Lithium carbonate

Endpoint summary

Key value for chemical safety assessment

Effects on fertility

Description of key information

The NOAEL for systemic toxicity is based on various findings at 45 mg/kg/bw regarding in vivo changes on body weight and water consumption, but also pronounced morphological changes in liver and kidneys. Some variations in one or the other sex on adrenals and thyroid glands (P generation however not in the F1 or F2 generation) were noted. A value of 15 mg/kg bw was derived as the systemic NOAEL.
The NOAEL for reproductive toxicity and foetal toxicity is considered to be 45 mg/kg bw/day as no obvious reproductive changes were observed in both generations.

Effect on fertility: via oral route

Endpoint conclusion:

no adverse effect observed

Dose descriptor:

NOAEL

45 mg/kg bw/day

Study duration:

chronic

Species:

rat

Quality of whole database:

GLP and guideline compliant 2-generation study with highest reliability.

Effect on fertility: via inhalation route

Endpoint conclusion:

no study available

Effect on fertility: via dermal route

Endpoint conclusion:

no study available

Additional information

A study according to OECD 416 was initiated by the Lead Registrant and a Member Registrant for non-European regulatory purposes. A summary of the findings is written below. Further publications found during literature search were also evaluated and discussed below.

Rat
The potential to affect reproductive performance including fertility and development of the progeny was studied in a two-generation reproduction toxicity study in rats with lithium carbonate according to OECD Guideline No. 416 (2001, Advinus, 2012). The test item was dissolved in Milli-Q water and administered orally by gavage to Wistar rats at dose levels of 5, 15 and 45 mg/kg bw/d. Similarly, concurrent vehicle control group animals were administered Milli-Q water (vehicle) alone. Each group consisted of 25 male and 25 female rats. Animals from all groups were observed for clinical signs, behaviour, physical abnormalities and changes in body weight, food and water consumption during various phases of the experiment. The estrous cycle length and pattern was evaluated by vaginal smears examination for all females during 2 weeks prior to mating. After a minimum of 10 weeks of treatment, females were cohabitated with males in a 1:1 (one male to one female) ratio. The number, weight, survivability and mortality of pups were observed during the lactation period. Physical signs of postnatal development were observed daily until the criterion was met. Vaginal opening and preputial separation were also observed in pups selected for the F1-generation.
The animals were subjected to detailed necropsy at sacrifice and study plan specified organs were weighed. Andrological assessment like sperm motility was evaluated for all groups, whereas the sperm morphology, enumeration of homogenization resistant testicular spermatids and caudaepididymal sperm counts were carried out only in control and high dose groups.
Histopathological examination of parents was initially carried out for the preserved organs including gross lesions from control and high dose group animals. Further, based on the microscopic changes observed in the high dose, liver, kidneys and adrenals from males and liver, kidneys and thyroid from females of P generation and liver, kidneys and thyroid from males, liver and kidneys from females of F1 generation were considered as target organs and were examined in lower dose groups. The reproductive organs of non-pregnant females were also examined in the low and mid dose groups.
The left testis was collected in modified Davidson’s fixative and one 4-5 µm thick section was prepared and stained with PAS and Hematoxylin for microscopic examination.
The post lactational ovaries were examined for qualitative depletion of primordial follicles. A quantitative evaluation of primordial and primary follicles was done in F1 females. Ovarian follicle count was carried out for the control and high dose groups and all the not littered females of F1 generation suspected of reduced fertility.
For F1 and F2 weanlings, histopathological examination of the organs of reproductive system and kidneys as potential primary target was carried out for the available one randomly selected pup/sex/ litter in all the groups. All gross lesions were also examined for the pups with external abnormalities or clinical signs.
At 5 mg/kg bw/d had no effects on general health, body weights, food and water intake, estrus cyclicity, precoital time, gestation length, pups survivability, mating, fertility, fecundity or sperm parameters in both generations. There were no treatment-related changes with regard to any absolute or relative organ weights including reproductive organs and other gross or microscopic findings of parents, offspring or weanlings in both the generations.
At 15 mg/kg bw/d, treatment significantly increased the water intake periodically in males of both generations. There were no effects on general health, body weights, food intake, estrous cyclicity, precoital time, gestation length, pups survivability, mating, fertility, and fecundity or sperm parameters in both the generations. There were no treatment-related changes in reproductive and other organ weights and gross findings of parents or weanlings in both the generations. Microscopically, slightly dilated tubules of kidneys were seen in both generation males and females, however they were considered to be an adaptation to the pharmacology of lithium carbonate (vasopressin-downregulation) and therefore not considered as a toxicological effect.
At 45 mg/kg bw/d, treatment-related findings included increased body weights and net body weight gains in males of P generation and increased water intake in both P and F1 generations in males. Apparently higher net body weight gains were observed in both P and F1 generations premating females. There were no treatment-related changes in reproductive organ weights and gross findings of parents or weanlings in both the generations. There were also no relevant treatment-related or consistent changes in estrous cyclicity, pre-coital time, gestation length, pups survivability, mating, fertility, and fecundity or sperm parameters in both the generations when dose response and historical control ranges were taken into account. Postmortem examination in P generation demonstrated a higher body weight in males, a significant increase in the absolute and relative liver weight in males and in the relative liver weight in females. Furthermore a marginal increase in absolute and relative adrenal weight and an increase in absolute but not in relative weight of thyroid in males only was noted.
In F1 generation, the terminal body weight was not affected. A significant increase in the absolute and relative liver weight was observed in males only.
Microscopically, increased cytoplasmic rarefaction of hepatocytes in liver in males was observed, whereas in females, hepatocellular hypertrophy and focal basophilic hepatocytes were observed. Increased colloids in thyroid follicles of females were also observed. However, these changes were not present in the F1 parental rats. In F1 generation, the terminal body weight was not affected. A significant increase in the liver weight was observed in males. Microscopically, increased cytoplasmic rarefaction of hepatocytes in liver in males was observed. In females, focal basophilic hepatocytes were observed in the liver. Finally, pronounced and severely dilated tubules of kidneys were observed in both generations. Taking into account the steep dose response curve of lithium carbonate, the changes and histopathological findings in kidneys and liver as well as the variations noted with regard to adrenals and thyroids, are considered as an early onset of lithium carbonate systemic toxicity.
Evaluation of pups showed that in both generations, the mean weight of male, female and total pups per litter at all the doses tested were unaffected by treatment and that there were no external abnormalities in live or dead pups in any of the groups. No treatment-related changes were observed in the survival data of pups up to lactation day 21 at all the doses tested. No relevant effects were seen for postnatal developmental observations in F1 and F2 pups such as pinna detachment, incisor eruption, ear opening, and eye opening. The mean age and body weights at acquisition of balano-preputial separation and vaginal opening in F1 were not affected by treatment when compared to vehicle control group. Finally, no test item related microscopic findings were observed in both male and female pups of F1 and F2 litters.
The “No Observed Adverse Effect Level (NOAEL)” for systemic toxicity in parental rats is considered to be 15 mg/kg bw/d. The effects observed at 45 mg/kg are considered to be of toxicological relevance. At this dose, not only various in vivo changes on body weight and water consumption but also pronounced morphological changes in liver and kidneys and some variations noted in one or the other sex on adrenals and thyroid glands (in P generation), however, not in the F1 or F2 generation were noted.
The “No Observed Adverse Effect Level (NOAEL)” for reproductive toxicity and fetal toxicity is considered to be 45 mg/kg bw/d as no clear substances related and biologically relevant effects on reproductive parameters were observed in the P, F1 and F2 generations.
In principle, this study is the key study for investigating reproductive performance including fertility and development including sexual maturation and is reliable without any restrictions (RL1). Treatment prior to mating for 10 weeks covered more than sufficiently the time needed for complete spermatogenic cycle in males. Moreover, it has to be considered that the F1 generation animals of both sexes were already exposed in utero and after birth for a sufficient period to elucidate any possible effect on functional or morphological effect on fertility including sexual maturation. The comprehensive investigations performed in two successive generation showed clearly neither biological nor toxicological relevant or consistent effects on fertility and fecundity, especially not for any sperm, spermatid parameter or primary and secondary reproductive organs. Moreover, additional endocrinologically susceptible organs (e.g., pituitary, thyroid, adrenals) revealed no clear or consistent findings indicative for adverse toxicity. The only issue which could be of some concern is the fact that the overall overt systemic toxicity might be considered to be not sufficient enough for meeting the criteria for a maximum tolerable dose (MTD). However, in this respect the known steep-dose-response relationship has to be considered especially with regards to the rather long exposure period of two successive generations with exposure in utero and during postnatal development and maturation.

There is supporting information available from a rather old long-term exposure study performed with lithium chloride (Trautner et al., 1958). In this study in general male and female rats were exposed to lithium chloride via the drinking water at concentrations of 20 mmol/L and 50 mmol/L. Within this study a subset of rats was examined for effects of prolonged subtoxic lithium exposure on pregnancy and development of its progeny in rats in a group of 52 rats and 100 controls. The animals were administered LiCl in a concentration of 20 mmol in drinking water resulting in plasma Li levels of 1.5-2.0 mmol. None of the lithium exposed animals responded with any signs of toxicity or with noticeable behaviour changes. Normal pregnancies of lithium-treated females and controls were recorded (with respect to incidence and progress of pregnancy, birth and lactation, and the health and development of the progeny). No malformations or other defects in the lithium exposed litters were recorded. There were neither any differences in size and weight among these and untreated controls. In case the progeny was maintained at the same lithium concentration in the drinking water, they showed only initially slightly lower growth but finally there was no difference to the control and the overall development can be considered as normal. For worst case considerations, the daily lithium intake was 2.0 mmol/kg bw/d corresponding to 13.9 lithium mg/kg bw/ day and representing also the NOAEL for systemic toxicity as well as for the investigated reproductive performance and progeny development. Overall, with regards to the age of the performed study, this study is finally considered as supporting information with no major concern regarding reproductive performance and development with respect to the limited parameters investigated.

Within the framework to study possible impairments on the estrous cycle in rodents as side effects of lithium therapy, Sheikha et al. (1989) examined the effects of this antimanic drug on plasma and pituitary levels of luteinizing hormone (LH) and follicle stimulating hormone (FSH) in rats following ovariectomy (OVX), i.e. an altered endocrine state. Adult female OVX rats were intraperitoneally (ip) injected with lithium, 40 days post-operation, at a dosage of 3.0 and 2.0 mEq/kg bw/d for 3 and 7 days, respectively (twice daily at 08.00 and 16.00 h). Control OVX rats received nothing or saline injections, whereas an intact control (C) received no surgical manipulation or drug injections.

The authors observed that the levels of plasma LH and FSH in non-treated OVX (only) group showed nearly 6-fold and 75-fold increase respectively compared to those in the control. Intraperitoneally lithium injections in OVX rats for 3 and 7 days resulted in a significant reduction in plasma LH and FSH levels, when compared with those in the OVX control groups. Lithium also led to a significant reduction in the levels of pituitary LH after both 3 and 7 days. In contrast, the levels of pituitary FSH remained unchanged. The authors assumed that the observed results suggested that the pituitary gonadotropes constitute a target for lithium's action, either directly or via the hypothalamus. Overall, it is concluded that beside deficiencies in reporting and study design, the obtained results in an unknown number of animals after inappropriate ip injection should be considered as not reliable or relevant for risk assessment purposes regarding possible reproductive toxicity and consequently was finally disregarded (RL3).

Ghosh et al. (1991a) performed a screening study on effect of lithium chloride on testicular steroidogenesis and gametogenesis in groups of each 8 immature male Wistar rats. Subcutaneous injections of lithium chloride at a daily dose of 2.0 mg/kg for 15 days resulted in significant inhibition of spermatogenesis at stage VII of the seminiferous epithelial cycle. Spermatogonia A, preleptotene spermatocytes and step 7 spermatids were decreased in number in comparison to controls. Serum levels of follicle stimulating hormone (FSH), luteinizing hormone (LH), prolactin (PRL) and testosterone were decreased. Activities of testicular 3beta-hydroxysteroid dehydrogenase and 17beta-hydroxysteroiddehydrogenase activities were suppressed along with a low caudal epididymal sperm count in comparison with controls. When the treatment was prolonged for 20 and 25 days, it showed an additional reduction in accessory sex organ weights and number of midpachytene spermatocytes at stage VII. Based on the observed results and the applied study conditions, the authors concluded that lithium has an adverse effect on testicular function in immature rats.

