Agnes van der Heide

The Effectiveness of Early Treatment with “Second-Line” Antirheumatic Drugs: A Randomized, Controlled Trial

In recent years, therapeutic strategies for patients with rheumatoid arthritis have been discussed in detail [1-8]. Traditional therapy, usually referred to as the pyramid model, begins with nonsteroidal anti-inflammatory drugs (NSAIDs). If these are insufficiently effective, they are later replaced or supplemented with second-line antirheumatic drugs, which are distinguished from NSAIDs primarily by their assumed disease-modifying potential and delayed onset of action. These second-line drugs are also referred to as slow-acting antirheumatic drugs (SAARDs) or disease-modifying antirheumatic drugs (DMARDs). Traditionally, the SAARDs of first choice have been hydroxychloroquine or intramuscular gold followed by D-penicillamine, methotrexate, or azathioprine. The traditional therapeutic pyramid is based on the principle of primum non nocere: Because of their potential toxicity, SAARDs should only be given when milder therapies have failed. Recently, the beneficial effects of the pyramid strategy have been questioned because the long-term outcome of rheumatoid arthritis continues to be disappointing. Patients with rheumatoid arthritis have increased mortality [9, 10], and their quality of life is seriously affected by functional impairment and loss of employment [11, 12]. The pyramid model might be undesirable because administering SAARDs only after milder regimens prove to be insufficiently effective delays the suppression of inflammation. It is preferable that the disease process be controlled as soon as possible, because radiologic abnormalities appear in the joints early in the course of the disease [13, 14] and are related to the extent of inflammation [15, 16]. Toxicity indices for NSAIDs and most SAARDs have recently been shown to be similar; thus, exposure of patients to possible adverse reactions is probably not in itself a reason to withhold treatment with SAARDs [17]. However, the ability of SAARDs to modify the course of disease and prevent radiologic damage is often questioned [18, 19]. The poor long-term outcome of rheumatoid arthritis might be caused not by a delay in effective treatment but by the low disease-modifying power of both first- and second-line drugs. The short courses of SAARDs (short because of adverse reactions or impatience with a delay in effect) might also play a role [20, 21]. Combinations of SAARDs have not been shown to be superior to single drugs [21, 22]. Current clinical practice is shifting toward the earlier introduction of second-line treatment for rheumatoid arthritis, but thus far no clinical trials have shown this strategy to be beneficial. We investigated the consequences of ignoring the treatment pyramid for patients with recent-onset rheumatoid arthritis. In this report, we describe the first-year results of a randomized clinical trial in which the delayed and the immediate introduction of SAARDs were compared. Effectiveness was measured by disease activity and progression of radiologic abnormalities. The various SAARDs are considered as a single group in this report; comparisons among the second-line therapies will be reported in detail later. Methods Patients As of January 1990, all patients with recent-onset rheumatoid arthritis (diagnosed according to the 1987 American College of Rheumatology criteria [[23]] from six rheumatologic centers in the region of Utrecht, the Netherlands, were asked to participate in a randomized, prospective clinical trial. Disease duration had to have been less than 1 year, and most patients were enrolled shortly after diagnosis. The following exclusion criteria were applied: 1) age younger than 17 years, 2) comorbid conditions that might interfere with one of the therapeutic strategies, 3) previous or current treatment with any SAARDs, glucocorticosteroids, or cytotoxic or immunosuppressive therapy, 4) the possibility of pregnancy or breast feeding, and 5) psychiatric or mental disturbances that would make adherence to the study protocol unlikely. All patients gave informed consent before study entry and were told that the study was intended to compare the unsure outcome of treatment with NSAIDs alone with that of treatment with NSAIDs and a SAARD for rheumatoid arthritis at an early stage. The study design was approved by the ethical committees of all participating hospitals. The baseline characteristics of patients who were eligible for the study but who objected to randomization were compared with those of randomly assigned patients to determine whether any selection bias had occurred. Treatment Patients entering the study were randomly assigned to one of two therapeutic groups. All randomization procedures were done by drawing sealed envelopes from blocks of 100 with equal numbers of patients for each of the four treatments per hospital; the person doing the procedures was blinded to treatment assignments. This method was used because eligible patients were not distributed equally among the centers. The assigned therapeutic