Umeå University's logo

Aim: The selective use of defunctioning stomas in anterior resection for rectal cancer hinges on accurately predicting anastomotic leakage. The aim of this study was to develop a prediction model for use in a prospective randomized clinical trial.

Method: Colorectal Cancer Database (CRCBaSe) Sweden was used to identify patients who underwent low anterior resection for rectal cancer 2007–2021. Eligibility criteria mirrored the forthcoming SELective defunctioning Stoma Approach in low anterior resection for rectal cancer (SELSA) trial, including patients <80 years of age and with American Society of Anaesthesiologists' (ASA) physical status grade of <III; further, patients without a defunctioning stoma were excluded. The outcome comprised anastomotic leakage within 30 days or in-hospital. Candidate predictors included age, sex, ASA grade, cardiovascular disease, diabetes, body mass index (BMI), tumour stage, tumour height, and neoadjuvant therapy. Seven models were developed and internally validated using bootstrapping. A threshold of a predicted leakage risk of ≤10% was chosen for trial implementation. Validation was conducted using chart-reviewed data from a nested cohort.

Results: Of the 2727 eligible patients, 199 (7.3%) were registered with an anastomotic leakage. All models demonstrated similar performance, with prediction instability observed for risks exceeding 12.5%. The preferred model included three significant predictors: male sex (OR 2.00; 95% CI: 1.45–2.75), BMI >30 kg/m2 (OR 1.82; 95% CI: 1.21–2.74), and radiotherapy (OR 1.90; 95% CI: 1.35–2.69). The bootstrapped area under the curve (AUC) was 0.64 (95% CI: 0.62–0.65), with a negative predictive value of 94.6% (95% CI: 93.7%–95.6%). For the validation cohort, the corresponding estimates were 0.66 (95% CI: 0.59–0.74) and 89.5% (95% CI: 86.2%–92.5%).

Conclusion: Accuracy of anastomotic leakage prediction using registry-based data is moderate; however, the model's ability to rule out a >10% risk is considered appropriate for trial use.

Background: Anastomotic leakage after rectal cancer surgery is linked to reduced survival and higher recurrence rates. While an aggravated inflammatory response may worsen outcomes, few studies have explored the combined effects of leakage and inflammation.

Methods: This is a retrospective multicenter cohort study including patients operated with anterior resection for rectal cancer in Sweden during 2014–2018. Anastomotic leakage within 12 months was exposure and primary outcome was recurrence-free survival. Mediation analysis was performed to evaluate the potential effect of systemic inflammatory response, as measured by the highest postoperative C-reactive protein (CRP) level within 14 days of surgery. Confounders were chosen using a causal diagram.

Results: Some 1036 patients were eligible for analysis, of whom 218 (21%) experienced an anastomotic leakage. At the end of follow-up at a median of 61 months after surgery, recurrence-free survival amounted to 82.6% and 77.8% in the group with and without leakage, respectively. The median highest postoperative CRP value after surgery was higher in the leakage group (219 mg/l), compared with the group without leakage (108 mg/l). Leakage did not lead to worse recurrence-free survival (HR 0.66; 95% CI 0.43–0.94), and there was no apparent effect through postoperative highest CRP (HR 1.12; 95% CI 0.93–1.29).

Conclusions: In conclusion, anastomotic leakage, with its accompanying CRP increase, was not found to be associated with recurrence-free survival after anterior resection for rectal cancer in this patient cohort. Larger, even more detailed studies are needed to further investigate this topic.

