Hospital autopsy rates in the United States declined from

Transcription

1 Original Article Comparison of Clinical Diagnoses and Autopsy Findings Six-Year Retrospective Study Hyejong Song Marshall, MD; Clara Milikowski, MD Context. The frequency of autopsies has declined in most developed countries beginning in the latter half of the 20th century. During this time period the technology of medicine made significant advances; however, it is important to regularly reevaluate the role of the autopsy to confirm suspected diagnoses and identify unsuspected findings. Objective. To determine what portion of autopsies reveal clinically meaningful unexpected findings. Design. Reports that included clinical histories of autopsies performed at Jackson Memorial Hospital during the 6 years between 2009 and 204 were reviewed by 2 pathologists. Each case was classified using the Goldman Classification. Results. In the given time period, 923 autopsies were performed; 52 patients (55.5%) were adults. A total of 334 cases were subject to review after excluding those with a short (, day) hospital stay, restriction to a single Hospital autopsy rates in the United States declined from 30% to 40% in the 960s to 8% in the early 2000s. 2 Multiple factors, such as the elimination of the hospital autopsy rate requirement in 972, lack of reimbursement, cultural and religious objections by families, and increasing workload for house staff, account for the decrease. 3 Even as sophisticated medical technologic tools have been developed and made available, the role of the autopsy for confirming clinical diagnoses and identifying unsuspected findings should be reevaluated regularly. Herein, we report our findings of the importance of the autopsy by comparing the clinical diagnosis with the anatomic findings in a single large academic institution between the years of 2009 and 204. MATERIALS AND METHODS Clinical information of all adult (8 years or older) autopsy cases performed in Jackson Memorial Hospital (Miami, Florida) for 6 Accepted for publication December 29, 206. Published as an Early Online Release June 28, 207. From the Department of Pathology and Laboratory Medicine, Jackson Memorial Hospital/University of Miami, Miami, Florida. The authors have no relevant financial interest in the products or companies described in this article. The data and abstract were presented at the annual meeting of the United States and Canadian Academy of Pathology (USCAP); March 4, 206; Seattle, Washington. Reprints: Clara Milikowski, MD, Pathology and Laboratory Medicine, Jackson Memorial Hospital/University of Miami, 6 NW 2th Ave, Holtz Children s Hospital, 2nd Floor, Suite 2042, Miami, FL 3336 ( CMilikowski@med.miami.edu). organ or body cavity, and cases referred from other facilities. A total of 33 of 334 cases (9.9%) were identified as class I discrepancy, where the autopsy revealed a discrepant diagnosis with a potential impact on survival or treatment. Critical findings, such as untreated infection (5 of 33 cases; 45.5%), pulmonary embolism (8 of 33 cases; 24.2%), and undiagnosed malignancy (6 of 33 cases; 8.2%), were found in these cases. Major significant findings that had not been clinically detected, whether clinically manageable or not (class I and II), were found in 65 of 334 cases (9.5%). Conclusion. Despite intensive modern clinical investigations, autopsies continue to reveal major antemortem diagnostic errors in a significant number of cases. (Arch Pathol Lab Med. 207;4: ; doi: /arpa OA) years (from January 2009 to December 204) was retrospectively evaluated. Only cases with a stay longer than 24 hours in Jackson Memorial Hospital were included in the study in order to ensure enough time for initial clinical evaluation and management. Exclusion criteria were: () a hospital stay of less than 24 hours, (2) restrictions to single-organ or body cavity dissections, and (3) cases referred from other facilities. Cases from the