SEARCH · PubMed Health

Results for “Prognostic model”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 883 records · Page 49Linked to original sources

PubMed2006

Prognostic index score and clinical prediction model of local regional recurrence after mastectomy in breast cancer patients.

PURPOSE: To develop clinical prediction models for local regional recurrence (LRR) of breast carcinoma after mastectomy that will be superior to the conventional measures of tumor size and nodal status. METHODS AND MATERIALS: Clinical information from 1,010 invasive breast cancer patients who had primary modified radical mastectomy formed the database of the training and testing of clinical prognostic and prediction models of LRR. Cox proportional hazards analysis and Bayesian tree analysis were the core methodologies from which these models were built. To generate a prognostic index model, 15 clinical variables were examined for their impact on LRR. Patients were stratified by lymph node involvement (<4 vs. >or =4) and local regional status (recurrent vs. control) and then, within strata, randomly split into training and test data sets of equal size. To establish prediction tree models, 255 patients were selected by the criteria of having had LRR (53 patients) or no evidence of LRR without postmastectomy radiotherapy (PMRT) (202 patients). RESULTS: With these models, patients can be divided into low-, intermediate-, and high-risk groups on the basis of axillary nodal status, estrogen receptor status, lymphovascular invasion, and age at diagnosis. In the low-risk group, there is no influence of PMRT on either LRR or survival. For intermediate-risk patients, PMRT improves LR control but not metastases-free or overall survival. For the high-risk patients, however, PMRT improves both LR control and metastasis-free and overall survival. CONCLUSION: The prognostic score and predictive index are useful methods to estimate the risk of LRR in breast cancer patients after mastectomy and for estimating the potential benefits of PMRT. These models provide additional information criteria for selection of patients for PMRT, compared with the traditional selection criteria of nodal status and tumor size.

Adult↗

PubMed1985

Postoperative survival of patients with potentially curable cancer of the colon.

The association between histopathologic findings and postoperative survival in 442 patients with potentially curable carcinomas of the colon has been analyzed using Cox's regression model. The prognostic variables included in the study were age, sex, stage of disease according to Dukes' classification and Broders' grading, as well as presence/absence at the time of operation of venous and nerve invasion. The overall five-year postoperative survival rate was 46.6 percent. Using a model including all prognostic factors, sex and Dukes' classification were not found to be associated with survival. Broders' grading and/or nerve invasion yielded only a borderline statistical significance in the model that included all factors. The invasion of veins was almost always associated with invasion of nerves.

Adult↗

PubMed2005

[The prognostic value of end-stage liver disease model in liver cirrhosis].

OBJECTIVES: The prognostic ability of the model for end-stage liver disease (MELD) has been validated in many countries, but its abilities remain uncertain in China. Our aim is to evaluate the abilities of MELD in prognosis of liver cirrhosis. METHODS: A cohort of 315 patients with liver cirrhosis were retrospectively studied and followed up at least for one year. MELD score was obtained for each patient according to the modified formula by Kamath P.S.. The area under the receiver operating characteristic (ROC) curve (AUC) was used to compare MELD and Child-Turcotte-Pugh (CTP) score and classification in predicting accuracy. Kaplan-Meier survival analysis was used to compare the mortality in subgroups ranked by the MELD score. RESULTS: The AUC values generated by the ROC curves for the MELD were 0.95, 0.85 and 0.83 respectively in predicting 3, 6 month and 1 year survival, and were all more than 0.80 in predicting longer time survival, whereas the AUC was 0.82, 0.78, 0.74 for CTP score and 0.70, 0.66, 0.61 for CTP classification respectively in predicting 3, 6 month and 1 year survival. The differences of AUCs between the MELD and CTP classification were significant in predicting 6 month and 1, 3 and 4 year survival, but were not significant in predicting other time point survival. The differences of AUCs between the MELD and CTP score were not significant in predicting survival. In the subgroups of patients ranked by MELD score lower than 10, 10 to 20, 20 to 30, and more than 30, the survival rate was significant different (P = 0.000). CONCLUSIONS: MELD is a useful prognosis indicator for the liver cirrhosis. The ability of prognosis by MELD is similar to Child-Turcotte-Pugh classification and score.

Area Under Curve↗

PubMedDate not supplied

A prognostic score for prostatic adenocarcinoma based on clinical, histological, biochemical and cytometric data from the primary tumour.

