Volume 77 • Number 12

Case Series
Postoperative Complications Following Gingival
Augmentation Procedures
Terrence J. Griffin,* Wai S. Cheung,* Athanasios I. Zavras,† and Petros D. Damoulis*

Background: Postoperative pain, swelling, and KEY WORDS

bleeding are the most common complications follow- Allograft; connective tissue; postoperative
ing soft tissue grafting procedures; however, detailed complications; smoking; transplantation,
documentation is sparse in the literature. The aims autologous.
of this prospective study were as follows: 1) to com-
pare the frequency of complication occurrence after

T
he free soft tissue graft (FSTG) and subepi-
free soft tissue grafting (FSTG) or subepithelial con- thelial connective tissue graft (SCTG) are
nective tissue grafting (SCTG) procedures; 2) to eval- among the most commonly used and widely
uate the use of an acellular dermal matrix (ADM) studied gingival augmentation procedures.1-12 Both
as the donor tissue alternative to an FSTG or SCTG; techniques have several applications, including in-
and 3) to identify possible predictors for these compli- creasing keratinized tissue,1,2 soft tissue thickness,3,4
cations. and vestibular depth;5 reducing gingival recession;6-8
Methods: Seventy-five FSTG and 256 SCTG proce- and replacing pigmented and pathological oral mu-
dures were performed in 228 patients by a single cosa.9,10 The predictability and long-term stability of
operator. In five free soft tissue and 84 bilaminar the surgical outcome for these procedures have also
graft procedures, an ADM was used instead of autog- been well established.11,12
enous tissue. Variables such as the duration and Complications and atypical healing responses after
location of procedures, smoking history, gender, FSTG or SCTG procedures have been reported in the
and age were recorded. Patients were asked to fill literature,13-17 albeit in a non-systematic way. Most
out a questionnaire 1 week after the surgeries regard- documented complications are associated with the
ing postoperative pain, swelling, and bleeding. donor site. Harvesting of a free soft tissue graft can re-
Data were analyzed using the x2 test and logistic sult in excessive hemorrhage, postoperative bone ex-
regression analysis. Odds ratios were calculated for posure, and recurrent herpetic lesions associated with
moderate and severe adverse outcomes grouped to- a second surgical procedure leaving a painful, open
gether. palatal wound.13 Cases of a mucocele14 and an arterio-
Results: The duration of surgical procedures was venous shunt15 have also been reported. Complica-
highly correlated with pain or swelling post-surgi- tions associated with the SCTG include necrosis of
cally (P = 0.001). Current smokers were three times graft and palatal tissue, excessive bleeding, prolonged
more likely to experience post-surgical swelling (P = pain/discomfort or infection at the donor and/or recip-
0.01). Patients who underwent FSTG procedures ient sites, and graft shrinkage.18 The main concern for
were three times more likely to develop post-surgical the donor site is tissue necrosis when palatal thickness
pain (P = 0.002) or bleeding (P = 0.03) compared to is inadequate or primary closure is not achieved.19,20
those who received SCTG procedures. When an An association between extensive necrosis and post-
ADM was applied instead of autogenous tissue, the operative pain has been suggested, when two SCTG
probability of swelling or bleeding was significantly harvesting techniques were evaluated.21 As a result,
reduced (odds ratio [OR] = 0.46, P = 0.02 and OR = several modifications for harvesting donor tissue
0.3, P = 0.001, respectively). have been proposed in an effort to preserve blood
Conclusions: Long surgical procedures and smok- supply and palatal tissue thickness and achieve pri-
ing may increase the severity and frequency of certain mary wound closure.22-24 Furthermore, a number of
post-surgical complications after gingival augmenta-
tion procedures. FSTG procedures incur a higher like- * Department of Periodontology, Tufts University School of Dental Medicine,
lihood for postoperative pain or bleeding than SCTG Boston, MA.
† Department of Health Policy and Epidemiology, Harvard University School
procedures, whereas the application of an ADM may of Dental Medicine, Boston, MA.
significantly reduce the probability of swelling and
bleeding. J Periodontol 2006;77:2070-2079. doi: 10.1902/jop.2006.050296

2070
J Periodontol • December 2006 Griffin, Cheung, Zavras, Damoulis

complications have been described for recipient sites. Table 1.

