Technology, Computing, and Simulation

Section Editor: Maxime Cannesson

E NARRATIVE REVIEW ARTICLE
Drug Infusion Systems: Technologies, Performance,
and Pitfalls
Uoo R. Kim, MD,* Robert A. Peterfreund, MD, PhD,† and Mark A. Lovich, MD, PhD*
Downloaded from http://journals.lww.com/anesthesia-analgesia by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsI

This review aims to broadly describe drug infusion technologies and raise subtle but important
Ho4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdtwnfKZBYtws= on 02/19/2024

issues arising from infusion therapy that can potentially lead to patient instability and morbidity.
Advantages and disadvantages of gravity-dependent drug infusion are described and compared
with electromechanical approaches for precise control of medication infusion, including large-
volume peristaltic and syringe pumps. This review discusses how drugs and inert carriers interact
within infusion systems and outlines several complexities and potential sources of drug error.
Major topics are (1) the importance of the infusion system dead volume; (2) the quantities of
coadministered fluid and the concept of microinfusion; and (3) future directions for drug infusion.
The infusion system dead volume resides between the point where drug and inert carrier
streams meet and the patient’s blood. The dead volume is an often forgotten reservoir of drugs,
especially when infusion flows slow or stop. Even with medications and carriers flowing, some
mass of drug always resides within the dead volume. This reservoir of drug can be accidentally
delivered into patients. When dose rate is changed, there can be a significant lag between
intended and actual drug delivery. When a drug infusion is discontinued, drug delivery continues
until the dead volume is fully cleared of residual drug by the carrier. When multiple drug infusions
flow together, a change in any drug flow rate transiently affects the rate of delivery of all the oth-
ers. For all of these reasons, the use of drug infusion systems with smaller dead volumes may
be advantageous.
For critically ill patients requiring multiple infusions, the obligate amount of administered fluid
can contribute to volume overload. Recognition of the risk of overload has given rise to micro-
infusion strategies wherein drug solutions are highly concentrated and infused at low rates.
However, potential risks associated with the dead volume may be magnified with microinfusion.
All of these potential sources for adverse events relating to the infusion system dead volume
illustrate the need for continuing education of clinical personnel in the complexities of drug
delivery by infusion.
This review concludes with an outline of future technologies for managing drug delivery by con-
tinuous infusion. Automated systems based on physiologic signals and smart systems based
on physical principles and an understanding of dead volume may mitigate against adverse
patient events and clinical errors in the complex process of drug delivery by infusion.  (Anesth
Analg 2017;124:1493–505)

T
he operating room, intensive care unit (ICU), inter-
ventional radiology suite, cardiac catheterization lab-
oratory, and ancillary procedural sites are operated
and managed by intensivists, nurses, and anesthesiologists.
Clinical care often requires continuous infusion of vasoac-
tive, positive and negative inotropic, analgesic, and seda-
tive medications to critically ill patients. As a result of their
profound physiologic effects, relatively short half-lives, and
high initial concentrations of these infused medications,
combined with dynamic clinical circumstances of many
of these patients, infusion therapy can be extraordinarily
From the *Department of Anesthesia, Critical Care and Pain Medicine, St.
Elizabeth’s Medical Center, Boston, Massachusetts; and †Department of An-
esthesia, Critical Care and Pain Medicine, Massachusetts General Hospital,
Boston, Massachusetts.
Accepted for publication September 26, 2016.
Funding: None.
The authors declare no conflicts of interest.
Reprints will not be available from the authors.
Address correspondence to Mark A. Lovich, MD, PhD, Department of An-
esthesia, Critical Care and Pain Medicine, St. Elizabeth’s Medical Center, 736
Cambridge St, Boston, MA 02135. Address e-mail to mark.lovich@steward.org.
Copyright © 2017 International Anesthesia Research Society
DOI: 10.1213/ANE.0000000000001707
complex. Anecdotal evidence, albeit much of it unpub-
lished, suggests that the complexity of drug infusions can
lead to patient morbidity and mortality. Indeed, morbid-
ity and mortality conferences year after year are filled with
examples of harm to patients or near misses that directly
arise from the complexity of drug infusion and the manner
in which human users interact with the technology. This
review aims to broadly describe various options for drug
infusion devices and technologies; it is not intended to pro-
vide an exhaustive list of all available technologies or mate-
rials. It is our intent to outline subtle but important issues
that arise with infusion therapy that can potentially lead to
patient instability and morbidity.
VERY BRIEF HISTORY OF INTRAVENOUS
INFUSION
A complete history of IV infusion is beyond the scope of this
review. IV infusion technologies advanced in the wake of the
need for resuscitation. Ancient texts allude to infusing sub-
stances for therapeutic benefit that include “solutions” and
“donated” blood.1,2 The first reported instance of IV infusion
was in 1492, when Pope Innocent VIII was given a blood trans-
fusion by vein–vein anastomosis from several live donors
after a stroke, but unfortunately all subjects involved died.2,3

May 2017 • Volume 124 • Number 5 www.anesthesia-analgesia.org 1493

Copyright © 2017 International Anesthesia Research Society. Unauthorized reproduction of this article is prohibited.
 E NARRATIVE Review article
In 1616, William Harvey completed the description of the cir-
culatory system in “De moto cordis” started by predecessors
that provided a rationale to attempt drug infusion. In 1656, Sir
Christopher Wren fashioned the first documented drug infu-
sion device using a quill as an IV catheter and a pig bladder
as a reservoir, giving IV injections of wine, ale, opium, and
liver of antimony into his dog’s vein. In 1662, Johann D. Major
injected an unpurified compound into the vein of a man; how-
Downloaded from http://journals.lww.com/anesthesia-analgesia by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsI
ever, the outcome was suboptimal. In 1665, Richard Lower
Ho4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdtwnfKZBYtws= on 02/19/2024
achieved the first successful blood transfusion in animals.
This was followed by the transfusion of blood from a sheep
into a human by Jean Baptiste in 1667, and shortly thereaf-
ter, the first documented transfusion reaction, death from
transfusion, and trial for manslaughter allegedly caused by
transfusion. Given these failures, blood transfusion was dis-
couraged or banned in some European countries. Transfusion
was not attempted again until 1795 when Philip Syng Physick
(credited as the father of modern surgery) suggested human-
to-human transfusion and likely attempted it in an unpub-
lished series of patients. The first successful human-to-human
transfusion was performed by James Blundell in 1830 to save
a woman with postpartum hemorrhage.4 The cholera epidem-
ics in the 1830s motivated physicians to advance IV resuscita-
tion, initially unsuccessfully with water and then successfully
in 1833 when Thomas Latta infused dilute saline to transiently
save a patient near death.5
All of these advances in transfusion and resuscitative
therapies were coupled to advances in drug infusion tech-
nologies, including the hollow-bore metallic needle by the
mid-19th century and adoption of antiseptic techniques in
the late 19th century. In 1930, the first plastic IV catheter
was demonstrated. In 1960, central line cannulation was
first described. Since then, there have been a multitude of
advances in catheters, IV tubing systems, connectors, reser-
voirs, and fluid propulsion devices.
The Present Era
Modern drug infusion systems have common basic elements:
reservoir, tubing, propulsion mechanisms, flow controllers,
and an IV catheter.6 Drug dissolved in solutions start in some
sort of reservoir, a bottle, plastic bag, or syringe. All drug
infusions end in an IV catheter designed for insertion into
a peripheral or central vein. (This review does not discuss
intraosseous or subcutaneous infusion.) Tubing systems con-
nect the reservoir to the catheter. Every infusion system needs
a method of propelling fluid. The simplest, least capital-inten-
sive, but perhaps least precise are gravity-driven infusions.
For gravity-driven infusions, a flow control mechanism such
as a roller clamp on the tubing regulates the rate of fluid or
drug administration. On the other end of the spectrum, elec-
tromechanical syringe pumps and peristaltic infusion pumps
are designed to deliver fluid with reliable precision.
Most infusion systems have built-in ports for merg-
ing other drug flows or inert crystalloids into a common
fluid pathway. Back check valves on these ports are often
included to prevent retrograde flow into the tubing or res-
ervoirs of other medications. A multitude of connectors,
stopcocks, and manifolds can be inserted between the tub-
ing and the catheter to make connections for additional
drugs before flowing into a common fluid pathway. A very
1494   
www.anesthesia-analgesia.org
Copyright © 2017 International Anesthesia Research Society. Unauthorized reproduction of this article is prohibited.
common arrangement is to establish a primary infusion of
inert crystalloid carrier into a catheter port or peripheral IV
catheter and then to “piggyback” all other medications into
the main line as secondary infusions. This inert flow can be
designed for volume resuscitation or can be modified for
more precise control to propel all coinfused medications.
Gravity-Driven Infusion
Before the development and adoption of electromechanical
pumps, most drug infusions were propelled by gravity. Drug
infusions today may still be given by gravity, especially in
environments with fewer resources available. Specialized
clear tubing systems allow direct visualization for rough
approximation of flow rates. Such tubing systems have a drip
chamber in which fluid from the reservoir exits a narrow hol-
low tube with the orifice pointing downward. The droplets
of fluid are presumed to be a fixed volume. Two models are
widely available, either 15-drop/mL or 60-drop/mL tubing
(Figure 1). In the former, each drop is approximately 1/15th
of 1 mL and in the latter each drop is approximately 1/60th of
1 mL. The drop size of 1/60th of 1 mL is convenient, because
the flow rate in mL/h is equal to the drops per minute.
Figure 1. Drip chambers within the IV tubing set allows direct visu-
alization of the fluid droplets for rough approximation of flow rates
by counting the drops per unit time. A, Macrodrip chamber with com-
mon IV tubing sets delivering 10, 15, or 20 drops (gtt) per milliliter.
B, Microdrip chamber with the most widely used IV tubing set deliver-
ing 60 drops (gtt) per milliliter.
The flow rate of fluid in gravity-driven systems is
often controlled by a simple roller clamp. This type of
clamp has a nonlinear response to roller position, that is,
the resistance of the tubing, and therefore, the flow rate
from a bag suspended at a fixed height above the patient
has little effect at the extremes of the roller excursion
but changes abruptly at intermediate positions. This can
make selecting an exact flow rate imprecise.7 The preci-
sion of counting the drops to determine the flow rate is
predicated on each drop being exactly 1/60th or 1/15th
of 1 mL. There is a relationship, however, between drop-
let radius r and surface tension (γ ) of the fluid8:
3γ d t
r= 3
(1)
4ρ g 
ANESTHESIA & ANALGESIA
 Drug Infusion Systems

