Potential limitations of diagnostic standard codes to distinguish polycythemia vera and secondary erythrocytosis: A retrospective case series

doi:10.21203/rs.3.rs-957923/v1

Download PDF

Research Article

Potential limitations of diagnostic standard codes to distinguish polycythemia vera and secondary erythrocytosis: A retrospective case series

https://doi.org/10.21203/rs.3.rs-957923/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Red cell overproduction is seen in polycythemia vera (PV), a bone marrow myeloproliferative neoplasm characterized by trilinear cell proliferation (WBC, platelets), as well as in secondary erythrocytosis (SE), a group of heterogeneous disorders characterized by elevated EPO gene transcription.

We aimed to verify the concordance of the International Classification of Diseases (ICD) code-based diagnosis of “polycythemia” or “erythrocytosis” with the true clinical diagnosis of these conditions.

We retrospectively reviewed the electronic medical records (January 1, 2005, to December 31, 2016) of adult patients with ICD codes of polycythemia and/or erythrocytosis who had testing done for the presence of the JAK2V617F mutation. We verified the accuracy of the ICD code-based diagnoses by meticulous chart review and established whether these patients fulfilled the criteria by the evaluating physician for PV or SE and according to diagnostic guidelines at that time.

We identified and chart reviewed a total of 578 patient records. Remarkably, 11% of the patients had concurrent diagnosis codes for PV and SE and were unable to be classified appropriately without individual chart review.

The ICD code-based diagnostic system led to misidentification in an important fraction of cases. This represents a problem for the detection of PV or SE cases by ICD-based registries and their derived studies. Research based exclusively on ICD codes could have a potential impact on patient care and public health, and limitations must be weighed when research findings are conveyed.

Health Economics & Outcomes Research

Medical Genetics

Erythrocytosis

secondary polycythemia

secondary erythrocytosis

iathrogenic erythrocytosis

chronic hypoxic states

The term polycythemia denotes an increased total red cell volume, sometimes used interchangeably with erythrocytosis, the former recommended by some experts to be used only in the context of primary myeloproliferative diseases.¹Erythrocytosis describes the increase above 2 SD of the values that evaluate red cell mass (hematocrit or hemoglobin). However, the measurement of the previous values is based on whole blood; therefore, it is dependent on plasma volume and subject to changes in it.²

Osler's case series created a different perspective in terms of erythrocytosis etiology.³^] The more distinctive example of primary polycythemia is polycythemia vera (PV), described as the outcome of alterations in the bone marrow that are typically acquired; its characteristic feature is trilineage hematopoietic cell hyperplasia caused by hypersensitivity of the immature red cell line to EPO.² It usually presents with leukocytosis, thrombocytosis, and splenomegaly. WHO criteria of PV in 2016 refer to erythrocytosis as hemoglobin above 16.0 g/dL for females and 16.5 g/dL for males.⁴

Secondary erythrocytosis (SE) is a group of heterogeneous disorders caused by an upregulation of molecular pathways that ultimately leads to the abnormal activation of EPO gene transcription (located on chromosome 7) in response to undesirable tissue oxygenation or the alteration of factors that are involved in erythropoiesis (e.g., hypoxia-induced factor pathway). ⁵Therefore, as an outcome of cardiopulmonary or hypoxia-related issues [smoking, chronic pulmonary disease (COPD), obstructive sleep apnea syndrome (OSA)], a side effect of certain drugs (diuretics, androgens or anabolic steroids, erythropoietin), some renal disorders, autonomous EPO production, and finally, as a diagnosis of exclusion, idiopathic erythrocytosis can be mentioned.⁶The prevalence, mortality and morbidity can be difficult to quantify due to dependent relationships with primary etiologies and paucity of data.

Erythrocytosis is a common reason for hematology services consultation. The prevalence of PV is estimated to be 47-55 per 100,000 in the United States⁷, which contrasts with the paucity of accurate data on SE. Benign hematologists and/or leukemia specialists care for those patients. Despite the expertise and forward understanding of pathophysiology, for some cases, the classification and management of the disease can be far from straightforward and challenging. Moreover, documentation of the appropriate diagnosis is an important task of current medical practice. This implies the use of the international diagnostic classification standard for reporting diseases (ICD codes). ICD codes are widely used in medical and epidemiology research. The introduction of errors at the steps of diagnosis assignment has the potential to introduce systematic mistakes and bias in those types of research studies.

