Higher Rates of Psoriatic Arthritis in Patients with Calcium Pyrophosphate Deposition Disease than in Controls: A Retrospective Cohort Study

Abstract

Calcium pyrophosphate deposition disease (CPPD) has been shown to be associated with inflammatory arthritis such as rheumatoid arthritis. However, few studies have investigated the correlation between CPPD and psoriatic arthritis (PsA). Our study aimed to determine whether there were higher rates of PsA in patients with CPPD than controls. A retrospective cohort study was conducted using the Veterans Affairs’ Corporate Data Warehouse. Individuals with a CPPD ICD code were matched with controls and diagnoses of PsA and psoriasis were collected. A total of 41,084 CPPD patients were matched with 119,192 controls. The proportion of CPPD patients with PsA diagnosis was more than double that of controls (1.07% vs. 0.37%; p < 0.0001), and more CPPD patients were diagnosed with psoriasis (3.05% vs. 2.52%; p < 0.0001). Those with CPPD had higher odds of a PsA diagnosis (OR 3.550, 95% CI 2.602–4.844). A total of 61.59% of PsA diagnoses preceded the CPPD diagnoses by at least one year. This is the first case–control study demonstrating an association between CPPD and PsA, potentially related to the fact that both PsA and CPPD could be triggered by trauma, and are closely associated with osteoarthritis. It also is possible that inflammatory pathways contribute to CPP crystal deposition in joints.

1. Introduction

Calcium pyrophosphate deposition disease (CPPD) is an acute and chronic arthritis triggered by calcium pyrophosphate (CPP) crystals. While CPPD typically presents acutely as monoarticular or oligoarticular arthritis as its own entity [1], CPPD has also been shown to be a late complication of inflammatory arthritis such as rheumatoid arthritis [2,3]. Gout has also been shown to complicate psoriatic (PsA) arthritis; several studies have established that hyperuricemia is more common among patients with psoriasis compared to the general population [4,5]. However, few studies have investigated the correlation between CPPD and PsA [4].

There have been two case reports of PsA patients with CPPD that shared certain HLA alleles. The authors speculated that the association might be tied to an underlying immunogenetic predisposition to autoimmunity, although they suggested that this association might also be serendipitous [6]. Other studies have examined synovial fluid (SF) samples from PsA patients. One small study found calcium pyrophosphate (CPP) crystals in 21.62% of patients with PsA [7]. Similarly, two retrospective studies found CPP crystals in the synovial fluid of PsA patients in 3.8% of samples and 3.82% of samples [8,9]. Due to the retrospective nature of the study, it was not possible to determine whether the CPP crystals were associated with prior joint trauma [8]. The authors hypothesize that CPP crystals could be linked to alterations of biochemical pathways in PsA patients leading to the production of inorganic pyrophosphate [8,10].

Given that the studies surrounding the association between CPPD and PsA have been limited, our study aimed to determine whether there were higher rates of PsA in patients with CPPD than age and sex matched controls, and as well as the timeframe between the separate diagnoses.

2. Materials and Methods

A retrospective case–control study was conducted using the national Veterans’ Health Administration’s Corporate Data Warehouse. Patients who had one or more encounters with an ICD code for CPPD (ICD-9: 712.[123]% or 275.49; ICD-10: M11.2*) between January 1 2010 and March 1 2020 were pair-matched to control patients based on sex, year of birth, vital status at the time of the CPPD patient’s CPPD index date, and having had an inpatient or outpatient encounter the same month as the matched case’s first encounter with the CPPD ICD code, defined as the index date. Up to four controls were matched to each case without replacement via custom SAS script; cases with no control matches were excluded.

Diagnoses of PsA and psoriasis were collected, identified by the presence of one or more ICD codes at any time between January 1 2010 and March 1 2020. The absence of any of these ICD-9/10 codes was assumed to indicate an absence of these diseases. The date of first coding event was recorded for both PsA (ICD-9: 696.0; ICD-10: L40.5*) and psoriasis (ICD-9: 696.[123458]; ICD-10: L40.[013489]*).

Statistical Analysis

Descriptive statistics were performed, followed by univariates observing the differences in PsA between CPPD and control groups via Chi square, Fisher’s Exact test, and one-way ANOVA, as appropriate. Point and relative frequencies were reported within each of the case and control groups. Conditional logistic regression with stratification by case–control matching was performed to determine whether the diagnosis of PsA was mathematically predicted by CPPD after also controlling for psoriasis and race. Lastly, among those with CPPD, the time between the initial dates of the CPPD and PsA diagnoses was described.

p-values less than 0.05 were considered significant. The statistical analysis was conducted using SAS 9.4 (SAS, Cary, NC, USA) [11].

3. Results

A total of 52,775 unique CPPD patients were identified based on having had one or more ICD code, of whom 41,084 had at least one eligible matching control, and were thus retained. A pool of 327,031 fully matched, qualifying control candidates was identified, of which 119,192 were retained. The resultant sample included 160,276 patients (

Table 1. Baseline characteristics of PsA and psoriasis diagnoses in CPPD and control cohorts.

Category	CPPD		Control		Total		p-Value
	%	n	%	n	%	n
Asian	0.58%	240	0.40%	471	0.44%	711	<0.0001
Black	15.29%	6283	21.16%	25,222	19.66%	31,505
Hawaiian or Pacific Island	0.90%	370	0.77%	918	0.80%	1288
Native American or Alaskan Native	0.90%	371	0.83%	988	0.85%	1359
White	76.30%	31,348	71.58%	85,320	72.79%	116,668
Female *	4.90%	2013	6.38%	7610	6.00%	9623	<0.0001
Male *	95.10%	39,071	93.62%	111,582	94.00%	150,653

* Matching criterion; not exactly the same proportion across cohorts because not all cases had 4 controls.

