Advanced Drug Delivery Reviews 54 (2002) 1015–1039

www.elsevier.com / locate / drugdeliv

Poly(ortho esters): synthesis, characterization, properties and

uses
Jorge Heller a , *, John Barr a , Steven Y. Ng a , Khadija Schwach Abdellauoi b ,
Robert Gurny b
a
A.P. Pharma, 123 Saginaw Drive, Redwood City, CA 94063, USA
b
University of Geneva, School of Pharmacy, 30 Quai E. Ansermet, CH1211, Geneva 4, Switzerland
Received 22 March 2002; accepted 19 June 2002

Abstract

Over the last 30 years, poly(ortho esters) have evolved through four families, designated as POE I, POE II, POE III and
POE IV. Of these, only POE IV has been shown to have all the necessary attributes to allow commercialization, and such
efforts are currently underway. Dominant among these attributes is synthesis versatility that allows the facile and
reproducible production of polymers having the desired mechanical and thermal properties as well as desired erosion rates
and drug release rates that can be varied from a few days to many months. Further, the polymer is stable at room temperature
when stored under anhydrous conditions and undergoes an erosion process confined predominantly to the surface layers.
Important consequences of surface erosion are controlled and concomitant drug release as well as the maintenance of an
essentially neutral pH in the interior of the matrix because acidic hydrolysis products diffuse away from the device. Two
physical forms of such polymers are under development. One form, solid materials, can be fabricated into shapes such as
wafers, strands, or microspheres. The other form are injectable semi-solid materials that allow drug incorporation by a simple
mixing at room temperature and without the use of solvents. GMP toxicology studies on one family of POE IV polymers has
been concluded, an IND filed and Phase I clinical trials are in progress. Important applications under development are
treatment of post-surgical pain, osteoarthritis and ophthalmic diseases as well as the delivery of proteins, including DNA.
Block copolymers of poly(ortho ester) and poly(ethylene glycol) have been prepared and their use as a matrix for drug
delivery and as micelles, primarily for tumor targeting, are being explored.
 2002 Elsevier Science B.V. All rights reserved.

Keywords: Poly(ortho ester); Bioerodible polymer; Injectable semi-solid; Drug delivery; Protein delivery; Post-surgical pain; Periodontal
disease; Ophthalmic delivery; Micelles; Tumor targeting

Contents

1. Introduction ............................................................................................................................................................................ 1016

2. Poly(ortho ester) I.................................................................................................................................................................... 1016
3. Poly(ortho ester) II .................................................................................................................................................................. 1017

*Corresponding author. Tel.: 1 1-605-366-2626; fax: 1 1-650-365-9452.

E-mail address: jheller@appharma.com (J. Heller).

0169-409X / 02 / $ – see front matter  2002 Elsevier Science B.V. All rights reserved.
PII: S0169-409X( 02 )00055-8
1016 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039

4. Poly(ortho ester) III ................................................................................................................................................................. 1018

5. Poly(ortho ester) IV ................................................................................................................................................................. 1020
5.1. Polymer synthesis ............................................................................................................................................................ 1020
5.2. Polymer hydrolysis .......................................................................................................................................................... 1022
5.3. Control of mechanical and thermal properties..................................................................................................................... 1024
5.4. Polymer fabrication .......................................................................................................................................................... 1025
5.5. Polymer stability .............................................................................................................................................................. 1025
5.6. Control of erosion rate...................................................................................................................................................... 1026
5.7. Control of drug delivery rates ........................................................................................................................................... 1027
5.7.1. Low molecular weight, water-soluble drugs ............................................................................................................. 1027
5.7.2. Macromolecular drugs ............................................................................................................................................ 1027
6. Semi-solid polymers ................................................................................................................................................................ 1029
6.1. Polymer molecular weight control ..................................................................................................................................... 1029
6.2. Polymer stability .............................................................................................................................................................. 1030
6.3. Drug release .................................................................................................................................................................... 1031
6.3.1. Triethylene glycol semi-solids................................................................................................................................. 1031
6.3.1.1. Post-surgical pain control ........................................................................................................................... 1031
6.3.2. 1,10-decanediol semi-solids .................................................................................................................................... 1032
6.3.2.1. Periodontal disease .................................................................................................................................... 1032
6.3.2.2. Ocular applications .................................................................................................................................... 1033
6.3.2.3. Estrus synchronization in sheep .................................................................................................................. 1033
7. Block copolymers with poly(ethylene glycol) ............................................................................................................................ 1033
7.1. Synthesis ......................................................................................................................................................................... 1033
7.1.1. Release of proteins ................................................................................................................................................. 1033
7.1.2. Micelles ................................................................................................................................................................ 1034
8. Conclusions ............................................................................................................................................................................ 1036
Acknowledgements ...................................................................................................................................................................... 1036
References .................................................................................................................................................................................. 1037

1. Introduction 2. Poly(ortho ester) I

Poly(ortho esters) have already been extensively Poly(ortho ester) I, the first such polymer pre-
reviewed so no attempt will be made to provide a pared, has been developed at the Alza Corporation
further exhaustive review. Instead, the reader is and described in a series of patents by Choi and
referred to a comprehensive article in Advances in Heller [3–7]. This polymer was originally designated
Polymer Science that covers all work with poly(ortho as Chronomer 姠 but the name was later changed to
esters) prior to and including 1992 [1], and a chapter Alzamer  . The polymer is prepared as shown in
in Encyclopedia of Controlled Drug Delivery that Scheme 2.
includes all work up to 1998 [2]. When placed in an aqueous environment, the
Poly(ortho esters) have been under development polymer will hydrolyze as shown in Scheme 3.
since 1970 and during that time, four polymer Because ortho ester linkages are acid sensitive and
families have been developed. These are shown in hydrolysis of this polymer produces g -butyrolactone
Scheme 1. which rapidly opens to g -hydroxybutyric acid, the
In this review, we will concentrate on POE IV and polymer must be stabilized with a base such as
on the latest developments in the use and characteri- Na 2 CO 3 to avoid an uncontrolled, autocatalytic
zation of this material. However, for the sake of hydrolysis reaction.
completeness, for each poly(ortho ester) family, we The polymer has been used in the treatment of
will present a brief, updated review with pertinent burns [8], in the delivery of the narcotic antagonist
references, and refer the reader to the original naltrexone [9] and in the delivery of the contracep-
literature for additional details. tive steroid levonorgestrel [10]. The polymer has
 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039 1017

Scheme 1.

Scheme 2.

Scheme 3.

also been investigated by Sudmann in a number of 3. Poly(ortho ester) II

orthopedic applications [11–19]. However, the au-
tocatalytic nature of the hydrolysis, as well as the
rather low glass transition temperature of the poly-
mer has severely limited its application and this
polymer is no longer under development.
Poly(ortho ester) II was developed at the Stanford
Research Institute, now known as SRI International,
under contract with the National Institute of Child
Health and Human Development. A history of the
1018 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039
Scheme 4.
development of this polymer has been published
[20].
Poly(ortho ester) II is prepared by the addition of a
diol to the diketene acetal 3,9-diethylidene 2,4,8,10-
tetraoxaspiro[5.5]undecane, as shown in Scheme 4
[21,22].
This synthesis represent a very significant im-
provement over that used to prepare poly(ortho ester)
I since it is merely necessary to dissolve the mono-
mers in a polar solvent such as tetrahydrofuran and
to add a trace of an acidic catalyst. The reaction
proceeds to completion virtually instantaneously, is
highly reproducible and polymer molecular weights
can be readily adjusted by controlling stoichiometry.
Polymer hydrolysis occurs as shown in Scheme 5
[23]. Even though the hydrolysis produces acidic
hydrolysis products, unlike POE I, it is not au-
tocatalytic because the initial hydrolysis products are
neutral. Details of the hydrolysis mechanism have
been published [24].
Even though the synthesis is via a polycondensa-
Scheme 5.
tion reaction, no small molecule by-products are
evolved so that dense, crosslinked materials can be
prepared by using triols, or higher functionality
hydroxy-containing monomers, as shown in Scheme
6 [25]. The ability to form dense, crosslinked
matrices that completely biodegrade to small, water-
soluble molecules is unique to poly(ortho esters).
Another significant advantage is that mechanical
and thermal properties of the polymer can be ad-
justed by using diols having different degrees of
chain flexibility, so that materials that range from
hard, glassy materials, to materials that are semi-
solids at room temperature can be obtained.
However, because poly(ortho esters) are extremely
hydrophobic materials [26], the amount of water
available to react with the hydrolytically labile ortho
ester linkage is limited and for this reason, under
physiological conditions, the polymer is very stable.
Therefore, means of controlling erosion rates of the
polymer must be devised if devices having erosion
rates that can be varied from a few days to months
are desired.
Because ortho ester linkages are acid-labile [27],
one means of controlling polymer erosion rates is by
lowering the pH at the polymer–water interface. A
considerable amount of work has been expended in
attempts to achieve control of erosion rates by taking
advantage of the acid-sensitivity of ortho ester
linkages by the addition of acidic excipients, for
example suberic acid, to the polymer matrix. This
approach was only marginally successful and for that
reason devices using acidic excipients have never
reached commercialization. Because that work has
been reviewed in considerable detail [1,2], it will not
be summarized here.
4. Poly(ortho ester) III
This polymer was originally developed at the
Stanford Research Institute [28] using venture capital
 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039 1019

