Discordance between ICD-Coded Myocardial Infarction and Diagnosis according to the Universal Definition of Myocardial Infarction.
The ICD coding system is the standard diagnostic tool used to enable the storage and retrieval of diagnostic information for clinical, epidemiological, and quality healthcare purposes, as well as for reimbursement and resource allocation decision-making (8). The WHO endorses the use of the ICD coding system across its members, and in the US, the National Center for Health Statistics (NCHS) is responsible for regulating the use, interpretation, and periodic revision of ICD codes (8, 9).
Although the Universal Definition of MI subtypes, including T2MI, are not recognized by the ICD coding system, clinicians in practice may use the terms "T1MI" and "T2MI" as recommended by the Global Task Force consensus document (1). The primary purpose of our study was to assess the concordance between T1MI and T2MI according to the Universal Definition of MI classification system and ICD-9 coding for the diagnosis of MI. Our secondary purpose was to identify the ICD-9-coded non-MI diagnoses that were adjudicated as MI using the Universal Definition of MI.
STUDY DESIGN AND POPULATION
Following institutional review board approval, we prospectively included a total of 1927 consecutive, unselected patients [Use of TROPonin In Acute coronary syndromes (UTROPIA); NCT02060760] in whom serial cardiac troponin I (cTnI) measurements were ordered on clinical indication in the emergency department (ED) at Hennepin County Medical Center (Minneapolis, MN) from February 4, 2014, through May 9, 2014. Patients who were <18 years old, pregnant, or had trauma, those who declined to participate in research as documented on information disclosure, and those who did not present through the ED or were transferred from an outside hospital were excluded.
Hospital coders classified each patient in primary, secondary, and tertiary diagnoses following the ICD-9 coding system, in use at the time of study enrollment. Following the approach of the Third Universal Definition of MI (2), all patients with at least 1 cTnI result over the 99th percentile were adjudicated by 2 clinicians, with a third senior adjudicator consulted in case of discrepancy, based on all available clinical data. We compared primary, secondary, and tertiary ICD-9 diagnosis code 410 vs adjudicated T1MI and T2MI based on both contemporary (cTnI) and high-sensitivity (hs-cTnI) assay adjudications. Patient diagnoses and cTnI concentrations for all discrepant diagnoses in those cases that were (a) adjudicated as an MI but were not coded as an MI and (b) coded as an MI but not adjudicated as an MI were determined following chart review based on ICD-9 coding.
Fresh EDTA plasma samples were simultaneously measured with both the contemporary cTnI and hs-cTnI assays on the Abbott ARCHITECT i1000SR or i2000SR analyzer. The contemporary cTnI assay was used by providers for clinical decision-making, because the US Food and Drug Administration (FDA) has not yet cleared the hs-cTnI assay for clinical use. Sex-specific 99th-percentile cutoffs for the hs-cTnI assay were used at 16 ng/L for females and 34 ng/L for males (10-12). Total imprecision (%CV) was 5.3% at 15 ng/L (10), with a 20% CV at the assay limit of detection (LoD) of 1.9 ng/L. For the contemporary cTnI assay, the %CV was 18.8% at 0.030 [micro]g/L 99th percentile, with an LoD of 0.009 [micro]g/L.
We calculated the percent of agreement for diagnostic accuracy for MI and used the Kappa index and McNemar tests to assess the concordance between ICD-9-coded MI and adjudicated MI according to the Third Universal Definition of MI. Statistical estimations were performed for both cTnI and hs-TnI assays.
