Models to estimate biological variation components and interpretation of serial results: strengths and limitations

AbstractBiological variation (BV) has multiple applications in a variety of fields of clinical laboratory. The use of BV in statistical modeling is twofold. On the one hand, some models are used for the generation of BV estimates (within- and between-subject variability). Other models are built based on BV in combination with other factors to establish ranges of normality that will help the clinician interpret serial results for the same subject. There are two types of statistical models for the calculation of BV estimates: A. Direct methods, prospective studies designed to calculate BV estimates; i. Classic model: developed by Harris and Fraser, revised by the Working Group on Biological Variation of the European Federation of Laboratory Medicine. ii. Mixed-effect models. iii. Bayesian model. B. Indirect methods, retrospective studies to derive BV estimates from large databases of results. Big data. Understanding the characteristics of these models is crucial as they determine their applicability in different settings and populations. Models for defining ranges that help in the interpretation of individual serial results include: A. Reference change value and B. Bayesian data network. In summary, this review provides an overview of the models used to define BV components and others for the follow-up of patients. These models should be exploited in the future to personalize and improve the information provided by the clinical laboratory and get the best of the resources available.

Download Full-text

Weight loss and incident dementia in elderly Yoruba Nigerians: a 10-year follow-up study

International Psychogeriatrics ◽

10.1017/s1041610210001390 ◽

2010 ◽

Vol 23 (3) ◽

pp. 387-394 ◽

Cited By ~ 18

Author(s):

Adesola Ogunniyi ◽

Su Gao ◽

Frederick W. Unverzagt ◽

Olusegun Baiyewu ◽

Oyewusi Gureje ◽

...

Keyword(s):

Weight Loss ◽

Normal Subjects ◽

Smoking History ◽

Close Monitoring ◽

Cognitive Assessments ◽

Mixed Effect ◽

Mixed Effect Models ◽

Dementia Risk ◽

Incident Dementia

ABSTRACTBackground: The relationship between weight and dementia risk has not been investigated in populations with relatively low body mass index (BMI) such as the Yoruba. This study set out to achieve this objective using a prospective observational design.Methods: The setting was Idikan Ward in Ibadan City, Nigeria. The participants were all aged 65 years or older and were enrolled in the Indianapolis-Ibadan Dementia Project. Repeated cognitive assessments and clinical evaluations were conducted to identify participants with dementia or MCI during 10 years of follow-up (mean duration: 5.97 years). BMI measures, information on alcohol, smoking history, cancer, hypertension, diabetes, heart attack, stroke and depression were collected at each follow-up evaluation. Mixed effect models adjusted for covariates were used to examine the differences in BMI among participants who developed dementia or MCI and those who remained cognitively normal during the follow-up.Results: This analysis included 1559 participants who had no dementia at their first BMI measurements. There were 136 subjects with incident dementia, 255 with MCI and 1168 with normal cognition by the end of the study. The mean BMI at baseline was higher for female participants (22.31; SD = 4.39) than for male (21.09; SD = 3.61, p < 0.001). A significantly greater decline in BMI was found in those with either incident dementia (p < 0.001) or incident MCI (p < 0.001) compared to normal subjects.Conclusion: Decline in BMI is associated with incident MCI and dementia in elderly Yoruba. This observation calls for close monitoring of weight loss in elderly individuals which may indicate future cognitive impairment for timely detection and tailored interventions.

Download Full-text

Twins in Guinea-Bissau have a 'thin-fat' body composition compared to singletons

10.1101/2021.07.07.21260161 ◽

2021 ◽

Author(s):

Rucha H Wagh ◽

Morten Bjerregaard-Anderson ◽

Souvik Bandopadhyay ◽

Pranav Yajnik ◽

Rashmi B Prasad ◽

...

