Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

Highlighted lines and questions below provide links to the pertinent description of criteria in The EBM User's Guide, now available at the Canadian Centres for Health Evidence

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

Clinical Validity of a Negative Computed Tomography Scan in Patients With Suspected Pulmonary Embolism: A Systematic Review.

Quiroz R, Kucher N, Zou KH, Kipfmueller F, Costello P, Goldhaber SZ, Schoepf UJ.

JAMA. 2005 Apr 27;293(16):2012-2017 [abstract]

Reviewed by Sheila Hanson MD MS, Medical College of Wisonsin, Milwaukee, WI

Review posted October 5, 2005

---

I. Are the results of the study valid?

A. Primary questions:

1. Did the overview address a focused clinical question?

Yes. This meta-analysis reviewed clinical trials that used chest CT to diagnose acute pulmonary embolus (PE) and determined the frequency of subsequent venous thromboembolic events (VTE) during the clinical follow up period after an initial negative or inconclusive chest CT.

2. Were the criteria used to select articles for inclusion appropriate?

Yes. Inclusion criteria were: clinical follow-up of office visits, phone interviews or questionnaires, 2) minimum follow-up of 3 months, 3) study population of more than 30 patients and 4) chest CT performed on all patients.

Studies were excluded if D-dimer testing was used as an initial triage tool, if follow-up was inappropriate or absent, if the quality score was less than 5 (discussed later), or if the article was a review or editorial. Patients were also excluded if they were placed on anticoagulants for any reason other than VTE.

B. Secondary questions:

3. Is it unlikely that important, relevant studies were missed?

No. The authors made a significant effort to capture relevant studies. Databases including PubMed, MEDLINE, EMBASE, CRISP, meta-register of Controlled Trials, and Cochrane were searched for English language articles published Jan 1990 to May 2004. In addition, the search was strengthened by hand searching of relevant journals, correspondence with investigators and experts in the field, and cross-referencing using the Science Citation Index. Search terms were: negative predictive value, pulmonary embolism, deep vein thrombosis, venous thromboembolism, computed tomography, chest CT and spiral CT.

The search criteria were broad enough that most clinically relevant studies should have been captured. Personal contacts should have revealed unpublished or other published studies that had been missed, lessening potential publication bias. Though only English language articles were included, the 15 selected articles originated from 7 countries.

Meta-analysis studies are at risk of publication bias, given that negative studies are less likely to be published, have a longer lag to publication, and less likely to be cited if they are published. These authors attempted to assess publication bias by evaluating funnel plot asymmetry. This was non-significant (p>0.2) by both Egger and Begg techniques (1,2).

4. Was the validity of the included studies appraised?

Validity of the component studies was assessed by assigning a quality score, with one point for each of the following:

1. Published in peer-reviewed journal
2. Prospective design
3. Imaging technique explicitly described
4. Inclusion and exclusion criteria accurately described
5. Patient demographics collected
6. Follow-up included
7. Recurrences and mortality included

Of the 22 potential studies only 15 were included in the analysis. Seven studies were excluded because they did not meet the selection criteria or because the quality score was less than 5. It would have been helpful to know which criteria rejected studies in order to fully assess validity. Meta-analysis cannot compensate for flaws in the individual studies. Only 8 of the 15 included studies were prospective. Additionally, only symptomatic patients were included as patients were not screened for PE during the follow up period. Undetected VTE may have contributed to mortality and morbidity either as an unrecognized event in the follow-up group (n=3533) or in the significant number of patients lost to follow up (n=199) or excluded for anticoagulation (n=309). The actual incidence of PE could be higher than reported.

5. Were assessments of studies reproducible?

It is unclear from the article how reproducible was the data extraction. Two reviewers abstracted the data independently with a third investigator to arbitrate any discrepancies. The frequency of discordance was not reported.

6. Were the results similar from study to study?

All of the studies showed that the probability of a PE being present is decreased after a negative chest CT, though the magnitude of this decrease varied between studies. The negative likelihood ratio of posttest VTE after negative chest CT ranged from 0.03 to 0.22 for the individual studies, showing moderate to large reduction in pretest probability of VTE. The pre-test probability was set as the prevalence of PE in each study and ranged 15-38%.

