Emerging Treatments and Technologies

O R I G I N A L A R T I C L E

Efficacy of Injected Liquid Silicone in the

Diabetic Foot to Reduce Risk Factors for
Ulceration
A randomized double-blind placebo-controlled trial
CARINE H.M. VAN SCHIE, MSC TRUDY WIGNALL, FRCR tar tissue thickness (4). Both plantar pres-
ALEXANDRA WHALLEY, BSC SALLY HOLLIS, MSC sures and tissue thickness can be objec-
LORETTA VILEIKYTE, MD ANDREW J.M. BOULTON, FRCP tively quantitatively assessed, whereas
calluses, which are more difficult to quanti-
tatively assess, have also been reported to be
predictive of foot ulceration (5). Calluses act
as foreign bodies, and their removal leads to
OBJECTIVE — To investigate the effectiveness of injecting liquid silicone in the diabetic foot reduced plantar pressure in most cases (6).
to reduce risk factors for ulceration in a randomized double-blind placebo-controlled trial. Furthermore, neuropathic ulcers are com-
monly found beneath plantar calluses.
RESEARCH DESIGN AND METHODS — A total of 28 diabetic neuropathic patients
without peripheral vascular disease were randomized to active treatment with 6 injections of Traditionally, calluses are removed
0.2 ml liquid silicone in the plantar surface of the foot or to treatment with an equal volume when they are excessively formed under the
of saline (placebo). No significant differences were evident regarding age or neuropathy status diabetic foot; however, only 2 preliminary
between the 2 groups. All injections were under the metatarsal heads at sites of calluses or high studies have addressed how callus buildup
pressures. Barefoot plantar pressures (pedobarography) and plantar tissue thickness under the can be minimized (7,8). Preventive care to
metatarsal heads (Planscan ultrasound device) were measured at baseline and at 3, 6, and 12 reduce the incidence of foot ulceration also
months after the first injection. Injection sites were photographed at all stages, and callus for- includes the provision of pressure-reducing
mation was scored as a change from baseline. Throughout the study, patients were treated by insoles and therapeutic footwear (9).
the same podiatrist for all podiatry treatment. Although effective, insoles must be replaced
after they lose their cushioning properties,
RESULTS — Patients who received silicone treatment had significantly increased plantar tis-
sue thickness at injection sites compared with the placebo group (1.8 vs. 0.1 mm) (P  0.0001) and padded inserts may move in the shoe.
and correspondingly significantly decreased plantar pressures (232 vs. 25 kPa) (P  0.05) Furthermore, neither device provides any
at 3 months, with similar results at 6 and 12 months. A trend was noted toward a reduction of protection for noncompliant patients.
callus formation in the silicone-treated group compared with no change in the placebo group. Previous reports have suggested the
therapeutic use of liquid silicone injections
CONCLUSIONS — The results confirm the efficacy of plantar silicone injections in reduc- in the foot to replace fat padding at callus
ing recognized risk factors associated with diabetic foot ulceration. sites, corns, and localized painful areas
(10,11). To date, more than 20,000 injec-
Diabetes Care 23:634–638, 2000 tions have been given, and this has provided
anecdotal evidence of reduced callus forma-
tion, reduced diabetic foot ulcer recurrence,
iabetic foot disease remains one of the function and structure (prominent meta- and relief of localized pressure-related foot

