow tube between the heater element and the heat sensor is Q0. If the pressure and viscosity are also at the nominal value, then the flow rate will be F0. According to the Poisieulle theory of laminar fluid flow in a tube (equation (9)), fluid flow rate varies linearly with a change in applied pressure. Thus, as the flow rate changes, the time that it takes the heated fluid increment under the heater element to move to the heat sensor will change to a new, measured value, TM in a linear fashion. Since the volume of the fluid, Q0, between the heater element and the heat sensor is unchanged, the new flow rate can be calculated as: 2

[0056] The new time, TD′, that fluid flow should be permitted to deliver the desired dose of drug QD to the patient under the new conditions of different pressure can be calculated as follows: 3

[0057] It can thus be seen that the only measurement needed to compensate for the change in pressure is the new time required for the heated increment of liquid to pass from the heater block to the sensor block, TM. The desired delivery, QD is specified since that is the delivery desired, and Q0 is known from the system design.

[0058] Case 2. Viscosity Variations.

[0059] The compensation of viscosity variations is essentially the same as for pressure variations. The only difference is that it is a temperature change that causes the viscosity of the drug solution to change—a viscosity change due to a drug solution formulation change is virtually impossible because of the product inspections required by the FDA. As temperature rises, the fluid becomes less viscous, and the flow rate increases. As temperature falls, the viscosity increases, and the flow rate decreases. As in the case of pressure change, for ease of description, assume that the volume of fluid in the flow tube is the nominal volume, Q0. Again, according to the Poisieulle theory of laminar fluid flow (equation (9)), the flow rate varies linearly but inversely with viscosity. As in the case for pressure, a new time for the heated fluid increment to flow from heating element to the heat sensor, TM is measured. A new flow rate, FM is calculated the same way using equation (5). And, the new time, TD′, required for fluid flow to deliver the desired dose of drug, QD, each cycle is calculated using equation (6).

[0060] Case 3. Combinations of Pressure and Viscosity Variation

[0061] The treatment is exactly the same as for either a pressure variation or a viscosity variation since the flow rate varies linearly with both pressure and viscosity. The system measures a new time, TM, for a fluid increment to flow from the heating element to the heat sensor. A new flow rate, FM, is calculated according to equation (5). And the new time, TD′, required for fluid flow to deliver the desired dose of drug, QD, is calculated from equation (6).

[0062] Case 4. Variations in Flow Tube Diameter

[0063] Variations in flow tube diameter can be compensated for in a manner similar to the method used for variations in pressure and viscosity. However, because the flow rate varies with the fourth power of the diameter, according to the Poisieulle theory of laminar flow (equation (9)), the compensation formulas are different. For the purpose of clarity of description, again assume that the pressure in the reservoir, P, is nominal, as is the fluid viscosity, ν. Further, let the total length of the flow tube equal L, and let the distance between the heating element and the heat sensor equal ½L. Let the diameter of a nominal flow tube be D1, and the diameter of a new replacement flow tube be D2. We then have the volume of the fluid in the nominal flow tube between the heating element and the heat sensor, Q0, and the similar volume of the fluid in the new replacement flow tube, Q0′, using straight geometry, as: 4

[0064] By Poisieulle's theory of laminar flow in tubes, we have the following equations for liquid flow rate in the nominal tube, F0, and for the liquid flow rate in the new replacement flow tube, F0′: 5

[0065] The nominal flow time between the heating element and heat sensor for a nominal flow tube, T0, is given by the ratio of Q0 and F0, and can be derived as follows: 6$\begin{array}{cc}{T}_0=\frac{16�\; ��\pi L^2}{PD_1^2}v& \left(11\right)\end{array}$

[0066] Similarly, the measured flow time between the heating element and the heat sensor in the new replacement flow tube, TM is given by the ration of Q0′ and F0′, and can be derived as follows: 7$\begin{array}{cc}{T}_M=\frac{16�\; ��\pi L^2}{PD_2^2}v& \left(12\right)\end{array}$