(ρ is the fluid density, g is the gravitational constant, and
d t is the diameter is the dropper tube). It is reasonable to
suspect that the droplet size may vary with molecular prop-
erties of drug solutions, pharmaceutical formulations, and
additives.9,10 Temperature plays a role in flow variability as
a result of its effect on surface tension that affects the droplet
size.11 Manufacturer differences between drip chamber tip
diameters of disposable IV sets may also account for varia-
Downloaded from http://journals.lww.com/anesthesia-analgesia by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsI
position by limiting the effects of creep.15 However, the lack
of accuracy and flow rate variations between brands may
limit clinical use.15,16
The main advantages of gravity-driven drug delivery is
the low cost and simplicity of use (Table 1) and an instan-
taneous drug flow response when flow controllers are
opened, which is why this propulsion method is still widely
used today. In addition, the pressure of fluid emitted from

tions in droplet size.12 Variations in flow over time may be the catheter is limited to what is imposed by the height of
Ho4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdtwnfKZBYtws= on 02/19/2024

the result of creep, in which the plastic tubing distorts, alter-
ing resistance long after the roller clamp is tightened or
relaxed, thus requiring periodic adjustment.13,14 These fac-
tors can make control and quantification of flow rate impre-
cise.7 Available technologies include manual flow regulators
(Figure 2) such as the Dial-A-Flo™ flow controller (Hospira,
Lake Forest, IL), which provide the ability to make finer
changes in flow resistance compared with roller clamps,
and may exhibit a more linear flow response to controller
Figure 2. Manual flow controllers allow finer changes in flow resis-
tance compared with roller clamps, such that resistance changes
linearly with controller position.
the reservoir. Thus, in theory, there is less likelihood of tis-
sue damage from an infiltrated IV compared with pressur-
ized sleeves for fluid bags (Figure 3A) or mechanical pumps
that may continue to infuse at higher pressures.
A limitation of gravity-driven delivery systems is that they
are subjected to the gravimetric head, the distance between
the height of the fluid line in the drip chamber and the patient,
and changes in downstream pressures. Changing the height
of the bag reservoir or repositioning the patient changes the
gravimetric head and, thus, the flow rate.17 The flow control
device (roller clamp or manual flow regulator) needs to be
adjusted to compensate. Because the flow rate is measured
by counting drops in a time interval, this is a slow, cumber-
some process and may take several minutes to complete.
Furthermore, as the central venous pressure of the patient
changes, the downstream backpressure alters the flow of
medications.11,18,19 Similarly, if the medication flow is into a
peripheral IV in the antecubital vein, simple flexion of the
elbow can abruptly slow gravity-driven infusion. The changes
in position of other joints can influence IV performance when
the catheter is inserted near those joint. Size of the IV cannula,
the target vein, temperature, viscosity, and density of fluid

Table 1.  Comparison of Gravity-Driven Infusion With Both Large-Volume and Syringe Pumps
Large-Volume Peristaltic and Cassette
Gravity-Driven Flow Pumps Syringe Pumps
Infusion category Macroinfusion Macroinfusion or microinfusion Microinfusion
Ease of use – Simple, fast setup – Pump programming required – Pump programming required
– Requires assembling multiple – Requires assembling multiple
components components
Complexity – Simple tubing/infusion set – Proprietary specialized tubing – Simple syringe (syringe
– No special equipment or and components specification must be in pump
electricity required library)
– Low compliance tubing
Cost – Low cost – Expensive proprietary components – Generic syringe
– Requires maintenance – Requires maintenance
Drug libraries No Yes Yes
Sensitivity to back pressure or CVP Yes No No
Sensitivity to gravitational head Yes No Noa
Precision Nob Yes Yes
Possible free flow error Yesc Nod Noe
Programming error possible No Yes Yes
Air detection Nof Yes No
Back pressure/occlusion alarm No Yes Yes
Alert when reservoir is near empty? No Yes Yes
Duration of transition to new Short Short but requires reprogramming Potentially long with possible drug
reservoir VTBI flow interruption
Abbreviations: CVP, central venous pressure; VTBI, volume-to-be-infused.
a
Flow rate may be transiently affected by vertical displacement of the pump.36,37,40,42
b
Drop counters (Figure 4) and manual flow controllers (Figure 2) may assist in more precisely regulating flow rates.15,21
c
Inadvertently leaving the clamp open can bolus drug.
d
Most inserts have integrated safety clamp fitments that automatically clamp tubing when removed from the pump (Figure 7).
e
Leaving the clamp open with the plunger disengaged may cause siphoning into or out of the syringe, leading to bolus of drug if elevated above the patient or
possible dilution if positioned below the patient.41
f
Unless used with dedicated air detectors.