).

The sample demographic was largely White (72.29%) and male (94.00%), with an average age of 69.350 (standard deviation: 12.989) (Table 1 and

Table 2. Prevalence and age at PsA and psoriasis diagnoses in CPPD and control cohorts.

Category	CPPD		Control		Total		p-Value
	%	n	%	n	%	n
Alive	76.79%	31,548	57.37%	68,381	62.35%	99,929	<0.0001
Psoriatic Arthritis	1.07%	440	0.37%	444	0.55%	884	<0.0001
Psoriasis	3.05%	1254	2.52%	3001	2.65%	4255	<0.0001
	Mean	STD	Mean	STD	Mean	STD
Age at CPPD Index *	69.5537	12.3671	69.2804	13.1965	69.3504	12.9894	<0.0002
Total	41,084		119,192		160,276

* Matching criterion; not exactly the same proportion across cohorts because not all cases had 4 controls.

). Those who identified as Black were less likely to have CPPD (15.29% v 21.16%) than the subjects identifying as other races.

The proportion of CPPD patients with a PsA diagnosis was more than double that of controls (1.07% vs. 0.37%; p < 0.0001), and more CPPD patients were diagnosed with psoriasis than controls (3.05% vs. 2.52%; p < 0.0001) (Table 1).

After controlling for comorbid psoriasis and race, those with CPPD had higher odds of a PsA diagnosis than controls (OR 3.550, 95% CI 2.602–4.844) (

Table 3. Logistical regression of PsA diagnosis among those with CPPD compared with controls, after controlling for comorbid psoriasis and demographic factors.

		OR	Lo95	Up95	Sig
CPPD	Control	3.550	2.602	4.844	*
Psoriasis	No Psoriasis	240.840	121.011	479.327	*
Asian	White	0.207	0.008	5.475
Black		0.451	0.264	0.771	*
Island		0.278	0.033	2.313
Native		0.290	0.036	2.355

* denotes significance.

). Among those with CPPD, 61.59% of PsA diagnoses preceded the CPPD diagnoses by at least one year; PsA diagnoses ranged from preceding CPPD by 10 years to following the CPPD diagnosis by 6 years (N = 440) (Figure 1).

4. Discussion

This study is notable as, to our knowledge, it is the first case–control study examining the relationship between CPPD and PsA. This study demonstrates a significant association between CPPD and PsA, with the proportion of CPPD patients with a PsA diagnosis being more than double that of controls (1.07% vs. 0.37%; p < 0.0001). The prevalence of 0.37% of PsA in the general population is in line with other studies; a recent systematic review and meta-analysis estimated the PsA prevalence to be 0.11%, with estimates ranging from 0.01% to 0.19% depending on the geographical location [12]. Moreover, our results suggest that the PsA diagnosis precedes the CPPD diagnosis, which has previously been observed with gout.

Similar reasons might explain the association between PsA and CPPD. There is much literature linking trauma, specifically physical injury, to being a potential trigger in both PsA and CPPD patients [1,9,13]. In regard to trauma-induced PsA, the release of self-antigens from injured joints and substance P, which may be released from peripheral nerve terminals, is hypothesized to lead to the development of PsA [9]. Acute flares of CPPD often occur in the context of joint trauma, such as hip-fracture repairs or meniscectomies [1]. It is also possible that inflammatory pathways, such as those present in PsA, RA, OA, and PsO, contribute to CPP crystal deposition in joints. Understanding the association between CPPD and PsA can not only aid in screening and diagnosis of either condition, but also lead to an understanding of the crystal-stimulated inflammatory cascade.

A similar association has been noted between gout and PsA, and researchers have donned this association “psout”, proposing that it could be thought as a novel overlap syndrome. In this vein, authors make the distinction between “multimorbidity” (defined as the coexistence of two or more chronic conditions in a same individual) and “comorbidity” (defined by the occurrence of diseases in addition to a main disease) [4]. Authors ultimately consider the gout and PsA association as a “multimorbidity/comorbidity entity” [4]. Recognizing multimorbid/comorbid conditions, like CPPD and PsA, is important when it comes to considering which disease-specific guidelines to implement [4].

The strengths of this study include that it is the first case–control study demonstrating an association between CPPD and PsA, and the fact that it utilized the Veterans’ Affairs database of patients which is a national database. One possible limitation is that the ICD10 code used for CPPD is chondrocalcinosis, which is a radiographic finding prevalent in 7.0–18.5% of the general population, possibly including patients in the CPPD cohort who were asymptomatic [14]. In other words, it cannot be ruled out that patients diagnosed with PsA prior to CPPD diagnoses were in fact affected by CPPD, as the presence of chondrocalcinosis is typically used in conjunction with positively birefringent crystals in accordance with the Ryan–McCarty diagnostic criteria to make a definitive diagnosis of CPPD [15]. Another limitation of the study given the utilization of the claimed data is the possible misclassification of codes. Additionally, this study only included the data available within the Veterans’ Affairs database, potentially defining patients as controls in this study who were being treated for CPPD in the community. The VA population is also a predominantly male population, who are more likely to have experienced joint trauma; this potentially might have overestimated the association between CPPD and PsA.

5. Conclusions

In conclusion, this case–control study suggests an association between CPPD and PsA, and that the PsA diagnosis precedes the CPPD diagnosis, a novel finding that merits future research. The coexistence of both CPPD and PsA can pose a challenge in management in regard to both the diagnosis and the choice of therapy. It is important to distinguish between a PsA vs. a CPPD flare, as both can present with monoarticular inflammatory arthritis. Understanding this association would lead to earlier identification, which is crucial in implementing appropriate treatment strategies.