Review of case-mix corrected survival curves

Survival is an end point of immense interest in cardiothoracic research. In observational studies, the comparison of survival between groups of patients is usually accomplished using a toolbox that includes Kaplan-Meier survival curves and the Cox model. The Cox model yields comparisons between groups adjusted for case-mix differences, whereas the Kaplan-Meier is a plot of survival over time without adjustment. During the past decade, new methods have emerged for case-mix adjustment of survival curves and are increasingly being used in cardiothoracic research. The purpose of this report is to describe, illustrate, and review several approaches to case-mix adjusted survival (or event-free) curves.     

Immortal Time Bias in the Analysis of Time-to-Event Data in Orthopedics

Many outcomes in arthroplasty research are analyzed as time-to-event outcomes using survival analysis methods. When comparison groups are defined after a time-delayed exposure or intervention, a period of immortal time arises and can lead to biased results. In orthopedics research, immortal time bias often arises when a minimum amount of follow-up is required for study inclusion or when comparing outcomes in staged bilateral vs unilateral arthroplasty patients. We present an explanation of immortal time and the associated bias, describe how to correctly account for it using proper data preparation and statistical techniques, and provide an illustrative example using real-world arthroplasty data. We offer practical guidelines for identifying and properly handling immortal time to avoid bias. Please visit the following https://youtu.be/58p8w5o-ci4 for a video that explains the highlights of the paper in practical terms.     

Competing Risk Analysis: What Does It Mean and When Do We Need It in Orthopedics Research?

Most orthopedic studies involve survival analysis examining the time to an event of interest, such as a specific complication or revision surgery. Competing risks commonly arise in such studies when patients are at risk of more than one mutually exclusive event, such as death, or when the rate of an event depends on the rates of other competing events. In this article, we briefly describe the survival analysis censoring methodology, common fatal and nonfatal competing events, and define circumstances where standard survival analysis can fail in the setting of competing risks with real-world examples from orthopedics. Please visit the following https://youtu.be/ifj_Mm3eGu8 for a video that explains the highlights of the paper in practical terms.     

Avoiding Systematic Bias in Orthopedics Research Through Informed Variable Selection: A Discussion of Confounders,Mediators, and Colliders

There are 3 common variable types in orthopedic research—confounders, colliders, and mediators. All 3 types of variables are associated with both the exposure (eg, surgery type, implant type, body mass index) and outcome (eg, complications, revision surgery) but differ in their temporal ordering. To reduce systematic bias, the decision to include or exclude a variable in an analysis should be based on the variable’s relationship with the exposure and outcome for each research question. In this article, we define 3 types of variables with case examples from orthopedic research. Please visit the following https://youtu.be/V-grpgB1ShQ for videos that explain the highlights of the article in practical terms.     

Living With Survival Analysis in Orthopedics

Time to event data occur commonly in orthopedics research and require special methods that are often called “survival analysis.” These data are complex because both a follow-up time and an event indicator are needed to correctly describe the occurrence of the outcome of interest. Common pitfalls in analyzing time to event data include using methods designed for binary outcomes, failing to check proportional hazards, ignoring competing risks, and introducing immortal time bias by using future information. This article describes the concepts involved in time to event analyses as well as how to avoid common statistical pitfalls. Please visit the following https://youtu.be/QNETrx8B6IU and https://youtu.be/8SBoTr9Jy1Q for videos that explain the highlights of the paper in practical terms.     

Measurement Error and Misclassification in Orthopedics: When Study Subjects are Categorized in the Wrong Exposure or Outcome Groups

Datasets available for orthopedic research often contain measurement and misclassification errors due to errors in data collection or missing data. These errors can have different effects on the study results. Measurement error refers to inaccurate measurement of continuous variables (eg, body mass index), whereas misclassification refers to assigning subjects in the wrong exposure and/or outcome groups (eg, obesity categories). Misclassification of any type can result in underestimation or overestimation of the association between exposures and outcomes. In this article, we offer practical guidelines to avoid, identify, and account for measurement and misclassification errors. We also provide an illustrative example on how to perform a validation study to address misclassification based on real-world orthopedic data. Please visit the following https://youtu.be/9-ekW2NnWrs or videos that explain the highlights of the article in practical terms.     

