Evidence-based statistical analysis and methods in biomedical research (SAMBR) checklists according to design features

Randomized clinical trials and related concepts in clinical research

Clinical research is a key factor in healthcare progress, as it contributes toward improving our knowledge on the prevention, etiology, and treatment of different conditions. Healthcare professionals and researchers should be familiar with this specific terminology and procedures of clinical research to understand and be able to evaluate clinical trial results and make decisions using up-to-date recommendations. To do so, they must be familiar with different methodological aspects: from establishing the type of design, the study population, and the groups to be studied, to understanding the randomization and blinding processes. Additionally, when it comes to communicating the results and publishing them, it is also necessary to know how to do it adequately to ensure transparency. This work includes a description of different concepts commonly used in clinical research, particularly in the clinical trials field, in an attempt to compile different topics by providing a brief and accessible overview.

Association of exposure to nonpersistent endocrine disruptors with sex hormones and metabolic health in US females

Endocrine disruptive chemicals (EDCs) are considered as the potential attributes for the increasing trend in obesity and metabolic syndrome (MS) through disruption of sex hormones, particularly in women. We evaluated the association of understudied EDC compounds with total testosterone (TT), sex hormone-binding globulin (SHBG), obesity, and MS. A population-based cross-sectional study was conducted using the National Health and Nutrition Examination Survey datasets collected during the years 2013-2016. Women of age ≥15 years with urinary measurements of nonpersistent EDCs, including bisphenol, triclosan, triclocarban, dichlorophenol, and paraben compounds were included in this study. Data were analyzed using the modified Poisson models to estimate the adjusted relative risk (RR) and 95% confidence interval (CI). The associations were also validated by considering TT and SHBG concentrations as the outcomes. The study included 1974 women with 11% high TT, 10.5% low SHBG, 40% obesity, and 46.2% MS. A medium to high exposure to bisphenol-A (RR = 1.64; 95% CI: 1.14, 2.35, p = 0.009), bisphenol-F (RR = 1.83; 95% CI: 1.35, 2.49, p < 0.001), bisphenol-S (RR = 1.83; 95% CI: 1.35, 2.49, p = 0.041) and 2, 4- dichlorophenol (RR = 1.61; 95% CI: 1.06, 2.45, p = 0.026) were associated with low SHBG but not with high TT. In addition, high exposure to triclosan was also inversely associated with SHBG concentrations (regression coefficient = -0.09; 95% CI: -0.15, -0.02, p = 0.013). However, these EDCs were found to be associated with SHBG, obesity, and MS according to menopausal status. High exposure to certain nonpersistent EDCs was associated with low SHBG, obesity, and MS according to menopausal status.

Increases in amyloid-β42 slow cognitive and clinical decline in Alzheimer's disease trials

Positive effects of new anti-amyloid-β (Aβ) monoclonal antibodies in Alzheimer's disease (AD) have been attributed to brain amyloid reduction. However, most anti-Aβ antibodies also increase the CSF levels of the 42-amino acid isoform (Aβ42). We evaluated the associations of changes in CSF Aβ42 and brain Aβ-PET with cognitive and clinical end points in randomized trials of anti-Aβ drugs that lowered (β- and γ-secretase inhibitors) or increased CSF Aβ42 levels (anti-Aβ monoclonal antibodies) to test the hypothesis that post-treatment increases in CSF Aβ42 levels are independently associated with cognitive and clinical outcomes. From long-term (≥12 months) randomized placebo-controlled clinical trials of anti-Aβ drugs published until November 2023, we calculated the post-treatment versus baseline difference in ADAS-Cog (cognitive subscale of the Alzheimer's Disease Assessment Scale) and CDR-SB (Clinical Dementia Rate-Sum of Boxes) and z-standardized changes in CSF Aβ42 and Aβ-PET Centiloids (CL). We estimated the effect size [regression coefficients (RCs) and confidence intervals (CIs)] and the heterogeneity (I2) of the associations between AD biomarkers and cognitive and clinical end points using random-effects meta-regression models. We included 25 966 subjects with AD from 24 trials. In random-effects analysis, increases in CSF Aβ42 were associated with slower decline in ADAS-Cog (RC: -0.55; 95% CI: -0.89, -0.21, P = 0.003, I2 = 61.4%) and CDR-SB (RC: -0.16; 95% CI: -0.26, -0.06, P = 0.002, I2 = 34.5%). Similarly, decreases in Aβ-PET were associated with slower decline in ADAS-Cog (RC: 0.69; 95% CI: 0.48, 0.89, P < 0.001, I2 = 0%) and CDR-SB (RC: 0.26; 95% CI: 0.18, 0.33, P < 0.001, I2 = 0%). Sensitivity analyses yielded similar results. Higher CSF Aβ42 levels after exposure to anti-Aβ drugs are independently associated with slowing cognitive impairment and clinical decline. Increases in Aβ42 may represent a mechanism of potential benefit of anti-Aβ monoclonal antibodies in AD.