We aimed to describe the concordance of a ten-year ICD code-based cohort of adult patients consulted to our center with a diagnosis of “polycythemia” or “erythrocytosis” (and had testing done for the presence of a JAK2V617F mutation) with the true clinical diagnosis of these conditions, as recorded in the medical records by the evaluating hematologist and supported by the available diagnostic data.

This study is a subanalysis of a previously conducted study that aimed to explore the performance of EPO and JAK2 mutations. Our population was selected if they had both tests performed.⁸ A retrospective review of electronic medical records was conducted from January 1, 2005, to December 31, 2016 at MD Anderson Cancer Center (MDACC) of patients with ICD-9-CM or ICD-10-CM code diagnosis corresponding to polycythemia and/or erythrocytosis, or similar. We did not include patients with codes for familial erythrocytosis or polycythemia neonatorum.

Patients were selected if they had at least 2 consecutive billing codes for any of those diagnoses to avoid selecting patients who had an erroneous code entry and those who did not have follow-ups or recurrent visits to complete diagnostic work-up. Additionally, we also excluded patients who did not have JAK2 mutation studies in their medical records.

We then reclassified the cases as SE or PV after a meticulous chart review of clinical notes, hematopathology, and other diagnostic data. (Figure 1). We collected demographic data, comorbidity conditions and medications, relevant laboratory studies including hemoglobin, serum erythropoietin (EPO) level, JAK 2 mutation status, and bone marrow studies. Records of phlebotomy, polysomnography, tobacco use, and methemoglobin tests were noted.

Data were analyzed using Bioinformatics Core using SAS software and Stata IC statistical software. Variables were described by their frequency, median values, and interquartile range when appropriate.

The study is a retrospective evaluation of the patients and involved minimal risk to subjects, therefore informed consent was waived for this study. The institutional review board at MD Anderson Cancer Center approved the study. All methods were performed in accordance with the relevant guidelines and regulations

A total of 1,092 patients were identified as having at least 2 billing codes for a diagnosis of PV (ICD D45) and SE (ICD D75.1), or both. Of those, 578 corresponded to adult patients (age greater than 18 years of age) who had JAK2 mutation studies. Of those, 480 cases were given an ICD code diagnosis of polycythemia 67 cases with ICD code of erythrocytosis, and 61 cases had both ICD codes used concurrently.

We then chart-reviewed the medical records of the 578 selected cases. Of the patients who had a listed ICD diagnosis of PV, 95% (427) met the WHO criteria for PV; importantly, 4% (19) of those patients had an underlying condition from which the SE diagnosis could have been made (Figure 1). From the group of cases with only secondary erythrocytosis ICD code, 94% (63) had confirmed SE. Overall, those patients with a single diagnosis code (PV or SE) were “miscoded” in only 5% of cases.

The most notable finding was that 11% of all cases had concurrent codes for both PV and SE. Only through chart review were those cases able to be classified appropriately (Figure 1). Of those cases, 5 corresponded to PV, 54 to SE and 2 did not have erythrocytosis.

After medical record review and reclassification of cases, the final population of 434 patients with a confirmed diagnosis of PV had a median age of 59.2 years, 57.72% (329) were male, 86.3% (491) were Caucasian, 3.3% (19) were African American and 10.53% (60) corresponded to other racial groups. A total of 136 patients had a confirmed diagnosis of SE. The median age for that group was 56 years, and 58% of the patients were male (Table 1).

Only two patients with SE were noted to have a positive somatic JAK2 V617F mutation by PCR, and they did not have bone marrow morphology criteria compatible with PV. Additionally, 5 patients with SE had other variants (predominantly of germline origin) in the JAK2 gene. JAK2 missense germline variants were reported in 4 cases: 2 (L393 V) gain function, 1 (R1063H), 1 (E846D), and a variant of uncertain origin (R715G). One of those patients had a concurrent germline variant in the TET2 gene. These findings are similar to recent observations from large European cohorts that have demonstrated the presence of clonal hematopoiesis in some general populations of patients with erythrocytosis without a diagnosis of PV.⁹ In those cohorts, JAK-2 abnormalities were the most frequent among the identified clonal abnormalities.

