Peter Felding

Eosinophilia in routine blood samples and the subsequent risk of hematological malignancies and death

Eosinophilia may represent an early paraclinical sign of hematological malignant disease, but no reports exist on its predictive value for hematological malignancies. From the Copenhagen Primary Care Differential Count (CopDiff) Database, we identified 356,196 individuals with at least one differential cell count (DIFF) encompassing the eosinophil count during 2000–2007. From these, one DIFF was randomly chosen and categorized according to no (<0.5 × 109/L), mild (≥0.5–1.0 × 109/L) or severe (≥1.0 × 109/L) eosinophilia. From the Danish Cancer Registry and the Danish Civil Registration System, we ascertained hematological malignancies and death within 3 years following the DIFF. Using multivariable logistic regression odds ratios (ORs) were calculated and adjusted for previous eosinophilia in a DIFF, sex, age, year, month, C‐reactive protein, previous cancer, and comorbidity. ORs for developing Hodgkins lymphoma (HL) was significantly increased in individuals exhibiting severe eosinophilia, ORu2009=u20099.09 (C.I. 2.77–29.84), Pu2009=u20090.0003. The association with classical myeloproliferative neoplasms (cMPNs) showed an increasing risk with ORu2009=u20091.65 (1.04–2.61) Pu2009=u20090.0322 and ORu2009=u20093.87 (1.67–8.96) Pu2009=u20090.0016 for mild and severe eosinophilia. Eosinophilia was in a similar fashion associated with chronic lymphatic leukemia (CLL), ORu2009=u20092.57 (1.50–4.43), Pu2009=u20090.0006 and ORu2009=u20095.00 (1.57–15.94), Pu2009=u20090.0065, and all‐cause death, OR of 1.16 (1.09–1.24), Pu2009<u20090.0001 and 1.60 (1.35–1.91), Pu2009<u20090.0001. We confirm associations between eosinophilia and HL and cMPNs, and in addition for the first time demonstrate a dose‐dependent association between eosinophilia and CLL as well as death. Unexplained eosinophilia should prompt clinicians to consider conditions where early diagnosis may improve prognosis. Am. J. Hematol. 88:843–847, 2013.

The Copenhagen Primary Care Differential Count (CopDiff) database.

Background The differential blood cell count provides valuable information about a person’s state of health. Together with a variety of biochemical variables, these analyses describe important physiological and pathophysiological relations. There is a need for research databases to explore associations between these parameters, concurrent comorbidities, and future disease outcomes. Methods and results The Copenhagen General Practitioners’ Laboratory is the only laboratory serving general practitioners in the Copenhagen area, covering approximately 1.2 million inhabitants. The Copenhagen General Practitioners’ Laboratory has registered all analytical results since July 1, 2000. The Copenhagen Primary Care Differential Count database contains all differential blood cell count results (n=1,308,022) from July 1, 2000 to January 25, 2010 requested by general practitioners, along with results from analysis of various other blood components. This data set is merged with detailed data at a person level from The Danish Cancer Registry, The Danish National Patient Register, The Danish Civil Registration System, and The Danish Register of Causes of Death. Conclusion This paper reviews methodological issues behind the construction of the Copenhagen Primary Care Differential Count database as well as the distribution of characteristics of the population it covers and the variables that are recorded. Finally, it gives examples of its use as an inspiration to peers for collaboration.

Prevalence and clinical significance of neutropenia discovered in routine complete blood cell counts: a longitudinal study

Neutropenia, defined as an absolute blood neutrophil count (ANC) <1.5 G L−1, may accompany a variety of diseases. However, the clinical significance of neutropenia detected in a routine complete blood cell count is poorly understood.

