British Journal of Rheumatology 1998;37:1029–1035

LETTERS TO THE EDITOR the possibility that the anti-PRL autoantibody could be the cause of HPRL in SLE patients. Moreover, the interpretation is different from the finding by Hattori et al. [4] in idiopathic HPRL patients with anti-PRL autoantibody without autoimmune rheumatic disease. Hattori et al. found that the autoantibodies to PRL interfere with the real values of serum PRL using the radioimmunoassay (RIA) technique, showing that the measured levels of this hormone are lower than the real values, i.e. false low levels. In contrast, the immunoradiometric assay (IRMA) technique expressed more confident results. We are running a transverse design study with a large number of SLE patients. At this time, we have studied 260 SLE patients; of those, 42 have shown HPRL (16.2%, 95% CI = 11.7–20.7%). We consider the values >20 ng/ml (520 mUI/ml ) as HPRL. Moreover, we have found 14 females with autoantibodies against PRL; this autoantibody is an IgG detected by affinity chromatography (Sepharose coated with Protein A and G), and none of the SLE patients with anti-PRL autoantibody have displayed normal values of PRL. Our results, as well as those shown by Hattori et al., show that the presence of antibodies to PRL could interfere with the real values of PRL when the technique used is RIA (double antibody). In contrast, the IRMA technique appears to be more specific and confident. In our study, the range of direct PRL using RIA was 17.8–69.4 ng/ml; using IRMA it was 31.1–229.7 ng/ml. The proportions of direct PRL to total serum PRL using the RIA in the presence of serum anti-PRL autoantibody were 64.3 ± 38.1% and, without anti-PRL autoantibody, 106.4 ± 12.5% (P = 0.0003). In contrast, when using IRMA, the proportions were the same with or without autoantibodies to PRL (116.2 ± 23.3 vs 120.6 ± 17.7%, P = 0.55). We also found a positive correlation between the titres of anti-PRL autoantibody and the serum levels of PRL (r = 0.98, P = 0.0001). Our explanation s of these phenomena is similar to that discussed in previous papers by Hattori et al. using PRL [4] and Schneider and Pervos [5] using the thyroglobulin assay. There appears to be enough evidence that in some SLE patients with HPRL, the anti-PRL autoantibody could be the cause of HPRL. However, it is necessary to carry out more experimental studies in order to know more about the role of the anti-PRL autoantibody in autoimmune rheumatic disease.

Re: Prolactin and Systemic Lupus Erythematosus SWe read with interest the article by Mok et al. [1]. In this paper, as well as in the previous report, the authors did not find any correlation between prolactin (PRL) and lupus activity in systemic lupus erythematosus (SLE ) patients, in spite of the fact that experimental studies using animal models have shown sufficient evidence that PRL plays a role in autoimmune rheumatic disease [2]. In contrast, the clinical trials carried out in humans, including the article by Mok et al., have shown contradictory data concerning the participation of PRL in SLE and its relationship with disease activity. The debate has not yet yielded a satisfactory conclusion. Recently, we carried out an analysis using the results from the English language literature concerning the correlation between PRL and lupus activity in SLE patients [3]. Briefly, we found five papers on this topic; the same articles alluded to in the references in Mok’s article. In only three of those papers was enough information found to perform an analysis of the power of the study. Of these papers, only one found a correlation between hyperprolactinaemia (HPRL) and disease activity; the other two articles did not find statistically significant differences between HPRL and disease activity. However, the power of those studies was 10 mmHg between the arms; (5) audible bruit on auscultation over one or both subclavian arteries or abdominal aorta; (6) ateriogram abnormality (narrowing or occlusion of the aorta or one of its branches). There were 11 females (73%) and four males, the mean age at diagnosis was 29.5 yr (range 12–64 yr) and 12 patients (80%) were below the age of 40 yr. Twelve patients (80%) had claudication of an extremity on presentation, also 12 (80%) patients had a decreased brachial artery pulse. Ten patients (67%) had a >10 mmHg difference between the arms, and nine patients (60%) had a bruit over the subclavian arteries or the aorta. Total aortography was performed in all patients and the results were classified according to the new classification of Hata et al. [2] as follows: five patients (33%) had involvement of the branches of the aortic arch only (type I ), one patient (7%) had involvement of the ascending aorta, aortic arch and its branches only (type IIa), three patients (20%) had involvement of the ascending aorta, aortic arch with its branches and thoracic descending aorta (type IIb), one patient (7%) had involvement of the thoracic descending aorta, abdominal aorta and/or renal arteries (type III ), two patients (13%) had involvement of the abdominal aorta and/or renal arteries only (type IV ), and finally three patients (20%) had the combined features of both type IIb and IV (type V ). Table I shows a comparison of the angiographic findings between our patients and a group of Japanese and Indian patients. Types I and II were significantly more common in the Japanese compared to the Indian