Scheme 6.

funding and following the transfer of the technology
to the University of Geneva [29], was then exten-
sively characterized and investigated by this group in
ocular and other applications. This work has recently
been reviewed [30].
The polymer is prepared as shown in Scheme 7
[28]. Because this polymer has a very flexible
backbone, it is a semi-solid material at room tem-
perature. Such materials offer a number of important
advantages. Dominant among these is the possibility
of incorporating therapeutic agents by a simple
mixing procedure at room temperature without the
use of solvents and the fact that such materials can
be easily injected. However, molecular weight must
be limited so that injection through a needle size no
larger than about 20 gauge is possible.
Unlike the synthesis of POE II which is virtually
instantaneous after dissolution of the monomers in
tetrahydrofuran and the addition of a trace of an
acidic catalyst, the synthesis shown in Scheme 6
requires long reflux times and azeotroping out etha-
nol to drive the reaction to high molecular weight.

Scheme 7.
1020 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039
Scheme 8.
Also, unlike the synthesis of POE II, it is very
difficult to achieve reproducible results. The polymer
hydrolyzes as shown in Scheme 8 [31].
A detailed study of the hydrolysis mechanism
using model compounds has established that only the
1 and 2 isomeric esters are produced. Thus, the
absence of the 6-isomer indicates that cleavage
occurs exclusively by an exocyclic cleavage [31].
Even though this material offers a number of
significant advantages, and has shown excellent
biocompatibility in rabbit eyes [32], difficulties in
synthesis and especially difficulties in reproducibly
preparing materials of chosen molecular weights,
has made practical applications difficult and this
material is no longer under development.
5. Poly(ortho ester) IV
POE IV is a modification of POE II that allows
control over erosion rates without the need to add
acidic excipients.
5.1. Polymer synthesis
As already mentioned, POE II is an extremely
hydrophobic material and despite the reactivity of
ortho ester linkages when exposed to water, polymer
hydrolysis is slow and devices fabricated from the
polymer are highly stable. This stability is shown in
Fig. 1, which shows weight loss of a wafer as a
function of time when placed in a pH 7.4 buffer at
37 8C [33].
To make this polymer system adaptable to a wide
range of delivery times, it is essential to devise a
means of reproducibly controlling erosion rates. This
has been achieved by the incorporation of a short
segment based on glycolic acid, or lactic acid into
the polymer backbone [34]. Such segments act as
latent acids, because they readily hydrolyze to gly-
colic, or lactic acids, which then catalyze hydrolysis
of ortho ester linkages in the polymer backbone.
Then, by controlling the concentration of such
segments in the polymer, rate of erosion can be
accurately controlled. Such polymers have been
 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039 1021

Fig. 2. Effect of polymerization time on molecular weight of a

Fig. 1. Weight loss as a function of time for a polymer prepared
from 3,9-dimethylene-2,4,8,10-tetraoxaspiro[5.5]undecane and
1,6-hexanediol. 0.05 M phosphate buffer, pH 7.4, 37 8C (from
Ref. [33]).
poly(ortho ester) prepared from 15 mole% trans-cyclohex-
anedimethanol, 40 mole% 1,6-hexanediol, 40 mole% triethylene
glycol and 5 mole% triethylene glycol monoglycolide (from Ref.
[35]).

after addition of a trace of an acidic catalyst to a

designated as autocatalyzed poly(ortho esters). They tetrahydrofuran solution of the monomers [35]. The
are prepared as shown in Scheme 9 [34]. virtually instantaneous rate of reaction is shown in
The condensation reaction between a diol and a Fig. 2. A detailed characterization of the polymer has
diketene acetal proceed virtually instantaneously been carried out [36].

Scheme 9.
1022 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039

Scheme 10.

The latent acid diol is prepared by the reaction
between a diol and either lactide, or glycolide, as
shown in Scheme 10 [36]. This reaction produces a
range of products in varying concentrations with
n-values varying between 1 and 7, as shown by the
gel permeation chromatogram trace in Fig. 3. To
control erosion rates, the exact structure of the latent
acid diol is not important, and it is the total con-
centration of the a -hydroxy acid segments in the
polymer that determines erosion rates.
5.2. Polymer hydrolysis
Polymer hydrolysis occurs as shown in Scheme 11
[37].
The hydrolysis proceeds in three consecutive
steps. In the first step, the lactic acid or glycolic acid
segment in the polymer backbone hydrolyzes to
generate a polymer fragment containing a carboxylic
acid end-group that will catalyze ortho ester hy-
drolysis. A second cleavage produces free a -hydroxy
acid that also catalyzes hydrolysis of the ortho ester
links. Further hydrolysis of ortho esters then
proceeds in two steps to first generate the diol, or
mixture of diols used in the synthesis and penta-
erythritol dipropionate, followed by ester hydrolysis
to produce pentaerythritol and propionic acid.
Fig. 4 shows polymer weight loss and release of
lactic acid [37]. The most significant finding of this
study is the linearity of weight loss and the concomi-
tant release of lactic and propionic acid. While linear
rate of weight loss alone does not necessarily
indicate surface erosion [38], the concomitant linear
weight loss and release of lactic acid argues convinc-
ingly for a process confined predominantly to the
surface layers of the polymer matrix. The process is
not pure surface erosion because there is a significant
drop in molecular weight of the remaining polymer,
indicating that some hydrolysis is taking place in the
bulk material.
The following process would account for the
observed facts. Initially, due to the highly hydro-
phobic nature of the polymer surface, no hydrolysis
takes place, and an induction period is observed.

Fig. 3. Gel permeation chromatograph of reaction products between lactide and triethylene glycol.
 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039
Scheme 11.
Fig. 4. The relationship between lactic acid release ( ? ) and
weight loss (j) for a poly(ortho ester) prepared from 3,9
diethylidene-2,4,8,10-tetraoxaspiro [5.5] undecane, and a 100 / 70 /
30 mixture of 1,10-decanediol and 1,10-decanediol lactide. 0.13
M, pH 7.4 sodium phosphate buffer at 37 8C (from Ref. [37]).
Gradually, as a result of hydrophilic species gener-
ated by the hydrolysis process, sufficient water
penetrates the surface layers to induce a more rapid
1023
hydrolysis and once steady state has been reached,
constant erosion with concomitant generation of
hydrolysis products takes place. In parallel with this
process, some water penetrates the bulk material and
induces limited polymer chain cleavage. However,
due to the limited number of polymer chains cleaved,
the material remains very hydrophobic and water
concentration in the bulk remains very low.
An erosion process that is confined predominantly
to the surface layers has a number of important
consequences. First, if the drug is well immobilized
in the matrix, its release is controlled by polymer
erosion so that an ability to control polymer erosion
translates into an ability to control rate of drug
release. Second, because drug release is controlled
by polymer erosion, drug release and polymer ero-
sion take place concomitantly and when drug release
has been completed, no polymer remains. Third,
because most of the hydrolysis occurs in the outer
layers of the device, acidic hydrolysis products
diffuse away from the device and do not accumulate
in the bulk material. Thus, the interior of the matrix
1024 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039

