Contemporary issues relating to calcium pyrophosphate deposition disease

Abstract

Calcium pyrophosphate deposition (CPPD) disease is a common crystalline arthritis that has historically been under-studied compared to other rheumatic diseases including gout. For decades, the presence of CPPD in joints has been consider of uncertain clinical importance, with conflicting reports on whether CPPD leads to osteoarthritis progression or other adverse outcomes. A growing body of literature in the past five years has challenged the prevailing view of CPPD disease as “just chondrocalcinosis.” These findings have arrived together with recent developments in classifying CPPD disease for research studies, advances in quantification of CPPD on imaging studies, consensus on outcomes to be measured in CPPD disease clinical trials, newly-identified genetic associations, and updates in the nomenclature for this crystalline arthritis. This review summarizes key updates in each of these realms.

Introduction

The trickiest question around calcium pyrophosphate deposition (CPPD) is generally not whether crystals are present, but what their presence means. Why do calcium pyrophosphate crystals cause severe acute inflammatory arthritis in some joints, co-exist with or worsen osteoarthritis in other joints, trigger chronic inflammation in other joints, and cause no symptoms in yet other joints? The answers may lie in factors beyond the crystal itself, including cartilage- or synovium-specific factors, immune cell responses, or genetic polymorphisms causing increased crystal burden. This review covers updates in CPPD disease classification and imaging, outcome measurement in CPPD disease clinical research, genetics of disease and long-term outcomes, and considerations for targeted treatments.

1. Nomenclature: a brief update

A number of publications highlight challenges in CPPD disease nomenclature.^1–3 The Gout, Hyperuricemia and Crystal Arthritis Network (G-CAN) recently published updated consensus-based terminology for CPPD disease for use in clinical care and research.⁴ The G-CAN terminology addresses the many clinical manifestations of CPPD disease, including acute calcium pyrophosphate (CPP) crystal arthritis, chronic CPP crystal arthritis, and osteoarthritis (OA) with CPPD. The next step will be development of lay terminology for CPPD disease. Patient-friendly language that is short and simple is critical for increasing patient and public awareness of, and understanding of, this common crystalline arthritis.

2. CPPD disease clinical diagnosis and classification for research: related, but different

Clinical diagnosis of CPPD disease is important to patients and clinicians; a diagnosis provides an explanation for symptoms and guides treatment choices. The only diagnostic criteria for CPPD disease, referred to as Ryan and McCarty’s criteria,⁵ were published over 50 years ago and were not validated in a larger population, but they continue to provide a guiding light for CPPD disease diagnosis. In that frameworks, CPPD disease is diagnosed in the presence of joint symptoms, plus CPP crystals in synovial fluid (especially the more inflammatory monoclinic crystals⁶), plus radiographic evidence of CPPD, e.g. chondrocalcinosis. Most rheumatologists would agree that the presence of all three of these findings is classic for CPPD disease.

However, as most rheumatologists would also agree, CPP crystals can be quite tricky to identify using compensated polarized light microscopy. In the decades since the Ryan and McCarty diagnostic criteria were published, ultrasound has become a quite important tool in CPPD assessment, and computed tomography (CT) and to a lesser extent dual-energy CT have also been increasingly used for CPPD evaluation (see Section 3). CPPD disease diagnosis currently centers around joint symptoms, synovial fluid crystal analysis (when joint aspiration is performed and when CPP crystals are observed), and presence of CPPD on conventional radiograph (chondrocalcinosis), ultrasound, CT, or dual-energy CT. Still, there is nuance: what if an individual has minimal knee pain and chondrocalcinosis on knee x-ray, but has significant wrist pain and swelling with no CPPD on x-ray or ultrasound (which is more sensitive for CPPD) of the wrist? Are the wrist symptoms attributable to CPPD disease? This writer prefers evidence of CPPD in the symptomatic joint to attribute symptoms to CPPD disease, even if the patient has CPPD in other anatomic sites.