May 2017 • Volume 124 • Number 5 www.anesthesia-analgesia.org 1495

Copyright © 2017 International Anesthesia Research Society. Unauthorized reproduction of this article is prohibited.
 E NARRATIVE Review article
Downloaded from http://journals.lww.com/anesthesia-analgesia by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsI
Ho4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdtwnfKZBYtws= on 02/19/2024
Figure 3. A, Pressure bags may be used for rapid infusion of blood
products or fluid. B, In the pediatric population, burettes are used to
better control fluid administration. Many burettes have a ball valve at
the bottom that prevents air from entering the infusion line.
all affect the rate of infusion.20 Finally, the composition of the
delivered fluid can affect the flow rate. For example, human
serum contains a factor that causes sustained venoconstric-
tion with resulting slower flow compared with whole blood,
although whole blood has over double the viscosity.18
In ICU and operating room settings, multiple medica-
tions are often infused via a single fluid pathway, usually as
a result of the limited IV access and the multitude of infused
medications. Changes in the flow rate of any of the medi-
cations or the inert carrier will change the pressure at the
manifold and alter the flow rates of all the others. Clearly,
much time is consumed iteratively adjusting all the flow
controllers. Given the short half-lives and rapid onset of
many vasoactive and inotropic compounds and the time
required to adjust all of the flow controllers, much potential
for iatrogenic hemodynamic instability exists.
Figure 4. A, Drop counters utilize an optical sensor that counts the drops in the drip chamber to calculate the rate of flow. B, Integration of
drop detectors into IV sets that can control an adjustable clamp give the ability to set gravity-driven flow to a desired rate. This may aid in more
accurate administration of fluid and medications, including compensation for changes in gravimetric head.
1496   
www.anesthesia-analgesia.org
Copyright © 2017 International Anesthesia Research Society. Unauthorized reproduction of this article is prohibited.
Gravity-driven drug delivery systems can be subjected
to human error. Roller clamps can be inadvertently left open
leading to drug or fluid overdose. For pediatric patients,
multichambered tubing systems, known as burettes, such as
the Buretrol (Baxter Healthcare, Deerfield, IL) and Soluset
(Hospira, Lake Forest, IL) are available to release fluid in mea-
sured stages and thus prevent drug or fluid overdose (Figure
3B). Another human error associated with gravity-driven
flows is the failure to recognize an empty reservoir, because
the simple nonelectronic system is unable to generate audi-
ble alarms. Not only would drug delivery cease, but air may
enter the tubing possibly resulting in a potentially dangerous
venous air embolism. Some of these dangers may be miti-
gated through newer technologies. Drop counters such as the
DripAssist™ have an optical sensor that quantifies the drop
rate in the drip chamber and therefore the flow rate (Figure
4A). In addition, devices such as the AutoClamp™ integrate
a drop counter (Ace Medical, Seoul, South Korea) with an
automated adjustable flow rate controller (Figure 4B).21 Such
devices can compensate for changes in reservoir height or
changes in downstream pressure and circumvent some of the
limitations of gravity-driven systems.
Electromechanical Pumps
The first medical use of an electromechanical device to pump
fluid was in 1951, when Sigmamotor, Inc. developed a linear
peristaltic pump used in cross-circulation open heart sur-
gery.22,23 This technology was then adapted for drug infusion
in 1961. The precision provided by electromechanical pumps
allowed development of total parenteral nutrition and IV
chemotherapies.24 Successive iterations introduced numer-
ous safety and ergonomic features. The principal advantages
of mechanical pumping over gravity-driven drug adminis-
tration include improved accuracy and precision in volume
and infusion rate. Mechanical pumps can also overcome
small changes in resistance to flow or back pressure. The flow
rate from these devices is less subject to physiologic changes
in downstream pressure, perturbations from adjustment of
other medications or inert carriers, and changes in patient or
reservoir positioning than is gravity-driven drug delivery.
ANESTHESIA & ANALGESIA
 Drug Infusion Systems
Some electromechanical pumps have utilized drop counters
to ensure accurate flow rate.24 These features make reposi-
tioning patients on operating room tables and ICU beds, and
also transport between care units, more reliable and safer.
Perhaps the most important and useful safety feature
added to electromechanical infusion pumps is software that
incorporates drug libraries that include “guard rails” for dos-
ing that are programmable by local hospital pharmacies with
Downloaded from http://journals.lww.com/anesthesia-analgesia by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsI
upper and lower limits for the rate of infusion.25 This may min-
Ho4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdtwnfKZBYtws= on 02/19/2024
imize human errors in programming the pumps. However,
it is important to note that these drug dose limits are “soft”
alerts that may be overridden by the user. Clinicians may also
choose generic infusion modes where the rate is solely deter-
mined by the clinician without soft alerts. In addition, pres-
sure alarms can detect resistance (such as kinks in tubing) and
possible obstruction in the infusion system from catheter infil-
tration, thrombus, or blockage from drug precipitation. Air
detectors have the capability of sensing bubbles as small as
0.1 mL of volume. Volume-to-be-infused settings reduce the
potential for air aspiration and alert the clinician when reser-
voirs are near empty, thus minimizing the potential for abrupt
cessation of drug flow. Some infusion pumps are capable of
reading bar-coded medications and may decrease the likeli-
hood of “incorrect drug” errors. Certain pain/patient-con-
trolled analgesia pumps are passcode-protected, so that the
settings are not changed without physician or nurse consent.
Although electromechanical devices provide quantita-
tive and reliable drug infusion, they do so at considerable
cost, which includes purchase, maintenance, and special-
ized disposable tubing systems. They are not free from
potential sources of error. Even with drug libraries, pro-
gramming errors by clinicians are still possible.26 User inter-
faces may be unintuitive and the multitude of alarms and
safety mechanisms can be complex. When using low flows,
there may be significant delays in alerting clinicians that the
catheter or tubing is occluded, and thus, critical medications
may not be administered in a timely manner.27–29 Excessive
infusion line compliance may increase the time until the
pump alarms from an occlusion and also creates a poten-
tial drug bolus upon occlusion release.30 Therefore, infusion
tubing should be made from stiff noncompliant materials.
Mechanical infusion pumps are classified by the res-
ervoir and mechanism of action. Syringe pumps deliver
steady flow at low rates but have a limited reservoir vol-
ume, typically ≤60 mL. Conversely, volumetric pumps have
larger reservoirs, often up to 1 L. Peristaltic and cassette
pumps are volumetric pumps that utilize different drive
Figure 5. Syringe pump schematic: a rotary step-
per motor is coupled to a carriage by way of a
drive screw, which turns rotation of the motor into
linear excursion. The plunger plate of the syringe
is secured to the carriage through a fitting mount
that slides along a guide. A clutch mechanism
decouples the carriage from the drive screw to
allow manual movement of the carriage for inser-
tion and removal of syringes.
May 2017 • Volume 124 • Number 5
Copyright © 2017 International Anesthesia Research Society. Unauthorized reproduction of this article is prohibited.
mechanisms whose advantages and disadvantages are dis-
cussed below.
Syringe Pumps
Syringe pumps use a rotary stepper motor that turns a drive
screw coupled to a carriage such that rotation is converted
to linear excursion. The carriage secures the plunger plate
of a syringe with a fitting mount and slides along a guide,
which drives the plunger forward (Figure 5). A mechani-
cal clutch mechanism allows the carriage to be decoupled
from the drive screw so that the carriage can be extended
backward for syringe insertion. The carriage is then moved
in approximation to the plunger flange of the syringe before
releasing the clutch. Modern clinical-grade syringe pumps
have a software library of syringe inner diameters from dif-
ferent manufacturers. The user selects the manufacturer and
size, so the software utilizes the relationship between rota-
tional movement of the drive screw, the linear movement
of the plunger, and volume flow rate of fluid. Because these
pumps can accommodate standard syringes, the dispos-
ables are potentially less expensive than the cartridge–tub-
ing combinations used in peristaltic or cassette-type pumps.
As a result of the mechanics of the stepper motor that
turns the drive screw, this pump delivers fluids in small,
discrete drug volumes, which at higher flows appears to be
continuous but at ultralow flows may appear intermittent.
This is especially important in the pediatric population in
which medications are often administered at low-flow rates
with the syringe plunger moving incrementally with each
step in drive screw rotation, causing intermittent regular
small boluses. In the case of inotropic drugs such as dopa-
mine or epinephrine, this may manifest as regular repeated
hemodynamic fluctuations.31,32 In addition, the friction in
the plunger head on the wall of the syringe barrel may cause
it to intermittently stick (“stiction”), with drug delivery and
hence hemodynamic implications.31 Recently, the U.S. Food
and Drug Administration released a safety communication
report outlining concerns of flow discontinuity at low rates
that may result in dosing variability, potentially leading to
adverse clinical outcomes.33 To minimize flow rate variabil-
ity, some recommend using the highest possible flow and
therefore lowest concentration.34
There are multiple sources of delays and inaccura-
cies inherent in syringe pump systems, many of which
occur at the time of first starting an infusion.35 On user
initiation of flow, the stepper motor turns to translate
the carriage assembly. A delay in drug flow may result
www.anesthesia-analgesia.org 1497
 E NARRATIVE Review article
if the carriage assembly is not snuggly applied to the
plunger of the syringe. Syringe pump design also affects
the start-up delay times.36 Various brands of disposable
syringes may differ in the thickness of the plunger flange.
Pump manufacturers have introduced gaps in the fitting
mount to accommodate this variable thickness, which
may lead to free play between the fitting mount and the
plunger flange. This gap has been shown to cause an ini-
Downloaded from http://journals.lww.com/anesthesia-analgesia by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsI
tial start-up delay between 7 minutes and nearly 1 hour,
Ho4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdtwnfKZBYtws= on 02/19/2024
depending on the syringe, pump, and flow rate.36 For this
reason, manufacturers recommend pump priming, deliv-
ering a bolus of fluid out the end of the tubing via the
pump before connecting the infusion line to the patient,
to ensure that the carriage mechanism is fully engaged
with the syringe plunger. A small bolus of 2 mL has been
shown to greatly decrease or even eliminate many start-
up delays.36
When the user starts a syringe pump, the pressure in the
syringe and tubing rises; however, fluid will not flow at the
full programmed rate until all of the distensible elements of
the infusion system become pressurized, including the rub-
ber plunger head and the infusion tubing itself.30,35 These
delays are more significant at lower infusion pump flow
rates.37 At very high flows, inaccuracies may result from lim-
ited power in the pump motor, and thus intermediate flows
may be most accurate.35 Prewetting or lubricating syringes
may paradoxically exacerbate delays in drug reaching
desired flow rates.35 Syringe size also seems to play a role in
variability and start-up delay. The time to reach target flow
using a 50-mL syringe compared with a 10-mL syringe was
significantly longer, and this difference was exacerbated at
lower flows.38 Pump priming ensures that the compressible
or distensible components of the system are pressurized,
thus minimizing start-up delay. However, this pressure will
dissipate if the pump is paused even for a few minutes.39
Drug delivery after restarting an infusion may be delayed
because of the need to repressurize the compliant elements
of the infusion system.
The flow rate from a syringe pump may be affected by
the height above or below the central venous pressure of the
patient. Lowering a syringe pump may lead to a decreased
infusion rate,37 and raising a syringe pump may lead to an
inadvertent bolus.40,41 Therefore, the height of the syringe
pump should not be abruptly altered during transport.42
The Panomat P-10 is a microvolumetric infusion pump
(Roche Diagnostics, Mannheim, Germany) that consists of
a specially designed noncompliant syringe unit combined
with stiff tubing. This pump has lower susceptibility to
start-up delays and errors caused by vertical displacement,
even at minimal flow rates.43
Peristaltic Pumps
Linear peristaltic pumps have a rotary stepper motor that is
converted to linear movement by the use of a camshaft, which
in turn activates a series of linear peristaltic actuators (fin-
ger-like projections or rollers). Alternating vertical displace-
ment of the actuators in a stepwise fashion compresses the
tubing that causes a positive forward displacement of fluid
(Figure 6A). This is the most common type of infusion pump
used today. Rotary peristaltic pumps have rotating cams that
1498   
www.anesthesia-analgesia.org
Copyright © 2017 International Anesthesia Research Society. Unauthorized reproduction of this article is prohibited.
turn in a circular fashion and compress a tube aligned in a cir-
cular geometry to produce forward displacement (Figure 6B).
The reservoir for peristaltic pumps is usually sus-
pended over the pump and can be of almost unlim-
ited size. These pumps demand a specialized cartridge
embedded within the infusion tubing to transmit the
peristaltic motion to the fluid and thus drive it forward,
which drives up the cost of each disposable. Desired char-
acteristics of pump tubing include durable materials with
known and consistent mechanical properties and little
potential for degradation after tens-of-thousands of peri-
staltic compressions. For these reasons, silicone is most
often used for pump tubing. Many peristaltic pumps
can be easily adapted to use a syringe as the reservoir
wherein the syringe is mounted inverted over the pump
and fluid is merely drawn from the negative pressure
generated by the peristaltic motion. In addition, built
into many cartridges are gravity-free-flow-stop actuators
that automatically clamp the tubing when it is removed
from the pump, thus avoiding uncontrolled drug deliv-
ery (Figure 7).44
Peristaltic pumps are able to flow over a broad range of
infusion rates (Table 1). In fact, they can usually flow faster
than many gravity-driven drug infusion systems. The
Alaris™ infusion system (Becton, Dickinson and Company,
Franklin Lakes, NJ) is an example of linear peristaltic posi-
tive displacement pumps that has a range of flows from 0.1
to 999 mL/h. Indeed, the AlarisTM pump reports an accu-
racy of ± 5.5% at rates <0.1 mL/h.44 However, at ultralow
flow rates, there is concern for pulsatile ejection from the
pump, which can lead to pharmacologic and hemody-
namic lability.32,45–48 This is because the drive mechanism
works through a series of small discrete steps, causing a
pause in flow when the linear actuators lift off the tubing
(Figure 6A).
Cassette Pumps
Cassette pumps, otherwise known as volumetric pumps,
infuse a known volume of fluid in 2 separate sequential
cycles through a valve mechanism coupled to a mechani-
cal piston.49 In the first cycle, a set volume (usually 5 mL)
from the infusate bag is directed by the valve mechanism
into the pump chamber with downward movement of the
piston, and with the next cycle, upward movement of the
piston redirects the valve so that the collected fluid is then
administered to the patient (Figure 8A).50 The Plum A+™
infusion system (Hospira, Lake Forest, IL) is an example
of a cassette pump. The major disadvantage is intermittent
interruptions in flow, up to 1 second, when the piston fills.51
Cost also plays a role because each pump has manufacturer-
specific cassettes that can be expensive. As a result of these
factors, cassette pumps have been largely replaced by peri-
staltic and syringe-type pumps.
Passive Mechanical Pumps
Spring-powered passive syringe pumps such as the
Springfusor™ (Go Medical Industries, Ltd, Subiaco, Western
Australia) are portable, low-cost, simple to use, and do not
require a power source. They drive fluid out of a syringe
over a limited time period (approximately 10 minutes), but
ANESTHESIA & ANALGESIA
 Drug Infusion Systems