Racial/ethnic differences in the clinical presentation and survival of breast cancer by subtype

Background

Breast cancer (BC) affects racial and ethnic groups differently, leading to disparities in clinical presentation and outcomes. It is unclear how Hispanic ethnicity affects BC outcomes based on geographic location and proximity to the United States (U.S.)/Mexico border. We hypothesized that the impact of race/ethnicity on BC outcomes depends on geographic location and country of origin within each BC subtype.

Methods

We analyzed BC data from the Texas Cancer Registry by race/ethnicity/birthplace according to BC subtype (luminal A/luminal B/human epidermal growth factor receptor 2 [HER2]/triple-negative breast cancer[TNBC]). Other covariates included age, geographic location (U.S., Mexico), residency (border, non-border), treatments, and comorbidities. Crude and adjusted effects of race/ethnicity and birthplace on overall survival (OS) were analyzed using Cox regression methods.

Results

Our analysis of 76,310 patient records with specific BC subtypes revealed that Hispanic and non-Hispanic Black (NHB) patients were diagnosed at a younger age compared with non-Hispanic White (NHW) patients for all BC subtypes. For the 19,748 BC patients with complete data on race/ethnicity/birthplace/residency, Hispanic patients had a higher mortality risk in the Luminal A subtype, regardless of birthplace, whereas U.S.-born Hispanics had a higher risk of death in the TNBC subtype. In contrast, NHB patients had a higher mortality risk in the Luminal A and HER2 subtypes. Residence along the U.S./Mexico border had little impact on OS, with better outcomes in Luminal A patients and worse outcomes in Luminal B patients aged 60-74 years.

Conclusion

Race/ethnicity, geographic birth location, and residency were significant predictors of survival in BC. Migration, acculturation, and reduced healthcare access may contribute to outcome differences.

Association of heavy menstrual bleeding with cardiovascular disease in US female hospitalizations

BACKGROUND

Heavy menstrual bleeding (HMB) is a common menstrual disorder associated with multiple risk factors of cardiovascular disease (CVD) in women. In addition, HMB is often present with irregular menstruation (IM) which is a risk factor for CVD outcomes. However, the relationship between HMB and CVD outcomes is unexplored in the presence or absence of IM. We determined the association of HMB with multiple CVD outcomes using a nationally representative sample of female hospitalizations in the US.

METHODS

All hospitalizations of females with HMB diagnosis and normal menstrual cycles from ages of 18 to 70 years were extracted from the National Inpatient Sample Database, 2017. The HMB was defined using the International Classification of Diseases (ICD)-10 for excessive and frequent menstruation bleeding and included any current or history of HMB diagnosis. Outcomes including major adverse cardiovascular events (MACE), coronary heart disease (CHD), stroke, heart failure (HF), atrial fibrillation (AF) or arrhythmia, myocardial infarction (MI), and diabetes (DM) were defined using ICD-10 codes. Adjusted logistic regression and prosperity scores-matched logistic regression analyses were conducted to summarize adjusted associations with an odds ratio (OR) and a 95% confidence interval (CI).

RESULTS

Among 2,430,851 hospitalizations, HMB was observed in 7762 (0.68%) females with age ≤ 40 years and 11,164 (0.86%) females with age > 40 years. Among hospitalizations with age ≤ 40 years, HMB was significantly associated with increased odds of CVD outcomes including MACE (OR = 1.61; 95% CI: 1.25, 2.08), CHD (OR = 1.72; 95% CI: 1.10, 2.71), stroke (OR = 1.95; 95% CI: 1.12, 3.40), HF (OR = 1.53; 95% CI: 1.15, 2.03), and AF/arrhythmia (OR = 1.84; 95% CI: 1.34, 2.54). These associations were confirmed in multiple sensitivity analyses. In contrast, HMB was not robustly associated with CVD events among hospitalizations of women with age > 40 years. HMB without IM was strongly associated with DM, HF, AF, and MACE outcomes while HMB with IM was strongly associated with CHD and AF outcomes in hospitalizations of young women.