D most serious and severe long-term

complications of diabetes and is the
most common cause of hospitalization for
diabetic patients in Western countries. Foot
tarsal heads) and dryness of the skin, which
can lead to excessive callus formation (1,2).
An important predictive risk factor for the
development of foot ulceration is high plan-
pain in nondiabetic patients (10,11). Only
minimal side effects have been reported.
Experience with this technique to date
has not only been anecdotal but also has
ulceration occurs as a consequence of the tar foot pressure (3), which usually occurs been restricted to one center. Therefore, the
interaction of several contributory factors: at sites with bony prominences and has present study was designed to investigate for
diabetic neuropathy causes changes in foot been strongly associated with reduced plan- the first time the efficacy of injecting liquid
silicone in reducing risk factors for diabetic
foot ulceration in a randomized placebo-
From the Department of Medicine (C.H.M.v.S., A.W., L.V., A.J.M.B.), Manchester Royal Infirmary, Manchester;
the Radiology Department (T.W.) and the Diabetes Foot Clinic (C.H.M.v.S.), Disablement Services Center, With-
controlled trial. We hypothesized that injec-
ington Hospital, Manchester; and the Medical Statistics Unit (S.H.), Lancaster University, Lancaster, U.K. tion with silicone would lead to increased
Address correspondence and reprint requests to Carine H.M. van Schie, MSc, Diabetes Foot Clinic, Dis- localized plantar thickness, reduced plantar
ablement Services Centre, Withington Hospital, Cavendish Road, Manchester M20 8LB, U.K. E-mail: pressure, and reduced callus formation.
carine@footclinic.demon.co.uk.
Received for publication 26 August 1999 and accepted in revised form 12 January 2000.
Abbreviations: ABPI, ankle-brachial pressure index; CVA, cerebrovascular accident; NDS, neuropathy dis-
ability score; VPT, vibration perception threshold. RESEARCH DESIGN AND
A table elsewhere in this issue shows conventional and Système International (SI) units and conversion METHODS — The study was approved
factors for many substances. by the local ethics committee. All patients

