Reconstituted botulinurn toxin type A

does not lose potency in humans if it is

refrozen or refrigerated for 2 weeks
before use
R. Richard Sloop, MD; Bradley A. Cole, MD; and Rodolfo 0. Escutin, MD

Article abstract-Botulinum toxin type A (BTX-A) (Botox, Allergan, Irvine, CA) labeling recommends its use within 4
hours of reconstitution. Since BTX-A is available only in 100-unit vials, a substantial quantity is often discarded. Using
eight volunteers, we measured the percent decline in extensor digitorum brevis (EDB) M-wave amplitude (percent
paralysis) following injection of freshly reconstituted BTX (right EDB) and compared this with the decline following
injection of BTX that was refrozen (-20" C) or refrigerated (+4"C) for 2 weeks (left EDB) after reconstitution. When
analyzed as paired data, there was essentially no difference in the muscle paralysis resulting from fresh BTX compared
with refrozen or refrigerated BTX, and no statistical difference between groups was noted. Reconstituted BTX-A that is
subsequently refrigerated or refrozen for 2 weeks does not lose potency in humans.
NEUROLOGY 1997;48:249-253

Botulinum toxin type A (BTX-A) has been effectively
used for treating a variety of neurologic conditions,
including torticollis, spasmodic dysphonia, blepharo-
spasm, strabismus, hemifacial spasm, oromandibu-
lar dystonia, focal hand dystonia, and tremor.'
BTX-A available in the United States (Botox purified
neurotoxin complex, Allergan, Irvine, CA) is supplied
as 100-mouse unit (MU) vials of freeze-dried Clos-
tridium botulinum toxin type A, with each MU corre-
sponding to the calculated median lethal intraperito-
neal dose (LD50) injected in mice.2 Reconstitution is
performed by the clinician with preservative-free
normal saline prior to intramuscular injection. Cur-
rent FDA-approved product labeling recommends ad-
ministration within 4 hours of reconstitution. Be-
cause only 100-MU vials of BTX-A are available, a
substantial quantity of this expensive toxin may be
discarded if these recommendations are strictly fol-
lowed. As a result, some clinicians have resorted to
refreezing the unused reconstituted toxin for later
use.3 However, the potency of this refrozen toxin and
its efficacy in producing muscle weakness has not
been established.
We have developed a human model for quantify-
ing the degree of muscle paralysis following botuli-
num toxin injection by injecting 13 healthy volun-
teers with seven varying doses of BTX-A in the
extensor digitorum brevis (EDB) and measuring the
fall in EDB M-wave amplitude and area4(figure 1). A
logarithmic dose response relationship was noted be-
tween increasing doses of BTX and fall in EDB
M-wave amplitude and area. Using this method, we
compared freshly reconstituted BTX with reconsti-
tuted BTX that was subsequently refrozen or refrig-
erated for 2 weeks.
Methods. Eight paid volunteers, five nonpregnant
women and three men (age range, 21 to 33 years), without
neuromuscular disease or diabetes, provided informed con-
sent, as approved by the Institutional Review Board a t
Loma Linda University. A subclinical neuropathy was ex-
cluded by obtaining a baseline sural SNAP amplitude and
distal latency, as well as a peroneal M-wave amplitude,
distal latency, and conduction velocity. The point of maxi-
mal M-wave amplitude over the surface of the muscle was
located and marked with indelible ink to assure recording
from the same site on subsequent testing. The EDB
M-wave amplitude and area were examined before and
after injection with BTX-A (Botox). Three pre-injection
nerve conduction studies were averaged to obtain a base-
line EDB M-wave amplitude and area. BTX-A was recon-
stituted using nonpreserved normal saline a t a concentra-
tion of 25 units per mL. This was divided into three
aliquots. The first aliquot was injected within 2 hours of
dilution, the second was immediately refrozen a t -20" C,
and the third was immediately refrigerated a t +4" C.
The right EDB in each of the eight subjects was injected
with 2.5 units (in 0.2 mL of normal saline) of freshly recon-
stituted BTX-A. Two weeks later the left EDB was injected
with 2.5 units of the refrozen preparation in four subjects,
while the other four received 2.5 units of the refrigerated
preparation. In all subjects the BTX was injected intra-