University of Miami Hospital (Miami, Florida), which shares attending physicians and residents in the adjacent campus and whose full medical records were available, were considered in-house cases. All cases had consent for autopsy from the legal next of kin, and each autopsy was performed by pathology residents who observed restrictions, if there were any. Attending pathologists directly supervised each case. For each eligible autopsy the autopsy report, including clinical history, was reviewed by 2 pathologists (H.M. and C.M.). Medical records were reviewed if the clinical information provided in the autopsy report was unclear or inadequate. Each case was classified using the Goldman Classification (Table ). 4 When a case could be classified with more than one discrepancy, the higher class (smaller number) was regarded as the final classification. The pathologists reviewed the cases independently, compared their assessments, and came to consensus by discussion if there was discordance. The study design was reviewed by the Institutional Review Board of the University of Miami and approved to be conducted as nonhuman research. RESULTS Of 923 autopsies performed in Jackson Memorial Hospital for the 6 years ( ), 52 cases (55.5%) were performed on adults (age.8 years). A total of 334 of Arch Pathol Lab Med Vol 4, September 207 Autopsy Discrepancy Rates Marshall & Milikowski

2 Table. Criteria of Goldman Classification Class Type of Discrepancy Definition Example: Death Due to: I Major Directly related to death; if recognized, may have altered treatment or survival Unsuspected myocardial infarction presenting with chest pain II Major Directly related to death; if recognized, would not have altered treatment or survival Unsuspected myocardial infarction presenting with cardiac arrest III Minor Incidental autopsy finding not directly related to death but related to terminal disease process IV Minor Incidental autopsy finding unrelated to cause of death V No error Clinical and autopsy diagnoses in complete agreement Known myocardial infarction with unsuspected left ventricular mural thrombus Known myocardial infarction with unsuspected lung cancer Table 2. Numbers of Cases Included and Excluded by Year Year Total Total number of autopsies (A) Cases of age older than 8 y (B) Outside case (C) Short (, d) stay (D) Single organ or single cavity only (E) Cases subject to review (B [C þ D þ E]) cases (65.2%) were subject to review after excluding those with a short (, day) hospital stay, restriction to a singleorgan or body cavity dissection, and those referred from other facilities (Table 2). The demographics of the patients included are presented in Table 3. A total of 33 of 334 cases (9.9%) were identified as a class I discrepancy, where the autopsy revealed a discrepant diagnosis with a potential impact on survival or treatment (Table 4). Various critical findings, such as untreated infection (5 of 33 cases; 45.5%), pulmonary embolism (8 of 33 cases; 24.2%), and undiagnosed malignancy (6 of 33 cases; 8.2%) were identified in cases of this class (Table 5). For the cases with class I discrepancies, the clinical history or impression and critical laboratory or imaging results are listed in Table 6. Major significant findings that had not been clinically detected, whether they might or might not have clinical impact, were found in a total of 65 of 334 cases (class I and II; 9.5%). Table 3. Patient Demographics Item No. (%) Sex Male 89 (56.6) Female 45 (43.4) Total 334 (00) Age, y (.4) (6.6) (2.9) (26.6) (25.4) (2.6).80 5 (4.5) Total 334 (00) DISCUSSION This study is a retrospective review of the adult autopsy reports for recent years in the third largest public hospital (550 beds) in the United States. It excluded cases from outside institutions, patients with short hospital stays, and those in which the autopsy was restricted to a single organ or body cavity, in order to optimize the clinical and pathologic correlation. Even though the exclusion made the size of the study smaller, the entire set of included cases was comprehensively reviewed, unlike in some previous studies which only reviewed randomly sampled cases. 