Prof Richard Montgomery, MB BS, FRCSEd, FRCSEng, Honorary Consultant Trauma & Orthopaedic Surgeon
One’s perspective on septic arthritis (SA) may differ depending on whether you are an orthopaedic surgeon (large joints), a hand surgeon (small joints), a foot surgeon (whose experience will include diabetic feet), an infectious disease specialist/microbiologist, or a rheumatologist. The variation in opinion reflects real differences in these subspecialties’ experiences of the condition.
All orthopaedic surgeons are taught that septic arthritis (SA) is an urgent condition. But how urgent is it really, and does that urgency differ depending on which endpoint you look at?
There are a number of adverse endpoints of septic arthritis (SA) to consider: death; spread into adjacent bone, into nearby soft tissues, or into the bloodstream; and finally joint destruction.
Every sufferer is an individual, and each individual/host sits somewhere on a spectrum of ability to localise infection, or to resist widespread sepsis. The patient who rapidly succumbs to overwhelming sepsis may figure in overall figures, but not as a long term joint destruction case in orthopaedic follow-up.
It is useful to think of SA as involving an interaction between organism & host, modified by a number of characteristics:
- The virulence & characteristics of the organism(s)
- The immune competence of the host (normal/local compromise/systemic compromise/local & systemic compromise)
- The particular characteristics of the joint(s) (small/large/intra-articular metaphysis etc)
- Haematogenous (single/multiple/ secondary to septic focus in skin/UTI/ endocarditis etc) or direct inoculation
In 1976, Newman wrote ‘Review of septic arthritis throughout the antibiotic era’. Ann. rheum. Dis. (1976), 35, 198. He reviewed 130 patients accumulated over a 30 year period at the Nuffield Orthopaedic Centre, covering the period from the advent of effective antibiotics just after World War 2. Hip & knee infections accounted for 111/130 infections; strep & staph accounted for 93/130; organism unknown occurred in 32.
He classified it as:
A: Organism isolated from joint
B: Organism isolated elsewhere
C: No organism isolated but (i) histological or radiological evidence of
infection, or (ii) turbid fluid aspirated from joint.
The above classification is still in use.
Newman concluded:
The problems and necessity of rapidly establishing a diagnosis are
stressed. Overall, 70% attained a good result though infection in infants’ hips and all joints in the elderly carried a poor prognosis. Once a good result was achieved the joint did not deteriorate with the passage of time.
The last sentence there is not widely accepted these days, as more recent studies show that subsequent deterioration of the joint occurs much more frequently than in the general population.
The most common causative organisms remain staphylococcal and streptococcal species comprising 40.6% and 28% of cases respectively (Ryan et al, BACTERIAL JOINT INFECTIONS IN ENGLAND AND WALES. ANALYSIS OF BACTERIAL ISOLATES OVER A FOUR YEAR PERIOD.British Journal of Rheumatology 1997; 36: 370-373).
Who gets SA?
It is impossible to understand SA in adults without understanding a bit about individual susceptibility. It is a common observation that patients suffering the condition are often in poor health beforehand, due to chronic illness, immunosuppression, age, or addiction of one form or another.
A prospective 2-year study of septic arthritis identified 75 patients with synovial fluid culture positive (Newman A) patients. Gupta, Sturrock & Field. ‘A prospective 2‐year study of 75 patients with adult‐onset septic arthritis’. Rheumatology, Volume 40, Issue 1, January 2001, Pages 24–30, https://doi.org/10.1093/rheumatology/40.1.24
Underlying joint disease was found in 46/75 (including 25 rheumatoid patients, & 15 osteoarthritis patients).
IV drug users accounted for 11/75. 15% involved more than 1 joint.
Staphylococci and streptococci represented 90% of organisms cultured.
56% involved the knee.
78% lived in areas of high social deprivation. Whether this reflects the association of SA with chronic illness & addiction, or is a separate effect, is not clear.