As shown in Table 1, among the 249 MIs adjudicated using the contemporary cTnI assay, only 69 (28%) were ICD-9-coded MIs. Of the 180 MI-adjudicated patients that were not ICD-9 coded as MI, 34 (19%) were adjudicated as T1MI and 146 (81%) as T2MI. For those coded as MI using ICD codes, 55 (79%) were T1MI and 14 (21%) were T2MI. A fair Kappa index of 0.386 (95% CI 0.319-0.452) and a McNemar difference of 0.0892 (P < 0.001) were found. Among the 207 MIs adjudicated based on the hs-cTnI assay, only 67 (32%) were ICD-9-coded MIs. Of 140 MI-adjudicated patients that were not ICD-9 coded as MI, 27 (19%) were adjudicated as T1MI and 113 (81%) as T2MI. For those coded as MI using ICD codes, all were adjudicated as MI: 57 (85%) as T1MI and 10 (15%) as T2MI. A moderate Kappa index of 0.439 (95% CI 0.367-0.511) and a McNemar difference of 0.0674 (P < 0.001) were found.
Tables 2 and 3 show the ICD-9 -coded diagnoses for patients that were adjudicated as MI based on the contemporary cTnI assay (72%; Table 2) and the hscTnI assay (68%, Table 3), but not coded as MI. Agreement in cases adjudicated as non-MI was excellent, with the ICD-9-coded findings disagreeing in fewer than 1% of adjudicated patients for both assays.
Several findings are unique to our study assessing the concordance of MI diagnoses according to ICD codes and the Universal Definition of MI. First, a substantial number of patients with increased contemporary cTnI (72%) and hs-TnI (68%) who were adjudicated as MI were not ICD-9 coded as MI. ICD-9 codes did not capture the majority of MIs, with 80% being T2MI. These data strongly support including T2MI under the ICD-10 coding system. The inclusion of T2MI in ICD-10 is important for several reasons. First, in current US practice, clinicians may opt to not label patients as having T2MI owing to concerns over being penalized by coders, quality review programs, or both for deviating from accepted guideline-driven therapies geared toward T1MI (e.g., aspirin on arrival and discharge, [beta]-blockers, statin prescribed on discharge) that may not be appropriate for T2MI and/or lack evidence supporting their use in the absence of atherothrombosis (3). Second, we have shown that in clinical practice, ICD codes represent a mixture of both T1MI and T2MI. WHO endorses the ICD coding system as the standard diagnostic tool for epidemiology, health management, and clinical purposes (7-9). Before introducing T2MI in the 2007 Universal Definition of MI consensus document (1), clinicians primarily equated MI to an event occurring due to an atherothrombotic mechanism, and therefore the ICD codes served as a robust, simple tool to track the incidence of MI, as well as to assess compliance with evidence-based therapies.
Since 2007, several studies have demonstrated that in real-life cohorts (instead of selected patients from randomized trials or presenting exclusively with chest pain), T2MI has a substantial incidence (3-6). We have demonstrated that ICD-coded MIs represent a heterogeneous mixture of atherothrombotic and non-atherothrombotic events, potentially confounding the use and interpretation of such information. A paucity of data exists that addresses the use of ICD-coded MI since the introduction of the 5-MI subtype classification by the Universal Definition of MI. Lofthus et al. retrospectively examined every inpatient encounter for 1 year with a final primary coded diagnosis of acute MI at 2 hospitals (289 and 139 patient encounters at each hospital) and adjudicated each case according to the Universal Definition of MI (13). These investigators demonstrated that approximately 25% of patient encounters with a primary coded diagnosis of acute MI did not have T1MI. Similar to our findings, ICD-coded MI represented a mixture of T1MI and T2MI (13).
In our current study, the percentage of patients with T2MI was similar regardless of whether the cTnI or hscTnI assay was used for adjudication. Few patients who were ICD coded as MI were not adjudicated as MI. Five patients coded as MI were adjudicated by hs-cTnI as non-MI because all hs-cTnI results were below the 99th percentile. However, cases with adjudication by cTnI results, which did include values above the 99th percentile, were classified as MI. This likely demonstrates a slight adjudication discrepancy based on cTnI assays alone.
