SHORT COMMUNICATION

The cytospin technique improves the detection of calciumpyrophosphate crystals in synovial fluid samples with a lowleukocyte count

Christoph Robier • Franz Quehenberger •

Manfred Neubauer • Mariana Stettin •

Franz Rainer

Received: 8 December 2012 / Accepted: 28 January 2013

� Springer-Verlag Berlin Heidelberg 2013

Abstract In synovial fluids (SF) with low leukocyte or/

and crystal counts, important features may be missed, if

exclusively smears are examined by polarized microscopy.

That may be overcome by cytocentrifuges, which use low-

speed centrifugal force to concentrate cells onto a glass

slide and thus enhance the number of cells per high power

field (HPF). We compared the calcium pyrophosphate

(CPP) crystal counts in cytospin preparations with those in

common smears of SF. The number of CPP crystals was

counted in 50 SF samples by polarized microscopy, and

statistical comparisons of the mean values of the cytospin

and smear preparations were performed using the Wilco-

xon test. The reproducibility within the slides of the

cytocentrifuge and smear samples was determined by

Spearman’s rank correlation. The crystal counts were

significantly higher in the cytospin than in the smear

preparations (median 96/10 HPF vs. 2.5/10 HPF,

p \ 0.0001). The correlation in the crystal count between

the slides 1 and 2 was significantly higher within the

cytocentrifuge than in the smear group (0.97 vs. 0.73,

p = 0.0004). CPP-negative cytospin preparations in

initially smear-positive slides were not observed. We

confirmed that the cytospin technique significantly enhan-

ces the number of examinable crystals per HPF, compared

to common smears.

Keywords Calcium pyrophosphate � Synovial fluid �Cytocentrifuge � Cytospin

Introduction

The diagnosis of crystal-related arthropathies is based on

the detection of crystals in the synovial fluids (SF) of the

affected joints. Thereby, polarized microscopy of fresh SF

smears is still regarded as the gold standard by many lab-

oratories [1]. However, in several situations, such as in SF

with low leukocyte or/and crystal counts, important mor-

phological features, which are crucial for the correct

diagnosis, may be missed if exclusively smears are

examined. In particular, in calcium pyrophosphate (CPP)

crystal-related disorders [2], in which a varying and often

low concentration of CPP crystals is frequently observed

[1], the crystals may be easier overseen than monosodium

urate (MSU) crystals in patients with gout [1, 3]. This

limitation may be overcome by cytoconcentration using the

cytospin technique. Cytocentrifuges (CC) use low-speed

centrifugal force to place a monolayer of cells onto a

defined area of a glass slide and thus enable the analyst to

examine a higher number of cells per high power field

(HPF), compared to common smears. Cytocentrifuges spin

at lower speeds and show a more gradual way of acceler-

ation and deceleration than normal centrifuges. The tech-

nique provides thin-layer cell preparations from any liquid

materials, especially from hypocellular fluids such as SF,

cerebrospinal fluid, aszites, or pleural effusions. Based on

the method’s technical background, we hypothesized that

CPP crystals, which are mostly found within the cytoplasm

of SF macrophages, but also extracellular [4], should be

easier detected in cytospin preparations. The aim of this

study was to compare the CPP crystal counts in CC

C. Robier (&) � M. Neubauer � M. Stettin � F. Rainer

Central Laboratory, Department of Internal Medicine,

Hospital Barmherzige Brueder Graz-Eggenberg,

Bergstrasse 27, 8020 Graz, Austria

e-mail: christoph.robier@bbegg.at

F. Quehenberger

Institute of Medical Informatics, Statistics and Documentation,

Medical University of Graz, Graz, Austria

123

Rheumatol Int

DOI 10.1007/s00296-013-2689-0

preparations with those in common smears of SF, in order

to confirm our hypothesis and thus to improve the quality

of SF crystal diagnosis.

