Article Text

Download PDFPDF

Aspiration of normal or asymptomatic pathological joints for diagnosis and research: indications, technique and success rate
  1. E Pascual1,
  2. M Doherty2
  1. 1
    Facultad de Medicina, Universidad Miguel Hernández, Hospital General Universitario de Alicante, Alicante, Spain
  2. 2
    University of Nottingham, Nottingham, UK
  1. Professor M Doherty, Academic Rheumatology, Clinical Sciences Building, City Hospital, Nottingham, NG5 1PB, UK; Michael.Doherty{at}


Although joint aspiration is a basic clinical skill, aspiration of normal joints, or asymptomatic clinically quiescent joints, is only rarely undertaken. There are two main indications for this procedure. Firstly, for definitive diagnosis of crystal-associated arthritis (gout and pseudogout) during the intercritical period and for subsequent monitoring of treatment success of gout; and secondly, to obtain normal synovial fluid for biomarker research. The justification for these indications, the success rate and the technical aspects related to this procedure are presented in this article.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Joint aspiration (arthrocentesis) is considered a core skill for practicing rheumatologists and orthopaedic surgeons. It is also commonly undertaken by doctors, general practitioners and specialised allied health professionals. However, although arthritic joints, especially the knee, are commonly aspirated and injected with corticosteroid in routine clinical practice, aspiration of asymptomatic and clinically normal human joints is undertaken only very occasionally. Indeed, many experienced rheumatologists may never have attempted this.

There are two main indications for this procedure. Firstly, for the diagnosis of crystal-associated arthritis (principally gout) during the intercritical period and for subsequent monitoring of treatment success. Secondly, for research related to synovial fluid (SF) biomarkers.

This article summarises the justifications and success rates for this procedure and the technical issues relating to the aspiration of normal or clinically quiescent joints.


For diagnosis and monitoring of gout

The first images of monosodium urate (MSU) crystals were obtained by the inventor of the microscope Antonio Van Leeuwenhoek in the 17th century. However, it was the regular finding of MSU crystals in SF aspirated from joints affected by acute gout and the possibility of a simple unequivocal diagnostic test that first brought crystal detection and identification to the attention of rheumatologists in the early 1960s.1 The importance of this finding was emphasised in Wallace et al2 published in 1977, in which MSU crystal identification was proposed as the gold standard for diagnosis. Demonstration that the injection of MSU crystals into healthy canine and human joints caused the rapid onset of florid inflammation that is so characteristic of acute gout35 confirmed MSU crystals as primary pathogenic agents, but also prompted the hypothesis that acute gout results from “injection” or “shedding” into the joint cavity of preformed MSU crystals from neighbouring tissues such as cartilage.6 The corollary of this was that the presence of inflammatory crystals within the joint cavity would always cause synovitis and that SF would be free of crystals during intercritical periods. Therefore it was presumed that a definitive diagnosis of gout based on MSU crystal identification could only be made by examining SF during an acute attack, or by aspirating a tophus from which MSU crystals can be obtained even in the absence of clinical inflammation.

From 1979 onwards, however, several reports confirmed that in patients with gout MSU crystals, many of them intracellular,7 could be identified in the majority of SF samples aspirated between acute attacks from asymptomatic but previously affected first metatarsophalangeal joints (MTPJs)810 and knees.11 12 In a study of 37 patients with gout who had received no urate lowering therapy (ULT) MSU crystals were identified in 36/37 (97%) asymptomatic but previously inflamed knees, but also from 7/37 (26%) knees that had never had acute gout in that joint.13 In a further study involving intercritical aspiration of previously affected joints (80 knees, 21 first MTPJs) 100% (43/43) of joints from patients who had received no ULT had MSU crystals in their SF, whereas 71% (34/48) of joints from patients on ULT were positive; the longer the patient had received ULT and the lower their serum urate the less likely they were to have SF crystals.14 Thus, in untreated patients with gout the presence of MSU crystals in SF is highly consistent during and between attacks, but effective treatment reduces this detection rate. Maintaining the serum urate below the therapeutic target of 6 mg/dl (360 umol/litre: the saturation point for MSU crystal formation) results in eventual disappearance of SF MSU crystals.15 The longer the duration of gout prior to ULT the longer this takes to occur,16 presumably reflecting the size of the initial crystal load. Interestingly, after successful ULT, the concentration of MSU crystals in SF markedly decreases16 perhaps explaining why reduction of serum uric acid (SUA) quite quickly results in a decrease in the number of gouty attacks.17 18

It is noteworthy that SF from an asymptomatic knee that contains MSU crystals shows a higher total cell count and higher percentage of polymorphs than SF from knees of patients with gout that are free of crystals.13 In crystal positive asymptomatic joints about one in every five cells (almost all large mononuclear cells, but also occasional polymorphs) contains intracellular MSU crystals.7 These findings suggest that MSU crystals do cause subclinical inflammation. The modestly raised cell count and proportion of polymorphs are both reduced by colchicine;19 prophylactic doses of this drug may reduce acute attacks by diminishing this subclinical inflammation and stabilising the joint environment.

