Healthcare resource utilization in patients with pulmonary hypertension associated with chronic obstructive pulmonary disease (PH-COPD): a real-world data analysis

RATIONALE

There is a lack of real-world characterization of healthcare costs and associated cost drivers in patients with pulmonary hypertension secondary to chronic obstructive pulmonary disease (PH-COPD).

OBJECTIVES

To examine (1) excess healthcare resource utilization (HCRU) and associated costs in patients with PH-COPD compared to COPD patients without PH; and (2) patient characteristics that are associated with higher healthcare costs in patients with PH-COPD.

METHODS

This study analyzed data from the IQVIA PharMetrics® Plus database (OCT2014-MAY2020). Patients with PH-COPD were identified by a claims-based algorithm based on PH diagnosis (ICD-10-CM: I27.0, I27.2, I27.20, I27.21, I27.23) after COPD diagnosis. Patients aged ≥40 years and with data available ≥12 months before (baseline) and ≥6 months after (follow-up) the first observed PH diagnosis were included. Patients with other non-asthma chronic pulmonary diseases, PH associated with other causes, cancer, left-sided heart failure (HF), PH before the first observed COPD diagnosis, or right-sided/unspecified HF during baseline were excluded. Patients in the PH-COPD cohort were matched 1:1 to COPD patients without PH based on propensity scores derived from baseline patient characteristics. Annualized all-cause and COPD/PH-related (indicated by a primary diagnosis of COPD or PH) HCRU and costs during follow-up were compared between the matched cohorts. Baseline patient characteristics associated with higher total costs were examined in a generalized linear model in the PH-COPD cohort.

RESULTS

A total of 2,224 patients with PH-COPD were identified and matched to COPD patients without PH. Patients with PH-COPD had higher all-cause HCRU and annual healthcare costs ($51,435 vs. $18,412, p<0.001) than matched COPD patients without PH. Among patients with PH-COPD, costs were primarily driven by hospitalizations (57%), while COPD/PH-related costs accounted for 13% of all-cause costs. Having a higher comorbidity burden and a prior history of COPD exacerbation were major risk factors for higher total all-cause costs among patients with PH-COPD.

CONCLUSIONS

Treatment strategies focusing on preventing hospitalizations and managing comorbidities may help reduce the burden of PH-COPD.

Hemodialysis Access Type and Access Patency Loss: An Observational Cohort Study

Rationale & Objective

Access patency outcomes for arteriovenous fistulas (AVFs) as compared with arteriovenous grafts (AVGs) in patients receiving hemodialysis (HD) who have achieved a functioning permanent access are not fully explored.

Study Design

Observational cohort study.

Setting & Population

Fee-for-service Medicare beneficiaries aged ≥18 years with kidney failure who were newly using a permanent access for maintenance HD from the United States Renal Data System (2010-2015). Patients using an oral anticoagulant were excluded.

Exposure

AVG or AVF.

Outcomes

Loss of primary unassisted, primary assisted, and secondary patency.

Analytical Approach

Outcomes were characterized using cumulative incidence curves, and HRs adjusted for sociodemographic and clinical factors were estimated for the comparison of AVF versus AVG.

Results

The cohort included 60,329 and 17,763 patients newly using an AVF and AVG, respectively, for HD. Over 3 years of follow-up, AVG users, compared to AVF users, had a higher cumulative incidence of loss of primary unassisted patency (87% vs 69%; HR, 1.56; 95% CI, 1.52-1.60), loss of primary assisted patency (69% vs 25%; HR, 3.79; 95% CI, 3.67-3.92), and loss of secondary patency (22% vs 10%; HR, 2.03; 95% CI, 1.92-2.16). Stratified analyses revealed differences by subgroups; in particular, incidence of patency loss was higher among patients who underwent prior interventions to maintain prefunctional access patency and Black patients.

Limitations

This analysis focused on outcomes occurring after first successful use of a permanent access and thus does not inform about risk of patency loss during access maturation.

Conclusions

Among patients with kidney failure who successfully used a permanent access for HD, patency loss was consistently substantially higher in those using AVGs compared with AVFs. New interventions, such as prophylactic drugs, are needed to improve access longevity and reduce the need for invasive interventions, particularly among patients unable to receive a fistula.

