SEARCH

SEARCH BY CITATION

Keywords:

  • Chondrocalcinosis;
  • Magnesium;
  • Hypomagnesemia;
  • Calcium pyrophosphate dihydrate crystals

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Objective

To determine an association between magnesium (Mg) depletion and chondrocalcinosis, which has been reported but not investigated in a cross-sectional study.

Methods

Prevalence of chondrocalcinosis was investigated in 144 individuals: 72 patients receiving home parenteral nutrition (HPN) compared with 72 age- and sex-matched controls. Presence of chondrocalcinosis was assessed by knee radiographs. Blood serum and globular Mg levels and 24-hour urinary Mg content were compared.

Results

Mean ± SD age for both patients and controls was 51 ± 17 years, and 51% in both groups were women. Mean duration of HPN was 6.4 years. Prevalence of chondrocalcinosis was markedly higher in patients receiving HPN than controls (16.6% versus 2.7%; P = 0.006, odds ratio [OR] 7.0, 95% confidence interval [95% CI] 1.45–66.1). Mean ± SD serum and globular Mg levels were significantly lower in patients than controls (serum: 0.75 ± 0.09 mmoles/liter versus 0.81 ± 0.08 mmoles/liter, P = 0.0006; globular Mg: 1.8 ± 0.31 mmoles/liter versus 2.0 ± 0.35 mmoles/liter, P = 0.0003). Twenty-four-hour urinary Mg level was lower in patients than controls (mean ± SD 3.85 ± 1.50 mmoles versus 5.37 ± 3.71 mmoles; P = 0.001). Prevalence of chondrocalcinosis was significantly higher in patients with a low serum Mg level (OR 13.5, 95% CI 2.76–127.3, P < 0.0001), with a similarly high but not significant occurrence of chondrocalcinosis in patients with a low globular Mg level (OR 4.09, 95% CI 0.603–20.26, P = 0.08) and in patients with a low 24-hour urinary Mg level (OR 3.9, 95% CI 0.77–16.34, P = 0.05).

Conclusion

Long-lasting Mg depletion is strongly associated with chondrocalcinosis.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Articular chondrocalcinosis is characterized by the deposition of calcium pyrophosphate dihydrate (CPPD) crystals in articular cartilage. Most often, chondrocalcinosis is an incidental radiographic finding in elderly patients. Alternatively, it can present as acute synovitis or chronic arthritis or as a chronic arthropathic feature (1, 2).

Prevalence of chondrocalcinosis increases with age in the most common sporadic disease subset (1, 2). Chondrocalcinosis can also be a familial disease, frequently characterized by early onset and florid polyarticular involvement (3). Chondrocalcinosis may also be related to several metabolic diseases, including hypophosphatasia, hemochromatosis, or primary hyperparathyroidism (4).

Several other diseases were first suggested to be associated with chondrocalcinosis based on observational studies in which several biases may have been introduced. For instance, diabetes mellitus and hypothyroidism were once believed to be a source of secondary chondrocalcinosis, but their link to CPPD deposition was subsequently discounted by controlled studies involving age-matched controls (4, 5).

Hypomagnesemia belongs to the list of widely admitted causes of secondary chondrocalcinosis because features of CPPD deposition disease have been repeatedly reported in young patients with hypomagnesemia induced by renal leakage of magnesium (Mg) (6–14). Chondrocalcinosis has also been described in patients with Mg depletion due to short bowel syndrome (15). However, evidence for a definitive link between hypomagnesemia and chondrocalcinosis is incomplete, as no controlled study has been performed to the best of our knowledge.

Mg is the most prevalent intracellular divalent cation in the body. Its distribution is approximately equally divided between the skeleton and the soft tissues (16–18). The serum Mg concentration is the most clinically available test for assessing Mg status. However, it may not always reflect the true total body Mg content because low intracellular Mg level has been documented in patients with normal serum concentration. Globular Mg, which is the intracellular level of Mg in red blood cells, appears to be a more accurate assessment of body Mg status (18, 19). In patients with Mg depletion, measurement of 24-hour Mg excretion can help to distinguish between gastrointestinal or renal losses. A high daily Mg excretion indicates renal Mg wasting because normal renal response to Mg depletion is to lower magnesium excretion to very low levels (16, 17, 20).

