Self-reported systemic complaints in patients with metal-on-metal hip arthroplasty

Abstract

Introduction

An increase in blood serum metal ion levels is seen after implantation of all metal-on-metal (MoM) hip prosthesis. Systemic complaints contributed to raised cobalt ion concentrations in patients with MoM arthroplasty may lead to a variety of symptoms. The aim of this study is to investigate self-reported systemic complaints in association with cobalt ion concentrations in patients with any type of MoM hip prosthesis.

Methods

A cohort study was conducted. Patients with both unilateral and bilateral, resurfacing and large head metal on metal total hip arthroplasties (LHMoM THA) were included. Cobalt ion concentrations were measured by inductively coupled plasma mass spectrometry. Based on the known cobalt toxicity symptoms of case-reports and toxicology reports a new non-validated questionnaire was developed. Analysis was performed on two groups; a low cobalt ion concentration group and a high cobalt ion concentration group.

Results

A total of 62 patients were included with a mean age at surgery of 60.8 ± 9.3 years and a mean follow up of 6.3 ± 1.4 years. Mean cobalt ion concentrations were 104 ± 141 nmol/L (9–833). Based on the different thresholds (120–170 or 220 nmol/L) the low cobalt ion concentration group consisted of 44 (71%), 51 (82%) or 55 (89%) subjects respectively. In the 120 nmol/L and 170 nmol/L thresholds a significant difference in age was found. The composite score for OVS increased from 54% to 57%–68% with rising threshold value, a hint at the correlation between ion concentration and symptom prevalence

Discussion

Ocular-vestibular symptoms were more common in high cobalt ion concentration groups for the three threshold levels tested and with increasing prevalence for higher threshold values. With regards to proactively inquired, self-reported symptoms the threshold where effects may be present could be lower than values currently applied in clinical follow-up

1. Introduction

An increase in blood serum metal ion levels is seen after implantation of all MoM hip prosthesis due to release from the metal surface directly, more so during articulation and corrosion of the bearing surfaces. Peak concentrations are reached at 9–12 months postoperative, hereafter an equilibrium is attained between metal ion release from the implant and metal ion discharge from the body (e.g. via urine).¹ Design of the prosthesis, component positioning, contact patch to rim (CPR) distance, smaller joint sizes and renal impairment are factors which might influence metal ion concentrations after this period.^2,3 The continuous exposure to metal ions can result in soft tissue reactions and raised metal ion concentrations in blood, serum and urine.^4,5 The bearing surfaces in MoM prosthesis consist of cobalt, chromium and molybdenum. One of the most used hip resurfacing systems, Birmingham Hip Resurfacing (BHR), consists of ±60% cobalt, ±30% chromium, ±5% molybdenum and ±1% of nickel, manganese, carbon, silicon and iron.⁶ The toxicity of the trivalent (III) form of chromium is low, as opposed to the hexavalent (VI) form.⁷ The chromium ions released by MoM prosthesis are of the trivalent (III) form, and therefore non-hazardous.⁸ This in contrast to released cobalt ions. Cobalt intoxication or cobaltism is mostly known from industrial exposure, medicinal use and beer additives.⁹ This was further substantiated by a systematic review showing a correlation between symptom severity and cobalt ion concentrations but not with chromium ion concentrations.¹⁰ Several case-reports of cobalt toxicity after hip arthroplasty have been published in the last decade, even fatal cases.^11,12 Systemic complaints in patients with MoM arthroplasty is also known as ‘Prosthetic Hip-Associated Cobalt Toxicity’ (PHACT) and may lead to a variety of symptoms: neuro-ocular toxicity (tinnitus, vertigo, deafness, blindness, convulsions, headaches and peripheral neuropathy), cardiotoxicity and thyroid toxicity. Nausea, anorexia and unexplained weight loss have been described.13, 14, 15, 16 It was shown that subjects with well functioning MoM hips have more structural and functional differences in organs (heart, liver, spleen) and a changed structure of the visual pathways and the basal ganglia compared with a conventional total hip control group.17, 18, 19 Thus, systemic effects from metal ions even with well functioning implants or with ion concentrations lower than those associated with known adverse effects may exist and warrant investigation. However, little is known in current literature about the clinical interpretation of raised metal ion concentrations and there potential systemic effects.¹⁵ One can expect a relevant difference between complaints pro-actively self-reported or complaints specifically investigated. In the first case, patients may not report because these events are not so common, not so strong or not associated with the procedure to become reported, although they may exist and indicate an effect. The aim of this study is to investigate self-reported systemic complaints in association with cobalt ion concentrations in patients with any type of MoM hip prosthesis.

