Blood metal concentrations and cardiac structure and function in total joint arthroplasty patients

Abstract

BACKGROUND

There is concern regarding potential long-term cardiotoxicity with systemic distribution of metals in total joint arthroplasty (TJA) patients.

AIM

To determine the association of commonly used implant metals with echocardiographic measures in TJA patients.

METHODS

The study comprised 110 TJA patients who had a recent history of high chromium, cobalt or titanium concentrations. Patients underwent two-dimensional, three-dimensional, Doppler and speckle-strain transthoracic echocardiography and a blood draw to measure metal concentrations. Age and sex-adjusted linear and logistic regression models were used to examine the association of metal concentrations (exposure) with echocardiographic measures (outcome).

RESULTS

Higher cobalt concentrations were associated with increased left ventricular end-diastolic volume (estimate 5.09; 95%CI: 0.02-10.17) as well as left atrial and right ventricular dilation, particularly in men but no changes in cardiac function. Higher titanium concentrations were associated with a reduction in left ventricle global longitudinal strain (estimate 0.38; 95%CI: 0.70 to 0.06) and cardiac index (estimate 0.08; 95%CI, -0.15 to -0.01).

CONCLUSION

Elevated cobalt and titanium concentrations may be associated with structural and functional cardiac changes in some patients. Longitudinal studies are warranted to better understand the systemic effects of metals in TJA patients.

Key Words: Total joint arthroplasty, Metal-on-metal, Cardiotoxicity, Heart failure, Echocardiography, Cobalt

Core Tip: This study evaluated echocardiographic measures in 110 prospectively recruited total joint arthroplasty patients who had a recent history of elevated metal concentrations. Elevated cobalt and titanium concentrations were positively associated with some structural and functional cardiac changes.

INTRODUCTION

Total joint arthroplasty (TJA) is one of the most common surgical procedures in the United States and is expected to increase further in the coming years with a projected yearly incidence of nearly 5 million total hip arthroplasty (THA) and total knee arthroplasties (TKA) by the year 2040[1]. More than 7 million Americans are currently living with at least one TJA[2].

Although the TJA implant designs and materials can vary, most are made of metals including cobalt-chromium alloys, stainless steel, and titanium combined with ceramics and polymers[3]. Chronic exposure to these aforementioned metals can occur as metal containing implants are subjected to repeated mechanical and oxidative stresses that subsequently release metal ions locally and systemically[4]. Locally, metal ions released from both metal-on-metal (MoM) implants and metal-on-polyethylene are associated with adverse local tissue reactions[5,6]. In particular, cobalt was first identified as cardiotoxic in 1967 when Quebec beer drinkers developed acute cardiomyopathy after drinking beer that used cobalt as a foam-stabilizing agent[7]. Considering the history of cobalt-related cardiotoxicity, several studies over the last decade investigated the potential for cardiotoxicity associated with MoM implants with a focus on cobalt-chromium ion release[8-12]. In a postmortem study, cobalt concentration was significantly higher in the myocardial cells of patients who underwent a THA and this was postulated to arise from alterations in cardiac calcium handling[7]. In particular, three echocardiographic studies in THA patients showed higher metal concentrations in patients with MoM implants but no clinically significant associations between increased metal concentrations and cardiac dysfunction[13-15]. Even titanium, a biocompatible inert material, was attributed to osteolysis through the formation of titanium dioxide[16]. Due to the high prevalence of TJA and the potential of metal ions to be distributed systemically, concerns remain regarding the potential for long-term cardiotoxicity stemming from the systemic effects of metal ions[4,7].

With this background, the aim of the current study was to examine the association between three commonly used implant metals (cobalt, chromium, titanium) with echocardiographic measures and cardiac biomarkers in THA and TKA patients regardless of their implant type.

MATERIALS AND METHODS

The study population comprised 110 TJA patients prospectively recruited at a large tertiary care medical center between March 2019 and December 2022. Patients were invited to participate based on having elevated blood metal concentrations of either titanium, cobalt, and/or chromium, as flagged in the electronic medical records (concentrations higher than 1 ng/mL for cobalt and chromium, and higher than 2 ng/mL for titanium) over the previous one-year period and had not undergone a revision procedure since their elevated laboratory measurements.