TABLE I Angiographic findings in Lebanese vs Japanese and Indian patients Lebanon

I IIa IIb III IV V Total

India

Japan

n

%

n

%

n

%

5 1 3 1 2 3

33 7 20 7 13 20

7 1 6 3 29 56

7 1 6 3 28 55

19 9 8 0 1 43

24 11 10 0 1 54

15

102

80

population [2, 3]. In addition, Indian patients had a rather higher frequency of type IV disease [2, 3]. Similarly to the Japanese patients, Lebanese patients showed (using a x2 analysis) a significantly higher frequency of type I and II involvement compared to the Indian patients (P < 0.001). Our data suggest that the vascular lesions of Takayasu’s arteritis in the Lebanese patients occur primarily in the ascending aorta, aortic arch and/or its branches, and extend into the abdominal aorta. I. U, A.-F. M, K. K,* S. A

Department of Internal Medicine, Faculty of Medicine and *Department of Health Services Administration, Faculty of Health Sciences, American University of Beirut, Beirut, Lebanon Accepted 9 April 1998 Correspondence to: I. Uthman, American University of Beirut, Medical Center, PO Box 113-6044, Beirut, Lebanon. 1. Arend WP, Michel BA, Bloch DA et al. The American College of Rheumatology 1990 criteria for the classification of Takayasu arteritis. Arthritis Rheum 1990;33:1129–34. 2. Hata A, Noda M, Moriwaki R, Numano F. Angiographic findings of Takayasu arteritis: new classification. Int J Cardiol 1996;54(suppl.):155–63. 3. Numano F. Differences in clinical presentation and outcome in different countries for Takayasu’s arteritis. Curr Opin Rheumatol 1997;9:12–5.

Spondylodiscitis Caused by Streptococcus equisimilis SInfectious non-tuberculous spondylodiscitis may be caused by a wide variety of micro-organisms. Grampositive microbes are responsible for the majority of cases, with staphylococci accounting for about onehalf to three-quarters of these cases [1]. Group C streptococci are rare pathogens and include the large colony of Voges–Proskauer-negative bacteria (S. equi, S. equisimilis, S. zooepidemicus and S. dysgalactiae), as well as the minute colony of Voges–Proskauerpositive S. anginosus (‘S. milleri’). Streptococcus equisimilis is the strain most frequently isolated in human infections and has been identified as an aetiological agent of meningitis, brain abscess, endocarditis, pneumonia, cellulitis, neutropenic sepsis, puerperal sepsis, septic arthritis and osteomyelitis [2].

LETTERS TO THE EDITOR

We present the case of an 83-yr-old man with infectious spondylodiscitis due to S. equisimilis with apparent primary focus in the oral cavity. To our knowledge, no previous cases of S. equisimilis spondylodiscitis have been described. An 83-yr-old Caucasian man with a medical history of non-insulin-dependent diabetes mellitus and moderate aortic insufficiency was hospitalized with complaints of fever and back pain of 6 weeks duration. On physical examination at admission, the patient had a temperature of 38°C, a blood pressure of 150/60 mmHg, diffuse gingivitis and localized midline tenderness over the lower thoracic spine. No rash or lymphadenopathy were found. Neurological examination revealed only mildly hypoactive patellar and Achilles reflex. Investigations revealed: WBC (6.9 × 109/l ), normocytic normochromic anaemia (Hb: 102 g/l ), platelets (127 × 109/l ), ESR (58 mm/h), C-reactive protein (52 mg/l ) and glucose (115 mg/dl ). Alkaline and acid phosphatase, alanine aminotransferase, aspartate aminotransferase, creatine kinase, lactic dehydrogenase, bilirubin, calcium, phosphorus and urea nitrogen were normal. A tuberculin test (2 UI-RT 23) was negative. X-ray films of the thoracic spine disclosed narrowing T8–T9 disc space. Magnetic resonance imaging (MRI ) was performed. The disc space and the vertebral bodies showed a decreased signal in T1-weighted images and an abnormal increased signal intensity in the T2-weighted images (Fig. 1). These findings were consistent with infectious spondylodiscitis. Three of six blood cultures performed at admission yielded a group C Streptococcus later identified as S. equisimilis. Culture of the urine, as well as agglutination tests for Brucella and Salmonella microorganisms were negative. A CT-guided needle aspiration of the T8–T9 disc space was performed. Microscopic examination of stained specimens disclosed a moderate number of Gram-positive cocci, but cultures were negative. The hospitalization period was uneventful and after 16 days of satisfactory ceftriaxone therapy (2 g once daily), the patient was discharged and continued this therapy on an ambulatory basis for three more weeks. On 4 month follow-up, ESR and C-reactive protein returned to normal, and the patient did not suffer from any thoracic pain. There is agreement that vertebral osteomyelitis is a process involving bacterial seeding; there is less agreement as to whether the route is arterial or venous. Haematogenous vertebral osteomyelitis is usually associated with a single organism as the aetiological agent. In the absence of bacteraemia, the diagnostic procedure of choice is a bone biopsy, performed with a cutting needle, and guided by fluoroscopy or CT scanning. MRI is extremely sensitive in detecting and delineating infective lesions, regardless of their spinal location. Osteomyelitis is associated with oedema of the bone marrow, and is therefore seen as bright-intensity imaging on T2-weighted scans and as low-intensity imaging on T1-weighted scans. In our patient, MRI

1031

was consistent with infectious spondylodiscitis and three blood cultures yielded S. equisimilis, but subsequent cultures of specimens from the disc space were sterile. We suggest that the infection originated by a haematogenous route due to a primary bacteraemic event where the probable source of the organism was the oral cavity. The Lancefield group C streptococci are serologically identified by the group-specific carbohydrate rhamnose-N-acetylgalactosamine, located on the cell wall. The large-colony group C streptococci also demonstrate differences in their pathogenicity in humans and animals. They are common inhabitants of many animal species. In humans, they may be isolated from healthy skin, as well as the nasopharynx and oropharynx, intestinal tract and vagina [2]. In most cases, these organisms are sensitive to penicillin, although individual cases of penicillin and erythromycin resistance have been reported [3]. There have been several reports of septic arthritis due to group C streptococci [2–5]. This microorganism appears to affect joints with pre-existing rheumatological conditions, but no underlying

F. 1.—Sagittal T2-weighted magnetic resonance image demonstrates narrowing and high-intensity signal in the disc space at T8–T9 with spread to the adjacent vertebral end plates.

1032

BRITISH JOURNAL OF RHEUMATOLOGY VOL. 37 NO. 9

malignancy has been described. To our knowledge, no previous cases of S. equisimilis spondylodiscitis have been published. N. G-R, J. L. F-S, J. I-R, J. S-C*

Services of Rheumatology and *Microbiology, POVISA Medical Centre, Vigo, Spain Accepted 12 May 1998 Correspondence to: N. Go´mez-Rodrı´guez, Service of Rheumatology of POVISA Medical Centre, C/Salamanca no 5, 36.211 Vigo, Spain. 1. Norden C, Gillespie WJ, Nade S, eds. Infections in bones and joints. Boston. Blackwell Scientific Publications 1994;231–48. 2. Ortel TL, Kallianos J, Gallis HA. Group C streptococcal arthritis: case report and review. Rev Infect Dis 1990;12:829–37. 3. Barnham M, Kerby J, Chandler RS, Millar MR. Group C streptococci in human infection. A study of 308 isolates with clinical correlations. Epidemiol Infect 1989;102:379–90. 4. Sobrino J, Bosch X, Wennberg P, Villalta J, Grau JM. Septic arthritis secondary to group C streptococcus typed as Streptococcus equisimilis. J Rheumatol 1991;18:485–6. 5. Steinfeld S, Galle C, Struelens M, De Gheldre Y, Farber CM, Appelboom T et al. Pyogenic arthritis caused by Streptococcus equisimilis (Group C streptococcus) in a patient with AIDS. Clin Rheumatol 1997;16:314–6.

Joint Physiology and Synovial Cell Proliferation SWe read with interest the articles by Edwards and ˇ eponis et al. [2]. The latter have Morris [1] and C contributed importantly to the now extensive evidence that rheumatoid arthritis is associated with increased synovial cell proliferation. Proliferation indices, using cell cycle antigens or nucleotide incorporation, indicate rates of synovial cell proliferation comparable with some cancers [2–6 ]. We agree that more recently developed antibodies to Ki67 antigen label proportions of synovial cells which are similar to other markers of proliferation, such as proliferating cell nuclear antigen, and that low labelling indices observed in some earlier studies may represent technical difficulties [6, 7]. Synovial cell proliferation may continue for many years in rheumatoid arthritis, with only limited expansion of the synovium. This apparent paradox can be explained by concurrent synovial cell death [6, 8, 9]. A picture is emerging of rapid turnover both of synovial fibroblasts and of endothelial cells, far in excess of the rate of change in total cell number. This increased cell turnover is associated with the synovium acquiring an immature phenotype. Vascular structures in rheumatoid synovitis express antigens such as integrin a b , which would be v 3 downregulated in mature tissues [6 ]. We suggest that vascular immaturity is a likely explanation of the deficient innervation of synovium in rheumatoid arthritis which is noted by Edwards and Morris. Innervation of the neovasculature begins only several weeks after the initiation of angiogenesis [10]. Increased release of neuropeptides such as substance P remains a plausible hypothesis, but cannot explain the observed depletion in inflamed synovium of nonsecreted neuronal proteins such as protein gene product

9.5 [11, 12]. Substance P-like immunoreactivity has frequently been detected in synovial fluids, but more stringent attempts to identify this with intact substance P using high-performance liquid chromatography have had little success [13]. This is not surprising considering the abundance of membrane peptidases within the inflamed synovium which would be expected rapidly to degrade peptides such as substance P [14, 15]. Immunoreactivity in synovial fluid is likely to be a very imprecise measure of the release of metabolically unstable peptides. The synovium in rheumatoid arthritis should be viewed as a highly dynamic tissue, in which stroma, blood vessels and nerves continue to grow over long periods, and where there is concurrent cell death and neuronal damage. The histological snapshot reflects a balance in which nerves and vessels have immature phenotypes and abnormal distributions, being deficient in the superficial synovial layers, but present in the deeper synovium [11, 16, 17]. The physiology of the normal joint differs importantly from the pathophysiology of the persistently inflamed joint. D. A. W, P. I. M*

Rheumatology Unit, Nottingham University, City Hospital, Hucknall Road, Nottingham NG5 1PB and *School of Postgraduate Medicine, University of Bath, Claverton Down, Bath BA2 7AY Accepted 9 April 1998 Correspondence to: D. A. Walsh, Rheumatology Unit, Nottingham University, City Hospital, Hucknall Road, Nottingham NG5 1PB. 1. Edwards JCW, Morris V. Joint physiology: relevant to the rheumatologist? Br J Rheumatol 1998;37:121–5. ˇ eponis A, Konttinen YT, Imai S et al. Synovial lining, endothel2. C ial and inflammatory mononuclear cell proliferation in synovial membranes in psoriatic and reactive arthritis: a comparative quantitative morphometric study. Br J Rheumatol 1998; 37:170–8. 3. Mohr W, Beneke G, Mohing W. Proliferation of synovial lining cells and fibroblasts. Ann Rheum Dis 1975;34:219–24. 4. Qu Z, Hernandez Garcia C, O’Rourke LM, Planck SR, Kohli M, Rosenbaum JT. Local proliferation of synoviocytes contributes to synovial hyperplasia. Arthritis Rheum 1994;37:212–20. 5. Nagashima M, Yoshino S, Ishiwata T, Asano G. Role of vascular endothelial growth factor in angiogenesis of rheumatoid arthritis. J Rheumatol 1995;22:1624–30. 6. Walsh DA, Wade M, Mapp PI, Blake DR. Focally regulated endothelial proliferation and cell death in human synovium. Am J Pathol 1998;152:691–702. 7. Lalor PA, Mapp PI, Hall PA, Revell PA. Proliferative activity of cells in the synovium as demonstrated by a monoclonal antibody, Ki67. Rheumatol Int 1987;7:183–6. 8. Firestein GS, Yeo M, Zvaifler NJ. Apoptosis in rheumatoid synovium. J Clin Invest 1995;96:1631–8. 9. Sugiyama M, Tsukazaki T, Yonekura A, Matsuzaki S, Yamashita S, Iwasaki K. Localisation of apoptosis and expression of apoptosis related proteins in the synovium of patients with rheumatoid arthritis. Ann Rheum Dis 1996;55:442–9. 10. Walsh DA, Hu D-E, Mapp PI, Polak JM, Blake DR, Fan T-PD. Substance P and new vessel formation in the rat subcutaneous sponge granuloma. Histochem J 1996;28:759–69. 11. Mapp PI, Kidd BL, Gibson SJ et al. Substance P-, calcitonin gene-related peptide, and C-flanking peptide of neuropeptide Y-immunoreactive fibres are present in normal synovium but depleted in patients with rheumatoid arthritis. Neuroscience 1990;37:143–53.

LETTERS TO THE EDITOR 12. Pereira da Silva JA, Carmo-Fonseca M. Peptide containing nerves in human synovium: immunohistochemical evidence for decreased innervation in rheumatoid arthritis. J Rheumatol 1990;17:1592–9. 13. Joyce TJ, Yood RA, Carraway RE. Quantitation of substance P and its metabolites in plasma and synovial fluid from patients with arthritis. J Clin Endocrinol Metab 1993;77:632–7. 14. Mapp PI, Walsh DA, Kidd BL, Cruwys SC, Polak JM, Blake DR. Localisation of the enzyme neutral endopeptidase to the human synovium. J Rheumatol 1992;19:1838–44. 15. Walsh DA, Mapp PI, Wharton J, Polak JM, Blake DR. Neuropeptide degrading enzymes in normal and inflamed human synovium. Am J Pathol 1993;142:1610–21. 16. Stevens CR, Blake DR, Merry P, Revell PA, Levick JR. Comparative study by morphometry of the microvasculature in normal and rheumatoid synovium. Arthritis Rheum 1991; 34:1508–13. ˇ eponis A, Konttinen Y, Mackevicius Z, Solovieva SA, 17. C Hukkanen M, Tamulaitien M et al. Aberrant vascularity and von Willebrand factor distribution in inflamed synovial membrane. J Rheumatol 1996;23:1880–6.

