mal ECG, which was done in this patient and noted in our report. Fiona E. Gallahue, MD Department of Internal Medicine Division of Emergency Services Harborview Medical Center Seattle, Washington doi:10.1016/j.jemermed.2009.01.002 e D-DIMER MEASUREMENT AND LABORATORY FEEDBACK e To the Editor: We read with interest the recent article by Breen et al., a cautionary tale showing that placing overt reliance on D- dimer testing results in potentially missing patients with suspected venous thromboembolism (VTE), because false “negative” cases might be occasionally encountered (1). We broadly agree with their conclusions. Their patient case study is paradigmatic insofar as many factors might have compromised the clinical usefulness of D-dimer measure- ment in this situation. VTE is a common condition with significant morbidity and mortality if not diagnosed and treated in a timely fashion. D-dimers are fibrinolytic prod- ucts produced at the site of thrombosis, and the presence or absence of D-dimer in patient blood might assist in con- firming or refuting a diagnosis of suspected thrombosis. In particular, a “negative” D-dimer by enzyme-linked immu- nosorbent assay (ELISA) is generally considered the bio- chemical gold standard for ruling out acute episodes of VTE (2– 4). However, both the sensitivity and specificity of D-dimer measurements depend on the prevalence of disease in the study population, and this explains the variety of results in different reports. Although rare, “false negative” D-dimer results can be encountered in patients with VTE, mostly using methods based on turbidimetry or latex- enhanced nephelometry. Non-diagnostic values are ob- served because D-dimer is measured too early before or too late after the development of thrombosis. Falsely negative results can also occur after initiation of anticoagulant ther- apy, because anticoagulants limit clot propagation and di- minish D-dimer production. Also, thrombus location and size are important determinants of D-dimer levels, inas- much as lower sensitivities have been reported in diagnos- ing calf deep vein thrombosis (DVT) vs. above-the-knee DVT and peripheral vs. central pulmonary embolism (PE) (2,4). Most of these confounding factors may coexist in specific clinical contexts, such as the one reported by Breen et al. (1). Pre-analytical variability arising from misidenti- fication, troublesome phlebotomy, and inadequate or poor specimen collection or handling are additional sources of inaccuracy in laboratory testing (5). It is also worth men- tioning that no two D-dimer assays are identical. The value obtained by Breen et al.’s case on admission (495 ng/mL) using an unspecified ELISA was alarmingly close to the diagnostic cut-off, which was reported to be 500 ng/mL. Assay imprecision (expressed as coefficient of variation, CV%) of the most widely used ELISAs at this value ranges from 4.9% to 9.9% (6,7). Due to this (unavoidable) impre- cision, a test result approaching the diagnostic threshold always must be considered with caution. Given the 5–10% imprecision of the assay, most laboratorists would know that a D-dimer value lying very close to the diagnostic cut-off can be either “positive” or “negative” in a dichoto- mous division, although most physicians may not know this, lacking specific knowledge on the analytical assay performance. The lack of feedback between clinicians and the clinical laboratory is an old, well-recognized dogma, which might have a deleterious impact on patient outcomes, the largest barrier being the increasing shortage of labora- tory specialists able to provide analytical and, especially, clinical interpretations (8). Whereas analytic interpretation involves examining the raw data and elaborating a conclu- sion about the reliability of test results, clinical interpreta- tion involves describing what the result means for the patient, either in general or based on specific knowledge of that patient’s situation. Probably, interpretative comments added to clinical reports would help to assist a medical dialogue on individual findings, as long as this process reflects best practice and reliable guidelines (9). Therefore, although we essentially agree with critics who claim that laboratories and researchers should look at continuous test results in a manner that more usefully describes the data, as laboratorists, we counter that cli- nicians should better utilize laboratorists as skilled con- sultants and not mere “test executors.” Laboratory pro- fessionals will not decline their responsibilities, but the relationship between clinical laboratory specialists and physicians should be viewed as a tight partnership in which both parties are communicating efficiently in the interest of patients and health care providers (10). In the end, the case report of Breen et al. may reflect a combi- nation of a small PE clot on initial presentation, timing and specimen issues, but also a lack of communication with the laboratory at the time of initial testing, thereby preventing positive professional consultation that would likely have agreed with their view that a result of 495 ng/mL could be consistent with a “positive” event. Giuseppe Lippi, MD Sezione di Chimica Clinica Dipartimento di Patologia Università di Verona Verona, Italy 82 Letters to the Editor