Introduction
The number of individuals aged older than 65 years is increasing rapidly, and ageing is associated with alterations in numerous physiological functions. Poor nutritional status is one of the main risk factors for frailty. This geriatric syndrome is associated with alterations in multiple physiological functions and reduced functional reserves( Reference Bonnefoy, Berrut and Lesourd 1 ), in which sarcopenia, the age-related loss of skeletal muscle mass and strength, is considered as a key component of frailty( Reference Clegg, Young and Iliffe 2 ). The decline in nutritional status, functional ability and the increased risk of falls in older adults have been associated with the occurrence of poly-pharmacy which increases with age( Reference Jyrkka, Enlund and Lavikainen 3 , Reference Maher, Hanlon and Hajjar 4 ). However, a causal relationship has not been clearly established, partly because (i) some diseases by themselves promote malnutrition( Reference Jyrkka, Mursu and Enlund 5 ) and (ii) the age-associated anorexia could also be a confounding factor. Indeed, reduced food intake can result not only from medication but also from multiple factors including declines in physiological functions (smell and taste, central and peripheral drive to eat, gastric emptying), social factors (poverty, loneliness) and pathological conditions (depression, dementia, somatic diseases, oral-health status)( Reference Bailly, Maitre and Van Wymelbeke 6 , Reference Wysokiński, Sobów and Kłoszewska 7 ). It is nevertheless well documented that many drugs have unpleasant tastes or odours of their own and can further alter the sensory perceptions of dietary products through various mechanisms involving peripheral receptors, chemosensory neural pathways, brainstem and brain or hypo-salivary side effects leading to a poor oral cavity health( Reference Schiffman and Zervakis 8 ). These aspects will not be developed in the present review but we will focus on other types of drug–food interactions that could affect intestinal transport and metabolism, systemic transport or tissue/cellular distribution of nutrients( Reference Genser 9 ). A typical example of drug action on a nutrient bioavailability is paracetamol (acetaminophen; APAP) that interferes with cysteine (Cys).
APAP is the most frequently used analgesic in older adults and is the first-line treatment of a large variety of pain such as headache, muscle pain and chronic pains such as lower back pain and arthritis( Reference Airaksinen, Brox and Cedraschi 10 – Reference Makris, Abrams and Gurland 12 ). APAP is usually considered to be safe when administered within the therapeutic range, but in over-dosages it can cause severe toxicity in the liver and more rarely in the kidneys( Reference Jaeschke 13 , Reference Chen, Lin and Dai 14 ). A recent meta-analysis including eight observational studies highlighted increases in cardiovascular and gastrointestinal disorders and mortality with regular intake of high therapeutic doses( Reference Roberts, Delgado Nunes and Buckner 15 ). In connection with nutrition, the detoxification of APAP requires Cys( Reference Forrest, Clements and Prescott 16 , Reference Hodgman and Garrard 17 ), a conditionally indispensable amino acid (IAA)( Reference Obled, Papet and Breuillé 18 ). Cys deficiency has already been shown to be partly responsible for the decreased health and quality of life with ageing( Reference Dröge 19 ). Indeed, Cys deficiency can deplete and oxidise Cys and glutathione (GSH) pools, which promote the progressive increase in ageing-related oxidative stress associated with various chronic diseases and sarcopenia( Reference Jones 20 – Reference Meng and Yu 22 ). So, it is relevant to question whether APAP treatment could decrease Cys availability necessary to sustain protein reserves, notably in skeletal muscle, and consequently worsen sarcopenia and frailty in older adults. The purpose of the present review was to provide bibliographic data, and evidence leading to our hypothesis; then to discuss recent results concerning the effect of APAP treatment on muscle mass and metabolism.
Sarcopenia: definition, prevalence and mechanisms
The term sarcopenia was first proposed by Rosenberg in 1989( Reference Rosenberg 23 ). It comes from the Greek sarx ‘flesh’ and penia ‘poverty’. Nowadays, sarcopenia is defined by a loss of muscle mass, muscle strength and muscle quality( Reference Cruz-Jentoft, Baeyens and Bauer 24 – Reference Fielding, Vellas and Evans 27 ). It is closely correlated with morbidity and increased mortality( Reference Fearon, Evans and Anker 28 ). In 2015, one in eight individuals was aged 60 years and over and they will represent more than one in five individuals in 2050( 29 ). Inevitably with this increase in the ageing population, the prevalence of sarcopenia will increase too and is estimated that it will affect more than 200 million individuals in 2050( Reference Janssen 30 ). Currently, 25–50 % of the elderly aged 65 years and older are probably sarcopenic( Reference Janssen, Shepard and Katzmarzyk 31 ). In the USA, the cost of sarcopenia has been evaluated to be about $18·4 billion( Reference Janssen, Shepard and Katzmarzyk 31 ). Sarcopenia is now a major cost in terms of healthcare. It is an important component of the frailty syndrome( Reference Clegg, Young and Iliffe 2 ) and could lead to dependency( Reference Cooper, Dere and Evans 32 , Reference Cesari, Landi and Vellas 33 ). Thus, the prevention of sarcopenia is of prime importance.
