• 
• 

Three Part Question

In an [adult patient with pleuritic chest pain], does the presence of [chest wall tenderness] help to rule out [Pulmonary Embolism]?

Clinical Scenario

A 30 year old man presents to the ED with a 3 days history of right sided chest pain that increases in intensity with breathing, lying on the right side and application of local pressure. His BP is 130/70mm Hg, heart rate 90 beats per minute, respiratory rate 23/min, and temperature 37.3°. He denies any history of trauma. Pulmonary embolism is one of the differential diagnoses, but you question whether the presence of the chest wall tenderness is enough to rule out pulmonary embolism before carrying out further tests.

Search Strategy

OVID Medline 1950 - August Week 4 2010
exp pulmonary embolism/ OR exp thromboembolism/ OR pulmonary infarct$.mp. / OR pulmonary embol$.mp./ OR PE.mp
AND
Tender$.mp/ OR tenderness.mp/OR palpat$.mp/ OR exp palpation/ OR clinical feature.mp/ OR exp chest pain
Limit Humans and English
Limit Diagnosis (Sensitivity)

Google Scholar Search- “Pulmonary embolism” “chest wall tenderness”
Cochrane database- “Pulmonary embolism” “chest wall tenderness”

Search Outcome

Medline 304 papers- 3 relevant papers (2 study papers and 1 case report) found
Google Scholar- 1 extra article (study paper) found
References from selected papers- 1 relevant article found (1 study paper selected from Medline was a subgroup analysis of this article)
Cochrane- no relevant articles found
Overall, 3 relevant study papers have been critically appraised.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Le Gal et al 2005 Switzerland	965 consecutive outpatients who presented to EDs of three Hospitals between Oct 2000 and June 2002 Exclusion criteria- allergy to contrast iodine, ongoing treatment with anticoagulants, creatinine clearance<30ml/min, pregnancy, life expectancy<3 months Average age- 61±19 years Female 562 (58%), Male 403 (52%)	2b- Subgroup analysis from a prospective cohort study Signs and symptoms recorded in standardized form and presence of chest pain reproducible on palpation recorded	Prevalence of PE in patients with chest wall tenderness vs no chest wall tenderness	Overall prevalence of PE 222/965 (23%)	No standardised way of testing chest wall tenderness, relying entirely on the clinician’s discretion and assessment, and patients’ subjective response. No blinding between reference standard results and clinical details. Misleading calculations This calculation is misleading as the authors have misinterpreted the definition of a positive and negative diagnostic test. The positive diagnostic test for PE should be the absence of chest wall tenderness and the negative diagnostic test should be the presence of chest wall tenderness. Hence the correction was made to the following calculations as shown below; For absence of chest wall tenderness in PE- Sensitivity 0.828829 (0.77271 to 0.875931), 82.88% (77.27% to 87.59%) Specificity 0.205922 (0.177387 to 0.236812), 20.59% (17.74% to 23.68%) PPV 0.237726 (0.208137 to 0.269329), 23.77% (20.81% to 26.93%), NPV 0.801047 (0.737306 to 0.855195), 80.1% (73.73% to 85.52% Likelihood Ratio LR (positive test) = 1.043762 (0.966944 to 1.114293) LR (negative test) = 0.831243 (0.599654 to 1.137801)
				Prevalence of PE in patients with chest wall tenderness 38/191 (19.9%) vs no chest wall tenderness 184/774 (23.8%) P=0.25
				For presence of chest wall tenderness in PE: Sensitivity 17.1% (CI 12.4-22.7%) Specificity 79.4% (CI 76.3-82.3%) Positive likelihood ratios 0.83 (CI 0.60 to 1.14) Negative likelihood ratios 1.04 (CI 0.97- 1.11)
Fennerty et al 1989 UK	250 consecutive patients (mean age 53.6 range 18-95 who underwent V/Q scans between 1983 and 1987)	2b- Retrospective cohort study of case notes	Presenting clinical features of patients with suspected PE in high probability and non-high probability V/Q scan groups	96 patients with high probability V/Q scans. Local chest wall tenderness 9/96	Retrospective analysis, hence not all positive and negative findings may have been documented. Small sample size. The gold standard diagnostic test (pulmonary angiogram) not used hence the true incidence of PE not known in the study population. The study assumes that those with high probability V/Q scans were positive for PE, but we do not know the false-positive rates. Unclear where the cohort group was recruited from 250 consecutive patients with suspected PE who attended ED? Outpatients? Inpatients?
				86 with low probability V/Q. Local chest wall tenderness 1/86 (1.16%)
				68 with indeterminate V/Q. Local chest wall tenderness 0/68
				For absence of chest wall tenderness in PE: Sensitivity 91% (CI 83- 96%) Specificity 0.65% (CI 0.02- 3.56%) PPV 36.3% (CI 30.2- 42.7%) NPV 10% (CI 0.25- 44.5%) +LR 0.91 (CI 0.84-0.96) -LR 14.4 (CI 2.42-87.4)
Hull et al 1987 USA	173 consecutive patients( mean age 47, range 17-88) presenting to EDs of two hospitals with pleuritic chest pain underwent hx, ex, ECG, CXR, impedance plethysmography, V/Q scans and pulmonary angiogram if V/Q abnormal.	2b- Prospective cohort study with 3 months follow up.	Positive pulmonary angiogram for PE	Overall prevalence of PE 36/173 (21%)	Small sample size. Uncertain from the study if clinicians performing primary hx and examination were blinded to radiology reports, although radiologists interpreting films were blinded to clinical details. Only hospital outpatients took part. A large proportion of patients with chest wall tenderness may not have been referred.
			PE at autopsy	32/173 (18%) had chest wall tenderness
				For absence of chest wall tenderness in PE: Sensitivity 89% (CI 74-97%) Specificity 20%(CI 14-28) PPV 22.7% (CI 16.1- 30.5%) NPV 87.5% (CI 71-96.5%) +LR 1.12 (CI 0.93- 1.27) -LR 0.54 (CI 0.21- 1.33)

