The aim of our study was to determine the current incidence and outcome of infected total knee arthroplasty (TKA) in our unit comparing them with our earlier audit in 1986, which had revealed infection rates of 4.4% after 471 primary TKAs and 15% after 23 revision TKAs at a mean follow-up of 2.8 years. In the interim we introduced stringent antibiotic prophylaxis, and the routine use of occlusive clothing within vertical laminar flow theatres and 0.05% chlorhexidine lavage during arthroplasty surgery.

We followed up 931 primary TKAs and 69 revision TKAs for a mean of 6.5 years (5 to 8). Patients were traced by postal questionnaire, telephone interview or examination of case notes of the deceased.

Nine (1%) of the patients who underwent primary TKA, and four (5.8%) of those who underwent revision TKA developed deep infection. Two of nine patients (22.2%) who developed infection after primary TKA were successfully treated without further surgery. All four of the patients who had infection after revision TKA had a poor outcome with one amputation, one chronic discharging sinus and two arthrodeses.

Patients who underwent an arthrodesis had comparable Oxford knee scores to those who underwent a two-stage revision. Although infection rates have declined with the introduction of prophylactic measures, and more patients are undergoing TKA, the outcome of infected TKA has improved very little.

Our aim in this study was to determine the outcome of hip arthroplasty with regard to infection at our unit. Infection after total joint arthroplasty is a devastating complication. The MRC study in 1984 recommended using vertical laminar flow and prophylactic antibiotics to reduce infection rates. These measures are now routinely used. Between 1993 and 1996, 1727 primary total hip arthroplasties and 305 revision hip arthroplasties were performed and 1567 of the primary and 284 of the revision arthroplasties were reviewed between five and eight years after surgery by means of a postal questionnaire, telephone interview or examination of the medical records of those who had died.

Seventeen (1.08%) of the patients who underwent primary and six (2.1%) of those who underwent revision arthroplasty had a post-operative infection. Only 0.45% of patients who underwent primary arthroplasty required revision for infection.

To our knowledge this is the largest multi-surgeon audit of infection after total hip replacement in the UK. The follow-up of between five and eight years is longer than that of most comparable studies. Our study has shown that a large cohort of surgeons of varying seniority can achieve infection rates of 1% and revision rates for infection of less than 0.5%.

Patients who undergo total knee replacement (TKR)are at high risk of venous thromboembolism. Low-molecular-weight heparins (LMWH) are the most suitable chemical prophylactic agents but there are some uncertainties about their safety and effectiveness. The foot pump offers an alternative.

We randomised 229 patients undergoing primary, unilateral TKR to receive either the A-V Impulse foot pump or enoxaparin, a LMWH. Ascending venography was undertaken between the sixth and eighth postoperative day in 188 patients without knowledge of the randomisation category. The prevalence of venographic deep-vein thrombosis was 58% (57/99) in the foot-pump group and 54% (48/89) in the LMWH group which was not statistically significant. There were four cases of proximal thrombi and two of fatal pulmonary emboli in the foot-pump group and none in the LMWH group. There were fewer haemorrhagic complications and soft-tissue effects in the foot-pump group.

We conclude that the neither method provides superior prophylaxis.

Chemical prophylaxis is known to reduce the venographic prevalence of deep-vein thrombosis (DVT) after total knee replacement (TKR), but it is uncertain whether this affects the incidence of symptoms. Further analysis depends on the basic epidemiology of thromboembolic symptoms. We therefore studied the pattern of such symptoms in a consecutive series of 1000 patients with primary TKR, with particular reference to risk factors and prophylaxis.

We reviewed all the clinical records and contacted all the patients individually, noting risk factors, prophylaxis, symptomatic pulmonary embolus (PE) or DVT and its timing, death and its causes, and all complications. All the patients wore antiembolism stockings, 83% had regional anaesthesia and 33.9% had chemical prophylaxis.

One patient died from PE on the day of surgery, having had no prophylaxis giving a rate of 0.1% (95% CI 0.003% to 0.56%). Symptomatic, radiologically confirmed thromboembolism (VTE) was common with a rate of 10.6% (95% CI 8.7% to 12.5%). There was a similar incidence of VTE in those with and without chemical prophylaxis (10.1% v 10.5%, RR 0.96, NS). VTE was more common in patients with risk factors (15.1% v 9.5%, RR 1.59, p = 0.02) and tended to occur earlier in this group (median day of onset 5 v 7, p = 0.01). Chemical prophylaxis did not reduce the frequency of symptomatic thromboembolism in either those with risk factors (RR 0.81, p = 0.5) or those without them (RR 0.94, p = 0.8). Haematoma or wound dehiscence was more common in those having chemical prophylaxis (11.9% v 6.9%; RR 1.73 95% CI 1.16 to 2.60). Readmission for symptomatic, radio-logically confirmed thromboembolism involved 1.1% of patients (95% CI 0.55% to 2.1%). Four patients were readmitted with proven non-fatal PE and six with proven DVT (the latest on day 40).

Our results show that the main risk factor for thromboembolism was TKR itself; chemical prophylaxis did not reduce the incidence of symptomatic thromboembolism but gave an increased perception of side-effects. New prophylactic methods or combinations of methods are needed, with their efficacy compared by randomised controlled studies of both the clinical and the radiological effect.