634 DIABETES CARE, VOLUME 23, NUMBER 5, MAY 2000

 van Schie and Associates

received complete information about the sured for each foot. The peak pressure of all were carried out without knowledge of the
study before giving their consent. injection sites was used for analysis. To results from previous visits.
ensure standardization, pressure was mea-
Patients sured during the same step after initiation Method of injection
A total of 28 diabetic patients attending the of gait for all subjects for all measurements. Patients were randomized according to a
Manchester Diabetic Foot Clinic (Manches- The plantar tissue thickness was mea- random number sequence just before the
ter, U.K.) were enrolled in the study. Inclu- sured at each injected site by using the first injection. All investigators and patients
sion criteria were established neuropathy Planscan, which is an ultrasound device remained blinded to the treatment regimen
defined as a vibration perception threshold (Department of Medical Physics and Clin- throughout the study, with the exception of
(VPT) of 25 V (12) or a neuropathy dis- ical Engineering, Royal Hallamshire Hospi- the podiatrist administering the injections
ability score (NDS) of 6 (13) and the pres- tal) (15). The Planscan is a scanning who did not participate in any of the assess-
ence of a callus under at least 1 metatarsal platform that holds a high-resolution ultra- ments or analyses. A total of 6 injections
head. Patients with peripheral vascular dis- sound probe. The ATL Ultramark 9 ultra- were given per site at 2 weekly intervals.
ease (i.e., the absence of more than 1 foot sound scanner with a 5-MHz linear array Between 1 and 5 sites were selected for injec-
pulse in both feet or an ankle-brachial pres- transducer (Advanced Technology Labora- tion, depending on the number of callus
sure index [ABPI] of 0.9) and with an tories, Bothwell, WA) was used for the sites. Thus, each patient received between 6
active or previous ulcer during the past assessment. The measurement was per- and 30 injections. The volume of silicone or
6 months were not selected for the study. formed while the subjects stood barefoot saline per injection was 0.2 ml for each site;
Patients were randomized to silicone on the platform with equal weight bearing the total maximum volume injected was 1.2
treatment (n = 14) or to equal amounts of between the 2 feet (4). The foot was placed ml/injection site. An area with previous
placebo treatment (n = 14) according to a directly above the ultrasound probe, which ulceration was only chosen if it had been
random number sequence. Areas of injec- was attached to the frame under the top healed for a minimum of 6 months. Before
tion were chosen under metatarsal head surface of the platform. Scanning the length injection, all areas to be injected were
sites with calluses or widened skin striae. of the foot allows more accurate and repro- debrided of calluses and were cleaned. A
ducible results of the minimum depth of an skin refrigerant (fluro-ethyl) was sprayed
Study design irregular-shaped bone prominence below over the site of injection for 2–3 s before
At baseline, all patients underwent a neu- the sole of the foot (15). A minimum of 3 injection. Any discomfort was noted, and, if
ropathic assessment, including the modi- measurements were taken of the metatarsal necessary, the needle was withdrawn, and
fied NDS (13) and VPT measurement on depth at each injected site, and these were local anesthesia was administered (mepiva-
the hallux, by using a Neurothesiometer averaged per site and were used for analy- caine 3%). A needle guide was attached to
(Horwell, Nottingham, U.K.) (12), and vas- sis. A combined intra- and interobserver the syringe to assist in a more precise injec-
cular status was determined by checking coefficient of variation of 10% has been tion. The liquid silicone or saline was
foot pulses and by assessing the ABPI. reported for this measurement (4). implanted subcutaneously in equal amounts
Clinical photographs were taken of The clinical photographs of the beneath and within 1–2 mm of the central
the areas to be injected, and plantar pres- injected sites taken at all visits were point of the callus. After injection, the site
sure and tissue thickness were assessed at assessed as a trend of change from baseline. was covered with a sterile bandage, and the
the chosen injection sites after calluses A special scoring system was designed for patient was advised to keep the injection site
were debrided. this assessment because no other estab- dry for 24 h and to check the site for any
After all assessments, the first injection lished method exists. A negative score indi- signs of inflammation or infection. Patients
was given, and 5 subsequent injections cates a worsening of callus formation, a were allowed to resume regular activities
were given at 2 weekly intervals. Follow-up score of 0 indicates no change from base- immediately after injection.
visits were scheduled at 3, 6, and 12 line, and a positive score indicates a reduc-
months after baseline, during which all tion of callus formation. A maximum Statistical analysis
measurements of efficacy were repeated. negative score of 3 was given for an ulcer The change from baseline was used for data
Throughout the study, the patients were or hemorrhage, and a maximum positive analysis of the plantar tissue thickness and
treated by the same study podiatrist, and all score of 3 was given for the complete dis- pressures to eliminate the effect of natural
study patients continued to receive the appearance of a callus. One score was given differences in baseline plantar thickness and
same treatment that is offered to all patients for each subject that represented the pressures at different sites. Data are medians
at high risk for foot ulceration, including change in callus buildup over all injection (interquartile ranges). The changes in peak
receiving specialty footwear and regular sites per subject during the 1-year follow- plantar pressure and tissue thickness per
podiatry treatment as needed. up period. Thus, a composite score was injected site were averaged over the total
given for photographs taken at 3, 6, and 12 number of injected sites per patient because
Outcome measures months of follow-up when compared with treatment was randomized per patient and
Dynamic plantar pressures were measured the baseline photographs. Two indepen- not per injection site. The average change
during barefoot walking by using an opti- dent observers blinded to patient groups per patient was then used for further analy-
cal pedobarograph (Department of Medical scored all photographs, and an average of sis by using the Mann-Whitney U test for
Physics and Clinical Engineering, Royal the 2 observers was used for analysis. differences between the 2 treatment groups
Hallamshire Hospital, Sheffield, U.K.), All outcome measurements were per- at each follow-up visit. An intention-to-treat
which is a pressure platform built into a formed completely blinded to the treatment analysis was carried out by using a conser-
5-m walkway (14). Five steps were mea- regimen, and all follow-up measurements vative carry forward analysis in which miss-

DIABETES CARE, VOLUME 23, NUMBER 5, MAY 2000 635

Efficacy of injected liquid silicone

Table 1—Demographic and neurological characteristics of patients by treatment group

Diabetes Type of History of

n Sex (F/M) Age (years) duration (years) diabetes (1/2) ulceration NDS VPT (V) ABPI
Silicone 14 5/9 58.1 ± 12.3 10.5 (9.3–17.8) 5/9 8 8.0 (7.3–9.5) 29.5 (25.3–41.5) 1.23 (1.0–1.28)
Placebo 14 3/11 55.0 ± 7.8 15.0 (7.3–22.0) 6/8 7 8.0 (8.0–10) 28.0 (25.0–34.8) 1.18 (1.11–1.38)
Data are n, means ± SD, and medians (interquartile ranges). No significant differences were observed between the 2 groups in any variable.