Keyword(s):

Body Composition ◽

Early Life ◽

Fat Body ◽

Mixed Effect ◽

Guinea Bissau ◽

Mixed Effect Models ◽

Restricted Environment ◽

Health Project ◽

Lower Height

Undernourished Indian babies are paradoxically more adipose compared to European babies. In utero, twins are in a growth restricted environment and share nutritional sources, therefore, they might have a 'thin-fat' body composition compared to singletons. We compared anthropometry of twins (n=209, 97 males) and singletons (n=182, 86 males), using linear mixed-effect models in the Guinea-Bissau Twin Registry at the Bandim Health Project. Twins had lower birth weight (2420 vs 3100 g, p<0.001); and at follow-up, lower height (HAZ mean Z-score difference, -0.21, p=0.055), weight (WAZ -0.73, p=0.024) and BMI (BAZ -0.22, p=0.079) compared to singletons but higher adiposity (skinfolds: 0.33, p=0.001). Twins also had higher fasting (0.38, p<0.001) and 2-hr OGTT glucose concentrations (0.29, p<0.05). Data on prematurity was not available in this cohort. The thin-fat phenotype of Guinea-Bissau twins could suggest that it is a manifestation of early life undernutrition and is not exclusive to Indians. Keywords: Twins, adiposity, thin-fat, early life undernutrition, Guinea-Bissau.

Download Full-text

Learning and cognitive fatigue trajectories in multiple sclerosis defined using a burst measurement design

Multiple Sclerosis Journal ◽

10.1177/1352458513477922 ◽

2013 ◽

Vol 19 (11) ◽

pp. 1518-1525 ◽

Cited By ~ 9

Author(s):

Roee Holtzer ◽

Frederick Foley ◽

Vanessa D’Orio ◽

Jessica Spat ◽

Melissa Shuman ◽

...

Keyword(s):

Multiple Sclerosis ◽

Performance Measures ◽

Cognitive Fatigue ◽

Mixed Effect ◽

Mixed Effect Models ◽

Relapsing Remitting ◽

Measurement Design ◽

Relapsing Remitting Multiple Sclerosis ◽

Linear Mixed Effect Models

Background: Compromised learning and cognitive fatigue are critical clinical features in multiple sclerosis. Objectives: This study was designed to determine the effect of repeated exposures within and across study visits on performance measures of learning and cognitive fatigue in relapsing–remitting multiple sclerosis (RRMS). Methods: Thirty patients with RRMS and 30 controls were recruited. Using a burst measurement design (i.e. repeated assessments within and across study visits) the oral version of the Symbol Digit Modalities Test (SDMT) was administered three times during the baseline and two consecutive monthly follow-up visits for a total of nine test administrations. Learning was assessed within and across study visits whereas cognitive fatigue was assessed during the course of each test administration that was divided into three 30-second intervals. Results: Linear mixed-effect models revealed compromised learning within (95% CI: 2.6355 to 3.9867) and across (95% CI: 1.3250 to 3.1861) visits and worse cognitive fatigue (95% CI: −2.1761 to −0.1720) in patients with RRMS compared with controls. Among patients with RRMS, worse self-rated cognitive dysfunction predicted poor learning within (95% CI: −0.1112 to −0.0020) and across (95% CI: −0.0724 to −0.0106) visits. Conclusions: Burst design is optimal to study learning and cognitive fatigue. This methodology, using the SDMT or other time-efficient tests as outcome measures, can be successfully implemented in longitudinal studies and clinical trials.

Download Full-text

Monitoring Anti-Pythium insidiosum IgG Antibodies and (1->3)-β-D-Glucan in Vascular Pythiosis

10.1101/300764 ◽

2018 ◽

Author(s):

Navaporn Worasilchai ◽

Nitipong Permpalung ◽

Pakawat Chongsathidkiet ◽

Asada Leelahavanichkul ◽

Alberto Leonel Mendoza ◽

...