5 There was a limitation to calculate the exact autopsy rate because: () some cases were declared to be under the jurisdiction of the medical examiner, and therefore dropped from the initial data set, and (2) Jackson Memorial Hospital has many branches and affiliated hospitals from which autopsies are requested, and the total number of deaths in each hospital was not available. Despite intensive modern clinical investigations, autopsies have continued to reveal major antemortem diagnostic errors in as many as 30% of cases. 6 8 Follow-up or metaanalysis studies show that the rate of clinically significant discrepancy is decreasing over decades; however, 4% to 7% of cases still have class I discrepancies.,5 This study shows a Table 4. Result of Case Classification Class No. (%) of Cases I 33 (9.9) II 32 (9.6) III 2 (3.6) IV 0 (32.9) V 47 (44.0) 334 (00) Arch Pathol Lab Med Vol 4, September 207 Autopsy Discrepancy Rates Marshall & Milikowski 263

3 Table 5. Findings With Potential Impact on Survival or Treatment (Class I Discrepancy) Item and Findings No. (%) Infection 5 (45.5) ( case with coexisting 6 fungal myocarditis) Viral pneumonia 2 Bacterial pneumonia 2 Miliary tuberculosis and bacterial meningitis Meningitis of unknown etiology Toxoplasmosis involving lung, liver, and brain Candidemia confirmed by postmortem blood culture Renal abscess 8 (24.2) Malignancy 6 (8.2) Lymphoma involving multiple organs 3 Diffuse large B-cell lymphoma Follicular lymphoma Peripheral T-cell lymphoma Pulmonary carcinoma 2 Small cell carcinoma Squamous cell carcinoma Gastric adenocarcinoma Cardiovascular 3 (9.) Retroperitoneal and/or intraabdominal 2 hemorrhage Cartilaginous emboli Immunologic (3.0) Anaphylactic laryngeal edema Total 33 (00) comparable result of 9.9% of clinically significant findings to the previous studies. Undiagnosed infections, especially pneumonia, and pulmonary embolism were the 2 most common significant unexpected findings. These are common complications of hospitalization and subsequent immobilization. The character of the study, which excluded the short hospital stay, may make the frequency of those complications more notable by a possible exclusion of sudden cardiovascular accidents, which might have a significant portion in autopsies but have a short time period between the initial clinical presentation and the patient s demise. Because early suspicion and detection may make for a significantly better prognosis in those potentially fatal complications, vigilant management for preventing these conditions is suggested. Beyond the direct clinical or administrative advantages, autopsies have other virtues, such as educational and epidemiologic values. 9 The autopsy should continue to be the gold standard for quality control in clinical management, including radiologic evaluation, 0 5 in spite of declining requests. 6 The most crucial factors influencing attitudes toward the autopsy have been shown to be the clinician s level of experience with autopsy in training and practice. Therefore, the importance of the autopsy should be emphasized in medical education and postgraduate training so the number of significant diagnostic discrepancies can be reduced and patient care can be optimized. Table 6. Clinical History or Impression and Critical Labs or Imaging Compared With Class I Autopsy Findings Serial Age, y/sex Clinical History or Impression Critical Lab or Imaging Significant Autopsy Findings 34/M History of stroke, shortness of breath, flulike symptoms, possible pneumonia 2 8/M DM, valvular heart disease, fever and chills after procedure for benign prostatic hyperplasia 3 68/F DM, HTN, ischemic cardiomyopathy Diffuse nodularity on chest radiograph, MSSA in blood Blood and urine culture negative