Sarcopenia is a multifactorial event affected by intrinsic (for example, age, hormonal change) and extrinsic (for example, disease, nutrition) factors( Reference Walrand, Guillet and Salles 34 ). Muscle loss arises from an imbalance between proteolysis and protein synthesis. Various mechanisms are involved in sarcopenia related to either a decrease in the availability of amino acids (AA) for muscle or an intrinsic impairment of muscle protein metabolism (Fig. 1). In the following, only factors affecting AA availability will be described.
The first major determinant of AA availability for muscle is dietary protein intake. Ageing can be associated with the reduction of nutrient intake, known as ‘anorexia of ageing’, representing a physiological feature of old age associated with decreased energy expenditure and loss of muscle mass( Reference Landi, Calvani and Tosato 35 ). Epidemiological data have shown that 22–41 % of individuals older than 50 years did not reach the RDA for proteins of 0·8 g/kg body weight per d( Reference Calvani, Miccheli and Landi 36 ). The ingestion of a small bolus of IAA (7 g) has been shown to induce a lower muscle protein accretion in the elderly than in young individuals( Reference Katsanos, Kobayashi and Sheffield-Moore 37 ). Furthermore, the quality of ingested protein is of prime importance and is closely linked to its composition of IAA. In fact, IAA are primary stimuli of protein synthesis and some of them have a specific role in muscular anabolism, such as branched-chain AA. This is the case for leucine, nowadays recognised as a regulator of protein renewal, particularly by its action on protein synthesis( Reference Anthony, Yoshizawa and Anthony 38 – Reference Rieu, Balage and Sornet 45 ) and protein degradation( Reference Nakashima, Ishida and Yamazaki 46 , Reference Combaret, Dardevet and Rieu 47 ). So it appears of prime importance to recommend the consumption of high-quality proteins, i.e. containing a large proportion of IAA( Reference Dardevet, Rieu and Fafournoux 48 – Reference Paddon-Jones, Sheffield-Moore and Katsanos 50 ). Currently many authors agree on this point and recommend that older adults consume 1–1·5 g proteins/kg per d( Reference Morais, Chevalier and Gougeon 51 – Reference Deutz, Bauer and Barazzoni 56 ) rather than the 0·8 g proteins/kg per d usually recommended for adults whatever their age( Reference Wolfe 57 , Reference Volpi, Campbell and Dwyer 58 ).
Not only the quantity and the quality of ingested proteins are of importance, but also the repartition of the dietary protein intake in the course of the day. Indeed, protein pulse feeding (80 % of daily proteins concentrated in one meal) has been shown to be more efficient in stimulating N balance and protein synthesis( Reference Arnal, Mosoni and Boirie 59 , Reference Arnal, Mosoni and Dardevet 60 ). These results have recently been confirmed in undernourished elderly individuals( Reference Bouillanne, Curis and Hamon-Vilcot 61 ), for whom protein pulse feeding allowed protein mass gain. The efficacy of protein pulse feeding can be explained by the concept of the anabolic threshold( Reference Dardevet, Remond and Peyron 49 ). This threshold represents the minimum quantity of AA necessary to trigger protein synthesis stimulation. When compared with healthy young adults, in the elderly, this threshold is thought to be higher, leading to a loss of protein synthesis stimulation for a same amount of ingested proteins. Many studies have indicated that 25 to 30 g of high-quality protein is necessary to pass the anabolic threshold( Reference Paddon-Jones and Leidy 62 – Reference Deer and Volpi 64 ). The increase in the anabolic threshold with ageing could be explained by many factors associated with ageing such as inflammation, insulin resistance or oxidative stress that have all been shown to lower the muscle anabolic response following food intake( Reference Dardevet, Remond and Peyron 49 , Reference Boirie 65 ).
This concept is in accordance with studies performed on the speed of protein digestion, a feature that determines the post-ingestion kinetics of plasma aminoacidaemia. The post-ingestion peak of plasma aminoacidaemia occurs earlier and is higher with fast-digestible proteins than slow-digestible proteins. In young adults, slow digestible proteins were more effective than fast proteins in stimulating postprandial anabolism( Reference Mosoni and Patureau Mirand 66 ). In contrast, in the elderly, fast-digested proteins were more effective( Reference Dangin, Boirie and Garcia-Rodenas 67 – Reference Pennings, Boirie and Senden 69 ). Indeed, fast-digestible proteins generated a hyperaminoacidaemia that exceeded the anabolic threshold and then allowed a muscle anabolic response to meals in the elderly.
Another important determinant of AA availability for muscle is the first-pass extraction by the splanchnic area. The splanchnic extraction of AA has been shown to be higher in the elderly than in young adults( Reference Boirie, Gachon and Beaufrere 70 , Reference Volpi, Mittendorfer and Wolf 71 ). With ageing a larger amount of AA are sequestrated in the splanchnic area leading to a reduced availability of AA for peripheral tissues such as muscles. Consequently, an increased splanchnic extraction may also contribute to a decreased AA availability for muscle. With a usual amount of protein ingested, the AA level may not reach the anabolic threshold and, as a consequence, muscle anabolism may be reduced in the elderly, compared with younger adults.