These include unexpected post-surgical swelling and
Patient Demographics
ecchymosis,25 external root resorption,26 gingival sur-
gical cysts,27-29 exostosis,16,17,30 and graft loss.31 Re-
Total subjects 228
cently, three cases were reported in which soft tissue
abscesses were found at the recipient sites following Total procedures 331
an SCTG anywhere from 6 days to 7 weeks.32 Reaction
to the suture material (‘‘stitch abscess’’) or possible Age (years)
epithelial implantation was suspected to be the cause. Mean – SD 48.5 – 15.2
One apparent shortcoming of both procedures is Range 19-84
the autogenous graft harvesting. It creates a second Gender
wound site, causing longer chair time, higher possibil- Male 79
ity of tissue morbidity, and intra- and/or postoperative Female 149
discomfort, which can result in lower patient accep-
tance. Acellular dermal matrix (ADM) was introduced Smoking (data available for 225 subjects)
to gingival augmentation procedures in the late Non-smoker 198
1990s.33 It is a processed, cell-free material that is Current smoker 27
obtained from human skin and composed of struc-
turally integrated basement membrane complex
tion to the above criteria, teeth with lack of attached
and extracellular matrix. ADM has been used as an
gingiva and/or gingival recession were assigned for
alternative to the autogenous soft tissue grafts in the
treatment when at least one of the following conditions
treatment of various mucogingival deformities in an
applied: 1) the defects were progressively receding
effort to avoid the second surgical site and overcome
with sensitivity or root decay; 2) the apico-coronal di-
graft size limitations.4,12,34-39
mension of the defect was ‡2 mm; 3) the tooth (teeth)
Postoperative complications after gingival aug-
with recession was (were) scheduled to be restored
mentation procedures, such as pain, swelling, and
with prosthesis or become an abutment for a remov-
bleeding, have often been documented and associ-
able partial prosthesis; 4) the tooth (teeth) with reces-
ated with poor surgical outcomes.25,40 However, no
sion had old prosthesis(es) or root restoration(s) that
systematic effort has been made to date to quantify
were planned to be renewed; 5) the affected tooth
them or identify any potential modifiers that could
(teeth) was (were) esthetically unpleasant to the
help reduce their frequency and/or severity. The aims
patient; and 6) the teeth were treatment planned to
of this prospective study were as follows: 1) to com-
be labially moved for orthodontic reasons. Smokers
pare the frequency of occurrence for pain, swelling,
were included, but subjects who were pregnant or
and bleeding after FSTG or SCTG procedures; 2) to
presented with poor oral hygiene were excluded.
evaluate any effect the application of an ADM as
Each subject received a comprehensive periodon-
the donor tissue alternative to an FSTG (ADMF) or
tal examination. Surgical options were discussed and
SCTG (ADMS) might have on frequency and/or se-
a treatment plan was formulated based on treatment
verity of these complications; and 3) to identify pos-
needs and preferences of subjects. Oral hygiene in-
sible predictors for these complications.
structions, full-mouth debridement or scaling and root
MATERIALS AND METHODS planing under local anesthesia, and occlusal adjust-
ment were rendered as indicated before the surgical
Study Population
phase to all subjects. All surgical procedures were
A total of 228 subjects (79 males and 149 females; age
performed by a single operator (TJG).
range, 19 to 84 years) diagnosed with mucogingival
deformities from January 1997 to December 2002 Surgical Procedures
participated in this study (Table 1). The study protocol Procedures were performed as previously described
was approved by the Tufts-New England Medical Cen- with minor modifications.34,36
ter Institutional Review Board with patient consent. In- FSTG and ADMF. A periosteal bed free from muscle
clusion criteria were as follows: 1) absence of any attachment and of sufficient size was prepared
uncontrolled local or systemic disease where peri- through sharp dissection. For an FSTG, a graft of
odontal surgery might be contraindicated; 2) presence the appropriate size was dissected from the palate
of at least one tooth with Miller Class I or II gingival re- using an angled blade holder.‡ For an ADMF, ADM
cession,41 which also presented with £1 mm of at- allografts§ were rehydrated in sterile saline for 10 to
tached gingiva; 3) presence of clinical probing
depths £3 mm; and 4) absence of clinically apprecia- ‡ Hartzell & Son, Concord, CA.
ble gingival inflammation in the affected teeth. In addi- § Alloderm, LifeCell, Branchburg, NJ.