Intestinal failure (IF) in patients receiving home parenteral nutrition (HPN) is a unique condition for studying the association between chondrocalcinosis and Mg depletion, because in these patients severe hypomagnesemia frequently develops despite Mg supplementation. The present prospective cross-sectional study was designed to assess the prevalence of chondrocalcinosis in patients receiving HPN and to investigate the link between Mg status and chondrocalcinosis in these patients.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Patients and controls.

A total of 81 consecutive patients attending the HPN center at our institution between November 2002 and March 2005 with a diagnosis of IF requiring HPN were considered for inclusion. Inclusion criteria required HPN of at least 2 years' duration, regardless of age and underlying causes of IF. One patient refused to enter the study, and the duration of HPN was too short in 8 patients. Therefore, we enrolled 72 patients. HPN was tailored to the needs of each patient. Patients received a median of 5 cyclic nocturnal HPN infusions per week, each infusion containing 14 mmoles of Mg chloride. In addition, all patients received daily oral vitamins and mineral supplements, including 9 mmoles of Mg lactate. Further details of the HPN protocol in our center have been described elsewhere (21).

The control group comprised 72 consecutive patients hospitalized for low back pain in our rheumatologic department from November 2002 to March 2005 who were matched by sex and age (within 5 years) with the HPN group. The study protocol was approved by the local research ethics committee, and all participants provided informed consent.

Data collection.

Weight-bearing, fully extended anteroposterior radiographs of both knees were obtained from each patient and control. All tibiofemoral joints were assigned a Kellgren/Lawrence (K/L) grade for osteoarthritis (OA) (22). Chondrocalcinosis was defined by the presence of a definite linear calcification seen on the radiographs of either knee. The search for chondrocalcinosis was restricted to the tibiofemoral compartments because knee radiographs have been shown to detect calcifications in 90% of patients with chondrocalcinosis (23). Chondrocalcinosis was considered absent if both knees showed no evidence of calcification. Two independent readers (TB and GA) who were unaware of the patient details read all the radiographs in a randomly ordered sequence. Any discrepancies were resolved by a consensus reading with a third observer (PR). The day the radiographs were obtained, Mg status was assessed in both groups by measurement of blood level of serum and globular Mg (normally 0.75–0.98 mmoles/liter and 1.5–2.20 mmoles/liter, respectively) and determination of 24-hour urinary Mg level (normally 2–6 mmoles). To exclude patients with hemochromatosis or hyperparathyroidism, levels of calcium, phosphate, transferrin saturation, and ferritin were determined in all participants. Moreover, patients or controls with a known personal or family history of CPPD disease were excluded. Joint effusion, when present, was aspirated and synovial fluid was analyzed under compensated polarized light microscopy.

Statistical analysis.

We determined that 120 patients (60 in the HPN group and 60 in the control group) would allow an 80% power to demonstrate, at a 5% significance level, a difference in prevalence of chondrocalcinosis of at least 20% between the 2 groups, considering a chondrocalcinosis prevalence of <5% in the control group. Student's t-test was used to compare quantitative variables between patients and controls. Because the criteria for asymptotic calculations were not fulfilled, exact 95% confidence intervals (95% CIs) of odds ratios (ORs) were used to estimate prevalence of chondrocalcinosis in the 2 groups, and exact conditional stratified logistic regression was used to test the relationship between HPN and chondrocalcinosis. In addition, because it cannot be excluded that the presence and/or the severity of OA might interfere with the diagnosis of chondrocalcinosis, we compared the K/L grade in the 2 groups. We then checked that the conclusions were robust when the analysis was stratified according to the K/L grade of each knee using Mantel-Haenszel and Breslow-Day tests for testing the heterogeneity of OR between the different grades of OA. In a second step, we analyzed, in the entire data sample, the prevalence of chondrocalcinosis according to the presence/absence of hypomagnesemia or abnormally low 24-hour urinary Mg level. The cutoff points we used to dichotomize the Mg levels (blood, globular, and urinary) were the lower value of normal range in our laboratory. Prevalence of chondrocalcinosis was compared using Fisher's exact test and exact 95% CIs of ORs were calculated (note that exact CIs are wider than asymptotic CIs). All criteria for categorization of patients regarding their level of Mg were defined a priori. All analyses were performed using SAS software, version 9.13 (SAS Institute, Cary, NC). Two-sided P values less than 0.05 were considered significant.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Demographic features.

