Comparing Smoothing Techniques for Fitting the Nonlinear Effect of Covariate in Cox Models

Roshani, Daem; Ghaderi, Ebrahim

doi:10.5455/aim.2016.24.38-41

Cited by 32 publications

(25 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To test whether the association with age differed between the menopausal status groups, we included an interaction term of menopausal status with age in the model and tested its significance with an analysis of variance (ANOVA). Furthermore, in order to take into account a potential non-linear relationship of age with CVD risk factors, restricted cubic splines for age were added to the model [22, 23]. The model was then tested for non-linearity with an ANOVA analysis.…”

Section: Methodsmentioning

confidence: 99%

Unraveling the associations of age and menopause with cardiovascular risk factors in a large population-based study

Kat

Dam

Onland‐Moret

et al. 2017

BMC Med

119

View full text Add to dashboard Cite

BackgroundAlthough the association between menopause and cardiovascular disease (CVD) risk has been studied extensively, the simultaneous role of chronological aging herein remains underexposed. This study aims to disentangle the relationships of menopausal status and chronological aging with CVD risk factors in the largest study population to date.MethodsIn this cross-sectional study, CVD risk factors were compared between women with a different menopausal status within the same yearly age strata. The study population comprised female participants of the baseline visit of the population-based LifeLines Cohort Study. A total of 63,466 women, aged between 18 and 65 years, was included. Of them, 39,379 women were considered to be premenopausal, 8669 were perimenopausal, 14,514 were naturally postmenopausal, and 904 were surgically postmenopausal.ResultsCompared to postmenopausal women aged 45 years, average total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-c) were 0.5 and 0.4 mmol/L higher, respectively, in postmenopausal women aged 50. Systolic and diastolic blood pressure levels were 4 and 1 mmHg higher, respectively. At all ages between 46 and 55 years, and after adjustment for confounders, naturally postmenopausal women had 0.2 to 0.4 mmol/L higher TC and 0.1 to 0.3 mmol/L higher LDL-c levels compared to premenopausal women in the same age range. Systolic blood pressure levels were up to 4 mmHg lower in naturally post- compared to premenopausal women at all ages between 29 and 52 years. Body mass index levels were up to 3.2 kg/m2 higher in women with surgical menopause compared to all other women between the ages 32 and 52 years. All aforementioned results were statistically significant.ConclusionsChronological age and menopausal status are both independently associated with CVD risk factors. Based on the comparatively smaller observed differences associated with menopausal status than with chronological aging, the significance of a more unfavorable lipid profile in a later reproductive stage may be less obvious than previously thought.Electronic supplementary materialThe online version of this article (doi:10.1186/s12916-016-0762-8) contains supplementary material, which is available to authorized users.

show abstract

Section: Methodsmentioning

confidence: 99%

Unraveling the associations of age and menopause with cardiovascular risk factors in a large population-based study

Kat

Dam

Onland‐Moret

et al. 2017

BMC Med

119

View full text Add to dashboard Cite

show abstract

“…In order to identify the optimal cutoff; time to treatment from the date of diagnosis and prostatectomy was introduced as a continuous variable in a multivariate cox regression model. Restricted cubic spline (RCS) function with four knots was used in cox model to allow nonlinear association between HR and time to treatment (2,25). RCS-based cox model was adjusted using all the pertinent pretreatment clinicodemographic variables including clinical GS, PSA, clinical T stage, race, age, and year of prostatectomy.…”

Section: Discussionmentioning

confidence: 99%

Optimizing Time to Treatment to Achieve Durable Biochemical Disease Control after Surgery in Prostate Cancer: A Multi-Institutional Cohort Study