Figure 6. A, Peristaltic infusion devices

utilize a rotary stepper motor that
is converted to linear movement by
the use of a camshaft, which in turn
activates a series of multiple linear
peristaltic actuators. Up and down
movements of the rollers in a step-
Downloaded from http://journals.lww.com/anesthesia-analgesia by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsI

wise fashion compress the tubing that

Ho4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdtwnfKZBYtws= on 02/19/2024

causes a positive displacement of

fluid. B, In rotatory peristaltic pumps,
rotating cams compress the tubing by
turning in a circular fashion, which pro-
pels the fluid forward.

Figure 7. A, Safety clamp fitment integrated

into pump tubing for the Alaris™ device,
which is a gravity-free-flow-stop mechanism
that automatically clamps the tubing when
it is removed from the pump, preventing
uncontrolled drug flow. B, When the tub-
ing is initially removed from the package,
the safety clamp is in the open position.
However, when the tubing is inserted into the
Alaris™ device and subsequently removed,
the safety clamp will be in the closed posi-
tion, preventing uncontrolled drug flow.

the rate is imprecise and variable. These devices are most
often used for infusions of medications whose delivery is
not time-critical such as antibiotics (Figure 8B).
Elastomeric pumps are not electromechanically operated
and do not require a motor or microprocessor system. These
devices utilize the elastic properties of the drug reservoir
to provide flow when pressurized. A desired rate can be
achieved by adjusting a mechanical flow controller. These
pumps are commonly used to deliver local anesthetics to
implanted catheters along peripheral nerves or surgical
incisions. An example is the On-Q pump (Halyard Health,
Alpharetta, GA) that has wide use in ambulatory orthope-
dic surgery (Figure 8C).
Drug Infusion Systems Subtleties
Drug infusion systems all have tubing, drive mechanisms
with flow controllers, ports and connectors, and an IV cath-
eter.6 Even if all elements of the infusion system, including
electromechanical pumps, work perfectly as users expect,
there is nuance in how drugs come together inside infusion
systems and interact with inert carrier flows in a common
fluid pathway. Unanticipated delays or transient perturba-
tions in drug administration into the patient can lead to iat-
rogenic morbidity.
The dead volume, also known as the common volume,
of a drug infusion system is defined as the volume between
the point where drug and inert carrier streams meet and the
patient’s blood. Contributions to this volume can be from dis-
tal parts of infusion tubing, connectors, manifolds, and cath-
eter lumens. When flows are stopped, one can think of this
volume as a “dead volume”: a reservoir for forgotten drug
to cause havoc when released into the patient at some later
time or a void to be traversed when drug flow first starts. If
multiple infusions are running, it can be considered to be a
volume common to all of the medications and inert carrier
fluids. Both terms, common volume and dead volume, are
used essentially interchangeably in the published literature.
Dead volume is a seemingly easy concept to grasp;
however, incident reports demonstrate that the lack of
understanding continues to cause harm in daily practice.
In a review of over 4000 safety incidents reported to the
Australian Incident Monitoring Study, 8 cases were the
result of tubing dead volume; 6 of those 8 cases were the
result of succinylcholine remaining in the dead volume
causing unintended muscle paralysis with the adminis-
tration of medication or fluid later in the perioperative
period.52 Additional reports described drug precipitation
after inadvertent mixing of incompatible drugs (thiopen-
tal and vecuronium) in the dead volume, causing occlu-
sion of the vascular access device.53 In an analysis of safety
incidents involving neuromuscular blockade in the United
Kingdom, 1 case details a patient who developed respira-
tory distress during the recovery period immediately after
an IV tubing system was flushed, implicating vecuronium
in the dead volume.54
In addition to the risk of unintended drug administra-
tion at some later time, residual drug in the dead volume
is undelivered. This may have implications for compounds
that need to rapidly reach threshold serum concentrations
such as some antibiotics.55 Clearly, much continuing educa-
tion is needed to prevent these kinds of events related to the
dead volume of drug infusion systems.

