Authors' objectives

To evaluate methods used in the emergency department (ED) to assess patients with sore throat, earache and sinusitis.

Searching

Sore throat: MEDLINE was searched from January 1966 to July 1997 for studies published in the English language. The keywords used were 'pharyngitis', 'strep', 'sore throat', 'streptococcal pharyngitis', 'group A streptococcus' and 'Streptococcus Pyogenes'. The bibliographies of identified studies and several textbook chapters were checked.

Ear pain: MEDLINE was searched from January 1975 to July 1998 using the keywords 'ear pain', 'otitis media', 'diagnosis', 'tympanometry' and 'otoscopy'. The bibliographies of identified studies and several textbook chapters were checked. Case reports and letters were excluded.

Sinusitis: MEDLINE was searched from January 1985 to July 1998 using the keywords 'sinusitis', 'diagnosis', 'clinical guidelines' and 'CT scan-sinus'. The bibliographies of identified studies were checked. Case reports and letters were excluded.

Study selection

Assessment of study quality

The quality of the evidence was classified according to study design. Level I evidence was from well-designed RCTs or well-designed meta-analyses of RCTs. Level II evidence was from well-designed clinical studies (non-randomised controlled trials, case-control studies, cohort studies, time series, or uncontrolled prospective studies). Level III evidence was from descriptive studies, case series or case reports, studies using non-randomised historical controls, or reports of expert opinion or expert committees. Articles were reviewed against the classification of evidence definitions. The authors did not state how many reviewers performed the quality assessment.

Data extraction

Sore throat: two reviewers extracted the data. Data were extracted on: the use of culture and 'rapid strep' methods, degree of blinding, study population location, prevalence of GABHS by the chosen 'gold' standard for the study, and results. Sensitivity, specificity, and PPVs and NPVs were evaluated or calculated from the data presented.

Ear pain and sinusitis: the authors did not state how the data were extracted for the review, or how many reviewers performed the data extraction. Tabulated data included author, year of publication, population, location, blinding and results.

Methods of synthesis

Results of the review

Ninety-two studies evaluated diagnostic tests for sore throat (38,792 patients). Sixteen studies evaluated diagnostic tests for ear pain (the number of patients was unclear). Twenty-nine studies evaluated diagnostic tests for sinusitis (at least 1,786 patients).

Sore throat (85 level II studies and 7 level III studies).

The sensitivities and specificities of available tests varied widely.

Latex agglutination-based rapid streptococcal tests: there were 19 studies with 8,138 patients. The mean values were sensitivity 79.47% (SD=10.80), specificity 95.36% (SD=5.37), PPV 88.65% (SD=13.4) and NPV 90.43% (SD=7.02). The median values were sensitivity 80.10%, specificity 96.50%, PPV 96.00% and NPV 91.75%. Sensitivity ranged from 58% (specificity 92%) to 96.7% (specificity 97.9%), specificity from 76% (sensitivity 72%) to 100% (sensitivity 80.6%), the PPV from 57 to 100%, and the NPV from 74 to 99%.

Enzyme-linked immunosorbent assay-based rapid strep tests: there were 24 studies with 13,169 patients. The mean values were sensitivity 78.79% (SD=8.88), specificity 94.99% (SD=7.37), PPV 87.35% (SD=11.18) and NPV 92.58% (SD=5.24). The median values were sensitivity 78.00%, specificity 97.05%, PPV 90.00% and NPV 93.00%. Sensitivity ranged from 60.8% (specificity 98.3%) to 95.6% (specificity 96.7%), specificity from 63% (sensitivity 79%) to 100% (sensitivity 79.4%), the PPV from 3.6 to 100%, and the NPV from 7.1 to 98.3%.

Optical immununoassay: there were 10 studies with 7,560 patients). The mean values were sensitivity 89.07% (SD=7.35), specificity 92.57% (SD=9.51), PPV 79.17% (SD=22.09) and NPV 95.24% (SD=1.90). The median values were sensitivity 91.40%, specificity 95.40%, PPV 87.00% and NPV 95.00%. Sensitivity ranged from 77% (specificity 62% and 97%) to 98.9% (specificity 98.4%), specificity from 62% (sensitivity 77%) to 98.8% (sensitivity 94.8%), the PPV from 22 to 91.5%, and the NPV from 93 to 99%.

Ear pain (8 level II studies and 29 level III studies).

Otoscopy: there were 5 studies involving more than 700 ears. Most of the studies did not calculate the sensitivity and specificity. Other problems included the use of physical examination as the 'gold' standard, and the possibility of selection bias: all but one study were conducted in children with severe disease.

Tympanometry: there were 8 studies involving at least 1,859 ears. The sensitivity ranged from 71% (specificity 38%) to 95% (specificity 57%), while specificity ranged from 37% (sensitivity 78%) to 93% (sensitivity 79%). The studies were predominantly conducted in children with a strong clinical indication for myringotomy, and consequently high disease prevalence. Several studies suggested that general anaesthesia may influence the results.

Acoustic reflectometry: there were 3 studies involving at least 1,156 ears. The sensitivity ranged from 79% (specificity 86%) to 90% (specificity 86%), while the specificity ranged from 79% (sensitivity 86%) to 86% (sensitivity 79 and 90%).