2. Materials and methods

A cohort study was conducted between October 2012 and December 2016. The cohort was initially selected for another study, the methods are described in detail elsewhere.²⁰ Patients with both unilateral and bilateral, resurfacing and large head metal on metal total hip arthroplasties (LHMoM THA) were included for the current study. Systemic symptoms were not a presenting complaint in any of these subjects.

Cobalt and chromium ion concentrations were measured by inductively coupled plasma mass spectrometry (ICP-MS) using the NexION® 300X ICP-MS (PerkingElmer, Waltham, Massachusetts, United States) at Ziekenhuis Groep Twente, The Netherlands. The first 5 mL of each vena punction was disposed, hereafter the blood was collected in a tube (BD Vacutainer® Trace Element, Franklin Lakes, New Jersey, United States).²¹ The blood was shipped within a week from collection and until then stored in a refrigerator at a temperature of 2°–8 °C. Concentrations are described in nmol/L. Conversion from nmol/L to ug/L (ppm) is by multiplication with 0.052 for chromium and 0.059 for cobalt.

Based on the known cobalt toxicity symptoms of case-reports and toxicology reports a new non-validated questionnaire was developed. The subjects were asked to answer 17 questions on general health since placing of the MoM hip prosthesis. The questions were subdivided in general questions/symptoms, ocular-vestibular symptoms (OVS), neurological symptoms, emotional health and cardio- and thyroid toxicity symptoms. The answer options were yes and sometimes (score 1) and no (score 0). Combined scores were obtained per group and a combined neuro-ocular-vestibular composite score (NOVCS). The Hip disability and Osteoarthritis Outcome Score Physical function Shortform (HOOS-PS) was obtained and scored from no difficulty (0) to extreme difficulty (100).²²

This study was performed in compliance with the 1975 Declaration of Helsinki, as revised in 2000, was studied and approved by the IRB (METC Zuyd, Heerlen, The Netherlands, IRB Nr. 10N72) and conducted in accordance with the guidelines for Good Clinical Practice (GCP).

2.1. Statistical analysis

Statistical analysis was performed using IBM SPSS Statistics 22 (International Business Machines Corporation, Armonk, New York, United States). Independent samples T test and Fisher's Exact Test were used. Analysis was performed on two groups; a low cobalt ion concentration group and a high cobalt ion concentration group. The upper limit of well functioning prosthesis was set at 170 nmol/L (10 μg/L), as stated by the Dutch orthopedic society (Nederlandse Orthopaedische Vereniging (NOV)) and described by Verhaar.²³ We also performed analysis with two arbitrarily chosen thresholds, a lower limit of 120 nmol/L and a higher one of 220 nmol/L to investigate whether for systemic effects another concentration level may be better suited than one which is established for known adverse effects. Results are shown as means ± standard deviation and [range] or (percentage) with p < .05 as significance level.

3. Results

A total of 62 patients (58% men) with 71 prosthesis were included with a mean age at surgery of 60.8 ± 9.3 years (41.6–78.1) and a mean follow up of 6.3 ± 1.4 years (3.7–9.6). There were 44 resurfacing and 27 LHMoM THA. The mean interval between the metal ion concentration measurement and fulfilling of the questionnaires was 1.4 ± 0.8 years (0.2–2.5). Mean cobalt and chromium ion concentrations were 104 ± 141 nmol/L (9–833) and 95 ± 130 nmol/L (6–592), respectively (Table 1).

Table 1.

Baseline characteristics.

Gender	M 36 (58%)/F26 (42%)
Age at OR (years)	60.8 ± 9.3 [41.6–78.1]
Follow-up (years)	6.3 ± 1.4 [3.7–9.6]
Resurfacing (number of)	44
LHMoM (number of)	27
Unilateral (number of)	53
Bilateral (number of)	9
Interval (years)a	1.4 ± 0.8 [0.2–2.5]
Cobalt (nmol/L)	104 ± 141 [9–833]
Chromium (nmol/L)	95 ± 130 [6–592]
HOOS-PS(0–100)	18.9 ± 17.9 (0–67.9)

^aInterval between measuring of metal ion concentrations and completing questionnaires.