Medication use by US patients with pulmonary hypertension associated with chronic obstructive pulmonary disease: a retrospective study of administrative data

BACKGROUND

Pulmonary hypertension (PH) is a serious complication of chronic obstructive pulmonary disease (COPD). While clinical guidelines recommend specific drug therapies for pulmonary arterial hypertension (PAH), these drug therapies are not recommended for PH due to lung disease.

METHODS

This was a retrospective cohort study using the Optum® Clinformatics® Data Mart from January 2009-September 2019. An algorithm was designed to identify adults with ≥ 2 ICD-9-CM or ICD-10-CM diagnosis codes for PH and with ≥ 2 diagnosis codes for COPD. Sensitivity analyses were conducted among subgroups of patients with evidence of a right heart catheterization (RHC) or pulmonary function test (PFT). Patient characteristics, medications used, and durations of use of PAH and COPD medications were analyzed.

RESULTS

A total of 25,975 patients met the study inclusion criteria. Their mean age was 73.5 (SD 10.0) years and 63.8% were female. Medications targeting PAH were prescribed to 643 (2.5%) patients, most frequently a phosphodiesterase-5 inhibitor (2.1%) or an endothelin receptor antagonist (0.75%). Medications for COPD were prescribed to 17,765 (68.4%) patients, most frequently an inhaled corticosteroid (57.4%) or short-acting beta agonist (50.4%). The median durations of use ranged from 4.9 to 12.8 months for PAH medications, and from 0.4 to 5.9 months for COPD medications. Of the subgroup of patients with RHC (N = 2325), 257 (11.1%) were prescribed a PAH medication and 1670 (71.8%) used a COPD medication. Of the subgroup with a PFT (N = 2995), 58 (1.9%) were prescribed a PAH medication and 2100 (70.1%) a COPD medication.

CONCLUSIONS

Patients with PH associated with COPD were identified in a US administrative claims database. Very few of these patients received any of the medications recommended for PAH, and only about two thirds received medications for COPD.

Validation of a claims-based algorithm to identify patients with chronic thromboembolic pulmonary hypertension using electronic health record data

This study aimed to validate an algorithm developed to identify chronic thromboembolic pulmonary hypertension (CTEPH) among patients with a history of pulmonary embolism. Validation was halted because too few patients had gold-standard evidence of CTEPH in the administrative claims/electronic health records database, suggesting that CTEPH is underdiagnosed.

Patient and physician factors influence decision-making in hypercholesterolemia: a questionnaire-based survey

Background

Goal attainment of guideline-recommended low-density lipoprotein cholesterol (LDL-C) is suboptimal. Little is known about how patient factors influence physicians’ treatment decision-making in hypercholesterolemia. We examined physicians’ treatment recommendations in high-risk patients whose LDL-C remained uncontrolled despite statin monotherapy.

Methods

Physicians completed a questionnaire prior to randomization into period I of a two-period randomized controlled trial evaluating LDL-C goal attainment in patients whose LDL-C remained ≥100 mg/dL after 5 weeks’ treatment with atorvastatin 10 mg/day (NCT01154036). Physicians’ treatment recommendations were surveyed for two hypothetical and one real scenario: (1) LDL-C presumed near goal (between 100–105 mg/dL), (2) LDL-C presumed far from goal (~120 mg/dL), and (3) observed baseline LDL-C of enrolled patients. Prognostic factors considered during decision-making were identified by regression analysis. Observed lipid outcomes at the end of period I (following 6 weeks’ treatment with ezetimibe 10 mg plus atorvastatin 10 mg, atorvastatin 20 mg, or rosuvastatin 10 mg) were compared with estimated LDL-C outcomes for physicians’ treatment recommendations after 6 weeks (based on individual patients’ pre-randomization LDL-C and expected incremental change).

Results

Questionnaires were completed for 1,534 patients. No change in therapy, or double atorvastatin dose, were frequently recommended, even when LDL-C was far from goal (6.5% and 52.2% of patients, respectively). Double atorvastatin dose was commonly recommended in all scenarios (43–52% of patients). More intensive LDL-C-lowering regimens were recommended infrequently e.g. double atorvastatin dose and add ezetimibe only <12% in all scenarios. Overall, cardiovascular risk factors and desire to achieve a more aggressive LDL-C goal were prominent factors in decision-making for treatment. Comparison of observed and estimated LDL-C levels showed that physicians tended to overestimate the effectiveness of their recommendations.

