Concave likelihood-based regression with finite-support response variables.

We propose a unified framework for likelihood-based regression modeling when the response variable has finite support. Our work is motivated by the fact that, in practice, observed data are discrete and bounded. The proposed methods assume a model which includes models previously considered for interval-censored variables with log-concave distributions as special cases. The resulting log-likelihood is concave, which we use to establish asymptotic normality of its maximizer as the number of observations n tends to infinity with the number of parameters d fixed, and rates of convergence of L1 -regularized estimators when the true parameter vector is sparse and d and n both tend to infinity with log ( d ) / n → 0 $\log (d) / n \rightarrow 0$ . We consider an inexact proximal Newton algorithm for computing estimates and give theoretical guarantees for its convergence. The range of possible applications is wide, including but not limited to survival analysis in discrete time, the modeling of outcomes on scored surveys and questionnaires, and, more generally, interval-censored regression. The applicability and usefulness of the proposed methods are illustrated in simulations and data examples.

Impact of modern systemic therapies and clinical markers on treatment outcome for metastatic melanoma in a real-world setting.

BACKGROUND: The survival in metastatic melanoma has dramatically improved after the introduction of immune checkpoint- (ICIs) and MAPKinase inhibitors (MAPKis). OBJECTIVE: Our aim was to describe therapy response and survival in a real-world population as well as to assess the associations between clinical variables and therapy outcome for patients with metastatic melanoma receiving first-line ICIs or MAPKis. METHODS: A total of 252 patients with metastatic (stage IV) melanoma were prospectively followed between 1 January 2010 and 3 December 2017 with follow-up until 31 March 2019, at the Karolinska University Hospital, Sweden. Hazard ratios (HRs) for progression-free survival (PFS) and overall survival (OS) were analysed with Cox regression, and logistic regression was used to estimate odds ratios (ORs) for therapy response. RESULTS: Patients receiving ICIs (n = 138) experienced longer PFS compared to patients that received MAPKis (n = 114; median PFS for ICIs was 6.8 months, and median PFS for MAPKis was 5.3 months). In the multivariable analyses of clinical markers, increasing M-stage (OR 0.65; 95% CI 0.45-0.94; P = 0.022) and male sex (OR 0.41; 95% CI 0.19-0.90; P = 0.027) were significantly associated with lower response to ICIs. Lower baseline albumin levels (OR 0.90; 95% CI 0.83-0.98; P = 0.019) and male sex (OR 0.33; 95% CI 0.12-0.93; P = 0.036) were related with lower response to MAPKis. For ICIs, increasing M-stage (HR 1.34; 95% CI 1.07-1.68; P = 0.010), increasing LDH (HR 1.73; 95% CI 1.19-2.50; P = 0.004) and decreasing albumin (HR 1.06; 95% CI 1.01-1.10; P = 0.011) were significantly associated lower PFS in the adjusted model. The corresponding markers for MAPKis were increasing LDH (HR 1.44; 95% CI 1.08-1.92; P = 0.013) and decreasing albumin (HR 1.05; 95% CI 1.02-1.09; P = 0.005) for PFS. CONCLUSION: ICIs and MAPKis were effective in this real-world population, and we could confirm the importance of previously reported clinical prognostic markers. Albumin values may be associated with therapy outcome but need further validation.

Impact of comorbidity on disease characteristics, treatment intent and outcome in diffuse large B-cell lymphoma: a Swedish lymphoma register study.