Awasthi

Gerke

Park

et al. 2019

Cancer Epidemiology, Biomarkers &Amp; Prevention

View full text Add to dashboard Cite

Background: The impact of treatment delays on prostate cancer-specific outcomes remains ill-defined. This study investigates the effect of time to treatment on biochemical disease control after prostatectomy. Methods: This retrospective study includes 1,807 patients who received a prostatectomy as a primary treatment at two large tertiary referral centers from 1987 to 2015. Multivariate cox model with restricted cubic spline was used to identify optimal time to receive treatment and estimate the risk of biochemical recurrence. Results: Median follow-up time of the study was 46 (interquartile range, 18-86) months. Time to treatment was subcategorized based on multivariate cubic spline cox model. In multivariate spline model, adjusted for all the pertinent pretreatment variables, inflection point in the risk of biochemical recurrence was observed around 3 months, which further increased after 6 months. Based on spline model, time to treatment was then divided into 0 to 3 months (61.5%), >3 to 6 months (31.1%), and 6 months (7.4%). In the adjusted cox model, initial delays up to 6 months did not adversely affect the outcome; however, time to treatment >6 months had significantly higher risk of biochemical recurrence (HR, 1.84; 95% confidence interval, 1.30-2.60; P < 0.01). Conclusions: The initial delays up to 6 months in prostate cancer primary treatment may be sustainable without adversely affecting the outcome. However, significant delays beyond 6 months can unfavorably affect biochemical disease control. Impact: Time to treatment can aid clinicians in the decisionmaking of prostate cancer treatment recommendation and educate patients against unintentional treatment delays.

show abstract

“…The proportional hazard assumptions of the Cox model were tested for each covariate by visually examining the parallelism of stratified survival curves, but no violations were detected. The relationship between number of ≥4 mm deep periodontal pockets and incident liver disease was examined using the penalized spline smoothing method adjusted for age, sex and number of teeth (edentulous subjects were excluded). P values <0.05 were considered statistically significant.…”

Section: Methodsmentioning

confidence: 99%

Periodontitis is associated with incident chronic liver disease—A population‐based cohort study

Helenius‐Hietala

Suominen

Ruokonen

et al. 2018

Liver International

View full text Add to dashboard Cite

Background & Aims Chronic liver disease is a major health concern worldwide and the identification of novel modifiable risk factors may benefit subjects at risk. Few studies have analyzed periodontitis as a risk factor for liver complications. We studied whether periodontitis is associated with incident severe liver disease. Methods The study comprised 6165 individuals without baseline liver disease who participated in the Finnish population‐based Health 2000 Survey (BRIF8901) during 2000‐2001, a nationally representative cohort. Follow‐up was until 2013 for liver‐related admissions, liver cancer and mortality from National Hospital Discharge, Finnish Cancer Registry and Causes of Death Register, Statistics Finland. Mild to moderate periodontitis was defined as ≥1 tooth with periodontal pocket ≥4 mm deep, and advanced periodontitis as ≥5 teeth with such pockets. Multiple confounders were considered. Results A total of 79 subjects experienced a severe liver event during follow‐up. When adjusted for age, sex and number of teeth, hazards ratios by Cox regression regarding incident severe liver disease were, for mild to moderate periodontitis, 2.12 (95% CI 0.98‐4.58), and, for advanced periodontitis, 3.69 (95% CI 1.79‐7.60). These risk estimates remained stable after additionally adjusting for alcohol use, smoking, metabolic risk, serum gamma‐glutamyltransferase, dental‐care habits, lifestyle and socioeconomic status. Periodontal disease‐associated liver risk was accentuated among subjects with non‐alcoholic fatty liver disease or heavy alcohol use at baseline. Conclusions Periodontitis was associated with incident liver disease in the general population independently of various confounders. As a preventable disease, periodontal disease might present a modifiable risk factor for chronic liver disease.

show abstract

Comparing Smoothing Techniques for Fitting the Nonlinear Effect of Covariate in Cox Models

Cited by 32 publications

References 15 publications

Unraveling the associations of age and menopause with cardiovascular risk factors in a large population-based study

Unraveling the associations of age and menopause with cardiovascular risk factors in a large population-based study

Optimizing Time to Treatment to Achieve Durable Biochemical Disease Control after Surgery in Prostate Cancer: A Multi-Institutional Cohort Study

Periodontitis is associated with incident chronic liver disease—A population‐based cohort study

Contact Info

Product

Resources

About