From the Department of Neurology, Lorna Linda University School of Medicine, Loma Linda, CA.
Presented in part a t the 47th annual meeting of the American Academy of Neurology, Seattle, WA, May 1995.
Received December 6, 1995. Accepted in final form May 14, 1996.
Address correspondence and reprint requests to Dr. R. Richard Sloop, Department of Neurology, Lorna Linda University Medical Center, 11234 Anderson,
Rm. 1580, Loma Linda, CA 92354.
Copyright 0 1997 by the American Academy of Neurology 249
 .I- I T

K,

50 . I ___ ~-
/
)/
~

+Average M-Wave Amplitude -~

- Y - Average M - Wave AN^ _ _ _ ~ ~

0 1.25 2.5 5 7.5 10 15 20

Botulinum Toxin Dose (Mouse Units)

Figure 1. EDB M-wave amplitude and area decrement (muscle "paralysis'? with increasing doses of intramuscularly in -
jected botulinum A toxin. Each dose was injected in two subjects indicated by the bars above and below the average value
of the two responses. Although each amplitude and area data set is for the same dose, the data points were staggered
slightly to allow better visualization of the data,

muscularly directly beneath the site of maximal M-wave
amplitude using a 27-gauge needle.
Following the BTX injection, the EDB M-wave ampli-
tude and area were recorded on six occasions (days 2, 4,6,
9, 11,and 13).The decline in EDB M-wave amplitude and
area were plotted for each subject and the effect of BTX
expressed as percent decline in the M-wave amplitude and
area. The three pre-injection values were averaged to cal-
culate the baseline; the post-injection values from days 9,
11, and 13 were averaged to calculate the decline.
A Nicholet Viking I1 electromyograph was used to gen-
erate the stimuli and to record all responses (filter settings
2 Hz to 10 kHz). Eleven-mm-diameter disposable surface
electrodes (TECA Corporation) were used for all record-
ings. M-wave amplitude and area were obtained using su-
pramaximal stimulation of the peroneal nerve at the ankle
with a stimulation distance to the EDB of 8 cm. The refer-
ence electrode was placed on the lateral aspect of the foot
approximately 6 cm distal to the recording electrode. All
subjects were warmed in a water bath prior to the record-
ing, and surface skin temperatures were above 32" C post-
warming. The same examiner performed all studies.
The four subjects who received fresh toxin into the right
EDB followed by refrozen toxin into the left EDB formed
group 1. The other four subjects who received fresh toxin
followed by refrigerated toxin formed group 2. The data for
each group were analyzed separately as paired data.
260 NEUROLOGY 48 January 1997
Results. The results of each group are shown in figure 2.
There was no significant difference between fresh and re-
frozen toxin or between fresh and refrigerated toxin in
these paired samples (table). The average E:DB muscle
paralysis in all eight subjects was 58.5 with fresh toxin
and 58.0 with non-fresh toxin (frozen and refrigerated
combined). The decline in M-wave area post-injection was
virtually identical to the decline in amplitude.
One subject experienced some weakness of toe exten-
sion and an "achy" feeling in her right foot while doing
aerobics the day after injection with the fresh BTX. This
resolved within 2 days. The other subjects did not experi-
ence any side effects and could not identify weakness in
any foot muscle.
Discussion. Although some clinicians have at-
tempted to preserve the BTX that remains following
a treatment session by refreezing the reconstituted
toxin and using it at a later date,:j it has been diffi-
cult to determine whether there has been a loss of
potency as a result of this refreezing. The methods
for measuring the response t o BTX are inexact and
rely to a large degree on subjective patient observa-
tion. Gartlan and Hoffman6 attempted to measure
the stability of reconstituted BTX-A (Botox) by in-
jecting mice and determining the LU50 using BTX
that had been reconstituted and preserved for vary-
 90

80

.-p
u)