BACKGROUND: Comorbidity impacts overall survival amongst patients with diffuse large B-cell lymphoma (DLBCL). However, associations of comorbidity with lymphoma characteristics, treatment selection and lymphoma-specific mortality are less well known. OBJECTIVE: To examine the impact of comorbidity on DLBCL characteristics, treatment intent and cause of death. METHODS: We identified 3905 adult patients diagnosed with DLBCL 2007-2013 through the Swedish Lymphoma Register. We assessed comorbid disease history according to the Charlson comorbidity index (CCI). Comorbidity data and causes of death were collected through register linkage. Associations were estimated using multinomial regression and flexible parametric survival models. RESULTS: Overall, 45% of the patients (n = 1737) had a history of at least one comorbidity at DLBCL diagnosis (cardiovascular disease, diabetes and solid cancer were most frequent), and 997 (26%) had a CCI score of ≥2. The relative probability of presenting with poor performance status (PS > 2) was higher amongst comorbid patients [Relative Risk Ratio (RRR)PS>2 : 2.02, 95% CI: 1.63-2.51]. Comorbid patients had a substantially lower relative probability of receiving curative treatment (RRR: 0.48, 95% CI: 0.38-0.61). Amongst all patients, CCI ≥ 1 was associated with a significantly increased risk of all-cause and lymphoma-specific death after adjustments. Amongst patients selected for curative treatment, comorbidity was associated with an increased risk of all-cause death (HRCCI>1 : 1.54, 95% CI: 1.32-1.80), but not with lymphoma-specific death (HRCCI>1 : 1.05, 95% CI: 0.86-1.28). CONCLUSION: Comorbidity is associated with inferior DLBCL outcome, mainly due to a lower likelihood of receiving treatment with curative intent. Possibly, more comorbid DLBCL patients could be treated with curative intent if comorbid conditions were optimized in parallel.

Sex differences in repair rates and outcomes of patients with ruptured abdominal aortic aneurysm.

BACKGROUND: Data are conflicting on sex differences in ruptured abdominal aortic aneurysm (rAAA) repair rates and outcomes have rarely been addressed. The aim of this study was to investigate differences in the management and outcome of rAAA in men and women, and to describe time trends over a 15-year interval. METHODS: Data on patients with rAAA were extracted from the Swedish National Patient Registry and the Cause of Death Registry for the interval 2001-2015. The study included patients with rAAA whether or not they were admitted to any hospital in Sweden. A propensity score-matched model was used to determine sex differences in repair type and outcome after rupture. Time trends for rAAA events and mortality were investigated. RESULTS: Some 10 724 patients were identified. A higher percentage of men were admitted to hospital (79·8 versus 77·5 per cent; P = 0·011). Of those admitted, a higher percentage of men than women were treated (56·6 versus 40·4 per cent, P < 0·001). Women were less likely to be treated when diagnosed with rAAA (average treatment effect (ATE) in the model -0·080, 95 per cent c.i. -0·106 to -0·055; P < 0·001). Thirty-day mortality was also higher in women (ATE 0·094, 0·053 to 0·135; P < 0·001); this effect persisted to 1 year (ATE 0·095, 0·052 to 0·137; P < 0·001). Time trends indicated a decrease in rAAA incidence, mostly owing to a decrease among men. CONCLUSION: In this study, fewer women with rAAA received surgery and 30-day mortality was higher than in men. There was an overall decrease in rAAA incidence, principally in men.