2071
Complications After Mucogingival Surgery Volume 77 • Number 12

30 minutes and trimmed to fit the recipient beds prior compiled by one of the authors (WSC) who was not
to use. Grafts were transferred to the prepared re- involved in formulating the treatment plan. The levels
cipient site and sutured with a 5-O bioabsorbable of complications were classified as none to minimum
polyglactin materialk and a P-3 needle. if the score was ‘‘0 to 3,’’ moderate for ‘‘4 to 6,’’ and
SCTG and ADMS. After an initial horizontal right- severe for ‘‘7 to 10.’’ None to minimal pain meant little
angle incision beveled into the adjacent interdental or no discomfort; moderate was any pain that both-
papillae at or slightly coronal to the cemento-enamel ered the patient and mildly affected normal function;
junction (CEJ), a full-thickness flap was reflected be- and severe was considered any pain that could not
yond the mucogingival junction and further released be tolerated and even disrupted the patient’s daily
by sharp dissection to allow for coronal displacement functions.42 None to minimal swelling ranged from
with minimal tension. The buccal part of the papilla no abnormal feeling or visible change in appearance
was deepithelized, and the exposed root surface was to a feeling of ‘‘fat’’ or enlargement of intra- or extra-
planed meticulously with hand and ultrasonic instru- oral soft tissue correlating to the surgery; moderate
ments. For the SCTG, a connective tissue graft was swelling indicated a slight visible change in the size/
obtained from the palate in the area of the first molar shape of the soft tissue in addition to the feeling;
to the lateral incisor. Two horizontal incisions were and severe swelling was defined as a very noticeable
made: 1) ;2 mm apical to the gingival margin; and change in the size/shape of the soft tissue. None to
2) 1 to 2 mm apically. No vertical incisions were used. minimal bleeding ranged from no detectable bleeding
The second incision was made parallel to the surface to a trace of blood clotted without any treatment;
of the palate and was carried far enough apically (;12 moderate bleeding was any oozing or mild bleeding
to 13 mm) to provide a sufficient height of connective which would stop by ice application; severe bleeding
tissue. A small periosteal elevator was used to raise a was considered any bleeding that could not be stop-
full-thickness periosteal connective tissue graft. The ped at home upon telephone instructions. The scoring
small band of epithelium and all adipose tissue were scales for bleeding and swelling followed the same
removed and the width and thickness of the graft were design as the pain scale, because no systematic
appropriately adjusted. For an ADMS, the ADM allo- approach for evaluating the former was found in the
grafts were managed as previously described and ori- literature.
ented in such a way that the basement membrane All demographic and procedure-related informa-
faced the root surface. The grafts were placed over tion (e.g., gender; age; type, location, and length of
the denuded roots and stabilized by continuous sling the procedure; and number of teeth involved) was
sutures with 5-0 bioabsorbable polyglactin material¶ taken from the patients’ dental charts. All patients
and a P-3 needle. The flap was coronally positioned who stated at the initial consultation appointments
to completely cover the graft and secured by using that they were smoking were educated to its adverse
continuous vertical mattress and sling sutures into effects to periodontal tissue and surgery. In addition,
the mesial and distal papillae. At the same time, a con- smoking cessation was suggested. Smoking status
tinuous locking suture was applied to the donor site was assessed again just before the surgery. Patients
with the same material to approximate the edges of who never smoked, quit smoking at least 30 days be-
the donor flap and achieve hemostasis. fore the surgery, or smoked less than five cigarettes
For all procedures, a non-eugenol periodontal per day were categorized as ‘‘non-smokers.’’ Patients
dressing# was placed over recipient and donor sites. who smoked more than five cigarettes per day were
Subjects were placed on 0.12% chlorhexidine gluco- categorized as ‘‘current smokers.’’
nate mouthrinse** and given standard postoperative
Statistical Analysis
instructions. No systemic antibiotics were used. The
Descriptive statistics were expressed as means –
dressing was replaced at the first week post-surgical
standard deviations and/or frequency distributions.
visit. The dressing and sutures were removed 2 weeks
Chi-square tests were performed to explore pos-
after the surgery, and subjects received routine peri-
sible risk factors that might associate with severe
odontal care afterwards.
complications for each procedure. Predictors that
may introduce severe postoperative complication risk
Questionnaire and Information Collection
were assessed by logistic regression analysis. Multi-
At the 1-week follow-up appointment, every patient
variable logistic regression calculates the effect of
was given a questionnaire to rate postoperative pain,
each risk factor while simultaneously controlling for
swelling, and bleeding for the previous week, and over-
all discomfort on day 7 on a ‘‘0 to 10’’ scoring scale.
The questionnaire was given by one of the practice k Vicryl, Ethicon, Johnson & Johnson, Sommerville, NJ.
¶ Vicryl, Ethicon, Johnson & Johnson.
assistants who was not involved in the study. The in- # Coe-Pak, GC America, Alsip, IL.
formation from patient charts and questionnaires was ** Peridex, Procter & Gamble, Cincinnati, OH.