- 70 (5)
E
h Figure 2. EDB M-wave amplitude
2
0
60 decrement resulting from a fixed 2.5-
0
unit intramuscular injection of botu-
z 50
h u m toxin type A (Botox) (the bar
w
reflects the mean of each group of
+. responses, and each subject is num-
f! 40
bered). Note the essentially identical
Q)
n effect of refrozen and refrigerated
toxin compared with reconstituted
30 and immediately injected (fresh)
toxin. (The difference in mean mus-
cle paralysis between group 1 and
20
group 2 is probably explained by
Fresh Toxin Refrozen Toxin Fresh Toxin Refrigerated Toxin
variation in individual sensitivity to
Group 1 Group 2 B TX.)

ing lengths of time, either by refrigeration or refreez-
ing. They found a 70% loss in potency when the BTX
was reconstituted, immediately refrozen? and then
injected 2 weeks later and a 44% drop in potency of
BTX after 12 hours of refrigerator storage. However?
the mouse LD50 may not be a good predictor of the
human response.
Our technique relies on the response of human
EDB muscle to intramuscular botulinum toxin injec-
, ~ demonstrated a log-
tion. In our previous r e p ~ r twe
arithmic dose response relationship between decre-
ment in M-wave amplitude and area and increasing
dose of BTX-A.4 For the present study we chose to
inject 2.5 units of BTX as this point lies on the
steeper slope of the curve and allows a greater sensi-
tivity in detecting differences in the potency of the
different toxin preparations. (The mean 2.5-unit in-
jection produced a larger M-wave decrement [59%,
see figure 21 than was found in the original dose
response curve [48%, see figure 11. This may be the
result of a different potency of the two toxin prepara-
tions. The studies were done at different times and a
different preparation of toxin was used in each study.)
As each subject received fresh toxin into one EDB
and preserved toxin into the other, the data could be
analyzed as paired data. This lends to the strength
of the study substantially as there is individual vari-
ation in sensitivity to botulinum toxin (see figure 2),
but each individual could serve as his or her own
control. Thus, the difference between the mean mus-
cle paralysis from fresh toxin in group 1 compared
with fresh toxin in group 2 can be explained by the
slight preponderance of more sensitive individuals in
group 1.
We chose the one-tailed probability to compare
fresh and preserved toxin in each paired group as
this would be the most sensitive method to detect a
difference between the two groups. A two-tailed
probability could have been used but would be less
sensitive and the toxin would not likely gain potency

Table Statistical comparisons of fresh and non-fresh toxin in group I and in group 2 and between groups 1 and 2. (Group 1 is fresh
toxin versus refrozen toxin. Group 2 is fresh toxin versus refrigerated toxin.)
~ _ _ _ _ _ _ ~ _ _ _ _ _ ~ .
- ~

Group 1 Fresh toxin Refrozen toxin One-tailed p value for difference: 0.40"
Mean paralysis: 62.0 Mean paralysis: 60.5
Lower 95% confidence limit: 53.0
~ _ _ _ ~~ ~ __
_ _ _____ ~~
-
Group 2 Fresh toxin Refrigerated toxin One-tailed p value for difference: 0.46'''
Mean paralysis: 55.0 Mean paralysis: 55.5
Lower 95%Jconfidence limit: 43.5
~

Two-tailed p value for difference Two-tailed p value for difference Two-tailed p value for difference between group
between group 1 and group 2 between group 1 and group 2 1 change in paralysis and group 2 change in
fresh toxin: 0.411 non-fresh toxin: 0.53t paralysis: 0.79t
_ _ _ _ _ . _ ~ _ _ ~ ~~ _ _