Atrial fibrillation 4 83/F Myasthenia gravis, pneumonia Pneumonia and pulmonary edema, Proteus bacteremia 5 64/F COPD aggravation Hilar lymphadenopathy on imaging, BAL with Candida 6 56/M DM, HTN, nausea, vomiting, chest pain Hepatitis C virus positive with elevated liver enzymes, cardiac enzyme negative 7 68/M Pneumonia Blood culture negative, pleural effusion, consolidation of lung 8 48/F Mixed connective tissue disease, Rhabdomyolysis, CMV, Klebsiella, sepsis by multiple organisms Acinetobacter, Enterococcus 9 65/F Leukemia, s/p bone marrow Elevated liver enzymes, negative transplant blood culture and BAL 0 65/F s/p liver transplant, transfusion of Bibasilar opacity and pleural red blood cells to correct effusion on chest radiograph anemia 55/M Admitted for schizophrenia, COPD 2 58/M Adrenal mass, steroid therapy for numbness Adrenal mass, pleural effusion on imaging, hypercalcemia 3 54/M s/p kidney transplant, hematuria and abdominal pain, chest pain Nonsignificant electrocardiogram and troponin 4 43/F HIV, HTN, hypertensive crisis Pulmonary edema and enlarged cardiac silhouette, positive for cocaine CMV pneumonia and colitis, not bacterial Small cell carcinoma of lung with metastasis to lymph nodes and bone Acute retroperitoneal hemorrhage extending into abdominal cavity HSV pneumonia, not bacterial or fungal, not just viral and bacterial Toxoplasmosis involving lung, liver, and brain Laryngeal edema Squamous cell carcinoma of lung Diffuse large B-cell lymphoma, meningeal malignant lymphomatosis Chronic meningitis of undetermined etiology 264 Arch Pathol Lab Med Vol 4, September 207 Autopsy Discrepancy Rates Marshall & Milikowski

4 Table 6. Continued Serial Age, y/sex Clinical History or Impression Critical Lab or Imaging Significant Autopsy Findings 5 77/M Dementia, HTN, normal pressure Bacterial pneumonia hydrocephalus, hematemesis 6 59/M POEMS syndrome, vomiting Hepatomegaly, spontaneous Follicular lymphoma bacterial peritonitis 7 39/M Neck pain, complicated hospital course including pulmonary embolism Edema of spinal cord Cartilaginous emboli to cervical spinal cord, recent hematoma of thalamus 8 57/M DM, s/p kidney and pancreas transplant, nausea and abdominal pain 9 69/M Lymphoma, s/p chemotherapy, rash, fever 20 23/F 36-wk pregnancy, pulmonary infiltrates, emergency cesarean delivery 2 58/M Inguinal hernia, persistent HTN after surgery 22 30/F Anti-NMDA receptor autoimmune/ paraneoplastic encephalitis 23 54/M HTN, DM, s/p kidney transplant, pulmonary edema, acute renal failure 24 58/F DM, HTN, weight loss, lymphadenopathy, shortness of breath 25 50/F Bronchiectasis, cirrhosis due to probable autoimmune hepatitis, hemoptysis 26 4/M HTN, sickle cell trait, after elective lumbar disk surgery, abdominal pain and respiratory failure, renal injury, disseminated intravascular coagulation, possible septicemia 27 62/F HTN, autoimmune hepatitis overlap with primary biliary cirrhosis, varices, hepatorenal syndrome, acute respiratory distress syndrome 28 74/F Cirrhosis, pleural effusion, aspiration pneumonia 29 54/F Asthma, cough, chest pain, antibiotic treatment, diarrhea, biopsy was planned but not done 30 20/M Cerebral palsy, hydrocephalus s/p shunt, hip dislocation s/p replacement. Septic shock. Treated with antibiotics 3 42/M HIV, DM, congestive heart failure, atrial fibrillation 32 58/M Gout, asthma, HTN, atrial fibrillation, extremity weakness, and altered speech 33 76/F Bladder and upper urinary tract cancer, DM, coronary artery disease, sudden collapse in hospital stay for urinary tract infection Leukocytosis Pulmonary interstitial edema on chest radiography, gramnegative rods in bone marrow culture workup with negative results Lung consolidation with mild pleural effusion Infection workup with negative results Leukopenia workup with negative result Soft tissue mass encasing the celiac vessels and pericardium, mild elevation of aspergillosis glucomannan level Decreasing hemoglobin and