Different competing risks models for different questions may give similar results in arthroplasty registers in the presence of few events

Background and purpose — In arthroplasty registry studies, the analysis of time to revision is complicated by the competing risk of death. There are no clear guidelines for the choice between the 2 main adjusted analysis methods, cause-specific Cox and Fine–Gray regression, for orthopedic data. We investigated whether there are benefits, such as insight into different aspects of progression to revision, to using either 1 or both regression methods in arthroplasty registry studies in general, and specifically when the length of follow-up is short relative to the expected survival of the implants.

Patients and methods — Cause-specific Cox regression and Fine–Gray regression were performed on total hip (138,234 hips, 124,560 patients) and knee (139,070 knees, 125,213 patients) replacement data from the Dutch Arthroplasty Register (median follow-up 3.1 years, maximum 8 years), with sex, age, ASA score, diagnosis, and type of fixation as explanatory variables. The similarity of the resulting hazard ratios and confidence intervals was assessed visually and by computing the relative differences of the resulting subdistribution and cause-specific hazard ratios.

Results — The outcomes of the cause-specific Cox and Fine–Gray regressions were numerically very close. The largest relative difference between the hazard ratios was 3.5%.

Interpretation — The most likely explanation for the similarity is that there are relatively few events (revisions and deaths), due to the short follow-up compared with the expected failure-free survival of the hip and knee prostheses. Despite the similarity, we recommend always performing both cause-specific Cox and Fine–Gray regression. In this way, both etiology and prediction can be investigated.     

Conditional Survival After Surgical Treatment of Melanoma: An Analysis of the Surveillance, Epidemiology, and End Results Database

Background

Survival curves following surgical treatment of cutaneous melanoma are heavily influenced by early deaths. Therefore, survival estimates may be misleading for long-term cancer survivors. We examined whether conditional survival (CS) is more accurate in predicting long-term melanoma survival.

Methods

We used the Surveillance, Epidemiology, and End Results database (1992–2005) to identify patients who underwent surgical treatment for melanoma. We included patients with T2–T4 disease and with known nodal status. Patients were stratified into low-risk (T2-3N0M0) and high-risk (T4N0M0 or T2-4N1-3M0) categories. We defined CS as time-specific estimates conditioned on living to a certain point in follow-up, and calculated 10-year cancer-specific survival curves conditioned on annual survival. We adjusted for potential confounders using a Cox proportional hazards regression model (α = 0.05).

Results

A total of 8647 patients met inclusion criteria (low-risk, 5987 [69.2%]; high-risk, 2660 [30.8%]). At diagnosis, low-risk patients had a significantly better 10-year survival rate (low-risk, 79.6%; high-risk, 41.2%; P < 0.001). On CS analysis, survival differences remained until 8 years after treatment, after which 10-year cancer-specific survival rates were no longer significantly different (P = 0.51) for low-risk (95.4%) and high-risk (91.7%) groups. Multivariate analysis demonstrated that age, gender, location, and ulceration (initial predictors of survival) were no longer predictive after 8 years of survival.

Conclusions

For patients who survive 8 years after surgical treatment of melanoma, CS data become discordant with traditionally used estimates. Our findings have important implications for patient counseling, as high-risk melanoma survivors may require no more intensive surveillance than low-risk survivors 8 years after treatment.     

Functional status‐based risk–benefit analyses of high‐KDPI kidney transplant versus dialysis

Yearly, over half of deceased‐donor kidneys with kidney donor profile index (KDPI) > 85 were discarded, yet they could improve survival outcomes for dialysis patients. The potential risk of high‐KDPI kidney transplant (KT) depends on the patient's overall health summarized by functional status, which should be examined. The analyzed cohort consisted of adult deceased‐donor KT candidates on dialysis listed in 2005–2014. A multivariate Cox proportional hazards model was fitted with functional status, measured using Karnofsky Performance Score (KPS), and transplant status as time‐varying covariates. Derived from the Cox model, survival curves were analyzed to compare the survival outcomes between dialysis and transplant with different kidney qualities across three different KPS strata: 10–40, 50–70, and 80–100. With KDPI 0–99 KT, KPS 10–40 patients will survive ≥4.38 years median compared with 3.21 years median if they remained on dialysis. For KPS 50+ patients, the median survival years increase from 5.82 to 6.60 years on dialysis to ≥7.83 years after KDPI < 100 KT. The risk‐adjusted analyses suggested that patients are expected to benefit more from KDPI 81–99 KT than from remaining on dialysis.     