Concurrent malignancy was found in 56 patients with SE, of which 35.7% had hematological malignancies (non-PV) and 64.2% had solid malignancies. Of the latter, 16.6% had breast cancer, and 13.8% had hepatocellular cancer. The most frequently identified causes of SE in our patients were idiopathic erythrocytosis (IE) in 56 (41.18%), chronic hypoxic states in 44 (32.4%), iatrogenic erythrocytosis (IaE) in 15 (11%), and autonomous EPO production in 10 (7.35%). (Table 1)

The World Health Organization, International Classification of Diseases, Clinical Modification (ICD-CM) code system, has simplified the extensive amount of information on medical records, improving the identification of diseases, and has an impact on quality-of-care assessment, allocation of resources, evaluation of management patterns and outcomes of diseases. The need for specificity after ICD 9^th edition due to the broadness of concepts was overcome with the establishment of the tenth revision in 2014. ¹⁰Despite the above, determining the extent of hematological diseases based solely on hospital discharge codes could be problematic.

Our report highlights the potential limitations of the use of ICD coding for its use in studies. Remarkably, 11% of the patients had a concurrent diagnosis code for PV and SE. Although the percentage of misdiagnosis in this study may appear to be modest, these chronic illnesses have a high cost on health expenses. An analysis made by Metha et al revealed an age-adjusted prevalence of 56.5 per 100 000 of patients who suffer from polycythemia vera and reported that in 2010, the annual cost reached $14 903 dollars of overall health care, outlining a considerable increase in comorbidities in this population.¹¹ However, these results, as their authors describe, may have limitations and biases for the same reason as this study. The impact of erroneous adjudication of diagnostic codes may have an impact on interventions using data from public health registries, surveillance and disease control at a population level. It has been proven that a correct diagnosis after profound data analysis could lead to a tremendous improvement in health costs. ¹²

The numerous variables, the high number of unexplained cases reported, as shown in a cross-sectional study from NHANES 2007-2008¹³ and the misidentification of erythrocytosis seen in some cases¹⁴, make data collection complex. Misidentification of these cases may be consistent with the low reported data rate of polycythemia cases at present. Larger-scale studies are needed to further identify the impact on healthy areas, as the outcomes of the above erythrocytosis research are hitherto unknown.

Challenges of diagnosis between etiologies of erythrocytosis (PV versus SE) could be clarified after JAK2 somatic mutation genotyping is performed, as it prevails as a determinant tool in diagnosis. ⁸ In our cohort, the prevalence of the JAK2 V617F mutation was estimated to be 1.5% among cases of SE. Interestingly, we found that 3.7% of the SE patients had other clonal abnormalities of germline origin in the JAK2 gene. Even though the significance of these mutations and the risk of developing PV or leukemia phenotypes are unknown, tools to predict functional effects, such as the DANN score, could establish a probable risk of pathogenicity in some of these mutations. ^15,16^] With the development of more sensitive molecular diagnostic techniques, the prevalence of clonal hematopoiesis abnormalities seems to have increased in recent years and explains (at least partially) the differences seen across cohorts, including ours. ¹⁷

While the number of patients without erythrocytosis was small in our study, the fact that bias can be introduced systematically in the identification of cases through disease code-based registries questions the validity of studies done with only disease codes as the strategy to identify cases.

This observation has been highlighted in previous publications, which increased the awareness of potential error sources and recommended code users to better evaluate the applicability and limitations of codes for their study of a particular disease or medical conditions. ¹⁸

In our cohort, on the other hand, the code described as IE was given if the possibility of any etiology of erythrocytosis, whether primary or secondary, was excluded, and no further explanation was found for the presence of this phenomenon. ¹⁹^]Consistent with the literature, aside from IE cases, the most commonly identified etiology of SE in this study was hypoxia, and approximately 24 (54.54%) patients had SE caused by obstructive sleep apnea (OSA).