The effectiveness of computer reminders for improving quality assessment for point-of-care testing in general practice—a randomized controlled trial

BackgroundComputer reminders are increasingly being applied in efforts to improve quality and patient safety. However, research is still needed to establish the effectiveness of different kinds of reminders in various settings. This study aimed to evaluate the effectiveness of computer reminders for improving adherence to a quality assessment scheme for point-of-care testing in general practice.MethodThe study was conducted as a randomized controlled crossover trial among general practices in the Capital Region of Denmark. The intervention consisted of sending computer reminders (ComRem) to practices not adhering to the guideline recommendations of split testing for hemoglobin and glucose. Practices were randomly allocated into two groups. During the first follow-up period, one of the groups received the ComRem intervention together with the general implementation activities (GIA), while the other group only received the GIA. For the second follow-up period, the intervention was switched between the two groups. Outcomes were measured as split test procedure adherence.ResultsA total of 142 practices were randomly allocated to the early intervention group and 144 practices to the late intervention group (the control group in the first follow-up period). In the first intervention period, the mean number of split tests performed in the group receiving ComRem group increased from 1.22 to 3.76 (out of eight possible tests) while the mean number of split tests increased from 1.11 to 2.35 in the group targeted by GIA only (p = 0.0059). After the crossover, a similar effect of reminders was observed. Furthermore, the developments in outcome measures over time showed a strong effect of computer reminders beyond the intervention periods.ConclusionThere was a significant effect of computer reminders on adherence to the quality assessment scheme for point-of-care testing. Thus, computer reminders seem to be useful for supporting the implementation of relatively simple procedures for quality and safety.Trial registrationClinicalTrials.gov: http://NCT01152177

Is thrombocytosis a valid indicator of advanced stage and high mortality of gynecological cancer

Abstract Objective Thrombocytosis has been associated with higher stage and mortality of cancer, however, the evidence is conflicting. We examined the stage distribution and prognosis of gynecologic cancer according to levels of prediagnostic platelet count. Methods In a primary care resource with blood cell counts from more than 500,000 individuals, we identified 581 women with a primary diagnosis of gynecological cancer. We divided the pre-diagnostic mean platelet count derived from the 3-year period prior to cancer diagnosis into three categories of thrombocytosis (no, 150–400×10 9 /L; mild, >400–550×10 9 /L; severe, >550×10 9 /L). Logistic regression models were used to calculate odds ratios (ORs) for the association of prediagnostic platelet counts with stage at diagnosis. Subsequently, we estimated hazard ratios (HRs) for all-cause or gynecological cancer-specific mortality by level of thrombocytosis using Cox proportional hazard regression models. Results Patients with non-localized disease had higher levels of prediagnostic platelet count [mild thrombocytosis: OR, 2.36 (95% CI, 1.33–4.19); severe thrombocytosis: 4.54 (95% CI, 1.55–13.3); compared with no prediagnostic thrombocytosis]. The median overall survival was 1.04years among patients with severe prediagnostic thrombocytosis and 3.25years among those with mild thrombocytosis, P Conclusions Prediagnostic thrombocytosis was associated with advanced stage of gynecological cancer at diagnosis and increased all-cause and cancer-specific mortality. The platelet count may have an important role in diagnosis and post-diagnostic control of gynecological cancer.

The efficacy of computer reminders on external quality assessment for point-of-care testing in Danish general practice: rationale and methodology for two randomized trials

AbstractBackgroundPoint-of-care testing (POCT) is increasingly being used in general practice to assist general practitioners (GPs) in their management of patients with diseases. However, low adherence to quality guidelines in terms of split test procedures has been observed among GPs in parts of the Capital Region in Denmark. Computer reminders embedded in GPs electronic medical records (ComRem) may facilitate improved quality control behaviour, but more research is needed to identify what types of reminders work and when. The overall aim of this study is to evaluate the efficacy of ComRem to improve GPs adherence to quality guidelines. This article describes the rationale and methods of the study that constitute this research project.Methods/designThe study is conducted as two randomised controlled trials (RCTs) among general practices in two districts of the Capital Region in Denmark. These districts contain a total of 739 GPs in 567 practices with a total of 1.1 million patients allocated to practice lists. In the first RCT (RCT A), ComRem is compared to postal reminder letters. In the second RCT (RCT B), ComRem is compared to usual activities (no reminders) with a crossover approach. In both of these studies, outcomes are measured by the number of split tests received by the laboratory.ConclusionsThis study will contribute to knowledge on the efficacy of ComRem in primary care. Because the study does not explore GPs perceptions and experiences with regard to ComRem, we will subsequently conduct a qualitative survey focusing on these aspects.Trial registrationsStudy A: ClinicalTrials.gov identifier: NCT01152151nStudy B: ClinicalTrials.gov identifier: NCT01152177