A total of 144 patients and controls (51% women in both groups) were recruited, equally distributed in the HPN and control groups. The mean ± SD age for both patients and controls was 51 ± 17 years. None of the included patients showed evidence of hemochromatosis or hyperparathyroidism. The 72 patients with IF had been treated with HPN for a mean duration of 6.4 years (range 2–16 years). These patients' intestinal diseases are listed in Table 1. The most frequent causes of IF were radiation-induced enteritis, mesenteric arterial thrombosis, and Crohn's disease.

Table 1. Underlying intestinal diseases in 72 patients receiving home parenteral nutrition
DiseaseNo.
Radiation enteritis18
Mesenteric arterial thrombosis14
Crohn's disease13
Mesenteric venous thrombosis6
Chronic intestinal pseudoobstruction5
Volvulus4
Gardner disease3
Congenital origin3
Miscellaneous6

Prevalence of chondrocalcinosis.

Prevalence of chondrocalcinosis was significantly higher in patients receiving HPN than in controls (16.6% versus 2.7%; P = 0.006, OR 7.0, 95% CI 1.45–66.1) (Table 2). None of the controls diagnosed with chondrocalcinosis showed clinical symptoms of CPPD deposition. By contrast, 25% of patients with chondrocalcinosis had experienced at least 1 pseudogout attack. The acute arthritis episodes were located in the knee (n = 2) or ankle (n = 1). In all of these cases, analysis of synovial fluid under polarizing light microscopy demonstrated CPPD crystals. The individuals with chondrocalcinosis were younger in the patient group (mean ± SD age 60.19 ± 16.7 years) than in the control group (mean ± SD age 80.5 ± 4.9 years).

Table 2. Characteristics of patients and controls*
CharacteristicPatients treated with HPN (n = 72)Age- and sex-matched controls (n = 72)OR (95% CI)P
  • *

    Values are the mean ± SD unless otherwise indicated. HPN = home parenteral nutrition; OR = odds ratio; 95% CI = 95% confidence interval.

  • n = 12.

  • n = 2.

  • §

    P value of exact conditional stratified logistic regression.

Age, years51 ± 1751 ± 17  
Women, %5151  
Magnesium level    
 Serum, mmoles/liter0.75 ± 0.090.81 ± 0.08 0.0006
 Globular, mmoles/liter1.8 ± 0.312.0 ± 0.35 0.0003
 24-hour urinary, mmoles3.85 ± 1.505.37 ± 3.71 0.001
Chondrocalcinosis, %16.62.77.0 (1.45–66.1)0.006§

K/L grade did not statistically differ between the HPN group and the control group. Moreover, the difference in prevalence of chondrocalcinosis between the 2 groups persisted when the analysis was stratified according to K/L grade (P = 0.003, OR 7.8 [adjusted for K/L grade], 95% CI 2.6–23.6), without any significant heterogeneity of OR between the different grades of OA (P > 0.6).

Mg values.