* Paired
t test.
Independent t test.
January 1997 NEUROLOGY 48 251
from the 2 weeks of preservation, and thus the one-
tailed p value is more logical as well.
The statistical comparison of the difference be-
tween potency change in group 1 and group 2 pro-
vides the best comparison of the two methods of
preservation (refrigeration and refreezing). This two-
tailed p value of 0.79 convincingly shows that there
is not any significant difference between these two
methods of toxin preservation.
The lower 95% confidence limit for both refrozen
and refrigerated toxin gives the best indication of
how robust our findings are. The slope of the first
portion of the dose response curve can be used to
estimate the sensitivity of this technique for changes
in toxin potency. If the portion of the curve between
0 and 2.5 units is treated as linear, an approximate
dose of toxin can be calculated that would produce a
given M-wave decrement. Using the lower 95% confi-
dence limit from the table, a 53%M-wave decrement
(refrozen toxin) would be produced by approximately
2.1 units of toxin, a 15%loss in potency. Similarly, at
the lower 95% confidence limit for the refrigerated
toxin, a 43.5% M-wave decrement would be the
equivalent of an approximately 20% loss of potency.
Thus, this study should be able to detect a loss of
potency of 15 to 20% with 95% certainty. Testing of a
single preparation of BTX-A by different investiga-
tors using the mouse bioassay results in a variability
of reported potency of -t30%,zand the 95% confi-
dence interval of Gartlan and Hoffman5 varies from
525 to 30% for fresh toxin to -43%/+ 140% for re-
frigerated and then refrozen toxin. Thus, the vari-
ability from the human model compares favorably
with that of the mouse bioassay. And although the
total number of muscles injected in our study was
small, the use of paired data yields a fairly tight 95%
confidence interval, lending to the certainty of our
results.
Animal responses to botulinum toxin may not be
accurate models of the human response for several
reasons. The species' differences in the sensitivity to
botulinum toxin is a known phenomenon. For exam-
ple, when injected subcutaneously in rats, BTX-F
produces much less muscle weakness than an equal
LD50 dose of BTX-A. In contrast, when injected in
humans in equal LD50 doses, these two toxins pro-
duce similar degrees of clinical That is,
although the mouse LD50 accurately predicts the
human response to BTX-F relative to BTX-A, it does
not accurately predict muscle weakness produced in
the rat. Although clinical benefit may not correlate
strictly with muscle weakness, the two are closely
related. For example, in the rat using BTX-F, a four-
fold increase in dose was necessary to obtain a simi-
lar degree of weakness as that obtained with the
BTX-A,Hand this four-fold change in dose does pro-
duce a substantially different clinical response in pa-
tients with blepharospasm.9*'0Even though it is pos-
sible that dissimilar degrees of muscle weakness
could produce a similar clinical response, one would
expect similar degrees of muscle weakness to pro-
252 NEUROLOGY 48 January 1997
duce a similar clinical response. Thus, the muscle
weakness measured in this study should be more
reliable than a less objective "clinical response".
Furthermore, although the mouse LD50 may ac-
curately predict the human response to BTX-F rela-
tive to BTX-A, our work has shown that the mouse
LD50 of BTX-B has little predictive value for the
dose necessary to produce human muscle paralysis, I
and the onset and duration of muscle paralysis re-
sulting from intramuscular injection of botulinum
toxin types A, B, and F is much different in the rat
than it is in h u m a n ~ . ~ . ~AlsoJ l using the rat model,
Shaari and Saunders14 found a 25-fold increase in
dose of type A toxin necessary to double paralysis,'"
whereas in our human model, a doubling of paralysis
occurred with a four-fold increase in dose."
In addition, Kauffman et al.R note that the rat
subcutaneous LD50 dose of BTX-F is 140 X that of
BTX-A, the intruperitoneal LD50 dose is 50-fold dif-
ferent for these same two toxins, and only a four-fold
difference in dose produces similar degrees of initial
muscle weakness after subcutaneous Thus,
not only is the species difference important, but also
the route of administration of each toxin type within
each species and the variable measured (local muscle
weakness versus animal death). Lastly, the mouse
LD50 model does not accurately predict the clinical
response to two different preparations of the same
botulinum toxin subtype (BTX-A, Botox & Dys-