Materials and methods

This study was approved by the institutional ethics com-

mittee of the Hospital Barmherzige Brueder in Graz-Eg-

genberg. We used residual material of CPP crystal-positive,

consecutive SF specimens obtained by diagnostic arthro-

centesis of 50 in- and out-patients of the Department of

Internal Medicine. Smears were regarded as CPP-positive, if

they contained at least two typical rhomboid or paralleleli-

pedic CPP crystals per ten HPF at 6309 magnification.

CPP-negative slides, SF with a too small sample size, and

hemorrhagic SF were excluded from the study. The majority

of the study participants were female (68 %), with a median

age of 72 years. Most of the samples were obtained from

symptomatic knee joints and shoulders (76, and 18 %), and

the median SF leukocyte count was 320 cells/ll, repre-

senting mostly non-inflammatory conditions. According to

current radiological findings and medical recordings, the

subjects were diagnosed with osteoarthritis (OA, n = 27),

chondrocalcinosis (n = 16), spondyloarthropathies (n = 4),

and rheumatoid arthritis (n = 3). The complete clinical

characteristics of the study population are shown in Table 1.

Immediately after arthrocentesis, the samples were antico-

agulated with unfractionated heparin (SF:heparin 5,000

IU = 4:1, Heparin Immuno, Ebewe Pharma, Unterach,

Austria) and were further pretreated with 50 ll/ml hyal-

uronidase V solution (1 lg hyaluronidase lyophilisate/ml

bidestillated water, Sigma-Aldrich, St. Louis, USA). After

the first crystal detection, the preparation of the smear and

cytospin samples was performed within 30 min to avoid the

dissolution of existing crystals and the formation of new,

crystal-resembling artifacts [5, 6]. Two smears and two CC

sediments on standard glass slides were prepared from each

SF sample. SF were cytocentrifuged for 10 min at 700 rpm

in a Shandon Cytospin 4 cytocentrifuge (Thermo Fisher

Scientific, Waltham, Massachusetts, USA). The smears and

sediments were dried on the slides at room temperature

[7, 8] and were examined in blinded fashion by polarized

microscopy (Zeiss Axioscope) within 24 h by an experi-

enced analyst (CR). The number of crystals per ten HPF of

each slide was counted, and statistical comparisons of the

mean values of the CC and smear preparations were carried

out. We further studied the correlation within the two slides

of the CC and smear samples.

Statistical analysis

Statistical data analysis was performed with the package

R2.12.0. The exact Wilcoxon test was used for group

comparisons between CC and smear preparations. A

p value \0.005 was considered statistically significant.

Statistical correlation analysis was assessed with Spear-

man’s rank correlation test.

Results

The CPP crystal counts were much higher in the CC than in

the smear preparations (median 96/10 HPF vs. 2.5/10

HPF), this difference was highly statistically significant

(p \ 0.0001). Spearman’s rank correlation test revealed

that the correlation in the determined CPP crystal count

between the slides 1 and 2 was significantly higher within

the CC group than in the smear group (0.97 vs. 0.73,

p = 0.0004). We did not observe any CPP-negative CC

preparations in initially smear-positive slides.

The crystal counts and the results of the group com-

parisons are summarized in Table 2.

Discussion

Our findings confirm that the cytospin technique signifi-

cantly enhances the examinable SF crystal count per HPFTable 1 Demographic and clinical characteristics of the study

population

Study participants, n 50

Age in years* 72 (43–88)

Female sex, n (%) 34 (68)

SF leukocyte count (cells/ll)* 320 (60–21,680)

Affected joints, n (%) 50 (100)

Knee 38 (76)

Shoulder 9 (18)

Talocalcanean joint 2 (4)

Elbow 1 (2)

SF synovial fluid

* Values are expressed as median (range)

Table 2 CPP crystal counts and the results of the comparison

between cytospin and smear preparations using the exact Wilcoxon

test

CPP crystal count Cytospin* Smear* p value

Slide 1, median (range) 93 (10–2,803) 2 (2–45)