The Third Global MI Task Force established the expert consensus document for the diagnosis of MI based on pathophysiological differences between MI types (2). The American College of Cardiology and the American Heart Association (ACC/AHA) Task Force on Data Standards formally developed the classifications as a controlled terminology for the purpose of creating a clinical standard (14). The revised MI definition has medical, individual, and societal ramifications that lead to the need for consistent identification using diagnostic codes for statistical review. Both the ACC and the AHA, among others, have requested ICD-10-CM (International Statistical Classification of Diseases and Related Health Problems, 10th revision, Clinical Modification) codes for specific MI subtypes to globalize the clinical profile of MIs, specifically with regard to T2MIs (15). The inclusion of T2MI in future ICD codes would facilitate research focused on the epidemiology, management, and outcomes for T2MI using electronic healthcare records and/or available databases.
Our study has several limitations. First, there is no gold standard methodology to adjudicate for T2MI, with a wide range of adjudication methods and frequencies reported in the literature owing to an absence of clearly defined criteria for diagnosis of this entity (3). Nearly all studies assessing T2MI rely on an adjudication process that is limited by the adjudicator's interpretation of the available data. In this study, all cases were adjudicated by at least 2 clinicians familiar with the Universal Definition of MI, with a third senior adjudicator reviewing all cases of discrepancy. Hence, although there may be potential limitations in all studies related to the adjudicated process, it should be noted that at present, there are no consistent, reproducible criteria to define this entity; and therefore, adjudications of T2MI are made on the basis of detailed review of available data and the best application of current knowledge according to the Third Universal Definition of MI. Therefore, we emphasize that the difference between ICD-coded MI diagnoses and adjudicated MI diagnoses according to the Third Universal Definition of MI is explained primarily by reluctance to apply the ICD MI code to these cases.
Second, the listed diagnoses in Tables 2 and 3 are on the basis of ICD codes (not adjudicated events). Several of these ICD-coded events represent an ongoing matter of debate as to whether they should be labeled only as acute myocardial injury or, more specifically, as acute MI (T1MI orT2MI).
Third, the ICD codes used in this study were employed on the basis of real-life ICD codes (ICD-9) used by our institution at the time of the study in 2014, and therefore they are not representative of the updated ICD-10 classification system. However, there remains no code for T2MI in the ICD-10, and therefore, we would not expect a change in our findings. Finally, our findings are representative of a single-site US center and may not be generalizable to other centers; further studies will be required both inside and outside the US to validate our findings.
In conclusion, ICD-9-coded MIs capture a small proportion of adjudicated MI diagnoses, primarily a result of not coding T2MI as ICD-coded MIs. ICD-9 codes represent a mixture of both T1MI and T2MI. These findings suggest the need for caution when using ICD-9 or ICD-10 codes to assess for quality review, resource utilization, and outcomes. Our findings highlight the need for an ICD-10 code for T2MI.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Authors' Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Disclosures and/or potential conflicts of interest:
Employment or Leadership: F.S. Apple, Clinical Chemistry, AACC.
Consultant or Advisory Role: Y. Sandoval, Roche Clinical Diagnostics; S.W. Smith, Alere, Siemens, and Roche; F.S. Apple, Philips Healthcare Incubator, Metanomics, and HyTest.
Stock Ownership: None declared.
Honoraria: F.S. Apple, Abbott POC and Instrumentation Laboratory.
Research Funding: F.S. Apple, research through Minneapolis Medical
Research Foundation (MMRF), not salaried: Abbott Diagnostics, Roche Diagnostics, Siemens Healthcare, Alere, Trinity, Nanomix, Becton Dickinson.
Expert Testimony: None declared.
Patents: None declared.
Other Remuneration: J. Diaz-Garzon, International Federation of Clinical Chemistry. Young Scientist Exchange Program.
Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, and final approval of manuscript.
(1.) Thygesen K, Alpert JS, White HD. Universal definition of myocardial infarction. J Am Coll Cardiol 2007;50: 2173-95.
(2.) Thygesen K, Alpert JS, Jaffe AS, Simoons ML, Chaitman BR, White HD, et al. Third universal definition of myocardial infarction. J Am Coll Cardiol 2012;60:1581-98.