Conclusions

This study provides insight into physicians’ perspectives on clinical management of hypercholesterolemia and highlights a gap in knowledge translation from guidelines to clinical practice. The need for lower LDL-C and cardiovascular risk were key drivers in clinical decision-making, but physicians’ treatment choices were more conservative than guideline recommendations, potentially resulting in poorer LDL-C reduction. When compared with actual outcomes, projected LDL-C control was better if physicians used more comprehensive strategies rather than simply doubling the statin dose.

Trial registration

Clinicaltrials.gov: NCT01154036

Electronic supplementary material

The online version of this article (doi:10.1186/s12944-015-0037-y) contains supplementary material, which is available to authorized users.

Incidence of cancer and mortality in patients from the Simvastatin and Ezetimibe in Aortic Stenosis (SEAS) trial

The Simvastatin and Ezetimibe in Aortic Stenosis (SEAS) clinical trial, including 1,873 patients found an increased risk for cancer with lipid-lowering therapy with ezetimibe/simvastatin 10/40 mg/day, relative to placebo. In a registry-based follow-up study over 21 months from the conclusion of the SEAS trial, new incident cancer and total mortality were investigated in the SEAS study cohort from Denmark, Finland, Norway, Sweden, and the United Kingdom. Among 1,359 subjects eligible for follow-up (73% of the original total cohort), 1,194 had no history of cancer (primary follow-up cohort). New cancers and deaths were identified in the national cancer and mortality registries and classified by an Expert Review Committee. Data were analyzed using Cox proportional-hazards models of new cancers and mortality during follow-up according to treatment group assigned in the SEAS base study and with age, gender, smoking history, and previous cancers as covariates. The primary follow-up cohort had 12 patients with new cancers in the ezetimibe/simvastatin group and 22 in the placebo group (hazard ratio 0.55, 95% confidence interval 0.27 to 1.11), indicating no significant difference between the treatment groups. During follow-up, 43 patients assigned to ezetimibe/simvastatin and 33 assigned to placebo died (hazard ratio 1.29, 95% confidence interval 0.82 to 2.03). In conclusion, in this registry-based observational follow-up study of the original SEAS study patient population, treatment with ezetimibe/simvastatin was not associated with an increased risk for cancer or mortality in the 21-month period after the completion of the original SEAS study.

LDL-C levels in US patients at high cardiovascular risk receiving rosuvastatin monotherapy

BACKGROUND

Statin therapy is recommended as the first-line pharmacotherapeutic approach for lowering LDL-C levels in patients at high cardiovascular risk.

OBJECTIVE

To assess LDL-C levels among patients at high cardiovascular risk treated with rosuvastatin monotherapy.

METHODS

This retrospective cohort study used patient records from the GE Centricity Electronic Medical Records (GE Centricity EMR) database and administrative claims data from Humana Medicare to identify patients at high cardiovascular risk with a first prescription for rosuvastatin monotherapy (index date) from January 1, 2008 through December 31, 2010. Eligible adult patients had an International Classification of Diseases, Ninth Revision (ICD-9) diagnosis or a Current Procedural Terminology (CPT) procedure code indicating coronary heart disease or atherosclerotic vascular disease, ≥1 LDL-C measurement 3 to 12 months after the index date, and continuous data during 1-year baseline and 1-year follow-up. Mean LDL-C levels and distribution of patients around <70 mg/dL and <100 mg/dL thresholds overall and by daily rosuvastatin dose were assessed.

RESULTS

Among 6004 GE Centricity EMR patients (mean [SD] age, 66 [10] years; 56% men) and 11,320 Humana Medicare patients (mean [SD] age, 74 [8] years; 44% men) who met selection criteria, the most frequently prescribed rosuvastatin dose was 10 mg, and the mean (SD) follow-up LDL-C level was 89 (37) mg/dL for GE Centricity EMR patients and 92 (36) mg/dL for Humana Medicare patients. Overall, lower mean LDL-C levels were observed as rosuvastatin dose increased. However, less than one-third of GE Centricity EMR and Humana Medicare patients had an LDL-C level <70 mg/dL, and approximately two-thirds had an LDL-C level <100 mg/dL.

CONCLUSION

More effective lipid-lowering strategies, such as statin up-titration or combination therapy, may be needed to achieve therapeutic goals in a substantial proportion of high-risk patients.