does not become highly acidic, as is the case with

poly(lactide-co-glycolide) copolymers, or poly(lactic
acid), and acid-sensitive drugs can be released
without loss of activity.
An ongoing study is designed to provide a direct
measurement of the pH in the matrix as a function of
depth [39]. To do so, pH-sensitive nitroxide radicals
having different pKa values are incorporated into the
matrix and the EPR signal analyzed. Such data
provide a direct measure of the pH in different layers
within the device and the mobility of the nitroxide
radicals provides a measure of water penetration into
the matrix. Preliminary data are consistent with a
surface erosion process and indicate that the pH in
the outer eroding layer is about 5.4 and the pH in the
matrix interior is about 6.4. The exact values may
require a slight adjustment pending a determination Fig. 5. Glass transition temperature of 3,9-diethylidene-2,4,8,10-
tetraoxaspiro[5.5]undecane, trans-cyclohexanedimethanol, 1,6-
of the pKa of the nitroxide radicals in water and in hexanediol polymer as a function of mole% 1,6-hexanediol (from
the interior of the hydrophobic polymer environment. Ref. [23]).

5.3. Control of mechanical and thermal properties

One of the most desirable attributes of POE IV

synthesis, shown in Scheme 9, is the ability to vary
independently thermal and mechanical properties and
erosion rates. Mechanical and thermal properties can
be varied by choosing the appropriate R-group in the
diol and in the latent acid. The use of rigid diols,
produces materials having high glass transition tem-
peratures and hence high fabrication temperatures,
while the use of flexible diols produces materials
having low glass transition temperatures and hence
low fabrication temperatures. The use of very flex-
ible diols produces materials that are semi-solid
materials at room temperature and these will be
covered later in this chapter.
The effect of diol structure on the glass transition
temperature is shown in Figs. 5 and 6. Fig. 5 shows
the effect of varying the ratio of a rigid diol, trans- Fig. 6. Effect of diol chain length on the glass transition tempera-
ture of polymers prepared from 3,9-diethylidene-2,4,8,10-tetraox-
cyclohexanedimethanol and a flexible diol, 1,6-hex-
aspiro[5.5]undecane and a,v -diols (from Ref. [40]).
anediol on the glass transition temperature [23].
Clearly, by an appropriate choice of diol ratio, any
glass transition temperature between 110 and 20 8C
can be selected. Fig. 6 shows the effect of chain does have a significant effect on the glass transition
length of a diol on the glass transition temperature temperature, as shown in Fig. 7 [36]. Thus, both diol
[40]. structure and latent acid diol structure must be
Figs. 5 and 6 were generated with polymers considered when designing polymers having desired
having no latent acid. However, the latent acid diol thermal and mechanical characteristics.
 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039
Fig. 7. Glass transition temperatures for poly(ortho ester) prepared
from 3,9-diethylidene-2,4,8,10-tetraoxaspiro[5.5]undecane and n-
octanediol, n-decanediol and n-dodecanediol, each with 5, 10 and
20 mole% of the corresponding lactide.
5.4. Polymer fabrication
Polymer fabrication is a crucial component in the
development of drug delivery devices and poorly
fabricated devices, even with the best of polymers,
can exhibit very poor drug release kinetics and a
significant burst and non-linear kinetics are a com-
mon consequence of a poorly fabricated device.
The most desirable situation is one where finely
micronized drug particles are individually sur-
rounded by polymer and there is minimal particle-to-
particle contact. Although this is not always the case,
drug loading should, in general, be no higher than
about 30 wt%, although devices with good release
characteristics were obtained with drug loading as
high as 70 wt%.
Poly(ortho esters) are excellent thermoplastic ma-
terials that can be easily fabricated by extrusion,
injection molding or compression molding. The
polymer is very thermally stable as shown in Fig. 8
[35].
The polymer is also soluble in tetrahydrofuran,
ethyl acetate and methylene chloride, so that mi-
croparticles using conventional microencapsulation
techniques can easily be prepared. In developing
1025
Fig. 8. Effect of thermal processing on molecular weight of a
poly(ortho ester) prepared from 3,9 diethylidene-2,4,8,10-tetraox-
aspiro[5.5]undecane, trans-cyclohexanedimehanol, 1,6-hex-
anediol, triethylene glycol and triethylene glycol glycolide (15 /
40 / 40 / 5). Processing parameters shown in legend (from Ref.
[35]).
microencapsulation methods, it is essential to fabri-
cated dense microspheres since porous microspheres
will, in general, show an initial drug burst and drug
release kinetics that are not always linear.
A highly desirable fabrication method is a sol-
ventless thermal fabrication method. One such pro-
cedure is an extrusion method where finely ground
polymer and a micronized protein are intimately
mixed and then extruded into thin strands at tempera-
tures low enough so that the stability of the incorpo-
rated therapeutic agent is not compromised. The
ability to control both mechanical and thermal
properties poly(ortho ester) makes an extrusion fabri-
cation method possible, and the method has been
applied to fabricating thin strands with incorporated
proteins without loss of activity [42].
5.5. Polymer stability
As shown in Fig. 9, poly(ortho esters) have
excellent stability, and when stored under anhydrous
conditions, they are stable at room temperature. The
particular polymer used in this stability study is a
hydrophilic polymer containing 40 mole% latent acid
that has been ground to produce microparticles, thus,
greatly increasing surface area. This is a very rapidly
1026 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039
Fig. 9. Stability of a polymer prepared from 3,9 diethylidene-
2,4,8,10-tetraoxaspiro[5.5]undecane, trans-cyclohex-
anedimethanol, triethylene glycol and triethylene glycol glycolide
(35 / 25 / 40) stored at room temperature and under anhydrous
conditions.
eroding polymer that would completely erode in a
matter of a few days if placed in an aqueous buffer.
The polymer is also relatively stable when steril-
ized by irradiation [35]. As shown in Fig. 10, there is
a decrease in molecular weight after irradiation, but
the decrease is of the same order of magnitude as
that observed with other bioerodible polymers. Be-
Fig. 10. Effect of g -irradiation at 24 kGy on a poly(ortho ester)
prepared from 3,9 diethylidene-2,4,8,10-tetraoxaspiro[5.5]un-
decane, trans-cyclohexanedimethanol, 1,6-hexanediol, triethylene
glycol and triethylene glycol glycolide (15 / 40 / 40 / 5). Polymer
stored post-irradiation at 5 8C in a dessicator (from Ref. [35]).
cause irradiation generates free radicals that in a
solid matrix can be long-lived, stability studies were
extended to 3 months to determine if post-irradiation
chain cleavage takes place. As shown in Fig. 10 [35],
the polymer is stable after the initial drop in molecu-
lar weight and EPR studies have shown that free
radicals dissipate in less than 24 h.
5.6. Control of erosion rate
From a consideration of the hydrolysis mecha-
nism, it can be inferred that two factors will affect
erosion rates. One is the concentration of the latent
acid in the polymer backbone and the other is the
hydrophilicity of the matrix that can be controlled by
using triethylene glycol as one of the diols used in
the synthesis.
Polymer hydrophilicity has a major influence on
erosion rate of the polymer, as illustrated in Fig. 11,
which shows the rate of erosion of two polymers
having the same concentration of latent acid, but
differing degrees of hydrophilicity [35].
Fig. 11. Weight loss as a function of time for a poly(ortho ester)
prepared from 3,9 diethylidene-2,4,8,10-tetraoxaspiro[5.5]un-