The effectiveness of computer reminders versus postal reminders for improving quality assessment for point-of-care testing in primary care: a randomized controlled trial

RATIONALE, AIMS AND OBJECTIVESnThis study aimed to evaluate the relative effectiveness of electronic and postal reminders for increasing adherence to the quality assurance programme for the international normalized ratio (INR) point-of-care testing (POCT) device in primary care.nnnMETHODSnAll 213 family practices that use the Elective Laboratory of the Capital Region, Denmark, and regularly conduct INR POCT were randomly allocated into two similarly sized groups. During the 4-month intervention, these practices were sent either computer reminders (ComRem) or computer-generated postal reminders (Postal) if they did not perform a split test to check the quality of their INR POCT for each calendar month. The adherence of the practices was tracked during the subsequent 8 months subdivided into two 4-month periods both without intervention. Outcomes were measures of split test procedure adherence.nnnRESULTSnBoth interventions were associated with an increase in adherence to the split test procedure - a factor 6.00 [95% confidence interval (CI) 4.46-7.72] and 8.22 [95% CI 5.87-11.52] for ComRem and Postal, respectively - but there is no evidence that one of the interventions was more effective than the other. In the ComRem group, the expected number of split tests (out of four) was 2.54 (95% CI 2.33-2.76) versus 2.44 (95% CI 2.24-2.65) in the Postal group, Pu2009=u20090.14. There was a slight decrease in adherence over the two follow-ups, but neither intervention was better than the other in achieving a lasting improvement in adherence.nnnCONCLUSIONnComputer reminders are as efficient as postal reminders in increasing adherence to a quality assurance programme for the INR POCT device in primary care.

Association of the blood eosinophil count with hematological malignancies and mortality.

Blood eosinophilia (≥0.5 × 109/l) may be an early sign of hematological malignancy. We investigated associations between levels of blood eosinophils and risks of hematological malignancies and mortality in order to provide clinically derived cut‐offs for referral to specialist hematology care. From the Copenhagen Primary Care Differential Count (CopDiff) Database, we identified 356,196 individuals with at least one differential cell count encompassing the eosinophil count during 2000–2007 and matched these laboratory data with Danish nationwide health registers. We used multivariable logistic regression to calculate odds ratios (ORs) for the 4‐year incidences of hematological malignancies and mortality between the eosinophil counts and a reference count of 0.16 × 109/l which was the median eosinophil count in our data. Risks of hematological malignancies and mortality increased above the median eosinophil count. At the 99th percentile, corresponding to an eosinophil count of 0.75 × 109/l, risks of hematological malignancies were increased more than twofold with OR (95% C.I.) of 2.39 (1.91–2.99). Interestingly, risks reached a plateau around an eosinophil count of 1.0 × 109/l. Risks also increased when the eosinophil count approached zero. Here, counts associated relatively more with acute myeloid leukemia and myelodysplastic syndromes whereas counts above 0.16 × 109/l associated more with myeloproliferative neoplasms. Eosinophil counts associate with hematological malignancies and mortality even below the definition of eosinophilia. The observed plateau of risks around 1.0 × 109/l is important for physicians encountering patients with eosinophilia since even mild‐to‐moderate eosinophilia according to traditional definitions confers maximally increased risks of subsequent/subclinical hematological malignancy. Am. J. Hematol. 90:225–229, 2015.

FRI0113 Rheumatoid arthritis and eosinophilia: the risk of lymphoproliferative malignancies and solid cancers. a study based on the copenhagen primary care differential count (copdiff) database