Total and low-density lipoprotein cholesterol in high-risk patients treated with atorvastatin monotherapy in the United Kingdom: analysis of a primary-care database

OBJECTIVE

British clinical guidelines recommend statins as first-line lipid-modifying treatment (LMT) for patients at high risk of cardiovascular disease (CVD). We undertook an observational study to assess total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-C) levels in high-risk patients who were treated with atorvastatin monotherapy by UK general practitioners.

METHODS

This retrospective database study included patients with a prescription for atorvastatin monotherapy between November 30, 2008, and November 30, 2011, with the index date defined as the first atorvastatin prescription during this period. Eligible high-risk patients with evidence of coronary heart disease (CHD), atherosclerotic vascular disease (AVD), diabetes mellitus (DM), or familial hypercholesterolemia (FH) were required to have ≥1 TC and LDL-C measurement between 3 and 12 months after the index date, and continuous enrollment 1 year before and 1 year after the index date. Cholesterol levels were assessed using the National Institute for Health and Care Excellence (NICE) guidelines: TC <4.0 mmol/L or LDL-C <2.0 mmol/L.

RESULTS

Of 2999 high-risk patients (60.2% men; mean [SD] age = 67.9 [10.6] years) meeting selection criteria, 23.9% 28.2%, 36.2%, and 11.6% received prescriptions for atorvastatin 10, 20, 40, and 80 mg, respectively (percentages do not sum to 100 because of rounding). Across all doses, the mean (SD) follow-up TC was 4.08 (0.80) mmol/L and LDL-C 2.08 (0.65) mmol/L. A large proportion of patients (88.8%) had TC < 5.0 mmol/L. However, only 45.8% had TC < 4.0 mmol/L, and 46.5% had LDL-C < 2.0 mmol/L. Although a larger proportion of patients with CHD/AVD + DM reached guideline-recommended lipid levels, only 63.7% of such patients had TC < 4.0 or LDL-C < 2.0 mmol/L, which are the current targets for this subgroup as recommended by NICE.

CONCLUSIONS

Less than half of UK high-CVD-risk patients receiving atorvastatin monotherapy achieved guideline-recommended treatment targets for TC, and less than two-thirds of patients with CHD/AVD + DM had values below TC (4.0 mmol/L) or LDL-C (2.0 mmol/L) targets. More effective lipid-lowering strategies may be warranted to optimize cholesterol lowering and target attainment in high-risk patients. Limitations of this study include its retrospective, observational nature.

Systematic review of trials of the effect of continued use of oral non-selective NSAIDs on blood pressure and hypertension

OBJECTIVE

To investigate the effects of continued use of non-selective NSAIDs (nsNSAIDs) on blood pressure and hypertension.

RESEARCH DESIGN AND METHODS

This was a systematic review of randomized clinical trials of oral nsNSAIDs used for at least a 4-week duration. Searches were conducted of PubMed and the Cochrane Database of Systematic Reviews, using key terms for nsNSAIDs and blood pressure or hypertension, to identify articles published in the English language peer-reviewed literature through March 2007.

MAIN OUTCOME MEASURES

Change from baseline to end of study in systolic blood pressure (SBP) and diastolic blood pressure (DBP), and the incidence of hypertension. Pooled statistics were computed using fixed and random-effects analyses.

RESULTS

Thirty-two articles were included. The mean change (95% confidence interval [CI]) in blood pressure (in mmHg) from baseline to end of study for five trials of ibuprofen was 3.54 (2.70, 4.39) for SBP and 1.16 (0.68, 1.64) for DBP (p < 0.001 for both changes). Results of four trials of indomethacin were similar to those for ibuprofen: 2.90 (-0.28, 6.08) for SBP (p = 0.07) and 1.58 (0.29, 2.87) for DBP (p = 0.02). Mean changes from baseline for two trials of diclofenac were -0.46 (-1.48, 0.56) for SBP (p = 0.38) and -0.56 (-1.19, 0.07) for DBP (p = 0.08) and were similar to those for placebo. Changes from baseline in SBP were positive but not statistically significant for naproxen, sulindac, and nabumetone. Compared with placebo, the risk ratio (95% CI) for hypertension was 2.85 (1.44, 5.65; p = 0.003) in two ibuprofen trials.

CONCLUSIONS

Continued use of ibuprofen increases blood pressure and raises the incidence of hypertension. There appears to be heterogeneity in such effects with continued use of other nsNSAIDs but, due to limitations in the data, results for naproxen, sulindac, and nabumetone are inconclusive.