ing data in both the silicone-treated and increased from baseline with 1.8 mm Intention-to-treat analysis
placebo groups were considered to be no (1.0–2.5), 2.0 mm (0.9–2.6), and 1.3 mm The results of the intention-to-treat analy-
different from baseline. (0.9–1.6) at 3, 6, and 12 months, respec- sis showed a significant increase in tissue
The scores of the clinical photographs tively, in the silicone-treated group compared thickness (1.8 mm [1.0–2.5], 1.8 mm
were averaged for the 2 observers and were with no change in plantar tissue thickness [0.8–2.5], and 1.2 mm [0.9–1.6] at 3, 6,
subsequently analyzed by using the Mann- (0.08 mm [–0.3 to 0.5], 0.2 mm [–0.2 to and 12 months, respectively) in the sili-
Whitney U test. SPSS software (Chicago) 0.6], and 0.25 mm [0.0 to 0.6], respectively) cone-treated group compared with the
was used for the statistical analysis. in the placebo group (P  0.005) (Fig. 1). placebo treatment group (0.08 mm [–0.3
A significant decrease in peak plantar to 0.5], 0.10 mm [–0.2 to 0.5], and 0.09
RESULTS — No significant differences pressures from baseline was measured in mm [0.0 to 0.4], respectively) (P  0.008).
were evident in baseline characteristics the silicone-treated group at 3, 6, and 12 The reduction in peak plantar pres-
between the 2 groups (Table 1). Not all fol- months (232 kPa [–372 to 84], sure, by using the intention-to-treat analy-
low-up visits were completed by all 182 kPa [–227 to 13], and 216 kPa sis, was significantly greater in the
patients because of foot ulcers (n = 2), [–300 to 30], respectively), whereas no silicone-treated group at 12 months (170
development of malignancy (n = 1), cere- change or a slight increase was evident kPa [–296 to 7]) compared with the
brovascular accident (CVA) (n = 1), repeat- from baseline in the placebo group (25 placebo group (0.0 [0.0 to 183] kPa) (P 
edly missed appointments (n = 1), technical kPa [–146 to 97], 58 kPa [–67 to 99], 0.02) and almost reached significance at 3
problems with equipment (n = 3), and and 145 kPa [–136 to 220], respectively) and 6 months (silicone vs. placebo 225
inability to analyze data (n = 1). Tissue (P  0.05) (Fig. 2). kPa [–370 to 21] vs. 25 kPa [–146 to
thickness data are missing for 2 silicone A significant correlation existed 97] [P = 0.06] and 172 kPa [–220 to 10]
and 5 placebo follow-up visits, and plantar between the percentage change in peak vs. 3.7 kPa [–35 to 89] [P = 0.09] at 3 and
pressure data are missing for 3 silicone and plantar pressure and plantar tissue thick- 6 months, respectively).
8 placebo follow-up visits (from a total of ness after injection with silicone. The cor-
84 follow-up visits). relation coefficients were 0.37, 0.34, Callus
A total of 62 sites were chosen for and 0.39 at 3, 6, and 12 months of fol- A nonsignificant trend was noted toward a
injection, of which 34 sites were injected low-up, respectively (P  0.05). reduction in callus formation in the sili-
with silicone, and 28 were injected with
saline (treatment was randomized per
patient). Between 1 and 5 sites (median 2)
per patient (depending on the number of
callus sites) were selected for injection.
Treatment was ceased at 3 sites (after 1 or
2 injections, all in the silicone-treated
group) because of the development of an
ulcer on the same foot as 2 of the sites and
1 site that looked fragile and seemed likely
to ulcerate according to the medical staff.
A nonsignificant difference was evident
in mean ± SD baseline thickness in the sil-
icone-treated group (0.66 ± 0.21 cm) com-
pared with the placebo group (0.85 ± 0.26
cm) (P = 0.06). The difference in thickness
was caused by the different number of first
and fifth metatarsal heads chosen as injec-
tion sites. The ratio of first/fifth metatarsal
head injection sites was 7/12 for the sili-
cone-treated group and 11/7 for the
placebo group. Figure 1—Changes in plantar tissue thickness (in millimeters) from baseline. The bars represent the
The median (interquartile range) plan- median values. Silicone treatment vs. placebo, P  0.0005 at 3 and 6 months and P  0.005 at 12
tar tissue thickness had substantially months of follow-up.