Mg values (mean ± SD) for both groups are indicated in Table 2. Mean ± SD serum and globular Mg level was significantly lower in patients than in controls (serum: 0.75 ± 0.09 mmoles/liter versus 0.81 ± 0.08 mmoles/liter, P = 0.0006; globular: 1.8 ± 0.31 mmoles/liter versus 2.0 ± 0.35 mmoles/liter, P = 0.0003). Similarly, 24-hour urinary Mg level was significantly lower in patients than in controls (3.85 ± 1.50 mmoles versus 5.37 ± 3.71 mmoles; P = 0.001).

Considering the entire sample of data, the prevalence of chondrocalcinosis was higher in patients with low serum Mg level (OR 13.5, 95% CI 2.76–127.3, P < 0.0001) (Table 3). A similar trend of increased prevalence of chondrocalcinosis, although not significant, was found in patients with low globular Mg level (OR 4.09, 95% CI 0.603–20.26, P = 0.08) and low 24-hour urinary Mg level (OR 3.9, 95% CI 0.77–16.34, P = 0.05).

Table 3. Association between Mg depletion and chondrocalcinosis*
Mg levelChondrocalcinosis, %OR (95% CI)P
  • *

    Mg = magnesium; OR = odds ratio; 95% CI = 95% confidence interval.

Serum 13.5 (2.76–127.3)< 0.0001
 <0.75 mmoles/liter, (n = 52)23.1  
 ≥0.75 mmoles/liter, (n = 92)2.2  
Globular 4.09 (0.60–20.26)0.08
 <1.5 mmoles/liter, (n = 13)27.3  
 ≥1.5 mmoles/liter, (n = 131)8.4  
24-hour urinary 3.9 (0.77–16.34)0.05
 <2 mmoles, (n = 16)25  
 ≥2 mmoles, (n = 128)7.8  

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

The observation of chondrocalcinosis in patients with severe hypomagnesemia due to short bowel syndrome (15) led us to hypothesize that long-lasting Mg depletion due to IF could be a cause of CPPD disease. Mg depletion is common in patients with IF because its main absorptive site, the distal small bowel, is most frequently defective (24). Despite increased Mg intake in patients receiving HPN, hypomagnesemia frequently develops because enteral or oral Mg supplementation may be a source of diarrhea and/or may be insufficient to prevent excessive intestinal Mg loss (25–27).

We therefore designed this prospective study of patients receiving long-term HPN therapy followed up in our hospital and controls matched for age, a well-known factor of chondrocalcinosis prevalence, and sex, because some studies (28, 29), although not all (30), have suggested a preponderance for CPPD deposition among women. Because estimates of the prevalence of chondrocalcinosis are predominantly based on radiographic findings of the disease (28, 30–32), we screened for CPPD calcifications on radiographs, focusing on the knees because these joints are the most frequently affected by chondrocalcinosis (1, 2, 33). The crude prevalence of chondrocalcinosis, as estimated by knee radiographs, in our control group (mean age 51 years) was 2.7%, which agrees with data from a large population-based radiographic survey (28) and from a recent cross-sectional study (30), which demonstrated a prevalence ranging from 3% to 5% in individuals younger than 70.

We found a significantly increased prevalence of chondrocalcinosis in the HPN group, and patients with chondrocalcinosis receiving HPN were younger than those in the control group, which further supports a causative link between HPN and CPPD deposition. Moreover, our study confirmed that reduced serum Mg level was strongly associated with chondrocalcinosis, with a trend, although not significant, toward a decreased globular Mg level and 24-hour urinary Mg level as related factors.

Diagnosing Mg deficiency may not be easy, because serum Mg accounts for only 0.3% of total body Mg. Although hypomagnesemia reliably indicates Mg deficiency, a normal plasma Mg concentration does not exclude Mg depletion (18, 20). Globular Mg concentration is usually considered a more sensitive parameter to evaluate Mg content (34), but some authors have recently highlighted the low sensitivity of globular Mg level in the diagnosis of Mg depletion (19). The Mg tolerance test, which is based on the measurement of a parenterally administered Mg load retention, is a more accurate method for the detection of Mg deficiency. However, the test is cumbersome and unsuitable for routine diagnostic purposes (18, 20). We found that 24-hour urinary Mg level was significantly lower in patients than in controls, which is consistent with Mg deficiency resulting from gastrointestinal losses in patients receiving HPN because the normal renal response to Mg depletion has been shown to be a decrease in Mg renal excretion (18, 20).

