Risk Stratification

The diagnosis of periprosthetic infection can be difficult to make and many tests are available to the clinician. Identification of a periprosthetic joint infection is important, as subsequent treatment is fundamentally different between cases that have a septic as opposed to an aseptic mode of failure. Specifically, a missed diagnosis of infection can lead to a high rate of subsequent failure if specific treatment for the infection is not considered.
 
Pre-test probability is weighted heavily in the performance of any diagnostic test. Thus the identification of a patient as having a higher or lower probability of periprosthetic joint infection, when initiating the diagnostic evaluation is important. Further, a determination of how likely or unlikely a diagnosis is, based on patient symptoms, risk factors, co-morbidities and the physical examination is typically an integral part of the evaluation that guides clinical decisions. Based on these factors, despite the paucity of high-level data that supports the specific identification of such factors as they relate to the probability of periprosthetic joint infection, it is important to outline factors that can assist the clinician in identifying patients who have a higher probability of periprosthetic infection and thus warrant a more extensive diagnostic evaluation. Similarly, in patients who are deemed to have a lower probability of periprosthetic infection, a less vigorous evaluation may be warranted. Clinicians oftentimes make such assessments of probability routinely in their practice that leads to more prudent selections among testing modalities that are based on specific individual patient characteristics. This routine assessment is of limited cost and low morbidity to the patient.
 
While the list compiled by the work group is not meant to be exhaustive, several risk factors were identified as supported by the available evidence as associated with the presence of periprosthetic joint infection as outlined below. Additional factors, not supported by the available literature, but unanimously agreed upon by the work group as additional risk factors to consider were also identified. Finally, several potential risk factors were identified that the literature did not support as being associated with a greater probability of periprosthetic infection including smoking status, obesity for patients who have undergone total knee arthroplasty, the use of drains, and immunocompromising states for patients who have undergone a total hip arthroplasty. Analysis of the literature revealed that prolonged post-operative wound drainage and hematoma formation appeared to increase the likelihood of periprosthetic joint infection on univariate analysis; however, this was not confirmed by multivariate analysis.
 
The work group proposes the following table for stratifying patients into higher or lower probability of infection:
 

Higher Probability of Infection	One or more symptoms, AND at least one or more: 1) risk factor;* OR 2) physical exam finding; OR 3) early implant loosening/osteolysis (as detected by x-ray)
Lower Probability of Infection	Pain or joint stiffness only and none of the following: 1) risk factors;* OR 2) physical exam findings; OR 3) early implant loosening/osteolysis (as detected by x-ray)

       *risk factor supported by evidence or expert opinion
 
Factors for Risk Stratification

Symptoms	Risk Factors Supported by Evidence	Risk Factors Supported by Consensus	Physical Exam Findings	Other
Pain in the replaced joint Stiffness in the replaced joint	Prior infection of the joint (knee) Superficial surgical site infection (hip and knee) Obesity (hip) Extended operative time (>2.5 hours, hip and knee) Immunosuppression* (knee)	Any recent (<1yr.) bacteremia or candidemia Metachronous prosthetic joint infection Skin disorders (psoriasis, chronic cellulitis, lymphedema, chronic venous stasis, skin ulcers) IV drug use Recent (<3 yrs.) MRSA infection or colonization Active infection at other site	Warmth, effusion, redness, swelling Sinus tract associated with the joint	Early (<5 yrs.) implant loosening or osteolysis (as detected by x-ray)

 
Factors not supported as a risk by evidence

1. Smoking
2. Obesity (knee)
3. Use of drains
4. Hematoma or use of anticoagulation (INR>2 or low molecular weight heparins)
5. Immunosuppression* (hip)

 
*The systematic review considered the following states as indicative of immunosuppression: HIV, diabetes, hepatitis, chemotherapy or other suppressive medication such as antimonoclonal antibodies (medications specified in search: prednisone, infliximab, adalimumab, methotrexate, etanercept), autoimmune diseases (lupus, rheumatoid arthritis, ankylosing spondylitis, Reiter’s syndrome, psoriatic arthropathy), inflammatory arthritis, renal disease (chronic renal failure, dialysis), liver failure, malnourishment, sickle cell disease, hemophilia, solid organ transplant