port).15J6Pearce et al.I7J8demonstrated a 2.7-fold dif-
ference in the relative potency of the two prepara-
tions of BTX-A (measuring regional muscle paralysis
in the mouse) which could not be detected by the
LD50 assay. These studies clearly establish that rat
and mouse models are not optimal ones to predict
the human response to botulinum toxin. Thus, it
should not be surprising that small changes in the
toxin induced by freezing or refrigerator storage
might affect its potency in one animal species by one
route of administration, or for one measured end
point, but that this effect would not be generalizable
t o other species or circumstances of injection.
The manufacturer of Botox has expressed concern
that the storage of reconstituted botulinum toxin
(with preservative free saline) could lead to contami-
nation and infection from injection (K.R. Aoki, Aller-
gan Pharmaceuticals, personal communication). We
have never encountered an infection fram stored
toxin, and the experience of others has been simi-
lar. 19~20
In summary, this study provides the first quanti-
fiable evidence that BTX-A does not lose significant
potency in human intramuscular injection when it is
reconstituted and then refrozen or refrigerated for 2
weeks before use. These findings coincide with clini-
cal ~ b s e r v a t i o n , ~
and
~ ~ we
~ " believe they reflect rea-
sonable certainty.
Acknowledgment
We would like to thank Grenith Zirnmerman, PhD, for her assis-
tance with statistical analysis.
References
1. Jankovic J, Brin M. Therapeutic uses of botulinum toxin.
N Engl J Med 1991;324:1186-1194.
2. Schantz EJ, Kautter DA. Standardized assay for Clostridium
botulinum toxins. J Assoc Off Anal Chem 1978;61:96-99.
3. Green P. Clinical trials for cervical dystonia 2. In: NIH con-
sensus development conference on clinical use of botulinum
toxin program. Bethesda, MD: NIH, November 199053-56.
4. Sloop RR, Escutin RO, Matus JA, Cole BA, Peterson GW.
Dose response curve of human extensor digitorum brevis mus-
cle function to intramuscularly injected botulinum A toxin.
Neurology 1996;46:1382-1386.
5. Gartlan MG, Hoffman HT. Crystalline preparation of botuli-
num toxin type A (BTX): degradation in potency with storage.
Otolaryngol Head Neck Surg 1993;108:135-140.
6. Greene PE, Fahn S. Use of botulinum toxin type F injections
to treat torticollis in patients with immunity to botulinum
toxin type A. Mov Disord 1993;8:479-483.
7. Mezaki T, Kaji R, Kohara N, et al. Comparison of therapeutic
efficacies of type A and F botulinum toxins for blepharospasm:
a double-blind, controlled study. Neurology 1995;45:506-508.
8. Kauffman JA, Way J F , Siege1 LS, Sellin LC. Comparison of
the action of types A and F botulinum toxin at the rat neuro-
muscular junction. Toxicol Appl Pharmacol 1985;79:211-217.
9. Frueh BR, Felt DP, Wojno TH, Musch DC. Treatment of
blepharospasm with botulinum toxin: a preliminary report.
Arch Ophthalmol 1984;102:1464-1468.
10. Borodic GE, Cozzolino D. Blepharospasm and its treatment,
with emphasis on the use of botulinum toxin. Plast Reconstr
Surg 1989;83:546-554.
11. Sloop RR, Escutin RO, Cole BA. Human muscle paralysis
from botulinum toxin injection: type A vs. type B [Abstract].
Neurology 1996;46(suppl 2):259.
12. Sellin LC, Thesleff S, Dasgupta BR. Different effect of types A
and B botulinum toxin on transmitter release a t the r a t neu-
romuscular junction. Acta Physiol Scand 1983;119:127-133.
13. Cull-Candy SG, Lundh H, Thesleff S. Effects of botulinum
toxin on neuromuscular transmission in the rat. J Physiol
1976;260:177-203.
14. Shaari CM, Saunders I. Quantifying how location and dose of
botulinum toxin injections affect muscle paralysis. Muscle
Nerve 1993;16:964-969.
15. Clarke CE. Therapeutic potential of botulinum toxin in neuro-
logical disorders. Q J Med 1992;299:197-205.
16. Erin MF, Blitzer A. Botulinum toxin: dangerous terminology
errors. J R SOC Med 1993;86:493-494.
17. Pearce LB, Borodic GE, First ER, MacCallum RD. Measure-
ment of botulinum toxin activity: evaluation of the lethality
assay. Toxicol Appl Pharmacol 1994;128:69-77.
18. Pearce LB, Borodic GE, Johnson EA, First ER, MacCallum
RD. The median paralysis unit: a more pharmacologically rel-
evant unit of biological activity for botulinum toxin. Toxicon
1995;33:217-227.
19. Garcia A, Fulton J E . Cosmetic denervation of the muscles of
facial expression with botulinum toxin: a dose-response study.
Dermatol Surg 1996;22:39-43.
20. Greene P. Potency of frozedthawed botulinum toxin type A in
the treatment of torsion dystonia. Otolaryngol Head Neck
Surg 1993;109:968-969.