The aim of this study was to create a multivariate prognostic score for prostatic adenocarcinoma. A retrospective analysis of clinical, histological and cytometric prognostic factors in 325 cases of prostatic adenocarcinoma followed up on average over 13 years was performed. A multivariate prognostic score was built by using the independent prognostic factors. M-category, T-category, Gleason score and patient age were independent prognostic factors in the entire cohort. In MO tumors T-category, density of tumor infiltrating lymphocytes, the presence of apoptotic cells, and patient age were independent prognostic factors. In T1-2MO tumors the density of tumour infiltrating lymphocytes, mitotic index, standard deviation of maximum nuclear diameter and the serum level of acidic phosphatase were independent prognostic factors. By combining the coefficients of the regression model, a prognostic score was built for each of the tumour categories. The subsequent survival analysis based on the prognostic score indicated that it is a highly significant prognostic factor superior to individual prognostic parameters. The results show that the combination of prognostic data is a valuable tool in assessing the correct prognostic category for prostatic adenocarcinoma.

Adenocarcinoma↗

PubMed2026

Risk Prognostication After Hypomethylating Agents Combined With Venetoclax in AML: The PRISM Risk Model.

PURPOSE: As risk stratification for patients with AML treated with lower-intensity venetoclax-based therapy remains suboptimal, we developed and validated a prognostic model integrating clinical, cytogenetic, and molecular features. METHODS: We assembled a multinational data set comprising 2,092 adults with newly diagnosed AML treated with hypomethylating agents plus venetoclax (HMA + VEN). One thousand nine hundred eighteen patients with complete data were randomly divided into training (70%) and internal validation (30%) cohorts. Two independent external validation cohorts were assembled (n = 500 and n = 222). Modeling overall survival (OS), Elastic Net regression was applied in 1,000 bootstrap samples from the training cohort to select variables for a Ridge regression, which generated a continuous Prognostic Risk Integration for Survival Modeling (PRISM) score and risk categories based on tertiles (PRISM-3: low, moderate, high). These PRISM indices were then computed for the validation cohorts and compared with the 4-gene classifier (based on mutations in FLT3-ITD, N/KRAS, and TP53). RESULTS: PRISM integrated 17 clinical and genomic variables and demonstrated a linear association with OS. PRISM-3 stratified survival consistently across all cohorts (median OS: 25.1-28.8 months for low risk, 12.5-14.7 months for moderate risk, and 5.8-6.7 months for high risk; P < .001). Compared with the 4-gene classifier, PRISM-3 reassigned approximately 40% of patients (and >50% of those with favorable risk) and demonstrated significantly better discrimination in validation cohorts (C-index 0.63-0.65 v 0.59-0.61; P < .05). CONCLUSION: PRISM is a validated prognostic model for patients with AML receiving HMA + VEN that improves survival risk stratification beyond current standard tools and supports individualized, risk-adapted clinical decision making. The model, the PRISM-AML Risk Calculator, is publicly available.

Humans↗

PubMed2006

Prognosis in critical care.

Prognostic risk prediction models have been employed in the intensive care unit (ICU) setting since the 1980s and provide health care providers with important information to help inform decisions related to treatment and prognosis, as well as to compare outcomes across institutions. Prognostic models for critical care are among the most widely utilized and tested predictive models in healthcare. In this article, we review and compare mortality prediction models, including the APACHE (1981), SAPS (1984), APACHE-II (1985), MPM (1987), APACHE-III (1991), SAPS-II (1993), and MPM-II (1993). We emphasize the importance of model calibration in this domain. In addition, we present a brief review of the statistical methodology, multiple logistic regression, which underlies most of the models currently used in critical care.

Artificial Intelligence↗

PubMedDate not supplied

Effects of covariate omission and categorization when analysing randomized trials with the Cox model.

When analysing the survival of patients in comparative randomized clinical trials using the Cox proportional hazards model, important prognostic factors may be included for the adjustment of the treatment effect. In this paper we examine two of the most common misspecifications of the model: (i) an important prognostic factor is omitted from the analysis; and (ii) an important prognostic factor originally present on continuous scale is included in categorized form. Both situations may emerge from the occurrence of missing values. We investigate the properties of the maximum partial likelihood estimator of the treatment effect under this kind of misspecification. The estimate of the treatment effect is found to be asymptotically biased toward zero. For its asymptotic variance we obtain a quantity with the so-called 'sandwich' structure. Thus, variance estimation by the inverse of the second-order derivative of the likelihood is not consistent. The magnitude of overestimation or underestimation is evaluated numerically for specific settings. The precision of the treatment effect estimate under covariate omission or categorization is compared with the precision of the estimate in the correct and not misspecified model. It turns out that correct adjustment does not lead to a higher precision of the treatment effect estimate, but due to the resulting underestimation, covariate omission or categorization lead to loss of power of the test of no treatment effect.