Too much heterogeneity in the studies included in a meta-analysis may make pooled estimates less meaningful. The heterogeneity was assessed by Q-statistic with p<.001 showing significant heterogeneity between studies. This is not surprising as the individual studies involved used a variety of study designs and imaging protocols, including 7 retrospective reviews of Chest CT for PE, 2 studies where chest CT's were triggered by nondiagnostic ventilation/perfusion scan, 6 studies were prospective after negative chest CT with alternate imaging recorded. Meta-regression was performed to assess for differences in VTE's between studies using additional imaging prior to the CT and studies that used CT only, as well as for differences between studies using multidetector-row CT and those that used single-slice CT, and studies with 3 month follow up and those with longer follow up. No significant difference was detected based on these differences between studies. An influence analysis to evaluate the weight of the individual studies was performed, which showed no significant dominance by any one study.

II. What are the results?

1. What are the overall results of the review?

The pooled results of the 15 studies found a negative likelihood ratio (NLR) of 0.07, 95% CI (.05-.11), with a negative predictive value (NPV) of 99.1%, 95% CI (98.7%-99.55) for developing a clinically significant VTE after a negative chest CT. NLR indicates how much the odds of the disease decrease when a test is negative and is calculated by (1-Sensitivity)/Specificity. NLR of < 0.1 is very low, often reflecting conclusive changes from the pre-test probability. NPV is the probability that the patient will not have the disease with a negative test, that is the number of patients with the disease and a negative test result divided by the number of subjects with and without the disease with a negative test result.

Overall, out of 3500 patients, 36 PE's and 6 DVT's developed despite a negative CT. Mortality from these events was15 out of 36. The NPV for mortality from PE after a negative chest CT was 99.4% (95% CI, 98.7-99.9%).

2. How precise were the results?

The confidence intervals were narrow as reported above.

III. Will the results help me in caring for my patients?

1. Can the results be applied to my patient care?

Perhaps. Only adults were included in the studies, and the prevalence, etiology, and recurrence rate of VTE in children is significantly different than in adults. Other methods to diagnose PE such as V/Q scans and pulmonary angiograms are often technically difficult in the critically ill pediatric population. This study lends support to a commonly used method of diagnosing PE in children. The prevalence of PE in the populations studied of 15-38% is much higher than that in pediatrics, which ranges 1-79/100,000 hospitalizations in unselected hospitalized children. (3,4). So unless an unusually high-risk prevalence exists, the NPV should be similar to or less than that reported. However, this highlights the merit of using the likelihood ratio, since it is independent of prevalence.

2. Were all clinically important outcomes considered?

The study assessed clinically significant PE or DVT as detected during the follow-up period and mortality from VTE during this period. It is recognized that chest CT is limited in detection of peripheral pulmonary embolism compared to pulmonary angiography [5]. Using a clinical outcome as the standard in this meta-analysis is reasonable however, as even if peripheral PE's were missed it did not affect patient outcome, and this strategy is compatible with general clinical practice. It is possible, however, that small PE's that were missed resulted in a lost opportunity to look for DVT's, which could affect the patient's outcome many months later (beyond the period of follow-up.

3. Are the benefits worth the harms and costs?

The ease and availability of obtaining a chest CT in a critically ill child makes it ideal as a first line diagnostic image for PE. This meta-analysis reports a NPV of 99% for VTE after a negative chest CT, which is comparable to the gold standard of pulmonary angiography (NPV 98-100%). With-holding anticoagulation after a negative chest CT appears to be safe. The risk of additional bleeding, monitoring, and radiation would be avoided.

Only in a very high-risk patient population (e.g., critically ill thrombophilic patient with central venous access) where the prevalence is higher than that of the included studies (15-38%) would additional imaging be warranted.

References:

1. Egger M, Davey G, Schneider M, et al. Bias in meta-analysis detected by simple, graphical test. BMJ. 1997; 315:629-634. [abstract]
2. Begg CB, Berlin JA. Publication bias and dissemination of clinical research. J Natl Cancer Inst. 1989; 81:107-115.[abstract]
3. Stein PD; Kayali F; Olson RE. Incidence of venous thromboembolism in infants and children: data from the National Hospital Discharge Survey. J Pediatr 2004 Oct;145(4):563-5.[abstract]
4. Bernstein D; Coupey S; Schonberg SK. Pulmonary embolism in adolescents. Am J Dis Child 1986 Jul;140(7):667-71.[abstract]
5. Perrier A, Howarth N, Didier D et al. Performance of helical computed tomography in unselected outpatients with suspected pulmonary embolism. Ann Intern Med. 2001; 135:88-97.[abstract]

Comments

Back to the EB Journal Club Index