169 / 220
Pre-treatment and post-treatment care
All patients underwent pre-treatment nursing, dietetic and
speech and language therapy pre-treatment assessments as part
of routine practice. During chemoradiotherapy treatment, all
patients were routinely reviewed twice weekly by medical and
nursing teams, with additional reviews with the dietetic and
speech and language therapy teams as required. Post-treatment,
all patients were offered ongoing dietetic and speech and language
therapy support in post-treatment rehabilitation clinics.
Statistical analysis
Statistical analyses were performed using STATA software ver-
sion 12 (Statacorp, Texas, USA) and GraphPad Prism version 5. Pro-
phylactic gastrostomy versus NG tube as required matching was
performed using T stage (grouped as T1/2 and T3/4) and N stage
(grouped as N0/1, N2a/b, and N2c/3). Unmatched variables were
compared using the Student
t
-test, Fisher’s exact test and test for
trend. A
t
-test and chi-squared test were used as appropriate to
test for differences in the subgroups analysed. MDADI scores were
compared using a non-parametric Mann-Whitney
U
test. Univari-
ate non-parametric analyses were performed to determine if there
was any correlation between MDADI scores with clinical parame-
ters. Variables included were: age, T stage, N stage, use of induction
chemotherapy, number of induction chemotherapy cycles, type of
concurrent chemotherapy, number of concurrent chemotherapy
cycles, mean contralateral parotid dose, pre-radiotherapy diet. Sta-
tistical significance was declared at
p
< 0.05. A multivariate MAN-
OVA was performed on factors with a
p
-value < 0.20 on
univariate analysis (N-stage, number of induction chemotherapy
cycles, concurrent chemotherapy type, number of concurrent
chemotherapy cycles, pre-treatment oral intake and mean con-
tralateral parotid gland dose) The natural logarithm of MDADI
scores were used for the multivariate analysis to normalise the
data.
Results
From review of electronic records, 94 patients with a diagnosis
of squamous cell carcinoma of the oropharynx who had received
definitive concurrent chemoradiotherapy with curative intent
were identified. Of these patients the following were excluded:
18 patients had experienced disease recurrence, 3 patients had
died post-treatment with no evidence of disease recurrence, 5
had commenced therapeutic enteral feeding pre-treatment. Of
the remaining 68 patients, completed MDADI questionnaires were
received from 59 (87%) patients. 31/59 (53%) had been managed
with a prophylactic gastrostomy tube, and 28/59 (47%) had been
managed with a reactive approach with insertion of a nasogastric
tube if required during treatment.
A retrospective matched pair analysis was performed, matching
26/31 patients managed with a prophylactic gastrostomy with
26/28 managed with a nasogastric tube if required, on the basis
of T and N stage. On review of case notes, 24 of these matched
56 patients (43%) were documented as being offered a choice of
either a prophylactic gastrostomy or a reactive approach; 12/24
(50%) opted for a prophylactic gastrostomy. Of the 26 patients
managed with a prophylactic gastrostomy tube, the tube was doc-
umented as being used for at least supplemental nutrition in all
patients. Within the reactive NG tube if required group, 17/26
(65%) patients had an NG tube inserted and commenced enteral
feeding during treatment; no gastrostomies were used in this
group. Mean follow up from the last day of radiotherapy was
36 months (range 24–59) and 34 months (range 24–59) for
patients managed with a prophylactic gastrostomy or NG tube as
required respectively.
The baseline matched and unmatched characteristics, along
with treatment details and pre-treatment diet and shown in
Table 1
. With regard to the matched factors, the T stage distribu-
tion was similar between the two groups. For the purposes of
matching analysis, N0 and N1 were grouped together;
Table 1
shows that there was a small excess of patients with N0 disease
in the group of patients managed with a prophylactic gastrostomy,
although there is no statistically significant difference in N stage
between the two groups. With regard to unmatched factors,
Table 1
shows that there were no statistically significant differences in
baseline characteristics between the matched groups, including
Table 1
Patient, tumour and treatment details.
Prophylactic
gastrostomy (
N
= 26)
NG as needed
(
N
= 26)
p
-
value
Median follow up
34
30
0.31
Age
(
Mean
,
range
)
56 (36–66)
55 (38–68)
0.55
Sex
Male
20 (76.9%)
22 (84.6%)
0.48
Female
6 (23.1%)
4 (15.4%)
WHO PS
0
14 (53.8%)
20 (76.9%)
0.22
1
4 (15.4%)
2 (7.7%)
Not recorded
8 (30.8%)
4 (15.4%)
Smoking
Never
10 (38.5%)
10 (38.5%)
0.87
Ex
13 (50%)
11 (42.3%)
Current
2 (7.7%)
3 (11.5%)
Not recorded
1 (3.8%)
2 (7.7%)
Oropharynx subsite
Tonsil
13 (50.0%)
15 (57.7%)
0.55
BOT
12 (46.2%)
11 (42.3%)
Posterior pharynx
1 (3.8%)
0 (0%)
T stage
T1
5 (19.2%)
6 (23.1%)
0.94
T2
13 (50%)
12 (46.2%)
T3
2 (7.7%)
3 (11.5%)
T4
6 (23.1%)
5 (19.2%)
Nodal stage
N0
5 (19.2%)
0 (0%)
0.11
N1
1 (3.8%)
5 (19.2%)
N2a
2 (7.7%)
2 (7.7%)
N2b
15 (57.7%)
16 (61.5%)
N2c
3 (11.5%)
2 (7.7%)
N3
0 (0%)
1 (3.8%)
Induction chemotherapy
None
18 (69.2%)
24 (92.3%)
0.09
TPF
6 (23.1%)
2 (7.7%)
PF
2 (7.7%)
0 (0%)
RT dose
65 Gy in 30 fractions
1 (3.8%)
1 (3.8%)
0.6
70 Gy in 35 fractions
24 (92.3%)
24 (92.3%)
Mean contralateral parotid
dose
(
range
)/
Gy
37 (24–54)
33 (21–57)
0.06
Concurrent chemotherapy
Cisplatin
22 (96.2%)
24 (92.3%)
0.39
Carboplatin
4 (3.8%)
2 (7.7%)
No. of concurrent chemotherapy cycles
1
4 (15.4%)
6 (23.1%)
0.45
2
22 (84.6%)
19 (73.1%)
3
0 (0%)
1 (3.8%)
Pre-treatment oral intake
NBM
0 (0%)
0 (0%)
0.38
Sips
0 (0%)
0 (0%)
Pureed
2 (7.7%)
0 (0%)
Soft
3 (11.5%)
3 (11.5%)
Normal
21 (80.8%)
21 (80.8%)
Not recorded
0 (0%)
2 (7.7%)
B. Sethugavalar et al. / Oral Oncology 59 (2016) 80–85
149