May 2017 • Volume 124 • Number 5 www.anesthesia-analgesia.org 1499

Copyright © 2017 International Anesthesia Research Society. Unauthorized reproduction of this article is prohibited.
 E NARRATIVE Review article
Downloaded from http://journals.lww.com/anesthesia-analgesia by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsI
Ho4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdtwnfKZBYtws= on 02/19/2024
Figure 8. A, Cassette pumps withdraws a set volume from the drug reservoir into the cassette reservoir (usually 5 cc) in the first cycle, and
in the next cycle, the upward movement of the piston redirects the valve so that the collected fluid is then administered to the patient. B,
Spring-powered syringe pumps such as the Springfusor™ are portable, do not require electricity or battery, are simple to use, and are low
cost. C, Elastomeric pumps such as the OnQ® pump provide constant pressure from the elastic properties of the balloon-like reservoir that
contains medication.
Implications of the Dead Volume
Infusion system architectures with multiple infusions and
an inert carrier converging on a manifold connected to an IV
catheter give rise to many considerations for safe use. First, the
dead volume becomes a reservoir for forgotten drugs when all
of the flows stop. Second, even if medications and inert carri-
ers are flowing, some mass of drug always resides within the
dead volume (Vd) that is based on the concentration in the res-
ervoir (Cres), the drug (Qd), and carrier (Qc) flow rates56:
cresQd
Massof drugin dead volume = Vd (2)
Qd + Qc 
This drug mass can be accidentally delivered into
patients if medications (eg, muscle relaxant, dilute opiate
solution, propofol) are pushed upstream, carrier flows are
suddenly increased, or another drug infusion is started
at a high flow rate.56–59 Third, when a drug is started or a
dose rate changed, there can be a significant lag for the new
concentration at the upstream side of the dead volume to
propagate to the patient. This lag is usually up to 3 to 4 time
constants (τ )57,60:
Vd
τ=
Qd + Qc  (3)
This lag time can lead to delays in intended dose
delivery that, under some conditions, can be surprisingly
long.57,59,61,62 For infusions into standard central venous cath-
eters of sizeable dead volume (9-Fr introducer sheath, dead
volume ~2.5 mL) with inert carrier flow rates designed to
minimize excessive fluid administration (10 mL/h), steady-
state drug delivery rates may not be fully realized for over
20 minutes.61 Fourth, when drug is discontinued, drug
delivery continues for some time until the dead volume is
fully cleared of drug by the inert carrier.39,57,61,63 Restarting
drug or increasing the carrier may deliver residual drug to
the patient, perhaps as an unwanted bolus. Finally, when
1500   
www.anesthesia-analgesia.org
Copyright © 2017 International Anesthesia Research Society. Unauthorized reproduction of this article is prohibited.
multiple drug infusions are flowing, a change in dose of 1
drug transiently affects the rate of delivery of all the oth-
ers.58,61,64 If multiple drugs are flowing into a bank of con-
nected stopcocks, the dead volume for drugs inserted
upstream may be slightly larger than for those inserted
downstream.
The magnitude of these phenomena is proportional to the
dead volume of the infusion system and inversely propor-
tional to the total flow rate through it (Eq. 3). Use of signifi-
cantly higher carrier flow rates could minimize these issues
at the expense of more fluid administered. Although that
may be acceptable in operating room management where
fluid administration may be more liberal, high carrier flows
may be a source of morbidity for many ICU patients.65–67
There are technologies that minimize the contribution of the
manifold to the dead volume.58–60,63,68 One design shrinks
the common lumen of the manifold to a very small volume
(Figure 9A). Another approach is to run all infusions in par-
allel lumens of a single tube that meet at the downstream
connector to the catheter (Figure 9B), thereby minimizing
the dead volume from the manifold.
Many of these pitfalls become moot if each drug flows in
its own lumen without a carrier, thus eliminating the dead
volume. This is often an impractical option because many
patients have limited vascular access ports. Minimizing
dead volume will reduce all of the aforementioned potential
dangers; however, safe practices require full comprehension
of each of these pitfalls.
In Vivo Verification Using Living Simulators
Several studies have demonstrated in vitro the lag between
intended and actual drug delivery, generated by the need
for drugs to transverse the dead volume of infusion sys-
tems.14,57,59–62,64 Other studies using catecholamine infusions
in porcine models have verified that these lags may generate
delayed responses in blood pressure and contractility.58,63,69
One can, however, question whether these delays are of clini-
cal significance in ICU management. It is possible that some
clinicians can overcome the disadvantage of dead volume
ANESTHESIA & ANALGESIA
 Drug Infusion Systems

through vigilance and skill. Large multicenter prospective
double-blind clinical trials could determine whether ICU
patients suffer less iatrogenic morbidity with infusion sys-
tems designed to have minimal dead volumes. Such trials
might be difficult to conduct because of ethical constraints,
cost, and the challenge of establishing concrete endpoints in
a real-world setting.
As an alternative, a living swine simulator was used
Downloaded from http://journals.lww.com/anesthesia-analgesia by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsI
Microinfusion Versus Macroinfusion
Historically, medications were given by macroinfusion,
defined as administration of dilute medication solutions
infused at high flow rates. This approach was necessary
before technology evolved to provide accurate delivery of
concentrated medication at low flow rates, referred to as
microinfusion. The precision of electromechanical pumps
in the modern era allows clinicians to infuse medications

to determine whether the magnitude of the infusion sys- at flows as low as 0.1 mL/h. The choice between microin-
Ho4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdtwnfKZBYtws= on 02/19/2024

tem dead volume affected the ability of experienced ICU
nurses to maintain blood pressure within a specific range
in a model of hemodynamic instability.70 Occlusion bal-
loons were placed in the thoracic aorta and inferior vena
cava of anesthetized swine to rapidly raise or lower blood
pressure. Experienced ICU RNs were tasked to maintain
blood pressure within a narrow range using only infusions
of norepinephrine or sodium nitroprusside. The RNs were
blinded to balloon occlusions but had full access to hemo-
dynamic monitors. The data showed a wide range of RN
practice styles (indices included time to act, frequency of
actions, and magnitude of drug infusions). However, the
data also showed that hemodynamic stability was more
easily restored after balloon occlusion perturbation when
the medications were infused via an infusion system with a
smaller dead volume than a large one (Figure 10). Thus, one
might infer that infusion system architecture plays a role in
achieving therapeutic goals in a living simulator of hemo-
dynamic stability, even in very experienced hands.
All of the modeling of continuous drug infusion through
in vivo swine simulation58,63,70 has been performed using
drugs with short half-lives (epinephrine and norepineph-
rine) and therefore rapid onset of biologic effect. Plasma
concentrations of infused drugs with longer half-lives
reach steady-state within the circulation slower than drugs
with short half-lives.71 Therefore, for drugs with long half-
lives, although untested, delays in delivery from traversing
the dead volume may be insignificant compared with the
time needed to reach steady-state plasma concentration.
Examples include milrinone, vasopressin, some analgesic
agents, diuretics, and anticoagulants. For such drugs, the
effect of dead volume on delaying therapeutic goals is likely
to be less significant.
fusion and macroinfusion tends to be institution-specific
because familiarity, education, and training are essential to
the safe execution of either. A summation of the pros and
cons of each approach is summarized in Table 2.
The fast flow rate associated with macroinfusion has sev-
eral implications (Table 2). First, the concentrations of drugs
in the infused fluids are lower, so there is less liability that
large amounts of potent drug might be stored within the dead
volume and inadvertently bolus injected (Eq. 2). Second, the
start-up delays from pumps may be minimized because pres-
sure in the tubing rises faster. Third, there is less likelihood
of delayed occlusion alarms from pumps with faster flows.
In addition, there are shorter lags between intended dosing
and drug delivery at faster drug and total flow rates (Eq. 3).
The drug flow rates typically seen in macroinfusion may
be 7 to 50 mL/h, but under some circumstances may be much
higher. If multiple medications are infused, the fluid volume
delivered in a day may greatly exceed the patient’s maintenance
fluid requirements. This volume would be in addition to the
drug volume associated with intermittent dosing of antibiot-
ics, electrolyte supplementation, or analgesics. Although many
patients might compensate for this excess fluid, patients with
cardiac, pulmonary, renal, or neurologic illnesses may not tol-
erate this excess volume. Thus, the low concentration of drugs
mandates higher administered volumes of fluid, which can be
a major source of morbidity in the critically ill.65–67 Clinicians
and pharmacists may attempt to increase the concentration
of each medication (“double or quadruple strength”) to miti-
gate the problem; however, it is usually adopted in a reactive
manner after much excess fluid has already been given. In
addition, dosing errors can ensue when drugs are available in
multiple concentrations. Thus, some hospital intensivists and
pharmacists elect a microinfusion strategy.

Figure 9. A, Conventional 4-stopcock manifold (Top, Arrow International, Reading, PA) has a much larger internal volume than the 6-port Multi
Line Extension Set (Bottom, Summit Medical Products, Worcester, MA). B, Low-dead volume manifold (Edelvaiss Multiline, Doran International,
Toussieu, France) wherein drug infusions and crystalloid carriers flow coaxially down a 150-cm tube. Schematic of the tip of the Edelvaiss
Multiline visualizing the separation of multiple drug stream compartments so that both the carrier and the drug meet at the downstream con-
nector to the catheter, minimizing contribution of the manifold to infusion system dead volume.