CONCLUSIONS

HMB is associated with CVD events among US hospitalizations of young women. A routine investigation and screening of menstrual disorders, especially HMB, is useful for CVD risk stratification and management in young women.

Metal and oxidative potential exposure through particle inhalation and oxidative stress biomarkers: a 2-week pilot prospective study among Parisian subway workers

OBJECTIVE

In this pilot study on subway workers, we explored the relationships between particle exposure and oxidative stress biomarkers in exhaled breath condensate (EBC) and urine to identify the most relevant biomarkers for a large-scale study in this field.

METHODS

We constructed a comprehensive occupational exposure assessment among subway workers in three distinct jobs over 10 working days, measuring daily concentrations of particulate matter (PM), their metal content and oxidative potential (OP). Individual pre- and post-shift EBC and urine samples were collected daily. Three oxidative stress biomarkers were measured in these matrices: malondialdehyde (MDA), 8-hydroxy-2'deoxyguanosine (8-OHdG) and 8-isoprostane. The association between each effect biomarker and exposure variables was estimated by multivariable multilevel mixed-effect models with and without lag times.

RESULTS

The OP was positively associated with Fe and Mn, but not associated with any effect biomarkers. Concentration changes of effect biomarkers in EBC and urine were associated with transition metals in PM (Cu and Zn) and furthermore with specific metals in EBC (Ba, Co, Cr and Mn) and in urine (Ba, Cu, Co, Mo, Ni, Ti and Zn). The direction of these associations was both metal- and time-dependent. Associations between Cu or Zn and MDAEBC generally reached statistical significance after a delayed time of 12 or 24 h after exposure. Changes in metal concentrations in EBC and urine were associated with MDA and 8-OHdG concentrations the same day.

CONCLUSION

Associations between MDA in both EBC and urine gave opposite response for subway particles containing Zn versus Cu. This diverting Zn and Cu pattern was also observed for 8-OHdG and urinary concentrations of these two metals. Overall, MDA and 8-OHdG responses were sensitive for same-day metal exposures in both matrices. We recommend MDA and 8-OHdG in large field studies to account for oxidative stress originating from metals in inhaled particulate matter.

Associations of sex hormone ratios with metabolic syndrome and inflammation in US adult men and women

Background

Sex hormones play a critical role in sex differences and cardiovascular disease risk associated with metabolic syndrome (MS) and inflammation. However, the associations of sex hormone ratios with metabolic and inflammatory markers are unclear according to sex and age differences. We evaluated the associations of sex hormone ratios with MS and inflammation among males and females.

Methods

A retrospective cross-sectional study was conducted by including all adults from the National Health and Nutrition Examination Survey cycles 2013-2016 and excluding any pregnant women, heart disease, diabetes, and those currently taking insulin. MS was defined using the National Cholesterol Education Program criteria and a high-sensitivity C-reactive protein (CRP) level>3 mg/L was defined as a high CRP. Measures of MS components and CRP concentrations were also analyzed. The primary exposures were testosterone to estradiol (excess androgen index), testosterone to sex hormone-binding globulin (free androgen index), and estradiol to sex hormone-binding globulin (free estradiol index). The adjusted associations were summarized with a relative risk (RR) and 95% confidence interval (CI).

Results

This study included 9167 subjects with 4360 males and 4807 females. Increases in free estradiol index were positively associated with MS (RR=1.48; 95%CI: 1.39, 1.58; RR=1.31; 95%CI: 1.22, 1.40) and high CRP (RR=1.49; 95%CI: 1.25, 1.77; RR=1.26; 95%CI: 1.06, 1.50) in men with age<50 years and age≥50 years, respectively. Similarly, higher free estradiol index was also robustly associated with increased prevalence of MS (RR=1.22; 95%CI: 1.15, 1.28) and high CRP (RR=1.68; 95%CI: 1.48, 1.90) in women with age ≥50 years. Among women with age<50 years, a higher free androgen index was associated with MS (RR=1.34; 95%CI: 1.25, 1.42) and high CRP (RR=1.13; 95%CI: 1.02, 1.25). These associations were unchanged even after adjusting for all sex hormones.

