EUF Vol. 1 No. 1

87 96

due to other causes, but neither reached significance
(Supplementary Fig. 1B).
3.6. Pathway delays
The median time was 14 d (IQR 0–61 d) from initial onset
of symptoms to GP referral (time 1), 28 d (IQR 7–61 d)
from referral to hospital consultation (time 2), and 20 d
(IQR 0–50 d) from consultation to first treatment (time 3).
Patient characteristics by delay times are shown in
. Longer delays in Time 3 and Hospital Delay were
associated with smaller tumour size (both p < 0.05). A
longer total delay was observed for females (120 d vs
106 d, p < 0.05) and nonsmokers (118 d vs 105 d, p < 0.05)
when compared to other patients.
Analysis of survival by delay stratified for tumour stage
revealed no impact (Supplementary Table 2) except for
patients with MIBC, for whom a shorter delay in time 3
resulted in worse survival compared to those with a longer
delay (p < 0.05).
3.7. Predictors of survival
Univariate analysis identified histopathological criteria,
gender, delays, and smoking as factors associated with
UBC outcomes. Since these parameters are not necessarily
independent, multivariate analysis was used to determine
the impact of each factor. Cox regression of delay times
adjusted by a base model of independent factors revealed
that age, smoking status, and tumour stage (MIBC), grade
(G3), and size (>2 cm) were the most significant determi-
nants of poor outcome for UBC (Supplementary Table 3). No
significant influence was observed for delay, gender, or
occupational exposure.
4. Discussion
Here we report 17-yr outcomes for newly diagnosed cases
of UBC within a large geographic region in the UK. We have
updated an initial report and have now followed a large
proportion of cases (75%) until death. We are thus able to
examine the complex interaction among tumour character-
istics, patient gender, carcinogen exposure, pathway delays,
and mortality. Univariate and competing-risks analysis
revealed that many of these factors are associated with
disease-specific mortality. However, multivariate analysis
identified age, smoking status, and tumour stage, grade, and
size as the most significant determinants of poor outcome
for UBC. Notably, there was no significant influence of delay,
gender, or occupational exposure.
Comparison of CIFs revealed significant associations
between female gender and higher cumulative incidence of
death from grade 3 disease and MIBC, concurring with
previous reports of worse outcomes for females with
UBC Furthermore, there was a trend for
female patients to present more commonly with MIBC than
male patients. Importantly, female patients experienced
a significantly longer total delay than male patients. The
majority of this delay occurred in time 1, before GP referral
for investigation in secondary care; a significantly higher
proportion of female patients with visible haematuria
encountered longer delays in time 1 than equivalent male
patients. These data support observations of repeated
community-based treatments for suspected urinary infec-
tion in symptomatic females . As reported by
Hollenbeck et al there were no significant differences
in delays between the genders once patients were within
secondary care.
Female patients with MIBC had a significantly higher
cumulative incidence of death from UBC than male
patients; it is unlikely that differential utilisation of
radiotherapy or cystectomy between the genders would
cause this effect, but there is limited evidence to suggest
that female patients have worse outcomes from radiother-
apy compared to males However, such effects were not
large enough for gender to be an independent prognostic
factor in multivariate analysis when adjusted for pathway
delays.
It is a commonly held belief that more rapid cancer
diagnosis and treatment lead to better outcomes, and to
suggest otherwise is counterintuitive . However, the
relationship between delay and survival in UBC is complex
, with no direct linear relationship with any compo-
nents of delay . In the long-term follow-up of this
cohort, we confirmed this complex relationship, as noted by
others : no delay category had a significant influence
on survival, except for patients with MIBC, for whom a
shorter time for delay 3 was detrimental. This may
represent an anomaly, and there is no clear explanation
from our data, but it is feasible that patients with MIBC
with concerning features (eg, ongoing bleeding) or comor-
bidities were selected for expedited treatment , and
subsequently succumbed more rapidly as a result of those
features or comorbidities. This was also postulated by
Liedberg et al , who observed that a long treatment delay
had no influence on survival following cystectomy. It seems
that once patients are in secondary care, clinicians are good
at selecting those with the highest risk and treating them
rapidly . Similarly, Nielsen et al found that delay
from TURBT to radical cystectomy was not independently
associated with stage progression or decreased recurrence-
free or disease-specific survival. Likewise, for UBC patients
treated by radiotherapy, there is no significant influence of
treatment delay on survival However, Hollenbeck et al
investigated delay and survival in 29 826 patients with
UBC and found that longer delays from presentation to
diagnosis were associated with increased risk of bladder
cancer–specific mortality, a finding also observed by Gore
et al when assessing the interval between TURBT and
cystectomy.
Many patient-related factors analysed here are not
independent; for example, males are more likely to
smoke, to have occupational carcinogen exposure, and to
be more rapidly referred for investigation of haematuria
. Females are more likely to be nonsmokers, are
typically exposed to different occupational carcinogens,
and experience slower referral for investigation of hae-
maturia or lower urinary tract symptoms In
E U R O P E A N U R O L O G Y F O C U S 1 ( 2 0 1 5 ) 8 2 – 8 9 87