Finally, particular conditions can increase the use of AA in the splanchnic area with consequences on the bioavailability of AA for muscle. It is well known that hepatic protein metabolism strongly increases in acute inflammation, for example in sepsis( Reference Obled, Papet and Breuillé 18 ). Detoxification of APAP occurs in the liver and utilises Cys, meaning that the availability of Cys to muscle could be impacted and result in a pro-sarcopenic effect. APAP is a widely used analgesic drug to relieve pain especially in older individuals, a population at risk of low protein intake. The potential negative effects of APAP on Cys homeostasis and on muscle will be developed below.
Prevalence of paracetamol treatment in older individuals
APAP is the first-intention recommended medicine for the treatment of chronic pain especially in older individuals( Reference Makris, Abrams and Gurland 12 , Reference Bertin, Keddad and Jolivet-Landreau 72 – Reference Abdulla, Adams and Bone 77 ). Chronic pain is defined as daily pain which persists for 3 consecutive months( Reference Cavalieri 73 , Reference Ferrell, Argoff and Epplin 78 ). Pain prevalence varied from 25–40 % for home-dwelling to 28–93 % for the institutionalised elderly( Reference Abdulla, Adams and Bone 77 ). In the USA, at least 50 % of home-dwelling elderly individuals experienced chronic pain( Reference Cavalieri 73 , Reference Ferrell, Argoff and Epplin 78 ). In institutionalised elderly individuals this prevalence reached 49–84 %( Reference Cavalieri 73 , Reference Won, Lapane and Vallow 79 , Reference Maxwell, Dalby and Slater 80 ). The most common cause of chronic pain was arthritis whose prevalence reached 35–57 %( Reference Lawrence, Helmick and Arnett 81 , Reference Helmick, Felson and Lawrence 82 ) and low back pain with a prevalence of 49 %( Reference Lawrence, Helmick and Arnett 81 , Reference Leveille, Fried and Guralnik 83 ). The non-management of chronic pain in the elderly can have large repercussions in terms of health and well-being with a high risk of depression, altered physical activity, higher outcomes of falls and malnutrition, and so an altered quality of life( Reference Perrot 74 , Reference Ayres, Warmington and Reid 84 ). APAP is an over-the-counter medicine, which has been recommended by referent organisations such as the WHO, the Food and Drug Administration, the American Geriatrics Society( Reference Fine 75 , Reference Ferrell, Argoff and Epplin 78 ), the OsteoArthritis Research Society International( Reference Zhang, Moskowitz and Nuki 85 ), the European League Against Rheumatism( Reference Jordan, Arden and Doherty 86 , Reference Zhang, Doherty and Arden 87 ) and the National Institute for Health and Clinical Excellence( 88 ) to treat pain of small to moderate intensity. APAP has been the most sold and consumed drug for years in the USA and France( Reference Kaufman, Kelly and Rosenberg 89 , 90 ).
The maximum therapeutic dose of APAP is 4 g/d (four doses of 1 g each spaced by at least 4 h) for adults, whatever their age, and without hepatic or renal insufficiencies. At therapeutic doses APAP is usually considered safe even in long-term treatment. Notably, no sign of hepatotoxicity was reported when 4 g/d APAP was administered for up to 12 months to adult patients with osteoarthritis pain( Reference Temple, Benson and Zinsenheim 91 ). However, a recent meta-analysis including eight observational studies revealed increases in cardiovascular and gastrointestinal disorders, and in mortality with regular intake of high therapeutic doses( Reference Roberts, Delgado Nunes and Buckner 15 ).
Paracetamol detoxification induces an irreversible loss of cyseine
After oral ingestion, APAP is rapidly absorbed at the intestinal level( Reference Lu, Thomas and Tukker 92 ). APAP half-life varies from 90 min to 3 h after a unique dose( Reference Hodgman and Garrard 17 , Reference McGill and Jaeschke 93 ) with a maximum concentration reached at 60 to 90 min. These values depend on the fed state (i.e. post-absorptive or post-prandial) and the ingested dose( Reference Stillings, Havlik and Chetty 94 ). APAP undergoes an intense hepatic metabolism followed by renal elimination. APAP detoxification mainly occurs in the liver through phase I and II reactions( Reference Forrest, Clements and Prescott 16 , Reference Hodgman and Garrard 17 ). Briefly, APAP metabolism consists mainly (up to 90 %) of phase II reactions: direct sulfate (sulfation) or glucuronide conjugations (Fig. 2). The obligate coenzyme of the sulfotransferase reaction responsible for sulfation is the activated form of endogenous inorganic sulfate that is named PAPS (3’-phosphoadenosine 5’-phosphosulfate). PAPS synthesis depends on sulfate availability and sulfotransferase activities( Reference Klaassen and Boles 95 ). A marginal (about 10 %) phase I reaction consists of conjugation with GSH (γ-glutamyl-cysteinyl-glycine)( Reference Forrest, Clements and Prescott 16 , Reference Hodgman and Garrard 17 ). In this phase, APAP is converted by cytochrome P450 into a highly reactive compound (N-acetyl-p-benzoquinone imine; NAPQI). NAPQI is neutralised by GSH and then metabolised through the mercapturate pathway.