strategy was continued for at least 1 year. In the non-SAARD group, patients began receiving NSAID therapy, the dose and type of which could be modified at any time. For this group, initiation of SAARD treatment during the study was regarded as discontinuation of the therapeutic strategy. In the SAARD group, patients were randomly assigned (by persons blinded to treatment assignments) to treatment with one of the following SAARDs: hydroxychloroquine (400 mg/d), intramuscular gold (aurothioglucose, 50 mg/wk), or oral methotrexate (7.5 to 15 mg/wk). Therapy with the initial SAARD was continued for 12 months unless adverse reactions necessitated discontinuation; if this occurred, another SAARD was given. Hydroxychloroquine was followed by auranofin (6 to 9 mg/d); intramuscular gold was followed by D-penicillamine (500 to 750 mg/d); and methotrexate was followed by sulfasalazine (2000 to 3000 mg/d). Discontinuation of therapy with the second SAARD was regarded as discontinuation of the therapeutic strategy. Use of NSAIDs was allowed in the SAARD group; as in the non-SAARD group, the dose and type could be changed at any time. Patients were assigned to the non-SAARD group and the SAARD group in a ratio of 1:3, and power calculations for the primary end point disability ( equals 0.05; equals 0.20; difference to be detected, 25%) indicated that group sizes of 40 and 120, respectively, would be sufficient. The ratio of 1:3 was chosen to enable later comparisons of results within the SAARD group, which were thought to be similar. The use of analgesics was allowed in both groups; the use of oral glucocorticosteroids was avoided if possible; and intraarticular injections were not allowed within 2 months of a scheduled visit. Criteria for discontinuation or dose adjustment of a SAARD because of adverse reactions were described in detail in the study protocol. Discontinuation of any treatment because of sustained disease activity was done only if the patient and the attending physician (who had to discuss the patient with a colleague) judged it to be unavoidable. Primary End Points Assessments were done at the start of the trial and were repeated every 3 months. Clinical variables were assessed by the same physician or research nurse for each patient on each occasion. Primary end points were functional disability, pain, joint score, erythrocyte sedimentation rate, and radiologic abnormalities [24]. Radiologic abnormalities and erythrocyte sedimentation rate were determined by persons blinded to treatment assignments; functional disability and pain were self-assessed scores; and determination of the joint score was not blinded to treatment. Functional disability was assessed using a validated Dutch version of the Health Assessment Questionnaire Disability Score [25-27]. The questionnaire scores range from 0 to 3; 0 is the best score (no problems), and 3 is the worst. Pain was measured on two horizontal visual analog scales of 100 mm; the mean of the scores for pain during the night and pain during the morning was calculated. The joint score according to Thompson and colleagues [28, 29] assessed the simultaneous presence of joint tenderness and swelling in a selection of joints weighed for joint size (range, 0 to 534) [28, 29]. Erythrocyte sedimentation rate (mm/h) was measured using Westergren method. Radiography of hands and feet was done at the start of the study and at 12 months. A modified version of the method of Sharp and coworkers [30, 31] was used to score radiologic abnormalities. According to this method, erosions and joint space narrowing in hand and foot joints are scored and added together to obtain a total radiologic damage score (range, 0 to 448). All radiographs were evaluated by two investigators who were blinded to treatment assignments. The scores of the first investigator were used in the analyses, and the scores of the second investigator were used to validate the scores of the first. Correlation between the two sets of scores was satisfactory (Spearman correlation coefficients for scores at baseline and at 12 months were 0.84 for both); the second investigator was found to have given lower scores (means at baseline were 4.2 and 3.4, respectively; means at 12 months were 11.2 and 9.7, respectively). Therefore, in individual cases, differences in total scores of 25% or more were discussed until an agreement was reached on the appropriate score. That score was then used in the analyses. Additional Assessments Secondary end points were grip strength (mean of three measurements of each hand with a vigorimeter [Martin, Tuttlingen, Germany] in kPa), duration of morning stiffness (maximum, 720 min), and general well-being (horizontal visual analog scale of 100 mm). Additional laboratory variables were serum level of C-reactive protein (mg/L), hemoglobin concentration (mmol/L), and platelet count (times 109/L). Rheumatoid factor status was determined to be positive or negative, as indicated by either the Latex

Physician reports of terminal sedation without hydration or nutrition for patients nearing death in the Netherlands.