May 2017 • Volume 124 • Number 5 www.anesthesia-analgesia.org 1501

Copyright © 2017 International Anesthesia Research Society. Unauthorized reproduction of this article is prohibited.
 E NARRATIVE Review article
Downloaded from http://journals.lww.com/anesthesia-analgesia by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsI
Ho4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdtwnfKZBYtws= on 02/19/2024

Figure 10. A living swine simulator was used to determine whether the magnitude of the infusion system dead volume affected the ability of
experienced intensive care unit nurses to maintain blood pressure within a specific range in a model of hemodynamic instability. Occlusion
balloons were placed in the thoracic aorta and inferior vena cava to rapidly raise or lower blood pressure, and intensive care unit RNs were
tasked to maintain the mean arterial blood pressure (MAP) between 70 and 90 mm Hg using only infusions of norepinephrine or sodium nitro-
prusside. RNs were blinded to balloon occlusions but had full access to hemodynamic monitors. Representative samples of MAP data during
and after aortic and inferior vena cava (IVC) occlusion are shown.70 Hemodynamic stability was more easily restored after balloon occlusion
perturbation with a smaller dead volume infusion system than a large one. Reproduced with permission.70

Microinfusion is designed to minimize fluid delivery

Table 2.  Comparison of Microinfusion with
associated with drug administration (Figure 11). The low
Macroinfusion
drug flow rates, however, can lead to errors in drug dosing Macroinfusion Microinfusion
and patient morbidity. When flow is first initiated, all the Typical medication flow rates 7–50 mL/h 0.1–10 mL/h
factors that cause start-up delays with syringe and peristal- Medication preparation Bedside or pharmacy Pharmacy
tic pumps become more germane. A prototype microvolu- Pump Peristaltic large Syringe or
metric infusion pump device has been shown to decrease volume peristaltic large
start-up time and no drug delivery time through a special volume
Fluid volume administered Liberala Conservative
microsyringe designed with a rotating valve that mechani-
Lag time to steady-stateb Short Long
cally isolates the pump tubing into afferent and efferent Start-up delay Minimal Variable
infusion lines without a rubber plunger head, thus minimiz- Mass of drug residing in Lower Higher
ing system compliance.72 catheter and manifold
As discussed, the mass of drug residing in the infusion dead volume
system dead volume can be rapidly administered if a car- Potential for clinically Lower Higher
significant drug bolusc
rier rate increases, coinfusion of another drug is initiated,
or if a bolus of fluid is “pushed” upstream in the same fluid
a
Can be much greater than patients total fluid requirements and may be a
source of morbidity in patients with extreme cardiac, pulmonary, neurologic,
pathway. For a given dead volume, the drug mass avail- or renal disease.65–67
able for bolus is much greater with microinfusion than with b
When infused with an inert carrier typical of intensive care unit conditions
(adult: 7–20 mL/h or pediatric: 1–10 mL/h), delay in achieving steady-state
macroinfusion because the drug flowing from the reser- drug delivery after a change in drug or carrier flow rate.
voir is concentrated (Eq. 2). The use of inert carriers dilutes c
If a drug is pushed upstream or if a carrier flow rate or other drug flow rate
the concentrated drug within the infusion system tubing. is raised abruptly.
However, if carrier rates are slow or if drug solutions are
highly concentrated, the dilution may be insufficient to for propofol administration.73 Other medications such as
protect the patient from a large inadvertent bolus.56 Thus, remifentanil, sufentanil, and alfentanil have been studied
concentrated drug delivered by microinfusion strategies for use in TCI systems.74–77 The promise of TCI includes a
decrease the volume of fluid administration at the expense decreased incidence of recall, movement under anesthesia,
of potential dangers and iatrogenic sources of morbidity and delayed emergence. Studies have yet to show improved
from drug boluses. outcome of TCI compared with manual injection.78 This
may be attributable in part to limitations and inaccuracies
Future of Drug Infusion in the performances of available infusion pumps capable
Computer control of infusion rates was introduced in 1996 of external control by computers.35 Unrecognized delays in
as a method of tailoring drug infusion to produce a desired drug delivery resulting from the need for the drug to tra-
effect. A comprehensive review of such control methods is verse the infusion system dead volume may also contribute.
beyond the scope of this work. Briefly, these systems use The promise of TCI drug dosing may some day be fully real-
either pharmacokinetic models or feedback control from a ized with refinements in infusion pumps along with recog-
physiologic measurement. Target controlled infusion (TCI) nition of, and compensation for, dead volume-related drug
systems use population-based pharmacokinetic models to administration delays by TCI algorithms.
automatically drive drug delivery to achieve a theoretically Physiologic feedback technologies seek to titrate medi-
estimated target site drug concentration. Devices using cation infusion rate based on responses measured by moni-
these methods have been developed and are in clinical use toring systems such as the bispectral index monitor.79 Other
in some countries such as the Diprifusor™, a TCI system methods include electroencephalographic algorithms