Reply While agreeing with the substance of the comments made by Walsh and Mapp, which appear consistent with our own, we are unclear as to what point they are trying to put across. There is no doubt that rheumatoid arthritis ( RA) is associated with fibroblast proliferation and vascular and neuronal damage and regeneration, but these are all expected features of chronic inflammation. Proliferation does not imply a link to cancer. There has been a vogue for suggesting that RA synovium is quasi-neoplastic. However, as Walsh and Mapp point out, there is a balance between proliferation and cell death in RA of the sort typical of inflammation, but not neoplasia. Histopathologists do not confuse RA with cancer. Terms like ‘immature phenotype’ or ‘mesenchymoid transformation’ are loaded with implications and mean rather little. There is often confusion about which cells are proliferating. ‘Synovial hyperplasia’ is often used to mean that the whole tissue is bigger, with villous formation, when the phrase was created to describe the intima. The phrase is now so degraded by misusage that it should be laid to rest. The increase in number (‘hyperplasia’) of intimal cells in RA is mostly due to macrophage accumulation. There are extra fibroblasts, but alone they would not have been remarked on. The search for evidence of division was based on a misconception. Despite recent studies using PCNA or oncogenes, which give indirect evidence of potential to divide, older studies of mitotic figures and DNA content, which record replication, would still seem to be more reliable and are unremarkable. We are agreed that the relative paucity of nerves in comparison to vessels in rheumatoid synovium may relate to recent angiogenesis, but that is a different issue from the apparent discordance between the level of neuropeptides in the few nerves present and in synovial fluid, which suggests secretion. The conclusion that remains is that rheumatoid synovium is fairly ordinary chronic inflammatory tissue with a surface coating of activated FccRIIIa+ macro-

1033

phages and unusually marked lymphoid follicle formation. The physiology of rheumatoid synovium is, of course, different from that of the normal joint. It is dynamic, but then so is other granulation tissue. J. C. W. E, V. M

Centre for Rheumatology, Arthur Stanley House, Tottenham Street, London W1P 9PG An Open Study of Oxpentifylline in Early Rheumatoid Arthritis SThe phosphodiesterase (PDE) inhibitors are a family of drugs with many anti-inflammatory properties, including inhibition of TNF transcription [1–3]. Oxpentifylline (OXP), a non-specific PDE inhibitor, has been shown to reduce significantly the onset of mercuric chloride-induced arthritis in rats [4], and to have a significant ameliorative effect on severe RA in humans [5, 6 ]. Unfortunately, OXP is poorly tolerated in patients with long-standing RA [5, 6 ], possibly because severe systemic disease confers poor tolerance of drugs in general. We hypothesize that patients with either early or relatively mild RA might be more tolerant and therefore an ideal group for OXP treatment. We have conducted an open prospective study of OXP, 400 mg three times daily, over 12 weeks in 20 patients (median age 59 yr, range 26–75 yr, 16 female) fulfilling the 1987 ARA revised criteria for RA. The median disease duration was 2 yr (range 1–17 yr), seven had erosions and all were rheumatoid factor positive. At enrolment, all patients had three or more of the following disease activity criteria on two consecutive occasions (at least 1 week apart): early morning stiffness ( EMS) > 45 min (median 75, range 15–210), three or more tender or swollen joints (median tender 10, range 0–17; swollen 10, range 2–20), ESR > 25 mm/h (median 56, range 26–87), CRP > 20 mg/l (median 31, range 0–105). Five patients had had mild RA for >5 yr with a median swollen joint count of 10/28 (range 7–17), EMS of 60 min (range 20–60) and CRP of 46 mg/l (range 4–90). All patients were taking a non-steroidal antiinflammatory drug (NSAID) and two were taking a steady dose (9 weeks treatment (n = 11)

Baseline

End of trial

Baseline

Week 12

10 10 75 6 152 1.61 3 3 56 31

8.5 8 30** 4.4 165 0.875** 3.5 3.5 57 21

11 8 90 5.1 175 1.625 3 3 58 38

5** 4* 20** 2.2** 181 0.75** 4 4 58 27

Grip, mean score of both hands for each patient; HAQ, Health Assessment Questionnaire; patient and assessor global score: 1 = very poor–5 = very good. *P < 0.05; **P < 0.005 Wilcoxon signed rank test.