Update on the use of patient reported functional status and HRQOL measures in low back pain

Disease-specific functional status and generic health-related quality of life instruments used in low back pain studies are reviewed. A variety of instruments are available and many have been demonstrated to be valid and reliable. The Roland-Morris Disability Index and Oswestry Disability Index are the most commonly used low back pain-specific measures, while the Short Form-36 is the most commonly used generic quality of life tool. The Short Form-36 is valid and reliable in back pain but not necessarily intended to replace disease-specific measures. Studies have been inconsistent in their efforts to demonstrate a relationship between self-reported functioning and task performance.

Long-term morbidity, mortality, and economics of rheumatoid arthritis

OBJECTIVE

To estimate the morbidity, mortality, and lifetime costs of care for rheumatoid arthritis (RA).

METHODS

We developed a Markov model based on the Arthritis, Rheumatism, and Aging Medical Information System Post-Marketing Surveillance Program cohort, involving 4,258 consecutively enrolled RA patients who were followed up for 17,085 patient-years. Markov states of health were based on drug treatment and Health Assessment Questionnaire scores. Costs were based on resource utilization, and utilities were based on visual analog scale-based general health scores.

RESULTS

The cohort had a mean age of 57 years, 76.4% were women, and the mean duration of disease was 11.8 years. Compared with a life expectancy of 22.0 years for the general population, this cohort had a life expectancy of 18.6 years and 11.3 quality-adjusted life years. Lifetime direct medical care costs were estimated to be $93,296. Higher costs were associated with higher disability scores.

CONCLUSION

A Markov model can be used to estimate lifelong morbidity, mortality, and costs associated with RA, providing a context in which to consider the potential value of new therapies for the disease.

"Arthritis specific" global health analog scales assess "generic" health related quality-of-life in patients with rheumatoid arthritis

OBJECTIVE

Quality-of-life assessment is receiving increased attention as an outcome measure in rheumatoid arthritis (RA). The most widely used instruments use variations on a vertical visual analog scale (VAS). Since the Health Assessment Questionnaire (HAQ), Arthritis Impact Measurement Scales (AIMS), and other instruments have long included a "global" arthritis horizontal VAS (GLOB), we studied whether these 2 approaches assess the same concept.

METHODS

We studied 663 patients with RA from 4 ARAMIS (Arthritis, Rheumatism, and Aging Medical Information System) centers and had them complete, in different parts of the same HAQ, the GLOB and the Torrance "feeling thermometer" (FT).

RESULTS

The 2 scales were highly correlated (r = -0.676; p < 0.001). Reliability (estimated by 6 month test-retest) was 0.62 for the FT and 0.83 for the GLOB. The GLOB correlated more strongly than the FT with disability (r = 0.561 vs -0.507) and pain (0.630 vs -0.553). In stepwise regressions, pain and then disability were the dominant predictors of both GLOB and FT, followed weakly by joint count and then other variables. Patients with greater disability placed more emphasis on pain and patients with greater pain appeared to value more the contribution of disability. Change scores over 6 months between GLOB and FT correlated very well (-0.59).

CONCLUSION

"Health" and "health related quality-of-life" are nearly equivalent terms. Since large longitudinal rheumatology databases contain thousands of global health VAS values, data for longitudinal quality-of-life studies in arthritis are already available, and this dimension may readily be added to longterm outcome assessment. In patients with RA, "generic" and "disease specific" assessments yield very similar results.

Response to therapy in rheumatoid arthritis is influenced by immediately prior therapy

OBJECTIVE

Clinicians do not often employ washout periods before prescribing a change in therapy for rheumatoid arthritis (RA). As a result, the observed effectiveness or lack of effectiveness of a new drug actually represents the effectiveness of that drug had the patient been taking placebo minus the residual effectiveness of the old drug.

METHODS

We studied new starts of selected disease modifying antirheumatic drugs (DMARD) and prednisone in 2,898 patients with RA from 8 ARAMIS data bank centers, broken into subgroups on the basis of immediately prior therapy. Therefore, we examined the hypothesis that the chances of a treatment being observed effective depend upon the immediately preceding treatment. Using intent-to-treat analysis, we analyzed the effects upon Health Assessment Questionnaire (HAQ) disability and pain scores an average of 9 months after the new drug start.