Context Terminal sedation, the administration of sedating medications with cessation of nutrition and hydration, is an option for care of patients who are nearing death. However, little is known about physicians experience with terminal sedation. Contribution Of more than 400 physicians in the Netherlands who completed a survey about end-of-life care, just over half had ever used terminal sedation. Common reasons for using terminal sedation were relief of pain, agitation, or dyspnea. Hastening death was the primary intention in only 17% of reported cases. Caution Because attitudes and practices regarding terminal sedation vary geographically and culturally, it is unclear whether these results are generalizable outside of the Netherlands. The Editors Patients nearing death frequently have symptoms such as dyspnea, agitation, pain, and anxiety (1, 2). One of the most important goals of the medical care provided to these patients is the alleviation of these symptoms (3). If treatment with analgesic or anxiolytic agents is not effective, sedatives are sometimes used as an alternative to render patients unconscious and then oblivious to their symptoms (4, 5). Subsequently, if artificial nutrition and hydration are not given, death will follow soon. The ethical debate about this practice focuses on the extent to which it should be considered an end-of-life decision that possibly or certainly hastens death. Previous studies have explored the differences and similarities with other end-of-life decisions, such as euthanasia and physician-assisted suicide (6-19). However, little information exists on the medical practice of deep sedation with the forgoing of artificial nutrition or hydration in patients nearing death. Estimates about the frequency of deep sedation at the end of life vary from 15% to more than 60%, depending on the settings studied and the definitions used (4, 5, 20-26). The terminology used reflects these differences in definition of the practice of deep sedation at the end of life. Although terminal sedation is the most commonly used term, other frequently used terms, which demonstrate the different perspectives from which this practice is viewed, are sedation for intractable distress in the imminently dying, palliative sedation therapy, slow euthanasia, opioid coma, or anesthetic coma (6, 27-30). The present study describes the practice of terminal sedation in the Netherlands. This study was part of the evaluation of the notification procedure for physician-assisted death in the Netherlands, which was commissioned by the ministers of Health and Justice (31). Methods Respondent Characteristics We interviewed a nationwide sample of 410 physicians: 208 clinical specialists, 125 general practitioners, and 77 nursing home physicians. In the Netherlands, clinical specialists provide hospital care, general practitioners provide nonspecialized care outside the hospital, and nursing home physicians work in long-term care institutions mainly for elderly people. The proportions of deaths in these health care settings are approximately 35%, 42%, and 23%, respectively. The specialties involved in our study covered about 95% of all deaths in the Netherlands in 2001. The respondents were selected according to the following criteria: They were required to be in active practice at the time of the interview and to have actively practiced medicine within the registered specialty for the past 2 years in the same setting. All addresses were taken from the professional registries of the relevant specialties. To arrive at the desired number of 410 physicians, we sampled 482 physicians. Seventy-two physicians (15%) declined to take part in the study: 17% of clinical specialists, 18% of general practitioners, and 3% of nursing home physicians. Nonresponders did not differ in age from responders. Face-to-face interviews were conducted by experienced part-time working or recently retired physicians who were trained to administer the structured questionnaires. All interviews took place between March 2002 and October 2002. We applied strict rules to ensure the anonymity of all physicians and patients studied. Interview Process The interview schedule addressed experiences with end-of-life decision making (Appendix Figure [Questionnaire on Terminal Sedation]). Terminal sedation was defined as the administration of drugs to keep the patient in deep sedation or coma until death, without giving artificial nutrition or hydration. The respondents were first asked whether they had ever used terminal sedation and, subsequently, how often they had performed this practice in 2000 and 2001. Additional questions about the practice of terminal sedation concerned the physicians most recent patient to have received terminal sedation (n= 211). The physicians were asked about the patients characteristics; whether or not sedation or the forgoing of artificial nutrition or hydration had been discussed with the patient, family, or other health