1502   
www.anesthesia-analgesia.org ANESTHESIA & ANALGESIA
Copyright © 2017 International Anesthesia Research Society. Unauthorized reproduction of this article is prohibited.
 Drug Infusion Systems
Downloaded from http://journals.lww.com/anesthesia-analgesia by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsI
Ho4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdtwnfKZBYtws= on 02/19/2024
Figure 11. Phenylephrine, a commonly used vasoactive medica-
tion, is conventionally diluted in 250-mL bags. A more concentrated
microinfusion mix can be prepared in a syringe such as 60 mg/50
mL. Phenylephrine infused at 60 µg/min will require 90 mL/h from
the infusion bag (macroinfusion), whereas the concentrated syringe
preparation will achieve the desired dose at 3 mL/h (microinfu-
sion). One syringe (50 mL) has the same drug mass as 6 bags have
(1500 mL).
developed to recognize burst suppression,80–83 which have
been used for titration of propofol.84,85 Other systems have
used hemodynamic endpoints as a basis for titration of
medication during cardiac surgery.86–88 Such intelligent
feedback-driven drug delivery may one day allow devel-
opment of automated hemodynamic or electroencephalog-
raphy-based anesthetic delivery systems. Similar to TCI,
widespread deployment of effective physiologic feedback
systems will ultimately rely on development of infusion
pumps that are accurate and controllable, with incorpora-
tion of dead volume-related delays in drug delivery into
control software.
Automatically Meeting Clinicians’ Expectations
A strategy to accelerate the propagation of drug across the
infusion system dead volume to rapidly achieve desired
May 2017 • Volume 124 • Number 5
Copyright © 2017 International Anesthesia Research Society. Unauthorized reproduction of this article is prohibited.
steady-state drug administration after changes to drug dose
and carrier rates set by clinicians has been demonstrated.69,89
Syringe pumps were modified to allow external control from
central computers running software that use finite element
models of the movement of drug through the dead volume.
Algorithms designed to coordinate and control both the
drug and carrier flows speed the movement of drug across
the dead volume. This dramatically decreases the time to
achieve intended drug delivery in vitro and pharmacologic
response in vivo. It is possible that in the future, integrated
infusion systems will allow users to set desired drug deliv-
ery goals and software will coordinate and control drug and
carrier infusions to provide near instantaneous desired drug
delivery.
CONCLUSIONS
Drug infusion systems range in complexity, precision, and
cost from simple, inexpensive, gravity-driven drips to the
sophisticated electromechanical infusion technologies
featured in modern large-volume peristaltic and syringe
pumps. This review outlines the technology and high-
lights advantages and disadvantages of these very different
drug delivery systems. The distinction between microin-
fusion- and macroinfusion-based systems is specifically
emphasized along with a quantitative discussion of infu-
sion system dead volume and its effect on drug delivery.
Modern drug infusion systems provide precise drug deliv-
ery and important safety enhancements. However, all infu-
sion systems have the potential for iatrogenic morbidity
from unappreciated residual drug in the dead volume and
unappreciated delays in the kinetics of drug administra-
tion by infusion. New technologies may ultimately lead to
the implementation of “smart” systems capable of diffus-
ing some of these risks. Until then, continuing education is
needed to prevent adverse events related to the nuances of
drug infusion systems. E
ACKNOWLEDGMENT
The authors wish to thank Joon Kim for assisting with the
illustrations in many of the figures.
DISCLOSURES
Name: Uoo R. Kim, MD.
Contribution: This author helped read the reference literature, and
write and edit the manuscript.
Name: Robert A. Peterfreund, MD, PhD.
Contribution: This author helped read the reference literature, and
write and edit the manuscript.
Name: Mark A. Lovich, MD, PhD.
Contribution: This author helped read the reference literature, and
write and edit the manuscript.
This manuscript was handled by: Maxime Cannesson, MD, PhD.
REFERENCES
1. Peterfreund RA, Philip JH. Critical parameters in drug
delivery by intravenous infusion. Expert Opin Drug Deliv.
2013;10:1095–1108.
2. Barsoum N, Kleeman C. Now and then, the history of paren-
teral fluid administration. Am J Nephrol. 2002;22:284–289.
3. Millam D. The history of intravenous therapy. J Intraven Nurs.
1996;19:5–14.
4. Welck M, Borg P, Ellis H. James Blundell MD Edin FRCP
(1790–1877): pioneer of blood transfusion. J Med Biogr.
2010;18:194–197.
www.anesthesia-analgesia.org 1503
 E NARRATIVE Review article
5. MacGillivray N. Dr Latta of Leith: pioneer in the treatment of
cholera by intravenous saline infusion. J R Coll Physicians Edinb.
2006;36:80–85.
6. Philip JH. Intravenous access and delivery principles. In:
Longnecker DE, ed. Principles and Practice of Anesthesiology. St.
Louis, MO: Mosby Year Book; 1992:1183–1196.
7. Demoruelle JL, Harrison WL, Flora RE. Flow rate maintenance
and output of intravenous fluid administration sets. Am J Hosp
Pharm. 1975;32:177–185.
8. Tolman RC. The effect of droplet size on surface tension. J Chem
Downloaded from http://journals.lww.com/anesthesia-analgesia by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsI
Phys. 1949;17:333–337.
Ho4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdtwnfKZBYtws= on 02/19/2024
9. Flack FC, Whyte TD. Variations of drop size in disposable
administration sets used for intravenous infusion. J Clin Pathol.
1975;28:510–512.
10. Crass RE, Vance JR. In vivo accuracy of gravity-flow i.v. infu-
sion systems. Am J Hosp Pharm. 1985;42:328–331.
11. Flack FC, Whyte TD. Behaviour of standard gravity-fed
administration sets used for intravenous infusion. Br Med J.
1974;3:439–443.
12. La Cour D. Drop size in disposable sets for intravenous infu-
sion. Acta Anaesthesiol Scand. 1965;9:145–154.
13. Hamilton S. Intravenous therapy. In: Nettina SM, ed. Lippincott
Manual of Nursing Practice. 10th ed. Ambler, PA: Lippincott
Williams & Wilkins; 2014.
14. Décaudin B, Dewulf S, Lannoy D, et al. Impact of multiaccess
infusion devices on in vitro drug delivery during multi-infu-
sion therapy. Anesth Analg. 2009;109:1147–1155.
15. Horrow JC, Jaffe JR, Rosenberg H. A laboratory evalua-
tion of resistive intravenous flow regulators. Anesth Analg.
1987;66:660–665.
16. Rithalia SV, Rozkovec A. Evaluation of a simple device for reg-
ulating intravenous infusions. Intensive Care Med. 1979;5:41–43.
17. Pierce ET, Kumar V, Zheng H, Peterfreund RA. Medication
and volume delivery by gravity-driven micro-drip intravenous
infusion: potential variations during ‘wide-open’ flow. Anesth
Analg. 2013;116:614–618.
18. Pugsley HE, Farquharson RF. Factors influencing the rate
of flow of intravenous infusions (with special reference
to venoconstriction induced by serum). Can Med Assoc J.
1944;51:5–11.
19. Murphy RS, Wilcox SJ. High risk infusions—accuracy compro-
mised by changes in patient venous pressure. J Med Eng Technol.
2009;33:470–474.
20. Singh S, Randle LV, Callaghan PT, Watson CJ, Callaghan CJ.
Beyond poiseuille: preservation fluid flow in an experimental
model. J Transplant. 2013;2013:605326.
21. Choi GJ, Yoon IJ, Lee OH. Accuracy of automatic infusion
controller (AutoClamp) for intravenous administration. Open
Anesth J. 2015;9:23–28.
22. Hungerford, V. Sigmamotor Inc. Early history of Sigmamotor,
Sigma Incorporated and American Sigma. Available at: http://
www.sigmamotorinc.com/history_sigm-a_motor.html.
Accessed January 4, 2016.
23. Passaroni AC, Silva MA, Yoshida WB. Cardiopulmonary
bypass: development of John Gibbon’s heart-lung machine. Rev
Bras Cir Cardiovasc. 2015;30:235–245.
24. The San Diego Technology Archive. The IVAC legacy. Available
at: http://libraries.ucsd.edu/assets/sdta/etc/ivac-legacy.pdf.
Accessed August 10, 2016.
25. Scanlon M. The role of ‘smart’ infusion pumps in patient safety.
Pediatr Clin North Am. 2012;59:1257–1267.
26. Syroid N, Liu D, Albert R, et al. Graphical user interface simpli-
fies infusion pump programming and enhances the ability to
detect pump-related faults. Anesth Analg. 2012;115:1087–1097.
27. Dönmez A, Araz C, Kayhan Z. Syringe pumps take too long to
give occlusion alarm. Paediatr Anaesth. 2005;15:293–296.
28. Kim DW, Steward DJ. The effect of syringe size on the perfor-
mance of an infusion pump. Paediatr Anaesth. 1999;9:335–337.
29. Neal D, Lin JA. The effect of syringe size on reliability
and safety of low-flow infusions. Pediatr Crit Care Med.
2009;10:592–596.
30. Weiss M, Bänziger O, Neff T, Fanconi S. Influence of infusion
line compliance on drug delivery rate during acute line loop
formation. Intensive Care Med. 2000;26:776–779.
1504   
www.anesthesia-analgesia.org
Copyright © 2017 International Anesthesia Research Society. Unauthorized reproduction of this article is prohibited.
31. Capes DF, Dunster KR, Sunderland VB, McMillan D, Colditz
PB, McDonald C. Fluctuations in syringe-pump infusions: asso-
ciation with blood pressure variations in infants. Am J Health
Syst Pharm. 1995;52:1646–1653.
32. Klem SA, Farrington JM, Leff RD. Influence of infusion pump
operation and flow rate on hemodynamic stability during epi-
nephrine infusion. Crit Care Med. 1993;21:1213–1217.
33. U.S. Food and Drug Administration. FDA Safety
Communication 2016. Available at: http://www.fda.gov/
MedicalDevices/Safety/AlertsandNotices/ucm518049.htm.
Accessed September 12, 2016.
34. van der Eijk AC, van Rens RM, Dankelman J, Smit BJ. A lit-
erature review on flow-rate variability in neonatal IV therapy.
Paediatr Anaesth. 2013;23:9–21.
35. Sarraf E, Mandel JE. Time-delay when updating infusion rates
in the Graseby 3400 pump results in reduced drug delivery.
Anesth Analg. 2014;118:145–150.
36. Neff T, Fischer J, Fehr S, Baenziger O, Weiss M. Start-up delays
of infusion syringe pumps. Paediatr Anaesth. 2001;11:561–565.
37. Kern H, Kuring A, Redlich U, et al. Downward movement
of syringe pumps reduces syringe output. Br J Anaesth.