[0067] By dividing T0 by TM, that is, equation (11) by equation (12), the relationship between the nominal time, T0 and the measured time for the new replacement flow tube, TM can be discovered. After some algebraic manipulation, the following relationship is found: 8

[0068] Inserting equation (13) into equation (8) yields the following equation: 9

[0069] And since, from the basic flow equation 10

[0070] The new flow rate, F0′ can be shown to be 11$\begin{array}{cc}{F}_0^{\prime }=\frac{T0T_M^2}{}Q_{\mathrm{0<\; /mn>}}& (\mathrm{17<!--\; mn--><mo>)</mo>}\end{array}$

[0071] This is an important and unexpected result. The new flow rate for a new replacement tube can be expressed in terms of only nominal system parameters, T0 and Q0, and the new measured time TM. It is not necessary to know any of the physical properties of the new replacement flow tube (except that its total length, which can be accurately cut during manufacture).

[0072] From these results, the time, TD, required to permit fluid to flow down the flow tube to achieve the desired delivery of drug, QD, is 12$\begin{array}{cc}T_D=\frac{Q_D}{}\mathrm{And},\mathrm{finally},& \left(18\right)\\ T_D=\frac{Q_D}{Q_0<!--\; mo--><msub><mi>T</mi><mn>0</mn></msub>}& \left(19\right)\end{array}$

[0073] This is the result needed for correcting for a new replacement flow tube. Despite not knowing the diameter of the replacement flow tube, the correct time to permit fluid flow down this new tube to compensate for any differences in its diameter can be determined by simply measuring the time required for the increment of heated fluid to move from the heating element to the heat sensor and using equation (19).

[0074] One issue remains. Since pressure and viscosity compensations are linear with TM, and tube diameter compensations are quadratic with TM, an important issue is to know when to make either a linear or quadratic compensation. In reality, this is easily done. When a new replacement flow tube is placed in the system, this is the time to make the diameter compensation. At time of manufacture the pressure in a fresh reservoir can be set to nominal, and will remain at nominal over its shelf life. Temperature is easily and inexpensively measured (but even this may prove unnecessary since the product will be worn on the skin, and skin temperature is quite stable). Hence any differences between T0 and TM at time of replacement of the disposable component will be due to a change in flow tube diameter. With this new flow tube in use, the nominal time T0 can be reset to TM so that as the reservoir with the new flow tube is used, TM can be used as the new nominal flow time. As the system is used, the changes encountered will then only be due to viscosity and pressure changes that can be compensated using the linear correction.

[0075] In this manner, all of the expected variations that cause changes in fluid flow can be compensated, and accurate delivery of the desired dose of drug according to the stored schedule can be achieved.

[0076] Other embodiments to accomplish the invention may be known to those skilled in the art. For example, other methods of measuring the flow of the drug solution are known. One such method takes advantage of the fact that virtually all drugs solutions for administration to an animal are electrically conducting. If a magnetic field is placed perpendicular to the direction of drug flow, the induced flow of the ions in the drug solution results in a current flow in the direction of the magnetic field. Electrodes placed in appropriate positions can sense this current flow, which is directly proportional to actual volumetric flow rate. Alternatively, vanes or other mechanical devices can be placed in the flow path. When the liquid flows, the vanes will bend, resulting in a measure of flow rate.

[0077] Similarly, other methods of starting and stopping fluid flow are known. Valves may be placed in the flow path, and may be opened or closed as needed. Pressure on the medicament reservoir may be removed or replaced as needed.

[0078] While the above description of this automatic self-compensating liquid delivery system has been written in the context of a drug delivery system, the basic principles may be applicable elsewhere, for example, in an automobile carburetion system, in a paint dispensing system, in a gasoline station dispensing system, or any other system wherein an accurate and predictable liquid flow is required in a system where flow parameters may change.

[0079] The examples and embodiments described herein serve only to teach the invention and in no way serve to limit the scope of the invention. The scope is only limited by the following claims: