Statistical methods used in the calculation of geriatric reference intervals: a systematic review.

BACKGROUND: Geriatric reference intervals (RIs) are not commonly available and are rarely used. It is difficult to select a reference population from a cohort with a high degree of morbidity. Also important are the statistical approaches used to determine health-associated reference values. It is the aim of this study to examine the statistical methods used in the calculation of geriatric RIs. METHODS: A search was conducted on EMBASE and Medline for articles between January 1989 and January 2014. Studies were selected if they: 1) were English primary articles; 2) performed a clinical chemistry test on a blood fraction; 3) had a population sub-group consisting of individuals ≥65 years of age; and 4) calculated a RI for the subgroup ≥65 years of age. RESULTS: There were 64 articles identified, of which 78.1% described the RI calculation method used. RI calculation was performed by non-parametric (21.9%), parametric (42.2%), robust (3.1%), or other (17.2%) methods. Outlier detection (SD, Grubb's test, Tukey's fence, Dixon) was infrequently used and although most studies performed partitioning, only 57.8% tested the statistical significance of the partitions. Few studies (17.2%) reported confidence intervals for the RI estimates. Overall, only 14.1% of studies provided RI estimates which followed the CLSI guideline EP28-A3c. CONCLUSIONS: Statistical methods for RI calculation and partitioning varied considerably between studies and many failed to provide adequate descriptions of these methods. Challenges in analyses arose from insufficient sample sizes and heterogeneity in the elderly population. Geriatric RIs, although present in the literature, may not be properly calculated and should be carefully considered before applying them for clinical care.

Chapter 9: options for summarizing medical test performance in the absence of a "gold standard".

The classical paradigm for evaluating test performance compares the results of an index test with a reference test. When the reference test does not mirror the "truth" adequately well (e.g. is an "imperfect" reference standard), the typical ("naïve") estimates of sensitivity and specificity are biased. One has at least four options when performing a systematic review of test performance when the reference standard is "imperfect": (a) to forgo the classical paradigm and assess the index test's ability to predict patient relevant outcomes instead of test accuracy (i.e., treat the index test as a predictive instrument); (b) to assess whether the results of the two tests (index and reference) agree or disagree (i.e., treat them as two alternative measurement methods); (c) to calculate "naïve" estimates of the index test's sensitivity and specificity from each study included in the review and discuss in which direction they are biased; (d) mathematically adjust the "naïve" estimates of sensitivity and specificity of the index test to account for the imperfect reference standard. We discuss these options and illustrate some of them through examples.

Chapter 8: meta-analysis of test performance when there is a "gold standard".

Synthesizing information on test performance metrics such as sensitivity, specificity, predictive values and likelihood ratios is often an important part of a systematic review of a medical test. Because many metrics of test performance are of interest, the meta-analysis of medical tests is more complex than the meta-analysis of interventions or associations. Sometimes, a helpful way to summarize medical test studies is to provide a "summary point", a summary sensitivity and a summary specificity. Other times, when the sensitivity or specificity estimates vary widely or when the test threshold varies, it is more helpful to synthesize data using a "summary line" that describes how the average sensitivity changes with the average specificity. Choosing the most helpful summary is subjective, and in some cases both summaries provide meaningful and complementary information. Because sensitivity and specificity are not independent across studies, the meta-analysis of medical tests is fundamentaly a multivariate problem, and should be addressed with multivariate methods. More complex analyses are needed if studies report results at multiple thresholds for positive tests. At the same time, quantitative analyses are used to explore and explain any observed dissimilarity (heterogeneity) in the results of the examined studies. This can be performed in the context of proper (multivariate) meta-regressions.

Utility of urine myoglobin for the prediction of acute renal failure in patients with suspected rhabdomyolysis: a systematic review.

BACKGROUND: Urine myoglobin continues to be used as a marker of rhabdomyolysis, particularly to assess risk of developing acute renal failure and evaluate treatment success. We sought to determine the predictive validity of urine myoglobin (uMb) for acute renal failure (ARF) in patients with suspected rhabdomyolysis. METHODS: We performed a broad systemic review of the literature from January 1980 to December 2006 using the search terms myoglobin$ AND (renal OR ARF OR kidney). Only primary studies published in English where uMb measurement was related to ARF were included. RESULTS: Of 1602 studies screened, 52 met all selection criteria. The studies covered a wide spectrum of etiologies for rhabdomyolysis, dissimilar diagnostic criteria for ARF and rhabdomyolysis, and various methods of uMb measurement and were mostly case series (n = 32). There was poor reporting on the uMb method, and 17 studies failed to provide any information about the method. The reporting of clinical criteria for ARF with respect to timing, description, performance, and interpretation also lacked adequate detail for replication. Eight studies (total 295 patients) had data for 2-by-2 tables. Sensitivity of the uMb test was 100% in 5 of the 8 studies, specificity varied widely (15% to 88%), and CIs around these measures were high. Pooling of data was not possible because of study heterogeneity. CONCLUSIONS: There is inadequate evidence evaluating the use of uMb as a predictor of ARF in patients with suspected rhabdomyolysis.

Feasibility of biological specimen collection for the Canadian longitudinal study on aging (CLSA) biorepository.

ABSTRACTBiological specimen collection is an integral part of many longitudinal epidemiological studies. It is important to achieve high participant satisfaction for continuing involvement, and high sample quality for accurate biomarker measurement. We conducted a study to evaluate these issues on the sample collection proposed for the Canadian Longitudinal Study on Aging (CLSA). There were 85 participants recruited, and 65 attended either a hospital laboratory or private laboratory. Approximately 100 mL of blood and a random urine specimen were collected from each participant for a total of 2,108 sample aliquots. Quality standards were met for more than 90 per cent of samples and were similar for samples collected in both laboratories. More than 90 per cent of participants rated satisfaction with the collection as being good or excellent, and 84 per cent would be willing to repeat the collection in one to three years.

The Canadian longitudinal study on aging (CLSA).

ABSTRACTCanadians are living longer, and older persons are making up a larger share of the population (14% in 2006, projected to rise to 20% by 2021). The Canadian Longitudinal Study on Aging (CLSA) is a national longitudinal study of adult development and aging that will recruit 50,000 Canadians aged 45 to 85 years of age and follow them for at least 20 years. All participants will provide a common set of information concerning many aspects of health and aging, and 30,000 will undergo an additional in-depth examination coupled with the donation of biological specimens (blood and urine). The CLSA will become a rich data source for the study of the complex interrelationship among the biological, physical, psychosocial, and societal factors that affect healthy aging.

Physiological, pathological, pharmacological, biochemical and hematological factors affecting BNP and NT-proBNP.

OBJECTIVES: To gain a better understanding of the scope and breadth of factors associated with the B-type natriuretic peptides. DESIGN AND METHODS: Databases were searched from 1989 to February 2005 for primary studies that measured BNP or NT-proBNP for the purpose of diagnosis, prognosis and monitoring treatment. RESULTS: There were 103 factors identified in 72 studies. Most of the cardiac diseases were positively associated with BNP and NT-proBNP concentrations and of the non-cardiac conditions, dyspnea, diabetic nephropathy, and stroke were all associated with higher concentrations. Most biochemical and hematological markers showed positive associations. Factors that assessed heart function showed both positive and negative associations and drug therapy was either negatively associated or had no effect on BNP or BNT-proBNP concentrations. Few studies reported independent associations and of those that did age, female gender, and creatinine concentrations were positively associated with BNP and NT-proBNP. CONCLUSIONS: Various factors were found to be associated with BNP and NT-proBNP.

Diagnostic accuracy of BNP and NT-proBNP in patients presenting to acute care settings with dyspnea: a systematic review.

OBJECTIVE: We sought to compare the diagnostic performance of B-type natriuretic peptide (BNP) and N-terminal proBNP measurements in patients presenting to acute care settings with dyspnea, a common presenting symptom of heart failure. DESIGN AND METHODS: We conducted a systematic review of the literature. For all included studies, we applied the QUADAS 14-question quality assessment tool for systematic reviews of diagnostic accuracy and abstracted the data for every published cut point. RESULTS: We screened 4338 studies and included nine in the meta-analysis. All 9 studies scored positively on at least 50% of the QUADAS questions. The pooled estimates of sensitivity and specificity were the same for the BNP studies (0.97 (95% CI: 0.96, 0.98) and 0.70 (95% CI: 0.56, 0.85)) as for the NT-proBNP studies (0.95 (95% CI: 0.90, 1.01) and 0.72 (95% CI: 0.53, 0.90)). Tests for heterogeneity were significant in both subgroups: BNP (I(2)=97.9%, p<0.001) and NT-proBNP (I(2)=87.5%, p<0.001). Similar overall results were found for the likelihood and diagnostic odds ratios. CONCLUSIONS: BNP and NT-proBNP have very similar diagnostic performance characteristics and can be used to rule out heart failure as a cause of dyspnea in the acute clinical setting. However, there is no easily identifiable optimum cut point value for each peptide.

A systematic review of BNP as a predictor of prognosis in persons with coronary artery disease.

OBJECTIVE: This systematic review was conducted to examine whether B-type natriuretic peptide (BNP) can predict mortality and other cardiac endpoints in persons diagnosed with coronary artery disease (CAD). DESIGN AND METHODS: Databases were searched from 1989 to February 2005 for primary studies that measured BNP for the purpose of diagnosis, prognosis, and monitoring treatment. RESULTS: In 18 studies, concentrations of BNP were found to have consistent positive associations with poorer prognoses for persons with CAD. The overall range of effect (95% confidence interval) was 2.31 to 5.02, measured via a random effects meta-analysis on studies reporting an odds ratio. Prognostic ability was similar for mortality and non-fatal outcomes. Ranges of estimated measures of effect (i.e., odds ratio, relative risk, hazard ratio) were concentrated between 1.33 to 2.94 for mortality and 1.01 to 3.03 for non-fatal outcomes. CONCLUSIONS: Further research is needed to assess whether prognostic ability differs by comorbidity or concomitant treatment. As well, the importance and selection of cut points remains unresolved. Until greater clarity is given to these matters, it would be prudent for clinicians to employ caution when using concentrations of BNP to predict the prognosis of persons with CAD.

Evidence for the use of B-type natriuretic peptides for screening asymptomatic populations and for diagnosis in primary care.

OBJECTIVES: To determine the screening and diagnostic properties of BNP and NT-proBNP for heart failure in primary care. DESIGN AND METHODS: We conducted a systematic review of randomized control trials and observational (cohort or case-control) studies of heart failure detection using B-type natriuretic peptides published in English from January 1989 to February 2005. We extracted or calculated sensitivity, specificity, positive and negative likelihood ratios, area under the receiver-operator characteristic curve and diagnostic odds ratio (DOR). RESULTS: We included 17 studies (7 screening, 9 diagnosis in primary care or specialised clinic, 1 both). There was considerable heterogeneity within the study populations, reference standard for diagnosis, and B-type natriuretic peptide decision point. Sensitivity ranged from 26% to 98%; and specificity from 44% to 88%. For screening, the Diagnostic Odds Ratio (DOR) ranged from 2.7 to 29, and for diagnosis from 2.8 to 137. CONCLUSIONS: The performance characteristics of B-type natriuretic peptides measurement are not suitable for screening asymptomatic patients. For diagnosis in primary care, low B-type natriuretic peptide values may be used to rule-out heart failure but, due to poor specificity, high values cannot be used to rule-in the condition.

Monitoring the response to pharmacologic therapy in patients with stable chronic heart failure: is BNP or NT-proBNP a useful assessment tool?

OBJECTIVES: B-type natriuretic peptides are biomarkers of heart failure (HF) that can decrease following treatment. We sought to determine whether B-type natriuretic peptide (BNP) or N-terminal proBNP (NT-proBNP) concentration changes occurred in parallel to changes in other measures of heart failure following treatment. METHODS: We conducted a systematic review of the literature for studies that assessed B-type natriuretic peptide measurements in treatment monitoring of patients with stable chronic heart failure. Selected studies had to include at least three consecutive measurements of BNP or NT-proBNP. RESULTS: Of 4338 citations screened, only 12 met all of the selection criteria. The selected studies included populations with a wide range of heart failure severity and therapy. BNP and NT-proBNP decreased following treatment in nine studies and was associated with improvement in clinical measures of HF. CONCLUSIONS: There was limited data to support using BNP or NT-proBNP to monitor therapy in patients with HF.

Candidate recommendations for protein electrophoresis reporting from the Canadian Society of Clinical Chemists Monoclonal Gammopathy Working Group.

Protein electrophoresis is commonly used as an aid in the diagnosis of monoclonal gammopathies and is performed in many laboratories in Canada and throughout the world. However, unlike many other diagnostic tests, there is limited guidance for standardization and neither guidance nor specific recommendations for clinical reporting of serum (SPE) or urine (UPE) protein electrophoresis and immunotyping available in the literature. Therefore, a Canadian effort was undertaken to recommend standards that cover all aspects of clinical reporting with an ultimate goal towards reporting standardization. The Canadian Society of Clinical Chemists (CSCC) Monoclonal Gammopathy Interest Group (MGIG), which is composed of CSCC members with an interest in protein electrophoresis, has formed a Monoclonal Gammopathy Working Group (MGWG) to take initial steps towards standardization of SPE, UPE and immunotyping. Candidate standardization recommendations were developed, discussed and voted upon by the MGWG. Candidate recommendations that achieved 90% agreement are presented as consensus recommendations. Recommendations that did not achieve 90% consensus remain candidate recommendations and are presented with accompanying MGWG discussion. Eleven consensus recommendations along with candidate recommendations for nomenclature, protein fraction reporting, test utilization, interference handling and interpretive reporting options are presented.

Brain type carnosinase in dementia: a pilot study.

BACKGROUND: The pathological processes underlying dementia are poorly understood and so are the markers which identify them. Carnosinase is a dipeptidase found almost exclusively in brain and serum. Carnosinase and its substrate carnosine have been linked to neuropathophysiological processes. METHODS: Carnosinase activity was measured by a flourometric method in 37 patients attending a Geriatric Outpatient Clinic. There were 17 patients without dementia, 13 had Alzheimer's disease (AD) and 7 had mixed dementia (MD). RESULTS: The range of serum carnosinase activity for patients without dementia was 14.5 - 78.5 micromol/ml/h. There was no difference in carnosinase activity between patients without dementia (40.3 +/- 15.2 micromol/ml/h) and patients with AD (44.4 +/- 12.4 micromol/ml/h) or MD (26.6 +/- 15 micromol/ml/h). However, levels in the MD group were significantly lower than the AD group (p = 0.01). This difference remained significant after adjusting for gender, MMSE score, exercise, but not age, one at a time and all combined. The effect of other medical conditions did not remove the significance between the AD and MD groups. The MD group, but not the AD group, demonstrated a significant trend with carnosinase activity decreasing with duration of disease (from first recorded date of diagnosis to date of blood collection) (r = -0.76, p = 0.049). There was no association with carnosinase activity and MMSE score in the AD or MD group. Both AD and MD patients on any dementia medication (donepezil, galantamine, memantine) had higher carnosinase activity compared to those not taking a dementia medication. Carnosinase activity was higher in patients who regularly exercised (n = 20) compared to those who did not exercise regularly (n = 17)(p = 0.006). CONCLUSION: This exploratory study has shown altered activities of the enzyme carnosinase in patients with dementia.

National Survey of Adult and Pediatric Reference Intervals in Clinical Laboratories across Canada: A Report of the CSCC Working Group on Reference Interval Harmonization.

OBJECTIVE: Reference intervals are widely used decision-making tools in laboratory medicine, serving as health-associated standards to interpret laboratory test results. Numerous studies have shown wide variation in reference intervals, even between laboratories using assays from the same manufacturer. Lack of consistency in either sample measurement or reference intervals across laboratories challenges the expectation of standardized patient care regardless of testing location. Here, we present data from a national survey conducted by the Canadian Society of Clinical Chemists (CSCC) Reference Interval Harmonization (hRI) Working Group that examines variation in laboratory reference sample measurements, as well as pediatric and adult reference intervals currently used in clinical practice across Canada. DESIGN AND METHODS: Data on reference intervals currently used by 37 laboratories were collected through a national survey to examine the variation in reference intervals for seven common laboratory tests. Additionally, 40 clinical laboratories participated in a baseline assessment by measuring six analytes in a reference sample. RESULTS: Of the seven analytes examined, alanine aminotransferase (ALT), alkaline phosphatase (ALP), and creatinine reference intervals were most variable. As expected, reference interval variation was more substantial in the pediatric population and varied between laboratories using the same instrumentation. Reference sample results differed between laboratories, particularly for ALT and free thyroxine (FT4). Reference interval variation was greater than test result variation for the majority of analytes. CONCLUSION: It is evident that there is a critical lack of harmonization in laboratory reference intervals, particularly for the pediatric population. Furthermore, the observed variation in reference intervals across instruments cannot be explained by the bias between the results obtained on instruments by different manufacturers.

Reproducibility of impaired glucose tolerance (IGT) and impaired fasting glucose (IFG) classification: a systematic review.

BACKGROUND: The classifications of impaired glucose tolerance (IGT) and impaired fasting glucose (IFG) represent glucose levels above normal, but below the decision threshold for diabetes. We sought to determine what the reproducibility of these classifications was when repeat tests were performed by conducting a systematic review of the literature. METHODS: All primary studies published in English of any study design were included. Studies were excluded if they did not follow the World Health Organization or American Diabetes Association diagnostic criteria, used whole blood as the specimen type, a glucose meter for analysis, or performed repeat testing greater than 8 weeks apart. RESULTS: Five papers had reproducibility data for IGT or IFG, two of which where from the same population but sampled differently. The kappa coefficients, indicating agreement between repeat tests that exceeded chance, indicated poor to fair agreement for IGT (0.04, 0.22, 0.38, 0.42) and moderate agreement for IFG (0.44 and 0.56). Similarly, the observed reproducibility was slightly lower for IGT (33%, 44%, 47%, 48%) compared to IFG (51%, 64%). In two studies for which data were available for both IGT and IFG, the average reproducibility was lower (49%) for the prediabetes group compared to the diabetes group (73%) or the normal group (93%). CONCLUSIONS: Poor reproducibility of IGT and IFG classification suggests caution should be exercised when interpreting a single test result.