636 DIABETES CARE, VOLUME 23, NUMBER 5, MAY 2000

 van Schie and Associates

ulceration is not known. However, a

median pressure reduction of 30%, as
reported in this study, is similar to the
results of insole material and shoe design
studies (16–19). Thus, the injected silicone
provides an amount of cushioning similar
to the more conventional methods of pres-
sure relief; however, silicone has the advan-
tage of not being removable, thus offering
cushioning continuously.
The highly subjective nature of the
assessment of change in callus buildup
(scoring of clinical photographs) and con-
sequently high variability of this method
combined with the fact that calluses were
scored per patient (and not per site) may
have contributed to the nonsignificant dif-
ference in callus buildup between the
2 treatment groups.
Figure 2—Changes in peak plantar pressure (in kilopascals) from baseline. The bars represent the In conclusion, the results of this study
median values. Silicone treatment vs. placebo, P  0.05 at 3 and 12 months and P = 0.11 at 6 months have shown for the first time that the injec-
of follow-up. tion of liquid silicone in the diabetic foot can
substantially increase plantar tissue thick-
ness and reduce plantar peak pressures, thus
cone-treated group compared with the thermore, note that the only other center adding a new treatment method to diabetic
placebo group. The median score for change that used this technique injected individu- foot care. Whether this method can reduce
in callus buildup from baseline was 0.5 (0.0 ally tailored volumes; however, in the pres- foot ulceration remains to be answered in
to 1.0) in the silicone-treated group com- ent study, we used a standardized volume. larger trials. No side effects were reported in
pared with 0 (1.25 to 0.75) in the placebo No significant side effects were this study. Thus, a reduction of risk factors
group (P = 0.3). reported in this study, and only minimal strongly associated with foot ulceration has
side effects have been reported by the only been observed, and we believe that injection
Adverse events other center that has experience with inject- with liquid silicone may represent a poten-
A total of 7 patients developed foot ulcers ing silicone in the foot (10,11). Asympto- tial preventative intervention to reduce the
during the course of the study. A total of 3 matic fluid migration has been reported as incidence of neuropathic foot ulcers in high-
placebo and 3 silicone-treated patients devel- the main adverse response (10,11); how- risk diabetic patients. Further large-scale
oped ulcers at noninjected sites (toes, ever, this has only been observed in early studies are now indicated to confirm this
between digits, toe nails, heels, and Achilles cases when larger volumes were injected per observation.
tendons), whereas 1 placebo patient devel- single callus (3.0 ml). No such migration
oped an ulcer at an injection site. A total of 2 was evident in any of our study patients.
patients from the placebo group developed No infection, rejection, inflammation, or References
unrelated diseases (CVA and malignancy). allergic reaction has been reported. Mor- 1. Boulton AJM: Late sequelae of diabetic neu-
No clinical evidence of any migration of phological studies have confirmed that med- ropathy. In Diabetic Neuropathy. Boulton
injected silicone was observed throughout ical-quality liquid silicone injected in the AJM, Ed. Lancaster, U.K., Marius Press,
the study. foot is a safe procedure (10,11). In the pres- 1997, p. 63–76
2. Mayfield JA, Reiber GE, Sanders LJ, Janisse
ent climate, silicone is still a controversial D, Pogach LM: Preventive foot care in peo-
CONCLUSIONS — This study confirms topic, but no overt proof has been uncov- ple with diabetes. Diabetes Care 21:2161–
the efficacy of injecting liquid silicone in the ered linking silicone implants and any 2177, 1998
diabetic foot in this first randomized double- chronic medical condition. 3. Veves A, Murray HJ, Young MJ, Boulton
blind placebo-controlled trial. Even with a In this study, the cushioning effect of the AJM: The risk of foot ulceration in diabetic
conservative intention-to-treat analysis injected silicone was still significant at 12 patients with high foot pressure: a prospec-
(assuming no change from baseline for all months after the first injection, but how long tive study. Diabetologia 35:660–663, 1992
missing data), the increase in plantar tissue this effect remains or whether it will decline 4. Young MJ, Coffey J, Taylor PM, Boulton
thickness and the reduction in peak plantar is not known. Anecdotal evidence suggests AJM: Weight bearing ultrasound in dia-
pressure were significantly greater in the sili- that most patients stay free of calluses and betic and rheumatoid arthritis patients. Foot
5:76–79, 1995
cone-treated group than in the placebo group. pain for many years (10,11), although 5. Murray HJ, Young MJ, Hollis S, Boulton
Because this is the first study reporting 50% of plantar injection sites will require AJM: The association between callus for-
on objectively measured parameters before booster injections at a later date (11). mation, high pressures and neuropathy in
and after liquid silicone injections, com- Whether the amount of pressure diabetic foot ulceration. Diabet Med 13:
paring our data with the anecdotal reports reduction and the increase in plantar tissue 979–982, 1996
of Balkin (10,11) is extremely difficult. Fur- thickness is enough to prevent future foot 6. Young MJ, Cavanagh PR, Thomas G, John-