ANTECEDENTES: Los datos sobre las diferencias de sexo en las tasas de reparación del aneurisma de la aorta abdominal roto (ruptured abdominal aortic aneurysm, rAAA) son contradictorios y existe poca información de los resultados. El objetivo de este estudio fue investigar las diferencias en el tratamiento y resultados en varones y mujeres con rAAA y describir las tendencias a lo largo del tiempo durante un periodo de 15 años. MÉTODOS: Los datos de los pacientes con rAAA se obtuvieron a partir del Registro Nacional Sueco de Pacientes y del Registro de Mortalidad en el periodo 2001-2015. El estudio incluye pacientes con rAAA, tanto si fueron ingresados o no en cualquier hospital de Suecia. Se utilizó un modelo de emparejamiento por puntaje de propensión para determinar las diferencias de sexo en el tipo de reparación y los resultados tras la rotura del aneurisma. Se describen las tendencias temporales para los eventos de rAAA y de mortalidad. RESULTADOS: Se identificaron un total de 10.724 pacientes. Un elevado porcentaje de pacientes varones fueron ingresados en el hospital (79,8% versus 77,5%, P < 0,001). De aquellos que ingresaron, un mayor porcentaje de varones fueron tratados (56,6%, versus 40,4% P < 0,001) en comparación con las mujeres. Las mujeres tuvieron menos probabilidades de ser tratadas cuando se les diagnosticó de un rAAA, con un efecto medio del tratamiento (average treatment effect, ATE) en el modelo de -0,08 (i.c. del 95% -0,106 a -0,055, P < 0,01). La mortalidad a los 30 días fue también más elevada en las mujeres, ATE 0,094 (i.c. del 95% 0,053-0,135, P < 0,001) que persistió hasta un año, ATE 0,095 (i.c. del 95% 0,052-0,137, P < 0,001). Las tendencias temporales indicaron una disminución en la incidencia de rAAA, a causa principalmente de una disminución en los varones. CONCLUSIÓN: En este estudio, el número de mujeres con rotura de AAA operadas fue menor que en los varones, pero la mortalidad a los 30 días fue mayor en las mujeres. Hubo una disminución global en la incidencia de rAAA, principalmente en los varones.

Adherence to oxygen saturation targets increased in preterm infants when a higher target range and tighter alarm limits were introduced.

AIM: European consensus guidelines published in May 2013 recommended a target peripheral capillary oxygen saturation (SpO2 ) range of 90-95% for preterm infants. These were incorporated into guidelines at the Karolinska University Hospital, Sweden, in November 2013. This study compared clinical practice before and after those local guidelines. METHODS: We included infants who were born between 23 + 0 and 30 + 6 weeks from January 1, 2013 to December, 31 2015 and received intensive care in two Karolinska units. The lower saturation target of 88-92% and alarm limits of 85-95% used before November 2013 were compared to the new higher saturation target of 90-95% and alarm limits of 89-96%. RESULTS: Data from 399 infants were analysed. The mean SpO2 was 92.4% with the higher target (n = 301) and 91.1% with the lower target (n = 98). Using the higher instead of lower target meant that the SpO2 was within the prescribed target range more frequently (51% versus 30%) and the proportion of time with SpO2 >95% was increased by 9% (95% confidence interval 7-11%, p < 0.001). CONCLUSION: The higher saturation target and tighter alarm limits led to higher mean oxygen saturation, increased adherence to the target and increased time with hyperoxaemia.

Quantitative chromogenic immunohistochemical image analysis in cellprofiler software.

Visual grading of chromogenically stained immunohistochemical (IHC) samples is subjective, time consuming, and predisposed to considerable inter- and intra-observer variations. The open-source digital analysis software, CellProfiler has been extensively used for fluorescently stained cells/tissues; however, chromogenic IHC staining is routinely used in both pathological and research diagnostics. The current investigation aimed to compare CellProfiler quantitative chromogenic IHC analyses against the gold standard manual counting. Oral mucosal biopsies from patients with chronic graft-versus-host disease were stained for CD4. Digitized images were manually counted and subjected to image analysis in CellProfiler. Inter-observer and inter-platform agreements were assessed by scatterplots with linear regression and Bland-Altman plots. Validation comparisons between the manual counters demonstrated strong intra-observer concordance (r2 = 0.979), particularly when cell numbers were less than 100. Scatterplots and Bland-Altman plots demonstrated strong agreement between the manual counters and CellProfiler, with the number of positively stained cells robustly correlating (r2 = 0.938). Furthermore, CellProfiler allowed the determination of multiple variables simultaneously, such as area stained and masking to remove any nonstained tissue and white gaps, which also demonstrated reliable agreement (r2 = >0.9). CellProfiler demonstrated versatility with the ability to assess large numbers of images and allowed additional parameters to be quantified. CellProfiler allowed rapid high processing capacity of chromogenically stained chronic inflammatory tissue that was reliable, accurate, and reproducible and highlights potential applications in research diagnostics.