Because of the above data, the possibility of reaching an unequivocal diagnosis of gout by analysing SF samples drawn form asymptomatic joints is now included in the EULAR evidence-based recommendations for the diagnosis of gout.20 Especially for clinical studies, serial aspiration of asymptomatic joints until crystal disappearance is observed may also be considered as a means of determining successful elimination of MSU crystals,16 that is, “curative” treatment.21 22

For diagnosis of calcium pyrophosphate-associated arthritis

With calcium pyrophosphate dihydrate (CPPD) crystal-associated arthritis the diagnostic value of crystal identification in SF obtained from inflamed joints was apparent soon after recognition of the disease.23 As with gout, injection of CPPD crystals into canine and human joints resulted in acute synovitis,24 and crystal shedding into the joint space from the fibrocartilage and hyaline cartilage in which CPPD crystals form was proposed as the mechanism for acute “pseudogout”.25 As with MSU it was presumed that SF CPPD crystals would only be detected during inflammatory episodes, but it was soon apparent that CPPD crystals are regularly found in SF from uninflamed and even asymptomatic knees affected by CPPD arthropathy (chondrocalcinosis plus structural changes of osteoarthritis (OA)); these crystals are frequently intracellular and associate with modest elevation of SF cell counts.26 Contrary to gout, there is no treatment to eliminate CPPD crystals from joints so the use of sequential SF aspiration to monitor changes in concentration of CPPD has not been studied formally. Although radiographic chondrocalcinosis is most commonly due to CPPD it may also result from deposition of basic calcium phosphates. Furthermore, radiographs are relatively insensitive at detecting cartilage calcification so chondrocalcinosis may be absent despite CPPD presence in a joint, especially with modest crystal loads or when there is significant cartilage loss, or in complex joints such as the tarsus. Therefore as with gout, confirmation of CPPD in SF aspirated during acute synovitis or after the event is the gold standard for diagnosis of pseudogout and for chronic CPPD-associated arthropathy.

For research purposes

There is considerable research interest in measuring biomarkers of joint disease, especially OA. For eventual clinical purposes it would, of course, be most convenient to have a marker that can be measured in a blood or urine sample. However, there are potential disadvantages in quantifying biomarkers in these compartments, including: the contribution to blood and urine levels from non-articular tissues and from joints other than the index joint(s) of interest; hepatic and lymphatic clearance of the marker once it leaves the joint; and the influence of reduced renal function on urinary and blood levels. Therefore, although less applicable generally, synovial fluid (SF) studies are still undertaken in an attempt to get closer to the joint of interest to identify potential biomarkers and to gain better understanding of disease.

SF studies inevitably focus on the knee because: (a) it is a target site for almost all arthropathies; (b) it is the largest synovial joint in the body; and (c) it is the easiest joint to aspirate. However, relatively few SF studies compare SF from diseased joints to SF from normal joints. Many studies that focus on rheumatoid and other inflammatory arthritis include SF from OA joints as a surrogate for normal (“non-inflammatory” disease) controls. Even then, the OA knees that are studied are those with clinically detectable effusions resulting in a selection bias towards the more “inflammatory” end of the OA spectrum. However, the additional inclusion of normal knee SF data can greatly alter the interpretation of results obtained in rheumatoid arthritis and OA knees. For example, in an experiment measuring SF inorganic pyrophosphate levels,27 the mean concentrations in OA knees were significantly higher than in knees affected by rheumatoid arthritis. Given that both groups comprise individuals with abnormal arthritic knees, without knowledge of normal SF levels of pyrophosphate it might be assumed that both levels are abnormally raised, but to varying degrees. However, results obtained in normal knees showed that pyrophosphate levels are elevated in OA knees but are subnormal in rheumatoid knees. This is a very different conclusion, but one that fits well with the observed increased prevalence of CPPD crystal deposition in OA knees and the lower than expected prevalence of chondrocalcinosis in rheumatoid arthritis knees compared to the general population.28

The main reasons not to aspirate normal knees for research are probably the inconvenience of identifying normal SF donors compared to aspiration of patients, the reluctance to ask research ethics committees to sanction joint aspiration in normal volunteers and the presumption that normal knees will provide only “dry taps”. Certainly, there is added inconvenience in identifying normal participants, but it is important to realise that aspiration of a knee by an experienced practitioner is no bigger a procedure, and should be no more painful or dangerous, than venepuncture. Therefore, assuming a research study is well justified and designed, the inclusion of normal knee aspiration should carry no special ethical concerns. Furthermore, the ability to obtain SF from over 80% of normal adult knees (see below) is probably higher than general expectation and although volumes may be low they are certainly sufficient for many current assay systems. Therefore, we would suggest that inclusion of normal knee SF should be more commonly undertaken in SF studies.