(3.) Sandoval Y, Smith SW, Thordsen SE, Apple FS. Supply/ demand type 2 myocardial infarction: should we be paying more attention? J Am Coll Cardiol 2014;63: 2079-87.
(4.) Saaby L, Poulsen TS, Hosbond S, Larsen TB, Pyndt Diederichsen AC, Hallas J, et al. Classification of myocardial infarction: frequency and features of type 2 myocardial infarction. Am J Med 2013;126:789-97.
(5.) Shah AS, McAllister DA, Mills R, Lee KK, Churchhouse AM, Fleming KM, et al. Sensitive troponin assay and the classification of myocardial infarction. Am J Med 2015; 128:493-501.
(6.) Sandoval Y, Thordsen SE, Smith SW, Schulz KM, Murakami MM, Pearce LA, Apple FS. Cardiac troponin changes to distinguish type 1 and type 2 myocardial infarction and 180-day mortality risk. Eur Heart J Acute Cardiovasc Care 2014;3:317-25.
(7.) Mendis S, Thygesen K, Kuulasmaa K, Giampaoli S, Mahonen M, Ngu Bl, et al. World Health Organization definition of myocardial infarction: 2008-09 revision. Int J Epidemiol 2011;40:139-46.
(8.) World Health Organization, International Classification of Diseases (ICD). http://www.who.int/classifications/ icd/en/ (Accessed June 2016).
(9.) CDC, National Center for Health Statistics. Classification of diseases, functioning, and disability. http://www.cdc. gov/nchs/icd/index.htm (Accessed May 2016).
(10.) Sandoval Y, Smith SW, Schulz KM, Murakami MM, Love SA, Nicholson J, Apple FS. Diagnosis of type 1 and type 2 myocardial infarction using a high-sensitivity cardiac troponin I assay with sex-specific 99th percentiles based on the third universal definition of myocardial infarction classification system. Clin Chem 2015;61: 657-63.
(11.) Shah AS, Griffiths M, Lee KK, McAllister DA, Hunter AL, Ferry AV, etal. High sensitivity cardiactroponin and the under-diagnosis of myocardial infarction in women: prospective cohort study. BMJ 2015;350:g7873.
(12.) Shah AS, Anand A, Sandoval Y, Lee KK, Smith SW, Adamson PD, etal. High-sensitivity cardiac troponin I at presentation in patients with suspected acute coronary syndrome: a cohort study. Lancet 2015; 386:2481-8.
(13.) Lofthus DM, Khalili H, Raja VN, Regan CJ, Gadgil JY, Castillo D, etal. Accuracy of acute myocardial infarction clinical diagnosis and its implications. Int J Cardiol 2015;186:54-6.
(14.) Hicks KA, Tcheng JE, Bozkur B, Chaitman BR, Cutlip DE, Farb A, et al. 2014 ACC/AHA key data elements and definitions for cardiovascular endpoint events in clinical trials: a report of the American College of Cardiology/ American Heart Association Task Force on clinical data standards (Writing Committee to Develop Cardiovascular Endpoints Data Standards). J Am Coll Cardiol 2015; 66:403-69.
(15.) CDC. ICD-10 Coordination and Maintenance Committee Meeting: March 9-10, 2016: Diagnosis agenda: http://www.cdc.gov/nchs/data/icd/topic_packet_ 03_09_16.pdf (Accessed November 2016).
Jorge Diaz-Garzon,  Yader Sandoval, [2,3] Stephen W. Smith,  Sara Love,  Karen Schulz,  Sarah E. Thordsen, [2,3] Benjamin K. Johnson, [2,3] Brian Driver,  Katherine Jacoby,  Michelle D. Carlson, [2,3] Kenneth W. Dodd,  Johanna Moore,  Nathaniel L. Scott, [4,6] Charles A. Bruen,  Ryan Hatch,  and Fred S. Apple [1,5] *
 Department of Laboratory Medicine and Pathology, Hennepin County Medical Center and University of Minnesota, Minneapolis, MN;  Division of Cardiology, Hennepin County Medical Center, Minneapolis, MN;  Minneapolis Heart Institute, Abbott North western Hospital, Minneapolis, MN;  Department of Emergency Medicine, Hennepin County Medical Center and University of Minnesota, Minneapolis, MN;  Minneapolis Medical Research Foundation, Minneapolis, MN;  Department of Medicine, Hennepin County Medical Center, Minneapolis, MN.