All patients consented to the use of their medical data for research purposes. The Institutional Review Board approved the study and a listing was maintained on clinicaltrials.gov through the duration of the study. All patients had a previous surgical history of at least one prior THA, TKA, or hip resurfacing procedure. The surgical history was ascertained through self-report and the institutional Total Joint Registry.

Patients who agreed to participate traveled to a single day of in-person appointments where they underwent a single blood draw to measure metal (cobalt, chromium, titanium) concentrations and troponin, and N-terminal pro b-type natriuretic peptide (NT-proBNP) as general indicators of heart health. The cardiac biomarkers were measured using an electrochemiluminescent immunoassay and the metals assayed using inductively coupled plasma mass spectrometry within an International Organization for Standardization Class-7 Laboratory[17,18]. In the present study, cobalt and chromium were analyzed in whole blood while titanium was analyzed in serum. Due to ordering and lab errors, one patient was missing the chromium level, one patient was missing the NT-proBNP level, and two patients were missing the troponin level.

All patients also underwent comprehensive transthoracic echocardiogram (GE E95, General Electric Healthcare) consisting of two-dimensional (2D), three-dimensional (3D), Doppler and speckle-tracking systolic strain echocardiography on the same day as the blood draw. All patients received an ultrasound enhancement agent (Lumason, Bracco Diagnostics) and underwent enhanced measurement of 2D biplane volumes and unenhanced 3D volumes as recommended. All echocardiograms were performed according to the American Society of Echocardiography guidelines[19] and reviewed by a board-certified cardiologist with echocardiography expertise blinded to the results of metal levels.

Patient characteristics, laboratory values and echocardiographic findings were summarized using descriptive statistics [mean ± SD, median, interquartile range (IQR), count, percentage]. Many patients had “undetectable” metal concentrations (< 1 ng/mL), which were converted to numeric by using the highest numeric under the undetectable (e.g. 0.99 ng/mL). Metal measurements were treated as binary (detectable vs undetectable) or continuously on a log transformed scale (levels < 1 ng/mL was converted to 0.99 ng/mL). The association of log transformed metal concentrations (exposure) with echocardiographic measures and cardiac biomarkers (outcomes) were examined using age and sex-adjusted linear and logistic regression models. For significant results, β coefficients and odds ratio (OR) estimates from regression models were back-transformed to facilitate interpretation in original scales. All estimates represent the change corresponding to a doubling of the recorded metal level (e.g. from 1 ng/mL to 2 ng/mL).

RESULTS

The mean age of the 110 TJA patients was 68.8 (SD: 9.2 years) years, 45 (40.9%) were women and 108 (98.2%) were white (Table 1). Only 4 patients were current smokers and 1 had a previous history of heart failure. A total of 79 (71.8%) patients had a history of THA only, 11 (10.0%) TKA only and 20 (18.2%) patients had both surgeries. Of the 99 patients with at least 1 THA, 52 had a history of MoM implants, 5 had ceramic-on-poly implants and 3 had both MoM and ceramic-on-poly implants (remaining had metal-on-poly implants). The average time since the first THA or TKA surgery was 13.4 (SD: 5.7 years) years. A total of 59 patients (53.6%) had at least one revision for their TJA. Detectable metal concentrations were found for chromium (n = 43, 39.5%), cobalt (n = 51, 46.4%) and titanium (n = 89, 80.9%) (Table 2). Median (IQR) metal concentrations among detectable were 1.9 (1.2 ng/mL, 3.8 ng/mL) ng/mL for chromium, 3.7 (1.9 ng/mL, 6.3 ng/mL) ng/mL for cobalt and 3.0 (2.0 ng/mL, 5.0 ng/mL) ng/mL for titanium.

Table 1 Characteristics of the study population, n (%).