Sinusitis (1 level I study, 5 level II studies and 23 level 3 studies).

The criteria for the diagnosis of sinusitis varied greatly and there was no 'gold' standard.

Physical examination: there were 4 studies with 1,315 patients. Most of the studies did calculate the sensitivity and specificity. In the one study reporting sensitivity and specificity, the sensitivity ranged from 18% for maxillary toothache to 45% for sinus tenderness, whilst specificity ranged from 34% for cough or sneeze to 93% for maxillary toothache.

Radiography: there were 5 studies with 278 patients. The sensitivity ranged from 48 to 99%, while the specificity ranged from 75 to 100%.

Ultrasonography: there were 3 studies with at least 100 patients. Two studies reported sensitivity rates from 44 to 83% and specificity rates from 55 to 88%. One study using X-ray as the 'gold' standard found that sensitivity and specificity varied with the criteria used to classify X-ray as normal.

Limited CT: there were 2 studies with 93 patients. One study reported a sensitivity of 93% and a specificity of 89%, while the other study reported 88% overall concordance.

CT: the number of studies was unclear. One study found much variation in CT findings in patients diagnosed clinically as having acute sinusitis.

Cost information

Pharyngitis: 3 studies evaluated the cost-benefit ratios for treatment strategies for GABHS, and all supported the use of rapid testing or pharyngeal Gram's stain in patients with a high pre-test probability. The evidence on which this support was based was not critically appraised.

Sinusitis: 2 reviews found the cost of limited CT and plain radiography to be equivalent.

Authors' conclusions

Acute pharyngitis: patients with appropriate epidemiologic, historic, and physical examination characteristics should have rapid screens performed. Those patients with a positive test result may be treated for GABHS without confirmatory culture. Patients with negative results should receive confirmatory culture. Antibiotics may be started or withheld in these patients while waiting for the results of the culture. Follow-up in the ED is often sporadic and contacting patients is difficult, thus each practitioner must base treatment on a case-by-case basis.

Ear pain: the role of tympanometry and acoustic reflectometry in the ED is limited. There have been no studies in uncomplicated patient populations using myringotomy. Findings suggest these methods are no more reliable than history, physical examination, or pneumatic otoscopy.

Sinusitis: a positive antral aspiration has better diagnostic accuracy than a positive clinical examination, ultrasonography, or radiography. It appears that ultrasound lacks the sensitivity and specificity to be diagnostic. No imaging studies are recommended for the routine diagnosis of sinusitis presenting to the ED.

CRD commentary

The review contained no clearly stated objective or research question and the inclusion criteria were poorly defined. No definitions were given for classifying patients as having sore throat, ear pain, or sinusitis. By limiting the search to one database and bibliographies in textbooks, other relevant studies might have been omitted. There was no attempt to identify unpublished material, thus raising the possibility of publication bias, and the methods used to select the studies were not described. The quality assessment was limited to classifying the level of evidence according to study design and blinding, with some comments on the methodological problems of primary studies. No details of the methods used to classify the studies or extract data were given.

Some relevant information, including results, was presented clearly in tabular format. Diagnostic thresholds were not reported and were not considered when pooling the narrative summaries. The diagnostic accuracy measures varied considerably among the studies and this heterogeneity was not explored. The methods of pooling used were crude, and did not take the observed heterogeneity into account. In general, the pooling of the studies included in this review was unlikely to be appropriate. The evidence presented does not appear to support the strength of the authors' conclusions.

Implications of the review for practice and research

Practice: The authors stated the following.

Sore throat: 'Patients with appropriate epidemiologic, historic, and physical examination characteristics of GABHS should have rapid screens performed. Those patients with a positive test result may be treated for GABHS without confirmatory culture. Patients with negative results should receive confirmatory culture. Antibiotics may be started or withheld in these patients while waiting for the results of the culture. In the emergency department population, however, follow-up of culture is often sporadic and contacting patients difficult, so that each practitioner must base his or her treatment decision on a case-by-case basis.'

Ear pain: In the diagnosis of ear pain, the role of tympanometry and acoustic reflectometry in the ED is limited. Findings suggest these methods are no more reliable than history, physical examination, or pneumatic otoscopy.

Sinusitis: No imaging studies are recommended for the routine diagnosis of sinusitis presenting to the ED.

Research: The authors stated that future research into GABHS should include the search for testing methods with sensitivity and specificity equal to or better than culture, and explore methods of streptococcal testing at home that are cost-effective and safe for the patients.

Bibliographic details

Stewart M H, Siff J E, Cydulka R K. Evaluation of the patient with sore throat, earache, and sinusitis: an evidence based approach. Emergency Medicine Clinics of North America 1999; 17(1): 153-187. [PubMed: 10101345]

Indexing Status

Subject indexing assigned by NLM

MeSH

Adult; Child; Diagnostic Tests, Routine; Earache /diagnosis /etiology /therapy; Evidence-Based Medicine; Humans; Otitis Media /diagnosis /etiology /therapy; Pharyngitis /diagnosis /etiology /therapy; Sinusitis /diagnosis /etiology /therapy

AccessionNumber

11999003565

Database entry date

31/07/2005

Record Status

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.