However, it has to be considered that the investigation of one dose only and s.c. injection is not appropriate for risk assessment. Moreover, relatively low animal number were studied, no complete spermatogenic cycle was covered and there was a missing comparison to historical control data. Therefore, the transfer of these screening study results to humans is considered as questionable. Finally, since sperm effects by lithium in the rat depend on a mechanism not operating in humans this study was not regarded as applicable for humans (RL3). Due to the mentioned deficiencies and uncertainties with regards to reliability and relevance, this disregarded screening study should not lead to relevant concerns regarding male fertility.

In a further screening study, Ghosh et al. (1991b) studied the effect of prolactin (PRL) supplementation in lithium-treated rats on spermatogenesis, testicular 3-beta-hydroxysteroid dehydrogenase and 17-beta-hydroxysteroiddehydrogenase activities and serum levels of FSH, LH, PRL and testosterone in groups of each 10 male adult Wistar rats. It was shown that subcutaneous (s.c.) injections of lithium chloride at a dose of 2.0 mg/kg bw/d for 21 days resulted in a significant inhibition of spermatogenesis at stage VII of the seminiferous epithelial cycle, along with reduction of serum levels of FSH and LH and suppression of the activities of the investigated two testicular steroidogenic enzymes. Administration of bovine PRL at a dose of 0.25 mg/kg bw/d plus lithium treatment resulted in a remarkable protection of spermatogenic and steroidogenic activities of the testes, along with restoration of serum levels of FSH and testosterone. In addition, treatment resulted also in a decreased testicular weight but again, prolactin revealed a significant restoration of testicular weight. Body weights of the lithium-treated animals in all groups did not differ from that in controls.

However, it has to be considered that the investigation of one dose only and s.c. injection is not appropriate for risk assessment. Moreover, a relatively low animal number was studied, no complete spermatogenic cycle was covered and there was a missing comparison to historical control data. Therefore, the transfer of these screening study results to humans is considered as questionable. Finally, since sperm effects by lithium in the rat depend on a mechanism not operating in humans this study was not regarded as applicable for humans (RL3). Due to the mentioned deficiencies and uncertainties with regards to reliability and relevance, this disregarded screening study should not lead to relevant concerns regarding male fertility.

In an exploratory screening study, Ghosh and Biswas (1991) investigated the effect of lithium chloride on the activities of ovarian 3-beta and 17-beta hydroxysteroid dehydrogenase (HSD) and histology of ovary in sexually mature virgin Wistar (own breed) rats. After monitoring the 4-day estrous cycle, 16 females received LiCl at a dose of 200 µg/0.1 mL distilled water/100 g bw/d for 16 days. Another 16 animals received the same volume of water and served as control. On study day 17 at late pro-estrous, the animals were killed, body weights were determined and ovarian and uterine weight were recorded. Ovaries from 8 females were processed and used for enzymatic studies, while the ovaries from other 8 animals were processed for histopathology. The concentration of lithium was measured by Klinaflame photometer and a statistical analysis was performed (ANOVA, multiple tailed t-test).

There was no effect on body weight but the absolute ovarian and uterus weights were decreased. The activities of ovarian 3-beta and 17-beta - HSD were reduced and the numbers of Graafian follicles per square unit were decreased. The plasma levels of lithium were analyzed to be 0.6 ± 0.02 mEq/L. Based on the obtained results, the authors concluded that lithium treatment is associated with a reduction in the activities of ovarian steroidogenic dehydrogenases and inhibition of follicular maturation, when plasma levels of lithium remains in the therapeutic range. However, based on the few parameters investigated and limited information provided as well as the application route is not clearly specified beside further reporting deficiencies, the results are considered as not reliable and consequently the study was finally disregarded (RL3).

Jana and coworker (2001) performed a pharmaceutical side effect triggered explorative screening study in sexually mature female Wistar rats to study the effect of human chorionic gonadotrophin (hCG) coadministration on ovarian steroidogenic and gametogenic activities of lithium chloride.

Eighteen animals with regular 4-day estrous cycle were used. Twelve animals received s.c. injections of lithium chloride at a dose of 1.6 mg/kg bw/d for 28 days mimicking the therapeutic levels in humans. Of the twelve lithium-treated animals, six received hCG 25 mg/kg bw/d in 0.25 mL distilled water. HCG was injected s.c. 4 h after each lithium treatment. The remaining six animals served as controls and were treated with the same volume of distilled water by the s.c. route. The treatment schedule was started during the estrous phase. On the 29th day (the dioestrous phase in the control group), animals were sacrificed by light ether anesthesia after measurement of body weight. Blood was collected and plasma was separated, stored at – 20 °C until used for the measurement of plasma lithium concentration by atomic absorption spectrophotometry. The ovaries and uterus of each animal were dissected and their relative weights recorded. One ovary from each animal was kept at 4°C for enzymatic studies. The other ovary and both uterine horns were placed in Bouin’s fluid for histological study. Paraffin blocks at 5 mm thickness and stained with hematoxylin and eosin. Ovarian D-3-beta hydroxysteroid dehydrogenase (HSD) activity and the activity of ovarian 17-beta-HSD were measured. Histometric measurement of the diameter of the uterus, the thickness of the myometrium and endometrium and the height of the luminal epithelium were made from randomly selected sections. Quantification of folliculogenesis was performed by measuring healthy follicles and regressing follicles. Plasma lithium levels were evaluated in an atomic absorption spectrophotometer. Statistical analysis was performed by one-way ANOVA followed by a multiple two-tailed t test with Bonferroni modification.

The selected treatment led to decreases in relative ovarian and uterine weights, ovarian D-3-beta-hydroxysteroid dehydrogenase and 17-beta-hydroxysteroid dehydrogenase activities, folliculogenesis, uterine diameter, endometrial and myometrial thickness, and uterine luminal epithelial height. These parameters were not changed from the control level, when subcutaneous (s.c.) human chorionic gonadotrophin (hCG) at 25 mg/kg/day was co-administered with the lithium chloride. The duration of the estrous cycle was increased in lithium chloride-treated rat with longer met-estrous and di-estrous phases. Administration of hCG with lithium chloride prevented these estrous cycle alterations. Based on the obtained results, the authors concluded that hCG can protect ovarian steroidogenic and gametogenic function after lithium chloride treatment. However, an inappropriate route of exposure (s.c.) for toxicological risk assessment was selected in this exploratory screening study. Therefore, it is concluded that the obtained results should be considered as not reliable or relevant for risk assessment purposes regarding possible reproductive toxicity and consequently the study was finally disregarded (RL3).

Thakur et al. (2003) investigated specifically the possible adverse effect of subchronic exposure of lithium carbonate on reproductive organs of male rats. Each 20 adult males were exposed to lithium carbonate at doses of 500, 800, 1100 mg/kg of diet (corresponding to 25, 40, and 55 mg lithium carbonate /kg bw/d  by applying a diet factor of 0.05 as published by EPA in 1986 for rats) for 90 days. The weight of reproductive organs, histology of testis, epididymis, seminal vesicle, prostate, testicular interstitial fluid volume (IFV), testosterone level, sperm morphology and fertility index were analyzed. Treatment with higher doses of lithium carbonate (i.e. 800, 1100 mg/kg diet) significantly reduced testes, epididymis and accessory sex organs weights, whereas, lower dose (500 mg/kg diet) did not show any untoward effect. Similarly, the sperm number from cauda epididymis and daily sperm production was significantly decreased with higher doses of lithium carbonate. The serum testosterone levels and IFVs were also reduced significantly. Seminal vesicle and prostate secretions were completely blocked and spermatozoa were not seen in the lumen of epididymis and vas deference. Histological studies have revealed that lithium carbonate (1100 mg/kg) caused degeneration of spermatogenic cells and vacuolization of sertoli cells cytoplasm in the testis. The sperm transit rate and production of abnormal spermatozoa were significantly increased. When the lithium carbonate-treated males were mated with normal cyclic females, the fertility index declined to 50% even after 30 days of withdrawal of lithium carbonate treatment. Based on the obtained results with regards to the selected design, study conditions and parameter investigated, the authors concluded that the subchronic exposure of lithium carbonate promote reproductive toxicity and reduces fertility of male rats.

Finally, with regards to missing information on plasma level, any systemic toxicity or comparison to historical control data or deduction of a NOAEL, this specifically designed study was considered as not reliable or relevant and consequently disregarded (RL3). However, it cannot be completely excluded that the outcome of this study may trigger some concern with regards to impaired male fertility.
      
The working group of Allagui et al. (2005) investigated pharmacological side effects of low lithium concentrations on renal, thyroid, and sexual functions in male and female rats. Male and female mature rats (280 in total, no further information) were divided into three groups and fed on commercial pellets. Group (C) was control, group (Li1) received 2000 mg lithium carbonate/kg of food and group (Li2) 4000 mg lithium carbonate/kg of food (corresponding to about 212 mg (5.738 mmol Li) and 323 mg (8.742 mmol Li) in males, about 190 mg (5.142 mmol Li) and 289 mg (7.822 mmol Li) in females) for up to 28 days. After 7, 14, 21 and 28 days, serum concentrations of lithium, creatinine, free triiodothyronine (FT3) and thyroxine (FT4), testosterone and estradiol were measured.

The authors reported a dose-dependent loss of appetite and a decrease in growth rate associated with polydipsia, polyuria, and diarrhea. The lithium serum concentrations increased dose and time-dependently from 0.44 mM (day 7) to 1.34 mM (day 28) in Li1 rats and from 0.66 to 1.45 mM (day 14) in Li2 rats, respectively. However, treatment was terminated at day 14 in Li2 rats due to high mortality. The significant increase of creatinine at day 7 and 14 in Li2 and Li1 rats indicated that serum lithium concentrations ranging from 0.62 to 0.75 mM were able to induce renal insufficiency, secondarily to a time-dependent rise in lithium serum concentrations. A significant decrease of serum thyroxine (FT4) and triiodothyronine (FT3) was observed for lithium concentrations from 0.66 to 0.75 mmol/L (Li2 rats) to 1.27 mmol/L (Li1 rats). This effect was more pronounced for FT3 of FT4 / FT3 conversion. Furthermore, the testosterone level decreased and spermatogenesis was arrested. In treated female rats, estradiol level was found to be increased in a dose-dependent manner and animals were arrested in the diestrus phase at day 28. Noteworthy to mention that the investigated lithium levels were within the range of human therapies. Overall, this study is considered as not relevant for toxicological risk assessment as the effects were reported at lethal or severely toxic dose levels, leading to severe impairments of the general health and severe disturbance of homeostasis, especially for endocrinological active organs. Thus, the study was finally disregarded (RL3).

Zarnescu and Zamfirescu (2006) reported results of a male fertility screening study in which the effect of lithium carbonate on the ultrastructure of seminiferous tubules were examined. Ten Wistar rats were exposed to lithium carbonate dissolved in physiological saline at a dose of 35 mg/kg bw/d for 21 days. As a control group, physiological saline was administered to four animals. At the end of the experiment, all animals were killed, and testes were removed.

Under the conditions of the study, ultrastructural findings in rat seminiferous tubules were observed (e.g., loss of germ cell attachment, tunica propria effects, damage of spermatogenic cells, spermatids). However, the deduction of a NOAEL was not possible due to a single dose application only. There was a missing comparison to historical control data. Overall, the study has to be disregarded study (RL3) as only ultrastructural effects on rat seminiferous tubules were investigated, reporting deficiencies exist and no information on general subacute toxicity was reported. Moreover, the transfer of results to humans is questionable. Due to this very specific investigation including limited exposure duration considering the length of a normal spermatogenic cycle of 56 – 60 days in rats, the reported results with limited reliability should not lead to relevant concerns regarding fertility.

Sadeghipour et al. (2008) performed an exploratory screening study on possible pharmacological side effects to investigate specifically the effect on the neurogenic relaxation of isolated rat corpus cavernosum within the framework to investigated possible mechanisms of erectile dysfunction in patients, who received lithium chloride as pharmaceutical active ingredient. Groups of adult male Sprague-Dawley rats received 600 mg/L lithium chloride in water for 30 consecutive days, while controls received the tap water without supplement. No information was provided of the total number of treated or control animals. However, the individual results of the selected parameters showed results of 5 – 6 rats of each treated or control group animals. At termination the rats were sacrificed by cervical dislocation, dissected and strips of the corpus cavernosum of the rat penis were prepared. The corporal strips were precontracted with phenylephrine and electrical field stimulation (EFS) was applied to obtain relaxation.