Based on the different thresholds (120–170 or 220 nmol/L) the low cobalt ion concentration group consisted of 44 (71%), 51 (82%) or 55 (89%) subjects respectively. In the 120 nmol/L and 170 nmol/L thresholds a significant difference in age was found. The composite score for OVS increased from 54% to 57%–68% with rising threshold value, a hint at the correlation between ion concentration and symptom prevalence. Also, the NOVCS of the high cobalt ion concentration group in the 220 nmol/L threshold was significantly higher (p = .047). A composite score for cardio- and thyroid toxicity showed no significant differences between the groups (Table 2).

Table 2.

Occurrence of self-reported symptoms divided by different threshold.

	Occurrence at threshold 120 nmol/L			Occurrence at threshold 170 nmol/L			Occurrence at threshold 220 nmol/L
	<120	>120	p	<170	>170	p	<220	>220	p
Baseline
Number of subjects	44 (71%)	18 (19%)		51 (82%)	11 (18%)		55 (89%)	7 (11%)
Age at OR (years)	57.4 ± 7.4 [41.6–73.8]	68.4 ± 7.7 [55.2–78.1]	<.001	59.4 ± 8.7 [41.6–78.1]	65.9 ± 8.7 [55.2–77.7]	.029	60.2 ± 9.1 [41.6–78.1]	63.4 ± 8.5 [55.2–77.0]	.382
Follow-up (years)	6.7 ± 1.5 [3.7–9.6]	5.6 ± 1.1 [4.0–8.2]	.004	6.5 ± 1.5 [3.7–9.6]	5.9 ± 1.1 [4.8–8.2]	.239	6.4 ± 1.5 [3.7–9.6]	6.2 ± 1.3 [4.9–8.2]	.824
Cobalt (nmol/L)	38 ± 24 [9–100]	258 ± 178 [122–833]	<.001	52 ± 41 [9–158]	334 ± 193 [170–833]	.001	62 ± 54 [9–207]	418 ± 199 [266–833]	.003
HOOS-PS (0–100)	18.9 ± 19.4 [0–67.9]	17.9 ± 14.3 [0–46.1]	.846	19.2 ± 18.5 [0–67.9]	16.3 ± 15.1 [0–46]	.628	18.2 ± 18.2 [0–67.9]	21.8 ± 16.3 [4.6–46.1]	.621
General
Known thyroid disease	11%	6%	.662	10%	9%	1.000	11%	0%	1.000
Known CNS disease1	9%	11%	1.000	10%	9%	1.000	9%	14%	.528
Headache	36%	28%	.569	35%	27%	.735	35%	29%	1.000
Ocular-Vestibular symptoms
Vision/Eyes	52%	61%	.058	51%	73%	.317	53%	71%	.442
Hearing	34%	56%	.016*	39%	45%	.744	38%	57%	.425
Tinnitus	28%	39%	.545	28%	45%	.294	28%	57%	.189
Dizziness	36%	61%	.095	39%	64%	.186	40%	71%	.223
Composite	38%	54%	.023*	39%	57%	.043*	40%	68%	.008*
Neurological symptoms
Tremor	14%	29%	.263	16%	27%	.400	17%	29%	.599
Paresthesias phalanges	30%	18%	.518	29%	18%	.710	28%	14%	.663
Deafness phalanges	16%	24%	.481	16%	27%	.400	17%	29%	.599
Composite	20%	24%	.685	20%	24%	.638	20%	24%	.776
Emotional Health
Fear	16%	11%	1.000	16%	9%	1.000	15%	14%	1.000
Irritation	32%	44%	.390	37%	27%	.732	35%	43%	.690
Depression	11%	11%	1.000	10%	18%	.597	11%	14%	1.000
Composite	20%	22%	.693	21%	18%	.816	20%	24%	.774
Combined Neuro-Ocular-Vestibular Composite Score	28%	29%	.707	29%	38%	.228	29%	40%	.047*
Cardio- and Thyroid symptoms
Weight loss	12%	28%	.143	14%	27%	.367	17%	14%	1.000
Tiredness	68%	61%	.768	67%	64%	1.000	64%	86%	.406
Weakening	41%	33%	.775	37%	45%	.736	36%	57%	.186
Palpitations	20%	22%	1.000	24%	9%	.431	22%	14%	1.000
Composite Score	47%	36%	.140	35%	36%	1.000	35%	43%	.408

* Significant at p = .005.

