Anti-tumour necrosis factor (TNF) α treatment has had a tremendous effect on the managment of rheumatoid arthritis, but it needs to be monitored closely. Guidelines state that haematological complications are rare, so regular full blood counts are not recommended.^1,2 We noticed that a minority of our patients with rheumatoid arthritis receiving anti-TNF treatment developed neutropenia (<2×10⁹/l). Table 1 summarises the demographics of 133 patients with rheumatoid arthritis. Of these, 14.3% of patients had at least one episode of neutropenia, with no obvious cause other than the anti-TNF treatment. The lowest neutropenic episodes had counts that ranged from 1.02 to 1.94 (8/19 patients had an episode with counts <1.5). The time period between starting anti-TNF treatment and developing neutropenia ranged from 1 week (two patients) to 26 months, with a median of 3 months. Patients with a neutrophil count at baseline <4 had an odds ratio (OR) of 10.7 (95% confidence interval (CI) 3.6 to 31.4) for developing neutropenia. A history of neutropenia when receiving other disease-modifying antirheumatic drugs (DMARDs) gave an OR of 11.7 (95% CI 3.9 to 34.7).

Of the 19 patients, 16 who developed neutropenic episodes while receiving anti-TNF treatment have managed to stay on their original treatment. The usual approach has been temporary cessation, with reinstatement once the neutrophil count has recovered, or in some circumstances a lower tolerated threshold level being set, with no further problems encountered (>1.5). In our experience, patients gain so much benefit from anti-TNF treatment that they are reluctant to stop and are happy to continue, despite the potential risks of neutropenia. However, two of these patients have since had to stop anti-TNF, with one undergoing investigations into the aetiology of a pleural effusion, and another developing a resistant staphylococcal foot infection. One patient who was taking infliximab had recurrent episodes of neutropenia, which were managed with temporary cessation, but subsequently the patient was changed to etanercept and since then has had no further episodes. Another patient changed from etanercept to adalimumab without further problems.

There has been a small number of reports of cytopenias in patients receiving anti-TNF treatment.^1–7 However, no other series of patients have reported a rate of neutropenias affecting 1 in 8 patients, which seems to be the case in our experience. Only one of our patients who developed neutropenia developed a concomitant infection. Our experience of rates of development of neutropenia in patients receiving other commonly prescribed DMARDs such as methotrexate and leflunomide is 12.5% and 14.9%, respectively.⁸ In these patients, we perform regular full blood counts, in keeping with national guidelines. The prevalence of neutropenia in patients receiving anti-TNF treatment in our experience is similar at 14.3%, so that it would be inconsistent and unsafe for us not to monitor regular full blood counts in these patients. We would recommend regular full blood counts in all patients receiving anti-TNF treatment, whether they are on concomitant methotrexate or not, and believe that guidelines need to be modified to take this into account. Patients with a baseline neutrophil count <4 and those with a history of neutropenia when receiving DMARDs, need to be monitored particularly closely.