Globally, stroke is the third-leading cause of mortality and disability combined, and one of the costliest diseases in society. More accurate predictions of stroke outcomes can guide healthcare organizations in allocating appropriate resources to improve care and reduce both the economic and social burden of the disease. We aim to develop and evaluate the performance and explainability of three supervised machine learning models and the traditional multinomial logistic regression (mLR) in predicting functional dependence and death three months after stroke, using routinely-collected data. This prognostic study included adult patients, registered in the Swedish Stroke Registry (Riksstroke) from 2015 to 2020. Riksstroke contains information on stroke care and outcomes among patients treated in hospitals in Sweden. Prognostic factors (features) included demographic characteristics, pre-stroke functional status, cardiovascular risk factors, medications, acute care, stroke type, and severity. The outcome was measured using the modified Rankin Scale at three months after stroke (a scale of 0-2 indicates independent, 3-5 dependent, and 6 dead). Outcome prediction models included support vector machines, artificial neural networks (ANN), eXtreme Gradient Boosting (XGBoost), and mLR. The models were trained and evaluated on 75% and 25% of the dataset, respectively. Model predictions were explained using SHAP values. The study included 102,135 patients (85.8% ischemic stroke, 53.3% male, mean age 75.8 years, and median NIHSS of 3). All models demonstrated similar overall accuracy (69%-70%). The ANN and XGBoost models performed significantly better than the mLR in classifying dependence with F1-scores of 0.603 (95% CI; 0.594-0.611) and 0.577 (95% CI; 0.568-0.586), versus 0.544 (95% CI; 0.545-0.563) for the mLR model. The factors that contributed most to the predictions were expectedly similar in the models, based on clinical knowledge. Our ANN and XGBoost models showed a modest improvement in prediction performance and explainability compared to mLR using routinely-collected data. Their improved ability to predict functional dependence may be of particular importance for the planning and organization of acute stroke care and rehabilitation.

BACKGROUND: Previous studies on disparities in healthcare and outcome have shown conflicting results. The aim of this study was to assess differences in baseline characteristics, management, and outcome in myocardial infarction (MI) patients, by country of birth.

METHODS: In total, 194 259 MI patients (64% male, 15% foreign-born) from the nationwide SWEDEHEART registry were included and compared by geographic region of birth. The primary outcome was one-year major adverse cardiovascular events (MACE) including all-cause death, MI, and stroke. Secondary outcomes were long-term MACE (up to 12 years), the individual components of MACE, 30-day mortality, management, and risk factors. Logistic regression, Cox proportional hazard models and propensity score matching (PSM), accounting for baseline differences, were used.

RESULTS: Foreign-born patients were younger, often male, and had a higher cardiovascular (CV) risk factor burden, including smoking, diabetes, and hypertension. In PSM analyses, Asia-born patients had higher likelihood of revascularisation (OR 1.16, 95% CI 1.04-1.30), statins and betablocker prescription at discharge and a 34% lower risk of 30-day mortality. Furthermore, no statistically significant differences were found in the primary outcomes except for Asia-born patients having lower risk of one-year MACE (HR 0.85, 95% CI 0.73-0.98), driven by lower mortality (HR 0.72, 95% CI 0.57-0.91). The results persisted over long-term follow-up.

CONCLUSIONS: This study shows that in a system with universal healthcare coverage in which acute and secondary preventive treatments do not differ by country of birth, foreign-born patients, despite higher CV risk factor burden, will do at least as well as native-born patients.

Aim: Previous research has indicated that preoperative beta blocker therapy is associated with a decreased risk of complications after surgery for rectal cancer. This is thought to arise because of the anti-inflammatory activity of the drug. These results need to be reproduced and analyses extended to other drugs with such properties, as this information might be useful in clinical decision-making. The main aim of this work was to replicate previous findings of beta blocker use as a prognostic marker for postoperative leakage. We also investigated whether drug exposure might induce anastomotic leaks.

Method: This is a retrospective multicentre cohort study, comprising 1126 patients who underwent anterior resection for rectal cancer between 2014 and 2018. The use of any preoperative beta blocker was treated as the primary exposure, while anastomotic leakage within 12 months of surgery was the outcome. Secondary exposures comprised angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, statins and metformin. Using multivariable regression, we performed a replication analysis with a predictive aim for beta blockers only, while adjustment for confounding was done in more causally oriented analyses for all drugs. We estimated incidence rate ratio (IRR) and relative risk (RR) with 95% confidence intervals (CIs).