In further experiments, EFS were obtained (a) after a 30-min incubation with L-NAME (Nω-nitro-L-arginine methylester; 100 µM) or (b) after a 20-min incubation with L-arginine (0.1 mM). Additionally, for evaluating whether the cyclooxygenase (COX) pathway could be involved in the effect of lithium treatment on the NANC (Nonadrenergic noncholinergic) relaxation, in separate groups of either control or lithium-treated animals, EFS were obtained after a 20-min incubation with the cyclooxygenase inhibitor indomethacin (10 μM). In addition, concentration–response curves for sodium nitroprusside (SNP) were investigated in control and lithium-treated groups. Statistical analysis of the data was performed by one-way or two-way analysis of variance (ANOVA) followed by Tukey post hoc test.

There was no significant difference in the weight gain of control and chronic lithium-treated animals. Serum level of lithium was 0.31±0.02 mmol/L in chronic lithium-treated rats, while it was not detectable in control groups. The authors reported that the relaxation to EFS was significantly impaired in the treated rats. The nitric oxide (NO) synthase inhibitor Nω-nitro-L-arginine methyl ester (L-NAME; 100 µM) inhibited the relaxation to EFS in both, the treated and control rats. The NO precursor L-arginine enhanced the EFS-induced relaxation of the corporal strips of lithium chloride treated rats. The relaxation responses to the NO donor sodium nitroprusside were similar between two groups. The authors concluded that their data demonstrated that lithium treatment could impair the nitric oxide mediated neurogenic relaxation of rat corpus cavernosum, which could be prevented by L-arginine. In contrast, the SNP-induced relaxation was indistinguishable between control and lithium-treated animals.

Overall, it can be considered as very questionable, whether the obtained results of an isolated functional impairment, elucidated in a small number of animals without any investigation or correlation to gross or histopathological structures have any reliable impact on the assessment of reproductive toxicity. Consequently, the study was disregarded (RL3).

Ahmad et al. (2011) intended to investigate pharmacological side effects and investigated therefore, the possible toxic effect of small doses of lithium chloride in male Wistar rats. Groups of each 10 adult males received lithium chloride at dose levels of 10 and 30 mg/kg bw/d for 7 weeks via their drinking water. A group of 10 males served as control and received the carrier water. The body weight of all animals was recorded on day one of lithium treatment and on the last day of exposure after seven weeks. At termination, blood samples were taken for the examination of nearly standard clinical chemistry parameters. The wet weight of liver, kidney, heart, spleen and testis were recorded and these organs were processed for histopathology. In addition, erythrocyte lysate were prepared for the investigation of antioxidant enzymes. Reduced glutathione, lipid peroxidation and protein concentrations were measured. Statistical analyses was determined by 1-way analysis of variance (ANOVA), which was followed by Student-Newman-Keuls multiple comparison test.

The exposure of the male Wistar rats for 7 weeks led to a significant alteration in body weight and blood serum chemistry. The serum enzyme levels of alkaline phosphatase (ALP), high density lipoprotein (HDLP), and creatinine kinase (CK) were reduced. The serum urea and glucose were elevated in the lithium treated animals and were considered as cause for the disturbed general health status. Furthermore, a marked inhibition in the levels of serum alanine and aspartate transaminases (ALT and AST) was suggested to reflect a stimulating transamination reaction in hepatic and renal tissues. Lithium exposure reduced the glutathione (GSH) level and stimulated the lipid peroxidation (LPO) level in the rat blood cells as an indication for oxidative stress in the red blood cells. The histopathological observations of the liver and kidney tissues revealed several alterations indicative for severe hepatotoxicity and nephrotoxicity including tissue degeneration and necrosis due to lithium treatment. Based on the obtained results, the authors suggested that small doses of lithium induce overt signs of toxicity in rat blood as well as in liver and kidney tissues. Interestingly that neither the weight of the testes nor histopathological impairment of the testes were reported at these overt systemic toxic dose levels. However, as just one sex was exposed to two dose levels and only a limited number of parameters were investigated in comparison to standard requirements for subchronic toxicity studies, the value of this study is considered as somehow limited but in general supporting the known general toxicological profile of lithium chloride. With regards to possible indications for reproduction toxicity, the study was disregarded (RL3).

Toghyani et al. (2012, 2013) performed a subchronic screening study on possible adverse effect regarding male fertility. Groups of each 6 adult male Wistar rats were treated for 48 days with lithium carbonate doses of 0, 10, 20 and 30 mg/kg bw/d by gavage. One publication (Toghyani et al. 2012) reported effects on testicular tissue and LH, FSH and testosterone, while the other publication (Toghyani et al. 2013) reported findings with regards to spermatology. Twenty four hours after the last gavage, blood samples were taken, the animals were sacrificed and the testes were removed or sperm cells were isolated from the cauda epididymis, counted, motility was estimated and stained. Testes tissue was fixed with Bouin's and the section slides were stained with hematoxylin and eosin. Hormones were measured using a kit. The authors reported that under the condition of this screening study, all three doses resulted in a significant difference in the number of spermatogonia, primary spermatocytes, spermatid and spermatozoa cells and in a specific dose-dependent decrease. Additionally a reduction in LH, FSH and testosterone were reported in a dose-dependent manner. Separately, the authors reported that the rate of spermatogenesis and sperm quality were dose-dependently reduced. For none of the investigated parameters a dose without effect was achieved. Finally, the authors concluded that the exposure led to a spermatogenic dysfunction.

Overall, with regards to missing information on standard systemic toxicity, no plasma level determination, low animal numbers, too short exposure period for covering a complete spermatogenic cycle, deficiencies in reporting and evaluation and a missing comparison to historical control data, this specifically designed study was finally considered as not reliable or relevant and consequently disregarded (RL3). However, it cannot be completely excluded that the outcome of this study may trigger some concern with regards to impaired male fertility.

Mirakhori et al. (2013) carried out a pharmaceutical side effect triggered explorative screening study in sexually immature females to study the effects of lithium chloride (LiCl) on development of ovarian follicles in gonadotropin-induced rats by means of possible cellular effects, especially the balance between proliferation and apoptosis of granulosa cells. Each five animals of 23-day-old immature female rats were injected (route not specified) with 10 IU pregnant mare serum gonadotropin (PMSG), followed by injections of 250 mg/kg bw/d LiCl every 12 h for four doses. An untreated group served as control. Ovaries were removed 40 and 48 h after PMSG administration and prepared for histology, immunohistochemistry, Western blotting, and DNA laddering analysis.

The authors reported that in the ovaries of LiCl-treated rats, few antral but more atretic follicles were present compared to those of the control rats. The induction of atresia by LiCl was further confirmed by the presence of DNA fragmentation, accompanied by a reduced level of 17-beta-estradiol in the serum. At the cellular level, lithium significantly decreased the number of proliferating cell nuclear antigen (PCNA)-positive cells and conversely increased the number of TUNEL-positive cells in the granulosa layer of the antral follicles. At the molecular level, lithium increased the level of phosphorylated glycogen synthasekinase-3beta and decreased the expression of active (stabilized) beta-catenin. Based on these results, the authors concluded that there was evidence that lithium disrupts the balance between proliferation and apoptosis in granulosa cells, leading to follicular atresia possibly through the reduction in both the stabilized beta-catenin and 17-beta-estradiol synthesis.

However, the obtained results of this exploratory screening study should be treated with caution as the exact injection route was not specified and only a very small number of animals of an unspecified strain of rats was used. Therefore, it is finally concluded that the obtained results should be considered as not reliable or relevant for risk assessment purposes regarding possible impairment of fertility and was consequently disregarded (RL3).

Khodadadi and Pirsaraei (2013) investigated possible pharmaceutical side effects of lithium therapy in an exploratory screening study in immature (25-day old) female Wistar rats. The effect of lithium chloride (LiCl) on the progesterone synthesis, the main steroid produced by corpus luteum (CL), and steroidogenic acute regulatory protein (StAR) expression, the primary mechanism of the control of CL steroidogenesis were examined as endpoints. Immature female Wistar rats (25-day-old) were injected intraperitoneally (i.p.) with 2.0 mg/kg bw/d of lithium chloride (LiCl) or sterile distilled water (0.5 mL) for 15 days. The dose of LiCl was selected on the basis of the human therapeutic dose. The experiments were repeated two times with eight animals in each treatment and at each time point. All rats were treated with single i.p. injection of 10 IU pregnant mare’s serum gonadotrophin (PMSG) on the 13th day of experiment to induce follicular maturation. This was followed by single i.p. injection of 10 IU human chorionic gonadotropin (hCG) 48 h later to induce ovulation. The last injection of LiCl) or distilled water was at 12 h post-hCG injection. Rats injected only with distilled water and gonadotropins served as control group. All animals were killed by spinal dislocation at 4 h interval from 12 to 24 h post-hCG injection. Blood samples were collected by cardiac puncture. Serum levels of progesterone were measured by ELISA and CL formation was determined by histological analysis. The ovaries were rapidly removed, washed in the cold saline solution and weighed. One ovary from each rat was fixed in Bouin’s solution for histological studies, and the other was snap-frozen in liquid nitrogen and stored at 80 °C for RNA extraction. Then, StAR protein and gene expression were examined using immunohistochemistry (IC) and polymerase chain reaction. One-way ANOVA was used to analyze differences between groups.

As a result of treatment, there was a rapid increase in the ovarian weight and the number of corpora lutea observed between 16 and 20 h post hCG injection. LiCl-treated rats showed severe changes in CL formation, progesterone secretion and StAR expression during luteinization. However, folliculogenesis has not been affected. In summary, the authors concluded that the effect of LiCl on the rat ovary is reflected in the reduction of serum progesterone concentration during the luteal phase which could be attributed to the interference in the CL formation and steroidogenesis as evident from the decreased number of CL and disrupting StAR protein and mRNA expression. However, an inappropriate route of exposure (ip) for toxicological risk assessment was selected in this exploratory screening study. Therefore, it is concluded that the obtained results should be considered as not reliable or relevant for risk assessment purposes regarding possible reproductive toxicity. Thus, the study was finally disregarded (RL3).

In an explorative fertility and postnatal screening study, Gloeckner et al (1989) treated a group of 30 female Wistar rats with 20 mM lithium chloride (LiCl)/L drinking water beginning at an age of 70 days for at least 3 weeks before mating and continuously up to the 21st day of pregnancy. The control group consisted of 20 animals and received drinking water without LiCl. After spontaneous birth, the pups were reduced to 6 pups/litter and the F1-females were raised by their own dams, weaned with an age of 30 days end mated with untreated males overnight with an age of 100 - 120 days (day after mating = day 0 of pregnancy). For characterization of pregnancy performance of these F1-females, body weight gain during pregnancy, gestational duration, litter size, number of implantation sites and body mass of the newborns were recorded. Concentration of TSH was determined in serum and pituitary glands of dams end pups, for pups pooled per litter. Thyroxin (T4) was determined in serum of dams end pups and in thyroid glands of pups. Skeletal ossification of the pups after staining with alizarin red of perinatally developing ossification centers was investigated. Hepatic gamma-glutamyltranspeptidase activity (GGT) and glutathione (GSH) concentrations in the livers of the pups were also analyzed. Statistical analysis was performed using the χ2-test for comparison of portions of dams with distinct gestational durations and the Mann and Whitney test for comparison of all the other parameters. For newborn's body weight and ossification scores means per litter were used as units.

As a result of treatment, immediately after beginning the drinking water consumption decreased from 140 mL/kg (controls) to 80 mL/kg (lithium treated rats) per day. Three weeks after beginning of the treatment, serum levels of 0.96 ± 0.06 mM Li+/L were measured. The lithium consumption during pregnancy amounted to about 1.6, 1.98 and 2.08 mM/L in the first, second and third week of gestation. However, in the female offspring raised and mated with an age of 100 – 120 days (F1-dams), the reproductive performance (gestational duration, maternal body weight gain during pregnancy, number of implantations, litter size, body weight of newborns) of these prenatally lithium treated dams was not Influenced. Also the additional parameters in the form of hepatic GGT activity and GSH levels were unchanged in the progeny but skeletal ossification was reported to be slightly delayed. Maternal TSH levels (glandular and peripheral) were unchanged, but the known postpartal increase of serum T4 levels was delayed in F1-dams as was the known decrease of serum TSH in their offspring. There was also an extremely rapid postnatal increase of intraglandular T4 concentrations in offspring for unknown reasons.

Overall, standard reproductive and postnatal developmental parameters were not affected in this explorative screening study. Whether the isolated observations of a delay in T4 increase in dams exposed only in utero as well as the described postpartal TSH decrease delay and increase of only intraglandular T4 in the progeny is finally considered as highly unlikely as no other functional thyroid related parameters, no in-life and no morphological parameter was clearly affected too. Moreover, the study design consisted only of limited parameters, only one dose group and reporting and assessment deficiencies. Therefore, it is concluded that the obtained results should be considered as not reliable or relevant for risk assessment purposes regarding possible reproductive or postnatal developmental toxicity. Thus, the study was finally disregarded (RL3).
   