Conclusion

Free estradiol index was consistently and positively associated with MS and high CRP in males of all ages and older females. Free androgen index was positively associated with MS and high CRP in females with age<50 years.

Assessment of serum calcium, magnesium and zinc levels in patients with type 2 diabetes mellitus in the Ukrainian population

Background. The research deals with the provision of calcium, magnesium and zinc in patients with type 2 diabetes mellitus (T2DM). The purpose was to investigate the serum content of calcium, magnesium and zinc in patients with type 2 DM. Materials and methods. The open-label controlled study included 27 patients with T2DM. The control group consisted of 12 people without endocrine and somatic diseases. 70 % of patients had newly diagnosed DM; in others, the disease duration was up to 5 years. Most of the examined were of mature age, on average 61.83 ± 5.52 years for patients and 55.25 ± 5.52 years for controls. In the group with T2DM, women accounted for 62.96 %, among controls — 63.64 %. Anthropometric measurements and glucose indicators were determined according to a standard procedure. Concentrations of calcium, magnesium, zinc in serum were evaluated by atomic optical emission spectrometry with inductively coupled plasma (PerkinElmer Optima 2100 DV ICP-OES, USA) according to the original procedure approved by the Kundiiev Institute of Occupational Health of the NAMS of Ukraine. Results. The level of fasting blood sugar ranged from 7.1 to 17.2 mmol/l and indicated the presence of T2DM, glycated hemoglobin was from 7.1 to 11.2 %. According to anthropometric data, the group of patients almost completely corresponded to the control group. Although a slight increase in body weight and waist circumference was registered already at the stage of initial diabetes, the level of triglycerides increased significantly, and the level of high-density lipoprotein decreased. A significant reduction in the levels of calcium and magnesium in the blood serum of patients with T2DM was found. The zinc content was at the level of the lower edge of the reference value in 44 % of patients, in the control group — only in 16.6 % of cases. There were no significant differences in this indicator in the observation groups. A high correlation was found between the studied parameters in the blood, as well as an inverse correlation between the fasting glucose level and the content of calcium and magnesium. Conclusions. A significant decrease in the level of calcium and magnesium in the blood of Ukrainian patients with type 2 diabetes mellitus was found. There is a high correlation between serum calcium, magnesium and zinc levels in patients with T2DM. Significant risks of diabetes were revealed when the serum concentration of calcium and magnesium decreased.

Common Variables That Influence Sepsis Mortality in Mice

Introduction

Sepsis is characterized by a dysregulated host immune response to infection, leading to organ dysfunction and a high risk of death. The cecal ligation and puncture (CLP) mouse model is commonly used to study sepsis, but animal mortality rates vary between different studies. Technical factors and animal characteristics may affect this model in unanticipated ways, and if unaccounted for, may lead to serious biases in study findings. We sought to evaluate whether mouse sex, age, weight, surgeon, season of experiments, and timing of antibiotic administration influenced mortality in the CLP model.

Methods

We created a comprehensive dataset of C57BL/6J mice that had undergone CLP surgery within our lab during years 2015-2020 from published and unpublished studies. The primary outcome was defined as the time from sepsis induction to death or termination of study (14 days). The Log rank test and Cox regression models were used to analyze the dataset. The study included 119 mice, of which 43% were female, with an average age of 12.6 weeks, an average weight of 25.3 g. 38 (32%) of the animals died.

Results

In the unadjusted analyses, experiments performed in the summer and higher weight predicted a higher risk of mortality. In the stratified Cox model by sex, summer season (adjusted hazard ratio [aHR]=5.61, p=0.004) and delayed antibiotic administration (aHR=1.46, p=0.029) were associated with mortality in males, whereas higher weight (aHR=1.52, p=0.005) significantly affected mortality in females. In addition, delayed antibiotic administration (HR=1.42, p=0.025) was associated with mortality in the non-summer seasons, but not in the summer season.

Discussion

In conclusion, some factors specific to sex and season have a significant influence on sepsis mortality in the CLP model. Consideration of these factors along with appropriate group matching or adjusted analysis is critical to minimize variability beyond the experimental conditions within a study.

Association Between Tumor Mutation Profile and Clinical Outcomes Among Hispanic-Latino Patients With Metastatic Colorectal Cancer

In the United States, CRC is the third most common type of cancer and the second leading cause of cancer-related death. Although the incidence of CRC among the Hispanic population has been declining, recently, a dramatic increase in CRC incidents among HL younger than 50 years of age has been reported. The incidence of early-onset CRC is more significant in HL population (45%) than in non-Hispanic Whites (27%) and African-Americans (15%). The reason for these racial disparities and the biology of CRC in the HL are not well understood. We performed this study to understand the biology of the disease in HL patients. We analyzed formalin-fixed paraffin-embedded tumor tissue samples from 52 HL patients with mCRC. We compared the results with individual patient clinical histories and outcomes. We identified commonly altered genes in HL patients (APC, TP53, KRAS, GNAS, and NOTCH). Importantly, mutation frequencies in the APC gene were significantly higher among HL patients. The combination of mutations in the APC, NOTCH, and KRAS genes in the same tumors was associated with a higher risk of progression after first-line of chemotherapy and overall survival. Our data support the notion that the molecular drivers of CRC might be different in HL patients.

A systematic review and meta-analysis of the association between maternal polycystic ovary syndrome and neuropsychiatric disorders in children

There is emerging evidence demonstrating an association between maternal polycystic ovary syndrome (PCOS) and autism spectrum disorder (ASD) in children, however, the cumulative effect of maternal PCOS on the development of ASD or other neuropsychiatry disorders (NPD) in children and separately for males and females has not been examined. We sought to systematically evaluate the influence of maternal PCOS on a wide range of NPD including ASD, attention deficit hyperactivity disorder (ADHD), chronic tic disorder (CDT), other behavior disorders, anxiety, depression, bipolar disorder, schizophrenia in children as well as in women of reproductive age only. We queried electronic databases including PubMed, EMBASE, and Google Scholar, until March 2021. We used DerSimonian and Laird (D-L) random effects method to compute pooled effect size in terms of odds ratio (OR). Nineteen studies (1667851 mothers, 2260622 children) were included in this study. Mothers with PCOS had an increased odds of children diagnosed with ASD (OR = 1.40, p < 0.001), ADHD (OR = 1.42, p < 0.001), CTD (OR = 1.44, p = 0.001), anxiety (OR = 1.33, p < 0.001), as well as other behavioral symptoms (OR = 1.45, p < 0.001) in the adjusted analysis. The association between maternal PCOS and ASD (OR: 1.43 vs. 1.66), ADHD (OR: 1.39 vs. 1.54), and CTD (OR: 1.42 vs. 1.51) was found to be significantly consistent between males and females, respectively. Our data do not suggest increased fetal testosterone exposure is associated with increased autistic traits in children. However, PCOS was significantly associated with increased odds of a wide range of NPD in women themselves. Maternal PCOS is a risk factor for various NPD with a similar extent in their children regardless of their underlying comorbidities. Managing PCOS is essential for women's health as well as for their children's health. More research is needed to determine the mechanisms and links between maternal PCOS and NPD in children.

A systematic review and meta-analysis of geographic differences in comorbidities and associated severity and mortality among individuals with COVID-19

Several comorbidities have been shown to be associated with coronavirus disease 2019 (COVID-19) related severity and mortality. However, considerable variation in the prevalence estimates of comorbidities and their effects on COVID-19 morbidity and mortality have been observed in prior studies. This systematic review and meta-analysis aimed to determine geographical, age, and gender related differences in the prevalence of comorbidities and associated severity and mortality rates among COVID-19 patients. We conducted a search using PubMed, Scopus, and EMBASE to include all COVID-19 studies published between January 1st, 2020 to July 24th, 2020 reporting comorbidities with severity or mortality. We included studies reporting the confirmed diagnosis of COVID-19 on human patients that also provided information on comorbidities or disease outcomes. We used DerSimonian and Laird random effects method for calculating estimates. Of 120 studies with 125,446 patients, the most prevalent comorbidity was hypertension (32%), obesity (25%), diabetes (18%), and cardiovascular disease (16%) while chronic kidney or other renal diseases (51%, 44%), cerebrovascular accident (43%, 44%), and cardiovascular disease (44%, 40%) patients had more COVID-19 severity and mortality respectively. Considerable variation in the prevalence of comorbidities and associated disease severity and mortality in different geographic regions was observed. The highest mortality was observed in studies with Latin American and European patients with any medical condition, mostly older adults (≥ 65 years), and predominantly male patients. Although the US studies observed the highest prevalence of comorbidities in COVID-19 patients, the severity of COVID-19 among each comorbid condition was highest in Asian studies whereas the mortality was highest in the European and Latin American countries. Risk stratification and effective control strategies for the COVID-19 should be done according to comorbidities, age, and gender differences specific to geographical location.

