Background/Purpose: To better understand the variation in calcium pyrophosphate (CPP) crystal, size, shape and appearance under compensated polarized microscope, synovial fluid crystals were examined by 4 independent raters.

Methods: Over a 4-month period, synovial fluid samples were de-identified and collected from the university clinical laboratory when found to be positive for CPP crystals. All samples were centrifuged for 5 minutes (prior to clinical lab examination) and stored at 4°C. For this research effort, slides were dry mounted for enhanced stability.

Two experienced raters and two novice raters were trained using an atlas with sample CPP crystal images and a standardized protocol for describing each crystal’s shape, birefringence strength, color, and the raters’ certainty that the object was a CPP crystal. Raters examined each slide until a minimum of 10 CPP suspected crystals were identified and photographed with 100x objective, or a maximum of 30 minutes was reached.

For the purpose of crystal measurement, Adobe Photoshop CS6 was used to apply a high pass linear light filter to digital images to provide enhanced resolution and line discrimination. Crystal area (μm²), length (μm), and width (μm) (for rods), and diagonals (μm) (for rhomboids) were measured. Crystals rated with low certainty (1 on 1-3 scale) or crystals that had atypical shape or uncharacteristically bright birefringence were excluded. All remaining crystals were reviewed by a national CPP crystal expert (AR) for final inclusion in the dataset.

Results: Synovial fluid from 16 joint aspirates was reviewed by 4 raters. Of the 564 potential crystals identified by the 4 raters, the expert rater confirmed 293 (52%) were definite CPP crystals. Crystals that were rejected were more likely to be smaller, rhomboid, yellow rather than blue, and have weaker or no birefringence. After the brief training session, there were few differences in either the number of definite crystals or accuracy of crystal identification between the more experienced and lesser experienced raters.

Of the 293 definite crystals, 185 (63%) were categorized as rods with median area of 3.8 mm², (range, 1.0 – 22.9 mm²), and 108 rhomboids (37%) with median area of 5.8 mm², (range, 0.9 – 27.0 mm²), p <0.005. The expected frequency of color distribution ought to be 50:50; however, 60% of confirmed crystals were blue (suggesting that the yellow-orange crystals were not as easily identified). Compared to rhomboids, rods were more likely to be scored as having low birefringence (35% vs. 9%). Only 4 (1.4%) non-birefringent confirmed crystals were identified. Rod median length was 3.8 mm (range, 1.0 – 14.3 mm), median length to width ratio 3.4 (range, 1.1 – 13). Rhomboid median (maximum) diagonal length, 3.6 mm (range, 1.0 – 9.0 mm), median (minimum) diagonal length, 2.6 mm (0.9 – 7.0 mm) and median acute angle of 74^o (range, 41^o to 90^o).

Conclusion: After a brief training session, less experienced raters performed similarly to more experienced raters. CPPD crystals that are smaller, show weaker birefringent or are aligned perpendicular to the short-axis of polarization are harder to identify. There is significant variation in crystal size, shape and appearance under polarized microscopy.

« Back to 2017 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/calcium-pyrophosphate-crystal-size-shape-and-appearance-variability/