Results: Anastomotic leakage occurred in 20.6% of patients. Preoperative beta blockers were used by 22.7% of the cohort, while the leak distribution was almost identical between exposure groups. In the main replication analysis, no association could be detected (IRR 0.95, 95% CI 0.68–1.33). In the causally oriented analyses, only metformin affected the risk of leakage (RR 1.59, 95% Cl 1.31–1.92).

Conclusion: While previous research has suggested that preoperative beta blocker use could be prognostic of anastomotic leakage, this study could not detect any such association. On the contrary, our results indicate that preoperative beta blocker use neither predicts nor causes anastomotic leakage after anterior resection for rectal cancer.

Background: Electrolyte disturbances and dehydration are common after anterior resection for rectal cancer with a defunctioning loop ileostomy. High-quality population-based studies on the impact of a defunctioning loop ileostomy on renal failure are lacking.

Methods: This was a nationwide observational study, based on the Swedish Colorectal Cancer Registry of patients undergoing anterior resection for rectal cancer between 2008 and 2016, with follow-up until 2017. Patients with severe co-morbidity, with age greater than 80 years, and with pre-existing renal failure were excluded. Loop ileostomy at index surgery constituted exposure, while a diagnosis of renal failure was the outcome. Acute and chronic events were analysed separately. Inverse probability weighting with adjustment for confounding derived from a causal diagram was employed. Hazards ratios (HRs) with 95 per cent c.i. are reported.

Results: A total of 5355 patients were eligible for analysis. At 5-year follow-up, all renal failure events (acute and chronic) were 7.2 per cent and 3.3 per cent in the defunctioning stoma and no stoma groups respectively. In the weighted analysis, a HR of 11.59 (95 per cent c.i. 5.68 to 23.65) for renal failure in ostomates was detected at 1 year, with the largest effect from acute renal failure (HR 24.04 (95 per cent c.i. 8.38 to 68.93)). Later follow-up demonstrated a similar pattern, but with smaller effect sizes.

Conclusion: Patients having a loop ileostomy in combination with anterior resection for rectal cancer are more likely to have renal failure, especially early after surgery. Strategies are needed, such as careful fluid management protocols, and further research into alternative stoma types or reduction in stoma formation.

Background: The aim of this study was to investigate the health and economic outcomes of a universal early intervention for parents and children, the Salut Programme, from birth to when the child completed five years of age.

Methods: This study adopted a retrospective observational design using routinely collected linked register data with respect to both exposures and outcomes from Västerbotten county, in northern Sweden. Making use of a natural experiment, areas that received care-as-usual (non-Salut area) were compared to areas where the Programme was implemented after 2006 (Salut area) in terms of: (i) health outcomes, healthcare resource use and costs around pregnancy, delivery and birth, and (ii) healthcare resource use and related costs, as well as costs of care of sick child. We estimated total cumulative costs related to inpatient and specialised outpatient care for mothers and children, and financial benefits paid to mothers to stay home from work to care for a sick child. Two analyses were conducted: a matched difference-in difference analysis using the total sample and an analysis including a longitudinal subsample.

Results: The longitudinal analysis on mothers who gave birth in both pre- and post-measure periods showed that mothers exposed to the Programme had on average 6% (95% CI 3–9%) more full-term pregnancies and 2% (95% CI 0.03-3%) more babies with a birth weight ≥ 2500 g, compared to mothers who had care-as-usual. Savings were incurred in terms of outpatient care costs for children of mothers in the Salut area ($826). The difference-in-difference analysis using the total sample did not result in any significant differences in health outcomes or cumulative resource use over time.

Conclusions: The Salut Programme achieved health gains, as a health promotion early intervention for children and parents, in terms of more full-term pregnancies and more babies with a birth weight ≥ 2500 g, at reasonable cost, and may lead to lower usage of outpatient care. Other indicators point towards positive effects, but the small sample size may have led to underestimation of true differences.

