Assessing lithium nutritional status by analyzing its

Trace Elements and Electrolytes, Vol. ■00■ – No. ■00■/2019 (1-7)

Original©2019 Dustri-Verlag Dr. K. Feistle

ISSN 0946-2104

DOI 10.5414/TEX01551e-pub: February 10, 2019

Accepted for publicationAugust 28, 2018

Correspondence toProf. Berislav Momčilović, MD, PhD Institute for Research and Development of the Sustainable Ecosystems (IRES), Srebrnjak 59, 10000 Zagreb, Croatia [email protected]

Key wordslithium – nutritional status – hair – whole blood – safety limits

Assessing lithium nutritional status by analyzing its cumulative frequency distribution in the hair and whole bloodNinoslav Mimica1,2, Juraj Prejac3,4, Andrey A. Skalny5,6, Andrei R. Grabeklis7,8,9,10, and Berislav Momčilović11

1University Psychiatric Hospital Vrapče, 2 School of Medicine, University of Zagreb, 3Department of Oncology, University Hospital Center Zagreb, 4School of Dental Medicine, University of Zagreb, Zagreb, Croatia, 5Federal State Scientific Institution “Institute of Toxicology”, Federal Medico-Biological Agency, St. Petersburg, 6Russian Society of Trace Elements in Medicine, ANO “Center for Biotic Medicine”, Moscow, 7Orenburg State University, Orenburg, 8P. G. Demidov Yaroslavl State University, Yaroslavl, 9RUDN University, 10All-Russian Research Institute of Medicinal and Aromatic Plants, Moscow, Russia, and 11Institute for Research and Development of the Sustainable Ecosystems (IRES), Zagreb, Croatia

Abstract. Objective: Lithium was re-cently proclaimed to be an essential trace el-ement in human nutrition [1]. The aim of this study was to assess lithium nutritional status by analyzing lithium frequency distribution in the long-term biological indicator tissue of hair (H-Li) and in the short-term biologi-cal indicator tissue of whole blood (WB-Li). Materials and methods: Hair samples were analyzed in 1,073 apparently healthy adult Croats (339 males and 734 females) and the whole-blood lithium was analyzed in a random subsample of them (91 males and 143 females). Samples were analyzed using ICP-MS at the Center for Biotic Medicine, Moscow, Russia. Results: There were no gender-dependent differences in the lithium-adequate linear reference range for H-Li and WB-Li where lithium was (µg×g–1) 0.014 – 0.100 and 3.40 – 5.65, respectively. Conclusion: Lithium concentrations below 0.014 µg×g–1 for H-Li and 3.40 µg×g–1 for WB-Li would indicate lithium deficiency. The estimated upper safety lithium limits for H-Li and WB-Li were set at 0.100 and 6.65 µg×g–1, respectively. These preliminary data are important for a long-term biomonitoring of low-level lithium supplementation.

Introduction

Lithium was recently proclaimed to be an essential trace element for human nutrition [1]. Lithium is ubiquitously present in the human diet, and daily intake of lithium from food is estimated to vary in a range from

35 µg×d–1 in Finland, to 60 – 70 µg×d–1 in the USA, and ~ 105 µg×d–1 in Turkey [2]. However, since lithium was not considered to be an essential trace element [3], there are as yet no available recommended dietary al-lowances or upper safety limits for this el-ement like they exist for the other essential elements.

Lithium was shown to have a strong pharmacological potential to alleviate human depression and maniac depression when oth-er available antidepressants have failed [4]. However, the drawback for a wider usage of lithium in the psychiatric therapy of depres-sion is that it has to be used in high doses of 900 – 1,200 mg×d–1 [4]. Such high per os doses are close to the safety index limits of lithium, which has a very narrow therapeutic index, and could induce a plethora of toxic side effects [5]. Indeed, lithium is neurotoxic [6], it can induce nephrogenic diabetes insip-idus [7], and impair thyroid gland function [8]. One of many possible factors affecting lithium toxicity is its interference with so-dium metabolism [9].

Pharmacologic action of lithium in the treatment of manic-depressive psychoses is present when lithium intakes are sufficient to raise plasma lithium to 7 – 10 µg×mL–1 [2]. One of the specific peculiarities of the health effects of lithium is that suicidal rates are higher in the geographic areas with low-lithium than in the high-lithium areas [10].

Supplemental materialis available for free at:www.dustr i .com/nc/journals-in-english/mag/trace-elements-and-electrolytes.htmlVol. ■■■ – ■■■

• 1551Mimica / 10. February 2019, 4:01 PM

Mimica, Prejac, Skalny, et al. 2

Moreover, the lithium therapy of depression, and in difference to other antidepressant agents, is also associated with a lower inci-dence of suicidality [11].

The aim of this paper is to provide evi-dence for assessing the lithium nutritional status by analyzing the lithium frequency distribution in the long-term biological indi-cator tissue of human hair and in the short-term biological indicator tissue of whole blood. We also aimed to define the provi-sional hair and whole-blood levels indicating lithium deficiency and to assess the lowest observed effect level (LOEL) in these two bioindicator tissues.

Materials and methods

This prospective, observational, cross-sectional, and exploratory epidemiological study was approved by the Ethical Com-mittee of the Institute for Research and De-velopment of the Sustainable Eco Systems (IRES), Zagreb, Croatia. The study was con-ducted in accordance with the Declaration of Helsinki on Human Subject Research [12]. Every subject gave his/her written consent to participate in the study and filled out a short questionnaire on his/her health status and medical history (data not shown) [13]. Data on hair shampooing were also recorded to control for the presence of lithium (none was found).

Hair lithium (H-Li) was analyzed in a random sample of 1,073 apparently healthy adults (339 men, 734 women). Whole-blood lithium (WB-Li) was analyzed in a subset of 212 subjects (143 women and 91 men); the median age of women and men was 47 and 50 years, respectively. Our population consisted of subjects from the general Croa-tian population who were interested to learn about their health status; the majority of them were living in the capital city region of Zagreb, Croatia. All subjects consumed their usual home-prepared mixed mid-European diet, and none of them reported an adverse medical health condition.

H-Li and WB-Li were analyzed using inductively coupled plasma mass spectrom-etry (ICP-MS) (Elan 9000, Perkin Elmer, Waltham, MA, USA) at the Center for Biotic Medicine (CBM), Moscow, Russia (Sup-

plemental material 1). The CBM is an ISO Europe-certified commercial laboratory for analyzing bioelements (electrolytes, trace el-ements, and ultratrace elements) in different biological matrices. CBM is also a member of the exclusive External Quality Assess-ment of Surrey scientific group for the qual-ity control of trace element analysis.

Hair samples were collected over the pro-tuberantia occipitalis externa, an easily iden-tifiable bony bump at the back of the skull, by cutting 5- to 7-cm long strands; the longer parts of the hair strand were discarded. The collected strands were cut in short threads, repeatedly washed, and dried. H-Li analysis was performed following the recommenda-tions of the International Atomic Energy Agency [14] and other validated analytical methods and procedures [15].

Hair lithium analysis

A strand of hair, 5 – 7 cm long and weigh-ing less than 1 g was cut with titanium-coated scissors over the anatomically well-defined bone prominence at the back of the skull (protuberantia occipitalis externa). The indi-vidual hair samples were further minced into strands less than 1 cm long prior to chemical analysis, stirred for 10 minutes in ethylether/acetone (3 : 1, w/w), rinsed 3 times with the deionized H2O (18 MΩ×cm), dried at 85 °C for 1 hour to constant weight, immersed for 1 hour in 5% EDTA, rinsed again in the de-ionized H2O, dried at 85 °C for 12 hours, wet-digested in HNO3/H2O2 in a plastic tube, sonicated, and microwaved. The digested so-lutions were quantitatively transferred into 15 mL polypropylene test tubes. The liners and top were rinsed 3 times with deionized water, and the rinses were transferred into the individual test tubes. These test tubes were filled up to 15 mL with deionized water and thoroughly shaken to mix. The samples were run in a NexION 300 + NWR 2013 spectrometer (Perkin Elmer, Waltham, MA, USA). Graduation of the instrument was car-ried out with a Perkin Elmer reference solu-tion. We used certified GBW09101 Human Hair Reference Material (Shanghai Institute for Nuclear Research, Academia Sinica, Shanghai, China) to validate the quality of the analytical work.