Ibrahim and Canolty (1990) performed a preliminary cross-fostering screening study to investigate the possible effect of lithium carbonate on gestation and growth and development of the progeny. Two groups of female Sprague-Dawley rats were fed throughout gestation one of two diets (n=12 for each diet), either control or lithium-containing (0 or 1000 ppm lithium carbonate). At parturition, half of the rats (CL) that had been fed the control diet were switched to the lithium diet, and half (LC) that had been fed the lithium diet were switched to the control diet. The remaining rats in each group were fed during lactation the same diet as during gestation (CC and LL). Body weights and feed intakes were recorded daily. At parturition, after determining the total litter weight and the number of live and dead pups, the number of pups in each litter was reduced to six, standardizing by selecting males when possible. Dams in each of the two groups were assigned to one of four experimental groups during lactation. Half of the dams fed the control diet and half fed the lithium diet during gestation were continued on the same diet throughout lactation. At the end of the 21-day lactation period, blood and organs (brain, heart, liver, kidney and spleen) were obtained from each dam and one of her randomly selected pups. Statistical analyses for differences in the gestation period were tested by t-test. Dietary treatment effects during lactation were determined by one-way Analysis of Variance with significance of differences among means being determined by Duncan's Multiple Range Test.

The results indicated that 1000 ppm lithium carbonate was detrimental to pregnant and lactating rats and their progeny. Adverse effects included decreased growth in both dams and pups, as well as impaired reproductive performance in dams and increased mortality of pups. Lithium carbonate ingestion significantly decreased body weights of pregnant rats in comparison with the control group. At the end of the lactation period, the greatest reduction in pup weight was observed for group CL, the group ingesting 0 ppm lithium carbonate during gestation and 1000 ppm during lactation. The group ingesting 1000 ppm lithium carbonate during both gestation and lactation (group LL) showed no significant reduction in pup weight in comparison with the control group that was fed no lithium (group CC). However, this latter result is no doubt due to the fact that litter size in group LL was only 80% of that of group CC.

Feed intakes of dams were significantly decreased by the lithium treatment. The administration of 1000 ppm lithium carbonate in the diet caused decreased body weight gain in pregnant rats. There was an increase in pup mortality but no gross malformations in the newborn animals. Dams fed lithium carbonate during gestation had smaller kidney and liver weights than control dams fed no lithium during gestation or lactation. In contrast, weights of brain, heart and spleen were not affected by dietary lithium. At the end of the lactation period, there was a significant reduction in spleen weight of pups when their dams ingested lithium during the lactation period, while weights of other organs were not significantly affected. Organ weights were expressed relative to brain weight as well as body weight. When organ weights of dams were expressed relative to brain weight, affected weights of liver and kidney were noted. When these organ weights were expressed relative to body weight, significantly affected weights of all organs except spleen by the end of lactation period were recorded.

The authors concluded that the study showed that under their special cross-fostering design the dietary lithium carbonate concentration of 1000 ppm decreased body weights of pregnant and lactating dams as well as growth and survival of their pups and affected their organ weights.

Overall, it has to be considered that this was a very special non-standard cross-fostering study design using only 12 pregnant dams per group and one dose group only with limited parameters when compared to current standard test guideline requirements. Due to the obtained results, the selected dose of 1000 ppm was clearly above the maximum tolerated dose. Therefore, it is finally concluded that the obtained results should be considered as not reliable or relevant for risk assessment purposes regarding possible reproductive or postnatal developmental toxicity. Thus, the study was finally disregarded (RL3).
   
Sechzer et al. (1992) reported results of an explorative screening study, which is only available as abstract. Female rats (strain not specified) were maintained on natural lithium (Li) salts (3.0 milli-equivalents (mEq)/kg bw/d, not further specified) prior to breeding and during gestation and lactation. Li-salts were administered daily in saccharine sweetened drinking water as sole source of liquid. Control groups received only the sweetened water solution. It was reported that the Li-treated dams showed impaired maternal behaviour in the form of neglect during the immediate postpartum period and throughout the first three postpartum weeks. Maternal neglect was evident by the absence of nest building, short and infrequent periods of nursing, failure to retrieve pups and poor grooming of pups. This was accompanied by significant delays in the early development of the offspring of the treated dams. Although of normal gestation age, Li-exposed pups weighed 36% less than control pups; eyes and ears opened in control pups on day 12, but not until days 18 - 20 in treated pups. Depth perception appeared in untreated offspring between days 25 - 26 but by 33 days of age none of the treated pups showed any indication of perception of depth. The initiation of the startle response in treated pups was also delayed. At four months of age, spontaneous motility in the offspring of treated dams was 20% below that of control animals.

Overall, there is only very limited information available, which do not allow a reliable assessment. There is no information on the saccharine content of the drinking water. Therefore, it is finally concluded that the obtained results should be treated with caution and are considered as not reliable or relevant for risk assessment purposes regarding possible reproductive or postnatal developmental toxicity. Thus, the study was finally disregarded (RL3).
   
Teixeira et al., (1995) performed an explorative screening study to investigate the influence of chronic lithium chloride (Li+) administration on rat offspring. Pregnant Wistar rats drank either tap water ad libitum or 10 mM LiCl (corresponding to about 0.5 mEq/L) or were water restricted (paired to rats receiving LiCl) until pup weaning. Following birth, the pups were fostered to form five experimental groups: a) Control- S, stressed by water restriction (21 litter), b) Li+ during the prenatal and lactating periods (18 litter), c) Li+ during the prenatal period only (22 litter), d) Li+ during the lactating period only (15 litter), and e) Control-NS no treatment (13 litter).

Rat litters from dams treated during the prenatal and/ or postnatal periods were subjected to a range of tests until weaning in order to determine the functional integrity and development of the central nervous system. At birth, the number of stillborn offspring was recorded. All offspring were counted, weighed, examined for external malformations and the litters were then, reduced to eight pups per dam. In another experiment during gestation (samples pooled at 5, 15 and 20 days of pregnancy) and at weaning of the offspring, different groups of dams submitted (N =40) or not (N = 10) to Li+ treatment were killed by decapitation and the blood was collected for serum Li+, Na+ and K+ quantification by flame photometry. In pups behavioral test consisting of determination of physical landmarks of development, surface righting reflex, pup retrieval, motor coordination and sensory-motor performance were performed. Statistical analysis covered the two-tailed test. One-way analysis. of variance, followed by Duncan's test was applied to the results for serum lithium, sodium and potassium concentrations, litter size, number of male or female pups born, latency of positive righting reflex, and body weight. The chi-square test followed by Fisher's exact test were applied to the number of positive righting reflexes, pinna detachment, eye opening, cliff-avoidance, motor coordination, and the number of females that retrieved their pups.

As a result, no difference was noted in the serum sodium levels of the dams but there was an increase in serum potassium. Female pregnant rats treated with Li+ did not differ from the two controls in fertility. No malformation was observed in any of the pups. The stillborn incidence was not affected and there was no difference in litter size. No difference was found in the time to pup earflap opening or in the motor coordination test. Postnatal water restriction or Li+ treatment of the dams led in pups to an impairment of the righting reflex, an increase in the number of males born, a reduction in body weight at weaning and a delay in the eye opening date. The authors concluded that the treatment of dams with Li+ in amounts similar to those used in the prophylaxis of bipolar disorders in human aggravated the delay in physical and behavioral development of pups produced by stress associated with limited water intake and handling.

Overall, the study did not indicate a lithium chloride specific effect as limited water intake led to almost the same effects. Furthermore, only one concentration was tested and the cross-fostering study consisted only of a limited number of in-life investigations and did not comply with the current neurobehavioral testing battery (modified Irwin screen etc.) or standard reproductive and developmental toxicity study parameter. Therefore, the obtained results should be treated with caution and are considered as not reliable or relevant for risk assessment purposes regarding possible reproductive or postnatal developmental toxicity. Thus, the study was finally disregarded (RL3).

Mouse
Mroczka et al. (1983) studied the effect of continuously lithium chloride exposure to mating pairs of CFW mice in their drinking water ad libitum at doses of 0, 10, 20, 30, 50, 100 or 200 mM Li/L (corresponding to 0, 14, 28, 42, 70, 140, 208 mg/kg bw/d; 0, 2.3, 4.6, 6.9, 11.5, 23, 46 mg Li/kg bw/d ) on reproduction and postnatal development. Treatment started at 3 or 6 - 8 weeks of age and at least 2 weeks prior to mating and continued through gestation, lactation until weaning.

At a dose of 200 mM Li/L the mice died within 1 week and at 100 mM Li/L the mice did not reproduce. In mating pairs exposed to 50 mM Li/L and starting at 6-8 weeks of age, no reduction in litter size at birth but an increase in postnatal mortality and the length of time between litters, and reduction in the total number of litters per mating pair occurred. In mating pairs exposed to 50 mM Li/L and starting at 3 weeks of age, the treatment severely delayed postnatal growth and development of all pups per litter. Under the conditions and limitations of the present study, a NOAEL of 30 mM Li/L (corresponding to 42 mg/kg bw/d) was observed.

Overall, with regards to reporting and assessment deficiencies, no information on animal numbers or exact exposure period, non-standard exposure scheme with only limited parameters investigated and the missing comparison to historical control data, this reproduction and postnatal developmental toxicity screening study was finally considered as of questionable relevance and reliability and consequently disregarded (RL3). Finally, due to deficiencies and uncertainties this disregarded screening study should not lead to relevant concerns regarding fertility and postnatal development.

Nciri et al., (2009) investigated lithium carbonate for its potential to induce biochemical changes in blood, liver and testis tissue in an explorative screening study in male (Wistar?) mice. Each six animals per group received daily intraperitoneal (ip) injections of 20, 40, and 80 mg lithium carbonate/kg bw/d (corresponding to 3.77, 7.54 and 15.08 mg lithium/kg bw/d) for 14 or 28 days. The control group received daily sham injections of physiological saline (0.9% NaCl). The investigated parameters consisted of drinking water consumption, body weight, lithium and testosterone serum concentrations, activities of catalase (CAT), superoxide-dismutase (SOD), glutathione-peroxidase (GPX) and level of lipid peroxidation (expressed as TBARS) in liver. Each day, water consumption and body weight were measured. At days 14 and 28, animals were sacrificed 12 h after the last injections by rapid decapitation. Blood samples were taken and livers were processed and stored until use. The lithium concentration in serum was measured by flame atomic absorption spectrophotometry after 30 min, 1 h, 2 h, 3 h, 4 h, 5 h, 6 h, 12 h, and 24 h. For mice treated for 14 or 28 days, the lithium concentration was measured 12 h after injection. Testosterone level was measured in serum using a radioimmunoassay kit. Statistical analysis was made by the student’s t-test.

The authors reported that the first intraperitoneal injection was followed by a transitory peak of lithium in the blood, reaching 0.247+0.012 and 1.037+0.08 mM and disappearing 6 and 12 h later for the 20 and 80 mg/kg bw/d doses. From the first to the last day of treatment, lithium concentrations in the blood, measured 12 h after the injections, increased to 0.11, 0.122 and 0.25 mM for the 20, 40 and 80 mg/kg bw/d groups at the end of treatment. Lithium treatment was found to induce a weight gain, more markedly after 15 days of treatment in the top dose group. In parallel, a time- and dose-dependent polydipsia was observed. For instance, drinking water uptake was increased by 176% in the second week in the top dose group, while the water uptakes were increased in the last week by 170 and 233% in the low and mid dose groups, respectively. In addition, lithium treatment induced a decrease of blood testosterone levels in animals of the mid and top dose level, respectively. A disturbance of antioxidative status in liver cells as evidenced by the increase of TBARS level, a classical index of lipid peroxidation accompanied by an increase of both SOD and GPX activities occurred and were considered as indication for a lithium-induced oxidative stress.

Finally, the authors concluded that the lithium carbonate exposure by ip injections for up to 28 days, especially at the highest dose, was found to induce weight gain and polydipsia and a significant decrease of plasma testosterone level. The latter was considered as an indication for a damage of the male reproductive system. In addition, lipid peroxidation level and activities of SOD and GPX were increased in liver, which suggested oxidative stress in the liver.

Overall, an inappropriate route of exposure (ip) for toxicological risk assessment was selected in this exploratory screening study. Therefore, it is concluded that the obtained results should be considered as not reliable or relevant for risk assessment purposes regarding possible impairment of male fertility and consequently the study was finally disregarded (RL3).

