163 / 220
To account for the observed decline in HRQOL
before diagnosis, survival was evaluated according to the
timing of surveys before HNSCC diagnosis. Prediagnosis
HRQOL quartile was significantly associated with OS af-
ter HNSCC diagnosis after limiting analysis to surveys
from 1 to 5 years (N
5
428;
P
<
.001), 2 to 5 years
(N
5
239;
P
<
.001), and 3 to 5 years (N
5
155;
P
5
.006) before diagnosis.
The prognostic significance of clinically relevant 10-
point changes in HRQOL was also evaluated. A 10-point
increase in prediagnosis HRQOL was associated with a
14% reduced risk of death (HR, 0.86; 95% CI, 0.82-0.91)
(Table 5). Even after adjustment for other factors associated
with survival (age, smoking status, marital status, comorbid-
ities, household income, tumor site, stage, and radiother-
apy), each 10-point increase in the prediagnosis HRQOL
score was associated with a 9% reduction in the risk of death
(adjusted HR, 0.91; 95% CI, 0.85-0.97).
OPSCC
Most OPSCCs in the United States are human papillo-
mavirus (HPV)-related, and HPV-related OPSCC
(HPV-OPSCC) is considered a distinct disease from
HPV-negative HNSCC.
6
Therefore, we examined trends
in HRQOL and the prognostic significance of HRQOL
among individuals who had OPSCC compared with
those who had non-OPSCC. Two years before HNSCC
diagnosis, the HRQOL of individuals with OPSCC
(N
5
295) and non-OPSCC (N
5
1358) was similar
(Supporting Fig. 2, Supporting Table 3; see online sup-
porting information). However, 13 months after diagno-
sis, HRQOL for individuals with OPSCC was
significantly lower than for those with non-OPSCC (dif-
ference:
2
5.4; 95% CI,
2
9.0,
2
1.2). At 5 and 10 years
after diagnosis, HRQOL was again similar between the 2
groups. Overall, HRQOL was not significantly different
for individuals with OPSCC and those with non-OPSCC
(
P
5
.13). Higher prediagnosis HRQOL quartile was
nonsignificantly associated with improved survival after
OPSCC among 131 individuals who had prediagnosis
surveys available (HR, 0.95;
P
5
.32).
DISCUSSION
To our knowledge, this study is the first to examine both
prediagnosis and postdiagnosis HRQOL among individ-
uals with HNSCC. A significant and progressive decline
in HRQOL is observed before and after HNSCC diagno-
sis, and the magnitude of decline in HRQOL differs by
survival group. HRQOL in the years leading up to
TABLE 4.
Characteristics Associated With Health-Related Quality of Life (HRQOL) Among Individuals Diag-
nosed With Head and Neck Cancer in Which the Association With HRQOL Varies Significantly by Time
From Diagnosis
Difference in HRQOL:
D
(95% CI)
Univariate Analysis
a
Multivariate Analysis
a,b
Characteristic
2 Years
Prediagnosis
13 Months
Postdiagnosis
5 Years
Postdiagnosis
P
2 Years
Prediagnosis
13 Months
Postdiagnosis
5 Years
Postdiagnosis
P
Stage
<
.001
<
.001
Early
REF
REF
REF
REF
REF
REF
Late
0.3 (
2
3.2, 3.7)
2
10.1 (
2
13.5,
2
6.7)
2
0.5 (
2
4.1, 3.0)
0.2 (
2
2.8, 3.1)
2
6.8 (
2
9.8,
2
3.8)
1.5 (
2
1.5, 4.4)
Radiotherapy
.001
.01
No
REF
REF
REF
REF
REF
REF
Yes
2.2 (
2
1.2, 5.6)
2
7.3 (
2
10.6,
2
4.0)
2
0.8 (
2
3.9, 2.2)
1.9 (
2
1.1, 4.9)
2
5.1 (
2
8.1,
2
2.2)
2
1.2 (
2
3.9, 1.5)
Abbreviations:
D
, absolute difference in HRQOL compared with reference category; CI, confidence interval; REF, reference category.
a
Models include cubic spline terms for time relative to head and neck squamous cell carcinoma diagnosis.
b
Model includes all characteristics for which multivariate analysis results are reported in Tables 3 and 4, survey by proxy, and calendar year of diagnosis.
Figure 2.
Overall survival is illustrated according to health-
related quality-of-life quartiles assessed within 5 years before
head and neck cancer diagnosis.
HRQOL Before and After HNSCC/Rettig et al
Cancer
June 15, 2016
143