Better Prognosis of Senile Patients with Intertrochanteric Femoral Fracture by Treatment with Open Reduction Internal Fixation than by Hip Arthroplasty

ABSTRACT

Purpose: To compare the postoperative survival and mortality rates in intertrochanteric femoral fracture (IFF) patients who underwent either open reduction internal fixation (ORIF) or hip arthroplasty. Methods: Clinical data from senior patients who had IFF and underwent ORIF or hip arthroplasty were analyzed retrospectively. Survival curves were compared between groups with Kaplan–Meier method and log-rank test. Significant independent prognostic factors were identified by Cox multivariate regression analysis. Results: All patients recovered fully post-surgery. Although 31 patients died during the follow-up period (ORIF, mean 45.4 months; arthroplasty, mean 51.6 months), mortality rate did not differ significantly between the groups. The 1-yr and 2-yr survival rate estimates for the ORIF group were 92.2%, and 86%, respectively; they were 85% and 74% for the arthroplasty group. Average survival lengths for ORIF and arthroplasty groups were 88 and 67 months, respectively. The effect of surgical approaches on survival differed significantly (log-rank test c2 = 6.402, p = 0.011). Multivariate Cox regression model indicated that surgical choice (p = 0.036) was a significant independent risk factor for the prognosis of senile IFF, even with adjustment for age (p = 0.002). Conclusion: The overall postoperative prognosis was superior in senile IFF patients treated with ORIF.     

Model for end‐stage liver disease‐sodium and survival benefit in liver transplantation

There are currently no studies calculating the survival benefit of liver transplantation (LT) according to model for end‐stage liver disease‐sodium (MELD‐Na) and based on the competing risk (CR) method. We enrolled consecutive adult patients with chronic end‐stage liver disease entering the waiting list (WL) for primary LT (WL group = 337) and undergoing LT (LT group = 220) in the period 2006–2009. Two independent multivariable regressions (WL and LT models) were created to measure the prognostic power of MELD‐Na with respect to MELD. For the WL model, both Cox and CR multivariable analyses were performed. Estimates were finally included in a Markov model to calculate 3‐year survival benefit. WL Cox model: MELD‐Na (P < 0.0001) and MELD (P < 0.0001) significantly predicted survival. WL CR model: MELD‐Na (P = 0.0045) and MELD (P = 0.0109) significantly predicted survival. LT Cox model: MELD‐Na (P = 0.7608) and MELD score (P = 0.9413) had not correlation with survival. Benefit model: MELD and MELD‐Na had an overlapping significant impact on 3‐year survival benefit; CR method determined a significant decrease in 3‐year life expectancy (LE) estimations. MELD‐Na and MELD scores similarly predicted 3‐year LT survival benefit, but the gain in LE is significantly lower when a CR method is adopted.     

P-Values and Power in Orthopedic Research: Myths and Reality

The results of statistical tests in orthopedic studies are typically reported using P-values. If a P-value is smaller than the pre-determined level of significance (eg, < .05), the null hypothesis is rejected in support of the alternative. This automaticity in interpreting statistical results without consideration of the power of the study has been denounced over the years by statisticians, since it can potentially lead to misinterpretation of the study conclusions. In this paper, we review fundamental misconceptions and misinterpretations of P-values and power, along with their connection with confidence intervals, and we provide guidelines to orthopedic researchers for evaluating and reporting study results. We provide real-world orthopedic examples to illustrate the main concepts. Please visit the following https://youtu.be/bdPU4luYmF0 for videos that explain the highlights of the paper in practical terms.     

