95 / 236
Oakley et al.
review. Of these, 6 did not fulfill the CRS diagnostic criteria
by symptoms. Two of these 6 did not have nasal endoscopy
or CT evidence of CRS, whereas 4 who did not have qualify-
ing symptoms nonetheless had CT evidence of CRS. These
4 patients with “negative” cardinal symptoms but “posi-
tive” CTs had the following profiles: 1 had silent sinus syn-
drome, 1 had facial pain with no other symptoms, 1 had
a history of previous surgery elsewhere for CRS but was
asymptomatic at the time of the visit in the database, and 1
had altered mental status due to intracranial complications
of rhinosinusitis and symptoms could not be obtained. Of
the remaining 73 patients with qualifying symptom pro-
files, 2 did not have a confirming endoscopy or CT. This
review therefore resulted in 71 of 79 patients meeting the
2007 criteria by Rosenfeld et al.
13
for diagnosis of CRS,
with a positive predictive value (PPV) of 90%.
The charts were then reviewed for presence/absence of
polyps in the clinical record and compared to the coding re-
sults in the database for each patient. Of the 79 CRS charts
available for review, 62 carried the ICD-9 diagnosis 473.x
without 471.x and were considered CRSsNP, whereas 17
carried the ICD-9 diagnosis for CRSwNP, 471.x. Of the 62
putative CRSsNP from diagnosis codes, 48 were confirmed
on chart review whereas the remaining 14 were determined
to be clinical CRSwNP. Of the 17 putative CRSwNP from
diagnosis codes, 8 were confirmed on chart review whereas
the remaining 9 patients were determined to be CRSsNP.
Thus, the accuracy of using billing codes available in the
UPDB to identify CRSsNP or CRSwNP was 71% based on
our chart review.
Demographic characteristics
The characteristics of 1638 CRSwNP and 24,200 CRSsNP
adult cases in Utah and 5:1 matched population controls
are shown in Table 1. Patients with CRSwNP and CRSsNP
were similar in age, with a mean age at diagnosis of 44 years
and 43 years, respectively. In contrast, CRSsNP probands
were somewhat more likely to be female (52%), whereas
the majority of CRSwNP probands were male (55%). This
difference in gender percentage between CRSsNP and CR-
SwNP was statistically significant (
p
<
0.0001, chi-square
analysis). Consistent with the Utah population and similar
to other regions in the United States, cases and controls in
our study were predominantly non-Hispanic Whites; how-
ever, the proportion of non-Whites was higher in controls
compared to cases. The racial differences between cases
and controls, which were solely selected on sex and birth
year, were statistically significant for both CRSsNP and
CRSwNP (both
p
<
0.0001, chi-square analysis).
Familial risk of CRS
The familial risk of CRSwNP and CRSsNP is shown in Ta-
ble 2. First-degree relatives of case probands with CRSwNP
had a 4.1-fold increased risk (95% confidence interval [CI],
1.8 to 9.4;
p
<
10
−
3
) of having the same diagnosis com-
pared to population controls. Within this group, parents,
siblings, and children contributed to the overall 1stDRs
risk with similar HR estimates (data not shown). Second-
degree relatives of CRSwNP probands had a 3.3-fold in-
creased risk (95% CI, 1.5 to 7.5;
p
=
0.004). There was no
significantly increased risk beyond 2ndDRs, although the
HR in first cousins was suggestive of an increased risk in
3rdDR cases. We observed no increased risk in spouses of
CRSwNP probands compared to controls.
In the CRSsNP group (Table 2), 1stDRs had a 2.4-fold
increased risk (95% CI, 2.2 to 2.6;
p
<
10
−
15
) of carrying
the same diagnosis compared to controls, 2ndDRs had a
1.4-fold increased risk (95% CI, 1.3 to 1.4;
p
<
10
−
15
),
whereas 3rdDRs (first cousins) of cases exhibited a modest
but significant increased risk at 1.1-fold (95% CI, 1.08 to
1.2;
p
<
10
−
7
; and 95%CI, 1.0 to 1.2;
p
<
0.02). More dis-
tant cousins of cases (fourth-degree and fifth-degree) also
exhibited a slight increased risk of CRSsNP (HRs
=
1.06;
95% CI, 1.03 to 1.08;
p
<
10
−
12
). In contrast to CR-
SwNP, spouses of CRSsNP probands carried a 2-fold in-
creased risk of CRSsNP themselves (95% CI, 1.8 to 2.2;
p
<
10
−
15
).
We also calculated familial risk of CRSsNP in relatives
of CRSwNP case probands, to examine if the risk of a non-
polyposis phenotype was also elevated in their family mem-
bers (Table 3). In 1stDRs of CRSwNP cases, we observed a
2.5-fold increased risk of CRSsNP (95% CI, 2.1 to 3.0;
p
<
10
−
15
), whereas the increased risk of CRSsNP in 2ndDRs
of CRSwNP probands was 1.4-fold (95% CI, 1.2 to 1.7;
p
<
0.001). In the reverse comparison in which familial risk
of CRSwNP in CRSsNP probands was calculated, similar
risk estimates were observed.
Discussion
The complex nature of CRS has thus far limited our full un-
derstanding of the pathogenesis of this condition, and there-
fore our ability to treat it effectively and consistently. Not
only is the etiology believed to be multifactorial, with both
genetic and environmental influences, but it also presents as
an array of various phenotypes or endotypes with associ-
ated comorbidities. Prior research into the genetics of CRS
has targeted multiple levels of immunologic susceptibility
and comorbid links, including multiple human leukocyte
antigen (HLA) alleles, bitter taste receptor T2R38, Toll-
like receptors (TLRs), and the cystic fibrosis transmem-
brane regulator (CFTR) locus.
2,4,14
Although these studies
are promising, common limitations among them include
small sample sizes, unclear causality, and difficulty repli-
cating results.
4
In an effort to take a broader look at the familiality of
CRS, this population-based study assessed shared risk of
CRSwNP and CRSsNP based on observed familial cluster-
ing compared to that expected in the population over a 16-
year period. Although there are multiple studies that have
analyzed the association between various single nucleotide
polymorphisms (SNPs) and CRS in an effort to identify ge-
netic links to the condition,
9,15–23
the clinical importance
International Forum of Allergy & Rhinology, Vol. 5, No. 4, April 2015
73