care professionals; the drugs used; the intention of the physician; the estimated life-shortening effect; and whether euthanasia was discussed during the decision process. Appendix Figure. Terminal sedation: frequencies and case characteristics. Statistical Analysis We calculated all estimates about the occurrence of terminal sedation in the Netherlands by weighting the estimates of individual physicians. Weighting factors were based on differences in sampling fractions and response rates for the different specialties. These sampling fractions were 125 of 7027 for general practitioners, 77 of 810 for nursing home physicians, 34 of 394 for cardiologists, 34 of 545 for neurologists, 69 of 1321 for specialists in internal medicine, 35 of 325 for pulmonologists, and 36 of 769 for surgeons. The probabilities used to determine sampling weights were 1 in 56 for general practitioners, 1 in 11 for nursing home physicians, 1 in 12 for cardiologists, 1 in 16 for neurologists, 1 in 19 for specialists in internal medicine, 1 in 9 for pulmonologists, and 1 in 21 for surgeons. Data on the 211 most recent patients seen by physicians were not weighted. All analyses were done by using SPSS software, version 10.0 (SPSS, Inc., Chicago, Illinois). Role of the Funding Sources The sponsors approved the study design but were not involved in the collection, analysis, or interpretation of the data or in the decision to submit the manuscript for publication. Results Most of the 410 physicians interviewed (76%) were men; 51% were clinical specialists, 30% were general practitioners, and 19% were nursing home physicians (Tables 1 and 2). Of all physicians, a weighted percentage of 52% (95% CI, 48% to 57%) had ever practiced terminal sedation. This percentage was 55% (CI, 49% to 62%) for clinical specialists, 48% (CI, 39% to 57%) for general practitioners, and 75% (CI, 64% to 83%) for nursing home physicians. We asked all interviewed physicians to estimate the total number of times they performed terminal sedation in 2000 and 2001. These numbers were extrapolated to the total number of 140377 deaths in 2001 by multiplying them with the weighting factor for each specialty and assuming that the numbers were similar for the 5% of deaths covered by hospital doctors from specialties other than the ones included in our study. This extrapolation suggests that physicians used terminal sedation in 10.0% (CI, 9.1% to 10.8%) of all deaths in that year. Of the 10.0% of deaths preceded by terminal sedation, 5.5% (CI, 5.0% to 6.1%) were attended by clinical specialists, 2.5% (CI, 1.9% to 3.2%) by general practitioners, and 2.0% (CI, 1.7% to 2.2%) by nursing home physicians. Table 1. Characteristics of Interviewed Physicians Table 2. Proportion of Deaths per Specialty Of all physicians who had ever used terminal sedation, 211 provided information about their most recent cases of terminal sedation (103 clinical specialists, 53 general practitioners, and 55 nursing home physicians). Of these most recent cases, 78% (CI, 72% to 83%) involved patients 65 years of age or older and 54% (CI, 47% to 60%) involved patients who had cancer (Table 3). Clinical specialists and nursing home physicians also frequently reported practicing terminal sedation in patients with cardiovascular diseases. The most frequently mentioned reasons for using terminal sedation were the alleviation of pain (51% [CI, 44% to 58%]), agitation (38% [CI, 32% to 45%]), dyspnea (38% [CI, 32% to 45%]), and anxiety (11% [CI, 8% to 16%]). Table 3. Characteristics of the Sample Consisting of Each Physicians Most Recent Case of Terminal Sedation In 59% (CI, 52% to 66%) of the most recent cases seen by physicians, the physician had discussed the sedation with the patient (Table 4); in 33% (CI, 27% to 39%) of the cases, the patient had requested deep sedation. The main reasons for not discussing deep sedation with the patient were the fact that the patient was incompetent or subcomatose (25% [CI, 20% to 31%]). The decision to forgo artificial nutrition or hydration was discussed less frequently with the patient; the respondents reported discussing this topic in 34% (CI, 28% to 41%) of their most recent cases and receiving a request from the patient to forgo artificial nutrition or hydration in 9% (CI, 6% to 13%). Next to patient incompetence (37% [CI, 31% to 44%]), another frequently mentioned reason for not discussing the decision to forgo artificial nutrition or hydration was that many physicians perceived this not as optional but rather as a given; they considered terminal sedation to preclude the concomitant use of artificial nutrition and hydration (23% [CI, 18% to 29%]) (data not shown). Table 4. Discussion about Deep Sedation and Forgoing Artificial Nutrition or Hydration in Each Physicians Most Recent Case of Terminal Sedation, by Physician Specialty The decision to use sedation was discussed with relatives of the patient in 93% (CI, 89% to 96%) of the most recent cases seen by physician