Comment(s)

The diagnostic challenge of pulmonary embolism derives from interpreting the pre-test probability of the disease based on a set of non-specific symptoms and signs. The triad of haemoptysis, dyspnoea and pleuritic chest pain as classic presenting characteristics of pulmonary embolism are well documented. By contrast, in patients with chest wall tenderness, clinicians tend to concentrate on chest wall disorders of muscular, skeletal or neurological origin. The pathogenesis theory for this phenomenon could be deduced as follows: pleuritic chest pain is believed to arise from the irritation of the parietal pleura caused by the pulmonary infarct causing inflammation. The intercostal nerves supplying the parietal pleural also supply the intercostal skin in a dermatomal fashion. Therefore the palpation of the chest wall could in theory 1) apply pressure on the already inflamed and stretched parietal pleura, and/ or 2) elicit pain by pressing on the hyperanaesthetic area of the dermatomal skin supplied by the cutaneous branches of the intercostal nerves. There were only three relevant papers in the medical literature. These studies represented a wide variety of methodology and sample characteristics, ranging from prospectively studied cohorts of EDs’ and outpatients to retrospectively analysed patients with suspected PE. The lack of a validated objective method of detecting chest wall tenderness represents selection bias. Eliciting positive and negative clinical findings were entirely dependent on the clinicians’ discretion and their individual examination methods and also their willingness to document them. This affects the validity and reliability of the data. Likewise, the findings from these data cannot be applied to your local patients with confidence. According to the three papers reviewed, the overall prevalence of PE ranged from 20.8 to 38.4%. The sensitivity of the absence of chest wall tenderness (positive test) in PE varied from 0.83 to 0.91, the specificity of chest wall tenderness (negative test) in those not having PE varied from 0.0065 to 0.21, the positive predictive values 0.23 to 0.36 and negative predictive values 0.1 to 0.88. The absence of chest wall tenderness does not make the diagnosis of PE more likely (composite positive likelihood ratios 0.83-1.04), and the composite negative likelihood ratios cross 1 (range 0.54 to 14.44). By applying the rule of SnNout, the presence of chest wall tenderness (negative test) appears to suggest that PE is less likely, but does not effectively rule it out. The specificity is too low for the test to be used as a screening tool. Overall, methodology flaws preclude any meaningful comparison and conclusion of data.

Clinical Bottom Line

The presence of chest wall tenderness appears to reduce the clinical likelihood of PE. However, no definitive conclusion can be drawn about its diagnostic accuracy, due to heterogeneity of data and lack of methodology rigour. A further large study, using robust selection criteria with validated diagnostic algorithm is needed.

References

1. Le Gal G, Testuz A, Righini M, Bounameaux, Perrier A Reproduction of chest pain by palpation: diagnostic accuracy in suspected pulmonary embolism BMJ 2005; 330: 452-3
2. Fennerty AG, Shetty HGM, Paton D, Roberts G, Routledge PA and Campbell IA Clinical presentation and investigation of patients proceeding to isotope lung scanning for suspected pulmonary embolism Postgrad Med J 1990; 66: 285-289
3. Hull RD, Raskob GE, Carter CJ, Coates G, Gill GJ, Sackett DL, et al Pulmonary Embolism in Outpatients with Pleuritic Chest Pain Arch Intern Med 1988; 148: 838-844