Can Diagnostic and Therapeutic Arthrocentesis Be Successfully Performed in the Flexed Knee?

Sabeen Yaqub, MD; Wilmer L. Sibbitt, Jr, MD; Philip A. Band, PhD; James F. Bennett, MD; N. Suzanne Emil, MD; Monthida Fangtham, MD; Roderick A. Fields, MD; William A. Hayward, PhD; Scarlett K. Kettwich, MS; Luis P. Roldan, BS; Arthur D. Bankhurst, MD


J Clin Rheumatol. 2018;24(6):295-301. 

In This Article


As part of an institutional safety and quality improvement program to improve the success of arthrocentesis and decrease the exposure of the operator's hands to potential needlestick, we recorded baseline quality and outcomes of arthrocentesis in 20 consecutive clinically effusive knees using the standard extended knee superolateral portal (Figure 1), then introduced the quality intervention of mechanical compression of the superior knee in the next 35 consecutive effusive knees in the flexed knee position and again determined quality and outcome measures. The intervention was the use of a compressive knee brace to exert a constant circumferential mechanical compression on the superior knee (suprapatellar bursa and patellofemoral joint) during the intra-articular procedure, thereby forcing synovial fluid from the patellofemoral joint and suprapatellar bursa to the inferior knee (synovial reflections of the cruciate ligaments and femoral condyles), where it could be accessed by needle (Figs. 2–6). This study confirms the findings of Zhang et al.[18] that the standard extended knee superolateral approach to arthrocentesis is superior to the flexed knee approach and also demonstrates that mechanical compression of the superior knee mobilizes and displaces synovial fluid to the inferior knee, so that the extended and flexed knee arthrocentesis approaches become equivalent in terms of diagnostic and therapeutic arthrocentesis success and absolute fluid yield (Figure 7).[18]

The flexed knee positioning for needle procedures has been reported to be less painful than the extended knee positioning because the anterolateral portal is quite large, and the needle is less likely to contact the quadriceps tendon or the periosteum of the patella or femur during needle positioning.[13–15] However, in the present study, there was no significance difference in procedural pain between the extended and flexed knee groups, similar to previous comparisons of the flexed and extended knee positioning for knee injections.[12] This disparity with some prior reports may be related to the use of preprocedural lidocaine that might be providing immediate anesthesia that would tend to converge procedural pain in both groups regardless of technique.[13–15,30]

For standard arthrocentesis, experts recommend that the patient be supine with the knee extended, and the needle then enters the skin 1 to 2 cm superior to the superior third of the patella into the lateral recess of the suprapatellar bursa, and if there is no fluid return, the needle is directed into the patellofemoral joint toward the intracondylar notch.[1–3,10–26] Chen et al.[21] and Hirsch et al.[25] have reported that synovial fluid accumulates in the lateral suprapatellar bursa; thus, the lateral approach has advantages because of natural synovial fluid pooling. Jackson et al.[1] have demonstrated that the lateral approach is also routinely more accurate than medial approaches to place the needle into the intra-articular space even in a noneffusive knee. As noted by Roberts[9] and Roberts et al.,[10] when using the medial approach for arthrocentesis, the needle may be obstructed by medial plica, the medial triangular fat pad, cumulative particulate matter (gelatinous masses) that pools on the medial side, compressed dense hypertrophied synovial tissues, or fat-replaced soft tissues (lipoma arborescens) of the medial knee compartment. Consequently, extended knee position is used to maximize pooling in the lateral suprapatellar bursa.[8–10]

Arthrocentesis success can often be improved by compressing the opposing side of the knee using the operator's or assistant's hands to force fluid toward one side or the other or into patellofemoral joint where the fluid can be more predictably aspirated (Figure 1).[8–10,17–25] However, as a single operator applies pressure to the knee with 1 hand, the syringe must consequently be operated 1 handed. When a second operator compresses the knee, there is the increased expense associated with the second operator. Further, it is difficult to compress 360 degrees around the knee even with 2 hands, and bringing hands into the operative field increases the chance of a contaminated needlestick and blood-borne infection and thus is preferably avoided.[27]

In 2008, Meehan[31] presented the idea of an external pneumatic compression device to assist with arthrocentesis. Recently, Bhavsar et al.,[26] using a nonpneumatic system, have demonstrated that mechanical compression of the knee markedly accentuates arthrocentesis success while decreasing the risk of needlestick by removing the operator's hands from the procedural site. Unlike these previous reports that used external pressure to the medial and inferior knee, the present study utilized a mechanical knee brace that provided constant circumferential compression to the superior knee (suprapatellar bursa and patellofemoral joint) via mechanical compression that forces joint fluid to the inferior knee and toward a predefined inferior knee portal (anterolateral or anteromedial portal), where the fluid could be facilely accessed (Figs. 2–6). As this study demonstrates, constant compression by a constrictive mechanical brace mobilized fluid from the superior knee (suprapatellar bursa and patellofemoral joint) to the inferior knee, making the extended and flexed knees equivalent in arthrocentesis success and fluid yield (Figs. 5–7). Thus, this study demonstrates that the flexed knee with constant mechanical compression is a viable alternative technique to the standard superolateral extended knee arthrocentesis approach.