Lithium nutritional status 3

Whole-blood lithium analysis

Whole blood was drawn by venipuncture from the vena cubiti and collected into green-cup Vacuette collecting tubes (#454082 LotA13030M7m Greiner Bio On Interna-tional, Kremsmünster, Austria) which were randomly assigned for the ICP-MS analy-sis. Whole-blood samples of 0.5 mL were digested in a microwave oven with 0.1 mL of HNO3 at 175 °C. Blood standards were lyophilized SeronormTM Trace Elements Whole Blood Reference Standards level 1 (OK 0036), level 2 (MR 9067), and level 3 (Ok 0337) for lithium in the whole blood (SERO AS, Billingstad, Norway). A total of 5 mL of redistilled H2O were added to ev-

ery reference standard and stirred gently at room temperature for 2 hours to equilibrate. 1 mL of this equilibrated standard was pipet-ted into a 25-mL quartz glass vial and dried at 105 °C for 24 hours. The microwaved samples were dissolved in 5 mL of redistilled H2O with 0.1 mL of H2O2 added.

The detection limits for lithium in the hair and whole blood were 0.0105 and 0.00105 µg×g–1, respectively. All chemi-cals were of pro analysis grade (Khimmed Sintez, Moscow, Russia). Our detection limits (µg×g–1) are H-Li 0.0001 and WB-Li 0.00105. Lithium belongs to the pleiad of 14 elements sharing the same mass number (number of isotopes/name of the element): 3 He, 5 Li, 4 Be, and 2 B [16].

Median derivatives

The frequency distribution of lithium in the hair and whole-blood samples was ana-lyzed with the median derivative method of the log-transformed data after the Gaussian frequency distribution pattern was generated.

To scrutinize the H-Li and WB-Li fre-quency distribution, we used the median derivative model to fit the sigmoid logistic regression function (power function) for men and women separately (Supplemental mate-rial 1) A2 + (A1 – A2)/[1+(x/x0)p], where A1 is the initial value (lower horizontal asymp-tote), A2 is the final value (upper horizontal asymptote), x0 is the center of the median (M0 detected), p is power (the parameter that affects the slope of the area about the inflection point) (Supplemental material 2). The OriginPro 8.0 data analysis and graph-ing software was used for this analysis (OriginLab Corp., OriginPro Version 8.0., Northampton, MA, USA).

The bioassay sigmoid curves of log-trans-formed concentration data are widely used in pharmacology, toxicology, and radio-toxicol-ogy, and the semantic terminology of describ-ing such logistic curves is well developed [17, 18, 19, 20, 21]. Indeed, (a) there is no hair bioelement deposition (response if there is no bioelement availably (lithium in this case), (b) the linear hair bioelement deposition (re-sponse) is proportional to the available bioele-ment concentration, and (c) at high bioelement concentrations only small further increase in

Figure 1. A: Hair lithium frequency distribution in men (gray) and women (white) (log µg×g–1). B: Whole-blood lithium frequency distribution in men (gray) and women (white) (log µg×g–1).



hair bioelement deposition can be achieved by increasing the available bioelement concen-trations [18]. Hence, the terms lithium defi-cient, lithium adequate, and lithium excessive were chosen to describe the lithium hair and whole-blood deposition below the linear seg-ment of the logistic sigmoid curve, deposition at the linear range part of the logistic curve, and for the part of the sigmoid curve that is above the linear part, respectively. It should be noted that even the more complex division of the logistic bioassay sigmoid curve is pos-

sible, but such divisions are beyond the scope of this manuscript.

Results

Lithium was detected in all 1,073 ana-lyzed hair samples and in all 234 whole-blood samples. After the data were log trans-formed, the previous lithium data distribution (with skewed and kurtosis) was changed into standard Gaussian (bell shaped) frequency distribution curve for both the hair (Fig-ure 1A) and whole blood (Figure 1B).

Median derivatives (Supplemental mate-rial 2, 3, Table 1) were used to fit the bioas-say power function sigmoid curve. The data on the upward and downward arm of the me-dian derivatives are shown in Figure 2 (H-Li) and Figure 3 (WB-Li) for both men and women. The bioassay sigmoid curve [17] revealed that there is a linear segment of me-dian derivatives covering the range of d1 – u2 (for females) and D1 – U2 (for males) for H-Li and d3 – u4 (for females) and D3 – U4 (for males) for WB-Li. This linear range rep-resents the adequate lithium nutritional range where the rate of hair and WB-Li saturation is proportional to the dietary lithium intake. Adequate hair lithium concentrations of Cro-atian women had a linear range from 0.014 to 0.086 µg×g–1 (median 0.025 µg×g–1), and that for Croatian men ranged from 0.015 to 0.100 µg×g–1 (median 0.028 µg×g–1). The respective low linear region of the sigmoid power function curve below d1 for women and D1 for men was defined as deficient lith-ium nutritional status. Similarly, the respec-tive upper linear region of the sigmoid power function curve above the respective linear segments u2 for women (range u3 – u6) and segment U2 for men (range U3 – U6) was de-fined as an excessively high H-Li exposure. Evidently, on average, men and women have accumulated and retained the same amount of lithium in their hair.

