Introduction

Radiation synovectomy (RS) is used in the treatment of chronically inflamed joints refractory to treatment using conventional agents such as nonsteroidal anti-inflammatory drugs (NSAIDs), antirheumatoid drugs, or intra-articular corticosteroids. The technique is useful in delaying or, in some instances, avoiding surgical synovectomy.

Whilst the use of radioisotopes in managing chronically inflamed joints has been known for over three decades, RS is not commonly performed due to cost, difficulty in administration, and concerns of safety and effectiveness [1, 2]. There is reportedly a high dose delivery to surrounding normal structures and a high incidence of chromosomal aberrations following RS using certain radioisotopes. However, there are no reports of malignancy resulting from RS [3]. Yttrium90 (Y90) is a pure beta emitter with a mean tissue penetrance of 3.6 mm and mean half-life of 2.7 days. The use of this isotope in RS minimises the risk of irradiating the surrounding normal structures [1, 4].

In our hospital we have used Y90 RS since 1990 for inflammatory arthritis and synovitis that persists in one or both knees or recurs despite antirheumatoid drugs and several local corticosteroid injections. In this study, we analyse the results of therapy and complications in a group of patients so treated.

Materials and methods

All patients treated with RS to their knee joints at the Singleton NHS Trust between January 1990 and June 1999 were included in the study (Table 1). A total of 28 patients were referred for RS during this period, and a total of 38 joints were treated (including ten patients who had both knees treated). There were 11 males and 17 females in the study group, with a mean age of 55.9 years (range 32–80). The most common diagnosis (18 patients) was rheumatoid arthritis. Other varieties included ankylosing spondylitis, psoriatic arthritis, and seronegative inflammatory arthritis. The mean duration of treatment prior to RS was 6.7 years (range 1–20). All five patients with seronegative arthritis had involvement of only one joint, while the majority (n=23) showed evidence of polyarthritis.

Table 1. Patient characteristics
Full size table

All patients had been treated unsuccessfully with NSAIDs and intra-articular corticosteroids. Antirheumatoid drugs used included methotrexate, gold, and sulphasalazine (Table 2). One patient had had a partial surgical synovectomy prior to referral for RS. All patients continued their systemic drug therapy during and after the RS. Seventeen patients also had involvement of other joints at the time of Y90 therapy.

Table 2. Previous systemic treatment
Full size table

The joints treated were assessed before and after RS by the rheumatologist and clinical oncologist separately. Clinical assessments included joint pain, swelling, and range of movement. Follow-ups were also performed separately by the rheumatologist and oncologist, and response assessments by the different specialists were blinded. There was no disagreement on the assessments.

Radiation synovectomy was performed as an inpatient procedure in all patients jointly by the clinical oncologist and the rheumatologist. The joint was infiltrated with lidocaine anaesthetic with aseptic precautions. Synovial fluid was aspirated to ensure correct positioning of the needle within the joint space and, when a large effusion was present, to reduce its volume. A 10-mCi dose of Y90 was instilled through a three-way tap into the joint space, after which the needle was flushed with normal saline. Topical corticosteroid was not administered along with RS. The needle was removed and the wound sealed with sterile gauze. The treated joint was immobilised in a fibreglass shell for 72 h. Bed rest was not enforced.

Patients were followed up at 3-month intervals for 1 year and at 6-month intervals thereafter. Review of patients prior to 3 months was performed only for toxicity and not for response assessment. Response to therapy was first assessed at 3 months and performed clinically using a simple scoring system. Responses were recorded as good, very good, excellent, or none, based on assessments of pain, joint tenderness, joint swelling, and joint mobility (Table 3). An excellent or very good response to treatment was considered a satisfactory outcome. The endpoint for the study was relapse of symptoms. Patients were followed up for a minimum of 3 years after treatment or till relapse of symptoms.

Table 3. Response to radiation synovectomy
Full size table

Results

Treatment outcomes are summarised in Table 4 and Table 5.