2072
J Periodontol • December 2006 Griffin, Cheung, Zavras, Damoulis

the effects of the rest. The level of significance for re- incidence of moderate or severe pain compared to
jection of the null hypothesis was set at a = 0.05. free soft tissue grafting (27% versus 38.7%) but with
a higher incidence of moderate or severe swelling
RESULTS (31.6% versus 21.3%). Only a small percentage of
A total of 331 procedures were included in the analy- subjects (<6%) experienced moderate or severe
sis in 228 subjects (Table 1). Interestingly, the female- bleeding and only in the groups that autogenous tis-
to-male ratio was almost 2:1 in this population. The sue was used. Moderate or severe discomfort after
majority of subjects were non-smokers (198 non- 1 week was reported only after SCTG procedures by
smokers versus 27 current smokers). In 75 cases, relatively few subjects (7.6%).
an FSTG was used, 70 with an autogenous graft and To identify possible risk indicators for pain, swell-
five with an ADM (Table 2). A bilaminar technique ing, and bleeding, moderate and severe occurrences
was performed in 256 procedures, and an ADM was were grouped and analyzed in relation to procedure
used as the graft material in 84 of them. duration, number of teeth, location (maxilla versus
Duration, number and location of teeth, and the age mandible), age, gender, and current smoking (Tables
and gender of patients receiving each procedure are 4 through 6). The most significant risk indicator for
presented in Table 2. Overall, procedures using au- postoperative pain was duration of the procedure,
togenous grafts (FSTG or SCTG) involved fewer teeth particularly in subjects who received autogenous
and were relatively shorter in duration than the ones grafts (odds ratio [OR]: 1.08 for the FSTG and 1.04
involving ADM. Also, bilaminar procedures lasted for the SCTG per minute of procedure; both P
longer than the ones involving free soft tissue grafts. <0.05) (Table 4). Smoking was also significantly as-
However, those differences were not statistically sig- sociated with an increased risk for pain but only in
nificant, except when the number of teeth between the SCTG group (OR: 2.82; confidence interval [CI]
the SCTG and ADMS was compared (ADMS > SCTG; 1.06 to 7.47, P <0.05). Interestingly, in the FSTG
P <0.01). Age and gender distribution per procedure and ADMS groups, smoking appeared to decrease
reflected the overall sample demographics, with a ten- the risk for postoperative pain (however, not statisti-
dency for younger individuals to receive bilaminar cally significant). Subjects who underwent the ADMS
procedures (statistically significant only for SCTG procedure in the maxilla had a higher risk to develop
versus FSTG; P <0.001). The majority of the free soft pain compared to subjects who had this procedure in
tissue graft procedures were performed in the mandi- the mandible (OR: 2.1; CI: 1.34 to 3.26; P <0.01).
ble, whereas the bilaminar procedures showed a more There was no association between pain and age or
normal distribution between the two arches. gender. The exact same pattern was observed when
The type and severity of complications within risk indicators for swelling were assessed (Table 5).
the week after the surgery and the discomfort at the Once more, lengthy procedures were more likely to
1-week follow-up for each procedure are shown in result in moderate or severe swelling, when autoge-
Table 3. Pain and swelling were the most significant nous tissue was used. This time smoking was associ-
complications, with 27% to 40% of subjects reporting ated with increased likelihood of swelling for the three
moderate or severe pain, and 19% to 60% of them procedures analyzed (there were no data for ADMF);
reporting moderate to severe swelling. In general, however, statistical significance was reached again
bilaminar procedures were associated with a lower only in the SCTG group (OR: 4.88; CI: 1.75-13.63;

Table 2.
Descriptive Statistics for Each of the Four Procedures

Location Distribution Gender Distribution

per Procedure per Procedure
Procedures Average Duration of N Teeth
Type of Performed Procedure (minutes; Involved per Maxilla Mandible Both Age (years; Male Female
Procedure (N) mean – SD) Procedure (%) (%) (%) mean – SD) (%) (%)

FSTG 70 40.5 – 13.8 2.4 – 1.4 10 (14.3) 59 (84.3) 1 (1.4) 54.8 – 16.8 24 (34.3) 46 (65.7)

SCTG 172 45.1 – 19.3 2.6 – 1.4 91 (52.9) 65 (37.8) 16 (9.3) 46.1 – 15.1 57 (33.1) 115 (66.9)

ADMF 5 46.0 – 12.9 3.4 – 3.2 1 (20.0) 4 (80.0) _ 55.8 – 21.2 2 (40.0) 3 (60.0)

ADMS 84 49.5 – 17.8 3.5 – 2.6 31 (36.9) 37 (44.1) 16 (19.0) 45.4 – 11.2 25 (29.8) 59 (70.2)
– = not applicable.