Classification of CPPD disease is a bit more straightforward, as the validated ACR/EULAR 2023 CPPD Disease classification criteria provide a structured framework for determining if an individual fulfills a standardized definition for research studies.^7,8 Presence of synovial fluid CPP crystals in a symptomatic joint is sufficient for classification as CPPD disease, provided that the joint symptoms are not more likely due to another known condition such as seropositive rheumatoid arthritis. When sufficient criteria are not fulfilled, a scoring system provides points toward (or against) classification as CPPD disease based on clinical and imaging features including age at onset, presence of rare metabolic conditions linked with CPPD disease, inflammatory arthritis symptoms and the specific joints involved, osteoarthritis in particular hand/wrist joints, and number of peripheral joints with CPPD on imaging studies. Many patients with a clinical diagnosis of CPPD disease will fulfill classification criteria (which have an excellent sensitivity of 99%).

However, individuals with a high burden of CPPD on imaging and no inflammatory arthritis symptoms may not fulfill the classification criteria. For example, an individual with CPPD in the bilateral knees, bilateral wrists, and advanced osteoarthritis of the knees with lateral patellofemoral joint space narrowing may be clinically diagnosed with CPPD disease but would not meet classification criteria. This is an important group of patients that may still benefit from treatment for CPPD disease – particularly if therapies to prevent or reduce CPP crystals are developed – but may not be included in research studies of such potential therapies because they are not classified has having CPPD disease.

3. CPPD scoring systems and artificial intelligence (AI) to identify CPPD on imaging

Several recent reviews have covered updates in imaging advances in CPPD disease.^9–11 The EULAR 2023 recommendations for imaging in the diagnosis of crystalline arthritis include a number of recommendations specific to CPPD disease: conventional radiography and ultrasound are recommended for diagnostic assessment of CPPD (or CT in the case of axial involvement).¹² Additionally, EULAR suggests imaging the knee and wrist, in addition to imaging other symptomatic joints, given that CPPD often affects knee and wrist. Serial imaging in CPPD is not recommended unless there is an unexpected change in symptoms. An overarching principle that is particularly relevant to CPPD disease is that presence of crystal deposition may not always be related to clinical manifestations.¹²

Other notable updates in the past decade including the development and validation of ultrasound definitions by the OMERACT Ultrasound CPPD Subtask Force,^13,14 development of consensus-based definitions for CPPD on conventional radiography, CT and dual-energy CT,¹⁵ and demonstration of high diagnostic accuracy of the conventional radiography definition of CPPD.¹⁶ Novel imaging modalities including multi-energy spectral photon-counting CT have also demonstrated excellent accuracy for CPPD detection in several small studies.^17,18

The emergence of scoring systems to quantify CPPD is an exciting development in the research landscape. Validated scoring systems using ultrasound and CT – two of the most sensitive imaging techniques for CPPD identification – were recently published.^19,20 The ultrasound-based scoring system, developed by the OMERACT Ultrasound CPPD Subtask Force, is validated in both the knee and wrist, whereas the Boston University Calcium Knee Score (BUCKS) CT-based scoring system is validated only in the knee. The ultrasound score takes into account the number and size of deposits in hyaline cartilage and fibrocartilage structures.¹⁹ The CT score grades the size of deposits in hyaline cartilage and fibrocartilage, and additional scores the presence of deposits in synovial capsule and ligaments.²⁰ Conventional radiography-based scoring systems have also been developed but are not yet validated; one system grades the percentage of area of the medial and lateral tibiofemoral compartments with CPPD,²¹ while the other (presented as a conference abstract) used artificial intelligence (AI) to score the number and size of deposits using the parameters of the OMERACT ultrasound score.²² Validation of conventional radiography scoring systems will be an important step forward for measuring outcomes in CPPD disease clinical trials (see Section 6).