Restorative surgery after colectomy for ulcerative colitis in England and Sweden: observations from a comparison of nationwide cohorts.

AIM: A longstanding disparity exists between the approaches to restorative surgery after colectomy for patients with ulcerative colitis (UC) in England and Sweden. This study aims to compare rates of colectomy and restorative surgery in comparable national cohorts. METHOD: The English Hospital Episode Statistics (HES) and Swedish National Patient Register (NPR) were interrogated between 2002 and April 2012. Patients with two diagnostic episodes for UC (age ≥ 15 years) were included. Patients were excluded if they had an episode of inflammatory bowel disease or colectomy before 2002. The cumulative incidences of colectomy and restorative surgery were calculated using the Kaplan-Meier method. RESULTS: A total of 98 691 patients were included in the study, 76 129 in England and 22 562 in Sweden. The 5-year cumulative incidence of all restorative surgery after colectomy in England was 33% vs 46% in Sweden (P-value < 0.001). Of the patients undergoing restorative surgery, 92.3% of English patients had a pouch vs 38.8% in Sweden and 7.7% vs 59.1% respectively had an ileorectal anastomosis (IRA). The 5-year cumulative incidence of colectomy in this study cohort was 13% in England and 6% in Sweden (P-value < 0.001). CONCLUSION: Following colectomy for UC only one-third of English patients and half of Swedish patients underwent restorative surgery. In England nearly all these patients underwent pouches, in Sweden a less significant majority underwent IRAs. It is surprising to demonstrate this discrepancy in a comparable cohort of patients from similar healthcare systems. The causes and consequences of this international variation in management are not fully understood and require further investigation.

Diagnosis and classification of idiopathic inflammatory myopathies.

The idiopathic inflammatory myopathies (IIMs) are a heterogeneous group of diseases, collectively termed myositis, sharing symptoms of muscle weakness, fatigue and inflammation. Other organs are frequently involved, supporting the notion that these are systemic inflammatory diseases. The IIMs can be subgrouped into dermatomyositis, polymyositis and inclusion body myositis. The myositis-specific autoantibodies (MSAs) identify other and often more distinct clinical phenotypes, such as the antisynthetase syndrome with antisynthetase autoantibodies and frequent interstitial lung disease and anti-SRP and anti-HMGCR autoantibodies that identify necrotizing myopathy. The MSAs are important both to support myositis diagnosis and to identify subgroups with different patterns of extramuscular organ involvement such as interstitial lung disease. Another cornerstone in the diagnostic procedure is muscle biopsy to identify inflammation and to exclude noninflammatory myopathies. Treatment effect and prognosis vary by subgroup. To develop new and better therapies, validated classification criteria that identify distinct subgroups of myositis are critical. The lack of such criteria was the main rationale for the development of new classification criteria for IIMs, which are summarized in this review; the historical background regarding previous diagnostic and classification criteria is also reviewed. As the IIMs are rare diseases with a prevalence of 10 in 100 000 individuals, an international collaboration was essential, as was the interdisciplinary effort including experts in adult and paediatric rheumatology, neurology, dermatology and epidemiology. The new criteria have been developed based on data from more than 1500 patients from 47 centres worldwide and are based on clinically easily available variables.

Subthreshold and threshold attention deficit hyperactivity disorder symptoms in childhood: psychosocial outcomes in adolescence in boys and girls.