Keyword(s):

Residual Disease ◽

Synergistic Effects ◽

Clinical Information ◽

Mixed Effect ◽

Pythium Insidiosum ◽

Delay In Diagnosis ◽

Mixed Effect Models ◽

Monitoring Tools

AbstractBackgroundDespite aggressive treatment, vascular pythiosis has a mortality rate of 40%. This is due to the delay in diagnosis and lack of effective monitoring tools. To overcome this drawback, serum beta-D-glucan (BG) and P. insidiosum specific antibody (Pi-Ab) were examined as potential monitoring markers in vascular pythiosis.MethodsA prospective cohort study of vascular pythiosis patients was carried on during January 2010-July 2016. Clinical information and blood samples were collected and evaluated by the BG and Pi-Ab assays. Linear mixed effect models were used to compare BG and Pi-Ab levels. The in vitro susceptibility test was performed in all P. insidiosum isolates from culture positive cases.ResultsA total of 50 patients were enrolled; 45 survived and 5 died during follow-up. The survivors had a significantly shorter time to medical care (p<0.0001) and a significantly shorter waiting time to the first surgery (p<0.0001). There were no differences in BG levels among the groups at diagnosis (p=0.17); however, BG levels among survivors were significantly lower than the deceased group at 0.5 months (p<0.0001) and became undetectable after 3 months. Survivors were able to maintain ELISA value (EV) of Pi-Ab above 8, whereas EV among deceased patients was less than 4. In vitro susceptibility results revealed no synergistic effects between itraconazole and terbinafine.ConclusionsThis study showed that BG and Pi-Ab were potentially valuable markers to monitor the disease after treatment initiation. An unchanged BG level at 2 weeks after surgery should prompt an evaluation for residual disease.

Download Full-text

Influence of Cognitive Reserve on Cognitive Trajectories: Role of Brain Pathologies

Neurology ◽

10.1212/wnl.0000000000012728 ◽

2021 ◽

pp. 10.1212/WNL.0000000000012728

Author(s):

Xuerui Li ◽

Ruixue Song ◽

Xiuying Qi ◽

Hui Xu ◽

Wenzhe Yang ◽

...

Keyword(s):

Working Memory ◽

Cognitive Function ◽

Episodic Memory ◽

Cognitive Reserve ◽

Late Life ◽

Mixed Effect ◽

Mixed Effect Models ◽

Global Cognition

Background and Objectives:Evidence on the association of the cognitive reserve (CR) with the cognitive trajectories is limited. We aimed to examine the influence of CR indicator on domain-specific cognitive trajectories taking brain pathologies into account.Methods:Within the Rush Memory and Aging Project, 1,697 dementia-free participants (mean age: 79.6 years) were followed up to 21 years. CR indicator encompassing education, early-life, mid-life, and late-life cognitive activities, and late-life social activity was ascertained at baseline and categorized as tertiles (the lowest, middle, and highest). Global cognition, episodic memory, semantic memory, working memory, visuospatial ability, and perceptual speed were assessed annually with 19 tests, from which composite scores were derived. During the follow-up, 648 died and underwent autopsies to evaluate brain pathologies. Data were analyzed using linear mixed-effect models.Results:Among the participants, the score of the CR indicator ranged from -8.00 to 5.74 (mean: 0.00±2.23). In multi-adjusted mixed-effect models, compared to the lowest CR, the highest was related to a slower decline in global cognition (β=0.028, 95% confidence interval [CI]: 0.012 to 0.043), episodic memory (β=0.028, 95% CI: 0.010 to 0.047) and working memory (β=0.019, 95% CI: 0.005 to 0.033) during the follow-up. In brain pathological data analysis, the association of the highest CR with cognitive function changes remained significant among participants with high Alzheimer’s disease pathology or gross infarcts.Discussion:High CR indicator is associated with preserved global cognitive function, episodic memory, and working memory, even in the presence of brain pathologies. Our findings highlight the important role of high CR accumulation in the prevention of cognitive decline.