* Address correspondence to this author at: Hennepin County Medical Center, Clinical Labs P4, 701 Park Ave., Minneapolis, MN, 55415. Fax 612-904-4229; e-mail firstname.lastname@example.org.
Received July 7,2016; accepted September 13,2016.
Previously published online at DOI: 10.1373/clinchem.2016.263764
 Nonstandard abbreviations: MI, myocardial infarction; T1MI,type 1 MI; T2MI, type 2 MI; ICD, International Classification of Diseases; NCHS, National Center for Health Statistics; UTROPIA, Use of TROPonin In Acute coronary syndromes; cTnI, cardiac troponin I; ED, emergency department; hs-cTnI, high-sensitivity cTnI; FDA, US Food and Drug Administration; LoD, limit of detection; ACC, American College of Cardiology; AHA, American Heart Association; ICD-10-CM, International Statistical Classification of Diseases and Related Health Problems, 10th revision, Clinical Modification.
Table 1. Concordance between adjudicated and ICD-9-coded MI and non-MI classifications by MI subtypes 1 and 2. cTnI adjudication MI Non-MI Total ICD MI 69 (T1MI 79.2%; T2MI 20.8%) 8 77 Non-MI 180 (T1MI 18.9%, 10% F; 1670 1850 T2MI 81.1%, 58% F) Total 249 (T1MI 35.7%; T2MI 64.3%) 1678 1927 hs-TnI adjudication MI Non-MI Total ICD MI 67 (T1MI 85.1%; T2MI 14.9%) 10 77 Non-MI 140 (T1MI 19%, 15% F; 1288 1428 T2MI 81%, 58% F) Total 207 (T1MI 40.6%; T2MI 59.4%) 1298 1927 Table 2. Patient diagnoses that were adjudicated as MI based on the contemporary cTnI but were not ICD-9 coded as MI. MI adjudicated, not coded as MI (n = 146) Primary ICD-9 diagnosis n % Hypertensive disease 25 17.1 Acute respiratory failure 17 11.6 Arrhythmia 17 11.6 Pulmonary obstructive 13 8.9 disease and pneumonia Septicemia 11 7.5 Drug poisoning 8 5.5 Chest pain 7 4.8 Acute chronic systolic HF 4 2.7 Cerebral embolism 2 1.4 and cerebral infarction Takotsubo syndrome 2 1.4 Others 40 27.5 Table 3. Patient diagnoses that were adjudicated as MI based on the hs-cTnI assay but were not ICD-9 coded as MI. MI adjudicated, not coded as MI (n = 113) Primary ICD-9 diagnoses n % Hypertensive disease 16 14.2 Acute respiratory failure 15 13.3 Acute or chronic systolic HF 12 10.6 Drug poisoning 10 8.8 Arrhythmia 9 8.0 Septicemia 5 4.4 Pneumonia 4 3.5 Cerebral embolism and 2 1.8 cerebral infarction Chest pain 2 1.8 Syncope and collapse 2 1.8 Others 36 31.8
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Other Areas of Clinical Chemistry|
|Author:||Diaz-Garzon, Jorge; Sandoval, Yader; Smith, Stephen W.; Love, Sara; Schulz, Karen; Thordsen, Sarah E|
|Date:||Jan 1, 2017|
|Previous Article:||Rule-In and Rule-Out of Myocardial Infarction Using Cardiac Troponin and Glycemic Biomarkers in Patients with Symptoms Suggestive of Acute Coronary...|
|Next Article:||Trimethylamine N-oxide and Risk Stratification after Acute Myocardial Infarction.|