Bias↗

PubMed2005

Prognostic factors for low back pain in patients referred for physiotherapy: comparing outcomes and varying modeling techniques.

STUDY DESIGN: Data were derived from a randomized controlled trial on the (cost-) effectiveness of the implementation of the clinical guidelines on physiotherapy for low back pain in primary care. OBJECTIVES: To describe the course of low back pain in patients who are referred to physiotherapy, to identify clinically important prognostic factors on different outcomes, and to evaluate the influence of different statistical techniques in developing a prognostic model. SUMMARY OF BACKGROUND DATA: Several studies have aimed to identify prognostic factors for low back pain in primary care. These studies focused on different outcome measures and used various statistical techniques. METHODS: Primary outcomes were perceived recovery, improvement in pain, improvement in functioning, and presence of disabling low back pain at 3 and 12 months follow-up. Multivariate logistic regression analyses were performed for each outcome variable. Two cut-off points were used to determine significance with respect to the univariate analysis, and two selection methods were used to build the final multivariate models. The resulting prognostic models were compared. RESULTS: A total of 500 patients were included. Pain and disability reduced considerably in the first 3 months, but further reduction was only modest. Prognostic factors varied for different outcomes, but the duration of the current episode was included in all models generated. Varying the statistical techniques also resulted in a different prognostic model with some change to the amount of variance explained. CONCLUSIONS: A substantial proportion of patients still experienced some pain and disability at 12 months follow-up. The most stable predictor of prognosis in low back pain was the duration of the current episode. The choice of statistical method influenced the final model; however, changes in the explained variance were small.

Adult↗

PubMed1987

Serum bile-acids in liver cirrhosis: prognostic significance evidenced by a multivariate statistical model.

In order to evaluate the prognostic value of serum bile acids (SBA) in liver cirrhosis, we compared SBA levels with a theoretical expected survival length (ESL), computed on the ground of a previously proposed and validated multivariate statistical model. We demonstrated a strict correlation between SBA levels and ESL in subjects with liver cirrhosis by means of both standard linear correlation analysis and Logrank test. Our results account for a high predictive significance of SBA levels in liver cirrhosis, even in long term prognosis.

Adult↗

PubMedDate not supplied

Prognostic factors: guidelines for investigation design and state of the art analytical methods.

The proliferation of putative prognostic factors, derived prognostic indices and computerised prediction of outcome in surgical oncology has led to some confusion over the exact methods available for deriving clinically significant prognostic factors. The realisation that the interaction between factors is often complex and non-linear has led to the development of new statistical techniques. The aim of this article is to review the currently available methods of analysis. A review of the relevant literature available from statistical, medical and computer science sources was performed. Information has been conveyed at a level aimed at producing a practical understanding of the techniques involved rather than their underlying mathematical basis. There are now clear guidelines for the investigation of putative prognostic factors (Table 1). The established role of linear statistical models and prognostic indices remains vitally important for the majority of diseases with many derived prognostic indices having been validated in a prospective fashion. However, in order to improve the delineation of prognostic factors other more complex methods of analysis are now being utilised. Furthermore, the recognition of complex dynamic non-linearity within biological systems has led to the increasing use of non-linear statistical techniques and artificial intelligence. As such it is incumbent upon the modern clinician to be able to understand the basic assumptions required for multivariate analysis and also to realise when alternative statistical techniques should be employed.

Artificial Intelligence↗

PubMed2004

A comparison of the single and double factor high-risk models for risk assignment of prostate cancer treated with 3D conformal radiotherapy.