Messiha et al. (1993) studied the effect of orally exposed mice to lithium chloride (LiCl) on its progeny in an explorative screening study. Groups of adult female Sprague-Dawley mice (questionable strain) received either deionized distilled water (controls) or 1 mEq LiCl solution (Li-treated group) ad libitum. In a first experiment, the effects of postnatal exposure of the newborn to LiCl were studied. Ten female mice were individually mated with 10 males. After mating, the pregnant mice remained in the individual cages, each with their own litter. The LiCl solution was made available immediately upon delivery (N = 5), while the controls (N = 5) remained on distilled water. The newborn were breast-fed for 3 weeks until they were weaned and then the maternal animals were sacrificed. Thereafter, the weanlings of each litter were housed in separate cages according to their sex and were supplied for 2 consecutive weeks with food and distilled water ad libitum prior to sacrifice.

In a second experiment, the effect of continued maternal ingestion of LiCl from preconception to postnatal weaning of the newborn was studied as a function of the offspring's gender. The LiCl drinking water exposure in females only covered the time of mating, during pregnancy and continued until weaning of the offspring. The mice were housed as in the first experiment except that drinking fluids were withdrawn for 6 h periods during each day of mating. The mice offspring were removed from maternity cages and separated according to gender upon weaning. They had free access to distilled water and food and remained under this Li-free condition for a subsequent 2 weeks until they were sacrificed. Maternal body weight was determined after weaning of the offspring. The newborn body weight was also measured in 24-36 h old pups, at the time of weaning and when they were sacrificed. Maternal fluid intake was measured twice weekly and the litter size was recorded. The offspring were sacrificed by decapitation and the whole brain, left kidney, liver, spleen and testis were weighed. Heart and liver were kept frozen at -20°C until they were used for the enzymatic assays. The livers and hearts were processed to obtain cytoplasmic supernatants for the determination of L-ADH, L-ALDH and the H-LDH isoenzymes. The results for each experiment were analyzed by two tailed Student's t-test for independent means for statistical significance.

In general, effects in offspring but not in maternal animals during the two weeks of development of the weaning mice occurred. Continued maternal ingestion of LiCI from preconception to weaning of the newborn impaired the offspring organ development more than by postnatal exposure. This was particularly evident in developing female mice. There was a sex-dependent impairment in the offspring body weight by maternal treatment. A reduction of testis weight was recorded in the post of weaning offspring following exposure to LiCl both pre- and postnatally, but not after postnatal only exposure. The breast feeding resulted in a reduction of brain weight by the developing offspring, an enlargement of offspring spleen by postnatal maternal ingestion of LiCl during nursing and changes in kidney weight of developing mice.

The H-LDH determination showed an offspring isoenzyme sensitivity towards Li+ as a function of duration of maternal intake of LiCI. Maternal exposure to LiCI from preconception to weaning resulted in an induction of H-LDH5 in developing offspring of both sexes. The prolonged prenatal and postnatal maternal exposure to LiCI also induced offspring L-ADH of in both sexes. This was not evident when maternal LiCI drinking began only after birth. The induction of offspring L-ALDH in both sexes occurred only after postnatal maternal drinking of LiCI.

Based on the obtained findings under these specific study conditions, the authors concluded that the female offspring were more sensitive than the males to major organ weight changes by maternal exposure to LiCI. The maternal intake of LiCl from preconception until weaning of the nurslings induced offspring hepatic alcohol dehydrogenase and heart lactate dehydrogenase in both sexes, which was isoenzyme specific for the latter. The offspring also showed induction of liver aldehyde dehydrogenase but only as consequence of postnatal exposure to LiCl. These observations were considered by the authors to indicate offspring developmental toxicity as a consequence of maternal exposure to LiCl and breast feeding.

Overall, only one concentration was tested in a limited number of animals. This specifically designed explorative screening study consisted only of a limited number of in-life investigations and did not comply with the current reproductive and developmental toxicity study parameter. Therefore, the obtained results should be treated with caution and are considered as not reliable or relevant for risk assessment purposes regarding possible reproductive or postnatal developmental toxicity. Thus, the study was finally disregarded (RL3).

Non-standard rodent
Garacia Aseff et al. (1995) performed an exploratory screening study on effects of lithium chloride (LiCl) on male and female animals of vizcacha (Lagostomus maximus, a local south American rodent of the Chinchilla family). Groups of each 15 male and 15 female adult animals received intraperitoneal (ip) injections at a dose of 1 mEq/kg bw/d LiCl for one month. Each group was divided in three subgroups to study organ damages (Lot I), recovery of organ damages (Lot II) and the renal clearance (Lot III). The lithium content in the serum was analyzed by means of atomic absorption.

Based on the results obtained, the authors reported that microscopy revealed evidence for renal, gonadal, pituitary and adrenal damages. In addition, decreases in the female serum LH levels were observed, while testosterone and estradiol levels were not affected. In the males, in any case, the Li serum values were recorded to be significantly higher than those of female animals and therefore, the observed organ damage was more severe in the tissues of the males than in the females. In the recovery group (Lot II) the tissue damage recovered completely for the pituitary, partial in testes and kidneys and was noted to be enhanced in adrenals and ovaries. The renal clearance determined in subgroup Lot Ill revealed no differences for both sexes. Finally, the authors concluded that under the investigated conditions, the ip exposure of 1 mEq LiCl/kg bw/d for one month to male and female vizcachas led to alterations in different tissues. They assumed that the gonadal findings may be the consequence of a direct effect and not secondary due to a pituitary alteration, since in male vizcacha the LH serum level was not changed. Generally, the authors recommended also that the use of Vizcachas as experimental animals may have advantages for human risk assessment purposes.

Overall, the use of a non-standardized and non-validated animal model as well as the inappropriate route of exposure (ip) for toxicological risk assessment led to the conclusion that the obtained results should be considered as not reliable or relevant for risk assessment purposes regarding possible impairment of male and female reproductive toxicity. Thus, the study was disregarded (RL3).
   
Perez Romera et al. (2000) performed an exploratory and comparative screening study using adult males of a local South American rodent strain (vizcacha, Lagostomus maximus (a rodent of the Chinchilla family)). Results were compared with findings on rats that had received the same treatment. Groups of each 4 vizcachas received a dose of lithium chloride of 1 mmol/kg bw/day intraperitoneally for 35 days, while a control group consisting also of 4 males received the solvent sterile distilled water. At termination, the animals were anesthetized, blood samples were taken and serum obtained. The left testes of rats and vizcachas were removed, processed, fixed and stained. Concurrently, sperm from the rat and vizcacha caudae epididymis were suspended to determine several sperm parameters. A total of 200 sperms were evaluated per animal and the lithium concentration in serum was determined by atomic absorption spectrometry.

The authors reported, hypospermatogenesis and that the sperm number per mL decreased markedly in comparison with the controls in the local rodent. The sperm motility and viability were also affected at the plasm levels within the therapeutic range in humans. The testicular tissue and the sperm of rats were not damaged.

However, due to the limited number of animals, effects only in the local rodent and especially due to the not appropriate application route, the results of this explorative study are of very limited value and should be disregarded (RL3).

Case studies in humans (literature data)

Case reports on men under lithium therapy, also focus on sperm effects (Levin et al, 1981) or reduced libido sexualis (Blay et al, 1982). But the effects noted do not allow any conclusion as the number of cases is very low and confounding factors were not considered or the effects noted are most likely related secondary to the wished effect of lithium treatment.

Summary, discussion and overall conclusion
The general aim was to identify possible critical findings and issues of potential authority concerns with regards to the reproductive and developmental toxicity profile of Lithium Carbonate.

In principle, Lithium in the form as carbonate (CAS 554-13-2) and partly as chloride and hypochlorite is comprehensively investigated with regards to its reproduction/fertility and developmental toxicity profile. Both, guideline conform OECD TG 416 and TG 414 studies in rats as well as explorative screening studies, predominantly concerning prenatal developmental toxicity are available in rats, rabbits, mice, monkeys and pigs. Thus, the reproductive toxicity profile of lithium carbonate can be considered as sufficiently and appropriately examined.

With regards to toxicity of reproduction, the most reliable study is the two-generation reproduction toxicity study performed in Wistar rats (Advinus, 2012) according to the most recent OECD TG 416 under GLP conditions. Clear NOAELs were obtained. The NOAEL for systemic (parental) toxicity is 15 mg/kg bw/d and the NOAEL for reproductive including fertility and developmental toxicity is 45 mg/kg bw/d, demonstrating the lack of reproduction toxicity even at the systemic toxic dose level.

In principle, this study is the key study for investigating reproductive performance including fertility and development including sexual maturation and is reliable without any restrictions. Treatment over a dose range between 15 and 45 mg/kg bw/d prior to mating for 10 weeks covered more than sufficiently the time needed for complete spermatogenic cycle in males. Moreover, it has to be considered that the F1 generation animals of both sexes were already exposed in utero and after birth for a sufficient period to elucidate any possible effect on functional or morphological effect on fertility including sexual maturation. The comprehensive investigations performed in two successive generation showed clearly neither biological nor toxicologically relevant or consistent effects on fertility and fecundity, especially not for any sperm, spermatid parameter or primary and secondary reproductive organs. Moreover, additional endocrinologically susceptible organs (e.g., pituitary, thyroid, adrenals) revealed no clear or consistent findings indicative for adverse toxicity. The only issue which could be of some concern is the fact that the overall overt systemic toxicity might be considered to be not sufficient enough for meeting the criteria for a maximum tolerable dose (MTD). However, in this respect the known steep-dose-response relationship has to be considered especially with regards to the rather long exposure period of two successive generations with exposure in utero and during postnatal development and maturation.

In respect to reproductive performance and fertility, especially male fertility and fecundity, there are further published studies available, which are generally considered as explorative screening studies as they cover predominantly selective endpoints of male reproductive organs and spermatology including isolated investigations on selective hormones. In addition, these screening studies used mostly low animal numbers and partly an exposure route (subcutaneous and/or intraperitoneall injection or injection of unknown routes) not appropriate for risk assessment. A fact which prevents a meaningful and conclusive evaluation. Due to deficiencies in reporting and assessment, principally all of these studies were disregarded. Moreover, some of these screening studies described mechanisms not operating in humans or did not allow discrimination between primary and secondary effects (Gosh et al., 1991a, 1991b; Thakur et al., 2003; Zarnescu and Zamfirescut, 2006; Toghyani et al., 2012, 2013; Mroczka et al., 1983).

Furthermore, many further published studies, considered as explorative screening studies in rats, mice and in a non-standard rodent, were primarily intended to investigate pharmacological side effects of lithium therapy in male and/or female animals at concentrations or dose levels mimicking human drug exposure (Sheikha et al., 1989; Ghosh and Biswas et a., 1991, Garacia et al., 1995; Jana et al., 2001; Allagui et al., 2005; Sadeghipour et al.,2008; Nciri et al., 2009; Perez Romera et al., 2000; Ahmad et al., 2011; Mirakhori et al., 2013; Khodadad et al., 2013). However, very often an inappropriate route of exposure route (ip) for toxicological risk assessment was selected in these exploratory screening studies. Therefore, it is finally concluded that the obtained results should be considered as not reliable or relevant for risk assessment purposes regarding possible reproductive toxicity and consequently these studies were disregarded (RL3).

However, as within authorities there is a trend to consider also mechanistically screening studies (although with low reliability) for a weight of evidence approach, it cannot be completely excluded that the outcome of these questionable screening studies may trigger some concern with regards to impaired male fertility.

Effects on developmental toxicity

Description of key information

According to of a prenatal developmental toxicity study with Lithium carbonate, the

- NO(A)EL for developmental toxicity is 90 Lithium carbonate/mg/kg bw

- NO(A)EL for maternal toxicity is 30 mg Lithium carbonate/kg bw/day.

Link to relevant study records

Reference

Endpoint:

developmental toxicity

Type of information:

experimental study

Adequacy of study:

key study

Study period:

2010-02-18 to 2010-07-06

Reliability:

1 (reliable without restriction)

Rationale for reliability incl. deficiencies:

other: GLP and guideline compliant study.