Five-year cause-specific survival after meningioma surgery. A nationwide population-based study

BackgroundSurvival after meningioma surgery is often reported with inadequate allowance for competing causes of death.MethodsWe processed the French administrative medical database (Système National des Données de Santé: SNDS), to retrieve appropriate cases of surgically treated meningioma. Cause-specific survival in meningioma-related death was analyzed with the Fine & Gray (F&G) and cause-specific (CS) Cox models to identify associated factors.ResultsFive-year cumulative incidence was 2.85% for meningioma-related death and 6.3% for unrelated death (P < 0.001). In the adjusted F&G and cause-specific Cox regression models for meningioma-related death, gender, age at surgery, co-morbidities, neurofibromatosis type 2, tumor insertion, tumor grade, cerebrospinal fluid (CSF) shunt insertion, preoperative embolization and need for redo surgery for recurrence emerged as independent prognostic factors of cause-specific survival (CSS) in meningioma-related death.ConclusionAt 5 years, the risk of meningioma-unrelated death was 2.21-fold greater than the risk of dying from the meningioma disease. Five-year CSS after meningioma surgery was greater in younger adults with benign spinal meningioma with low comorbidity. Those with malignant cranial tumor requiring preoperative embolization or CSF shunting for associated hydrocephalus and with severely degraded overall health status showed a significantly increased risk of meningioma-related death. Redo surgery for recurrence failed to improve the risk of meningioma-related death. We recommend the use of net survival methods such as CSS in meningioma studies where unrelated mortality is predominant, as this approach results in more accurate estimates of disease risk and associated predictors.     

Former smoking and early and long‐term graft outcome in renal transplant recipients: a retrospective cohort study

Smoking is associated with unfavourable outcome in solid‐organ transplant recipients. Nicotine may predispose to kidney injury by increasing oxidative stress. We hypothesized that former smoking negatively affects graft outcome in kidney transplant recipients and especially those with delayed graft function (DGF). We included adult recipients of a kidney transplant between 1 January 2003 and 1 October 2015 at Ghent University Hospital and recorded outcomes until 31 October 2015. We used Kaplan–Meier and multivariable Cox proportional hazard analysis to examine the relationship between former smoking at the time of transplantation and the incidence of 10‐year graft loss with and without censoring for death in 1013 participants. We evaluated mean differences in eGFR over time by a random intercept and slope model, considering a linear time effect. After adjusting for potential confounders, a history of smoking was associated with an increased hazard of graft loss (adjusted hazard ratio (aHR) 1.60; 95% CI: 1.17–2.17; P = 0.003) and death‐censored graft loss (aHR 2.29; 95% CI: 1.41–3.72; P = 0.001). The linear time trend of eGFR was different between former and never smokers (P = 0.001). To conclude, former smoking exerts long‐lasting negative effects on graft outcome and this independent of DGF.     

Clinically resectable acinar cell carcinoma of the pancreas: Is there a benefit to adjuvant systemic therapy?

BackgroundPrior studies of adjuvant systemic therapy in pancreatic acinar cell carcinoma have been underpowered.MethodsWe queried the National Cancer Data Base to identify patients presenting with resectable (clinical stage I and II) acinar cell carcinoma between 2004 and 2015. Multivariable Cox Regression was used to evaluate the association between overall survival and systemic therapy.Results298 patients met inclusion criteria: 38 received no treatment; 60 received systemic therapy alone; 84 received surgical resection alone; 116 underwent resection followed by adjuvant systemic therapy. On univariate analysis, resection was associated with a survival benefit compared to no treatment and systemic therapy alone (3-year overall survival: 57% vs. 26%, p < 0.001). On Cox analysis, use of adjuvant therapy was associated with a survival benefit compared to resection alone (HR 0.54, 95% CI: 0.33–0.89).ConclusionsAdjuvant therapy is associated with a significant survival benefit in patients with resectable acinar cell carcinoma.     