Autosornal dominant hyaline body

myopathy presenting as scapuloperoneal
syndrome:
Clinical features and muscle pathology
S. Masuzugawa, MD; S. Kuzuhara, MD; Y. Narita, MD; Y. Naito, MD; A. Taniguchi, MD; and T. Ibi, MD

Article a b s t r a c t H y a l i n e bodies are rare subsarcolemmal aggregates in type 1 fibers of the skeletal muscle, stain pale
pink with hematoxylin-eosin and pale green with the modified Gomori trichrome, and lack reactivity for glycogen and
oxidative enzymes. We report clinical findings of autosomal-dominant hyaline body myopathy in seven members in four
generations and muscle biopsy findings in two of them. Slowly progressive muscle weakness and atrophy developed with
scapuloperoneal distribution; age at onset was from the first to the fifth decade. Muscle biopsy showed subsarcolemmal
hyaline bodies in approximately 20% of type 1 fibers. Hyaline bodies showed myofibrillar ATPase activity after acid
pre-incubation. Immunohistochemically, they stained intensely with myosin heavy chain (slow), but not with myosin
heavy chain (fast). Ultrastructurally, they consisted of granules sometimes in linear array, filaments, and amorphous
materials. These findings suggest that hyaline bodies may be products of degeneration of myosin heavy chain (slow).
NEUROLOGY 1997;48:253-257

Cancilla et al.' described a unique neuromuscular homogeneously hyalinized. These bodies retained re-
disorder characterized pathologically by discrete sub- activity for myofibrillar ATPase and lacked reactiv-
sarcolemmal bodies in type 1 fibers that appeared ity for glycogen and oxidative enzymes. Ultrastruc-

From the Department of Neurology (Drs. Masuzugawa, Kuzuhara, Narita, Naito, and Taniguchi), Mie University School of Medicine, Tsu City, Japan, and
the IVth Department of Internal Medicine (Dr. Ibi), Aichi Medical University, Japan.
Received February 29, 1996. Accepted in final form May 14, 1996.
Address correspondence and reprint requests to Dr. Masuzugawa, Department of Neurology, Mie University School of Medicine, 2-174 Edobashi, Tsu City,
Mie 514, Japan.
Copyright 0 1997 by the American Academy of Neurology 253
Reconstituted Botulinum Toxin Type A Does Not Lose Potency in Humans If It Is
Refrozen or Refrigerated for 2 Weeks Before Use
R. Richard Sloop, Bradley A. Cole and Rodolfo O. Escutin
Neurology 1997;48;249-253
DOI 10.1212/WNL.48.1.249

This information is current as of January 1, 1997

Updated Information & including high resolution figures, can be found at:
Services http://www.neurology.org/content/48/1/249.full.html

References This article cites 18 articles, 1 of which you can access for free at:
http://www.neurology.org/content/48/1/249.full.html##ref-list-1
Citations This article has been cited by 2 HighWire-hosted articles:
http://www.neurology.org/content/48/1/249.full.html##otherarticle
s
Permissions & Licensing Information about reproducing this article in parts (figures,tables)
or in its entirety can be found online at:
http://www.neurology.org/misc/about.xhtml#permissions
Reprints Information about ordering reprints can be found online:
http://www.neurology.org/misc/addir.xhtml#reprintsus

Neurology ® is the official journal of the American Academy of Neurology. Published continuously
since 1951, it is now a weekly with 48 issues per year. Copyright Copyright 1997 by Advanstar
Communications Inc.. All rights reserved. Print ISSN: 0028-3878. Online ISSN: 1526-632X.