The main limitations in our study relate to its retrospective design of selected cases referred to a cancer center. Moreover, our data analysis was restricted to patients who had JAK2 mutation molecular diagnostic studies. Furthermore, since several patients were evaluated as “second opinions” and their follow-up visits were carried out in various departments and medical centers, the etiology of erythrocytosis could not be established, accounting for approximately one-third of our cases. Another limitation includes the heterogeneity of the IE battery test used to evaluate secondary etiologies and the absence of follow-up data in several cases.

The use of ICD data capture modalities is critical to accurately identify specific populations of interest to conduct retrospective research; however, limitations are associated with possible misdiagnosed cases. Even though in our study the majority of the patients were correctly classified, individual medical chart review may be necessary until a better classification method is reached to decrease the possibility of introducing bias in such studies, especially in those cases with conflicting ICD diagnoses. Furthermore, the relationship of the codes with the true clinical diagnosis, as well as the identification and systematic management of dual codes in software healthcare systems, to decrease case misidentification should be a priority to improve public health and clinical studies. Research based exclusively on ICD codes could have a potential impact on public health and patient care, and limitations must be weighed when research findings are conveyed.

Funding source

The authors did not receive support from any organization for the submitted work.

Conflict of Interest

The authors do not have conflicts of interest to declare related to the content of this manuscript.

Availability of data and material

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Code availability

Not Applicable

Acknowledgments

None.

Authors contributions and disclosures

ABR, DDG, EF, LC, SV and CMRH contributed to the design and conceptualization of this study. LC performed the identification of the patients based on the IC code. EF, ABR, DDG and CMRH charts reviewed the medical records and reclassified the patients according to the WHO criteria for PV. All authors contributed to the writing and editing of the manuscript.

Ethics Approval

The study is a retrospective evaluation of the patients and involved minimal risk to subjects. The institutional review board at MD Anderson Cancer Center approved the study.