AI is increasingly used in medicine, and its ability to identify CPPD on imaging has been the topic of several recent studies, though the field remains relatively nascent. Deep learning methods have been applied to hand radiographs to identify CPPD at the triangulofibrocartilage complex and second and third metacarpophalangeal joints, using radiologist-labels as the reference.²³ In that study, a convolutional neural network performed well, with an area under the receiver operator curve (AUROC) 0.86. Several conference abstracts have reported on use of AI for identifying CPPD using ultrasound data.^22,24–26 This author expects to see more data emerge on the accuracy of AI for detecting CPPD in the near future.

4. Genetics of CPPD disease

Rare familial cases of CPPD disease, usually early-onset (e.g. well before age 60), led to identification of two causal genes: ANKH, which encodes a transmembrane ATP transporter, and TNFRSF11B, which encodes osteoprotegerin (OPG) (Table 1). The ANKH mutation is thought to be gain-of-function,²⁷ while the TNFRSF11B mutation is loss-of-function and increases osteoclast activity in subchondral bone.²⁸ A number of small cohort studies investigated polymorphisms in ANKH^29,30, ENPP1³¹, TNAP³¹ and HFE³² as possibly linked to CPPD disease, though overall their findings were inconclusive.

Table 1.

Published genes implicated in CPPD disease to date

Gene	Monogenic or polygenic risk	Rare or common	Loss or gain of function	Function of encoded protein
ANKH	monogenic	rare	gain	ATP transport channel (source of PPi)
TNFRSF11B	monogenic	rare	loss	Osteoprotegerin (OPG) inhibits bone resorption
ENPP1	polygenic	common	gain	Generates PPi from ATP
RNF144B	polygenic	common	gain	Ubiquitin-protein ligase

In 2025, the first CPPD disease genome-wide association study (GWAS) was published using nationwide data from >3000 individuals with CPPD disease in the Million Veterans Program.³³ Genome-wide significant signals for CPPD disease were detected at the same two genes in two different genetic cohorts (i.e., a cohort of African genetic ancestry and a cohort of European genetic ancestry): ENPP1, which generates inorganic pyrophosphate – a component of calcium pyrophosphate crystals – and RNF114B, which negatively regulates apoptosis. A meta-analysis of GWAS data from the Rotterdam Study, Multicenter Osteoarthritis Study, and Arthritis Research UK Osteoarthritis Genetics, included >11,000 knees with chondrocalcinosis and similarly identified a significant signal in RNF144B; additional signals were detected at ANKH, RGS7, and ZNF445.^34,35 These studies are a stepping stone to larger, multi-ethnic GWAS to better understand polygenic contributors to CPPD disease.

5. Long-term outcomes and disease burden in CPPD disease

Outcomes in joints with CPPD

CPPD disease pathophysiology begins in the cartilage, where calcium pyrophosphate (CPP) crystals form and then trigger activation of the NLRP3 inflammasome, mechanically deposit in cartilage and activate matrix metalloproteinaseses, contributing to cartilage degradation.^3,36 For decades, controversy has existed as to whether CPPD, represented as chondrocalcinosis on radiograph, instigates or worsens osteoarthritis (OA). However, several recent large cohort studies have identified deleterious effects of knee chondrocalcinosis on OA incidence and progression, and associations between chondrocalcinosis and persistent knee pain.^21,37,38 In the prospective Multicenter Osteoarthritis Study (MOST), chondrocalcinosis in a given knee subregion was significantly associated with greater risk for progressive cartilage damage on MRI in the same subregion over two years.³⁷ Additionally, in the prospective Rotterdam Study with up to 20 years of follow-up, chondrocalcinosis was associated with significantly higher risk for incident knee OA in individuals without baseline radiographic OA.²¹ MOST participants with chondrocalcinosis in the knee were also significantly more likely to report more frequent, persistent, and worsening knee pain over two years.³⁸

Prospective association between CPPD and diseases outside the joint

Many types of inflammatory arthritis, including gout and rheumatoid arthritis, are strongly associated with increased risk for cardiovascular disease, osteoporosis, and other adverse outcomes. A growing number of studies suggest that individuals with CPPD disease are also at greater risk for cardiovascular events and poor bone health compared to matched comparators without CPPD. Two prospective cohort studies, one using nationwide Veterans Administration data and one using tertiary care hospital data, have reported a significantly higher risk for non-fatal cardiovascular events in individuals with CPPD disease than age- and sex-matched comparators.^39,40 Individuals with CPPD disease had lower cortical bone mineral density in the hand and more frequent osteoporosis diagnosis codes in several studies.^41,42 Two recent studies have identified significantly higher risk for fragility fractures in individuals with CPPD disease than age- and sex-matched comparators, with similar risks in men and women with CPPD disease.^43,44 The pathophysiologic processes underlying cardiovascular risk and fracture risk are likely manifold, including consequences of intermittent (or in some cases, chronic) inflammation, dysregulated mineralization related to alterations in phosphate:PPi homeostasis, and greater joint pain leading to less physical activity.

6. Outcomes and endpoints for clinical research, including trials

Clinical trials in CPPD disease have been few and far between, and those that have been published have reported a variety of outcomes that limit comparability across studies. For example, a three-month methotrexate trial reported change in Disease Activity Score measured in 44 joints (DAS44) as the primary endpoint, while a six-month hydroxychloroquine trial reported the percentage of responders defined as at least a 30% decrease in tender or swollen joint count.^45,46 The Outcomes Measure in Rheumatology Trials (OMERACT) CPPD Disease Working Group formed in 2018 to establish core domain sets to be measured in future CPPD disease clinical research. Through a scoping literature review, qualitative interviews with patients and caregivers, and consensus-based work among clinicians, patients, and other stakeholders, two sets of core domains were established: one for studies focused on an individual flare of acute CPP crystal arthritis, and one for studies focused on recurrent flares or chronic CPP crystal arthritis.^47–51 Core domains for both acute and chronic/recurrent CPPD disease studies are: pain intensity, joint tenderness, joint swelling, overall function, patient global assessment of disease activity, and adverse events. Duration of acute flare is an additional core domains for acute CPP crystal arthritis, while additional core domains for chronic/recurrent CPP crystal arthritis are acute CPP crystal arthritis flares and physician global assessment of disease activity. For clinical trials, patient global assessment of response to treatment is also to be measured in trials for both acute and chronic/recurrent CPP crystal arthritis; use of rescue medications is to be measured in trials for acute CPP crystal arthritis; and joint damage and joint calcification on imaging tests are to be measured in trials for chronic/recurrent CPP crystal arthritis where appropriate.⁵¹

7. Toward targeted treatments for CPPD disease

Several recent reviews address current approaches for treating the various manifestations of CPPD disease, including acute CPP crystal arthritis and chronic CPP crystal arthritis; the reader is referred to those for a thorough discussion of clinical trial data and proposed treatment algorithms for the use of colchicine, glucocorticoids, non-steroidal anti-inflammatory drugs, methotrexate, hydroxychloroquine, IL-1 inhibitors and IL-6 inhibitors.^52,53 CPP crystals active the NLRP3 inflammasome and also can trigger inflammation through toll-like receptors or contact with cell membranes, leading to activation of innate immune cells and inflammatory cytokine production.^54,55 As in gout, another common crystalline arthritis, current treatment for CPPD disease aims to inhibit the inflammatory response to CPP crystals. However, unlike gout where urate-lowering therapy directly targets the underlying pathogenic process by reducing monosodium urate crystal formation, no equivalent disease-modifying therapies presently exist to inhibit the formation of, or to dissolve, CPP crystals.