Predicting Clinical and Economic Outcomes After Liver Transplantation Using the Mayo Primary Sclerosing Cholangitis Model and Child-Pugh Score

Issues in the selection and timing of liver transplantation for primary sclerosing cholangitis (PSC) remain controversial. Although the Child-Pugh classification (CP) score and Mayo PSC model have similar abilities to estimate pretransplantation survival, a comparison of these 2 scores in predicting survival after liver transplantation has not been conducted. The aim of this study is to compare the Mayo PSC model and CP score in predicting patient survival and related economic outcomes after liver transplantation. Data from 128 patients with PSC, identified from the NIDDK database, were used to calculate patient-specific Mayo PSC and CP scores before transplantation. Levels reflecting a poor outcome were defined a priori. Receiver operating characteristic (ROC) curves and regression methods (Cox proportional hazards and linear regression models) were used to assess the relationship between these 2 scores and 5 post liver transplantation outcome measures. CP score was found to be a significantly (P < .05) better predictor of death 4 months or less after liver transplantation than: (a) length of hospital stay >21 days (or death before discharge) and (b) resource utilization >200,000 units (measured by area under the ROC curve). The Cox model identified statistically significant (P < .05) associations between CP score and each outcome after adjusting for the Mayo PSC risk score. Similar results were not observed for the Mayo PSC model when adjusted for CP score. Among patients with PSC undergoing liver transplantation, CP score was a better overall predictor of both survival and economic resource utilization compared with the Mayo PSC model. (Liver Transpl 2000;6:753-758.)     

The survival benefit of postmastectomy radiotherapy for breast cancer patients with T1-2N1 disease according to molecular subtype

ObjectiveTo evaluate the significance of postmastectomy radiotherapy (PMRT) in female breast cancer patients with T1-2N1M0 disease according to molecular subtypes and other risk factors.MethodWe conducted a retrospective cohort-based study utilizing the Surveillance, Epidemiology, and End Results database. Patients who were diagnosed with T1-2N1M0 invasive breast cancer and received mastectomy between 2010 and 2014 were enrolled in our study. Overall survival (OS) was calculated with Kaplan-Meier method, and multivariant Cox hazard model was conducted to identify the impact of PMRT according to molecular subtypes and other risk factors. Propensity score matching (PSM) was applied to balance measurable confounders.ResultsOf all the 16,521 enrolled patients, 5775 (35.0%) cases received PMRT. The distribution of molecular subtype is 71.4% for Luminal A, 13.2% for Luminal B, 5.1% for HER2 enriched, and 10.3% for TNBC. The OS was significantly better for patients in PMRT group than the Non-PMRT group (P < 0.0001). Stratified by molecular subtype, PMRT significantly prolonged survival in Luminal A patients (HR: 0.759, 95% CI: 0.651–0.884, P < 0.001), Yet it brought no significant survival advantage in Luminal B, TNBC or HER2 enriched subtype (P = 0.914, P = 0.124, P = 0.103, respectively). Also, PMRT bore prognostic significance among those patients who were older than 56 years old, single, white, exempt from reconstruction and chemotherapy, and were with ductal, GradeⅡtumor (all P < 0.05). After PSM, the survival benefit of PRMT sustained in Luminal A patients with T1 tumor concomitant with one positive lymph node.ConclusionOur study demonstrates a beneficial impact for PMRT on overall survival among Luminal A subtype breast cancer patients with T1-2N1 disease. The selection of PMRT should be stratified by molecular subtype and other risk factors.     

Positive microscopic surgical margins: Is there an association with survival in resected small gastrointestinal stromal tumors?

BackgroundFew studies evaluate the relationships between surgical approach, histologic margin, and overall survival in gastrointestinal stromal tumor. We test the hypothesis that margin positive resection is associated with compromised overall survival.MethodsWe queried the National Cancer Data Base to identify patients undergoing resections for gastrointestinal stromal tumors ≤3 cm in size between 2010 and 2015. Multivariable logistic regression was used to identify factors associated with positive microscopic margins on final pathology. Cox proportional hazard methods were used to evaluate factors associated with overall survival.Results2064 patients met inclusion criteria; 135 (6.5%) had a microscopically positive surgical margin. On multivariable regression, minimally invasive approach was not associated with risk of a positive margin (OR 1.06 95% CI [0.71, 1.59]). On Cox analysis, positive margin status was not associated with OS (R1: 1.03, CI [0.46–2.31], reference R0).ConclusionsPositive microscopic surgical margins are not associated with compromised overall survival in patients undergoing resection of small gastrointestinal stromal tumors. Minimally invasive surgical approaches do not compromise oncologic outcomes in these cases.     