Download Full-text

Anxiety-impulsivity subtypes in adolescent internalizing disorder are characterized by distinguishable neurodevelopmental, neurocognitive and clinical trajectory signatures

10.1101/2021.10.30.21265692 ◽

2021 ◽

Author(s):

Huaxin Fan ◽

Nanyu Kuang ◽

Xinran Wu ◽

Gechang Yu ◽

Tianye Jia ◽

...

Keyword(s):

Emotional Disorders ◽

Sensation Seeking ◽

Internalizing Disorders ◽

Externalizing Disorders ◽

Regulatory Control ◽

Mixed Effect ◽

Neurodevelopmental Delay ◽

Mixed Effect Models ◽

Cortical Regions

Background Anxiety and impulsivity represent transdiagnostic pathology dimensions yet their interaction and contribution to emotional disorders in adolescence and to disease development remain controversial, and previous studies indicate heterogeneity within the broad category of internalizing disorders. Methods A combination of hierarchical and non-hierarchical clustering strategies was employed to determine impulsivity-related subtypes (based on the facets of negative urgency, lack of planning, lack of perseverance, sensation seeking and positive urgency in UPPS-P scales) in a large cohort of adolescents with internalizing disorders (n=2437) from Adolescent Brain Cognitive Development (ABCD) Study. Linear mixed-effect models were employed to determine cortical thickness alterations of the subtypes. Results Data-driven clustering identified two distinct subtypes of internalizing patients (subtype 1/subtype 2) with comparable levels of increased anxiety yet distinguishable levels of impulsivity, i.e., enhanced (subtype 1) or decreased (subtype 2) compared to healthy controls. Subtype 1 was further characterized by thicker prefrontal and temporal cortical regions involved in regulatory control and fear processing, while subtype 2 did not demonstrate significant thickness alterations. The differential neuroanatomical profiles remained stable over the two-year follow-up, while the two subtypes had different neurodevelopmental trajectories. Subtype 1 additionally reported more psychopathology and dysfunctionality including higher suicidal ideation, depressive symptoms and transition rates to externalizing disorders during follow-up as well as impaired neurocognitive and educational performance compared to subtype 2. Moreover, for subtype 1, anxiety at baseline (9-10 years) was significantly positively associated with impulsivity (lack of perseverance) at 2-year follow-up, while in subtype 2, baseline anxiety was significantly negatively associated with impulsivity (sensation seeking) at 2-year follow-up. Conclusions Our results demonstrate an impulsivity-dependent heterogeneity in adolescent internalizing disorders, with high-impulsivity patients being characterized by neurodevelopmental delay at the neural and cognitive levels. Individuals with elevated impulsivity are at a greater risk to develop behavioral dysregulation over the following two years and may thus require specific early interventions.

Download Full-text

Risk of suicide in patients who present to hospital after self-cutting according to site of injury: findings from the Multicentre Study of Self-harm in England

Psychological Medicine ◽

10.1017/s0033291721002956 ◽

2021 ◽

pp. 1-9

Author(s):

Galit Geulayov ◽

Deborah Casey ◽

Elizabeth Bale ◽

Fiona Brand ◽

Caroline Clements ◽

...

Keyword(s):