2073
Complications After Mucogingival Surgery Volume 77 • Number 12

Table 3.
Type and Severity of Complications per Procedure

Discomfort at 1-Week
Type of
Pain Swelling Bleeding Follow-Up
Complication
Severity* I (%) II (%) III (%) I (%) II (%) III (%) I (%) II (%) III (%) I (%) II (%) III (%)

Type of Procedure

FSTG† 42 (60.9) 26 (37.7) 1 (1.4) 57 (81.4) 13 (18.6) _ 66 (94.3) 3 (4.3) 1 (1.4) 70 (100) _ _

SCTG 126 (73.3) 40 (23.2) 6 (3.5) 112 (65.1) 52 (30.2) 8 (4.7) 170 (98.8) 1 (0.6) 1 (0.6) 159 (92.4) 11 (6.4) 2 (1.2)

ADMF 3 (60.0) 2 (40.0) _ 2 (40.0) 3 (60.0) _ 5 (100) _ _ 5 (100) _ _

ADMS 61 (72.6) 19 (22.6) 4 (4.8) 63 (75.0) 16 (19.1) 5 (5.9) 84 (100) _ _ 84 (100) _ _

– = not applicable.
* I = absent or mild; II = moderate; III = severe.
† One subject did not report pain status.

Table 4.
Risk Indicators for Moderate or Severe Pain per Procedure

Duration Location (maxilla Gender (female

Surgery Factors (minutes) N Teeth versus mandible) Age (years) versus male) Smoking

FSTG OR 1.08* 1.15 0.46 0.99 1.75 0.56

CI 1.04-1.14 0.81-1.65 0.17-1.23 0.97-1.02 0.60-5.08 0.10-3.12

SCTG OR 1.04* 1.3* 0.77 0.99 1.03 2.82*

CI 1.01-1.06 1.02-1.65 0.52-1.15 0.97-1.06 0.50-2.12 1.06-7.47

ADMF OR – 1.89 – 0.94 1.5 –

CI – 0.44-8.10 – 0.84-1.06 0.012-19.56 –

ADMS OR 1.02 0.98 2.1* 0.96 1.75 0.51

CI 0.8-1.65 0.81-1.18 1.34-3.26 0.92-1.01 0.56-5.41 0.56-4.61
– = not applicable.
* P <0.05.

P <0.05). Interestingly, none of the analyzed vari- significant (Table 7). This analysis confirmed that the
ables was associated with an increased risk for mod- duration of the surgical procedure was the most im-
erate or severe bleeding (Table 6). There were a few portant risk indicator for postoperative moderate or
interesting trends with smoking increasing the risk severe pain and swelling. For each minute of the
for bleeding in FSTG (OR: 2.4) and females being procedure, there was a 4% increase in the probability
more likely to bleed after an SCTG than males (OR: of developing moderate or severe pain (OR: 1.04;
3.62); however, both were not statistically significant. CI: 1.02 to 1.06; P = 0.001) and a 3% increase for
No postoperative bleeding cases were reported when moderate or severe swelling (OR: 1.03; CI: 1.02 to
the ADM was used as the graft material, indicating 1.05; P = 0.001). Current smoking was an important
that postoperative bleeding is usually associated with indicator for postoperative swelling, increasing the
the donor site surgery. risk almost three-fold (OR: 2.88; CI: 1.26 to 6.55;
To study predictors of moderate or severe out- P = 0.01) but not for pain or bleeding. Age and gender
comes for each type of complication, logistic regres- were not associated with any complication. The lo-
sion was used, controlling for possible confounders. gistic regression model also gave us the opportunity
In an effort to control for unknown or residual con- to compare the free soft tissue against the bilaminar
founders, all models contained gender, age, and surgical procedure and the use of autogenous versus
smoking, even if these variables were not statistically ADM graft tissues, while controlling for the other

2074