OBJECTIVE: To examine the association between different levels of childhood attention deficit hyperactivity disorder (ADHD) symptoms and sex differences in psychosocial outcomes during adolescence. METHOD: Swedish children (n = 4635) were screened for neuropsychiatric symptoms at age 9 or 12. ADHD symptoms were divided into three levels: screen-negative, screen-intermediate, and screen-positive. At follow-up (age 15), parents and teenagers filled out questionnaires regarding (i) hyperactivity/inattention, (ii) peer problems, (iii) school problems, (iv) internalizing problems, (v) antisocial behaviour, (vi) alcohol misuse, and (vii) drug misuse. All outcomes were controlled for symptoms of diagnostic categories other than ADHD. RESULTS: Increasing levels of ADHD symptoms in childhood were associated with higher proportions of adolescents who displayed negative psychosocial outcomes. More girls than boys reported internalizing problems (all levels) and risky drug use (screen-intermediate and screen-positive only). More boys reported antisocial behaviour at the screen-negative and screen-intermediate levels, but at the screen-positive level, similar proportions of girls and boys displayed antisocial behaviour. CONCLUSION: The findings support the view that ADHD symptoms, as well as their negative outcomes, are dimensionally distributed in the population and that adolescent girls and boys display different risk profiles. The findings confirm that ADHD symptoms are associated with higher risk of drug misuse in girls.

External validation of models predicting the individual risk of metachronous peritoneal carcinomatosis from colon and rectal cancer.

AIM: To externally validate previously published predictive models of the risk of developing metachronous peritoneal carcinomatosis (PC) after resection of nonmetastatic colon or rectal cancer and to update the predictive model for colon cancer by adding new prognostic predictors. METHOD: Data from all patients with Stage I-III colorectal cancer identified from a population-based database in Stockholm between 2008 and 2010 were used. We assessed the concordance between the predicted and observed probabilities of PC and utilized proportional-hazard regression to update the predictive model for colon cancer. RESULTS: When applied to the new validation dataset (n = 2011), the colon and rectal cancer risk-score models predicted metachronous PC with a concordance index of 79% and 67%, respectively. After adding the subclasses of pT3 and pT4 stage and mucinous tumour to the colon cancer model, the concordance index increased to 82%. CONCLUSION: In validation of external and recent cohorts, the predictive accuracy was strong in colon cancer and moderate in rectal cancer patients. The model can be used to identify high-risk patients for planned second-look laparoscopy/laparotomy for possible subsequent cytoreductive surgery and hyperthermic intraperitoneal chemotherapy.

Pain rather than induced emotions and ICU sound increases skin conductance variability in healthy volunteers.

BACKGROUND: Assessing pain in critically ill patients is difficult. Skin conductance variability (SCV), induced by the sympathetic response to pain, has been suggested as a method to identify pain in poorly communicating patients. However, SCV, a derivate of conventional skin conductance, could potentially also be sensitive to emotional stress. The purpose of the study was to investigate if pain and emotional stress can be distinguished with SCV. METHODS: In a series of twelve 1-min sessions with SCV recording, 18 healthy volunteers were exposed to standardized electric pain stimulation during blocks of positive, negative, or neutral emotion, induced with pictures from the International Affective Picture System (IAPS). Additionally, authentic intensive care unit (ICU) sound was included in half of the sessions. All possible combinations of pain and sound occurred in each block of emotion, and blocks were presented in randomized order. RESULTS: Pain stimulation resulted in increases in the number of skin conductance fluctuations (NSCF) in all but one participant. During pain-free baseline sessions, the median NSCF was 0.068 (interquartile range 0.013-0.089) and during pain stimulation median NSCF increased to 0.225 (interquartile range 0.146-0.3175). Only small increases in NSCF were found during negative emotions. Pain, assessed with the numeric rating scale, during the sessions with pain stimulation was not altered significantly by other ongoing sensory input. CONCLUSION: In healthy volunteers, NSCF appears to reflect ongoing autonomous reactions mainly to pain and to a lesser extent, reactions to emotion induced with IAPS pictures or ICU sound.

Cross-sectional relationships between dietary fat intake and serum cholesterol fatty acids in a Swedish cohort of 60-year-old men and women.