Possibly the largest group of normal human knees aspirated for a single research project was that recently undertaken in Nottingham, UK. As part of an OA biomarker study (unpublished), 216 people aged 45–80 with normal knees volunteered to undergo bilateral knee aspiration. These participants had no knee symptoms, no abnormal physical signs on knee examination and no radiographic features of tibiofemoral OA on weight-bearing semiflexed knee radiographs (24 had occult mild patellofemoral changes on x ray but were still included). The local research ethics committee approved the study and a single experienced clinician (MD) undertook all aspirations using a medial approach with no ultrasound guidance. A minimum volume of 0.5 ml of SF (the predetermined minimum for this study) was obtained from 180 (83%) of the 216 participants. SF was obtained from both knees in 84 (39%), from just the right knee in 71 (32%) and from just the left knee in 24 (11%). The SF volumes obtained from the 263 knees of these 180 participants ranged from 0.5 ml to 6.5 ml, with a median of 1.0 ml. Although not undertaken in this study, providing that both knees are normal the SF from each knee could be combined to make up a larger volume for analyses that require this.

Interestingly, not only was there a greater success rate with right knees (155 positive right vs 108 positive left knees) but the volume obtained from right knees also tended to be higher (median 1.5, range 0.5–6.5 vs median 0.5, range 0.5–4.5 for left knees). This may be of relevance in that knee OA also shows right-sided predominance.29 Only 24 participants had any radiographic evidence of OA, and this was only very mild, predominantly bilateral (20/24) and limited to the patellofemoral compartment. This suggests that the physiology of right and left knees may differ, perhaps as an adaptation to different biomechanical forces. An alternative explanation is that there was some systematic difference in the right-handed doctor’s technique, such as the angle of entry of the needle that favoured successful aspiration from right knees. This, however, would not explain the difference in the SF volumes obtained.

When considering smaller volumes for crystal identification (ie, just a drop or two) success rates for aspiration of asymptomatic knees and first MTP joints of patients with gout are even higher: for example, 73/80 (91%) of knees15 and 30/33 (91%) first MTP joints.30

Aspiration of asymptomatic joints of patients with other rheumatic disease may also be considered for research purposes. Asymptomatic joints of patients with rheumatoid arthritis (RA) can have occult inflammation as evidenced by synovial biopsy data31 or by sonography,32 and in the SF study of Pascual,33 analysis of fluids from 30 asymptomatic RA knees showed a mean (SD) cell count of 385 (638) (range: 20–3040) cells/μl while SF from normal knees had a mean (SD) cell count of 35 (17) (range:10–85) cells/μl. In the same report, SF from 14 asymptomatic systemic lupus erythaematosus (SLE) joints showed cell counts of 385 (503) (range: 30–1930) cells/μl.33 RA and SLE SF samples both contained a small percentage of polymorphs. All these data provide further evidence that in asymptomatic joints from at least some patients with RA and SLE there is subclinical inflammation, and that to study this arthrocentesis to collect SF is a feasible procedure.


The technique of arthrocentesis of asymptomatic knee and first MTP joints is simple and with exceptions can be performed with only modest discomfort for the patient. Asymptomatic but minimally inflamed joints usually contain increased volumes of SF, making a fruitful arthrocentesis even easier. The following recommendations are based solely on the personal experience of the authors. To our knowledge, comparative techniques have not been formally studied and the possible added value of ultrasound guidance to increase success rates, which are already high, has yet to be explored. As with routine arthrocentesis, the recommended setting (patient relaxed on a reclining couch, fully informed of the procedure) and use of aseptic technique and sterile equipment are essential.34 The operator should wear gloves, clean the overlying skin with an alcohol swab, use sterile equipment and apply a no touch technique. Prior to aspiration local anaesthetic can be applied to the skin using a spray or ointment rather than local infiltration (which itself is uncomfortable), although this is optional. The only contraindication is locally infected or broken skin; if the patient is on anticoagulants more prolonged pressure should be applied following aspiration. Possible complications are introduction of sepsis (very small if an aseptic technique is used) and induction of a post aspiration flare (no cases in the combined experience of the authors).