PURPOSE: Two models for stratification of prostate cancer aggressiveness predominate for the purposes of daily treatment decision making. This study investigates the relationships between these two clinically popular models. METHODS: Both risk stratification models use the same definition for low risk: Gleason score (GS) 10-20 ng/mL, without the presence of any high-risk feature; high risk (HR) was defined as the presence of GS 8-10, iPSA >20, or palpation stage T3. For the double factor high risk (DF) model, IR and HR were defined as one and more than one of the following: GS >or=7, iPSA >10, or stage T3. Between April 1989 and October 2001, 1,597 patients were treated definitively with 3D conformal radiation therapy (3D-CRT) alone for prostate cancer at our institution. The main clinical endpoint was freedom from biochemical failure (FFBF). RESULTS: The 5-year actuarial FFBF rate for the low-risk group was 83%. The SF model resulted in FFBF rates of 76% and 47% for IR and HR patients respectively. The DF model resulted in FFBF rates of 70% and 52% for IR and HR patients, respectively. The FFBF rate for patients defined as IR and HR by both models was 76% and 40%, respectively. Those classified as IR by the DF model and then further subdivided into IR and HR by the SF model had a 76% and 52% 5-year FFBF rate (p = 0.0004). Those classified as HR by the DF model and then further subdivided into IR and HR by the SF model had a 71% and 40% 5-year FFBF (p = 0.0014). CONCLUSIONS: The SF model created prognostic groups with a greater internal consistency than the DF model. The SF was also better at identifying patients with high-risk prostate cancer who may benefit from a more aggressive approach.

Aged↗

PubMed2002

Prognostic factors and scoring systems in chronic myelomonocytic leukemia: a retrospective analysis of 213 patients.

Chronic myelomonocytic leukemia (CMML) is a hematologic malignancy characterized by wide heterogeneity of clinical presentation and course. CMML shares myelodysplastic characteristics with features of myeloproliferative disorders. No treatment has proven effective in modifying the natural course of the disease. To improve the prognostic assessment of clinical outcome, the associations of patient and disease characteristics with survival times of 213 patients with CMML was investigated retrospectively. Median survival was 12 months. Univariate analysis identified low hemoglobin level; low platelet count; high white blood cell, monocyte, and lymphocyte counts; presence of circulating immature myeloid cells, high percentage of marrow blasts, low percentage of marrow erythroid cells, abnormal cytogenetics, and high levels of serum lactate dehydrogenase and beta(2)-microglobulin as characteristics associated with shorter survival. Hemoglobin level below 120 g/L (12 g/dL), presence of circulating immature myeloid cells, absolute lymphocyte count above 2.5 x 10(9)/L, and marrow blasts 10% or more were independently associated with shorter survival by multivariate analysis and were used to generate a prognostic score. The model identified 4 subgroups of patients with median survival of 24, 15, 8, and 5 months for low, intermediate-1, intermediate-2, and high risk, respectively. Researchers could not confer objective evidence suggesting that arbitrary divisions of CMML by white blood cell counts into "dysplastic" and "proliferative" categories reflect clinical entities differing in the risk of acute leukemia development, although a trend of shorter survival in patients with leukocytosis was observed. The prognostic model was compared with 6 previously published scoring systems for myelodysplastic syndrome/CMML. The reported results should provide an improved assessment of prognosis in CMML.

Adult↗

PubMed2025

Habitat radiomics predicts occult lymph node metastasis and uncovers immune microenvironment of head and neck cancer.

BACKGROUND: Occult lymph node metastasis (LNM) is a key prognostic factor for patients with head and neck squamous cell carcinoma (HNSCC). This study was to establish radiomics models derived from intratumoral, peritumoral, and habitat regions for identifying occult LNM in HNSCC. METHODS: Patients with pathologically confirmed HNSCC from three medical Centers (from March 2014 to April 2024) and The Cancer Genome Atlas (TCGA) were enrolled. Center 1 was split into training (n = 330) and internal test sets (n = 154), while Center 2 and Center 3 served as the external test set (n = 183). Genomic set (n = 50) from TCGA and single-cell RNA sequencing set (n = 6) from Center 1 were used for biological analysis. We used the intratumoral, peritumoral, and habitat volumes of interest (VOIs) to extract radiomics features, respectively. Based on Logistic Regression (LR), Support Vector Machine (SVM), and Random Forest (RF) classifiers, nine radiomics models were built to confirm the optimal predictive performance. The best-performing model, along with clinical-radiologic data, was combined to develop a hybrid model. The log-rank test was used to evaluate the model's prognostic performance. Additionally, bulk and single-cell RNA sequencing were applied for investigating the biological mechanisms underlying the optimal model. RESULTS: The RF-habitat radiomics model showed the best performance, achieving AUCs of 0.835-0.919 across all datasets. Survival analysis further confirmed the prognostic value of the RF-habitat radiomics model. The RF-habitat radiomics model and the hybrid model notably surpassed the clinical model in predictive performance. Moreover, the RF-habitat radiomics model was associated with the abundance level of exhaustion-associated CD8 + T cells, uncovering the immune microenvironment characteristics contributing to occult LNM in HNSCC. CONCLUSIONS: The RF-habitat radiomics model demonstrated excellent performance for predicting occult LNM in HNSCC across three cohorts, providing a non-invasive solution for occult LNM. Furthermore, radiogenomic analysis further revealed the biological associations of the model, primarily related to T cell dysfunction.

