Diagnostic errors in the new millennium: a follow-up autopsy study.

A systematic review of the second half of the last century suggested that diagnostic errors have decreased over time. Our previous study covering the years 1972-1992 was then the only time series showing a significant reduction of diagnostic errors from a single institution. We report here the results of a follow-up study a decade later. We analyzed discrepancies between clinical and autoptic diagnoses in 100 randomly selected medical patients who died in the wards and in the medical intensive care unit at a tertiary-care teaching hospital in Switzerland in the year 2002. Autopsy rate declined from around 90% in the years from 1972 to 1992 to 54% in the present study. Major diagnostic errors (class I and II) declined significantly from 30 to 7% (P<0.001) over the last 30 years. Class I errors decreased from 16 to 2% (P<0.001) in the year 2002. Sensitivity for cardiovascular diseases increased from 69 to 92% (P=0.006), for infectious diseases from 25 to 90% (P=0.013) and for neoplastic diseases from 89 to 100% (P=0.053). Specificity for cardiovascular diseases increased from 85 to 98% (P<0.001) but was unchanged at a high level for infectious diseases and neoplastic diseases. The number of diagnostic procedures increased from 144 to 281 (P<0.001) with an increase in the number of computer tomography investigations and of tissue sampling in the last decade. The frequency of major diagnostic errors has been further reduced at the beginning of the new millennium probably due in large part to new diagnostic tools.

Atheroembolic disease--a frequently missed diagnosis: results of a 12-year matched-pair autopsy study.

Diagnosis of atheroembolic disease (AD) is challenging, because no specific test is available and AD often masquerades as other clinical conditions. We conducted the current study to investigate the relative frequency of autopsy-proven AD over time, to describe its clinical presentation, and to identify risk factors for AD. We screened 2066 autopsy reports from 1995 to 2006 for AD. For each AD case, a control patient without AD was matched for age, sex, and autopsy year. Diagnostic and therapeutic interventions (surgery, catheter interventions, and drug treatment) in the last 6 months before death, as well as clinical and laboratory parameters during the last hospitalization, were retrieved from electronic charts. We identified 51 patients with AD. Among these only 6 (12%) had been diagnosed clinically. The organs most often affected were kidney (71%), spleen (37%), and lower gastrointestinal tract (22%). The relative AD frequency decreased over time from 3.5 to 0.5 per 100 autopsies, whereas the frequency of clinically suspected and biopsy-proven AD remained constant. Among clinical signs, skin lesions such as livedo reticularis and blue toe (33% vs. 14%; p = 0.04) were significantly increased in AD patients compared with the matched controls. We also observed a trend for higher incidence of eosinophilia and proteinuria in AD patients. Vascular interventions within 6 months before death were highly associated with AD (55% vs. 29%; p = 0.01), and in a multivariable analysis this remained the only significant risk factor for AD. Thus, the diagnosis of AD is frequently missed. Vascular interventions represent the most important risk factor for AD and should be performed restrictively in high-risk patients.

Mechanical complications after myocardial infarction reliably predicted using C-reactive protein levels and lymphocytopenia.

We assessed the accuracy of C-reactive protein (CRP) levels and lymphocyte counts to predict a mechanical complication (MC) after myocardial infarction (MI). Within 10 years, we identified 36 patients with 39 echocardiographically confirmed MC within 30 days of MI: ventricular septal defect (17 cases), papillary muscle rupture (10 cases), and left ventricular free wall rupture (12 cases). They were compared to 41 controls with an uncomplicated hospital course after MI. Peak CRP levels and minimum relative lymphocyte counts obtained within 96 h of the acute MI (AMI) and before diagnosis of the complication were compared with clinical parameters. Prior to the MC, peak CRP levels were significantly higher (p < 0.001) and relative lymphocyte counts lower (p < 0.001) than in controls while creatine kinase levels did not differ (p = nonsignificant). Using multivariate logistic regression, the following score was identified to have excellent prognostic significance for MC: CRP (mg/l) - 10 x Lyc (%). The area under the receiver-operating characteristic curve was 0.90 +/- 0.05 (p < 0.001). Combined use of CRP levels and relative lymphocyte counts may be helpful in accurately predicting an MC after AMI and should therefore be routinely assessed.