Consent for publication

Not applicable

Keohane, C., McMullin, M. F. & Harrison, C. The diagnosis and management of erythrocytosis. BMJ (Online), 347, https://doi.org/10.1136/bmj.f6667 (2013).
Patnaik, M. M. & Tefferi, A. The complete evaluation of erythrocytosis: Congenital and acquired., 23 (5), https://doi.org/10.1038/leu.2009.54 (2009).
Osler, W. Chronic cyanosis, with polycythaemia and enlarged spleen: a new clinical entity. 1903. The American journal of the medical sciences. 2008;335(6). doi:10.1097/MAJ.0b013e318175d13d
Barbui, T. et al. The 2016 WHO classification and diagnostic criteria for myeloproliferative neoplasms: document summary and in-depth discussion. Blood cancer journal, 8 (2), https://doi.org/10.1038/s41408-018-0054-y (2018).
Franke, K., Gassmann, M., Wielockx, B. & Erythrocytosis The HIF pathway in control., 122 (7), https://doi.org/10.1182/blood-2013-01-478065 (2013).
Hocking, W. G., Golde, D. W. & Polycythemia Evaluation and management. Blood Rev, 3 (1), https://doi.org/10.1016/0268-960X(89)90026-X (1989).
Mehta, J., Wang, H., Iqbal, S. U. & Mesa, R. Epidemiology of myeloproliferative neoplasms in the United States. Leukemia and Lymphoma, 55 (3), https://doi.org/10.3109/10428194.2013.813500 (2014).
Davila-Gonzalez, D. et al. Diagnostic Performance of Erythropoietin Levels in Polycythemia Vera: Experience at a Comprehensive Cancer Center. Clinical Lymphoma, Myeloma and Leukemia, https://doi.org/10.1016/j.clml.2020.11.002 (2020).
Wouters, H. J. C. M. et al. Erythrocytosis in the general population: Clinical characteristics and association with clonal hematopoiesis. Blood Advances, 4 (24), https://doi.org/10.1182/bloodadvances.2020003323 (2020).
Cartwright, D. J. ICD-9-CM to ICD-10-CM Codes: What? Why? How? Advances in Wound Care, 2 (10), https://doi.org/10.1089/wound.2013.0478 (2013).
Mehta, J., Wang, H., Fryzek, J. P., Iqbal, S. U. & Mesa, R. Health resource utilization and cost associated with myeloproliferative neoplasms in a large United States health plan. Leukemia and Lymphoma, 55 (10), https://doi.org/10.3109/10428194.2013.879127 (2014).
Pakhale, S., Sumner, A., Coyle, D., Vandemheen, K. & Aaron, S. (Correcting) misdiagnoses of asthma: A cost effectiveness analysis. BMC Pulmonary Medicine, 11, https://doi.org/10.1186/1471-2466-11-27 (2011).
Tremblay, D., Alpert, N., Taioli, E. & Mascarenhas, J. Prevalence of unexplained erythrocytosis and thrombocytosis–an NHANES analysis. Leukemia and Lymphoma, https://doi.org/10.1080/10428194.2021.1888377 (2021).
Guo, R., Liang, Y., Yan, L., Xu, Z. & Ren, J. Erythrocytosis caused by giant chromophobe renal cell carcinoma: a case report indicating a 9-year misdiagnosis of polycythemia vera. Chinese Journal of Cancer, 36 (1), https://doi.org/10.1186/s40880-017-0238-z (2017).
Benton, C. B. et al. Janus kinase 2 variants associated with the transformation of myeloproliferative neoplasms into acute myeloid leukemia., 125 (11), https://doi.org/10.1002/cncr.31986 (2019).
Lanikova, L. et al. Coexistence of gain-of-function JAK2 germ line mutations with JAK2V617F in polycythemia vera., 128 (18), https://doi.org/10.1182/blood-2016-04-711283 (2016).
Zapparoli, G. et al. Quantitative threefold allele-specific PCR (QuanTAS-PCR) for highly sensitive JAK2 V617F mutant allele detection. BMC Cancer, 13, https://doi.org/10.1186/1471-2407-13-206 (2013).
O’Malley, K. J. et al. Measuring diagnoses: ICD code accuracy. Health Serv. Res, 40 (5 II), https://doi.org/10.1111/j.1475-6773.2005.00444.x (2005).
McMullin, M. F. F. et al. A guideline for the management of specific situations in polycythaemia vera and secondary erythrocytosis: A British Society for Hematology Guideline. British Journal of Hematology, 184 (2), https://doi.org/10.1111/bjh.15647 (2019).

Table 1. Demographic and laboratory characteristics of 136 patients with a diagnosis of secondary erythrocytosis at MDACC

JAK-2 Mutation (V617F PCR) Status	No.	%
Positive	2	1.47
Negative	134	98.53
Characteristic	Median	IQR
Age (years)	56.2	66-49
Male %	57.7
Female %	42.3
Serum EPO value (mIU/mL)	19.3	8.8-22.6
Hematocrit %	49.1	24.1- 62
Hemoglobin (g/dL)	16.7	(8-20.70)
Etiology	Frequency	Percent
Idiopathic erythrocytosis	56	41.18
Hypoxia	44	32.35
Iatrogenic	15	11.03
Autonomous EPO	10	7.35
Multifactorial	6	4.41
Relative polycythemia	5	3.68
Total	136	100

IQR: Interquartile Range.

EPO: Erythropoietin.

No competing interests reported.

Download PDF

Editorial decision: Major revision
05 Jan, 2022
Reviews received at journal
21 Nov, 2021
Reviews received at journal
20 Nov, 2021
Reviewers agreed at journal
01 Nov, 2021
Reviewers agreed at journal
31 Oct, 2021
Reviewers invited by journal
30 Oct, 2021
Editor assigned by journal
30 Oct, 2021
Editor invited by journal
18 Oct, 2021
Submission checks completed at journal
18 Oct, 2021
First submitted to journal
06 Oct, 2021

You are reading this latest preprint version

Potential limitations of diagnostic standard codes to distinguish polycythemia vera and secondary erythrocytosis: A retrospective case series

Status:

Version 1

Abstract

Figures

Introduction

Methods

Results

Discussion

Declarations

References

Tables

Additional Declarations

Status:

Version 1