C-reactive protein was a useful investigation, being elevated in 98%. Fever was present in 64%; WBC was normal in 38%.
Leg ulcers were present in 11%, but in 38% of those who died. It suggests a poorer outlook. Abnormal renal function and raised WBC were predictors of poor outcome.
Most had IV therapy for 2 weeks, followed by 3 weeks oral medication.
37% required surgical intervention; a figure that seems low to this orthopaedic surgeon! Selection must be at work.
Several had had distal sites of infection; or recent joint instrumentation; or insertion of a pacemaker.
4 were diabetics, 4 had alcoholic liver disease.
In the 25 RA patients, 5 took prednisolone, 5 took methotrexate, 7 were on gold salts, hydroxychloroquine, sulphasalazine or penicillamine. 9 were not taking any DMARD.
(Author: Had this study taken place more recently, I suspect that anti – TNF therapy would have featured more strongly.)
Presentation
The classic symptoms & physical findings are well-known.
However, systemic signs are less common than expected. In a prospective study, fever was found in 34%; sweats in 15%; and rigors in 6%. At presentation, fever is present in ~60%. (Gupta MN, Sturrock RD, Field M. A prospective 2-year study of 75 patients with adult-onset septic arthritis. Rheumatology (Oxford) 2001; 40: 24–30).
In recent years, the use of ‘biologics’ (anti-TNF treatments, immunosuppressive disease-modifying antirheumatic drugs), has modified the presenting features. The normal findings of severe pain and tenderness may be absent or simply mimic an exacerbation of the chronic joint condition. A minimally raised white cell count and elevated C-reactive protein in the absence of systemic signs of infection may be interpreted as further evidence for the diagnosis of an exacerbation of inflammatory arthritis. (Salar et al, ‘Septic arthritis in the era of immunosuppressive treatments’ Ann R Coll Surg Engl 2014; 96: e11–e12 doi 10.1308/003588414X13814021678196).
This lack of physical signs in a group particularly prone to SA represents a bear-trap ready to ensnare the inexperienced or unwary.
Blood cultures may be positive in 24% in whom synovial fluid carries organisms, and in 9% they were the only positive microbiological test. (Matthews et al, ‘Bacterial septic arthritis in adults’. Lancet 2010; 375: 846–55).
Empiric Treatment before microbiological results are available.
The best advice one can give is to consult your clinical microbiologist or infectious diseases specialist to ascertain the up-to-date & local recommendations regarding likely organisms & sensitivities.
However, as a background, Clerc et al (2011) studied 11 years of cases of adult native septic arthritis. The main organisms were methicillin-susceptible Staphylococcus aureus (MSSA) and streptococci (respectively, 44.6% and 14.2% of cases). Only 11 cases (4.7%) of methicillin-resistant S. aureus (MRSA) arthritis were diagnosed, among which 5 (45.5%) occurred in known carriers.
They concluded: ‘Amoxicillin/clavulanate or cefuroxime would be adequate for empirical coverage of large-joint septic arthritis in our area. A broad-spectrum antibiotic would be significantly superior for small-joint infections in diabetics. Systematic coverage of MRSA is not justified, but should be considered for known carriers’.
Clerc et al, ‘Adult native septic arthritis: a review of 10 years of experience and lessons for empirical antibiotic therapy’. J Antimicrob Chemother 2011; 66: 1168–1173
doi:10.1093/jac/dkr047
Death from septic arthritis
Abram et al, ‘Mortality and adverse joint outcomes following septic
arthritis of the native knee: a longitudinal cohort study of
patients receiving arthroscopic washout’. Lancet Infect Dis 2020;
20: 341–49.
n= 12 132.
Of 10 195 with admission diagnosis of SA, 90 day mortality = 7.05% (but 22.69% in the over 79 year olds).
Shirtliff & Mader, CLINICAL MICROBIOLOGY REVIEWS, Oct. 2002, p. 527–544 Vol. 15, No. 4. 0893-8512/02/$04.000 DOI: 10.1128/CMR.15.4.527–544.2002.