Sulfation being a capacity-limited process, glucuronidation and, more significantly, oxidation to NAPQI increase at a supra-therapeutic dose of APAP( Reference Forrest, Clements and Prescott 16 , Reference Hodgman and Garrard 17 ). NAPQI is the compound responsible for the potential hepatic toxicity of APAP. After an APAP overdose, i.e. more than 7·5 g( Reference Hodgman and Garrard 17 ), the hepatic GSH pool is strongly depleted and NAPQI forms APAP–protein adducts causing important mitochondrial damage leading to cell death( Reference Jaeschke, McGill and Ramachandran 96 ). APAP overdoses have been treated with N-acetylcysteine for 40 years( Reference Prescott, Park and Ballantyne 97 ). Standardised administration of N-acetylcysteine should be started within 8 h of an acute overdose( Reference Hodgman and Garrard 17 ). N-acetylcysteine is a drug that replenishes GSH stores in vivo ( Reference Atkuri, Mantovani and Herzenberg 98 ). Co-administration of APAP and N-acetylcysteine could prevent acute APAP toxicity, as has been shown in a preclinical study( Reference Owumi, Andrus and Herzenberg 99 ).
We and others have recently shown that the formation of hepatic APAP–protein adducts occurs also at doses below the toxic level and even at therapeutic dosages. APAP–protein adducts do not lead to cell toxicity when they remain a small amount. APAP–protein adducts were present in serum from patients who received the maximum daily therapeutic dose of 4 g/d, for 10 d( Reference Heard, Green and James 100 ). APAP–protein adducts were also present in liver 1 h after a single injection of a very low dose of APAP (15 mg/kg) in overnight fasted mice( Reference McGill, Lebofsky and Norris 101 ). In both studies there was no relationship between APAP–protein adducts and toxicity. APAP–protein adducts were also observed without GSH depletion. As recently questioned( Reference McGill and Jaeschke 93 ), it seems that NAPQI simultaneously binds to GSH and proteins. We showed in a rat model that a 0·5 or 1 % APAP diet, equivalent to 2 and 4 g/d for humans, respectively( Reference Mast, Joly and Savary-Auzloux 102 ), during 17 d led to the formation of APAP–protein adducts( Reference Mast, Lyan and Joly 103 ). APAP–protein adducts were formed in the absence of hepatotoxicity and were detected with a large increase (218 %) with the 1 % dose compared with 0·5 %. So, even at low therapeutic dosage and before hepatic GSH depletion, APAP–protein adduct formation occurred and increased more than the increase in the dose.
After hepatic metabolism, all endproducts of APAP detoxification are excreted in the urine, 55–60 % as glucuronide, 20–30 % as sulfate conjugates and up to 10 % as GSH-dependent conjugates( Reference Forrest, Clements and Prescott 16 , Reference Hodgman and Garrard 17 ). As Cys is both the main source of sulfate and the limiting AA in GSH synthesis( Reference Taylor, Nagy and Bray 104 , Reference Courtney-Martin, Ball and Pencharz 105 ), APAP metabolism leads to an extensive use of Cys that is diverted from its physiological uses and finally definitively lost in the urine (Fig. 2). Based on the respective proportions of APAP metabolites, it appears that 30–40 % of the dose is metabolised using Cys-derived compounds (sulfate or GSH)( Reference Hodgman and Garrard 17 ). By equimolarity, the maximum therapeutic dose of 4 g/d APAP represents a net loss of 1·3 g Cys/d. Cys siphoned under APAP treatment is highly significant as it represents 20 % of the sulfur AA (SAA) intake in the elderly treated with 3 g/d APAP( Reference Pujos-Guillot, Pickering and Lyan 106 ). The requirement in AAS could be not achieved due to the low amount of ingested food in the elderly individuals chronically treated with APAP. The average SAA intake of the elderly was estimated to be 1·8 g/d, meaning that APAP detoxification would lead to urinary loss of the major part of SAA ingested, and that the main SAA metabolic needs would be uncovered( Reference Nimni, Han and Cordoba 107 ).
Cysteine, a conditionally indispensable amino acid
Cys and methionine (Met) are the two SAA used for protein synthesis but only Met is indispensable. Cys is provided by dietary proteins, GSH and protein degradation, with its endogenous synthesis occurring mainly in the liver from Met and serine( Reference Courtney-Martin and Pencharz 108 ) (Fig. 2). Cys becomes indispensable when its endogenous synthesis cannot be sufficient to cover all metabolic needs( Reference Obled, Papet and Breuillé 18 ). Both Cys supplied preformed in the diet and Cys formed from Met are equally partitioned toward the synthesis of taurine, sulfate and GSH( Reference Courtney-Martin and Pencharz 108 ) (Fig. 2), the last two compounds being key players in APAP detoxification.
Sulfate can be synthesised by two different pathways from Cys (Fig. 2). Sulfate is one of the endproducts of the cysteine sulfinate-dependent pathway of Cys catabolism, whose first step is catalysed by cysteine dioxygenase, a highly regulated enzyme( Reference Stipanuk 109 ). Other Cys catabolic pathways produce sulfide that is then oxidised into sulfate within the mitochondria( Reference Olson, Deleon and Gao 110 ). A significant depletion of GSH limits the production of sulfate through the sulfide oxidative pathway; and thiosulfate, an intermediate in this pathway, accumulates( Reference Huang, Khan and O’Brien 111 ). This observation highlights the importance of GSH, another Cys-related compound, in sulfate synthesis.