Slide 2, median (range) 97 (13–2,063) 3 (2–39)

Mean 96 (11.5–2,433) 2.5 (2–39.5) \0.0001

CPP calcium pyrophosphate

* The numbers of CPP crystals are expressed as crystals per ten high

power fields at 6309 magnification

Rheumatol Int

123

compared to common smears. Thus, especially in SF

specimens with a low crystal and/or leukocyte count, this

technique may be beneficial. For training purposes, as well

as for daily routine diagnosis by analysts with a lower grade

of experience, the CC may be of further value. The lack of

an adequate standard of SF examination for crystals among

teaching hospitals has been reported by McGill et al. [9]. In

that context, establishing a standardized procedure of pre-

analytical SF preparation, leading to a better diagnostic

performance, as it may be represented by the CC devices,

should be a future goal. Furthermore, in studies on the

morphological appearance of CPP crystals, the data collec-

tion will be much easier using CC. A prior study revealed

the evidence for a causal relationship between the structure,

size, and number of CPP crystals, and its distribution pat-

terns in various types of CPP-related disorders [10].

Overall, comparative data on the use of CC devices in

SF examinations have been rarely published. A recent

study using CC highlighted that a coincidence of CPP and

MSU crystals can be found in 7.7 % of patients with gout,

compared to a coincidence of 0.37 %, determined in

common smears in an older study [11, 12].

Most of the CC evaluations have been performed on fine

needle aspirations [13] and exfoliative cytology [14], in

order to improve the diagnostic outcome of tumor cytol-

ogy. Thereby, cytospins contributed further information

beyond that acquired from smears in 2 %, and in even

10 %, when smears were initially non-diagnostic [13].

However, based on the fact that CC are more expensive

than common smears, the authors of that study concluded

that it is not cost-effective to obtain CC in addition to

smears in all cases [13]. In our opinion, this statement is

also valid for SF analysis. In routine examinations of SF,

the CC technique should not be performed in all patients, it

should be restricted to hypocellular SF and to SF obtained

from inflamed joints which show clinical signs of crystal-

related arthropathies, but are found crystal-negative in

common smears. For the assessment of cytomorphological

features, CC preparations show a much better performance

in preserving cellular details by flattening and reducing the

overlapping of cells and can thus enable a more precise

interpretation of the leukocyte differential count [14].

In general, an enhancement of the crystal count per HPF

is more important for CPP than for MSU crystals, which

are usually found in high concentrations in gouty arthritis,

compared to the varying and often low concentration of

CPP crystals in CPP-associated disorders [1]. Several prior

studies confirmed that the results for CPP crystals were

much worse than for MSU crystals [1, 15]. It has been

clearly shown that the higher the crystal load in the SF is,

the more likely it is, that the analysts can perform a correct

detection and identification of the crystals [16]. Our

study had limitations that merit mention. First, the SF

examinations were exclusively performed by polarized

microscopy. A previous study showed that only one-fifth of

CPP crystals appears as birefringent under polarized light,

and a certain percentage of crystals may thus be missed, if

the search for CPP crystals is performed solely under

polarized light [3]. Ordinary light allows a better CPP

detection rate, but for the definitive CPP crystal identifi-

cation, polarized microscopy is still required [3]. That

potential bias may not be relevant, because on the one side,

it was apparent in both the CC and the smear group, and on

the other side, this study has not been performed to

determine the exact crystal count of the SF samples. Sec-

ond, whether there is an improvement in the crystal

detection rate due to the use of CC cannot be answered

based on our findings, because according to our study

design, crystal-negative samples were excluded. Further

studies incorporating data from comparisons between CC

and smear preparations of smear-negative SF samples are

needed to elucidate that question.

In conclusion, we have shown that the cytospin tech-

nique significantly enhances the number of examinable

crystals per HPF, compared to common smears of SF.