DIABETES CARE, VOLUME 23, NUMBER 5, MAY 2000 637

Efficacy of injected liquid silicone

son MM, Murray H, Boulton AJM: The 2:83–88, 1991 thickness in the foot: a new ultrasound
effect of callus removal on dynamic plantar 11. Balkin SW: Fluid silicone implantation of technique for loadbearing measurements
foot pressures in diabetic patients. Diabet the foot. In Neale’s Common Foot Disorders: and a metatarsal head depth study. Foot 2:
Med 9:55–57, 1992 Diagnosis and Management. 5th ed. Lorimer 71–78, 1992
7. Colagiuri S, Marsden LL, Naidu V, Taylor L: D, Ed. Edinburgh, U.K., Churchill Living- 16. Leber C, Evanski PM: A comparison of
The use of orthotic devices to correct plan- stone, 1997, p. 387–400 shoe insole materials in plantar pressure
tar callus in people with diabetes. Diabetes 12. Young MJ, Breddy JL, Veves A, Boulton relief. Prosthet Orthot Int 10:135–138, 1986
Res Clin Pract 28:29–34, 1995 AJM: The prediction of diabetic neuro- 17. Veves A, Masson EA, Fernando DJS, Boulton
8. Foster A, Eaton C, Dastoor N, Jones K, pathic foot ulceration using vibration per- AJM: Studies of experimental hosiery in dia-
Crofton B, Edmonds M: Prevention of neu- ception thresholds. Diabetes Care 17:557– betic neuropathic patients with high foot
ropathic foot ulceration: a new approach 560, 1994 pressures. Diabet Med 7:324–326, 1990
using subdermal injection of collagen 13. Young MJ, Boulton AJM, Macleod AF, 18. Perry JE, Ulbrecht JS, Derr JA, Cavanagh
(Abstract). Diabet Med 5 (Suppl. 5):A27, Williams DRR, Sonksen PH: A multicentre PR: The use of running shoes to reduce
1988 study of the prevalence of diabetic periph- plantar pressures in patients who have dia-
9. Uccioli L, Faglia E, Monticone G, Favales F, eral neuropathy in the United Kingdom betes. J Bone Joint Surg Am 77:1819–1828,
Durola L, Aldeghi A, Quarantillo A, Calia P, hospital clinic population. Diabetologia 36: 1995
Menzinger G: Manufactured shoes in the 150–154, 1993 19. Schaff PS, Cavanagh PR: Shoes for the
prevention of diabetic foot ulcers. Diabetes 14. Veves A, Boulton AJM: The optical pedo- insensitive foot: the effect of a “rocker bot-
Care 18:1376–1378, 1995 barograph. Clin Podiatr Med Surg 10:463– tom” shoe modification on plantar pres-
10. Balkin SW, Kaplan L: Injectable silicone 470, 1993 sure distribution. Foot Ankle 11:129–140,
and the diabetic foot: a 25-year report. Foot 15. Bygrave CJ, Betts RP: The plantar tissue 1990

638 DIABETES CARE, VOLUME 23, NUMBER 5, MAY 2000