Thus, our NOEL (no observed effect level) for lithium is composed of two parts; i.e., the lithium deficiency level and lithium adequate intake range. H-Li levels below 0.014 (µg×g–1) indicate nutritional lithium deficiency in both women and men, whereas our excessive LOEL values for H-Li are those above 0.100. Similarly, our WB-Li deficien-cy level is below 3.450 for men and 3.243 for

Figure 2. Hair lithium power law bioassay sigmoid function curve of hair lithium median derivatives. The difference between the H-Li median deriva-tives of men n = 339 (black) and women n = 734 (white). D = U men downward (D) and upward (U) median derivatives; d = u women downward (d) and upward (u) median derivatives.–– Logistic function Y = A2 + (A1-A2)/(1 + (X/X0)p); - - - 0.95 confidence limit; ••• 0.95 prediction limit.Men: Y = 0.991 + (0.026 – 0.991)/(1 + (X/0.029)1.640), r2 = 0.998;Women: Y = 1.006 + (0.021 – 1.006)/(1 + (X/0.027)1.661), r2 = 0.999.Box: Hair lithium linear saturation range for men and women (log concentration).See Supplemental material 2 for the model and Supplemental material 3 for median derivatives in-put numerical values.



women, whereas the excessive WB-Li levels are 5.480 for men and 5.634 for women, re-spectively. There was no correlation between the H-Li and WB-Li (Figure 3) values, and there was no effect of age on H-Li and WB-Li (Figure 4). However, the observed popu-lation concentrations of lithium in the hair were more than an order of magnitude lower than that in the whole blood.

Discussion

This is the first study to demonstrate what would be an adequate level of hair and

whole-blood lithium in the Croatian popula-tion. Indeed, hair and whole-blood lithium concentrations in this study were comparable to the hair and whole-blood lithium concen-trations observed by other authors [22, 23, 24, 25]. Hair lithium concentrations were lower than that in the whole blood and there was no gender difference in lithium content in either of the bioindicator tissues. Indeed, there was no difference in confidence inter-val and predictive limits within the same bio-indicator, i.e., H-Li and WB-Li, respectively. Apparently, lithium absorption from the gastrointestinal tract into the whole blood is much better than the further transport within the body of lithium into the hair. It should be noted that the therapeutic doses of lithium in clinical practice are 250 – 500 µg×mL–1, which is well above the observed dietary lithium intake [4].

The pharmacologic effects of lithium on depression are present when lithium in-takes are sufficient to raise plasma lithium to 7 – 10 µg×mL–1, and at slightly higher levels lithium becomes toxic [2]. However, in our healthy population, median lithium whole-blood levels were between 0.025 and 0.028 µg×mL–1 for women and men, respec-tively. It should be noted that lithium concen-trations are considerably lower in the plasma than in the whole blood [4]. We have esti-mated safety limits for whole-blood lithium at 6.5 µg×mL–1 for both men and women. The mere fact that the incidence of depres-sion is inversely related to the normally con-sumed dietary lithium, indicates the benefi-ciary health effect by much lower doses of lithium than those currently used in psychi-atric therapy [10, 11]. Our recent hair multi-element profile analysis has shown that the incidence of depression is strongly and posi-tively associated with the presence of met-als that make strong bonds with proteins [26, 27]. All of these metals were shown to impair the function of the Na-K-ATPase, which reg-ulates expulsion of sodium from the cell and influx of potassium into the cell in order to maintain the electrical cell membrane poten-tial in all the cells of our body [28]. Indeed, we have already shown how sodium and po-tassium concentrations were increased in the hair of depressed subjects [29], which would indicate some type of failure in the control of these two principal body cations [30].

Figure 3. The power law bioassay sigmoid function curve of whole-blood lithium median de-rivatives. The difference between the WB-Li me-dian derivatives of men n = 91 (black) and women n = 143 (white). D = U men downward (D) and up-ward (U) median derivatives; d = u women down-ward (d) and upward (u) median derivatives.–– Logistic function Y = A2 + (A1-A2)/(1 + (X/X0)p); - - - 0.95 confidence limit; ••• 0.95 prediction limit.Men: Y = 0.992 + (0.013 – 0.992)/(1 + (X/4.075)12.145), r2 = 0.999;Women: Y = 0.989 + (0.016 – 0.989)/(1 + (X/3.867)12.288), r2 = 0.999.Box: Whole blood lithium linear saturation range for men and women (log conc).See Supplemental material 2 for the model and Supplemental material 3 for median derivatives in-put numerical values.