Qualifier:

according to guideline

Guideline:

OECD Guideline 414 (Prenatal Developmental Toxicity Study)

Version / remarks:

2001

Deviations:

no

Qualifier:

according to guideline

Guideline:

EU Method B.31 (Prenatal Developmental Toxicity Study)

Version / remarks:

2004

Deviations:

no

GLP compliance:

yes (incl. QA statement)

Limit test:

no

Species:

rat

Strain:

other: Crl: CD(SD)

Details on test animals or test system and environmental conditions:

TEST ANIMALS
- Source: Charles River Laboratories Research, Models and Services Germany GmbH, Sandhofer Weg 7, 97633 Sulzfeld, Germany
- Age at study initiation: 8 - 9 weeks
- Weight at study initiation: 185 - 234 g
- Housing: MAKROLON cages (type III) with a basal surface of approx. 39 cm x 23 cm and a height of approx. 15 cm
- Diet: Commercial ssniff® R/Z V1324, ad libitum
- Water: tap water, ad libitum
- Acclimation period: 5 days

ENVIRONMENTAL CONDITIONS
- Temperature (°C): 22 +/- 3 degree C
- Humidity (%): 55 +/- 15 %
- Photoperiod (hrs dark / hrs light): 12 hrs dark/ 12 hrs light

Route of administration:

oral: gavage

Vehicle:

other: 0.5% aqueous hydroxypropyl methyl cellulose gel (Methocel)

Details on exposure:

PREPARATION OF DOSING SOLUTIONS:
The test item was suspended in the vehicle 0.5% aqueous hydroxypropyl methyl cellulose gel (Methocel)1 to the appropriate concentrations and was administered orally at a constant volume of 5 mL/kg b.w. once daily from the 6th to the 19th day of pregnancy. The dose of the test item was adjusted to the animal's body weight daily. The control animals received the vehicle at a constant volume of 5 mL/kg b.w. orally once daily in the same way. The test item mixtures were freshly prepared every day approx. 1h before use.

Applied volume: 5 mL/kg bw/day

Analytical verification of doses or concentrations:

yes

Details on analytical verification of doses or concentrations:

For the analysis of the test item formulations, samples of approx. 10 mL were taken at the following times:
At the beginning of the administration period: Analysis of concentration/homogeneity.
At start of administration, during (middle) administration and before administration to the last animal of each group (3 samples/dose level group).
Total number of samples: 9

At termination of the administration period ata time point when the majority of animals was dosed: Analysis of concentration/homogeneity.
At start of administration, during (middle) administration and before administration to the last animal of each dose level group (3 sample/dose level group).
Total number of samples: 9

Thus, the sum of all samples is 18.

The samples were labelled with the study number, species, type of sample, test item, concentration, sampling time and date and were stored immediately after withdrawal at -20 degree C or colder until dispatch.

The formulation samples were analysed for Lithium levels according to GLP by the Test Site AllessaChemie GmbH. The Phase Plan “Bestimmung des Lithiumgehaltes in Trägergemisch mittels ICP-OES (Teil-Prüfplan VP-Nummer 005/2010)” and any amendments to this Phase Plan are part of the LPT Study Plan 24635.

The analysis of the test item-carrier mixtures for Lithium levels revealed that the formulations used for the administrations in groups 2 to 4 were correctly prepared. The measured actual concentrations ranged from 96.45% to 103.64% of the nominal values. The results were within the expected range of the theoretical concentrations.

Details on mating procedure:

Sexually mature ('proved') male rats of the same breed served as partners. The female breeding partners were randomly chosen. Mating was monogamous: 1 male and 1 female animal were placed together in one cage during the dark period. Each morning a vaginal smear was taken to check for the presence of sperm. If findings were negative, mating was repeated with the same partner. The day on which sperm was found was considered as the day of conception (day 0 of pregnancy). This procedure was repeated until enough pregnant dams were available for all groups. Rats which did not become pregnant were excluded from the analysis of the results and replaced by other animals. A post-mortem negative staining according to SALEWSKI was carried out in the replaced animals in order to confirm the non-pregnancy status.

Duration of treatment / exposure:

From the 6th to the 19th day of pregnancy.

Frequency of treatment:

Once daily from the 6th to the 19th day of pregnancy.

Duration of test:

20 days after mating

No. of animals per sex per dose:

25

Control animals:

yes

Details on study design:

Summary on animals examined: 21 dams per dose group

Evaluated litters: 20 per dose group

Non pregnant dams: 1 in dose groups 0, 30 and 90 mg/kg bw/ day, i.e. 3 in total

Dams without viable fetuses: 1 (dose group 30 mg/kg bw/day)

Maternal examinations:

CAGE SIDE OBSERVATIONS: No data

DETAILED CLINICAL OBSERVATIONS: Yes
- Time schedule: Immediately after administration, any signs of illness or reaction to treatment were recorded. In case of changes, the animals were observed until the symptoms disappeared. In addition, animals were checked regularly throughout the working day from 7.00 a.m. to 3.45 p.m. On Saturdays and Sundays, the animals were checked regularly starting from 7.00 a.m. to 11.00 a.m. with a final check performed at approximately 3.30 p.m.

BODY WEIGHT: Yes
- Time schedule for examinations: The weight of each rat was recorded on day 0 of gestation (the day of detection of a positive mating sign), followed by daily weighings - always at the same time of the day. The body weight gain was also calculated in intervals (i.e. day 0-3, 3- 6, 6-9, 9-12, 12-15, 15-18 and 18-20).

FOOD CONSUMPTION: Yes
- Food consumption for each animal determined and mean daily diet consumption calculated as g food/kg body weight/day: Yes

Ovaries and uterine content:

The ovaries and uterine content was examined after termination: Yes
Examinations included:
- Gravid uterus weight: Yes
- Number of corpora lutea: Yes
- Number of implantations: Yes
- Number of early resorptions: Yes
- Number of late resorptions: Yes

Fetal examinations:

Weights of fetuses and weights of the placentae were determined (fetuses were considered as runts if their weight was less than 70% of the mean litter
weight). Fetuses were inspected externally for damages, especially for malformations. The fetuses were sacrificed by an ether atmosphere.

Statistics:

For all numerical values, homogeneity of variances was tested using the BARTLETT chi-square test. When the variances were homogeneous, the DUNNETT test (p <= 0.01) was used to compare the experimental groups with the control group. In case of heterogeneity of variances, the STUDENT's t-test was carried out, limit of significance was p <= 0.01. For the comparison of classification measurements (for example malformation-, resorption-, retardation- and variation rate) the FISHER's exact test (n < 100) or chi2-test with YATES' correction for continuity (n >= 100) (p <= 0.05 and p <= 0.01) was employed.

Details on maternal toxic effects:

Maternal toxic effects:yes

Details on maternal toxic effects:
Slight but significant reductions were noted for the net weight change and the food intake.

Dose descriptor:

NOEL

Effect level:

30 mg/kg bw/day

Based on:

test mat.

Basis for effect level:

other: maternal toxicity

Details on embryotoxic / teratogenic effects:

Embryotoxic / teratogenic effects:no effects

Dose descriptor:

NOEL

Effect level:

90 mg/kg bw/day

Based on:

test mat.

Basis for effect level:

other: embryotoxicity

Abnormalities:

not specified

Developmental effects observed:

not specified

Influence on the dam:

Mortality: None of the dams treated with 10, 30 or 90 mg Lithium Carbonate/kg b.w./day died prematurely during the course of the study.

Clinical signs: Pilo-erection was noted in four high-dosed dams treated with 90 mg Lithium Carbonate/kg b.w./day on two to four days, starting on gestation day 17 or 19 and lasting until laparotomy on gestation day 20. The drinking water intake of all high-dosed dams treated with 90 mg Lithium Carbonate/kg b.w./day was increased starting on gestation day 17, 18 or 19 and lasting until laparotomy on gestation day 20.

Body weight and body weight gain: Marginal reductions were noted for the mean body weights of the high-dosed dams (90 mg Lithium Carbonate/kg b.w./day) during the last gestation days. The increase in the mean body weight from the start value (day 0 of pregnancy) was 66.9% at the time point of laparotomy (control: 74.4%). Significant reductions (at p ≤ 0.01) were noted for the net weight change of the high-dosed dams from day 6 of gestation to laparotomy on gestation day 20 (carcass weight minus day 6 body weight).

Food consumption: Slight but statistically significant reductions (at p <= 0.01 or p <= 0.05) were determined for the relative food consumption of the high-dosed dams (90 mg Lithium Carbonate/kg b.w./day) on gestation days 7, 9, 11 to 13 and 19 (up to 18.3% below the control value).

Drinking water consumption: Increased intake of drinking water was noted in all high dosed females treated with 90 mg Lithium Carbonate/kg b.w./day on one to four days, starting on gestation day 17 (qualitative observation by visual appraisal).

Necropsy findings: No test item-related pathological findings were noted.

Uterus and carcass weights: The gravid uterus weight and the carcass weight were not influenced by the exposure to the test item. Influence on the fetus: No test item-related influence was noted on the prenatal fetal development at 10, 30 or 90 mg Lithium Carbonate/kg b.w./day with respect to the number of corpora lutea, implantation sites, resorptions, sex distribution, fetal and placental weights, number of live fetuses at birth and the values calculated for the pre- and postimplantation loss when compared to the control. No dead fetuses or runts were noted at laparotomy. Malformations No malformations were noted in the fetuses during external/ internal examination, skeletal examination (according to DAWSON) or soft tissue evaluation (according to WILSON). Variations No test item-related variations were noted in the fetuses during external / internal examination, skeletal examination (according to DAWSON) or soft tissue evaluation (according to WILSON). Retardations No test item-related influence was noted for the incidence of skeletal retardations.

Toxicokinetics: The toxicokinetic analysis based on Lithium plasma levels revealed a clear dose related systemic exposure to Lithium. Mean peak plasma levels of 1.66, 3.59 and 9.65 mg Li/L plasma, respectively, were observed at 10, 30 or 90 mg Lithium Carbonate/ kg b.w./day on gestation day 19. The plasma concentrations declined with a mean elimination half-life for Lithium between 8.4 to 12.0 hours. Toxicokinetics demonstrated dose proportional increases of Lithium plasma concentrations between 10 and 90 mg Lithium Carbonate/kg b.w./day. Peak time and half-life and increased with dose levels.

Analysis of test item formulation (performed by the Test Site AllessaChemie GmbH, Germany): The analysis of the test item-carrier mixtures for Lithium levels revealed that the formulations used for the administrations in groups 2 to 4 were correctly prepared. The measured actual concentrations ranged from 96.45% to 103.64% of the nominal values. The results were within the expected range of the theoretical concentrations.

Conclusions:

Under the present test conditions, the no-observed-effect level (NOEL) was 30 mg Lithium Carbonate/kg bw/day for the dams. The NOEL for the fetuses was >= 90 mg Lithium Carbonate/kg bw/day.

Executive summary:

An prenatal developmental toxicity study was performed in rats (strain: Crl CD (SD)) according to OECD guideline 414 and EU method B.31. In this rat embryotoxicity study, the test item Lithium Carbonate was administered to female rats at concentrations of 10, 30 or 90 mg/kg bw/day orally by gavage from the 6th to 19th day of pregnancy. Under the present test conditions, the no-observed-effect level (NOEL) was 30 mg Lithium carbonate/kg bw/day for the dams (maternal NOEL). At 90 mg Lithium carbonate/kg bw/day, pilo-erection was noted in a few dams. Furthermore, slight but significant reductions were noted for the net weight change and the food intake. The NOEL for the fetuses was >= 90 mg Lithium Carbonate/kg bw/day. There was no test item-related increase in the incidence of fetal malformations, external/ internal, skeletal or soft tissue variations or skeletal retardations. The toxicokinetic analysis revealed a clear dose-related systemic exposure to Lithium. In conclusion, no embryotoxic properties of the test item were noted during external/ internal, skeletal and soft tissue examinations. No test item-related increase was noted in the incidence of malformations, variations or retardations, not even at the materno-toxic dose level of 90 mg Lithium Carbonate/kg bw/day.

Effect on developmental toxicity: via oral route

Endpoint conclusion:

no adverse effect observed

Dose descriptor:

NOAEL

90 mg/kg bw/day

Study duration:

subchronic

Species:

rat

Quality of whole database:

GLP and guideline compliant study with highest reliability.

Effect on developmental toxicity: via inhalation route

Endpoint conclusion:

no study available

Effect on developmental toxicity: via dermal route

Endpoint conclusion:

no study available

Additional information

A study according to OECD 414 with lithium carbonate was conducted by the Lead Registrant and a Member Registrant for non-European regulatory purposes. This study is valuable for read-across to evaluate the potential toxicity of lithium and lithium compounds with respect to reproduction and fertility. The findings are summarised below. Further publications found during literature search were also evaluated and discussed below.