decane, and a mixture of (j) 25 mole% trans-cyclohex-
anedimethanol, 45 mole% 1,6-hexanediol, 5 mole% triethylene
glycol and 25 mole% triethylene glycol glycolide, (d) 75 mole%
trans-cyclohexanedimethanol and 25 mole% trans-cyclohex-
anedimehanol. 0.13 M, pH 7.4 sodium phosphate buffer at 37 8C
(from Ref. [35]).
 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039
Polymer hydrophilicity and latent acid concen-
tration are interrelated, and highly hydrophobic
matrices are relatively insensitive to the concen-
tration of latent acid, while hydrophilic matrices are
very sensitive to the concentration of latent acid.
This sensitivity is shown in Fig. 12. With this
particular system, there is no erosion for 30 days in
the absence of latent acid, but as little as 1 mole%
latent acid changes erosion times to about 3 weeks.
5.7. Control of drug delivery rates
In discussing drug delivery, it is of interest to
consider two types of drugs, low molecular weight
water-soluble drugs and macromolecular drugs.
5.7.1. Low molecular weight, water-soluble drugs
This represents the most difficult situation because
water-soluble drugs will increase the hydrophilicity
of the matrix and thus accelerate erosion rate.
Further, if there is matrix swelling, release by
diffusion is possible. However, as shown in Fig. 13,
this is not the case with poly(ortho esters) and the
erosion-controlled release of 5-FU can be demon-
strated by the concomitant 5-FU release and device
Fig. 12. Weight loss of a polymer prepared from 3,9 diethylidene-
2,4,8,10-tetraoxaspiro [5.5] undecane and (m) trans-cyclohex-
anedimethanol, triethylene glycol (70 / 30 and (j) trans-cyclo-
hexanedimethanol, triethylene glycol, triethylene glycol glycolide
(70 / 29 / 1). Extruded strands 1 3 10 mm, 0.01 M phosphate
buffered saline, pH 7.4, 37 8C.
1027
Fig. 13. Polymer weight loss (d) and 5-fluorouracil (5-FU)
release (j) from a polymer prepared from 3,9-diethylidene-
2,4,8,10-tetraoxaspiro [5.5] undecane and 1,3-propanediol / tri-
ethylene glycol diglycolide (90 / 10). Drug loading 20 wt%, 0.05
M phosphate buffer, pH 7.4, 37 8C (from Ref. [44]).
weight loss [41]. In this particular case, 5-FU
material balance is a little low, but there is little
doubt that the predominant mechanism controlling
drug release is erosion and not diffusion.
Further, as shown in Fig. 14, rate of drug release
is proportional to drug loading. These data are
consistent with a surface erosion process and provide
additional evidence for the hydrolysis and erosion
study described previously [37].
5.7.2. Macromolecular drugs
As already mentioned under fabrication, solvent-
less incorporation of peptides and proteins into thin,
extruded strands, is a particularly attractive method,
if the extrusion can be carried out at temperatures
lower than the temperature at which the protein
begins to denature.
Such a study has been carried out with a model
protein FITC-BSA that was mixed with polymer and
extruded into 1 mm strands cut to 10 mm lengths.
The extrusion was carried out at 75 8C. The strands
were then placed in a pH 7.4 phosphate buffer at
37 8C and weight loss and FITC-BSA release de-
terminations carried out. Results of this study are
shown in Fig. 15 [42].
1028 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039
Fig. 14. Effect of loading on amount of 5-fluorouracil (5-FU)
released from a poly(ortho ester) prepared from 15 mole% trans-
cyclohexanedimethanol, 40 mole% 1,6-hexanediol, 40 mole%
triethylene glycol and 5 mole% triethylene glycol glycolide. (j)
5.5 wt% 5-FU (14 mg), (h) 11.6 wt% 5-FU (28 mg), (d) 23.6
wt% 5-FU (56 mg), 0.01 M phosphate buffered saline, pH 7.4,
37 8C (from Ref. [44]).
Fig. 15. Release of FITC-BSA (d) and weight loss (j) from a
poly(ortho ester) prepared from 3,9-diethylidene-2,4,8,10-tetraox-
aspiro[5.5]undecane, 1,4-pentanediol and 1,6-hexanediol glycolide
(100 / 95 / 5). Strands, 1 3 10 mm, extruded at 70 8C. 0.01 M
phosphate buffered saline, pH 7.4, 37 8C. FITC-BSA loading 15
wt% (from Ref. [42]).
Three major features are notable. First, there is no
burst despite the fact that 15 wt% of a water-soluble
material has been incorporated, second, there is a
significant lag-time, and third, the rate of BSA
release is linear and polymer weight loss and BSA
release occur concomitantly. These are important
findings, but the long lag-time is not optimal in most
applications. For this reason, means of reducing the
length of the lag-time were investigated.
We have found that the addition of very small
amounts of poly(ethylene glycol), molecular weight
2 kDa, markedly changes kinetics of BSA release
[42]. Because poly(ethylene glycol) had to be phys-
ically mixed into the powdered polymer prior to
extrusion, we have chosen a solid poly(ethylene
glycol) and hence the choice of 2 kDa. The change
in kinetics after addition of 1 wt% poly(ethylene
glycol) is shown in Fig. 16. These data indicate that
poly(ethylene glycol), even in very small amounts is
effective in reducing the lag-time. As could be
anticipated, the incorporation of a hydrophilic moiety
onto the polymer will accelerate erosion rate and
hence release of FITC-BSA.
Fig. 16. Effect of 2 kDa poly(ethylene glycol) (PEG) on release
of FITC-BSA from a poly(ortho ester) prepared from 3,9-dieth-
ylidene-2,4,8,10-tetraoxaspiro[5.5]undecane, 1,3-propanediol and
triethylene glycol glycolide (100 / 85 / 15). (j) pure polymer, (d)
1 wt% PEG, (Strands, 1 3 10 mm, extruded at 70 8C, 0.01 M
phosphate buffered saline, pH 7.4, 37 8C. FITC-BSA loading 15
wt% (from Ref. [42]).
 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039
6. Semi-solid polymers
To prepare semi-solid materials, it is necessary to
use highly flexible diols, and in order to allow direct
injection of such materials, their viscosity must be
limited by preparing polymers having molecular
weights no higher than about 5 kDa. While it is, of
course, possible to prepare a great variety of semi-
solid polymers by using a range of highly flexible
diols, in order to limit toxicology studies preparatory
to regulatory approval, this work was confined to
two diols, triethylene glycol and 1,10-decanediol.
Semi-solids based on triethylene glycol produce
somewhat hydrophilic materials while semi-solids
based on 1,10-decanediol produce hydrophobic ma-
terials.
The most significant advantage of semi-solid
materials is that therapeutic agents can be readily
incorporated by mixing at ambient temperature and
without the use of solvents. Such conditions are mild
enough to incorporate even the most sensitive thera-
peutic agents. Mixing can be accomplished on a
small scale by using a mortar and pestle, but on a
somewhat larger scale it is better carried out using a
three-roll mill [43].
Scheme 12.
1029
6.1. Polymer molecular weight control
Polymer molecular weight control can be achieved
by using an excess of diol relative to the diketene
acetal, but can also be accomplished by using a
chain-stopper. In this approach, a calculated amount
of a monofunctional alcohol is used [44]. As shown
in Scheme 12, when n-decanol is used as a chain-
stopper in combination with 1,10-decanediol, both
polymer ends have n-decanol residues so that a
chain-stopped material is somewhat more hydro-
phobic relative to a stoichiometry-controlled material
that has terminal hydroxyl groups.
The use of chain-stoppers allows excellent and
reproducible molecular weight control by varying the
ratios of 1,10-decanediol to n-decanol as shown in
Fig. 17 [44]. The existence of terminal methyl
groups has been established by 1 H NMR studies
[44].
Because the principal means of administration of
semi-solid materials is by injection, preparation of
materials having reproducible viscosities is impor-
tant. Synthesis reproducibility as measured by
Brookfield viscosity, for a number of different
preparations is shown in Fig. 18 [45]. Clearly, the
1030 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039

