Preauricular skin tags: Is renal ultrasound warranted?
Urquhart DS, Nicholl RM
Neonatal Unit, Northwick Park Hospital, Harrow, Middlesex. UK
EDITOR:
Following a clinical discussion, we attempted to appraise the evidence as
to whether renal ultrasound is warranted in babies with isolated
preauricular skin tags. Using the search term ‘preauricular tag’, we
searched the OVID, EMBASE and MEDLINE databases and the metasearch engine
SUMSEARCH (http://sumsearch.uthscsa.edu). This search strategy found 2
papers1,2, one of which was the prospective study by Kohelet and Arbel1.
In this study of well newborn babies with preauricular tags, 6 out of
70 babies with skin tags had associated renal tract anomalies. The
importance of these findings is debatable, as 5 babies had grade 2-3
hydronephrosis which may be expected to resolve spontaneously, and indeed
none had required surgical intervention at follow-up. The remaining child
had a horseshoe kidney, which is a normal variant.
We believe the conclusions drawn by this paper may not be valid. The
control group used in this study were infants under investigation for
‘vomiting’ or ‘cyanosis’, and as such cannot be considered to be healthy
controls. The study group i.e. those with skin tags underwent ultrasound
on either day 3 or day 4 of life, but the median age for the control group
is not stated, but is likely to have been up to some weeks later when
hydronephrosis may have improved or resolved as part of the natural
history of this condition. Indeed the authors state that all of the
control group were studied after Day 5.
The issue of blinding was not explicitly stated, but one wonders whether
the sonographer would see the skin tag or have a parent present who would
remark on this, introducing a potential source of observer bias.
Using the population data supplied in the paper, we constructed a 2x2
table to interrogate the data provided (Table 1).
TABLE 1
TEST TARGET DISORDER
Presence of skin tag Renal tract ultrasound anomaly
Present Absent
Positive 6 64
Negative 31 17531
This gives a sensitivity of 16% (95% Confidence Intervals [CI] 4%,
28%), and a specificity of 100%. Prevalence is very low at around 0.002%.
Positive Predictive value is 9% (95% CI 2%, 15%), and Negative Predictive
Value 100%
We therefore feel that the authors recommendations (renal ultrasound
for all babies with preauricular skin tags) are not justified. Given the
above interpretation of the data, the test has very poor sensitivity (16%)
and a poor positive predictive value (9%). Following our critical
appraisal of this topic, local policy is not to routinely perform renal
ultrasound in such cases.
REFERENCES:
1. Kohelet D, Arbel E.A prospective search for urinary tract abnormalities
in infants with isolated preauricular tags. Pediatrics 2000; 105(5): E61.
2. Wang RY. Syndromic ear anomalies and renal ultrasounds. Pediatrics
2001; 108(2): E32.
Yours Sincerely
Dr Donald S Urquhart - Specialist Registrar in Paediatrics
Dr Richard M Nicholl - Consultant Neonatologist
eLetters: ![[Read eLetters]](/icons/misc/sectionDOWN.gif){kind=icon}