CPP crystals form around chondrocytes; cartilage is central to both CPPD pathogenesis and disease. Current understanding of CPP crystal formation centers around three proteins affecting extracellular inorganic pyrophosphate (PPi) concentration: (1) nucleoside triphosphate pyrophosophohydrolase enzymes such as ENPP1 that metabolize ATP, thereby generating inorganic pyrophosphate (PPi); (2) ANKH, a transmembrane ATP transporter that, by regulating extracellular ATP concentration, regulates extracellular PPi concentration; and (3) pyrophosphatase enzymes such as alkaline phosphatase (ALP) that degrade PPi.³ First steps toward targeted treatment of CPPD disease could center around any of these proteins.

ANKH inhibition was tested in CPPD disease in a proof-of-concept study of probenecid, which blocks the function of ANKH in vitro.^56,57 Among five participants randomized to probencid or no treatment for five days, synovial fluid ATP levels were numerically lower in the probenecid group (4.4 [SD 2.5] nanomoles/mL) than the untreated group (31.4 [SD 56.6] nanomoles/mL), though the small sample size limits results interpretation.

The central role of ENPP1 in this crystalline arthritis is underscored by the recently published first-ever GWAS in CPPD disease, which identified causal variants in ENPP1 in individuals of European and African genetic ancestry.³³ The lead ENPP1 variant conferred gain-of-function in skeletal muscle, suggesting that it causes increased generation of inorganic pyrophosphate and thus increased CPP crystal formation.⁵⁸ ENPP1 inhibitors repress ATP-induced CPP crystal formation in vitro.⁵⁹ ENPP1 inhibitors are being developed as cancer therapeutics;^60–62 clinical trials of these in CPPD disease would be of great interest.

Asfotase alpfa is an ALP repletion product that is FDA-approved for treatment of infantile/juvenile-onset hypophosphatasia. While some individuals with hypophosphatasia have CPPD disease -- and chondrocalcinosis is a minor diagnostic criterion for adult-onset hypophosphatasia⁶³ -- the prevalence of hypophosphatasia in CPPD disease has not been well-established. It remains unknown whether supplementing ALP could potentially prevent CPP crystal formation in individuals with normal serum alkaline phosphatase levels.

Drawing a parallel from therapeutic approaches in gout, potential therapies to dissolve CPP crystals might function in several ways. Treatments that directly dissolve or lyse CPP crystals would be akin to recombinant uricase therapy in gout; whether acute CPP crystal arthritis flares increase in frequency during such treatment would need to be closely monitored. Another theoretical approach could be to lower the concentration of inorganic pyrophosphate or CPP crystals in synovial fluid or cartilage, which might promote CPP crystal dissolution; this approach would be similar to using xanthine oxidase inhibitors to reduce the total body burden of urate. However, a challenge to CPP crystal dissolution is that in the absence of enzymes such as pyrophosphatase, inorganic pyrophosphate hydrolysis occurs extremely slowly.^64,65 A recent study described an in vitro cation exchange in which zinc ions in a synthetic zinc-pyrophosphate complex outcompeted/replaced calcium ions in calcium pyrophosphate crystals, resulting in dissolution of calcium pyrophosphate crystals.⁶⁵ In a study that screened bacteriophage libraries to identify peptides that could potentially lyse CPP crystals, the lead peptide (R25) lowered inorganic pyrophosphate by approximately 20%, in a magnesium-dependent process.⁶⁴ Histone deacetylase inhibitors were studied in vitro in human articular chondrocytes treated with TGF-beta1, with the observation that the histone deacetylase inhibitors reduced ENPP1 and ANKH activity and increased alkaline phosphatase activity, thus reducing CPP crystal formation.^66,67 These approaches could be pursued in animal models using synthetic CPP crystals, with a particular eye toward whether flares of acute monoarthritis are triggered as CPP crystals are dissolved.

Conclusion

This is an exciting time for CPPD disease, which is gaining attention with advances in understanding the genetic underpinning of disease, methods to quantify disease burden on imaging, recognition of consequences in and beyond the joint, and consensus on what to measure in future clinical studies. Development of targeted treatments and public awareness campaigns about CPPD disease are a much-needed next steps.