We observed certain subtleties to the mechanical compression-assisted flexed knee technique that deserve mention. First, even though constant mechanical compression of the suprapatellar bursa and patellofemoral joint forced fluid from the superior knee to the inferior knee where it could be accessed, the layering of the fluid over the femoral condyles and cruciate ligaments was quite thin dimensionally (Figs. 3–5) because of the limited compressibility of the overlying fat pad; thus, positioning of the needle tip and bevel depth-wise was important to maintain synovial fluid return compared with the typically larger dimensions of a grossly dilated lateral suprapatellar bursa in an extended knee (Figs. 1–5), where needle bevel positioning is correspondingly less important. Another aspect of the compression-assisted flexed knee technique is that a 2-inch (5.1-cm) needle (depending on knee subcutaneous and skeletal dimensions) is necessary for this technique to be predictably accurate so that the needle tip can predictably and effectively access the synovial space of medial femoral condyle or the cruciate ligaments. Because the 2-inch needle was hubbed or nearly hubbed against the skin in many individuals in order to obtain fluid return in the flexed knee, the use of the more common 1.5-inch (3.8-cm) needle would likely substantially reduce the arthrocentesis success in the flexed knee because the needle is too short to reach the synovial space overlying the femoral condyle and cruciate ligaments and would be useful only in smaller individuals.[12] Finally, for patients who have a tendency to experience syncope or have procedural hypotension and thus cannot be in the sitting position preprocedurally, the patient can be placed in the supine position on the table, and the hip flexed to 30 to 45 degrees and knee flexed to 90 degrees with the plantar foot on the table, mimicking the sitting position for the target flexed knee. The nontarget leg can also be flexed this way or fully extended on the table. This supine flexed knee positioning creates a stable flexed knee arthrocentesis position with the patient supine, and consequently, we have not observed syncope or procedural hypotension in this position.

After the compressive mechanical brace was placed on the superior knee, we found that the additional fluid return was not always immediate, but in many cases required 30 seconds to 3 minutes to permit movement and pooling of synovial fluid in the inferior knee before the additional fluid could be aspirated. Thus, during the procedure, it was also useful to wait 30 to 60 seconds after fluid flow apparently ceased to try again, and often additional fluid had since moved to the access point, and fluid yield continued at this point. This delay in fluid movement after constant compression is applied may be the result of the semisolid gel-like nature of synovial fluid, and its non-Newtonian viscosity and elasticity that tend to make the fluid move slowly and incompletely.[32–35] Because of synovial fluid's rheological properties that resist abrupt forces such as a solid, but deform and flow like a fluid with constant force, a constant low circumferential pressure is best suited to force the viscoelastic synovial fluid to move from one area of the joint to another. Hand squeezing or "milking" the joint creates sharp compressive forces that increase the elastic properties of synovial fluid over its viscous properties and causes increased resistance to flow.[34] Because of this, we recommend placing the compression brace on the flexed knee for several minutes to allow full synovial fluid pooling before attempting arthrocentesis, and during the procedure, be patient to permit additional synovial fluid to flow to the access point during the arthrocentesis procedure itself.

There are a number of limitations to this study. First, only the lateral approaches to flexed knee arthrocentesis were explored; it is possible that medial approaches might be just as successful with mechanical compression. Although the present study used an elastomeric constrictive mechanical brace, it would be anticipated that a pneumatic or other compressive brace placed similarly on the superior knee would have the same basic effect of fluid shift to the inferior knee where the fluid could then be accessed. Another potential limitation of our study is that we did not evaluate whether the fluid extracted with compression is identical to the fluid obtained during the standard arthrocentesis procedure, but it likely is. A final limitation to this study is the sequential rather than randomized study design, and this potentially could be a cause of consistent bias. However, a sequential study design, where the baseline is determined in consecutive patients, an intervention is introduced, and changes in outcome are determined in the postintervention consecutive patients, is a classic structure for quality improvement interventions and is a standard vehicle for change in all hospitals.