Fig. 17. Effect of n-decanol on the molecular weight of a Fig. 19. Room temperature stability for a polymer prepared from
poly(ortho ester) prepared from 3,9-diethylidene-2,4,8,10-tetraox- 3,9-diethylidene-2,4,8,10-tetraoxaspiro[5.5]undecane, triethylene
aspiro[5.5]undecane, 1,10-decanediol and 1,10-decanediol lactide glycol, and triethylene glycol glycolide (60 / 50 / 50). Material
(100 / 70 / 30 (from Ref. [44]). stored under anhydrous conditions (from Ref. [45]).

semi-solid polymer after storage at room temperature

under anhydrous conditions for 9 months [45]. As
with the solid polymers, within experimental error,
there is no change.
Fig. 20 shows the effect of irradiating the polymer
at a dose of 22.9–25.6 kGy [45]. Within experimen-
tal error, no changes in molecular weight could be
detected. This is consistent with previous finding for
POE III and indicates that low molecular weight

Fig. 18. Variation in Brookfield viscosity for nine typical prepara-

tions at 25 8C. Injectable formulation prepared from 3,9-dieth-
ylidene-2,4,8,10-tetraoxaspiro[5.5]undecane, triethylene glycol,
and triethylene glycol glycolide (60 / 50 / 50). Formulation contains
20 wt% monomethoxy poly(ethylene glycol), molecular weight
550 (from Ref. [45]).

synthesis is sufficiently reproducible to assure that

the same viscosity materials can be repeatedly
prepared.
Fig. 20. Stability of a polymer prepared from 3,9-diethylidene-
6.2. Polymer stability 2,4,8,10-tetraoxaspiro[5.5]undecane, triethylene glycol, and tri-
ethylene glycol glycolide (60 / 50 / 50) irradiated at 24 kGy (from
Fig. 19 shows changes in molecular weight of a Ref. [45]).
 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039
polymers, unlike their high molecular weight ana-
logues, do not significantly change molecular weight
on irradiation.
6.3. Drug release
In developing semi-solid formulations for drug
release, two formulations are under development.
One formulation is based on triethylene glycol, and
the other on 1,10-decanediol.
6.3.1. Triethylene glycol semi-solids
Semi-solid materials based on triethylene glycol
are somewhat hydrophilic and are designed for
relatively short delivery times.
6.3.1.1. Post-surgical pain control
Post-surgical pain originates at the site of surgery,
a relatively small area, so that treatment by the
systemic administration of analgesic agents with the
well-known side-effects such of drowsiness, consti-
pation and others does not represent optimal treat-
ment. Instead, the administration of a sustained
release formulation at the site of surgery is more
appropriate.
Poly(lactic acid) microspheres have been evalu-
ated as a sustained delivery system for local anes-
thetics such as butamben, tetracaine and dibucaine
[46,47]. The use of poly(lactic-co-glycolic acid)
copolymers was also investigated as a delivery
system for the release of bupivacaine and dexa-
methasone using a rat sciatic nerve blockage model
[48,49]. However, in many applications, such as
treatment of post-operative pain, an analgesic activi-
ty of only a few days would be desirable. Because
erosion times of poly(lactic acid) is measured in
months, and even years and the erosion time of
poly(lactic-co-glycolic acid) copolymers is measured
in weeks, these erosion times are clearly not optimal
for short term therapy. Because the erosion time of
poly(ortho ester) can be as short as a few days days
and injectable semi-solid formulations can be pre-
pared, such polymers are currently under investiga-
tion as delivery systems for analgesic agents, such as
bupivacaine, or mepivacaine.
Because anesthetic and analgesic agents such as
bupivacaine are cardiotoxic, it is important to control
the amount released at the post-operative site to
achieve analgesia while blood levels remain well
1031
Fig. 21. Release of bupivacaine from a polymer prepared from
3,9-diethylidene-2,4,8,10-tetraoxaspiro[5.5]undecane, triethylene
glycol, and triethylene glycol glycolide (60 / 50 / 50) containing 50
wt% monomethoxy poly(ethylene glycol) and 10 wt% bupivacaine
free base. Phosphate buffer, pH 7.4, 37 8C (from Ref. [45]).
below the cardiotoxic level. Fig. 21 [45] shows the
desired short time release using a formulation com-
prised of 40 wt% polymer, 50 wt% monomethoxy
poly(ethylene glycol) M Wt 550 and 10 wt% bupi-
vacaine free base. The poly(ethylene glycol) deriva-
tive was added to improve ease of injection.
The in vivo functionality of a bupivacaine delivery
system has been investigated using a rat sciatic nerve
model. In this model, the sciatic nerve in the hind
limb of the rat is exposed and the formulation
containing bupivacaine placed in the close proximity
of the nerve. The hind limb is then electrically
stimulated. No response is indicative of total nerve
block while dragging of the hind limb is indicative of
motor block. Results of that study are shown in Fig.
22 [50]. Clearly, sustained motor block has been
achieved while blood levels remained far below the
recognized cardiac and CNS toxic level of 2–4
mcg / ml [51]. Additional in vivo data are presented
in the following chapter.
Another important application of semi-solid ma-
terials is in the delivery of peptides and proteins.
Clearly, an appealing feature of semi-solids is that
therapeutic agents can be incorporated by a simple,
room-temperature mixing procedure without the need
for solvents. Such a procedure should not com-
1032 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039
Fig. 22. Bupivacaine blood levels and duration of motor block
using a polymer prepared from 3,9-diethylidene-2,4,8,10-tetraox-
aspiro[5.5]undecane, triethylene glycol, and triethylene glycol
glycolide (60 / 50 / 50) in a rat sciatic nerve model (from Ref.
[50]).
promise the integrity of even the most fragile
therapeutic agents.
This work is currently underway. A preliminary
result showing the release of FITC-BSA is shown in
Fig. 23 [50]. In this particular case a hydrophilic
polymer was used to release FITC-BSA, a water-
soluble material. Thus, immobilization of BSA is not
Fig. 23. Release of BSA from a semi-solid poly(ortho ester)
prepared from 3,9-diethylidene-2,4,8,10-tetraoxaspiro[5.5]un-
decane, and a 95 / 5 mole% mixture of triethylene glycol and
triethylene glycol glycolide. BSA loading 4 wt%, 0.1 M phosphate
buffer, pH 7.3, 37 8C (from Ref. [50]).
as good as when a hydrophobic, solid polymer was
used. Nevertheless, good linear kinetics were ob-
tained. Although no weight-loss determinations were
made, visually, there was no semi-solid materials
remaining after all the FITC-BA has been released.
6.3.2. 1,10 -decanediol semi-solids
Semi-solid materials based on 1,10-decanediol are
very hydrophobic, and their intended application is