4. Discussion

This study aimed to investigate if cobalt ion concentrations commonly observed with MoM hip arthroplasty may result in more frequent self-reported systemic complaints above various cobalt ion level thresholds. The main finding of this study was that OVS are more common in patients with high cobalt ion concentrations, independently of the thresholds investigated (120–170 or 220 nmol/l). In accordance, the composite score of OVS was significantly higher in the high concentration group for all three threshold values tested and the score increased with rising the threshold value.

Subjects in the current study reported tinnitus (31% overall) and hearing loss (40% overall). This is in agreement with Leikin et al. who reported tinnitus and/or hearing loss in 7 of their 26 (27%) subjects with MoM arthroplasty with a mean age of 55.6 years and median cobalt ion concentration of 14.1 μg/L (239 nmol/L). Patients were referred by self-referral, primary care physicians, orthopedic physicians, poison centers and hip-arthroplasty manufacturers. Identification of symptoms was performed by the use of a clinic standard medical toxicology health history form.¹⁵ Also Prentice et al. reported similar self-reported tinnitus (21–33%) and hearing problems (33%) in patients with MoM arthroplasty.²⁴ The average number of patients in the general population with tinnitus is far lower (14%).¹⁵ Rizzeti et al. and Bradberry et al. described tinnitus and hearing loss as symptoms of cobalt toxicity.^12,13 This is further substantiated by the fact that the current study reports more subjects with tinnitus in the high cobalt ion concentration group at higher thresholds and with increasing threshold values. In contrast to these results, van Lingen et al. did not find a correlation between self-reported neurotoxic complaints and cobalt ion concentrations.²⁵ Only 10% of subjects were found to have cobalt ion concentrations >170 nmol/L, compared with 18% in the current study. This might have resulted in an underpowered study to detect neurotoxic effects. Also, van Lingen et al. used the NCS-60 questionnaire which is not designed and not validated for cobalt toxicity. The NCS-60 is a validated Dutch questionnaire mainly used to determine exposure to neurotoxic compounds in workers and in the evaluation of suspected neurotoxicity.²⁶ The current study used a questionnaire which has not been validated but was designed to specifically detect symptoms of cobalt toxicity, as opposed to the generic NCS-60. Prentice et al. could not associate slightly elevated cobalt ion concentrations (<10 nmol/L), with clinically demonstrable visual or auditory dysfunction.²⁴ The cobalt ion concentrations in the cohort of Prentice et al.might be too low to generate visual or auditory dysfunction.

The clinically best-to-use cobalt ion concentration threshold is debatable, for well functioning hips, for the likely existence of pseudotumors and for possible cobalt toxicity. Tower advised neurological and cardiac assessment for patients with a cobalt concentration ≥7 μg/L (119 nmol/L).²⁷ In the 18 cobalt toxicity case reports reported by Bradberry et al. subjects showed a mean cobalt ion concentration of 398 μg/L (6754 nmol/L)[13.6–6521 μg/L].¹³ Those patients with a failed ceramic prosthesis, revised with a metal-containing prosthesis, showed blood cobalt concentrations of 506 μg/L (8587 nmol/L) [353–6521 μg/L]. Patients with a primary MoM prosthesis, like all subjects in the current study, showed a blood cobalt concentration of 34.5 μg/L (586 nmol/L)[13.6–398.6 μg/L]. In constrast, a review of Paustenbach et al. concluded that significant systemic complaints of cobaltism will not occur when concentrations of cobalt ions stay below 300 μg/L (5091 nmol/L).²⁸ Leikin et al. and Ho et al. could not correlate cobalt and chromium ion concentrations with the incidence of systemic symptoms.^15,16 The study by Ho et al. is a follow-up study of the aforementioned Leikin et al., a part retrospective and part prospective review of patients referred to toxicology clinics in London, UK and in the USA recorded in the Toxicology Investigators Consortium (ToxIC) Registry from June 2011 to June 2015.