Disturbances of metabolism of sodium and potassium have been observed in depression, and it is known that lithium can substitute for sodium in some active transport [4, 9]. Since lithium boosts the activity of Na-K-ATPase [31], we think that our results support the hypothesis that the impairment of the cell membrane Na-K-ATPase function is the ma-jor biological factor in the genesis of human depression [26, 32]. Indeed, recovery from depression was associated with an increased function of Na-K-ATPase [33].

Thus, monitoring of the hair lithium uptake may be a useful test for guiding the long-lasting low-level lithium supplementa-tion therapy in human depression. Unlike blood, which is in permanent equilibration with the surrounding tissue, hair lithium de-position is a one-way street and thus a reli-able time log of the actually absorbed lithium over time [34].

Acknowledgment

The authors would like to thank English language Prof. emeritus David F. Marshall for his English language editing of the manu-script.

Funding

The authors received no specific funding for this work.

Conflict of interest

The authors declare that they have no conflict of interest.

Note

The corresponding author will provide references he authored upon email request.

References[1] Schrauzer GN. Lithium: occurrence, dietary in-

takes, nutritional essentiality. J Am Coll Nutr. 2002; 21: 14-21.

[2] World Health Organization. Lithium. In: Trace elements in human nutrition and health. Copenha-gen: World Health Organization; 1996. p. 224-225.

[3] Schaffer U. Lithium. In: Merian E, Anke M, Ihnat M, Stoeppler M (eds.) Elements and their com-pounds in the environment. Occurrence, analysis and biological relevance, second ed., Vol. 2. Part 3, Metals and their compounds. Weinheim: Wi-ley-VCH; 2004. p. 479-496.

[4] Lazarus J. H. Endocrine and metabolic effects of lithium. New York – London: Plenum Medical Book Company; 1986.

[5] Merrill JC, Morton JJP, Soileau SD. Metals. In: Heyes AF (ed.) Principles and methods in toxicol-ogy, fourth ed., Ch. 14. Philadelphia PA: Taylor and Francis; 2001. p. 649-698.

Figure 4. Hair lithium, a long-term biological indi-cator, and whole-blood lithium (short-term biological indicator) are incommensurable. ○ Women, ■ men.

Figure 5. Age does not affect the gender-depen-dent difference in lithium distribution in men and women. ○ Women, ■ men.



[6] Emilien G, Maloteaux JM. Lithium neurotoxicity at low therapeutic doses Hypotheses for causes and mechanism of action following a retrospec-tive analysis of published case reports. Acta Neu-rol Belg. 1996; 96: 281-293.

[7] Bedford JJ, Weggery S, Ellis G, McDonald FJ, Joyce PR, Leader JP, Walker RJ. Lithium-induced nephrogenic diabetes insipidus: renal effects of amiloride. Clin J Am Soc Nephrol. 2008; 3: 1324-1331.

[8] Candy J. Severe hypothyroidism – an early com-plication of lithium therapy. BMJ. 1972; 3: 277.

[9] Mallinger AG, Hanin I, Himmelhoch JM, Thase ME, Knopf S. Stimulation of cell membrane sodi-um transport activity by lithium: possible rela-tionship to therapeutic action. Psychiatry Res. 1987; 22: 49-59.

[10] Ohgami H, Terao T, Shiotsuki I, Ishii N, Iwata N. Lithium levels in drinking water and risk of sui-cide. Br J Psychiatry. 2009; 194: 464-465, discus-sion 446.

[11] Müller-Oerlinghausen B, Berghöfer A, Ahrens B. The antisuicidal and mortality-reducing effect of lithium prophylaxis: consequences for guidelines in clinical psychiatry. Can J Psychiatry. 2003; 48: 433-439.

[12] Browne M. Charter of rights is adopted in the UN. The New York Times; 2005. http://www.nytimes.com/learning/general/onthisday/big/0618.html. Accessed January 9, 2016.

[13] Oppenheim AN. Questionnaire – design, inter-viewing, and attitude measurement. London: Continuum; 2004.

[14] International Atomic Energy Agency (IAEA). Ele-mental Analysis of biological materials, IAEA – TEC. DOC 197. Vienna: International Atomic Energy Agency; 1980.