Multicentre Study ◽

Odds Ratio ◽

Suicide Risk ◽

Adjusted Odds Ratio ◽

The Body ◽

Mixed Effect ◽

Risk Of Death ◽

Mixed Effect Models ◽

Self Harm

Abstract Background We compared the risk of death by suicide following hospital presentation for self-harm according to site of self-cut/stab. Method We included 54 999 self-harm presentations (involving 31 419 individuals) to hospitals in the Multicentre Study of Self-harm in England (1/1/2004–31/12/2014), with mortality follow-up to 31/12/2019. Information on method of self-harm was obtained through monitoring in hospitals. Information about mortality was obtained through linkage with NHS Digital. We assessed the association of site of self-cut with death by suicide using mixed effect models. Results In total, 10 790 (19.6%) hospital presentations involved self-cutting/stabbing, 7489 of which (69.4%) were due to laceration to the arm/wrist alone, 1846 episodes (17.1%) involved cutting elsewhere on the body, and 1455 (13.5%) were due to laceration to unknown site. Controlling for confounders, presentation to a hospital following self-cut/stab to bodily parts other than wrist/arm was associated with greater chance of subsequent suicide relative to presentation after self-poisoning alone [adjusted odds ratio (aOR) 1.75, 95% confidence interval (CI) 1.03–2.96, p = 0.038]. The likelihood of suicide after presentation for cutting/stabbing the wrist/arm alone was comparable to that of patients who had self-poisoned alone. Presentations after laceration involving the neck were associated with a four-fold greater chance of subsequent suicide relative to self-poisoning (aOR 4.09, 95% CI 1.80–9.30, p = 0.001). Conclusions Patients who attend hospital after self-cutting/stabbing are a heterogeneous group in terms of characteristics, methods of cutting/stabbing and risk of subsequent suicide. Risk of suicide is greater in individuals who self-cut/stab to parts of the body other than the wrist or arm, especially the neck.

Download Full-text

Switching to Integrase Inhibitors Unlinked to Weight Increase in Perinatally HIV-Infected Young Adults and Adolescents: A 10-Year Observational Study

Microorganisms ◽

10.3390/microorganisms8060864 ◽

2020 ◽

Vol 8 (6) ◽

pp. 864

Author(s):

Lucia Taramasso ◽

Antonio Di Biagio ◽

Francesca Bovis ◽

Federica Forlanini ◽

Elena Albani ◽

...

Keyword(s):

Weight Gain ◽

Strand Transfer ◽

Outpatient Services ◽

Weight Changes ◽

Mixed Effect ◽

Mixed Effect Models ◽

Significant Difference ◽

Integrase Strand Transfer Inhibitors ◽

Control Study

An unexpected increase in weight gain has recently been reported in the course of integrase strand transfer inhibitors (INSTI) treatment. The possibility of this effect in people who are perinatally infected with HIV (PHIV) and thus exposed to lifelong therapy needs to be explored. This is a retrospective multicenter case-control study. Adults with PHIV followed between 2010 and 2019 in two outpatient services in Northern Italy were included if they had at least two weight measures in two successive years of observation. Patients were considered as cases if they were switched to INSTI (INSTI group), or controls if they were never exposed to INSTI (non-INSTI group). The date of the switch in cases was considered to be the baseline (T0), while it was randomly selected in controls. Mixed effect models were used to assess the weight changes in INSTI and non-INSTI groups. A total of 66 participants, 50.0% women, 92.4% Caucasian, were included. Median follow-up was 9 years (range 2–10): 4 years (range 1–8) before and 3 (range 1–9) after-T0. Mean age at the last study visit was 27.3 (±4.8) years, and mean CD4+ T-cells were 820.8 (±323.6) cells/mm3. Forty-five patients were switched to INSTI during the study, while 21 remained in the non-INSTI group. The INSTI group experienced a mean increase (pre-post T0) in bodyweight of 0.28 kg/year (95% CI − 0.29; 0.85, p = 0.338), while in the non-INSTI group, the mean increase was 0.36 kg/year (95% CI − 0.47; 1.20, p = 0.391), without a significant difference between groups (p for interaction between time and treatment regimen = 0.868). Among patients on INSTI, the weight gain after T0 was higher than pre-T0, amounting to +0.28 kg/year (95% CI − 0.29; 0.85), although this difference did not reach significance (p = 0.337). PHIV switched to an INSTI-based regimen did not experience an excessive weight gain compared to those who were treated with a non-INSTI based regimen in our cohort.