Two controlled studies of biochemical abnormalities failed to demonstrate a significantly reduced serum Mg level in patients with sporadic chondrocalcinosis as compared with controls (35, 36). Although the number of patients included in these studies was small, these results suggest that hypomagnesemia was not associated with sporadic chondrocalcinosis. The occurrence of chondrocalcinosis in patients with hypomagnesemia due to renal waste has been mainly described in patients with Gitelman or Bartter syndrome (4, 6, 7, 12, 37). Finding chondrocalcinosis in individuals with hypomagnesemia due to causes other than renal genetic disorders has strengthened the hypothesis of a role for Mg in CPPD deposition disease. Two patients with chondrocalcinosis and hypomagnesemia while receiving tacrolimus therapy, a drug known to induce renal leakage of Mg, have been described (11). The association reported by Neame et al (30) between chondrocalcinosis and use of diuretics might be explained by increased urinary Mg loss because loop and thiazide diuretics inhibit Mg reabsorption by the renal tubule. In agreement with this hypothesis, no such association was found between chondrocalcinosis and the use of other antihypertensive agents (30). Our study is the first to demonstrate a significant association of chondrocalcinosis with Mg depletion due to IF. Only 1 patient with primary isolated Mg malabsorption, a rare intestinal disorder, who presented with acute synovitis of the knee presumably related to chondrocalcinosis has been previously reported (38).

The role of hypomagnesemia in chondrocalcinosis is not fully understood. Hypomagnesemia could favor chondrocalcinosis through intraarticular elevation of extracellular inorganic pyrophosphate (PPi) level and/or reduced level of the saturation product of CPPD (6). Increased joint fluid PPi level is also found during hypophosphatasia, hemochromatosis, and hyperparathyroidism and is presumably a common thread in the pathogenesis of chondrocalcinosis via cartilage PPi excess (1). The therapeutic potential of Mg in patients with pyrophosphate arthropathy is unknown. A 6-month, double-blind, placebo-controlled trial of Mg in patients with chondrocalcinosis demonstrated a trend toward improvement in pain and stiffness in patients taking Mg (39). Nevertheless, this study did not demonstrate any effect on radiographic evidence of chondrocalcinosis, but the study may have been underpowered with too short of a duration (39).

Our study has several limitations. Patients were not selected based on the presence or absence of hypomagnesemia but, rather, on the presence of a risk of hypomagnesemia due to IF. Radiographs of knees can be insensitive to the detection of CPPD crystal deposits and we did not perform specific crystal identification in all patients diagnosed with chondrocalcinosis. We did not assess chondrocalcinosis in other areas, such as the wrist and symphysis pubis, which may have slightly underestimated the actual prevalence of chondrocalcinosis in our 2 groups. Nevertheless, because both groups were assessed in a similar fashion, this is unlikely to have introduced bias in our results. Finally, although vitamins, minerals, and trace element levels were closely monitored in our patients receiving HPN, we cannot exclude the possibility of some other deficiencies in patients. However, micronutrient or vitamin deficiency has never been linked to chondrocalcinosis to date.

In conclusion, our data show that patients with IF receiving HPN therapy have a greater prevalence of chondrocalcinosis, and that hypomagnesemia is associated with CPPD deposition disease. Our work highlights the need to carefully detect Mg deficiency in patients with chondrocalcinosis.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Dr. Richette had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Richette, Bardin.

Acquisition of data. Richette, Ayoub, Lahalle, Badran.

Analysis and interpretation of data. Richette, Lahalle, Joly, Messing.

Manuscript preparation. Richette, Vicaut, Bardin.

Statistical analysis. Vicaut.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

We would like to thank the patients for their participation in this study.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES