Acute sinusitis is common enough, usually diagnosed by clinical examination, and
treated with a course of antibiotics. The trouble is that many of the symptoms of
acute sinusitis overlap with other common nasal complaints, so misdiagnosis and
over-treatment with antibiotics is thought to be common. Diagnostic interventions of
more invasive or high technology nature are available, like X-ray, CT scanning, or
sinus puncture. So when two systematic reviews of diagnostic tests for the same
condition appear in the same journal, almost adjacent to one another [1,2], clarity
may be expected. Unfortunately, it's not quite like that, and another two brains
problem rears its head. The clinical bottom line is that we don't have adequate
The review from Boston  used a MEDLINE search for studies that had to compare the
ability of at least two tests to diagnose acute sinusitis. This review included some
studies despite an inadequate description of eligibility criteria, for sensible
reasons. The sensitivity and specificity of one test had to be compared with another,
reference, test. The main analytical tool was the summary receiver operating
characteristics curve, when sensitivity is plotted against 1-specificity.
The review from Helsinki, Oulu and Copenhagen  also searched MEDLINE, and a
Finnish database, and hand searched four journals and contacted experts in the field.
Studies with the reference standard of sinus puncture or CT scanning were accepted.
Validity criteria from a Cochrane working group were used for validity, and excluded
some studies that failed some of these criteria. Summary receiver operating
characteristics were also used as an analytical tool, but likelihood ratios were also
Review 1 included 13 studies, six with patients from a hospital or emergency room
setting, three from general practice, one from both, and three did not say how
patients were recruited. Review 2 included nine studies, eight of which were included
in review 1 (though there were differences between the reviews in, for instance, the
year of publication of studies with otherwise identical references, as well as
Review 1 describes and criticises the studies and the methods, and gives overall
curves for several different comparisons. Review 2 also criticises the included
trials, and gives prevalence, sensitivity, specificity and likelihood ratios for each
method compared with sinus puncture (Table). Using the weighted prevalence figure and
the likelihood ratios, radiography, ultrasound and clinical examination fail to
generate post-test probabilities of more than 80% or less than 20%.
Table: Diagnostic methods for acute sinusitis, compared with sinus puncture
Comparison (with sinus puncture)
Number of studies
Number of patients
Weighted prevalence (%, range)
Positive likelihood ratio
Negative likelihood ratio
These are cracking reviews, of high quality, and well written.
marginally prefers review 2 because it allows the calculation of post-test
probabilities, and shows clearly how radiography, ultrasound and clinical
examination fail to provide sufficiently high post-test probabilities to be truly
useful: no better than 80% or 20% is probably not good enough. Review 1, though,
has more details about the individual studies and the components of the
diagnostic tests, so that is more useful in understanding what is going on.
Both have excellent discussions of the problems of diagnostic test reporting,
and the particular problems about diagnosis of acute sinus infection, including
those of sinus puncture, regarded as the gold standard. Taken together, these
illuminate the problems of diagnostic test methods and quality, and their review.
They both conclude, rightly, that it's back to the drawing board. Review 2
additionally adds the visionary concept of combining the two aspects of clinical
management, diagnosis and treatment, as a unit.
Diagnostic testing evidence is in short supply. The importance of this
particular topic lies in the opportunities for research across the primary and
secondary care boundaries, and combining clinical examination with high-tech
- EA Engels, N Terrin, M Barza, J Lau. Meta-analysis of diagnostic tests for
acute sinusitis. Journal of Clinical Epidemiology 2000 53: 852-862.
- H Varonen, M Mäkelä, S Savolainen, E Läärä, J
Hilden. Comparison of ultrasound, radiography, and clinical examination in the
diagnosis of acute maxillary sinusitis: a systematic review. Journal of
Clinical Epidemiology 2000 53: 940-948.
story in this issue