Download Full-text

Estimates of Within-Subject Biological Variation Derived from Pathology Databases: An Approach to Allow Assessment of the Effects of Age, Sex, Time between Sample Collections, and Analyte Concentration on Reference Change Values

Clinical Chemistry ◽

10.1373/clinchem.2018.290841 ◽

2019 ◽

Vol 65 (4) ◽

pp. 579-588 ◽

Cited By ~ 4

Author(s):

Graham Ross Dallas Jones

Keyword(s):

Clinical Laboratory ◽

Biological Variation ◽

Clinical Environment ◽

Analyte Concentration ◽

Regression To The Mean ◽

Statistical Process ◽

Reference Change Value ◽

First Results ◽

The Mean ◽

Number Of Individuals

Abstract BACKGROUND Within-subject biological variation data (CVI) are used to establish quality requirements for assays and allow calculation of the reference change value (RCV) for quantitative clinical laboratory tests. The CVI is generally determined using a large number of samples from a small number of individuals under controlled conditions. The approach presented here is to use a small number of samples (n = 2) that have been collected for routine clinical purposes from a large number of individuals. METHODS Pairs of sequential results from adult patients were extracted from a routine pathology database for 29 common chemical and hematological tests. Using a statistical process to identify a central gaussian distribution in the ratios of the result pairs, the total result variation for individual results was determined for 26 tests. The CVI was then calculated by removing the effect of analytical variation. RESULTS This approach produced estimates of CVI that, for most of the analytes in this study, show good agreement with published values. The data demonstrated minimal effect of sex, age, or time between samples. Analyte concentration was shown to affect the distributions with first results more distant from the population mean more likely to be followed by a result closer to the mean. DISCUSSION The process described here has allowed rapid and simple production of CVI data. The technique requires no patient intervention and replicates the clinical environment, although it may not be universally applicable. Additionally, the effect of regression to the mean described here may allow better interpretation of sequential patient results.

Download Full-text

Overnight affective dynamics and sleep characteristics as predictors of depression and its development in women

SLEEP ◽

10.1093/sleep/zsab129 ◽

2021 ◽

Author(s):

Olga Minaeva ◽

Sandip V George ◽

Anna Kuranova ◽

Nele Jacobs ◽

Evert Thiery ◽

...

Keyword(s):

Sleep Duration ◽

Mixed Effect ◽

Sleep Diaries ◽

The Past ◽

Mixed Effect Models ◽

Depressed Group ◽

Carry Over ◽

Predictors Of Depression ◽

Linear Mixed Effect Models

Abstract Study Objectives We examined i) differences in overnight affective inertia (carry-over of evening affect to the next morning) for positive (PA) and negative affect (NA) between individuals with past, current, and no depression; ii) how sleep duration and quality influence overnight affective inertia in these groups, and iii) whether overnight affective inertia predicts depression development. Methods We used data of 579 women from the East-Flanders Prospective Twin Survey. For aim 1 and 2, individuals with past (n=82), current (n=26), and without (lifetime) depression (n=471) at baseline were examined. For aim 3, individuals who did (n=58) and did not (n=319) develop a depressive episode at 12-month follow-up. Momentary PA and NA were assessed 10 times a day for 5 days. Sleep was assessed daily with sleep diaries. Affective inertia was operationalized as the influence of evening affect on morning affect. Linear mixed-effect models were used to test the hypotheses. Results Overnight affective inertia for NA was significantly larger in the current compared to the non-depressed group, and daytime NA inertia was larger in the past compared to the non-depressed group. Overnight NA inertia was differently associated with shorter sleep duration in both depression groups and with lower sleep quality in the current compared to the non-depressed group. Overnight affective inertia did not predict depression development at 12-month follow-up. Conclusions Current findings demonstrate the importance of studying complex affect dynamics such as overnight affective inertia in relation to depression and sleep characteristics. Replication of these findings, preferably with longer time-series, is needed.

Download Full-text