RESULTS

Methotrexate reduced disability significantly except after intramuscular gold or hydroxychloroquine and it reduced pain significantly after all prior therapies. Hydroxychloroquine reduced disability significantly after nonsteroidal antiinflammatory drugs (NSAID) only, but disability increased after intramuscular gold; pain was decreased only after NSAID only. Prednisone had no consistent effect upon disability but was consistently associated with decreased pain. Greatest effectiveness was always seen with a new drug start after NSAID only treatment versus after DMARD treatment.

CONCLUSION

The effectiveness of a newly started RA treatment after 9 months may be substantially influenced by immediately prior treatment. This finding provides an additional reason for concern about direct extrapolation of clinical trial data into clinical practice.

Gastrointestinal tract complications of nonsteroidal anti-inflammatory drug treatment in rheumatoid arthritis. A prospective observational cohort study

BACKGROUND

Gastrointestinal tract (GI) complications associated with nonsteroidal anti-inflammatory drug (NSAID) use are the most common serious adverse drug reactions in the United States. Nonsteroidal anti-inflammatory drugs cause both minor GI side effects such as abdominal pain and vomiting and serious GI events such as ulcers and bleeding. This study evaluates the event rates for all NSAID-induced GI complications in patients with rheumatoid arthritis, describes the time course of these events, and evaluates the role of prophylactic therapy with antacids and H2 receptor antagonists.

METHODS

We studied 1921 patients with rheumatoid arthritis from 8 ARAMIS (Arthritis, Rheumatism, and Aging Medical Information System) centers. Patients were selected for the study if they were treated with NSAIDs and had at least 2.5 years of observation available. Information on GI complications attributed to NSAIDs was obtained from validated patient self-reports collected every 6 months and supplemented by review of hospital records for all hospitalizations.

RESULTS

Approximately 15% of the 1921 patients reported an NSAID-induced GI side effect during the 2.5 year observation period. Forty-two patients had a serious GI complication requiring hospitalization; 34 of these 42 patients did not have a preceding GI side effect. Patients who were taking antacids and H2 receptor antagonists did not have a significantly lower risk for serious GI complications than did those not taking such medications. Asymptomatic patients taking these medications had a significantly higher risk for GI complications compared with those who did not take these medications (standardized odds ratio, 2.14;95% confidence interval, 1.06-4.32).

CONCLUSIONS

A large majority of patients with serious GI complications do not have preceding mild side effects. Prophylactic treatment with antacids and H2 receptor antagonists is of questionable value and may increase the risk for subsequent serious GI complications.

Comparative toxicity of non-steroidal anti-inflammatory agents

Comparative toxicity of non-steroidal anti-inflammatory drugs was assessed using the Stanford Toxicity Index consisting of weighted symptoms, laboratory abnormalities and hospitalizations in 2976 consecutively enrolled rheumatoid arthritis patients from eight data bank centers with 27,936 patient-years of observation. Scores ranged from 1.77 (SE 0.20) for aspirin to 5.94 (SE 0.92) for meclofenamate, with many differences between drugs being 2- to 3-fold and highly statistically significant. Results are consistent with our prior data, persist when assessed by several different scoring algorithms, are consistent across data bank centers and are consistent with data of others. There are major and reproducible differences in the overall toxicity of different non-steroidal anti-inflammatory drugs.

A reevaluation of aspirin therapy in rheumatoid arthritis

BACKGROUND

Aspirin therapy has been largely superseded by prescription nonsteroidal anti-inflammatory drug (NSAID) therapy in rheumatoid arthritis, in part because of premarketing studies suggesting lesser toxic effects for NSAIDs than for aspirin. This study evaluates these toxic effects in a postmarketing population of patients with rheumatoid arthritis.

METHODS

We studied 1521 consecutive courses of aspirin and 4860 courses of NSAIDs in patients with rheumatoid arthritis from eight Arthritis, Rheumatism, and Aging Medical Information System Post-marketing Surveillance Centers. Toxicity index scores were generated from symptoms, laboratory abnormalities, and hospitalizations, weighted for variable severity and severity of side effect.

RESULTS

The toxicity index was only 1.37 (SE = 0.10) for aspirin and 1.87 to 2.90 for selected nonsalicylate NSAIDs. These differences were consistent across centers and remained after statistical adjustment for differing patient characteristics. There was a different toxicity with different aspirin preparations, with a score for plain aspirin of 1.36 (SE = 0.23), for buffered aspirin of 1.10 (0.20), and for enteric-coated aspirin preparations of 0.92 (0.14). Most important, there were strong dose effects, with a score of 0.73 (0.09) for 651 to 2600 mg daily, 1.08 (0.17) for 2601 to 3900 mg, and 1.91 (0.38) for more than 3900 mg. The average aspirin dose taken was only 2665 mg/d, approximately eight "tablets," compared with 3600 to 4800 mg/d used in the 16 pivotal premarketing studies reviewed. Average NSAID doses were, on the other hand, lower in premarketing trials (eg, naproxen 500 mg/d vs 773 mg/d in the Arthritis, Rheumatism, and Aging Medical System clinical practices).