Biosynthesis of GSH, which is tightly regulated, occurs in two steps catalysed by glutamate–Cys ligase and GSH synthetase( Reference Lu 112 ). Key factors of GSH synthesis are the activity of the rate-limiting enzyme, glutamate–Cys ligase, and the availability of Cys. GSH translocated out of the cells is degraded through the γ-glutamyl cycle, playing an important role in the inter-organ transport of Cys( Reference Lu 112 ). In the case of low SAA supply, GSH is used as a Cys supplier( Reference Cho, Sahyoun and Stegink 113 , Reference Cho, Johnson and Snider 114 ). Other studies have shown that the tissue GSH pool was depleted whereas protein synthesis was maintained in the case of low SAA supply( Reference Stipanuk, Coloso and Garcia 115 , Reference Lee, Londono and Hirschberger 116 ). These results highlight that protein synthesis has priority over GSH synthesis. In addition to its role in detoxification processes, GSH serves several vital functions including antioxidant protection. Indeed, GSH/GSSG (reduced GSH:glutathione disulfide ratio) is the major redox couple that determines the anti-oxidative capacity of cells, and GSH deficiency contributes to oxidative stress( Reference Wu, Fang and Yang 117 ). Total plasma Cys is also known as an indicator of the oxidative status( Reference El-Khairy, Ueland and Nygard 118 – Reference El-Khairy, Vollset and Refsum 121 ). The liver is quantitatively the most important organ in the metabolism of SAA and GSH( Reference Garcia and Stipanuk 122 ). The liver plays a key role in the regulation of peripheral plasma SAA/GSH and their bioavailability for peripheral tissues such as skeletal muscle( Reference Garcia and Stipanuk 122 ). Direct provision of Cys and indirect supply through GSH are indispensable for the muscle because Cys cannot be synthesised within muscle due to the lack of the enzymes necessary to synthesise Cys from Met( Reference Ishii, Akahoshi and Yu 123 , Reference Stipanuk and Ueki 124 ).
At the whole-body level, Cys can be synthesised from Met. Therefore, the total SAA requirement is defined as the Met intake, in the absence of Cys, that is needed to support all metabolic requirements of both Met and Cys( Reference Courtney-Martin and Pencharz 108 , Reference Huneau, Mariotti and Blouet 125 ). Nevertheless, Cys has been clearly demonstrated to have a sparing effect of 60 % of Met requirement when present in the diet. This sparing effect is clearly associated with a change in Met metabolic flux leading to the endogenous synthesis of Cys( Reference Rose and Wixom 126 – Reference Ball, Courtney-Martin and Pencharz 132 ). Other studies have shown that sulfate can also act as a sparing agent on Cys( Reference Smith 133 – Reference Zezulka and Calloway 135 ). However, the sulfate sparing of Cys, observed under Cys deficiency and very low sulfate ingestion, has been considered of primarily academic interest( Reference Baker 136 ). The mean requirement of total SAA is 12–15 mg/kg per d( Reference Courtney-Martin and Pencharz 108 ) and the population-safe intake is 21–27 mg/kg per d( Reference Di Buono, Wykes and Ball 137 , Reference Kurpad, Regan and Varalakshmi 138 ). The range of daily intake of SAA has been estimated from 6·8 g/d in the case of a high-protein diet to only 1·8 g/d in elderly individuals with low energy intake( Reference Nimni, Han and Cordoba 107 ).
Of note, SAA metabolism seems to be modified with ageing towards a higher SAA requirement( Reference Tuttle, Bassett and Griffith 139 , Reference Mercier, Breuillé and Buffière 140 ). Furthermore, Cys and GSH oxidation increases with age( Reference Dröge 19 ) as shown in human plasma and erythrocytes( Reference Samiec, Drews-Botsch and Flagg 141 ) and in rat liver and kidney( Reference Arivazhagan, Ramanathan and Panneerselvam 142 ), and Cys supplementation improves skeletal muscle function( Reference Dröge 19 ). Keeping in mind that APAP treatment leads to a significant irreversible loss of Cys, chronic treatment could further impair GSH and Cys homeostasis and muscle mass in the elderly.
Effects of long-term paracetamol treatment on glutathione and cysteine homeostasis
It is well known that acute APAP administration induces a time- and dose-dependent decrease in liver GSH concentration in mice or rats( Reference Mitchell, Jollow and Potter 143 , Reference Buttar, Chow and Downie 144 ). Decreases in liver concentrations of Cys and PAPS, as well as in serum sulfate, occur too( Reference Hjelle, Hazelton and Klaassen 145 ). A mathematical model of the dynamics of cellular changes in GSH homeostasis induced by APAP was recently built and tested in vitro using human liver-derived cells( Reference Geenen, du Preez and Snoep 146 ).