Rat
A prenatal developmental toxicity study was performed in rats (strain: Crl CD (SD)) according to OECD guideline 414 and EU method B.31. Lithium carbonate was administered to female rats at concentrations of 0, 10, 30 or 90 mg/kg bw/d orally by gavage from gestation day (GD) 6 through GD 19. The toxicokinetic analysis revealed a clear dose-related systemic exposure to lithium. At 90 mg/kg bw/d, pilo-erection was noted in a few dams. Furthermore, slight but significant reductions were noted for the net weight change and the food intake. With regards to prenatal developmental toxicity, there was no test item-related increase in the incidence of fetal malformations, external/ internal, skeletal or soft tissue variations or skeletal retardations, even not at the maternal toxic dose level of 90 mg/kg bw/d. Under the conditions of the study, the no-observed-effect level (NOEL) for maternal toxicity was 30 mg/kg bw/d, while the NOEL for prenatal developmental toxicity was 90 mg/kg bw/d, the highest dose investigated.

Overall, this study is the key study for investigating prenatal developmental toxicity in the rat and is reliable without any restrictions (RL1). Therefore, it is finally concluded that the results obtained under this full scale study should not lead to concerns regarding prenatal developmental toxicity.

Fritz (1988) performed three different prenatal developmental toxicity studies in Sprague-Dawley derived pregnant rats using varying exposure durations. But in any case, only one dose level was studied.

In the first study, Lithium carbonate was tested in a preliminary prenatal developmental toxicity screening study in rats exposed to 100 mg/kg bw/d of the test substance during 5 consecutive days of pregnancy (Group I: GD 6 - GD10; Group II: GD 11 - GD15; Group III: GD 16 – GD 20). On day 21 p.c. the dams were sacrificed and examined. This exposure regiment lead to polyuria and reduced body weight and food consumption in exposed dams. In group III, 7/25 females died one day before expected delivery. In the uterus of 7 females of group I, hemorrhagic alterations of the implantation sites were found. In group II and III enlargement of renal pelvis occurred in fetuses. Thus, fetal effects were noted at the severely toxic and lethal maternal dose. Under the conditions and limitations of this study, the determination of a NOAEL was not possible (single dose).

Overall, due to major methodological deficiencies as effects at maternal lethal dose are considered as of questionable relevance and in the form of missing control, reporting deficiencies and no comparison to historical control data amongst others, the study is of disregarded reliability (RL3). Finally, this preliminary prenatal developmental toxicity screening study should not lead to relevant concerns regarding postnatal development due to deficiencies and uncertainties of this disregarded study.
   
In the subsequently performed pre-, postnatal developmental toxicity screening study 20 animals were treated with either 100 mg/kg bw/d of the test substance or distilled water as a control on GD 16 through GD 20. About half of the animals were killed on day 21 p.c., the remaining rats were allowed to litter and raise their offspring. During the study 2/20 mortalities occurred in the dams of the dose group. The animals of the dose group showed decreased body weight and food consumption. Water consumption was increased. Polyuria was observed from one day after the commencement of treatment until the day of autopsy or parturition. The visceral examination of fetuses removed from uterus near term revealed enlarged renal pelvis at a per litter and per fetus incidence. Further, increasing postnatal deaths occurred in the dose group. Nearly half of the offspring from dams allowed to litter, died during the first 4 days after parturition and live offspring was found amongst 4/7 litters only. Thus, effects in fetuses/pups occurred only transiently at this clearly maternal toxic dose level. Under the conditions and limitations of this study, the determination of a NOAEL was not possible (single dose).

Overall, due to major methodological deficiencies as effects at maternal lethal dose are considered as of questionable relevance and in the form of short exposure period, reporting deficiencies, only one dose group, insufficient No. of animals, only transient kidney effects in progeny and missing comparison to historical control data amongst others, the study is of disregarded reliability (RL3). Finally, this prenatal developmental toxicity screening study should not lead to relevant concerns regarding postnatal development due to deficiencies and uncertainties of this disregarded study.

In the third study in the form of a postnatal developmental toxicity screening again 20 females were treated with either 60 mg/kg bw/d of the test substance or distilled water as a control on GD 16 through GD 20. Ten females each were allowed to raise their own offspring. The offspring of the remaining females were separated as soon as possible after birth and fostered by the control group or the experimental group, respectively. The exposure led to a statistically significant decrease in maternal body weight gain and food consumption. Water consumption was significantly increased and polyuria occurred. At autopsy after weaning of the offspring the kidney of the dams were macroscopically unchanged. The average initial litter size was determined day 1 post-partum and was significantly lower in the lithium-treated dams raising their own offspring. Under the conditions and limitations of the present study, no NOAEL deduction was possible (single dose).

Overall, due to major methodological deficiencies in the form of short exposure period, reporting deficiencies, only one dose group, insufficient No. of animals and missing comparison to historical control data amongst others, the study is of disregarded reliability (RL3). Finally, this prenatal developmental toxicity screening study should not lead to relevant concerns regarding postnatal development due to deficiencies and uncertainties of this disregarded study.

Gralla and McIlhenny (1972) performed a prenatal developmental toxicity study equivalent to former OECD guideline 414 in rats. Twenty pregnant Charles River female albino rats per group were dosed by gavage from GD 5 – GD 15 with lithium carbonate solutions at 49.88, 149.63 and 299.25 mg/kg bw/d (0.675, 2.025 and 4.05 mM/kg bw/d). The control group received tap water. A low incidence of maternal mortality occurred at the highest dose level administered. Two pregnant female rats, which had received 299.25 mg/kg bw/d (4.05 mM/kg bw/d) died unexpectedly for unknown reasons. Except of these mortalities, no maternal or developmental toxicity up to the highest dose level occurred. The NOAEL for maternal and prenatal developmental toxicity including teratogenicity of 299.25 mg/kg bw/d in the rat was observed. Overall, this prenatal developmental toxicity study is considered as supporting information with only minor restriction regarding reliability (RL2). Finally, the results obtained under the conditions of the study should not lead to concerns regarding prenatal developmental toxicity.

In addition, Gralla and McIlhenny (1972) performed a pre-/postnatal developmental toxicity screening study in Charles River albino rats. Lithium carbonate was administered to groups of 10 pregnant females at doses of 0, 0.675, 2.025, or 4.05 mM/kg bw/d (0, 0.9; 28, 57 mg Li/kg bw/d) orally by gavage from gestation day (GD) 14 through postnatal day (PND) 21 of lactation. Dams and their offspring were observed for mortality, normal body weight gain and general symptomatology after 1, 4 and 21 days of nursing plus gross external and internal examination at the conclusion of the study. Beside a minor effect in the form of a reduction in mean body weight gain of pups there was no maternal parameters such as fertility, average number of implantation sites, average litter size, body weight gain, offspring mortality and gross appearance after transverse sectioning or skeletal staining revealed no differences between treated and control groups. The plasma concentration at the dose of 57 mg Li/kg bw/d was 1.4 mM Li (0.9 – 2.8 mM Li). Overall, this pre-/postnatal developmental toxicity screening study is considered as supporting information with only minor restriction regarding reliability (RL2). Finally, the results obtained under the conditions of the study should not lead to concerns regarding prenatal developmental toxicity.

Hoberman et al. (1990) investigated the prenatal developmental toxicity of lithium hypochlorite as characteristically similar read across compound in Sprague-Dawley rats in a study similar to OECD guideline 414. Groups of 25 pregnant rats received lithium hypochlorite at doses of 0 (vehicle-reverse osmosis deionized water), 10, 50, 100, or 500 mg/kg bw/d, via oral gavage once daily on days 6 through 15 of gestation. Significant maternal toxicity was observed in the 500 mg/kg bw/d dosage group, which included maternal death, clinical signs, gross necropsy findings, and inhibited maternal body weight gain and feed consumption. At this clearly maternal toxic dose the only effects on embryo-fetal development were small reversible delays in skeletal growth. Average values for corpora lutea, implantation sites, litter sizes, live and dead fetuses, and resorptions were comparable in the five dose groups and/or were within the range observed in historical controls. The NOAEL for maternal and developmental toxicity for lithium hypochlorite was 100 mg/kg bw/d. The NOAEL for teratogenicity for lithium hypochlorite was 500 mg/kg bw/d, the highest dose investigated. Overall, the study is considered as supporting information with only minor restriction regarding reliability (RL2). This is predominantly due to the slightly shorter exposure period compared to current guideline requirements (gestation day (GD) 6 through GD 15 instead currently GD 6 through GD 19).

Marathe and Thomas (1986) studied the potential prenatal developmental toxicity of lithium carbonate in pregnant female Wistar rats in a respective screening study. The animals were treated orally (gavage) from GD 6 through GD 15 at doses of 0, 50 and 100 mg/kg bw/d. At the dose of 100 mg/kg bw/d there occurred reduction in number and weight of litter, increase in the number of resorptions, wavy ribs, short and deformed bones of the limbs, or an increased incidence of incomplete ossification of sternebrae and wide bone separation in the skull. Based on the results of this study, a NOAEL for prenatal developmental toxicity including teratogenicity of 50 mg/kg bw/d (9.4 mg Li/kg bw/day) was determined. However, the relevance of the effects in fetuses noted at 100 mg/kg bw/ day cannot be assessed as no information on maternal toxicity or incidences in historical controls were provided. Noteworthy to mention that in other studies in rats the dose of 100 mg/kg bw/d was shown to be severely toxic to dams including the occurrence of mortalities.

Overall, with regards to deficiencies in reporting and assessment in the form of low animal numbers in exposed groups, missing comparison to historical control data and especially, no information on maternal toxicity as in other studies mortalities in dams at comparable high doses occurred, this prenatal developmental toxicity screening study was finally considered as not assignable (RL4). Finally, this screening study should not lead to relevant concerns regarding pre-/postnatal development toxicity due to mentioned deficiencies and uncertainties.

Gralla and McIlhenny (1972) performed a prenatal developmental toxicity screening study with a study design in principle similar to former OECD guideline 414 but with lower animal numbers. Ten pregnant female New Zealand albino rabbits per group were dosed orally (capsule) with lithium carbonate in capsules from GD 5 through GD 18 at doses of 49.88 or 79.80 mg/kg bw/d (0.675 or 1.08 mM/kg bw/d). A low incidence of maternal mortality occurred at the highest dose level administered. Three female rabbits, which received 79.80 mg/kg bw/d (1.08 mM/kg bw/d) died late in pregnancy after prolonged anorexia and occasional tremors. One non-pregnant female rabbit receiving 49.88 mg/kg bw/d (0.675 mM/kg bw/d) died unexpectedly overnight. No test item related congenital abnormalities were detected. The NOAEL for maternal toxicity was 49.88 mg/kg bw/d (0.675 mM/kg bw/d) and the NOEL for prenatal developmental toxicity including teratogenicity was the highest dose tested of 79.80 mg/kg bw/d (1.08 mM/kg bw/d). Overall, this prenatal developmental toxicity screening study is considered as supporting information with only minor restriction regarding reliability (RL2). Finally, the results obtained under the conditions of the study should not lead to concerns regarding prenatal developmental toxicity.

In an explorative screening study, Canolty et al (1989) fed two groups of Sprague-Dawley rats ad libitum throughout gestation a control or lithium-containing diet (0 or 750 ppm lithium carbonate). In addition, there was also a control group pair-fed to the lithium group. The rats in the lithium group with a mean serum lithium concentration of 0.48 ± 0.06 mEq/L revealed an increase in fetal resorptions in comparison to both control groups. Beside this, there were no gross fetal malformations in any treatment group and no significant effects on the number of fetuses, total litter weight or mean fetal weight. With regards to maternal effects, no impairment on total body weight change or organ weights were recorded. Relative maternal organ weights expressed at difference between total body weight gain of the dam and the combined weights of her uterus, placenta and fetuses, was lower in dams fed the lithium-containing diet (43 ± 5 g) than in those fed the control diet ad libitum (65 ± 4 g), but not in those pair-fed the control diet (55 ± 6 g). Serum calcium concentration of dams in the lithium group (109 ± 2 µg/mL) was lower than that of dams in the ad libitum-fed (118 ± 1 µg/mL) but not the pair-fed (111 ± 5 µg/mL) control groups. The authors concluded that their results indicated a greater adverse impact on the dams rather than on the surviving fetuses when receiving a dietary concentration of 750 ppm lithium carbonate throughout pregnancy.

However, the reliability of this explorative screening study is considered as questionable. The results are only available in the form of an abstract. Only few experimental parameter were investigated/reported with very limited information regarding study design. There was only one dose group and no information on animal numbers. Thus, no reliable conclusion regarding possible developmental toxicity is possible and therefore, the study was finally disregarded (RL3).