in delivery times significantly longer than those
achievable using semi-solid materials based on tri-
ethylene glycol.
This material is under investigation in the treat-
ment of periodontal disease, in ocular applications
and in estrus synchronization.
6.3.2.1. Periodontal disease
Destructive periodontitis is associated with patho-
genic flora that resides in deep periodontal pockets
[52]. While conventional non-surgical treatments
such as mechanical scaling and root planing are in
many cases successful, there are situations where
refractory periodontitis is encountered and those
cases can best be treated by the systemic administra-
tion, or via rinses of an antimicrobial agents such as
tetracycline. However, the concentration of an anti-
microbial agent in the periodontal pocket can be
significantly enhanced by placing a sustained deliv-
ery device into the periodontal pocket and a number
of such devices are on the market [53].
In this application, an injectable 1,10-decanediol
formulation with 1,10-decanediol lactide latent acid
semi-solid formulation containing tetracycline was
used. Such a formulation can be injected directly into
the periodontal pocket and was investigated in
human volunteers.
The in vitro tetracycline free base release is shown
in Fig. 24 [45,54]. As the figure shows, excellent
linear release for the desired length of time has been
achieved. It is this formulation that was used in the
human volunteer study.
Preliminary findings indicated that the formula-
tions were very well tolerated with no pain during
application and no signs of irritation or discomfort
after treatment. Further, as shown in Fig. 25 [55],
concentrations of tetracycline in the periodontal
pocket were well above the minimum inhibitory
concentration (MIC) of 1 mg / ml even at day 11.
 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039
Fig. 24. Cumulative release of tetracycline free base from a
polymer prepared from 3,9-diethylidene-2,4,8,10-tetraoxas-
piro[5.5]undecane, decanediol and decanediol lactide (100 / 50 /
50). In PBS at pH 7.4 and 37 8C. Drug loading 20 wt% (from Ref.
[55]).
Fig. 25. Average tetracycline concentration in gingival crevicular
fluid of devices that have been retained in the periodontal pocket
after the placement of a delivery system prepared from 3,9-
diethylidene-2,4,8,10-tetraoxaspiro[5.5]undecane, 1,10-decanediol,
and 1,10-decanediol lactide (100 / 70 / 30) containing 10 wt%
tetracycline free base. Fraction of retained devices as shown above
bars (from Ref. [55]).
However, as also shown in Fig. 25, retention of the
device in the pocket was not satisfactory and at day
4, nine out of 24 sites were positive for tetracycline,
at day 7, 6 / 24 sites were positive for tetracycline and
1033
at day 11, two out of 24 sites were positive for
tetracycline. Results of this study suggest that the
injection of a tetracycline formulation into a
periodontal pocket is a promising treatment modali-
ty, but retention of the device must be significantly
improved before such a system can become a
successful treatment of periodontitis.
6.3.2.2. Ocular applications
The 1,10-decanediol semi-sold material with 1,10-
decanediol lactide as the latent acid has been investi-
gated in ocular applications. In a biocompatibility
study using rabbits, the polymer was injected sub-
conjunctivally, intravitreally, intracamerally and sup-
rachoroidally. In all of these applications, the poly-
mer exhibited outstanding biocompatibility. Addi-
tionally, the polymer exhibits surprisingly long life-
times and it appears that devices having a lifetime of
a few months are feasible [56]. Final results of this
study are not available at this writing.
6.3.2.3. Estrus synchronization in sheep
The objective of estrus synchronization in sheep is
to make all female sheep fertilizable at chosen times.
While a number of non-degradable devices are on
the market either as vaginal sponges or ear-implants,
a bioerodible device would clearly be a significant
labor-saving device. An injectable 1,10-decanediol
formulation containing fluorogestone acetate is cur-
rently under investigation as a 14 day delivery device
and in vivo testing is in progress [57].
7. Block copolymers with poly(ethylene glycol)
Block copolymers of poly(ortho esters) and poly-
(ethylene glycol) were prepared and are currently
under investigation as a matrix for drug delivery and
as micelles.
7.1. Synthesis
AB and ABA block copolymers can be prepared
as shown in Schemes 13 and 14.
7.1.1. Release of proteins
As discussed under 5.5.2, the addition of very
small amounts of poly(ethylene glycol), molecular
1034 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039
Scheme 13.
weight 2 kDa, was effective in reducing the lag-time
in BSA delivery from extruded strands. However,
poly(ethylene glycol) had to be physically mixed into
the powdered polymer prior to extrusion, and the
addition of very small amounts of an excipient
represents an additional complications. For this
reason, we have investigated the use of an AB-block
copolymer containing 6 wt% of a 2-kDa poly-
(ethylene glycol) in the polymer chain as a matrix for
BSA release.
Release of FITC-BSA from this AB block co-
polymer is shown in Fig. 26 [42]. Clearly, there has
been a very significant improvement, and the use of
AB and ABA block copolymers comprised of poly-
(ethylene glycol) and poly(ortho esters) is now under
active development as release matrices for the deliv-
ery of proteins and other therapeutic agents.
7.1.2. Micelles
When amphiphilic polymers are dispersed in
water, they spontaneously self-assemble into micellar
structures [58]. Such structures have a number of
important applications, but perhaps the most im-
portant application is tumor targeting using the
enhanced permeability and retention effect [59].
Briefly, this effect is based on differences in per-
meability between normal vasculature and that feed-
ing tumors. Because vasculature feeding tumors in
newly formed, it has an incompletely formed endo-
thelium and is thus more permeable than normal
vasculature. Thus, macromolecules, or micelles hav-
ing dimension in the order of 50 nanometers can not
escape normal vasculature, but are able to escape
tumor vasculature and accumulate in the tumor, and
because tumors have poor lymphatic drainage, such
agent will be retained in the tumor.
In the actual application of micelles in tumor
targeting, a hydrophobic anticancer agent is phys-
ically entrapped in the hydrophobic core of the
micelle and the formulation injected intravenously.
The physical entrapment of drugs is preferable to one
where the drug is chemically attached to the hydro-
 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039 1035

