Classifying radiographic progression status in early rheumatoid arthritis patients using propensity scores to adjust for baseline differences

A latent class model to multiply impute missing treatment indicators in observational studies when inferences of the treatment effect are made using propensity score matching

Analysts often estimate treatment effects in observational studies using propensity score matching techniques. When there are missing covariate values, analysts can multiply impute the missing data to create m completed data sets. Analysts can then estimate propensity scores on each of the completed data sets, and use these to estimate treatment effects. However, there has been relatively little attention on developing imputation models to deal with the additional problem of missing treatment indicators, perhaps due to the consequences of generating implausible imputations. However, simply ignoring the missing treatment values, akin to a complete case analysis, could also lead to problems when estimating treatment effects. We propose a latent class model to multiply impute missing treatment indicators. We illustrate its performance through simulations and with data taken from a study on determinants of children's cognitive development. This approach is seen to obtain treatment effect estimates closer to the true treatment effect than when employing conventional imputation procedures as well as compared to a complete case analysis.

Optimal methotrexate dose is associated with better clinical outcomes than non-optimal dose in daily practice: results from the ESPOIR early arthritis cohort

Background

Although methotrexate (MTX) is the consensual first-line disease-modifying antirheumatic drug (DMARD) for rheumatoid arthritis (RA), substantial heterogeneity remains with its prescription and dosage, which are often not optimal.

Objective

To evaluate the symptomatic and structural impact of optimal MTX dose in patients with early RA in daily clinical practice over 2 years.

Methods

Patients included in the early arthritis ESPOIR cohort who fulfilled the ACR-EULAR (American College of Rheumatology/European League against Rheumatism) criteria for RA and received MTX as a first DMARD were assessed. Optimal MTX dose was defined as ≥10 mg/week during the first 3 months, with escalation to ≥20 mg/week or 0.3 mg/kg/week at 6 months without Disease Activity Score in 28 joints remission. Symptomatic and structural efficacy with and without optimal MTX dose was assessed by generalised logistic regression with adjustment for appropriate variables.

Results

Within the first year of follow-up, 314 patients (53%) with RA received MTX as a first DMARD (mean dose 12.2±3.8 mg/week). Only 26.4% (n=76) had optimal MTX dose. After adjustment, optimal versus non-optimal MTX dose was more efficient in achieving ACR-EULAR remission at 1 year (OR 4.28 (95% CI 1.86 to 9.86)) and normal functioning (Health Assessment Questionnaire ≤0.5; OR at 1 year 4.36 (95% CI 2.03 to 9.39)), with no effect on radiological progression. Results were similar during the second year.

Conclusion

Optimal MTX dose is more efficacious than non-optimal dose for remission and function in early arthritis in daily practice, with no impact on radiological progression over 2 years.

Prognostic Factors of Patients with Gastric Gastrointestinal Stromal Tumor after Curative Resection: A Retrospective Analysis of 406 Consecutive Cases in a Multicenter Study

BACKGROUND/PURPOSE

Gastric gastrointestinal stromal tumors (GISTs) have a highly variable clinical course, and recurrent disease sometimes develops despite curative surgery. This study was undertaken to investigate the surgical role in treating gastric GISTs and evaluate the clinicopathological features of a large series of patients who underwent curative resection for gastric GISTs to clarify which features were independent prognostic factors.

METHODS

The clinicopathological data of 406 patients with gastric GISTs who underwent curative resection at 4 university hospitals in Daegu, South Korea, from March 1998 to March 2012 were reviewed. All cases were confirmed as gastric GISTs by immunohistochemical staining, in which CD117 or CD34 was positive. Clinical follow-up was performed periodically, and disease-free survival rates were retrospectively investigated using the medical records.

RESULTS

The mean follow-up period was 42.9 months (range: 2-166). There were 11 recurrent patients (2.7%). Due to the small number of recurrences, age, sex and location were controlled using propensity score matching before performing any statistical analysis. Tumor size, mitotic count, NIH classification, and cellularity were judged to be independent prognostic factors for recurrence by univariate analysis. In a multivariate analysis, tumor size and mitotic count were significantly and independently related to recurrence, and tumor size was determined to be the most important prognostic factor for recurrence after curative resection (hazard ratio: 1.204; p < 0.01).

CONCLUSIONS

The results of this multicenter study demonstrate that disease-free survival rates are good. Tumor size was disclosed as the most important factor for recurrence in gastric GIST patients who underwent radical resection.

A comparison of two methods of estimating propensity scores after multiple imputation

In many observational studies, analysts estimate treatment effects using propensity scores, e.g. by matching or sub-classifying on the scores. When some values of the covariates are missing, analysts can use multiple imputation to fill in the missing data, estimate propensity scores based on the m completed datasets, and use the propensity scores to estimate treatment effects. We compare two approaches to implement this process. In the first, the analyst estimates the treatment effect using propensity score matching within each completed data set, and averages the m treatment effect estimates. In the second approach, the analyst averages the m propensity scores for each record across the completed datasets, and performs propensity score matching with these averaged scores to estimate the treatment effect. We compare properties of both methods via simulation studies using artificial and real data. The simulations suggest that the second method has greater potential to produce substantial bias reductions than the first, particularly when the missing values are predictive of treatment assignment.