Outcome data were collected twice at baseline and every 2 weeks for 12 weeks during treatment: EMS, pain score (10 cm visual analogue scale), tender and swollen joint counts (total 28), grip strength (mean of right and left hands, mmHg), Stanford Health Assessment Questionnaire (HAQ, modified for UK patients), patient’s and assessor’s global assessment (five-point scale), ESR, CRP, full blood count, urea, electrolytes and liver enzymes. Compliance was assessed by performing a tablet count at each assessment. No change in NSAID or other treatment was permitted. Responders were classified according to the 20% and 50% Paulus response criteria [7]. A comparison of mean baseline and end of the trial (the time a patient withdrew or the 12 week assessment) outcome measures was performed with the Wilcoxon signed rank test. The number of patients remaining in the trial, withdrawals and Paulus responses at fortnightly intervals during the trial are shown in Table I. Ten patients (50%) completed 12 weeks of OXP treatment and seven (35%) opted to continue with OXP after the study. Adverse events were reported by nine patients (45%), most frequently neck or occipital pain (n = 5), which resolved in four cases whilst OXP was continued. The five patients who withdrew early due to adverse events (mainly gastrointestinal ) were significantly older than the remaining 15 patients (median age 71.5 vs 56.5 yr, respectively; P = 0.007, Mann–Whitney U ) and had a higher tender joint count (median count 13.5 vs 8, respectively; P = 0.01, Mann–Whitney U ). There were no other significant differences in baseline demographic, clinical or laboratory scores between patients who completed the trial and those withdrawn due to inefficacy or adverse events. On Paulus criteria, there was a 20% improvement in 7/20 patients (35%) and a 50% improvement in 3/20 patients (15%). Comparison of individual baseline and end of trial outcome measures (see Table II ) revealed a significant improvement in EMS and HAQ in the whole study group. In contrast, a more substantial

clinical response was seen in the patients who remained on OXP treatment for >9 weeks (n = 11). In this group, the 20 and 50% Paulus response rate was 64 and 27%, respectively, and there were significant improvements in 5/6 clinical scores, but not in ESR or CRP (see Table II ). In conclusion, OXP treatment was not associated with any serious adverse events; however, tolerability was poor despite the selection of patients with early or mild RA. On an intention-to-treat basis, efficacy was disappointing, but there was an impressive clinical response, particularly in the HAQ, in those patients who remained on OXP for >9 weeks (including four with side-effects). However, other than by age, the responders could not be distinguished at baseline from those destined to withdraw due to adverse events or inefficacy. The problem of poor tolerability may be overcome by the next generation of isozyme-specific PDE inhibitors. Lymphocytes and macrophages preferentially express the type III and IV isozymes of PDEs [8, 9], and Rolipram, a type IV inhibitor, is reported to ameliorate established collagen II-induced arthritis in rats both clinically and radiologically [10]. Our findings in humans lend clinical support to the theoretical rationale to develop PDE inhibitors as a safe means of targeting the immune response in the inflamed joint. We thank Dr O. Duke for helpful discussions and access to patients, Mrs J. Dunwoody for help with data collection and Hoechst Marion Roussel for the supply of oxpentifylline. P. D. W. K, D. J,* B. E. B*

Department of Rheumatology, St Helier NHS Trust, Carshalton, Surrey and *Department of Rheumatology, St George’s Healthcare NHS Trust, London Accepted 15 May 1998 Correspondence to: P. D. W. Kiely, Department of Rheumatology, St George’s Healthcare NHS Trust, Blackshaw Road, London SW17 0QT.

LETTERS TO THE EDITOR 1. Kammer GM. The adenylate cyclase-cAMP-protein kinase A pathway and regulation of the immune response. Immunol Today 1988;9:222–9. 2. Moore AR, Willoughby DA. The role of cAMP regulation in controlling inflammation. Clin Exp Immunol 1995;101:387–9. 3. Anaya J-M, Espinoza LR. Phosphodiesterase inhibitor pentoxifylline: an anti-inflammatory/immunomodulatory drug potentially useful in some rheumatic diseases. J Rheumatol 1995; 22:595–9. 4. Kiely PDW, Gillespie KM, Oliveira DBG. Oxpentifylline inhibits tumour necrosis factor a mRNA transcription and protects against arthritis in mercuric chloride treated Brown Norway rats. Eur J Immunol 1995;25:2899–906. 5. Maksymowych WP, Avina-Zubieta A, Luong MH, Russell AS. An open study of pentoxifylline in the treatment of severe refractory rheumatoid arthritis. J Rheumatol 1995;22:625–9. 6. Huizinga TWJ, Dijkmans BAC, van der Velde EA, van de Pouw Kraan TCTM, Verweij CL, Breedveld FC. An open study of pentoxifylline and thalidomide as adjuvant therapy in the treatment of rheumatoid arthritis. Ann Rheum Dis 1996;55:833–6. 7. Paulus HE, Egger MJ, Ward JR, Williams HJ. Analysis of improvement in individual rheumatoid arthritis patients treated with disease-modifying antirheumatic drugs, based on the findings in patients treated with placebo. Arthritis Rheum 1990;33:477–84. 8. Essayan DM, Huang S-K, Undem BJ, Kagey-Sobotka A, Lichtenstein LM. Modulation of antigen- and mitogen-induced proliferative responses of peripheral blood mononuclear cells by non selective and isozyme selective cyclic nucleotide phosphodiesterase inhibitors. J Immunol 1994;153:3408–16. 9. Gantner F, Kupferschmidt R, Schudt C, Wendel A, Hatzelmann A. In vitro differentiation of human monocytes to macrophages: change of PDE profile and its relationship to suppression of tumour necrosis factor-alpha release by PDE inhibitors. Br J Pharmacol 1997;121:221–31. 10. Nyman U, Mussener A, Larsson E, Lorentzen J, Klareskog L. Amelioration of collagen II-induced arthritis in rats by the type IV phosphodiesterase inhibitor Rolipram. Clin Exp Immunol 1997;108:415–9.

1035

Depression in Rheumatoid Arthritis SI read, with great interest, the study of depression in rheumatoid arthritis [1]. With the availability of newer diagnostic tools, depression can be more accurately measured and characterized. Nonetheless, that it played a part in RA symptoms has long been assumed; the depressive aspects seem to parallel the degree of chronic pain, as profiles in tests are similar to those seen in other chronic pain syndromes, not rheumatological in nature. The respected authors of this paper state that further research is needed to determine whether RA is accompanied by more depression than osteoarthritis. That study was performed some time ago [2] and is referred to in a later paper [3]. I think that is because RA patients characterize their symptoms as coming from a disease (RA), whilst osteoarthritis patients depersonalize the pain (my knee hurts). That distinction probably sets up a different prognostic algorithm for the sufferer. G. E. E

241 South Sixth Street, Philadelphia, PA 19106-3731, USA Accepted 28 May 1998 1. Abdel-Nasser AM, El-Azim S, Taal E, El-Badawy SA, Rasker JJ, Valkenburg HA. Depression and depressive symptoms in rheumatoid arthritis patients: an analysis of their occurrence and determinants. Br J Rheumatol 1998;37:391–7. 2. Spergel P, Ehrlich GE, Glass DD. The rheumatoid personality— A psychodiagnostic myth. Psychosomatics 1978;19:79–86. (Abstract: Arch Phys Med Rehabil 1976;57:561.) 3. Ehrlich GE. Arthritis and its problems. Clin Rheum Dis 1981;7:302–20.