CONCLUSIONS

Aspirin therapy, in doses commonly employed in practice, has an excellent safety profile in rheumatoid arthritis, and it is the least costly NSAID. The safety advantage is explained primarily by a dose effect and secondarily by possible differences between formulations. Newer management strategies for rheumatoid arthritis emphasize NSAID use as symptomatic therapy and use of disease-modifying anti-rheumatic drug therapy for anti-inflammatory objectives. Thus, the original recommendation for "anti-inflammatory" doses of aspirin now is less easily justified. Aspirin therapy merits reconsideration as adjunctive therapy for the management of rheumatoid arthritis.

The relative toxicity of alternative therapies for rheumatoid arthritis: implications for the therapeutic progression

The traditional pyramid for the therapeutic progression in rheumatoid arthritis (RA) is based on assumptions that RA is a mild disease, that nonsteroidal antiinflammatory drugs (NSAIDs) have low toxicity, and that disease-modifying antirheumatic drugs (DMARDs) are extremely toxic. This article reviews data from ARAMIS (Arthritis, Rheumatism, and Aging Medical Information System), casting strong doubt on these assumptions. NSAIDs result in 1.3% excess gastrointestinal hospitalizations per year. Mortality rates in RA are far above those expected from age- and sex-matched populations. Individual NSAIDs show widely different overall quantitative toxicity indices, as do individual DMARDs. However, the ranges of toxicity of the two classes of drugs show nearly complete overlap. It is suggested that the new therapeutic progression in RA should emphasize initial use of DMARDs, beginning with the least toxic.

The health assessment questionnaire 1992: status and review

Over 100 papers describing and utilizing the Stanford Health Assessment Questionnaire (HAQ) have been published since 1980. A brief overview of the HAQ is presented along with a guide to the accumulated literature. The topics covered include: studies using the disability, pain, economic, and drug side effect dimensions of the HAQ; reliability and validity studies; applications to various rheumatic diseases; language adaptations; modifications and derivative scales; studies correlating the HAQ with sociodemographic, health status, laboratory, and physical measures; and randomized controlled trials and observational studies using the HAQ. A few comments regarding future directions for research are also presented.

Rectal thiopental versus an intramuscular cocktail for sedating children before computerized tomography

The sedative effects of rectal thiopental sodium and an intramuscular cocktail of meperidine hydrochloride, chlorpromazine hydrochloride and promethazine hydrochloride were compared in 72 pediatric patients undergoing computerized tomography (CT). Pediatric patients scheduled were randomly assigned to receive either the i.m. cocktail (2.0 mg/kg of meperidine, 1.0 mg/kg of both chlorpromazine and promethazine) or 25--45 mg/kg of thiopental rectally before scanning. Side effects, and onset, duration and depth of sedation were recorded by 14 unblinded investigators. Clarity of CT scans was rated by two blinded radiologists. Additional doses of sedatives were administered as necessary. Sedation was not achieved in 3% of the thiopental group or in 14+ of the i.m. cocktail group. Additional sedatives were required by eight patients in the thiopental group and by five patients in the cocktail group. The mean time for onset of sedation was 8 minutes with thiopental and 18 minutes with the cocktail. The mean duration of sedation was 7 hours for the cocktail group and 2.75 hours for the 25-mg/kg rectal thiopental group. All scans were diagnostic (acceptable) in the rectal thiopental group, but 14% of those in the i.m. cocktail group were nondiagnostic. Rectal thiopental is an effective alternative to an i.m. cocktail for sedating children before CT scans and may reduce the number of nondiagnostic scans.

Rectal thiopental for sedation of children for computed tomography

Rectally administered thiopental was evaluated for sedation in pediatric computed tomography and compared with "cardiac cocktail" sedation. The drug produced sedation as effective as the "cocktail" with easier administration, more rapid onset, and shorter duration of sedation. Although no complications were observed, careful observation for respiratory depression is suggested.