Table 4. Time to response of radiation synovectomy
Full size table
Table 5. Outcome of radiation synovectomy
Full size table

Response to treatment

Assessment of response to treatment began at 3 months, and 29 of 30 joints showing response to RS had done so by 6 months (Table 4). Nine joints failed to show any response within this period, of which one showed response by 9 months. The maximum observed responses were achieved within 18 months and recorded as excellent in ten patients, very good in nine, and good in 11. Eight patients failed to show any response to treatment. No differences in response were noted, depending on the duration of arthritis, therapy before RS, or the number of joints involved.

Long-term control

Four patients were lost to follow-up after a mean of 21 months. Two of these patients had control of symptoms at the time of last follow-up; one was recorded as 'good' and the other as 'very good'. The other two patients had no response to treatment when last assessed.

Follow-up status at 3 years was available for the remaining 34 joints. Of these, six joints showed no response to the treatment, and in two joints the response was recorded as 'good'. The remaining 26 joints (68%) showed a satisfactory response to treatment (15 very good and 11 excellent) at 3 years. Of these, 11 joints (29%) showed control of all symptoms beyond 3 years. Their responses were scored as excellent and very good in four and seven joints, respectively. Two joints continued their 'good' response beyond 3 years.

In all, 15 joints showed recurrence of symptoms after a period of initial control (mean 13.6 months, range 9–36). The best response before relapse was recorded as very good in eight patients and excellent in seven.

The mean duration of response in patients showing a satisfactory outcome after RS was 28.4 months (range 10–84). The mean duration of long-term control of symptoms in the 11 joints was 48.6 months (range 36–84). The mean duration of symptom control observed in the entire patient group was 24.4 months (range 0–84).

Management of failures

Fifteen joints failed after an initial period of response following RS, and eight joints showed no response at all, making a total of 23 failures. Eleven of these failures were treated by surgery and six were treated using other agents. Two patients were lost to follow-up. Retreatment with Y90 was performed in four patients who showed no response to initial therapy. All these patients were treated with 10 mCi of Y90. Two showed an excellent response to reinjection which was maintained after 28 months and 34 months of follow-up, respectively. The other two patients did not respond.

Toxicity

The RS was well tolerated in all patients, with no evidence of acute pain, tenderness, or swelling. Two patients had extravasation of the isotope and needle track ulcers, which were recorded as major toxicity. One of these patients showed a necrotic area of skin and subcutaneous tissue which required skin grafting. In three other patients, there was evidence of minor pigmentation at the injection site.

Discussion

Patients with chronic arthritis refractory to standard drug therapy and intra-articular corticosteroid injections are sometimes managed with surgical synovectomy. The procedure is costly and may be associated with surgical complications including joint stiffness resulting from prolonged joint immobilisation [5]. Radiation synovectomy is an easy and relatively cheap alternative for delaying or, in many instances, avoiding surgical synovectomy [3, 6].

Radioactive gold (Au198) was commonly used in RS, but currently the silicate colloid of yttrium90 (Y90) is preferred. Y90 is a pure beta emitter with a mean tissue penetrance of 3.6 mm and mean half-life of 2.7 days. Studies have shown very little migration of this isotope (especially the silicate colloid) from the joint space, and therefore minimal uptake occurs into regional lymph nodes [7, 8]. Thus, the substitution of Y90 for Au198 minimises the risk of irradiating normal tissues and reduces chromosomal aberrations and damage to normal tissues [1, 4, 9].

Following instillation, the radioactive isotope is concentrated in the synovial membrane from the injected colloid suspension, where it exerts its activity. It is common practice to drain synovial fluid prior to instilling Y90 for better contact of the radioactive isotope with the inflamed synovial membrane. It is estimated that, following an injection of 6 mCi to the joint space, a total radiation dose of between 5,500 rad and 11,000 rad is delivered to the synovium over 3 days [10]. Radiation synovectomy is performed as an inpatient or day-case procedure. Rigid splinting of the joint is advised following the procedure, but prolonged bed rest or inpatient stay is not mandatory [11]. These procedures help to prevent extravasation of the isotope from the joint space and reduce uptake by the lymphatics [12].

Most studies used a dose of 5 mCi for RS. Menkes [13] reports that 4 mCi may be sufficient to produce acceptable control of symptoms. It is advisable to limit the total dose of isotope to less than 15 mCi, especially in younger patients, for fear of long-term effects on normal tissues. Doses of 10 mCi were used in 35 of the treated joints in our series. Three patients receiving treatment in the initial period of the study were treated with 5-mCi doses, and their outcomes were not satisfactory. Our decision to use a higher, 10-mCi dose was based on this personal experience.