[15] Burges C. Valid analytical methods and proce-dures. Cambridge: The Royal Society of Chemis-try; 2000.

[16] Momčilović B, Prejac J, Momčilović R, Ivičić N, Veber D, Lykken GI. On the same element isotope mass number (Pleiad) and the clusters of elements bearing the same mass numbers in the Periodic system – the “chesshyja” (fish skin) model. Trace Elem Med (Moscow). 2008; 9: 5-20.

[17] Harpley FW, Stewart GA, Young PA. Principles of biological assay. In: Delanouis AL (ed.) Interna-tional Encyclopedia of Pharmacology and Thera-peutics, vol.1. Oxford: Pergamon Press; 1973.

[18] Grahame-Smith DG, Aronson JK. Oxford text-book of clinical pharmacology and drug therapy. Oxford: Oxford University Press; 1985.

[19] Rampell DM, Rosemberg J, Harrison JR. Biologi-cal monitoring in occupational medicine. In: La-Dou J (ed.) A Lange Medical Book. Norwalk CT: Appleton & Lange; 1990.

[20] Brody TM. Concentration-response relathionship, Ch.3, 2 nd ed. In: Brody M, Larner J, Minneman KP, Neu HC (eds.) Human pharmacology.Molec-ular to chemical. St Louis: Mosby; 1994.

[21] Gad S. C. Statistics for toxicologist, 4th ed. In: Hayes AW (ed) Principles and methods in toxicol-ogy. Philadelphia PA: Taylor & Francis; 2001.

[22] Caroli S, Alimonti A, Coni E, Petrucci F, Seno-fonte O, Violante N. The assessment of reference values for elements in human biological tissues

and fluids: A systematic review. Crit Rev Anal Chem. 1994; 24: 363-398.

[23] Chojnacka K, Zielińska A, Górecka H, Dobrzański Z, Górecki H. Reference values for hair minerals of Polish students. Environ Toxicol Pharmacol. 2010; 29: 314-319.

[24] Goullé JP, Mahieu L, Castermant J, Neveu N, Bonneau L, Lainé G, Bouige D, Lacroix C. Metal and metalloid multi-elementary ICP-MS valida-tion in whole blood, plasma, urine and hair. Refer-ence values. Forensic Sci Int. 2005; 153: 39-44.

[25] Kempson IM, Lombi E. Hair analysis as a biomon-itor for toxicology, disease and health status. Chem Soc Rev. 2011; 40: 3915-3940.

[26] Momčilović B, Prejac J, Skalny AV, Mimica N. In search of decoding the syntax of the bioelements in human hair – A critical overview. J Trace Elem Med Biol. 2018; 50: 543-553.

[27] Nieboer E, Richardson DH. S. The replacement of the nondescript term ‘heavy metals’ by a biologi-cally and chemically significant classification of metal ions. Environ Pollut. 1980; 1: 3-26.

[28] Adley DJ. The physiology of excitable cells, fourth ed. Cambridge: Cambridge Univ Press; 1998.

[29] Momčilović B, Prejac J, Brundić S, Skalny AV, Mimica N, Drmić S. An essay on human and ele-ments, multielement profiles, and depression. Transl Neurosci. 2010; 1: 322-334.

[30] Suhail M. Na+, K+,- ATPase: Ubiquitious multi-functional transmembrane protein and its rele-vance to various pathophysiological conditions. J Clin Med Res. 2010; 2: 1-17.

[31] Hesketh JE, Loudon JB, Reading HW, Glen AIM. The effect of lithium treatment on erythrocyte membrane ATPase activities and erythrocyte ion content. Br J Clin Pharmacol. 1978; 5: 323-329.

[32] el-Mallakh RS. The Na, K-ATPase hypothesis for manic-depression. I. The mechanism of action of lithium. Med Hypotheses. 1983; 12: 260-282.

[33] Naylor GJ, Dick DAT, Dick EG, Le Poidevin D, Whyte SF. Erythrocyte membrane cation carrier in depressive illness. Psychol Med. 1973; 3: 502-508.

[34] Bass DA, Hickock D, Quig D, Urek K. Trace ele-ment analysis in hair: factors determining accura-cy, precision, and reliability. Altern Med Rev. 2001; 6: 472-481.

[35] Smylevich L, Dougherty ER. Probabilistic Bool-ean network. Philadelphia; Society for Industrial and Applied Mathematics; 2002.


Documents

Assessing lithium nutritional status by analyzing its