Humans↗

PubMed1992

Comparison of prognostic indices for burns and assessment of their accuracy.

Logistic regression was used to create four individual models of prognostic indices based on our patient population. The simplest index using only three factors, age, extent of full skin thickness burn and inhalation injury, proved to be best. These models were also compared with established indices, like Baux's, Bull's and Roi's. For a specificity below 80 per cent, Baux's rule fits as well as Roi's index does. Changes in treatment modalities and improvement in therapy will influence the accuracy of an index. Indices must be adapted to changing conditions to avoid increasing false prediction of death.

Adolescent↗

PubMed2006

Prognostic variables and prognostic groups for malignant melanoma. The information from Cox and Classification And Regression Trees analysis: an Italian population-based study.

The common way to analyse the prognostic role of selected variables in cutaneous melanoma patients is by means of Cox proportional hazard model. The prognostic effect of the simultaneous presence of more than one independent variable in the same patient is, however, difficult to establish. This hampers the possibility of tailoring a survival expectance for a selected patient as well as to communicate it to the patient himself/herself. The objectives of the study were to compare information on cutaneous melanoma prognosis from multivariate Cox proportional hazard model and from Classification And Regression Trees analysis. Classification And Regression Trees analysis is an automatic method that splits data by means of a binary recursive process creating a 'tree' of groups with different profiles according to the analysed outcome, for example, the risk of death. This approach automatically produces data that is easily interpreted by clinicians. A total of 1403 invasive cutaneous melanoma patients, 1110 from the Tuscan Cancer Registry and 293 from the Reggio Emilia Cancer Registry, Italy, were included. Cases were incident during 1996-2001 and followed up at the end of 2003. Cox proportional hazard model and Classification And Regression Trees analysis were applied to the following variables: age, sex, Breslow thickness, Clark level, registry, subsite and morphologic type. The Classification And Regression Trees analysis identified 10 categories with statistically different survival; this results were summarized into six classes of different risks based on Breslow thickness, age and sex. The best prognostic group (5-year observed survival, 98.1%) included those subjected with Breslow less than 0.94 mm and age 19-44 years. The same thickness but an older age (50-69 years) was associated with a statistically significant different prognosis (5-year observed survival, 92.8%). The Cox proportional hazard model found sex, age, Breslow thickness, Clark and morphologic type to have a significant independent prognostic value. In conclusion, compared with the conventional approach based on Cox hazard model, Classification And Regression Trees analysis produces data closer to the clinical need of defining the prognostic profile of a specific patient. This may help the clinician both in the communication of risk and in the follow-up strategy.

Adult↗

PubMed2002

Prognostic-factors-based risk-stratification model for invasive urothelial carcinoma of the urinary bladder in Taiwan.

Objectives. To develop a prognostic-factors-based predictive model for invasive urothelial carcinoma of the urinary bladder derived from statistical comparison of clinical characteristics.Methods. The medical records for patients with invasive urinary bladder urothelial carcinoma were reviewed. Clinical data for age, sex, serum lactate dehydrogenase, creatinine, albumin, alkaline phosphatase, alanine aminotransferase, total bilirubin and hemoglobin levels, white blood cell and platelet counts, positive urine cytology, Eastern Cooperative Oncology Group performance status score, tumor size, histologic grading, T stage, presence of lymph node metastases, squamous differentiation, hydronephrosis, prostatic involvement, Charlson comorbidity index, surgical procedures, and adjuvant chemotherapy status were recorded. Univariate and multivariate analyses were performed to test independent factors for prediction of survival and disease recurrence.Results. After univariate and multivariate analyses, six independent prognostic factors were found: T stage, grading, prostatic involvement, Eastern Cooperative Oncology Group performance status score, and pretreatment serum creatinine and albumin levels. A scoring system was developed on the basis the relative risk associated with the proposed prognostic factors and patients were stratified into three groups according to their scores, with statistically significant prognostic differences revealed for each of the between-group comparisons. Independent factors affecting recurrence-free survival and best predicted disease recurrence were pretreatment serum creatinine, T stage, and surgical procedure.Conclusions. This prognostic-factors-based risk-stratification model for invasive urothelial carcinoma of the urinary bladder may help clinicians predict outcome and select the most appropriate therapeutic modalities. The incidence of recurrent disease is significantly higher for patients with poor renal function before treatment or advanced T stage and those undergoing transurethral tumor resection instead of radical cystectomy.