Background Cohort and population studies have indicated an association between rheumatoid arthritis (RA) and malignancies. Several mechanisms have been suggested; i) extrinsic pro-oncogenic effects of disease-modifying antirheumatic drugs, ii) intrinsic pro-oncogenic effects of disease activity and iii) surveillance bias. No study has been able to disentangle the possible effects of disease activity on lymphoma risk or identify the promoting factors. In this context, the eosinophil is an interesting candidate. Blood eosinophiliaoccurs in RA in outpatients with an estimated prevalence of 7.7% [1]and is linked to prognosis and severity of extra-articular manifestations [2]. Importantly, eosinophilia per se is associated with certain lymphoproliferative malignancies (unpublished data - CopDiff database to be presented at EULAR) Objectives To investigate whether RA is associated with risk of lymphoproliferative malignancies or solid cancers when adjusting for the potential mediator eosinophilia and several known confounders in a large primary care cohort largely unaffected by surveillance bias Methods From the Copenhagen Primary Care Differential Count (CopDiff)Database, we identified 359.950 individuals with a differential cell count (DIFF) during 2000-2007. From these individuals one DIFF was chosen at random. By linkage to the Danish National Patient Register, we identified baseline diagnoses of RA and categorized these according to time before baseline. From the Danish Cancer Registry we ascertained malignancies within three years following the DIFF. Odds ratios (ORs) and 95% confidence intervals (CIs) were computed using multivariate logistic regression and were adjusted for eosinophilia, sex, age, year, month, CRP and Charlson´s comorbidity index (CCI) Results 921 persons had recent onset (<3 years) RA and 2.583 had longer disease duration (≥3 years) RA. After multivariate adjustment, RA was not associated with incident lymphoproliferative malignancies. Compared with patients without RA, ORs (95% CI) for lymphoproliferative malignancies were 0.95 (0.39-2.32) for longer disease duration and 0.69 (0.10-5.00) for recent onset. Similarly, RA was not associated with an increased risk of incident solid cancers (ORs of 1.11 (0.89-1.38) and 0.76 (0.48-1.20), respectively). These risk estimates did not alter when eosinophilia, CRP and comorbidity were excluded from the models Conclusions RA with or without eosinophilia was not associated with an increased risk of lymphoproliferative or solid cancersduring 3 years of follow-upin a primary care population. In contrast to previous studies, this cohort is largely unaffected by surveillance bias References Panush, R.S. et al. Ann Intern Med, 1971. 75(2): p. 199-205 Kargili, A. et al. Rheumatol Int, 2004. 24(6): p. 321-4 Disclosure of Interest C. Lykkegaard Andersen: None Declared, H. Vestergaard Grant/research support from: BMS, Novartis, O. Bjerrum: None Declared, V. Siersma: None Declared, P. Felding: None Declared, H. Hasselbalch: None Declared, N. de Fine Olivarius: None Declared, H. Lindegaard Grant/research support from: Roche, MSD, BMS, Merck

Eosinophilia in routine blood samples as a biomarker for solid tumor development - A study based on the Copenhagen Primary Care Differential Count (CopDiff) Database.

Abstract Background. Eosinophilia may represent an early paraclinical sign of malignant disease and a host anti-tumor effect. The association between eosinophilia and the development of solid tumors has never before been examined in an epidemiological setting. The aim of the present study was to investigate eosinophilia in routine blood samples as a potential biomarker of solid tumor development in a prospective design. Material and methods. From the Copenhagen Primary Care Differential Count (CopDiff) Database, we identified 356 196 individuals with at least one differential cell count (DIFF) encompassing the eosinophil count during 2000–2007. From these, one DIFF was randomly chosen and categorized according to no (< 0.5 × 109/l), mild (≥ 0.5–1.0 × 109/l) or severe (≥ 1.0 × 109/l) eosinophilia. From the Danish Civil Registration System and the Danish Cancer Registry we ascertained all-cause death and solid tumors within the first three years following the DIFF. Using multivariable logistic regression, odds ratios (OR) were calculated and adjusted for previous eosinophilia, sex, age, year, month, C-reactive protein, previous cancer and Charlsons Comorbidity Index. Results. The risk of bladder cancer was increased with mild eosinophilia [OR 1.93 (CI 1.29–2.89), p = 0.0013]. No associations with eosinophilia were observed for the remaining solid cancers. Conclusion. We demonstrate that eosinophilia in routine blood samples associates with an increased risk of bladder cancer. Our data emphasize that additional preclinical studies are needed in order to shed further light on the role of eosinophils in carcinogenesis, where it is still unknown whether the cells contribute to tumor immune surveillance or neoplastic evolution.