Chronic administration of APAP also decreases liver GSH content in mice, the effect being dependent not only on APAP dose but also on the quantity of food ingested and SAA dietary content( Reference Mast, Joly and Savary-Auzloux 102 , Reference Reicks, Calvert and Hathcock 147 , Reference Kondo, Yamada and Suzuki 148 ) (Table 1). Long-term ingestion of APAP increases the Met requirement of mice for growth, and the maintenance of GSH level and protein synthesis in the liver( Reference Reicks and Hathcock 149 ). Similarly, feeding rats with a 1 % APAP diet inhibits growth( Reference McLean, Armstrong and Beales 150 ). The dose of APAP provided by the 1 % APAP diet is considered as an equivalent to the maximum therapeutic dose of 4 g/d for humans. Indeed, as the mean daily dry food ingested by humans is approximately 400 g/d, the 4 g of APAP represents 1 % of the daily DM ingested( Reference McLean, Armstrong and Beales 150 ). The addition of 0·5 % Cys or Met to the 1 % APAP diet or 1 % Met in drinking water restored growth but the addition of 1 % sodium sulfate to drinking water was ineffective( Reference McLean, Armstrong and Beales 150 ). In addition to the lowering effect of APAP on GSH in the liver, where detoxification occurs, APAP treatment decreases liver Cys concentration, plasma free cyst(e)ine (i.e. free cysteine plus free cystine) concentration, as well as intestine and muscle GSH contents( Reference Mast, Joly and Savary-Auzloux 102 , Reference Reicks, Calvert and Hathcock 147 , Reference Reicks and Hathcock 149 ). Based on the inter-organ relationships involved in GSH metabolism, mechanisms responsible for the low muscular content in GSH could result from the expected low export of GSH from the liver, the low availability of Cys for the muscle, or even from a significant export of GSH from the muscle. Whatever the mechanism, all these variations support the idea that chronic APAP can generate a lack of Cys/GSH at the whole-body level in conditions that could be encountered in humans.
PF, pair-fed; Cys, cysteine; Met, methionine; i.p., intraperitoneal; CySS, cystine; GSSG, glutathione disulfide.
Literature concerning the effect of chronic APAP on Cys and GSH in humans is rather scarce, apart from toxicology-related studies. Nevertheless, the direct link between APAP metabolism and alterations in hepatic SAA metabolic rates was made in 1987 by Lauterburg & Mitchell( Reference Lauterburg and Mitchell 151 ). Their work showed that single administrations of therapeutic doses of APAP triggered an elevation of Cys and GSH turnover in men. These results suggest a higher Cys need even at a therapeutic dosage to ensure APAP metabolism. More recently, a study conducted in young adults showed that a therapeutic dose of APAP induced an oxidation of plasma GSH when the diet was deficient in SAA( Reference Mannery, Ziegler and Park 152 ).
Ageing is known to disturb GSH homeostasis; accordingly, the hepatic GSH pool has been found to be lower in old than in young and adult mice( Reference Chen, Richie and Lang 153 ). More interestingly, after a single APAP administration the GSH hepatic pool strongly decreased at 4 h and was restored by only 41 % in old mice at 24 h; whereas adult and young mice showed better recovery of 66 and 94 %, respectively( Reference Chen, Richie and Lang 153 ). The elderly could be also more vulnerable to APAP-induced disorders in Cys/GSH homeostasis due to the pre-existence of the pro-oxidant Cys and GSH redox status and the low GSH pool described above. In this context, our study with home-dwelling elderly individuals taking 3 g/d APAP for 2 weeks revealed that chronic APAP led to an increase of protein consumption from 0·93 g/kg per d before treatment to 1·06 g/kg per d at the end of the 2-week experimental period( Reference Pujos-Guillot, Pickering and Lyan 106 ). The corresponding increase in SAA ingested was equivalent to half of the sulfur excreted in urinary APAP conjugates. These results supported an elevation of SAA requirement in the elderly treated with APAP. This study also revealed that metabolic disorders occurred such as an enhanced oxidation of sulfur-containing compounds. Moreover, ageing tended to amplify the depletion of blood GSH and modified the pattern of urinary APAP conjugates in favour of an increased loss of Cys in postoperative patients under intravenous APAP treatment (1 g every 6 h)( Reference Pickering, Schneider and Papet 154 ).
Altogether, it appears that therapeutic APAP treatment could generate perturbation of SAA and GSH metabolism leading to lower Cys availability for physiological functions. In the first part of the present review, we summarised the importance of an adequate AA supply to muscle to ensure protein homeostasis especially in the elderly. It is well known that if only one IAA is limiting or its metabolic need is increased (for example, Cys in sepsis or acute inflammation), this leads to decreased protein synthesis especially in muscle( Reference Obled, Papet and Breuillé 18 ). Consequently, the high prevalence of APAP treatment in the elderly could generate a risk for a frequent inadequate Cys supply to muscle and chronic treatment with APAP could be a risk factor for sarcopenia (Fig. 3).
Effects of long-term paracetamol treatment on skeletal muscle
The link between chronic APAP treatment and muscle status has been investigated in animals and human subjects (Table 2). In animals, Wu et al. studied very old rats (33 months) treated daily with APAP for 6 months( Reference Wu, Desai and Kakarla 155 – Reference Wu, Liu and Fannin 157 ). These studies showed an improvement in glycaemic status and the protein muscle signalling pathway due to an antioxidant effect of APAP. These positive effects were recorded with the very low dose administrated, 30 mg/kg per d, whereas the minimal dose requested for analgesic effect in rats is 200 mg/kg per d( Reference Muth-Selbach, Tegeder and Brune 158 , Reference Choi, Lee and Suh 159 ). So, at very low dose, APAP could have a beneficial effect on muscle but this amount is far below the chronic APAP prescription for its analgesic effect in humans.