Aged↗

PubMed1996

Predicting the long-term outcome of schizophrenia.

Predictors of long-term (13 year) outcome of schizophrenia are reported for a representative cohort of 'treated incidence' patients ascertained on their first contact with Nottingham psychiatric services between 1978-80. An initial (baseline) model including previously reported predictors of 2-year outcome (age, gender, ever married, acuteness of onset) and length of untreated illness was used to predict a range of outcome measures covering the domains of disability, psychopathology, hospitalization, employment, social activity, and global outcome. This model demonstrated significant prognostic ability across all non-hospitalization outcomes under both ICD-10 and ICD-9 diagnoses of schizophrenia, but was attenuated under broad (ICD-9 and CATEGO S, P or O) and restrictive (S+) diagnostic classifications. Female gender predicted more favourable outcome under all diagnostic classifications except S+. In an extended analysis, the addition of initial 2-year course type substantially increased the prognostic ability of the model under all diagnostic classifications and enabled over 30% of the variance in global ratings of disability and symptoms to be predicted. In this extended model female gender predicted more favourable outcome over and above the effect of course type, across most domains under ICD-10, and for disability and psychopathology under other diagnostic classifications. The inclusion of measures of psychopathology at the time of first assessment, pre-morbid functioning, and duration of index admission conferred only marginal additional predictive ability for respective outcomes in the domains of psychopathology, social activity, employment and hospitalization. Hospitalization during the past year was the most difficult outcome to predict under any model suggesting that resource utilization represents the 'administrative outcome' of schizophrenia and serves as a poor proxy for broader concerns in the era of community care. These data demonstrate that key demographic variables and the mode of onset influence the long-term course of schizophrenia, but that early course type is a particularly strong predictor.

Adult↗

PubMed1995

Peripheral neuropathy with necrotizing vasculitis in rheumatoid arthritis. A clinicopathologic and prognostic study of thirty-two patients.

OBJECTIVE: To examine the clinicopathologic features of the noncompressive neuropathies in rheumatoid arthritis (RA). METHODS: We studied 32 patients with RA and peripheral neuropathy whose nerve and/or muscle biopsy specimens exhibited necrotizing vasculitis. Morphologic analysis of nerve specimens included light and electron microscopy studies and teased fiber preparation. Survival was evaluated, and the prognostic values of clinical, biologic, and pathologic features were assessed by Cox proportional hazards model. A prognostic assessment based on the significant variables was devised to estimate the probability of survival of any individual patient. RESULTS: Epi- and/or perineurial vasculitis was observed with the same frequency in the 17 patients with sensory and motor deficit and the 15 patients with sensory neuropathies and was associated with axonal degeneration of an average of 77.7% of the nerve fibers. The mean followup was 7.2 years, and the overall survival rate at 5 years was 57%. A full prolonged remission of the vasculitis was observed in 53% of the patients; relapse occurred in 25%. The factors correlated with mortality, in decreasing order of significance, were clinical cutaneous vasculitis (P = 0.0003), neuropathy affecting 3 or 4 limbs (P = 0.03), and depressed level of C4 (P < 0.05). The prognostic assessment indicated a wide range of 5-year probabilities of survival, from < 1% to 93%. CONCLUSION: Necrotizing vasculitis is responsible for the different patterns of noncompressive neuropathies in RA, including mononeuritis multiplex and distal symmetric sensory or sensorimotor neuropathy. Cutaneous vasculitis, multifocal neuropathy, and depressed C4 level were the 3 independent variables which best predicted mortality. We propose a prognostic assessment according to these variables, to stratify patients to receive more aggressive or less aggressive therapy.

Adult↗