hematocrit Chronic cholestatic hepatitis on liver biopsy, Klebsiella bacteremia MSSA bacteremia, blood culture positive for Klebsiella, aspiration pneumonia on chest radiograph Lung opacity concerning for infection, multiple enlarged lymph nodes Hypernatremia, coagulopathy Hyponatremia, marked and diffuse cerebral edema, MSSA bacteremia, pulmonary opacity; consolidation versus atelectasis Brain CT with no intracranial hemorrhage Mild interstitial pulmonary edema on chest radiograph Intra-abdominal and retroperitoneal hemorrhage Abscesses of transplanted kidney Poorly differentiated adenocarcinoma of stomach with metastasis to vertebral bone marrow and myocarditis Peripheral T-cell lymphoma Candidemia confirmed by postmortem blood culture Miliary tuberculosis. Acute meningitis suggestive of bacterial etiology Bacterial pneumonia Abbreviations: BAL, bronchoalveolar lavage; CMV, cytomegalovirus; COPD, chronic obstructive pulmonary disease; CT, computed tomography; DM, diabetes mellitus; HIV, human immunodeficiency virus; HSV, herpes simplex virus; HTN, hypertension; MSSA, methicillin-sensitive Staphylococcus aureus; NMDA, N-methyl-D-aspartate; POEMS, polyneuropathy, organomegaly, endocrinopathy, monoclonal gammopathy and skin changes; s/p, status post. Arch Pathol Lab Med Vol 4, September 207 Autopsy Discrepancy Rates Marshall & Milikowski 265

5 References. Shojania KG, Burton EC, McDonald KM, Goldman L. Changes in rates of autopsy-detected diagnostic errors over time: a systematic review. JAMA. 2003; 289(2): National Center for Health Statistics. Autopsy Patterns in 2003: Data on Mortality. Hyattsville, MD: US Dept of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics; Scordi-Bello IA, Kalb TH, Lento PA. Clinical setting and extent of premortem evaluation do not predict autopsy discrepancy rates. Mod Pathol. 200;23(9): Goldman L, Sayson R, Robbins S, Cohn LH, Bettmann M, Weisberg M. The value of the autopsy in three medical eras. N Engl J Med. 983;308(7): Schwanda-Burger S, Moch H, Muntwyler J, Salomon F. Diagnostic errors in the new millennium: a follow-up autopsy study. Mod Pathol. 202;25(6): Spiliopoulou C, Papadodima S, Kotakidis N, Koutselinis A. Clinical diagnoses and autopsy findings: a retrospective analysis of 252 cases in Greece. Arch Pathol Lab Med. 2005;29(2): Tavora F, Crowder CD, Sun CC, Burke AP. Discrepancies between clinical and autopsy diagnoses: a comparison of university, community, and private autopsy practices. Am J Clin Pathol. 2008;29(): Gonzalez-Franco MV, Ponce-Camacho MA, Barboza-Quintana O, Ancer- Rodriguez J, Cecenas-Falcon LA. Discrepancies between clinical and autopsy diagnosis: a study of 33 autopsies performed over a 7 years period. Medicina Universitaria. 202;4(54): Burton JL, Underwood J. Clinical, educational, and epidemiological value of autopsy. Lancet. 2007;369(957): Bauer TM, Potratz D, Goller T, Wagner A, Schafer R. Quality control by autopsy: how often do the postmortem examination findings correct the clinical diagnosis [in German]? Dtsch Med Wochenschr. 99;6(2): Grundmann E. Autopsy as clinical quality control: a study of 5,43 autopsy cases. In Vivo. 994;8(5): Hasan M, Woodhouse K. Autopsy: its role in clinical quality control in the elderly in the 990s. Arch Gerontol Geriatr. 995;2(2): Murken DR, Ding M, Branstetter BF 4th, Nichols L. Autopsy as a quality control measure for radiology, and vice versa. AJR Am J Roentgenol. 202;99(2): Sebok J, Magyar E, Csanadi C, Csaky A, Schonfeld T. The importance of the autopsy in quality control of medicine (possibilities in the Hungarian conditions). Orv Hetil. 2005;46(33): Solheim K. Quality control in the last round: autopsy is still necessary [in Norwegian]! Tidsskr Nor Laegeforen. 993;3(20): Hooper JE, Geller SA. Relevance of the autopsy as a medical tool: a large database of physician attitudes. Arch Pathol Lab Med. 2007;3(2): Arch Pathol Lab Med Vol 4, September 207 Autopsy Discrepancy Rates Marshall & Milikowski