In observational studies weighting techniques are often used to overcome bias due to confounding. Modeling approaches, such as inverse propensity score weighting, are popular, but often rely on the correct specification of a parametric model wherein neither balance nor stability are targeted. More recently, balancing approach methods that directly target covariate imbalances have been proposed, and these allow the researcher to explicitly set the desired balance constraints. In this study, we evaluate the finite sample properties of different modeling and balancing approach methods, when estimating the marginal hazard ratio, through Monte Carlo simulations. The use of the different methods is also illustrated by analyzing data from the Swedish stroke register to estimate the effect of prescribing oral anticoagulants on time to recurrent stroke or death in stroke patients with atrial fibrillation. In simulated scenarios with good overlap and low or no model misspecification the balancing approach methods performed similarly to the modeling approach methods. In scenarios with bad overlap and model misspecification, the modeling approach method incorporating variable selection performed better than the other methods. The results indicate that it is valuable to use methods that target covariate balance when estimating marginal hazard ratios, but this does not in itself guarantee good performance in situations with, e.g., poor overlap, high censoring, or misspecified models/balance constraints.

Important advances have recently been achieved in developing procedures yielding uniformly valid inference for a low dimensional causal parameter when high-dimensional nuisance models must be estimated. In this paper, we review the literature on uniformly valid causal inference and discuss the costs and benefits of using uniformly valid inference procedures. Naive estimation strategies based on regularisation, machine learning, or a preliminary model selection stage for the nuisance models have finite sample distributions which are badly approximated by their asymptotic distributions. To solve this serious problem, estimators which converge uniformly in distribution over a class of data generating mechanisms have been proposed in the literature. In order to obtain uniformly valid results in high-dimensional situations, sparsity conditions for the nuisance models need typically to be made, although a double robustness property holds, whereby if one of the nuisance model is more sparse, the other nuisance model is allowed to be less sparse. While uniformly valid inference is a highly desirable property, uniformly valid procedures pay a high price in terms of inflated variability. Our discussion of this dilemma is illustrated by the study of a double-selection outcome regression estimator, which we show is uniformly asymptotically unbiased, but is less variable than uniformly valid estimators in the numerical experiments conducted. 

Background: Preoperative inflammation might cause and also be a marker for anastomotic leakage after anterior resection for rectal cancer. Available biomarker indices such as the modified Glasgow Prognostic Score (mGPS) or the C-reactive protein-to-albumin ratio (CAR) may be clinically useful for leakage assessment.

Methods: Patients who underwent anterior resection for rectal cancer during 2014–2018 from a multicentre retrospective cohort were included. Data from the Swedish Colorectal Cancer registry and chart review at each hospital were collected. In a subset of patients, preoperative laboratory assessments were available, constituting the exposures mGPS and CAR. Anastomotic leakage within 12 months was the outcome. Causally oriented analyses were conducted with adjustment for confounding, as well as predictive models.

Results: A total of 418 patients were eligible for analysis. Most patients had mGPS = 0 (84.7%), while mGPS = 1 (10.8%) and mGPS = 2 (4.5%) were less common. mGPS = 2 (OR: 4.11; 95% CI: 1.69–10.03) seemed to confer anastomotic leakage, while this was not seen for mGPS = 1 (OR 1.09; 95% CI: 0.53–2.25). A cut off point of CAR > 0.36 might be indicative of leakage (OR 2.25; 95% CI: 1.21–4.19). Predictive modelling using mGPS rendered an area-under-the-curve of 0.73 (95% CI: 0.67–0.79) at most.

Discussion: Preoperative inflammation seems to be involved in the development of anastomotic leakage after anterior resection for cancer. Inclusion into prediction models did not result in accurate leakage prediction, but high degrees of systemic inflammation might still be important in clinical decision-making.