Characteristic	Total
Age (years)
Mean (SD)	68.8 (9.2)
Median (IQR)	69.0 (63.0, 75.0)
Min, Max	44.0, 93.0
Women	45 (40.9)
Race
Pacific Islander	2 (1.8)
White	108 (98.2)
Smoking history
Current smoker	4 (3.6)
Previous smoker	36 (32.7)
Never smoker	58 (52.7)
History of heart failure	1 (0.9)
Body mass index (kg/m2)
Mean (SD)	30.3 (6.0)
Median (IQR)	30.0 (62.2, 34.3)
Min, Max	18.8, 47.0
Body surface area (m2)
Mean (SD)	2.1 (0.3)
Median (IQR)	2.1 (1.9, 2.2)
Min, Max	1.4, 2.7
Systolic blood pressure, mmHg
Mean (SD)	134.0 (18.9)
Median (IQR)	132.0 (122.0, 144.0)
Min, Max	90.0, 187.0
Diastolic blood pressure, mmHg
Mean (SD)	77.6 (8.8)
Median (IQR)	78.0 (72.0, 83.0)
Min, Max	60.0, 108.0
TJA history
Joint
Hip (total hip arthroplasty)	79 (71.8)
Knee (total knee arthroplasty)	11 (10.0)
Both	20 (18.2)
Revision surgery	59 (53.6)
Years between first TJA surgery and study visit
Mean (SD)	13.4 (5.7)
Median (IQR)	12 (10.0, 16.0)
Min, Max	2.0, 31.0

IQR: Interquartile range; TJA: Total joint arthroplasty.

Table 2 Metal concentrations by sex, n (%).

Metal concentrations (ng/mL)	Total	Men	Women
Detectable chromium,	43 (39.5)	28 (43.8)	15 (33.3)
Mean (SD)	3.8 (4.7)	3.0 (3.5)	5.2 (6.3)
Median (IQR)	1.9 (1.2, 3.8)	1.9 (1.3, 2.9)	2.8 (1.2, 5.8)
Min, max	1.1, 21.4	1.1,18.5	1.2,21.4
Detectable cobalt	51 (46.4)	36 (55.4)	15 (33.3)
Mean (SD)	4.9 (4.3)	5.2 (4.7)	4.1 (3.2)
Median (IQR)	3.7 (1.9, 6.3)	4.1 (2.1, 6.9)	3.2 (1.9, 5.2)
Min, max	1.1, 24.6	1.1, 24.6	1.1, 12
Detectable titanium	89 (80.9)	49 (75.4)	40 (88.9)
Mean (SD)	4.8 (5.1)	4.6 (6.1)	5.1 (3.5)
Median (IQR)	3.0 (2.0, 5.0)	3 (1,5)	4 (3,7)
Min, Max	1.0, 35.0	1,35	1,15

IQR: Interquartile range.

Chromium concentrations were not significantly associated with any echocardiographic findings (Table 3). Higher cobalt concentrations were associated with left and right ventricular dilatation. Higher cobalt concentrations were associated with increased left ventricular (LV) volumes as indicated with larger 2D biplane LV end-diastolic (estimate 5.09; 95%CI: 0.02-10.17), equating to a 5%-10% increase in LV volume per doubling in cobalt concentrations, and end-systolic volumes (estimate 3.13; 95%CI: 0.40-5.87) (Figure 1). For instance, patients with cobalt concentrations ≥ 4 ng/mL had an average 22.8 unit (95%CI: 7.4-38.1 P = 0.004) higher LV volumes than those with cobalt concentrations £1 ng/mL after adjusting for age and sex. This association was seen in men but not in women (Figure 2). Men had higher 2D biplane LV end-diastolic (estimate 6.86; 95%CI: 0.06-13.66; P = 0.05) and end-systolic volumes (estimate 4.58; 95%CI: 0.95-8.21; P = 0.01). Higher cobalt concentration was also associated with right ventricle enlargement in men (OR 1.94; 95%CI: 1.12-3.35; P = 0.02) but not in women (OR 0.91; 95%CI: 0.11-7.41; P = 0.93). Left atria were also larger in those with higher cobalt concentrations (estimate 4.16; 95%CI: 0.79-7.53), similarly in men and women. There was no association between cobalt concentrations and LV systolic function.

Figure 1 Distribution of two-dimensional biplane left ventricle end-diastolic volume by cobalt concentrations. 2D: Two-dimensional; LV: Left ventricle.

Figure 2 Forest plot of linear and logistic regression estimates 95%CI for the association between cobalt concentrations and echocardiographic parameters by sex. ¹Regression coefficients 95%CI retained from linear regression analyses, except the estimates for the right ventricle size (normal/enlarged) that is the odds ratio 95%CI retained from logistic regression analyses. LV: Left ventricle; RV: Right ventricle; 2D: Two-dimensional; 3D: Three-dimensional.

Table 3 Association of metal concentrations with echocardiographic parameters.

	Measure available	Mean (SD); Median (interquartile range)	Regression estimates (95%CI)1
Measure (unit)			Chromium	Cobalt	Titanium
LV size and systolic function
2D Biplane LV end-diastolic volume (mL)	100	134.04 (34.14); 129.5 (109, 153)	1.20 (-5.23 to 7.64)	5.09 (0.02-10.17)a	-3.67 (-8.55 to 1.22)
3D LV end-diastolic volume (mL)	95	130.38 (32.14); 126 (110, 147)	0.95 (-4.59 to 6.50)	4.09 (-0.57 to 8.74)a	-1.69 (-6.35 to 2.96)
2D Biplane LV end-systolic volume (mL)	99	50.76 (18.22); 48 (38, 60)	0.44 (-3.05 to 3.92)	3.13 (0.40-5.87)	-1.53 (-4.20 to 1.14)
3D LV end-systolic volume (mL)	94	49.24 (14.2); 48 (38, 57)	0.44 (-1.86 to 2.75)	1.46 (-0.50 to 3.42)	-1.16 (-3.09 to 0.76)
3D LV ejection fraction	94	62.12 (4.34); 62 (60, 65)	-0.10 (-0.88 to 0.68)	-0.51 (-1.17 to 0.15)	0.07 (-0.59 to 0.73)
LV global longitudinal strain	108	-19.74 (2.12); -20 (-21, -19)	-0.15 (-0.55 to 0.26)	-0.27 (-0.60 to 0.07)	-0.38 (-0.70 to 0.06)b
LV stroke volume (mL)	108	94.5 (17.22)/91 (82.5, 105)	-0.28 (-3.43 to 2.87)	1.59 (-0.99 to 4.16)	-2.69 (-5.24 to 0.15)b
LV cardiac output (l/minute)	109	6.12 (1.3); 5.84 (5.22, 6.8)	-0.01 (-0.35 to 0.16)	0.01 (-0.20 to 0.22)	-0.20 (-0.40 to 0.004)a
LV cardiac index (l/minute/m2)	108	2.92 (0.4)/2.82 (2.68,3.16)	0.01 (-0.08 to 0.09)	0.01 (-0.06 to 0.07)	-0.08 (-0.15 to -0.01)b
LV diastolic function
LV diastolic function (normal/abnormal)	110	Normal 42 (38.2)	Reference	Reference	Reference
		Abnormal 68 (61.8)	1.21 (0.78-1.87)	1.14 (0.80-1.62)	0.67 (0.47-0.97)b
Medial E/e' ratio	106	10.76 (3.7)/10 (8.3,12.5)	-0.38 (-1.02 to 0.27)	-0.05 (-0.60 to 0.49)	0.18 (-0.34 to 0.70)
LA biplane vol index (mL/m2)	107	31.32 (9.02); 30 (25, 37)	1.26 (-1.36 to 3.86)	2.72 (0.60-4.84)b	-0.24 ( -2.3 to 1.82)
LA reservoir strain	105	29.9 (6.9); 29.5 (26.2, 33.9)	0.002 (-1.38 to 1.39)	0.083 (-1.06 to 1.22)	1.602 (0.55-2.66)c
Right heart parameters
RV systolic pressure (mmHg)	80	28.34 (5.34); 28 (23.5, 32)	-0.41 (-1.54 to 0.73)	-0.22 (-1.19 to 0.75)	-0.28 (-1.15 to 0.58)
RV size (normal/enlarged)	110	Normal 99 (90)	Reference	Reference	Reference
		Abnormal 11 (10)	1.66 (0.91-3.05)	1.80 (1.08-2.99)b	1.22 (0.74-2.01)
RV basal diameter (mm)	53	40.48 (6.3); 41 (36, 46)	0.37 (-1.18 to 1.92)	1.17 (-0.18 to 2.51)a	-0.17 (-1.42 to 1.08)
Tricuspid annular plane systolic excursion (mm)	91	23.58 (3.7); 23 (21, 26)	-0.34 (-1.21 to 0.53)	0.11 (-0.61 to 0.83)	0.17 (-0.50 to 0.84)
RV Basal Lateral Systolic Strain	34	-25.06 (4.88); -26 (-29, -24)	1.19 (-0.64 to 3.01)	0.40 (-1.50 to 2.29)	-0.59 (-2.12 to 0.94)
Other parameters
Mid ascending thoracic aortic diameter (mm)	82	35.9 (4.08); 36 (33, 38)	0.16 (-0.65 to 0.98)	0.43 (-0.27 to 1.13)	-0.71 (-1.38 to -0.05)b
N terminal pro b-type natriuretic peptide (pg/mL)	109	153.4 (269.3); 85 (42, 138)	-20.47 (-72.32 to 31.38)	10.01 (-32.78 to 52.80)	13.41 (-27.74 to 54.55)
Troponin (ng/mL)	108	13.4 (11.3); 11 (7, 15)	-0.97 (-2.94 to 0.99)	-0.85 (-2.47 to 0.77)	0.22 (-1.34 to 1.78)

¹Estimates from age and sex-adjusted linear regression models indicate the β coefficients as the mean change in echocardiographic measures or cardiac markers per log-unit increase in metal concentrations. Odds ratio estimates from age and sex-adjusted logistic regression models indicate the odds of abnormal findings per log-unit increase in metal concentrations.

^a0.05 ≤ P < 0.1;

^b0.01 ≤ P < 0.05;

^cP < 0.01. LA: left atrium; LV: Left ventricular; RV: right ventricle.

Higher titanium concentrations were associated with measures of reduced LV systolic function as indicated by a reduction in LV global longitudinal systolic strain (estimate 0.38; 95%CI: -0.70 to -0.06) which is a sensitive measure of left ventricle contraction. For instance, patients with titanium concentrations ≥ 3 ng/mL had an average 1.3 (95%CI: 2.2-0.4, P = 0.005) reduction in LV global longitudinal strain than those with titanium concentrations £1 ng/mL after adjusting for age and sex. LV stroke volume, cardiac output, and cardiac index were similarly lower in patients with high titanium concentrations (Table 3). For instance, the normal range for cardiac index is 2.5-4.2 L/min/m². A 0.08 unit decrease in cardiac index equates from 1.9% to 3.2% decrease from normal cardiac function when titanium concentration was doubled (e.g. from 1 ng/mL to 2 ng/mL). On the contrary, higher titanium concentrations were associated with better left ventricle diastolic function (OR 0.67; 95%CI: 0.47-0.97) and higher left atrial reservoir strain (OR 1.60; 95%CI: 0.55-2.66). No significant associations were observed between metal concentrations and NT-proBNP and troponin levels.

DISCUSSION

There is concern regarding the systemic effects of increased metal concentrations in TJA patients. However, evidence has been limited to anecdotal case reports or large registry-based studies with limited data on clinical findings or echocardiographic abnormalities. In this study, we investigated the association of systemic chromium, cobalt, and titanium concentrations with echocardiographic measures of cardiac structure and function and cardiac labs troponin and NT-proBNP. Overall, we found that cobalt was associated with both left and right ventricular dilation in men and titanium was associated with significant changes in LV systolic function (i.e., cardiac index in Table 3). No significant associations were observed for chromium.

Besides case reports and registry-based studies, three previous studies are similar to ours and examined the correlation between metal concentrations and echocardiographic findings. Prentice et al[15] compared 35 MoM hip resurfacing patients with 35 matched asymptomatic THA patients with conventional implants. Patients with MoM implants had higher systemic cobalt and chromium concentrations (1.75 µg/L vs 0.38 µg/L and 1.27 µg/L vs < 0.30 µg/L for chromium), but relatively minor differences in cardiac ejection fraction and LV end-diastolic diameter.

In another study, 90 THA patients were split into three groups based on bearing type and blood metal concentrations, and cardiac function was evaluated using both cardiac magnetic resonance imaging and echocardiography[13]. There was no evidence for an association between metal concentrations and cardiac structure or function. Despite the large sample size similar to ours and 3 groups of THA patients with a wide spectrum of chromium and cobalt concentrations, linear associations between metal concentrations and LV structural measures were not reported. Splitting patients into 3 separate groups may have resulted in loss of power. Finally, 75 THA patients with the same modular implants (49 MoM bearings) were evaluated at two visits for cobalt, chromium, and titanium concentrations and an echocardiogram at the second visit[14]. Cobalt and chromium concentrations were higher in MoM patients, but titanium concentrations and cobalt-chromium ratios were not different between groups. Cobalt and chromium concentrations increased between time-points in the MoM patients with significantly lower global longitudinal strain than compared with the non-MoM group. Cobalt concentrations were associated with tricuspid annular plane systolic excursion, an indicator of right ventricular function. We similarly observed several associations between cobalt and titanium concentrations and cardiac function measures. Consistent with the previous study, we found an association between cobalt and right ventricular size as well as an association with LV volume, not described in earlier studies.

Cobalt and chromium concentrations were undetectable in about half of our patients and were not associated with echocardiographic differences in women. Yet, a relationship between LV dilation was found with cobalt concentrations in men. Furthermore, a doubling of cobalt levels in men was associated with a 1.9-fold increase in the prevalence if right ventricular enlargement. Prior studies have suggested a link between cobalt deposition and cardiac toxicity potentially related to intracellular calcium mishandling[7]. Here a tendency for abnormal ventricular dilation is a well-recognized finding that precedes the development of heart failure[20].

Titanium was the most commonly elevated metal in our study and was associated with several measures of decreased LV systolic function but not diastolic function. Better measurement of small concentrations in cobalt and chromium may allow detection of associations between these metals and heart function measures. Compared with cobalt and chromium, relatively little is known about the potential systemic effects of titanium. Evidence to date comes only from laboratory studies. For example, in an experimental laboratory study in hypertensive rats, intra-tracheal administration of titanium resulted in irreversible hemodynamic impairment and cardiac structural damage[21].

Our findings should be interpreted considering potential limitations. Due to the small sample size, we were unable to perform subgroup analyses by number of surgeries, surgery/implant types, potential reasons for high metal ions (e.g. trunnionosis) or among subgroups of patients such as those with smoking history, other lifestyle factors or selected comorbidities like heart failure. Adverse cardiac effects of high metal concentrations may vary in different ethnic and racial groups. With mostly Caucasian white patients, the generalizability our findings is limited. Our power calculations prior to the study indicated that with a sample size of 100 patients, the linear regression test of correlation = 0 (P = 0.05 two-sided) between the cardiac outcomes and metal concentrations would have 80% power to detect a correlation of 0.27 (with a corresponding r-square of 0.073). The main exposure variable in this study was the metal concentrations which were examined as a log-transformed continues exposure variable. Dichotomizing or categorization of continuous exposure variables for analysis is inefficient and misleading[22,23]. Importantly, we examined cross-sectional associations and observed some significant correlations with cobalt and titanium concentrations, but the magnitude was small. Therefore, these significant findings should be interpreted with caution and bearing in mind the evidence from other lines of research. Ideally, information on longitudinal changes in systemic metal concentrations and cardiac measures would be needed to further interpret causal implications of these cross-sectional findings. Furthermore, in addition to TJA implants, other environmental exposures, such as dental implants and food additives, nutritional status or alcohol intake can contribute to higher systemic levels[24,25]. Strengths include the use of 3D and speckle-tracking echocardiography, which provide much more sensitive markers of cardiac dysfunction than the traditional 2D echocardiography.

CONCLUSION

In conclusion, elevated cobalt and titanium concentrations may be associated with structural and functional cardiac changes in some patients. Given the high prevalence of TJA and limited evidence on the potential long-term cardiotoxicity of metal implants, more research is warranted to better understand how the systemic distribution of metal ions may affect cardiac function and the temporal relationship and quantitative thresholds for metal concentrations and cardiac pathology.