Also the current study has some limitations. The study is uncontrolled to the extent, that there is no matched group with a non-MoM bearing or a matched group without a THA to serve as a comparator in a general or similar population. However, comparison is made between the high and low ion concentration groups for three threshold values. Patients were informed on the cobalt ion concentrations which could potentially affect the reported symptoms. A patient reported outcome measure (PROM) asking for complaints may bias patients towards reporting something for which they only generate awareness via the questionnaire but this new approach also differentiates this study from previous reports relying on generic PROM's or proactively reported complaints. The aim was to correlate symptoms with cobalt ion concentrations, presumably due to a MoM prosthesis, but it might be possible that other factors influenced cobalt ion concentrations or self-reported symptoms. Patient related factors such as allergies, diabetes, diet, medication and occupation, were not (fully) taken into account. Toxicity of cobalt is related to the unbound (free) form of cobalt (Co²⁺), whereas most cobalt is bound to albumin. The refined biokinetic model of Unice et al. states that the concentration of unbound cobalt in the circulation raises and toxicity occurs at lower concentrations in susceptible subjects and in certain conditions such as kidney failure, iron deficiencies, sepsis, malnutrition and use of medication.²⁹ The used questionnaire was not formally validated but it was purely designed for the detection of symptoms of cobalt toxicity. It was not the intention of this study to create a new and validated questionnaire, but to study a possible effect yet unnoticed due to the lack of validated tools which are targeted and not generic.

The interval between obtaining the questionnaire and the metal ion concentrations was rather large with 1.4 ± 0.8 years (0.2–2.5). We aimed at measuring the steady-state concentrations so that the time difference between ion level assessment and PROM shall be less critical, than during the transient phase. The current study also did not measure renal functions, but severe renal insuffiencies can be ruled out and only milder forms which have produced no symptoms or have not yet been diagnosed may exist. Cobalt and chromium ions are rapidly excreted by the kidneys: 80–90% of cobalt ions within days and 60% of chromium ions within 8 h.^3,5 Although mild variations in metal ion concentrations might exist, the current study was designed to measure metal ion concentrations at a steady-state after the running-in period of 9–18 months.¹ Also, Lainala et al. showed no association between metal ion concentrations and mild or moderate renal insufficiency.³⁰

A significant difference in age between the groups was found for the 120 nmol/L and 170 nmol/L thresholds. This might have influenced the reported symptoms. Normal aging, with reduction in vision and hearing for example, might be confused with symptoms caused by increased cobalt ion concentrations and thus may have influenced the reported symptoms. The prevalence of hearing impairment for males in Dutch general practice is 3.9% in agegroup 55–59 years, 6.3% in agegroup 60–64 years, 9.7% for agegroup 65–69 years and 13.4% for agegroup 70–74. For females this is 3.6%, 5.1%, 6.9% and 9.0% respectively.³¹ Hearing impairment is thus more frequent with increasing age, but does not fully explain the differences found in OVS in the current study, although these specific confounding factors (e.g. age) was not controlled for.

Further research should focus on a larger cohort of MoM arthroplasties and investigate prevalence but also include symptom severity and frequency. The questionnaire used should be optimalized and focus on the main outcome of the current study. To optimize associations a shorter period of time between metal ion assessment and PROMs should be guaranteed. A long term follow-up like in this study must be considered as some systemic effect, e.g. depression, may need more time to manifest themselves.

This study aimed to detect a trend in self-reported systemic complaints in patients with metal-on-metal hip arthroplasty due to raised cobalt ion concentrations. Ocular-vestibular symptoms were more common in high cobalt ion concentration groups for the three threshold levels tested and with increasing prevalence for higher threshold values. With regards to proactively inquired, self-reported symptoms the threshold where effects may be present could be lower than values currently applied in clinical follow-up. It is unknown how exposure to elevated metal ion concentrations for a longer period of time affects the health of aging subjects. Further research with a larger cohort and more standardized questionnaire is necessary to uncover previously undiscovered or under-reported effects warranting investigation.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

Submission declaration

The authors declare is that this manuscript is not under consideration for publication elsewhere, and publication is approved by all authors.

Author contribution

JJ: Conceptualization, Methodology, Formal analysis, Writing - original draft.

MS: Methodology, Formal analysis, Writing - review & editing.

HK: Writing - review & editing.

BG: Conceptualization, Formal analysis, Writing - review & editing.

IH: Supervision, Writing - review & editing.

Declaration of competing interest

The authors declare that there is no conflict of interest.