Current trends and geographical differences in therapeutic profile and outcomes of COVID-19 among pregnant women - a systematic review and meta-analysis

BACKGROUND

Coronavirus disease (COVID-19) has been associated with adverse pregnancy outcomes. Due to the lack of effective treatments for COVID-19, it becomes imperative to assess the geographical differences and trends in the current clinical care and outcomes of COVID-19 in pregnant women.

METHODS

A PubMed search was performed to screen articles reporting therapeutics and outcomes of confirmed COVID-19 in pregnant women prior to August 27, 2020. We performed searches, quality assessments of eligible studies, extracted and reported data according to PRISMA guidelines. Meta-analyses and cumulative meta-analyses of proportions were performed for estimating each outcome and their pattern over time respectively.

RESULTS

One thousand two hundred thirty nine pregnant women with COVID-19 from 66 studies were analyzed. In case series analysis reflecting average-risk patients, the proportion of oxygen support, antibiotics, antivirals, and plasma therapy administration except for hydroxychloroquine was substantially higher in Asian studies (55, 78, 80, 6, and 0%) compared to the US (7, 1, 12, 0, and 7%) or European (33, 12, 14, 1, and 26%) studies, respectively. The highest preterm birth and the average length of hospital stay (35%, 11.9 days) were estimated in Asian studies compared to the US studies (13%, 9.4 days) and European studies (29%, 7.3 days), respectively. Even in case reports reflecting severe cases, the use of antivirals and antibiotics was higher in Asian studies compared to the US, Latin American, and European studies. A significant decline in the use of most therapeutics along with adverse outcomes of COVID-19 in pregnant women was observed.

CONCLUSIONS

Geographical differences in therapeutic practice of COVID-19 were observed with differential rates of maternal and clinical outcomes. Minimizing the use of some therapeutics particularly antibiotics, antivirals, oxygen therapy, immunosuppressants, and hydroxychloroquine by risk stratification and careful consideration may further improve maternal and clinical outcomes.

Evidence-based statistical analysis and methods in biomedical research (SAMBR) checklists according to design features

BACKGROUND

Statistical analysis according to design features and objectives is essential to ensure the validity and reliability of the study findings and conclusions in biomedical research. Heterogeneity in reporting study design elements and conducting statistical analyses is often observed for the same study design and study objective in medical literatures. Sometimes, researchers face a lot of predicaments using appropriate statistical approaches highlighted by methodologists for a specific study design either due to lack of accessibility or understanding of statistical methods or unavailability of checklists related to design and analysis in a concise format. The purpose of this review is to provide the checklist of statistical analysis and methods in biomedical research (SAMBR) to applied researchers.

RECENT FINDINGS

We initially identified the important steps of reporting design features that may influence the choice of statistical analysis in biomedical research and essential steps of data analysis of common studies. We subsequently searched for statistical approaches employed for each study design/study objective available in publications and other resources. Compilation of these steps produced SAMBR guidance document, which includes three parts. Applied researchers can use part (A) and part (B) of SAMBR to describe or evaluate research design features and quality of statistical analysis, respectively, in reviewing studies or designing protocols. Part (C) of SAMBR can be used to perform essential and preferred evidence-based data analysis specific to study design and objective.

CONCLUSIONS

We believe that the statistical methods checklists may improve reporting of research design, standardize methodological practices, and promote consistent application of statistical approaches, thus improving the quality of research studies. The checklists do not enforce the use of suggested statistical methods but rather highlight and encourage to conduct the best statistical practices. There is a need to develop an interactive web-based application of the checklists for users for its wide applications.