Evidence of the symptomatic and structural efficacy of methotrexate in daily practice as the first disease-modifying drug in rheumatoid arthritis despite its suboptimal use: results from the ESPOIR early synovitis cohort

OBJECTIVE

To describe the use of MTX in early arthritis (EA) in daily clinical practice and to evaluate its 6-month symptomatic efficacy and 12-month structural efficacy.

METHODS

Patients included in the French observational ESPOIR cohort were assessed. Evaluation of the symptomatic and structural efficacy was performed by generalized linear regression after adjustment on propensity score (PS) in the group of patients receiving at least 3 months of MTX vs the ones receiving any other treatment except LEF, SSZ or TNF inhibitors.

RESULTS

Within the first 6 months of follow-up of 777 EA patients, 59% received a DMARD, which was MTX in 68% (N = 313) of patients. The mean dose of MTX was 12.7 ± 3.8 mg/week. Only 53.7% of the patients received folic acid supplementation. MTX was initiated in patients with more active and severe disease. At 6 months, in unadjusted analysis, patients starting MTX had a significantly higher DAS-28 (3.58 vs 3.23; P = 0.001) and a significantly higher HAQ (0.60 vs. 0.48; P = 0.01) compared with controls. After adjustment by PS, there were no significant differences. Adjustment for the PS also revealed a statistically significant decrease in the radiological progression at 12 months in the MTX group [total Sharp-van der Heijde score (SHS), 1.05 ± 0.29 vs 2.02 ± 0.29, P = 0.025].

CONCLUSION

This study confirms the symptomatic and structural efficacy of MTX in EA in daily practice despite the non-optimal use of MTX, including low doses and infrequent concomitant folic acid supplementation.

Comparison of radiographic scoring methods in a cohort of RA patients treated with anti-TNF therapy

OBJECTIVE

To compare the ability of the simple erosion narrowing score (SENS) to classify radiographic progression relative to the Sharp/van der Heijde score (SHS) in a prospective cohort of anti-TNF-treated RA patients.

METHODS

Radiographs of the hands, wrists and feet of patients enrolled in a pharmacovigilance programme are performed every 2 years. These radiographs were read in chronological order by three rheumatologists and scored using the SHS. SENS scores were derived from the SHS. Additionally, one rheumatologist scored the radiographs using the SENS method only. Patients with radiographic progression in excess of the smallest detectable change were classified as progressors. The probability of agreement and κ-value between the SHS and SENS methods for determining progression was calculated.

RESULTS

A sample of 25 patients was selected from the database. The annualized mean (s.d.) change in SHS score was 6.61 U (7.48 U) and in SENS score was 2.27 U (2.17 U). Five patients were classified as progressors using SHS and seven using SENS, with a probability of agreement of 84% (κ = 0.565).

CONCLUSION

The SENS method captures radiographic progression reliably compared with the more detailed SHS method. SENS is suitable for application in clinical practice or in observational cohorts.

Equivalent responses to disease-modifying antirheumatic drugs initiated at any time during the first 15 months after symptom onset in patients with seropositive rheumatoid arthritis

OBJECTIVE

To evaluate responses by time to initiation of nonbiologic disease-modifying antirheumatic drugs (DMARD) in a DMARD-naive cohort of patients with early seropositive rheumatoid arthritis (RA).

METHODS

Subjects were categorized by the time from symptom onset to the first DMARD use (median 5.7 months, range 0.6-15.9). Subjects who started their first DMARD within 5 months of symptom onset were compared to subjects who started after 5 months. Disease Activity Scores (DAS-44) and total Sharp Score (TSS) progression rates were analyzed using Wilcoxon rank-sum and chi-square tests; multiple linear regression analysis adjusted for potential covariates. The slope of the least-squares regression line was calculated to estimate the annualized TSS progression rates.

RESULTS

Of 233 RA patients, 76% were female and mean age was 50 (SD 13) years. At DMARD start, DAS-44 was similar in all subsets within the 0.6 to 15 months' duration between symptom onset and DMARD initiation. Erosion scores tended to be higher in those who started DMARD later, but Health Assessment Questionnaire-Disability Index (HAQ-DI) scores were higher in those who started DMARD earlier. During the 2 years after DMARD initiation, improvements in HAQ-DI and DAS-44 were similar in the various duration subsets, with about 25% ever achieving DAS remission (DAS < 1.6). Radiographic progression tended to be numerically but not statistically more rapid in the earlier subsets.

CONCLUSION

Following initiation of nonbiologic DMARD therapy at various times within 15 months of symptom onset, improvements of DAS-44, HAQ-DI, remission rate, and radiographic progression rate were similar, although higher baseline erosion scores were present in those with later initiation of DMARD.