These authors said: ‘The mortality associated with this disease is usually between 5 and 20% and is often a result of the transient or chronic bacteremia that causes most cases of septic arthritis (5, 77–79, 86). This
high rate has not changed significantly over the last 40 years....’
McBride et al, ‘Epidemiology, Management, and Outcomes of Large and
Small Native Joint Septic Arthritis in Adults’. Clinical Infectious Diseases® 2020;70(2):271–9.
Small Joint SA, n=250.
7 day mortality = 0.4%; 1% thereafter
Large Joint SA, n= 302.
7 day mortality = 9/302 (3%), 30 day mortality 19/302 (6%), 90 day mortality 26/302 (9%).
In other words, one third had occurred in the first week, two thirds within the first month, and the rate reduced further after that.
Weston et al, ‘Clinical features and outcome of septic arthritis in a single UK Health District 1982–1991’. Ann Rheum Dis 1999;58:214–219:
Pre-existing joint disease was evident in 35% of cases. The mortality remains high at 11.5% with a significant additional morbidity of
31.6%. Multivariate analysis suggests that important predictors of death are: confusion at presentation, age >65 years, multiple joint sepsis or involvement of the elbow joint, and of morbidity are: age > 65 years,
diabetes mellitus, open surgical drainage, and Gram positive infections other than S aureus.
Overall, the studies above suggest death is typically an early phenomenon after SA, and higher in the elderly & previously sick. No surprise there.
Joint destruction
Studies in rodents with SA indicate very rapid damage to articular cartilage. Antibiotic prophylaxis in these animal models can prevent joint damage, otherwise some degree of joint damage seems inevitable even if an antibiotic was started 4 hours after onset:
Beginning the antibiotic treatment one day after infection reduced over-all loss of collagen by 37% and decreased the area of erosion of the infected articular surfaces. When antibiotic treatment was begun at four, eight, or twelve hours after infection, the loss of glycosaminoglycan averaged 18%.
From: ‘The effect of antibiotics on the destruction of cartilage in experimental infectious arthritis’. Smith et al, JBJS- A, Sep 1987, 69(7):1063-1068 PMID: 3654698.
Of course in humans, treatment within the short timescales above is simply not realistic. Many patients may delay for days before presentation, yet progressive joint damage does not seem inevitable. The majority of individuals recover after treatment within this timeframe. This fact alone suggests that there is a limit to which animal model studies can be used to make recommendations, or guide prognosis, regarding SA outcomes in humans.
A lot of the experimental work on SA has been done with knockout & transgenic mice, and selected arthritogenic staphylococci. These animals’ maturation & life expectancy are telescoped into 3 years. Whilst this work is valuable to elucidate the mechanisms involved, the extent to which this work can translate & be used to formulate clinical responses & management in humans is limited.
It seems to be the case that some of the damage to joint surfaces is caused by the bacteria, and some by the immune response of the host. Some of that immune response is helpful, but some is actively harmful.
E Sakiniene, T Bremell, A Tarkowski :’Addition of corticosteroids to antibiotic treatment ameliorates the course of experimental Staphylococcus aureus arthritis’. Arthritis & Rheumatism Volume39, Issue9, September 1996, Pages 1596-1605. These authors found the prevalence of arthritis 14 days after the onset of the disease was 22% in the corticosteroid and antibiotic–treated group, as compared with 81% in the control (nontreated) group and 48% in the antibiotic‐treated group. The severity of arthritis also decreased in the corticosteroid and antibiotic‐treated group, as did the mortality rate.
A 2015 systematic review & meta-analysis regarding the use of corticosteroids in septic arthritis (L Farrow. BMC Musculoskeletal Disorders volume 16, Article number: 241, 2015) concluded:
The current evidence base points strongly towards a beneficial effect for corticosteroids in septic arthritis. This is consistent with scientific rationale and data regarding glucocorticoid use in sepsis at other anatomical sites. There is however insufficient evidence to make treatment recommendations at present. Further large scale randomised controlled trials including adult subjects; safety profiles and long term efficacy are required for this promising therapeutic option.
It is clearly possible that progress on improving the outcomes may come from a combination of antibiotics, lavage, and inflammatory suppression in future.
Some clinicians challenge the timescales from animal models, as they relate to joint damage in humans.
Delay in diagnosis, and the effect of timing of treatment on the outcome of disease.
Kaandorp et al, ‘The outcome of bacterial arthritis. A prospective community‐based study’. Arthritis & Rheumatism, Volume 40, Issue5
May 1997, Pages 884-892. https://doi.org/10.1002/art.1780400516
In a prospective community study, data was acquired at presentation & 2 years later. n = 121 adults & 33 children (155 total). A poor patient outcome was defined as death due to SA or severe overall functional deterioration. A poor joint outcome was defined as amputation, arthrodesis, prosthetic surgery, or severe functional deterioration. The series contained native and previously operated joints.
One‐half of the adults had a preexisting joint disease and 29% of the infected joints contained synthetic material. The patient outcome was poor in 21% of all patients, and the joint outcome was poor in 33% of the surviving patients. Adverse prognostic factors were an older age, preexisting joint disease, and an infected joint containing synthetic material. These factors were interrelated. There was no association between a poor outcome and young age, comorbidity, immunosuppressive medication, functional class, multiple infected joints, type of microorganism, or treatment delay.
A problem with this study is that it includes prosthetic infections as well as native joints (39 prostheses and 11 fixation implants in a series of 174). I think prosthetic joint infections and native joint infections are best considered separately.
Chen et al, ‘Surgical Treatment for Septic Arthritis of the Knee Joint in Elderly Patients: A 10-year Retrospective Clinical Study’. Orthopedics, Vol 36, 4, e434-443, 2013. doi: 10.3928/01477447-20130327-19
n= 92 patients with knee SA. 33 had arthroscopic debridement, 45 underwent open debridement. 7 had arthrodesis (6 of whom had started treatment after 3 weeks of symptoms), 22 had arthroplasty (18 of whom had started treatment after 3 weeks of symptoms), 19 were waiting for arthroplasty (of whom 18 had started treatment after 3 weeks of symptoms). No patient in the arthrodesis/TKR/waiting for TKR groups had started treatment within the first week. 30 had no or only mild symptoms (27 of these had started treatment within the 1st 3 weeks).
Earlier presentation is the norm in western populations, but this study is still useful to us. It suggests that as far as knee joint destruction is concerned, timing is related to outcome if treatment is delayed beyond the first week from symptom onset. (NB This conclusion should not be extrapolated to other areas. For example, the hip may be vulnerable earlier because of potential tamponade effects on vasculature).
Lauper et al, ‘Native septic arthritis is not an immediate surgical emergency’. Journal of Infection 77 (2018) 47–53.
n=204. Sequelae followed in 83 (41%), being 15 recurrences, 30 secondary arthroses, 9 persistent pain, 9 Girdlestone procedures, 9 arthrodeses, 8 amputations, 8 stiffness, & 2 Chronic Regional Pain Syndrome (CRPS).
The median delay between the onset of symptoms & the first joint surgery was 6 days (range 0-214). The overall median delay between admission to the hospital and joint lavage was 18 h (range 1 h–19 days).
The following factors did not predict sequelae: age, steroid treatment, pre-existing arthropathy, duration of antibiotics, type of joint, number of operations, number of days of pre-hospital symptoms, or hours spent in ED. Patients who had lavage within 6 hours had similar outcomes to those who received lavage after longer intervals.
These authors concluded: Our data suggest that for native septic arthritis, in the absence of clinical sepsis, immediate joint drainage does not appear to reduce the risk of sequelae compared with delayed drainage.
Long term outcome.
The Abram study, reference above, concluded:
‘Within 15 years, 159 (8·76%; 95% CI 7·50–10·15) of 1816 patients had
received arthroplasty, corresponding to an annual risk of arthroplasty that was about six times that of the general population (risk ratio 6·14, 95% CI 4·95–7·62; p<0·0001).‘
Outcome of arthroplasty following septic arthritis.
Kunze et al, ‘Two-Stage Primary Arthroplasty of Native Hips and Knees That Had Previously Failed Treatment for Septic Arthritis: A Single-Center Experience.’ Arthroplasty Today 6 (2020) 431-436.
n=42. At a mean of 3.3-year follow-up, there were 14 (33.3%) complications and the infection cure rate was 95.2%.
On average, patients improved in the modified Harris Hip Score (42.9 ± 11.8 vs 83.3 ± 11.1, P <.001), KSS (35.9 ± 16.9 vs 80.1 ± 16.6, P < .001), KSSF (38.0 ± 15.1 vs 71.5 ± 24.0, P < .001), knee flexion (90.9 ± 14.9 vs 100.5 ± 17.1), and hip flexion (73.8 ± 21.2 vs 102.1 ± 11.8, P < .001).
Conclusion: Patients showed significant postoperative improvements but a high rate of complications after two-stage primary total joint arthroplasty. Some patients required a third operation.
Summary
Septic arthritis treatment is urgent. SA in large joints is still a killer.
Rapid diagnosis/microbiology sampling and empiric antibiotic treatment may save the lives of vulnerable patients with underlying pathologies or frailty who otherwise may succumb rapidly to the effects of generalised sepsis. The antibiotic prescription may need to be altered once the organisms have been identified & sensitivities have been determined.
Give thought to the possibility of endocarditis or another intrinsic focus being the source of the infection; this could save a life. Bring in the appropriate specialties to help.
Beware of patients who are on immunosuppressive treatment; symptoms & physical signs are suppressed leading to difficulty in diagnosis. Have a high index of suspicion.
CRP is one of the better indicators of SA.
Culture blood as well as joint fluid.
Large native joint SA is more likely to have treatment failure and result in death than small native joint SA. The only SNJSA patients who died in the McBride et al, Middlemore NZ study also had LNJSA.
As far as the joint itself is concerned, the sooner the diagnosis is made, the sooner the joint is drained, the sooner the patient can be relieved of pain/disability & the lower the cost of hospitalisation.
The issue of how rapid surgical drainage/washout needs to be, for the sake of the joint, is no longer as clear cut as once it seemed.
Looking at the joint outcome endpoint, delay in diagnosis, and the timing of treatment did not have a clear association with the outcome of disease in the studies by Lauper & Kaarndorp. In both studies the median delay in treatment from onset was within one week in both the poor outcome and favourable outcome groups.
Chen’s study showed a link of outcome to delay, but the delay was often longer than 3 weeks. Those treated within the first week did well.
It seems likely that, within the first week, delay is not associated with a poorer eventual outcome. Beyond the first week, delay is increasingly associated with poor outcome; beyond 3 weeks a poor outcome is likely.
Lauper et al said:
In conclusion, septic arthritis patients arriving at the hospital usually suffer from their infection for several days. Provided that they do not have signs of clinical sepsis, and especially if they are already receiving empiric antibiotic treatment, subjecting the patient to immediate surgical drainage does not appear to prevent long-term joint-related sequelae. Interventions can safely be organized for the next day, liberating resources during busy night shifts, and enabling patients to be operated by fresh teams the following day …..
Age and pre-existing joint disease seemed associated with poor outcome in in Kaandorp’s study, but not in Lauper’s.
The risk of subsequent arthroplasty after SA is 6X that of the general population.
Significant improvement can be expected after 2 stage arthroplasty for joint destruction secondary to SA, but a higher rate of complications is expected compared to a standard arthroplasty.