Intra-articular injection of Y90 is useful in reducing pain, tenderness, and swelling of the joints and improving mobility. In the series reported by Asavatanabodee et al., 82% of joints treated with Y90 showed improvements in clinical status at 6 months after injection [2]. Most other studies reported results ranging between 50% and 80% after Y90 synovectomy [14, 15, 16, 17, 18]. In a long-term retrospective study by Stucki et al. [6], an overall success rate of 50% was noted for 164 Y90 treatments. Their 3-year success rate was reported to be 30%.

Patients with established radiological abnormalities of the joints generally respond less well than those with less damage as seen on radiology [2, 17, 19, 20, 21]. Patients with short duration of disease generally respond better to Y90 RS than those with longer disease, in which there is likely to be greater irreversible joint damage. It is also reported that response to RS worsens in the presence of multiple joint involvement and active systemic disease.

Pain, swelling, febrile reactions, malaise, local tenderness, and infection are reported complications of RS [2, 19, 22]. Extravasation of the isotope, needle tract ulcers, and soft tissue necrosis are less frequent [8, 23]. In our series, no pain, fever, or joint swelling was reported despite routine avoidance of corticosteroids. However, in the initial period of the study, two patients developed extravasation of the isotope and needle tract ulcers. In one, the needle track ulcer was managed conservatively, while skin grafting was required in the other. By improving our technique using the three-way system and with adequate joint immobilisation, we were able to prevent further complications.

Responses to treatment were recorded by 6 months following therapy in most studies [2, 20]. Menkes et al. report that responses could be reliably assessed 6 months after therapy and that late responses are unlikely to occur [20]. Also in our study, 29 out of 30 joints which responded to therapy had done so by 6 months following RS. However, outcome of treatment continued to improve beyond 6 months to a maximum of 18 months. Response to RS was reported to be maintained for 24–49 months following therapy [16, 17, 24]. This is in agreement with the mean response duration of 28.4 months observed in the patients who responded to Y90 in our series.

Several studies have looked into the factors determining outcome with RS. Patients with established radiological abnormalities of the joints generally respond less well than those with less damage as seen on radiology [2, 15, 19, 21]. Patients with short duration of joint involvement also do better with RS [2]. It is likely that with a prolonged history of joint involvement, there may be more pronounced damage of the joint architecture, thereby accounting for worse response to therapy. Patients with multiple joint involvement and active systemic disease are also known to respond badly to local therapy [25]. It has also been attempted to study the outcome of RS by type of arthritis; psoriatic joints generally are reported to show a poor response [2, 16]. It should however be noted that most studies have too few patients to make any statistically significant comparison of outcome. In our study, too, a valid comparison of outcome based on number of affected joints, duration of arthritis, and nature of arthritis was not possible due to the limited number of patients within the different subgroups.

The role of reinjecting radioactive yttrium in relapsing or nonresponding patients is another matter of debate. Stucki [6] and Asavatanabodee [2] reported response rates of up to 53% following reinjection. In the series reported by Winfield et al., 20% and 63% of the patients undergoing retreatment with the isotope showed complete and partial response, respectively [26]. In our series, only four patients were reinjected with Y90, and there were two responses.

Nonresponders and patients who fail after a period of symptom control are managed by arthroscopic synovectomy, surgical synovectomy, or joint replacement. By the time of reporting, 13 patients in our series had been referred for surgical intervention and nine had undergone knee replacement. The RS helped to delay the procedure in several patients and gave long-term control of symptoms in 29% of the patients in this series.

Few randomised clinical trials examine the role of Y90 RS in the management of refractory arthritis [27]. Besides, randomised trials comparing RS with placebo or intra-articular steroid injection have used limited numbers of patients or short duration of follow-up [28, 29]. We find Y90 RS to be a safe and quick procedure with a favourable long-term outcome that can be useful in the management of patients with refractory synovitis. We feel its use is justified until its role is prospectively evaluated in a randomised clinical trial involving a large number of patients.