Based on our observations, we suggest that cytospin

devices should be used in selected clinical and scientific

questions, such as in hypocellular SF effusions, in SF with

a low crystal concentration, as well as for studies on the

morphological appearance of CPP crystals.

Conflict of interest The authors stated that there are no conflicts of

interest regarding the publication of this article.

References

1. Swan A, Amer H, Dieppe P (2002) The value of synovial fluid

assays in the diagnosis of joint disease: a literature survey. Ann

Rheum Dis 61:493–498

2. Zhang W, Doherty M, Bardin T, Barskova V, Guerne P-A, Jansen

TL, Leeb BF, Perez-Ruiz F, Pimentao J, Punzi L, Richette P,

Sivera F, Uhlig T, Watt I, Pascual E (2011) European league

against rheumatism recommendations for calcium pyrophosphate

deposition. Part I: terminology and diagnosis. Ann Rheum Dis

70:563–570

3. Ivorra J, Rosas J, Pascual E (1999) Most calcium pyrophosphate

crystals appear as non-birefringent. Ann Rheum Dis 58:582–584

4. Martinez Sanchis A, Pascual E (2005) Intracellular and extra-

cellular CPPD crystals are a regular feature in synovial fluid from

uninflamed joints of patients with CPPD related arthropathy. Ann

Rheum Dis 64:1769–1772

5. Kerolus G, Clayburne G, Schumacher HR Jr (1989) Is it man-

datory to examine synovial fluids promptly after arthrocentesis?

Arthritis Rheum 32(3):271–278

6. McGill NW, Swan A, Dieppe PA (1991) Survival of calcium

pyrophosphate crystals in stored synovial fluids. Ann Rheum Dis

50:939–941

7. Von Essen R, Holtta AMH, Pikkarainen R (1998) Quality control

of synovial fluid crystal identification. Ann Rheum Dis 57:

107–109

Rheumatol Int

123

8. Robier C, Neubauer M, Stettin M, Lunzer R, Rainer F (2012)

Dried cytospin preparations of synovial fluid are a stable material

for long-time storage and delayed crystal analysis. Clin Rheu-

matol 31:1115–1116

9. McGill NW, York HF (1991) Reproducibility of synovial fluid

examination for crystals. Aust N Z J Med 21(5):710–713

10. Swan A, Heywood B, Chapman B, Seward H, Dieppe P (1995)

Evidence for a causal relationship between the structure, size, and

load of calcium pyrophosphate dihydrate crystals, and attacks of

pseudogout. Ann Rheum Dis 54:825–830

11. Robier C, Neubauer M, Quehenberger F, Rainer F (2011)

Coincidence of calcium pyrophosphate and monosodium urate

crystals in the synovial fluid of patients with gout determined

by the cytocentrifugation technique. Ann Rheum Dis 70(6):

1163–1164

12. Jaccard YB, Gerster JC, Calame L (1996) Mixed monosodium

urate and calcium pyrophosphate crystal-induced arthropathy. A

review of 17 cases. Rev Rhum Engl Ed 63:331–335

13. Liu K, Dodge R, Glasgow BJ, Layfield LJ (1998) Fine needle

aspiration: comparison of smear, cytospin, and cell block prep-

arations in diagnostic and cost effectiveness. Diagn Cytopathol

19(1):70–74

14. Shah P, Deshmukh R (2012) Exfoliative cytology and cytocen-

trifuge preparation of oral premalignant and malignant lesions.

Acta Cytol 56(1):68–73

15. Hasselbacher P (1987) Variation in synovial fluid analyses by

hospital laboratories. Arthr Rheum 30:637–642

16. Gordon C, Swan A, Dieppe P (1989) Detection of crystals in SF

by light microscopy: sensitivity and reliability. Ann Rheum Dis

48:737–742

Rheumatol Int

123