BACKGROUND: The present study aimed to describe the relationship between self-reported dietary intake and serum cholesterol fatty acids (FAs) in a Swedish population of 60-year-old men and women. METHODS: Cross-sectional data collected in 1997-1998 from 4232 individuals residing in Stockholm County were used. Five diet scores were created to reflect the intake of saturated fats in general, as well as fats from dairy, fish, processed meat and vegetable oils and margarines. Gas chromatography was used to assess 13 FAs in serum cholesterol esters. The association between each diet score and specific FAs was assessed by percentile differences (PD) with 95% confidence intervals (CI) at the 10th, 25th, 50th, 75th and 90th percentile of each FA across levels of diet scores using quantile regression. RESULTS: Fish intake was associated with high proportions of eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA). For each point increase in fish score, the 50th PD in EPA and DHA was 32.78% (95% CI = 29.22% to 36.35%) and 10.63% (95% CI = 9.52% to 11.74%), respectively. Vegetable fat intake was associated with a high proportion of linoleic acid and total polyunsaturated fatty acids (PUFA) and a low proportion of total saturated fatty acids (SFA). The intake of saturated fats in general and dairy fat was slightly associated with specific SFA, although the intake of fat from meat was not. CONCLUSIONS: In the present study population, using a rather simple dietary assessment method, the intake of fish and vegetable fats was clearly associated with serum PUFA, whereas foods rich in saturated fats in general showed a weak relationship with serum SFA. Our results may contribute to increased knowledge about underlying biology in diet-cardiovascular disease associations.

Asthma phenotypes and lung function up to 16 years of age-the BAMSE cohort.

BACKGROUND: Asthma is a disease affecting many locations throughout the airway. Most studies have used spirometry as the primary assessment of airway obstruction, a method that may be less sensitive in regard to peripheral airway obstruction. The aim of this study was to elucidate the associations between asthma phenotypes based on age of onset and duration of symptoms, and (i) spirometry and (ii) small airway involvement measured by impulse oscillometry (IOS) in adolescence. METHODS: Children and adolescents taking part in BAMSE, a prospective birth cohort study, performed spirometry at 8 and 16 years and IOS at 16 years of age. Based on data collected in questionnaires, children were categorized into the following groups: 'never asthma', 'early transient asthma', 'early persistent asthma', and 'late onset asthma'. RESULTS: Compared with the never asthma group, all asthma groups were associated with lower FEV1 at 16 years of age (early transient-119 ml, 95% confidence interval -204 to -34; early persistent-410 ml, 95%CI -533; -287; and late onset-148 ml, 95%CI -237; -58). Between 8 and 16 years, significantly less increase in FEV1 was observed in the early persistent and late onset groups. The small airway index 'R5-20 ' was significantly associated with active asthma at 16 years, but not transient asthma. CONCLUSIONS: All asthma phenotypes studied were negatively associated with FEV1 in adolescence. IOS measurements indicated that active asthma could be associated with small airway impairments. These results provide new insights into the physiology underlying wheezing phenotypes based on age of onset and duration of disease.

Probability, rate and timing of reconstructive surgery following colectomy for inflammatory bowel disease in Sweden: a population-based cohort study.

AIM: Many patients with inflammatory bowel disease (IBD) need colectomy, but the rate of reconstructive surgery with restoration of intestinal continuity is unknown. The aim of this study was to investigate the probability, rate and timing of reconstructive surgery after colectomy in patients with IBD in a population-based setting. METHOD: The study cohort included all patients with IBD in Sweden who underwent colectomy from 2000 to 2009. Each patient was followed from admission for colectomy to admission for reconstructive surgery, date of death, migration or 31 December 2010. Kaplan-Meier survival curves and multivariable Poisson regression models were used to describe the probability, rate and timing of reconstructive surgery. RESULTS: Out of 2818 IBD patients treated with colectomy, 61.0% were male and 78.9% had ulcerative colitis. No reconstructive surgery had been performed in 1595 (56.6%) patients by the end of follow-up. Of the remaining 1223 patients, 526 underwent primary reconstructive surgery and 697 had a secondary reconstruction following a median interval of 357 days from primary surgery in the form of colectomy. The probability of reconstructive surgery was dependent on age (55.6% and 18.1% at ages 15-29 and ≥ 59 years, respectively), and the chance of reconstructive surgery was higher in hospitals that performed more than 13 colectomies for IBD per year [incidence rate ratio and 95% confidence interval 1.27 (1.09-1.49)]. CONCLUSION: Fewer than half of the patients having a colectomy for IBD underwent subsequent reconstructive surgery. Older age and low hospital volume were risk factors for no reconstructive surgery.