Survival of patients with small cell carcinoma of the prostate during 1973-2003: a population-based study

Study Type – Therapy (inception cohort) Level of Evidence 2a What's known on the subject? and What does the study add? Small cell carcinoma of the prostate is a lethal disease. Survival data is currently based on case reports and single institution case series which give limited information on its prognostic factors. In this large population‐based study, we provide more robust estimates of survival and have defined the prognostic factors.

OBJECTIVE

• ? To describe the survival of patients with primary small cell carcinoma (SCC) of the prostate and assess prognostic factors based on a large population sample.

PATIENTS AND METHODS

• ? A total of 241 cases of SCC of the prostate were reported to the Surveillance, Epidemiology, and End Results (SEER) registries from 1973 to 2003 of which 191 cases were included in our study.
• ? We used the Kaplan–Meier method for estimating survival, and Cox proportional hazard regression modelling to evaluate prognostic variables.

RESULTS

• ? The overall age‐adjusted incidence rate was 0.278 per 1 000 000 (95% confidence interval, 0.239–0.323).
• ? In all, 60.5% presented as metastatic disease compared with 39.5% who presented as local/regional disease (P= 0.012).
• ? The 12, 24, 36, 48 and 60 months observed survival rates were 47.9%, 27.5%, 19%, 17% and 14.3% respectively.
• ? On univariate analyses, age <60, concomitant low‐grade prostatic adenocarcinoma, absence of metastasis, prostatectomy and radiation therapy were favourable prognostic factors.
• ? In multivariate regression modelling, age, pathology and stage were strong predictors of survival.

CONCLUSIONS

• ? Using the SEER database, we present the largest study describing the epidemiology of primary SCC of the prostate.
• ? We found age, concomitant low‐grade prostatic adenocarcinoma, and stage of the disease to be the strongest predictors of survival for patients with prostatic SCC.
• ? Future studies evaluating a broader range of clinical and molecular markers are needed to refine the prognostic model of this relatively rare disease.

     

Management of HR+/HER2+ lobular breast cancer and trends do not mirror better outcomes

PurposeTreatment protocols for invasive lobular breast cancer (ILC) have largely followed those for invasive ductal breast cancer. This study compares treatment outcomes of endocrine therapy versus combined chemo-endocrine therapy in hormone-receptor-positive (HR+), HER2-positive (HER2+) ILC tumors in a large national registry.MethodsWe sampled the National Cancer Database (2010–2016) for female patients with stages I-III, HR+/HER2+ ILC who underwent surgery. Cochran-Armitage trend test examined trends of treatment regimen administration: Surgery only (S), chemotherapy (C), endocrine therapy (ET), and combined chemo-endocrine therapy (CET), with or without anti-HER2 therapy. Cox proportional hazard model were used to compare overall survival (OS) across ET and CET cohorts, stratifying for anti-HER2 therapy, before and after propensity score match of cohorts (2013–2016). Kaplan-Meier (KM) survival curves were also produced.ResultsN=11,421 were included. 58.7% of patients received Anti-Her2 therapy after 2013. CET conferred better OS over ET in the unmatched (adjusted-5-year-OS: 92.5% vs. 81.1%, p<0.001) and PS-matched (90.4% vs. 84.5%, p=0.001) samples. ET caused lower OS in patients who received Anti-Her2 therapy (HR: 2.56, 95% CI: 1.60–4.12, p<0.001) and patients who did not (HR: 1.84, 95% CI: 1.21–2.78, p=0.004), as compared to CET on multivariable analysis. KM modeling showed highest OS in the CET cohort who received Anti-Her2 (93.0%), followed by the CET cohort who did not receive Anti-Her2 (90.2%) (p=0.06).ConclusionChemotherapy followed by endocrine therapy and Anti-Her2 therapy was shown to be the most effective treatment modality in HR+/HER2+ ILC, contrasting previous data on the inconclusive benefit of chemotherapy in patients with ILC.