Akt, protein kinase B; mTOR, mammalian target of rapamycin; eIF2α, eukaryotic initiation factor 2α; MuRF-1, muscle ring finger protein.
We recently reported that consumption, for 14 d, of a 1 % APAP diet by adult rats leading to a daily dose equivalent to the maximum therapeutic posology for humans, decreased muscle mass( Reference Mast, Joly and Savary-Auzloux 102 ) (Table 2). APAP-induced loss of muscle mass occurred while (i) the sulfation of APAP was saturated and the GSH-dependent detoxification pathway was highly activated and (ii) plasma Cys concentration and GSH contents in liver and muscle were lowered. These results demonstrated for the first time that chronic therapeutic APAP treatment could contribute to muscle mass loss linked to a decreased Cys bioavailability even with an adequate-protein diet. We re-conducted the treatment with aged rats (24 months) submitted to repeated cures with APAP lasting 14 d and spaced by 14 d of washout to mimic treatment of chronic pain. Long-term consumption of a 1 % APAP diet in old rats created a net loss of muscle mass associated with decreased muscle protein synthesis( Reference Mast, Pouyet and Centeno 160 ). This loss of muscle mass and protein synthesis occurred simultaneously with the generalised decrease of GSH stores and plasma free cyst(e)ine but an elevation of liver protein synthesis. That suggests an inter-organ competition for the use of Cys under APAP treatment in favour of the liver and at the expense of muscle.
Clinical studies tested the effect of repeated APAP administration on muscle in combination with physical exercise( Reference Trappe, Fluckey and White 161 – Reference Trappe, Ratchford and Brower 166 ) (Table 2). They were carried out because APAP is a cyclo-oxygenase inhibitor( Reference Graham, Davies and Day 167 ) that could make an impact on prostaglandins, potential mediators of the muscle protein synthesis response to exercise( Reference Smith, Palmer and Reeds 168 ). The first study was performed after a high-intensity eccentric exercise in young adults receiving 4 g/d APAP, the first dose taken at the start of the exercise( Reference Trappe, Fluckey and White 161 , Reference Trappe, White and Lambert 162 ). At 24 h post-exercise, results showed an increased muscle fractional synthesis rate associated with increased prostaglandins. APAP attenuated the positive effects of exercise, but had no effect on muscle soreness over the following 9 d( Reference Trappe, Fluckey and White 161 , Reference Trappe, White and Lambert 162 ). The second study, in older individuals, consisted of three sessions of exercise per week for 12 weeks associated with 4 g/d of APAP( Reference Trappe, Carroll and Dickinson 163 , Reference Trappe, Standley and Jemiolo 164 ). Results showed a potentiation of muscle volume and force gain with APAP in exercised muscles, without any effect in non-exercised muscles. This effect seemed to be more pronounced in type I muscle fibres( Reference Trappe, Ratchford and Brower 166 ). The beneficial effect of long-term treatment with APAP on the exercised muscle was contradictory with the first study but could be explained by differences in the protocol design. Notably, all meals were standardised all through the first study, whereas only the evening meal before the biopsy was standardised in the second study. It is well known that protein intake is in large part responsible for protein muscle metabolism and in a previous study we showed that chronic APAP treatment in the elderly led to a spontaneous increase of dietary protein intake( Reference Pujos-Guillot, Pickering and Lyan 106 ). Thus, it is unknown whether the beneficial effect observed in exercised muscle was only attributable to the combined effect of exercise and APAP or also to a change in dietary protein intake. In the third study, APAP was administered at a lower dose (1 g/d) and only on days of exercise (3 to 5 d/week) for 16 weeks( Reference Jankowski, Gozansky and MacLean 165 ). APAP did not affect the exercise-induced increase in fat-free mass, or physical performance for several upper- and lower-body exercises, with the exception of knee flexion strength, which was increased. Altogether, these clinical studies were performed in healthy volunteers, whose dietary protein intakes were probably far above the value that theoretically places older individuals at risk of uncovered SAA requirement after APAP detoxification( Reference Nimni, Han and Cordoba 107 ). Accordingly, GSH and Cys homeostasis, which was not investigated, was probably unimpaired in these experimental conditions. We are aware of no study in patients suffering from myodystrophies or muscle disorders other than sarcopenia.
Further studies are still needed to determine whether repeated cures with APAP in the upper range of therapeutic doses in the elderly with low protein intakes could or not worsen sarcopenia (Fig. 3). Of course, this will be complicated by several confounding factors, such as (i) modification of dietary protein intake under APAP treatment, (ii) intrinsic effects of pathologies on muscle mass, (iii) modifications of physical activity that could be enhanced following the reduction of pain with APAP, and, of course, (iv) the difficulty to assess muscle mass variations over short periods of time. To overcome the latter, the determination of plasma and muscle Cys and GSH pools could be the first step to determine whether the availability of Cys for muscle decreases in older individuals under APAP treatment. If so, older individuals with suboptimal dietary protein intake could be supplemented with Cys in order to cut down the potential pro-sarcopenic effect of APAP. Indeed, in growing animals, supplementation with SAA was shown to be effective in counteracting the negative effects induced by APAP doses that are considered equivalent to the maximum therapeutic posology for humans.