Individualized prediction of risk of metachronous peritoneal carcinomatosis from colorectal cancer.

AIM: The purpose of the study was to develop a tool for predicting the individual risk of metachronous peritoneal carcinomatosis after surgery for non-metastatic colorectal cancer. METHOD: Independent predictors for metachronous colorectal carcinomatosis have previously been identified using a population-based database. Predictive models for colon and rectal cancer were developed from these data. The predictive models were based on multivariable Cox proportional hazard regression and were internally validated with bootstrapping. Performance was assessed by the concordance index and calibration plots. RESULTS: In all, 8044 patients who underwent abdominal resection of colorectal cancer Stage I-III were included. The colon and rectal cancer risk score models predicted metachronous peritoneal carcinomatosis with a concordance index of 80% and 78%, respectively. Factors in the models included age, pathological pT stage, pN stage, number of examined lymph nodes (0-11, 12+), type of surgery (emergency/elective), completeness of cancer resection (R0/R1/R2), adjuvant chemotherapy (yes/no), preoperative radiotherapy and tumour location. CONCLUSION: The proposed predictive models showed high internal validity and enabled individualized prediction of peritoneal recurrence of colorectal cancer. The models may help in the planning of treatment and follow-up of patients. However, external validation is warranted to assess generalizability of the predicted absolute risks.

Prediction of cardiovascular disease by abdominal obesity measures is dependent on body weight and sex--results from two community based cohort studies.

AIM: To study waist-hip ratio (WHR), waist circumference (WC), sagittal abdominal diameter (SAD), and waist-hip-height ratio (WHHR) as predictors of CVD, in men and women stratified by BMI (cut-off ≥25). METHODS AND RESULTS: A cohort of n = 3741 (53% women) 60-year old individuals without CVD was followed for 11-years (375 CVD cases). To replicate the results, we also assessed another large independent cohort; The Malmö Diet and Cancer study - cardiovascular cohort (MDCC, (n = 5180, 60% women, 602 CVD cases during 16-years). After adjustment for established risk factors in normal-weight women, the hazard ratio (HR) per one standard deviation (SD) were; WHR; 1.91 (95% confidence interval (CI) 1.35-2.70), WC; 1.81 (95% CI 1.02-3.20), SAD; 1.25 (95% CI 0.74-2.11), and WHHR; 1.97 (95% CI 1.40-2.78). In men the association with WHR, WHHR and WC were not significant, whereas SAD was the only measure that significantly predicted CVD in men (HR 1.19 (95% CI 1.04-1.35). After adjustments for established risk factors in overweight/obese women, none of the measures were significantly associated with CVD risk. In men, however, all measures were significant predictors; WHR; 1.24 (955 CI 1.04-1.47), WC 1.19 (95% CI 1.00-1.42), SAD 1.21 (95% CI 1.00-1.46), and WHHR; 1.23 (95% CI 1.05-1.44). Only the findings in men with BMI ≥ 25 were verified in MDCC. CONCLUSION: In normal weight individuals, WHHR and WHR were the best predictors in women, whereas SAD was the only independent predictor in men. Among overweight/obese individuals all measures failed to predict CVD in women, whereas WHHR was the strongest predictor after adjustments for CVD risk factors in men.