A narrow-barrelled tuberculin-like syringe (1 ml) or normal 1 or 2 ml syringe is easy to handle and, unlike larger volume syringes, produces only a modest vacuum that may reduce the likelihood of the synovial membrane being sucked into the bevel and occluding the needle. A 1–1.25 inch long 23 or 25 gauge (blue or orange) needle is preferred over a 21-gauge (green) needle, although the longer length of the 21-gauge needle may be required for knees of larger, more overweight individuals. Aspiration is undertaken via a medial approach (fig 1). This can be at the level of the proximal point of the patella, pointing downwards at approximately 30° and medially, to place the needle below the centre of the patella. Alternatively a more transverse approach can be made below the midpoint of the patella at the site of the medial “dimple” or gutter, again aiming to place the needle under the centre of the patella. Before piercing the skin move the patella from side to side to locate the landmarks and look for a “bulge” sign in the medial dimple, which may be present and help direct the site of aspiration. For a bulge sign stabilise the patella with one hand and massage in turn the medial and lateral dimples to move a small amount of fluid transversely from one dimple to another.

Figure 1 Aspiration undertaken via a medial approach.

As the needle is inserted a slight give is sometimes felt as the needle goes through the capsule. With the needle tip below the patella gently aspirate. If no SF is obtained (look particularly for filling of the semitransparent base of the needle) the needle can be slowly progressed a little, or withdrawn a little, while still applying gentle suction. Gentle massage of the lateral and medial aspects of the knee may encourage SF to move behind the patella and enter the barrel. Interestingly, SF sometimes enters the barrel in slow waves or pulses so it is worth being patient once the initial SF has entered the syringe. It is very important that the patient’s quadriceps are fully relaxed; conversing with the patient while the procedure is ongoing often helps.

First MTP joint

This joint can be entered medially on the superior aspect, using a 25-gauge (orange) needle and small syringe. If available, a small 0.5 ml syringe with a 29-gauge needle is ideal.30 To identify the joint line, distract and slightly flex the toe (fig 2). This opens up the joint line making it more readily palpable and often visible. Insert the needle into the upper side of the joint line below the extensor tendon. If the joint line does not easily open up palpate for the distal metaphysis of the first metatarsal, best felt with the joint slightly flexed, and insert the needle just distal to that landmark, repositioning it slightly if bony resistance is encountered. It is advisable to aspirate as soon as the needle has penetrated the skin since some SF is often superficially positioned at the superolateral aspect of the joint. A SF sample was obtained from 18 out of 21 (86%)14 and 30 out of 33 joints aspirated, but even in apparently negative aspirations SF can often be obtained by aiming the needle over the top of the joint beneath the extensor tendon.

Figure 2 Identification of the joint line.

SF analysis from asymptomatic joints

Even the smallest amount of SF usually contains sufficient crystals for diagnosis. Even with an apparently dry tap it is worth removing the syringe barrel, filling it with air, reconnecting it to the needle and injecting the air through it over a slide. This often empties just a very small spray of fluid that is contained within the needle onto the slide. At least for gout, SF from asymptomatic joints often contains a reasonably high crystal concentration. In a series by Pascual et al of 77 SF samples containing MSU crystals the crystals were identified in the first field that was examined (at 400× magnification) in 47 cases (61%).14 We advise examination of up to 30 microscope fields at 400×: if no crystals are seen, finding them is unlikely (although not impossible) by a contiued search. Use of a clean, well serviced polarised microscope and a two-step procedure are advised.35 The first step, to detect crystals, is carried out initially under ordinary light (400× magnification) to identify crystals (often intracellular) by their symmetrical, sharp line morphology and then under uncompensated polarised light to confirm birefringence. In the second crystal identification step, compensated polarised light is used to further distinguish MSU from CPPD. The combination of morphology and light characteristics permits distinction between MSU and CPPD: MSU are large and needle shaped, with strong (negative sign) birefringence, whereas CPPD are small, rhomboid or blunt needle shaped, with weak (positive sign) or absent birefringence.


Aspiration of normal or clinically quiescent knees and first MTP joints is a simple, safe procedure that provides sufficient SF for crystal identification in the vast majority of cases. It offers an elegant approach to diagnosis of gout and CPPD-associated arthropathy during an asymptomatic intercritical period. Aspiration of normal knees should also be considered for research purposes to provide useful true control SF data, given that sufficient SF for many analyses is obtained from over 80% of normal adult knees.



  • Competing interests: None declared.

  • Ethics approval: Ethics approval was obtained.