Nutritional implications of paracetamol treatment
APAP detoxification requires sulfate and GSH, both originating from Cys. Cys comes from dietary proteins, and its endogenous synthesis from Met. Theoretically, SAA requirement is enlarged by APAP treatment. The extent to which APAP treatment increases SAA requirement in humans is not yet known. That extent is not expected to vary linearly with the APAP dose because the contribution of the GSH-dependent pathway largely increases when sulfate and glucuronide conjugations are exceeded. The maximum increase of SAA requirement is still to be determined in human subjects, including older individuals, under treatment with the maximum therapeutic dose of APAP (4 g/d). The 24 h indicator AA oxidation technique, based on the breakpoint of the whole-body oxidation rate of an indicator AA when the intake of the test AA increases, could be used( Reference Courtney-Martin and Pencharz 108 ). Alternatively, GSH pools or APAP–protein adducts could be helpful indicators to determine the safe SAA requirement under the maximum dose of APAP. Indeed, plasma APAP–protein adducts are expected to be high when SAA intake is not sufficient and to progressively decrease with increasing SAA intake until a breakpoint corresponding to the requirement for minimum adverse effects of APAP. However, from an ethical point of view, such studies could probably not be realisable due to the potential toxic effect of the maximum therapeutic dose of APAP under low SAA intake. A compromise could be to set the lowest test intake of SAA at the mean established requirement. It would also be interesting to know whether the Met-sparing effect of Cys is affected by APAP treatment.
The frequency of prescriptions and self-medication uses of APAP that occur concomitantly with low protein intakes (i.e. below the mean protein requirement of 0·6 g/d) is not documented. It is also unknown whether individuals treated with high doses of APAP and having a low protein intake are aware of the risk they take regarding APAP toxicity. In addition, the potential pro-sarcopenic effect of APAP might occur insidiously due to the slow process of sarcopenia, and be undetected by physicians. We suggest that high plasma APAP–proteins could be an indicator of an APAP-induced Cys/GSH deficiency.
Meanwhile, it is prudent to recommend that clinical nutritionists consider APAP treatment, when it is chronic and at high dosage, as a potential determinant of sarcopenia, notably when liver and muscle pools of GSH are impaired. They should pay attention to the quantity and the origin of proteins in order to evaluate the SAA intake. Animal proteins contain more SAA than plant proteins( Reference Huneau, Mariotti and Blouet 125 ). If SAA intake is lower than the population-safe recommendation of 21–27 mg/kg per d, potential alternative drugs/strategies of chronic treatment with high doses of APAP should be discussed within multidisciplinary medical teams in order to evaluate benefit/risk of all the components of the care of patients, case by case.
Conclusion
APAP is the first-intention treatment of chronic pain of small to moderate intensity and is widespread in the elderly. Detoxification of APAP occurs in the liver and utilises sulfate and GSH, both of which are issued from Cys, a conditionally IAA. The detoxification-induced siphoning of Cys, which occurs under chronic treatment with APAP at the upper range of therapeutic doses, could reduce the availability of Cys for skeletal muscle. Since a lack of one IAA leads to lower protein synthesis especially in muscle, an APAP-induced decrease in Cys availability could exert a pro-sarcopenic effect. This potential negative drug–nutrient interaction deserves attention because sarcopenia is an important component of the frailty syndrome and can lead to dependency. It is of prime importance in the prevention of sarcopenia.
Preclinical and clinical studies gathered here provide evidence that APAP treatment can affect skeletal muscle. In combination with exercise, the effects of APAP could be related to either its role as a cyclo-oxygenase-inhibitor or linked to an APAP-induced shortage of Cys. Studies performed in growing animals have clearly shown that APAP treatment can impair growth rate and that supplementation with SAA was effective in restoring growth. Moreover, muscle mass decreased in adult rats and sarcopenia was worsened in old rats under an APAP treatment that activated the GSH-dependent detoxification pathway and impaired whole-body Cys and GSH homeostasis. These observations indicate that chronic treatment with a dose comparable with the maximum dose for humans increases the requirement of SAA. This is consistent with the spontaneous increase in protein ingestion observed in aged patients treated with APAP. The extent to which SAA requirement is increased is still to be determined.
Concerning the elderly, it is suggested that SAA requirement could not be covered when chronic treatment with the maximum therapeutic dose of APAP is associated with low protein intake. None of the clinical studies performed so far with chronic administration of APAP investigated the potential pro-sarcopenic effect of chronic/repeated APAP treatment in the elderly with low protein intake. Further clinical studies are needed to clarify the effect of chronic treatment on muscle, taking into account dosage and the heterogeneity of older populations regarding health, physical activity and nutritional status. It will be worthwhile to determine whether APAP can decrease Cys availability for muscle and consequently aggravate sarcopenia in the elderly. If so, the consumption of a high-protein diet or proteins rich in SAA (for example, eggs, cereals and lactoserum) will be the logical issue to be tested in order to cut down the pro-sarcopenic effect of chronic APAP treatment. Alternatively, nutritional supplements dedicated to the elderly with low dietary protein intake could be enriched in SAA to cover the specific APAP-induced increase in Cys requirement.
Acknowledgements
The present review received no specific grant from any funding agency, commercial or not-for-profit sectors.
C. M. planned the review and wrote the first draft, D. D. reviewed it and I. P. finalised the review.
The authors declare that they have no conflicts of interest.