Scheme 14.

phobic portion of the micelle, because the drug has

not been chemically modified thus making regulatory
approval easier.
Micelles based on poly(ethylene glycol) and poly-
(ortho esters) are of particular interest because poly-
(ortho esters) are highly hydrophobic materials so
that the entrapment of highly hydrophobic drugs
such as doxorubicin, or taxol, should be very effi-
cient. Additionally, poly(ortho esters) are pH-sensi-
tive and because the pH in the interior of tumors is
about 5.5, once internalized, the micelles should
rapidly degrade and deliver their payload.
Micelles having a structure shown in Scheme 15
are currently under investigation [60]. In this par-
ticular composition, a poly(ortho ester) segment
Fig. 26. Release of FITC-BSA from an AB-block copolymer
containing 6 wt% 2 kDa poly(ethylene glycol). Poly(ortho ester)
prepared from 3,9-diethylidene-2,4,8,10-tetraoxaspiro[5.5]un-
decane, and a 85 / 15 mole% mixture of 1,3-propanediol and
triethylene glycol glycolide. Strands, 1 3 10 mm, extruded at
70 8C, 0.01 M phosphate buffered saline, pH 7.4, 37 8C. FITC-
BSA loading 15 wt% (from Ref. [42]). Scheme 15.
1036 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039
based on trans-cyclohexanedimethanol was chosen
because this diol will produce the most hydrophobic
structure.
Ongoing work has shown that the entrapment
efficiency of ABA block copolymers is higher than
that AB block copolymers based on other biodegra-
dable segments. Optimization studies are currently
ongoing, but taxol has been entrapped at a loading as
high as 40 wt% which is significantly higher than
that achieved with other bioerodible micelles using
doxorubicin. Unfortunately, no direct comparison
with taxol is available.
The stability of these micelles measured by mi-
celle size as a function of pH and in the presence of
BSA has been investigated and is shown in Fig. 27.
Thus, at the pH encountered in tumor cells, the
micelles will rapidly loose their micellar structure
and deliver the entrapped drug, in this case taxol, to
the tumor.
When micelles are injected, they undergo a very
significant dilution. For this reason, they must have a
very low critical micelle concentration (CMC). We
have determined that this value is in the order of
10 24 g / l which is low enough to assure that upon
injection, micellar structure will be maintained. Light
scattering measurements have shown that micelles in
Fig. 27. Stability of micelles prepared from poly(ethylene gly-
col)–poly(ortho ester)–poly(ethylene glycol) (5000–4000–5000)
ABA block copolymers. (m) pH 7.4, 10 mM phosphate buffer,
37 8C. (j) pH 5.5, 10 mM citrate buffer, 37 8C. (d) pH 7.4, 10
mM phosphate buffer containing bovine serum albumin, 37 8C.
the desired range of 50–80 nanometers can be
prepared.
Thus, these micelles appear to be a promising
delivery system and are currently under active
development.
8. Conclusions
Poly(ortho esters) have evolved through a number
of families to the current version, POE IV and this
polymer has significant potential for producing use-
ful, commercially relevant bioerodible drug delivery
products. The success of POE IV is largely based on
the versatility of the synthesis that allows excellent
control over erosion rates by selecting the proper
ratios of diol to latent acid diol, and excellent control
of mechanical properties by appropriate selection of
diols used in the synthesis.
Also of great importance is ease of synthesis and
synthesis reproducibility, thermal stability that makes
compression, extrusion or injection molding possible
and ability to withstand radiation sterilization with-
out excessive loss of molecular weight. Most im-
portantly, the polymer can be stored under anhydrous
conditions at room temperature for long periods of
time. In vitro and in vivo studies, have shown that
the polymer erodes to completion and drug depletion
coincides with total polymer erosion.
GLP toxicology studies for a triethylene glycol
semi-solid polymer have been completed and an IND
for post-operative pain control filed. An initial
human clinical trial has been completed and a Phase
I human clinical trial using injectable, semi-solid
formulation is now in the advanced planning stages.
Acknowledgements
It has been over 30 years since poly(ortho esters)
have been under development, and a large number of
individuals have made major contributions to the
development of POE I, POE II and POE III. Because
this review has concentrated on POE IV, the authors
wish to acknowledge the contributors to the develop-
ment of this polymer system, at AP Pharma, at the
University of Geneva and at the University of Gent.
At AP Pharma, major contributions were made by Dr
 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039
Hui-Rong Shen, and Dr Ellen Qi Li. At the Universi-
ty of Geneva, major contributions were made by Dr
Suzanne Einmahl and Dr Alexandra Rothen-
Weinhold. At the University of Gent, the micelle
work was carried out by Professor Etienne Schacht
and Dr Veska Toncheva.
References
[1] J. Heller, Poly(ortho esters), Adv. Polym. Sci. 107 (1993)
41–92.
[2] J. Heller, R. Gurny, Poly(ortho esters), in: E. Mathiowitz
(Ed.), Encyclopedia of Controlled Drug Delivery, vol. 2,
John Wiley & Sons, 1999, pp. 852–874.
[3] N.S. Choi, J. Heller, Polycarbonates, U.S. Patent 4,079,038.
[4] N.S. Choi, J. Heller, Drug delivery devices manufactured
from poly(ortho esters) and poly(ortho carbonates), U.S.
Patent 4,093,709.
[5] N.S. Choi, J. Heller, Structured ortho ester and ortho
carbonate delivery devices, U.S. Patent 4,131,648.
[6] N.S. Choi, J. Heller, Erodible agent releasing device com-
prising poly(ortho esters) and poly(ortho carbonates), U.S.
Patent 4,138,344.
[7] N.S. Choi, J. Heller, Novel ortho ester polymers and
orthocarbonate polymers, U.S. Patent 4,180,646.
[8] L.M. Vistnes, E.E. Schmitt, G.A. Ksander, E.H. Rose, W.J.
Bankelhol, C.L. Coleman, Evaluation of a prototype thera-
peutic system for prolonged continuous topical delivery of
homosulfanilamide in the management of Pseudomonas burn
wound sepsis, Surgery 79 (1976) 690–696.
[9] R.C. Capozza, L. Sendelbeck, W.J. Balkenhol, Preparation
and evaluation of a bioerodible naltrexone delivery system,
in: R.J. Kostelnik (Ed.), Polymeric Delivery Systems, Gor-
don and Breach, New York, 1978, pp. 59–73.
[10] G. Benagiano, E. Schmitt, D. Wise, M. Goodman, Sustained
release hormonal preparations for the delivery of fertility
regulating agents, J. Polym. Sci., Pol. Symp. 66 (1979)
129–148.
[11] E. Solheim, E.M. Pinholt, G. Bang, E. Sudmann, Regenera-
tion of calvarial defects by composite of bioerodible poly-
orthoester and demineralized bone in rats, J. Neurosurg. 76
(1992) 275–279.
[12] E. Solheim, E.M. Pinholt, G. Bang, E. Sudmann, Inhibition
of heterotopic osteogenesis in rats by a new bioerodible
system for local delivery of indomethacin, J. Bone Joint
Surg. Am. 74 (1992) 705–712.
[13] E. Solheim, E.M. Pinholt, G. Bang, E. Sudmann, Effect of
local hemostatics on bone induction in rats: A comparative
study of bone wax, fibrin–collagen paste and bioerodible
polyorthoester with and without gentamicin, J. Biomed.
Mater. Res. 26 (1992) 791–800.
[14] E. Solheim, E.M. Pinholt, R. Andersen, G. Bang, E. Sud-
1037
mann, The effect of a composite of polyorthoester and
demineralized bone on the healing of large segmental defects
of the radius in rats, J. Bone Joint Surg. Am. 74 (1992)
1456–1463.
[15] E.M. Pinholt, E. Solheim, G. Bang, E. Sudmann, Bone
induction by composites of bioresorbable carriers and de-
mineralized bone in rats: A comparative study of fibrin–
collagen paste, fibrin sealant and polyorthoester with gen-
tamicin, J. Oral Maxillofac. Surg. 50 (1992) 1300–1304.
[16] B. Sudmann, O.G. Anfinsen, M. Rait, G. Bang, R. Koppang,
S.O. Stolen, H.S. Koppang, E. Sudmann, Assessment in rats
of a new bioerodible wax-like polymer, Acta Orthop. Scand.
64 (1993) 336–339.
[17] M. Aaboe, E.M. Pinholt, E. Hansen-Hjorting, E. Sondheim,
F. Praetorius, Guided tissue regeneration using degradable
and non-degradable membranes in rat tibia, Clin. Oral Impl.
Res. 4 (1993) 172–176.
[18] E. Solheim, E.M. Pinholt, R. Andersen, G. Bang, E. Sud-
mann, Local delivery of indomethacin by a polyortho ester
inhibits reossification of experimental bone defects, J.
Biomed. Mater. Res. 29 (1995) 1141–1146.