2001;86:828–831.
38. Schmidt N, Saez C, Seri I, Maturana A. Impact of syringe size
on the performance of infusion pumps at low flow rates. Pediatr
Crit Care Med. 2010;11:282–286.
39. Bartels K, Moss DR, Peterfreund RA. An analysis of drug deliv-
ery dynamics via a pediatric central venous infusion system:
quantification of delays in achieving intended doses. Anesth
Analg. 2009;109:1156–1161.
40. Lönnqvist PA, Löfqvist B. Design flaw can convert commer-
cially available continuous syringe pumps to intermittent bolus
injectors. Intensive Care Med. 1997;23:998–1001.
41. Keay S, Callander C. The safe use of infusion devices. Contin
Educ Anaesth Crit Care Pain. 2004;4:81–85.
42. Krauskopf KH, Rauscher J, Brandt L. Disturbance of continu-
ous, pump administration of cardiovascular drugs by hydro-
static pressure. Anaesthesist. 1996;45:449–452.
43. Neff TA, Fellmann C, Fuechslin RM, Gerber AC, Weiss M. The
Panomat P-10 micro-volumetric infusion pump is suitable for
continuous drug administration at minimal flow rates. Can J
Anaesth. 2002;49:1048–1052.
44. Becton, Dickinson and Company. Alaris™ System User Manual
2015. Available at: http://www.carefusion.com/Documents/
guides/user-guides/IF_Alaris-System-8015-v9-19_UG_
EN.pdf. Accessed January 8, 2016.
45. Schulze KF, Graff M, Schimmel MS, Schenkman A, Rohan P.
Physiologic oscillations produced by an infusion pump. J
Pediatr. 1983;103:796–798.
46. Hurlbut JC, Thompson S, Reed MD, Blumer JL, Erenberg A,
Leff RD. Influence of infusion pumps on the pharmacologic
response to nitroprusside. Crit Care Med. 1991;19:98–101.
47. Wood BR, Huddleston K, Kolm P. A comparison of infusion
devices at 1 ml/hr. Neonatal Intensive Care. 1993;6:20–22.
48. Stull JC, Erenberg A, Leff RD. Flow rate variability from elec-
tronic infusion devices. Crit Care Med. 1988;16:888–891.
49. Millam DA, Dufour JL. I.V. therapy. In: Mayer B, ed. Illustrated
Manual of Nursing Practice. 3rd ed. Springhouse, PA: Lippincott
Williams & Wilkins; 2002:114.
50. Voss GI, Butterfield RD. Parenteral infusion devices. In: David Y,
ed. Clinical Engineering. Boca Raton, FL: CRC Press LLC; 2003:278.
51. Diba A. Clinical Gate. Infusion equipment and intravenous anes-
thesia. Available at: http://clinicalgate.com/infusion-equip-
ment-and-intravenous-anaesthesia. Accessed February 24, 2016.
52. Singleton RJ, Kinnear SB, Currie M, Helps SC. Crisis manage-
ment during anaesthesia: vascular access problems. Qual Saf
Health Care. 2005;14:e20.
53. Taniguchi T, Yamamoto K, Kobayashi T. The precipitate formed
by thiopentone and vecuronium. Can J Anaesth. 1996;43:511–513.
54. Arnot-Smith J, Smith AF. Patient safety incidents involving neu-
romuscular blockade: analysis of the UK National Reporting
and Learning System data from 2006 to 2008. Anaesthesia.
2010;65:1106–1113.
55. Sherwin CM, McCaffrey F, Broadbent RS, Reith DM, Medlicott
NJ. Discrepancies between predicted and observed rates
ANESTHESIA & ANALGESIA
 Drug Infusion Systems
of intravenous gentamicin delivery for neonates. J Pharm
Pharmacol. 2009;61:465–471.
56. Lovich MA, Kinnealley ME, Sims NM, Peterfreund RA. The
delivery of drugs to patients by continuous intravenous infu-
sion: modeling predicts potential dose fluctuations depending
on flow rates and infusion system dead volume. Anesth Analg.
2006;102:1147–1153.
57. Lovich MA, Doles J, Peterfreund RA. The impact of carrier flow
rate and infusion set dead-volume on the dynamics of intrave-
nous drug delivery. Anesth Analg. 2005;100:1048–1055.
Downloaded from http://journals.lww.com/anesthesia-analgesia by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsI
58. Lovich MA, Pezone MJ, Maslov MY, Murray MR, Wakim MG,
Ho4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC1y0abggQZXdtwnfKZBYtws= on 02/19/2024
Peterfreund RA. Infusion system carrier flow perturbations and
dead-volume: large effects on drug delivery in vitro and hemody-
namic responses in a swine model. Anesth Analg. 2015;120:1255–1263.
59. Lannoy D, Decaudin B, Simon N, Barthelemy C, Debaene
B, Odou P. The impact on drug mass flow rate of interrupt-
ing and resuming carrier fluid flow: an in vitro study on
a very low dead-space volume infusion set. Anesth Analg.
2012;114:328–332.
60. Moss DR, Bartels K, Peterfreund GL, Lovich MA, Sims NM,
Peterfreund RA. An in vitro analysis of central venous drug
delivery by continuous infusion: the effect of manifold design
and port selection. Anesth Analg. 2009;109:1524–1529.
61. Lovich MA, Peterfreund GL, Sims NM, Peterfreund RA.
Central venous catheter infusions: a laboratory model shows
large differences in drug delivery dynamics related to catheter
dead volume. Crit Care Med. 2007;35:2792–2798.
62. Lannoy D, Décaudin B, Dewulf S, et al. Infusion set characteris-
tics such as antireflux valve and dead-space volume affect drug
delivery: an experimental study designed to enhance infusion
sets. Anesth Analg. 2010;111:1427–1431.
63. Lovich MA, Wakim MG, Wei A, et al. Drug infusion system
manifold dead-volume impacts the delivery response time to
changes in infused medication doses in vitro and also in vivo in
anesthetized swine. Anesth Analg. 2013;117:1313–1318.
64. Tsao AC, Lovich MA, Parker MJ, Zheng H, Peterfreund RA.
Delivery interaction between co-infused medications: an in vitro
modeling study of microinfusion. Paediatr Anaesth. 2013;23:33–39.
65. Lex DJ, Tóth R, Czobor NR, et al. Fluid overload is associated
with higher mortality and morbidity in pediatric patients under-
going cardiac surgery. Pediatr Crit Care Med. 2016;17:307–314.
66. Arikan AA, Zappitelli M, Goldstein SL, Naipaul A, Jefferson
LS, Loftis LL. Fluid overload is associated with impaired oxy-
genation and morbidity in critically ill children. Pediatr Crit
Care Med. 2012;13:253–258.
67. Tsai YC, Chiu YW, Tsai JC, et al. Association of fluid overload
with cardiovascular morbidity and all-cause mortality in stages
4 and 5 CKD. Clin J Am Soc Nephrol. 2015;10:39–46.
68. Foinard A, Décaudin B, Barthélémy C, Debaene B, Odou
P. Prevention of drug delivery disturbances during con-
tinuous intravenous infusion: an in vitro study on a new
multi-lumen infusion access device. Ann Fr Anesth Reanim.
2013;32:e107–e112.
69. Parker MJ, Lovich MA, Tsao AC, et al. Computer control of
drug delivery by continuous intravenous infusion: bridging the
gap between intended and actual drug delivery. Anesthesiology.
2015;122:647–658.
70. Pezone MJ, Peterfreund RA, Maslov MY, Govindaswamy RR,
Lovich MA. Infusion system architecture impacts the ability of
intensive care nurses to maintain hemodynamic stability in a
living swine simulator. Anesthesiology. 2016;124:1077–1085.
May 2017 • Volume 124 • Number 5
Copyright © 2017 International Anesthesia Research Society. Unauthorized reproduction of this article is prohibited.
71. Shafer S. Principles of pharmacokinetics and pharmacodynam-
ics. In: Longnecker DE, Tinker JH, Morgan GE, eds. Principles
and Practice of Anesthesiology. 2nd ed. St. Louis, MO: Mosby-
Year Book; 1997:1159–1210.
72. Weiss M, Gerber S, Füchslin RM, Neff TA. Accurate continuous
drug delivery at low infusion rate with a novel microvolumet-
ric infusion pump (MVIP): pump design, evaluation and com-
parison to the current standard. Anaesthesia. 2004;59:1133–1137.
73. Glen JB. The development of ‘Diprifusor’: a TCI system for pro-
pofol. Anaesthesia. 1998;53(suppl 1):13–21.
74. Engbers F. Practical use of ‘Diprifusor’ systems. Anaesthesia.
1998;53(suppl 1):28–34.
75. Moerman AT, Herregods LL, De Vos MM, Mortier EP, Struys
MM. Manual versus target-controlled infusion remifentanil
administration in spontaneously breathing patients. Anesth
Analg. 2009;108:828–834.
76. Slepchenko G, Simon N, Goubaux B, Levron JC, Le Moing JP,
Raucoules-Aimé M. Performance of target-controlled sufent-
anil infusion in obese patients. Anesthesiology. 2003;98:65–73.
77. Pérus O, Marsot A, Ramain E, et al. Performance of alfentanil
target-controlled infusion in normal and morbidly obese female
patients. Br J Anaesth. 2012;109:551–560.
78. Leslie K, Clavisi O, Hargrove J. Target-controlled infusion
versus manually-controlled infusion of propofol for general
anesthesia or sedation in adults. Cochrane Database Syst Rev.
2008;3:CD006059.
79. Powers KS, Nazarian EB, Tapyrik SA, et al. Bispectral index
as a guide for titration of propofol during procedural sedation
among children. Pediatrics. 2005;115:1666–1674.
80. Särkelä M, Mustola S, Seppänen T, et al. Automatic analysis
and monitoring of burst suppression in anesthesia. J Clin Monit
Comput. 2002;17:125–134.
81. Brandon Westover M, Shafi MM, Ching S, et al. Real-time seg-
mentation of burst suppression patterns in critical care EEG
monitoring. J Neurosci Methods. 2013;219:131–141.
82. Ching S, Liberman MY, Chemali JJ, et al. Real-time closed-loop
control in a rodent model of medically induced coma using
burst suppression. Anesthesiology. 2013;119:848–860.
83. Marchant N, Sanders R, Sleigh J, et al. How electroencepha-
lography serves the anesthesiologist. Clin EEG Neurosci.
2014;45:22–32.
84. Mukamel EA, Pirondini E, Babadi B, et al. A transition in brain
state during propofol-induced unconsciousness. J Neurosci.
2014;34:839–845.
85. Purdon PL, Pierce ET, Mukamel EA, et al. Electroencephalogram
signatures of loss and recovery of consciousness from propofol.
Proc Natl Acad Sci U S A. 2013;110:E1142–E1151.
86. Hoeksel SA, Blom JA, Jansen JR, Maessen JG, Schreuder JJ.
Computer control versus manual control of systemic hypertension
during cardiac surgery. Acta Anaesthesiol Scand. 2001;45:553–557.
87. Hoeksel SA, Schreuder JJ, Blom JA, Maessen JG, Penn OC.
Automated infusion of nitroglycerin to control arterial hyperten-
sion during cardiac surgery. Intensive Care Med. 1996;22:688–693.
88. Hoeksel SA, Blom JA, Jansen JR, Maessen JG, Schreuder JJ.
Automated infusion of vasoactive and inotropic drugs to con-
trol arterial and pulmonary pressures during cardiac surgery.
Crit Care Med. 1999;27:2792–2798.
89. Parker MJ, Lovich MA, Tsao AC, Deng H, Houle T, Peterfreund
RA. Novel pump control technology accelerates drug deliv-
ery onset in a model of pediatric drug infusion. Anesth Analg.
2017;124:1129–1134.
www.anesthesia-analgesia.org 1505