[19] O. Bush, E. Solheim, G. Bang, K. Tornes, Guided tissue
regeneration and local delivery of insulin-like growth factor I
by bioerodible polyorthoester membranes in rat calvarial
defects, Int. J. Oral Maxillofac. Implants 11 (1996) 498–505.
[20] J. Heller, Development of poly(ortho esters) – a historical
overview, Biomaterials 11 (1990) 659–665.
[21] J. Heller, D.W.H. Penhale, R.F. Helwing, Preparation of
poly(ortho esters) by the reaction of ketene acetals and
polyols, J. Polym. Sci., Polym. Lett. Ed. 18 (1980) 82–83.
[22] S.Y. Ng, D.W.H. Penhale, J. Heller, Poly(ortho esters) by the
addition of diols to a diketene acetal, Macromolecular
Synthesis 11 (1992) 23–26.
[23] J. Heller, D.W.H. Penhale, B.K. Fritzinger, J.E. Rose, R.F.
Helwing, Controlled release of contraceptive steroids from
biodegradable poly(ortho esters), Contracept. Deliv. Syst. 4
(1983) 43–53.
[24] J. Heller, S.Y. Ng, D.W.H. Penhale, B.K. Fritzinger, L.M.
Sanders, R.A. Burns, M.G. Gaynon, S.S. Bhosale, The use of
poly(ortho esters) for the controlled release of 5-fluorouracil
and a LHRH analogue, J. Controlled Release 6 (1987)
217–224.
[25] J. Heller, B.K. Fritzinger, S.Y. Ng, D.W.H. Penhale, In vitro
and in vivo release of levonorgestrel from poly(ortho esters)
II. Crosslinked polymers, J. Controlled Release 1 (1985)
233–238.
[26] T.H. Nguyen, K.J. Himmelstein, T. Higuchi, Some equilib-
rium and kinetics of water sorption in poly(ortho esters), Int.
J. Pharm. 25 (1985) 1–12.
[27] E.H. Cordes, H.G. Bull, Chem. Rev. 74 (1974) 581–604.
[28] J. Heller, S. Y Ng, B.K. Fritzinger, K.V. Roskos, Controlled
drug release from bioerodible hydrophobic ointments,
Biomaterials 11 (1990) 235–237.
[29] A. Merkli, J. Heller, C. Tabatabay, R. Gurny, Synthesis and
characterization of a new biodegradable semi-solid poly-
(ortho ester) for drug delivery systems, J. Biomatter. Sci.
Polym. Ed. 4 (1993) 505–516.
1038 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039
[30] S. Einmal, S. Capancioni, K. Schwach-Abdellaoui, M.
Moeller, F. Behar-Cohen, R. Gurny, Therapeutic applications
of viscous and injectable poly(ortho esters), Adv. Drug
Deliv. Rev. 53 (2001) 45–73.
[31] P. Wuthrich, S.Y. Ng, K.V. Roskos, J. Heller, Pulsatile and
delayed release of lysozyme from ointment-like poly(ortho
esters), J. Controlled Release 21 (1992) 191–200.
[32] S. Einmahl, F.F. Behar-Cohen, C. Tabatabay, M. Savoldelli,
F. D’Hermies, D. Chauvaud, J. Heller, R. Gurny, A viscous
bioerodible poly(ortho ester) as a new biomaterial for
intraocular applications, J. Biomed. Mater. Res. 50 (2000)
566–573.
[33] J. Heller, D.W.H. Penhale, R.F. Helwing, B.K. Fritzinger,
Release of norethindrone from poly(ortho esters), Polymer.
Eng. Sci. 21 (1981) 727–731.
[34] S.Y. Ng, T. Vandamme, M.S. Taylor, J. Heller, Synthesis and
erosion studies of self-catalyzed poly(ortho esters), Macro-
molecules 30 (1997) 770–772.
[35] S.Y. Ng, H.-R. Shen, E. Lopez, Y. Zherebin, J. Barr, E.
Schacht, J. Heller, Development of a poly(ortho ester)
prototype with a latent acid in the polymer backbone for
5-fluorouracil delivery, J. Controlled Release 65 (2000) 367–
374.
[36] M.B. Sintzel, J. Heller, S.Y. Ng, M.S. Taylor, C. Tabatabay,
R. Gurny, Synthesis and characterization of self-catalyzed
poly(ortho esters), Biomaterials 19 (1998) 791–800.
[37] K. Schwach-Abbellaoui, J. Heller, R. Gurny, Hydrolysis and
erosion studies of autocatalyzed poly(ortho esters) containing
lactoyl–lactyl acid dimers, Macromolecules 32 (1999) 301–
307.
[38] S.S. Shah, Y. Cha, C.G. Pitt, Poly(glycolic acid-co-DL-lactic
acid): diffusion or degradation controlled drug delivery?, J.
Controlled Release 18 (1992) 261–270.
[39] K. Schwach-Abbellaoui, R. Gurny, J. Heller, Determination
of pH within poly(ortho esters) by paramagnetic resonance
using pH-sensitive nitroxide radicals, study in progress
[40] J. Heller, A.-F. Rime, S.S. Rao, B.K. Fritzinger, S.Y. Ng,
Poly(ortho esters) for the pulsed and continuous delivery of
peptides and proteins, in: V.H.L. Lee, M. Hashida, Y.
Mizushima (Eds.), Trends and Future Perspectives in Pep-
tides and Protein Drug Delivery, Harwood Academic Pub-
lishers, Switzerland, 1995, pp. 39–56.
[41] J. Heller, J. Barr, S.Y. Ng, H.-R. Shen, K. Schwach-Abbella-
oui, S. Einmahl, A. Rothen-Weinhold, R. Gurny, Poly(ortho
esters) – their development and some recent applications,
Eur. J. Pharm. Biopharm. 50 (2000) 121–128.
[42] A. Rothen-Weinholod, K. Schwach-Abbellaoui, J. Barr, S.Y.
Ng, H.-R. Shen, R. Gurny, J. Heller, Release of BSA from
poly(ortho esters) extruded thin strands, J. Controlled Re-
lease 71 (2001) 31–37.
[43] J. Heller, J. Barr, S. Ng, H.-R. Shen, Injectable semi-solid
poly(ortho esters) for the controlled delivery of therapeutic
agents: Synthesis and applications, Drug Deliv. 2 (2002)
38–43.
[44] K. Schwach-Abdellaoui, J. Heller, J. Barr, R. Gurny, Control
of molecular weight for auto-catalyzed poly(ortho esters)
obtained by polycondensation reaction, Int. J. Polym. Anal.
Charact., in press.
[45] J. Heller, J. Barr, S.Y. Ng, H.-R. Shen, K. Schwach-Abdella-
oui, R. Gurny, N. Vivien-Castioni, P.J. Loup, P. Baehni, A.
Mombelli, Development and applications of injectable poly-
(ortho esters) in pain control and periodontal treatment,
Biomaterials, in press.
[46] N. Wakiyama, K. Juni, M. Nakano, Preparation and evalua-
tion in vitro of polylactic acid microspheres containing local
anesthetics, Chem. Pharm. Bull. 29 (1981) 3363–3368.
[47] N. Wakiyama, K. Juni, M. Nakano, Preparation and evalua-
tion in vitro of polylactic acid microspheres containing
bupivacaine, Chem. Pharm. Bull. 30 (1982) 3719–37276.
[48] J. Castillo, J. Curley, J. Hotz, M. Uezono, J. Tigner, M.
Chasin, R. Wilder, R. Langer, C. Berde, Glucocorticoids
prolong rat sciatic nerve blockage in vivo from bupivacaine
microspheres, Anestheology 85 (1996) 1157–1166.
[49] C. Drager, D. Benzinger, F. Gao, C.M. Berde, Prolonged
intercostal nerve blockage in sheep using controlled-release
of bupivacaine and dexamethasone from polymer micro-
spheres, Anestheology 89 (1998) 969–979.
[50] J. Heller, J. Barr, S. Ng, H.-R. Shen, R. Gurny, K. Schwach-
Abdellaoui, A. Rothen-Weinhold, M. van der Weert, De-
velopment of poly(ortho esters) and their application for
bovine serum albumin and bupivacaine delivery, J. Con-
trolled Release 78 (2002) 133–141.
[51] G.T. Tucker, L.E. Mather, Properties, absorption and dispo-
sition of local anesthetic agent, in: M.J. Cousins, P.O.
Bridenbaugh (Eds.), Neural Blockade in Clinical Anesthesia
and Management of Pain, Lippincott-Raven Publishers, New
York, 1998, pp. 55–95.
[52] A.J. van Winkelhoff, E.G. Winkel, Periodontal infections and
treatment, J. Periodontol. Implantol. Orale 15 (1996) 219–
232.
[53] K. Schwach-Abdellaoui, N. Vivien-Castioni, R. Gurny, Local
delivery of antimicrobial agents for the treatment of
periodontal disease, Eur. J. Pharm. Biopharm. 50 (2000)
83–99.
[54] K. Schwach-Abdellaoui, A. Monti, J. Barr, J. Heller, R.
Gurny, Optimization of a novel bioerodible device based on
autocatalyzed poly(ortho ester) for controlled delivery of
tetracycline to periodontal pocket, Biomaterials, in press.
[55] K. Schwach-Abbellaoui, N. Vivien-Castioni, P.J. Loup, P.
Baehni, A. Mombelli, J. Barr, J. Heller, R. Gurny, Bioerod-
ible injectable poly(ortho ester) for tetracycline delivery to
periodontal poicket: Preliminary trial in humans, http: /
www.pharmsci.org /, in press.
[56] F. Behar-Cohen, S. Einmahl, R. Gurny, unpublished work.
[57] S. Carpancioni, K. Schwach-Abbellaoui, P. Zanello, V. Kal-
satos, R. Gurny, Semi-solid autocatalyzed poly(ortho esters)
as a new veterinary drug delivery system for estrus
synchronization in ewes, Eur. J. Pharm. Biopharm, sub-
mitted.
[58] K. Kataoka, G.S. Kwon, M. Yokoyama, T. Okano, Y.
Sakurai, Block copolymer micelles as vehicles for drug
delivery, J. Controlled Release 24 (1993) 119–132.
 J. Heller et al. / Advanced Drug Delivery Reviews 54 (2002) 1015–1039 1039

[59] Y. Matsamura, H. Maeda, A new concept for macromolecu- [60] E. Schacht, V. Toncheva, J. Barr, S.Y. Ng, J. Heller, Synthesis
lar therapeutics in cancer chemotherapy: mechanism of and characterization of micelles based on ABA poly(ethylene
tumoritropic accumulation of proteins and antitumor agent glycol) and poly(ortho ester) block copolymers, work in
SMANCS, Cancer Res. 46 (1986) 6387–6392. progress.