Mouse

In vitro

Klug et al. (1991) developed an in vitro screening method to allow monitoring of the closure of the secondary palate. Various substances, which are known or suspected to induce cleft palates in mice in vivo including lithium were screened. Cultures of palates of NMRI mice were exposed to the substances within a culture period between day 13 – day 17. In this screening system, the tested lithium (not further specified) concentrations of 100 µg/mL and 300 µg/mL and other “positive” substances did not interfere with palate closure. However, the authors assumed that the sensitive phase may have not been covered and additional and modified assays may be necessary.

The reliability of this preliminary and non-validated in vitro screening, which was reported only in the form of an abstract, is regarded as very limited and not relevant for assessing possible developmental toxicity in vivo. Thus, this initial screening study was disregarded (RL3).

In vivo

Szabo (1970) preformed a prenatal developmental toxicity range-finding study in 3 - 4 female HaM/ICR mice. The animals were dosed oral (gavage) with 200, 300 and 465 mg/kg bw/d lithium carbonate from day 6 to 15 of pregnancy. No controls were included in the test. The dams were killed on day 18 of pregnancy and examined. No effects on the dams were reported. The highest dose of 465 mg/kg bw/d, caused increased incidence of death in utero (26 %) and among the 37 live fetuses, 11 were abnormal with cleft palate. The next lower dose of 300 mg/kg bw/d caused no fetal deaths, but 6/50 viable fetuses had cleft palate. The lowest dose of 200 mg/kg bw/d produced no effects of any kind. Thus, fetal effects including cleft palates were only observed at a very high dose, which caused severe maternal toxicity including deaths in a subsequent study. Under the conditions and limitations of the range-finding study, a developmental/teratogenicity NOAEL of 200 mg/kg bw/d in mice was determined.

Overall, due to major methodological deficiencies in the form of no control, no information on maternal toxicity, few animals, deficiencies in reporting and assessment amongst others, the study is of disregarded reliability (RL3). Finally, this prenatal developmental toxicity screening range-finding study should not lead to relevant concerns regarding postnatal development due to deficiencies and uncertainties of this disregarded study and especially as fetal effects were noted exclusively at excessive high dose levels, which caused mortality in the main study, (toxicity) stress in mice is known to induce malformations.

Subsequently Szabo (1970) performed a prenatal developmental toxicity screening study in female HaM/ICR mice, 16 - 20 animals were dosed with 200 and 465 mg/kg bw/d lithium carbonate suspended in 0.5 % tragacanth gel by gavage from day 6 to 15 of pregnancy (Szabo, 1970). The control group received the vehicle. The dams were killed on day 18 of pregnancy. The lowest dose of 200 mg/kg bw/d caused neither maternal nor fetal deaths and no relevant increase in cleft palate in the fetuses. The dose of 465 mg/kg bw/d was severely toxic and caused the death of 37 % of the dams. As a consequence of the severe maternal toxicity, 32 % fetal deaths were observed and 19/121 fetuses in 7/15 letters showed cleft palates. The association of severe maternal toxicity (stress) with an increased incidence of fetuses with cleft palates is known to be a common finding, especially in mice. Therefore, the fetal findings at a maternally lethal dose are considered to be not an indication for a substance specific teratogenicity but a consequence of severe maternal toxicity. The NOAEL for maternal and developmental toxicity including teratogenicity was 200 mg/kg bw/d.

Overall, the study is considered as sufficient for weight of evidence approach as it is comparable to guideline study with acceptable restrictions (e.g., slightly lower animals No.). There were some deficiencies in reporting and assessment but nevertheless, the study is regarded as scientifically reasonable. Fetal effects occurred exclusively at excessive high dose levels, which caused mortality in dams and (toxicity) stress in mice is known to induce malformations. Finally, the results obtained under the conditions of the study should not lead to concerns regarding prenatal developmental toxicity.

Smithberg and Dixit (1982) investigated the possible teratogenic potential of lithium carbonate in a prenatal developmental screening study in mice. Mated strain J/A female mice were intraperitoneally injected each separately on gestation day 8, 9 or 10 and on gestation days 12 through gestation days 14 with doses of 0.8 - 3.2 mg/mouse (32 - 128 mg/kg bw/d). Control females received either distilled water or sodium chloride of equal molarity. These females were autopsied on day 18 or 19 post coitum and their skeletons prepared as above and studied for malformations. Under the conditions of the present study and considering the limitations, a NOAEL for prenatal developmental toxicity including teratogenicity of 3.2 mg/mouse (128 mg/kg bw/d) was observed.

However, it has to be considered that the intraperitoneal exposure route is not appropriate for risk assessment. Moreover, due to major methodological deficiencies in the form of short exposure periods to investigate specific time-dependent effects, no information on maternal toxicity, deficiencies in reporting and assessment amongst other, the study is disregarded (RL3). Finally, the results obtained under the conditions of the study should not lead to relevant concerns regarding postnatal development due to deficiencies and uncertainties of this disregarded study.

The same working group (Smithberg and Dixit, 1982) performed an additional prenatal developmental toxicity screening study on lithium carbonate in another strain of mice with a slightly different study design. Mice of the 129 Sv/SL strain were intraperitoneally injected each separately on gestation day 8, 9 or 10 of gestation with doses of 0.8, 1.6, 3.2 and 5.0 mg/mouse (0, 32, 64, 128 and 200 mg/kg bw/d). The lowest dose of 0.8 mg/mouse (32 mg/kg bw/d) did not induce a significant increase in malformations. A dose of 5.0 mg/mouse (200 mg/kg bw/d) did significantly increase malformations (41.6 %), particularly when administered on day 9 of pregnancy. This same high dose produced malformations of 19.3 % and 17.3 % when administered on day 8 or 10, respectively. The malformations consisted of fused ribs and/or vertebral defects and exencephaly.

In this study very high doses were tested that might have caused maternal stress which is known to lead to an increase in malformations. The observed fetal findings cannot be assessed because no correlation to maternal toxicity is possible as maternal findings were not reported. Moreover, no information on historical control data were given. Under the conditions of the present study and considering the given limitations, a NOAEL for prenatal developmental toxicity including teratogenicity of 3.2 mg/mouse (128 mg/kg) was observed).

However, it has to be considered that the intraperitoneal exposure route is not appropriate for risk assessment. Moreover, due to major methodological deficiencies in the form of short exposure periods to investigate specific time-dependent effects, no information on maternal toxicity, deficiencies in reporting and assessment amongst other, the study is disregarded (RL3). Finally, the results obtained under the conditions of the study should not lead to relevant concerns regarding postnatal development due to deficiencies and uncertainties of this disregarded study, especially as fetal effects occurred exclusively at a dose level, which was overt toxic in other studies .

In a further prenatal developmental toxicity screening study of Smithberg and Dixit (1982), 16 mice of the 129 Sv/SL strain were treated with 2 mg/mL lithium carbonate (corresponding to 2129.4 mg/kg bw/d or 400 mg Li/kg bw/d) in drinking water from day 1 - 18 of gestation. Only two litters of 16 dams treated survived the treatment until day 18 post coitum. There were four fetuses in one, and two in the other dam. All were apparently normal following skeletal staining and inspection. The resorption level was extremely high, 60 % of all implantation sites counted.

However, it has to be considered that, due to major methodological deficiencies in the form of no control group, deficiencies in reporting and assessment amongst other, the study is disregarded (RL3). Finally, the results obtained under the conditions of the study should not lead to relevant concerns regarding postnatal development due to deficiencies and uncertainties of this disregarded study, especially as no fetal effects were observed at a dose level, which was overt toxic to dams.

Loevy and Catchpole (1973) investigated the potential prenatal developmental toxicity of lithium chloride monohydrate in a screening study in CDI mice. A dose of 15.5 mg of lithium chloride monohydrate/mouse (corresponding to 516.7 mg LiCl/kg bw/day or 450.3 mg/kg bw/d) in sterile water was injected subcutaneously each day for 2 or 3 days on days 11 through 13 of pregnancy. The mice were sacrificed on day 17 of pregnancy. The uteri were examined for resorption sites and the fetuses for malformations.

The subcutaneously injected high dose led to an increase in resorption sites (11 - 21), which was not further specified. This high dose led also to an incidence of cleft palate in the offspring of mice injected on days 11, 12, and 13 of 15.1 %; on days 12 and 13 of 7.2 %; and on days 11 and 12 of 3.4 %. The authors reported no maternal toxicity. However, an increase in resorption in mice can also be considered as an indication for maternal toxicity/stress. The latter is known to lead also to an increase in malformations, e.g. in cleft palates. Under the conditions and limitations of the present study, the deduction of a NOAEL was not possible (single dose).

However, it has to be considered that the investigation of one dose only and s.c. injection as exposure route  is not appropriate for risk assessment. Moreover, due to major methodological deficiencies in the form of short exposure periods to investigate specific time-dependent effects, deficiencies in reporting and assessment amongst other, the study is disregarded (RL3). Finally, the results obtained under the conditions of the study should not lead to relevant concerns regarding postnatal development due to deficiencies and uncertainties of this disregarded study, especially toxicity) stress in mice is known to induce malformations.

Monkey
Gralla and McIlhenny (1972) performed a prenatal developmental toxicity study in female rhesus monkeys similar to OECD 414 respectively ICH Segment II study, which were successfully mated with mature males. Six females were dosed with lithium carbonate at 49.51 mg/kg bw/d (0.67 mM/kg bw/d) by capsule during days 14 through 35. Five additional female monkeys served as controls and received empty capsules. The offspring were either developed by cesarean section or the females were allowed to deliver naturally on day 160±2. Lithium carbonate had no effect on reproduction in rhesus monkeys. A total of 7 normal progeny, 2 females and 5 males, were delivered from treated pregnant rhesus monkeys. All were normal and comparable to 3 females and 1 male delivered by the control group. All parameters examined in all infant monkeys were normal and all grew up without showing clinically any physical defects at 12 - 15 months of age. Thus, the dose of 49.51 mg/kg bw/d (0.67 mM/kg bw/d) investigated was a clear NOAEL for maternal and prenatal developmental toxicity including teratogenicity.

Overall, this prenatal developmental toxicity study is considered as supporting information with only minor restriction regarding reliability (RL2). The results obtained under the conditions of the study should not lead to concerns regarding prenatal developmental toxicity.

Pig
Kelley et al. (1978) exposed 12 pregnant pigs orally to 3000 mg/kg of lithium carbonate in the diet (corresponding to 36 mg/kg bw/d) during gestations days 30 through 114 in a pre- and postnatal developmental toxicity screening study. The control group consisted of 11 females. The pups were observed until postnatal day 21. The treated dams showed reduced body weights after 110 days. Five out of 12 pigs did not complete pregnancy. Average offspring born per litter and birth weight of the piglets did not differ. At this maternally toxic dose, there was a decrease in the number of piglets born alive and an increase in stillbirth and mummies in treated female pigs. In these females the litter weight was reduced. Postnatal mortality in piglets from treated females was increased but growth rates did not differ from control piglets. No abnormalities in the pups were reported.

Overall, with regards to the use of a single dose only, limited study design, deficiencies in reporting and assessment, missing comparison to historical control data, this pre-/postnatal developmental toxicity screening study was finally considered as not assignable (RL4). This screening study should not lead to relevant concerns regarding pre-/postnatal development toxicity due to mentioned deficiencies and uncertainties.

 

Human data (literature)

The effect of lithium carbonate was investigated in a prospective multicentre study of pregnancy outcome after lithium exposure during first trimester in pregnant women using lithium. The study showed that women with major affective disorders who wish to have children may continue lithium therapy, provided that adequate screening tests, including level 11 ultrasound and foetal echocardiography, are done (Jacobsen et al 1992). Babies of mothers treated with lithium carbonate in the first trimester were analysed for the potential malformation to the unborn. The data obtained did not reveal any increased frequency of physical or mental anomalies among the lithium children (Schou, M.; 1973; Schou, M.; 1976). Further studies on effects of lithium during pregnancy with ambiguous results (potential teratogenic target of lithium in humans: cardiovascular system) do not allow any conclusion with regard to the potential effects of lithium carbonate as all patients in this study were ill (manic depressive) and effects of confounding factors cannot be excluded. In addition, the cohort size was too small and bias effects are likely (Kallen, B. (1983). Possibly, cardiovascular malformations are specific to humans or to humans with coexisting psychiatric disorder (Giles, J.J., 2006) but based on a case-control studies, analysing cases of Ebstein's anomaly, also no clear conclusion could be drawn that lithium exposition during pregnancy is linked to an increase rate for Ebstein's anomaly (Zalzstein, E. et al, 1990, Correa-Villasenor, A. et al., 1994). Also in a study to quantify lithium exposure in nursing infants in 10 mother-infant pairs no serious adverse events were observed, and elevations of thyroid-stimulating hormone, blood urea nitrogen, and creatinine were few, minor, and transient and not considered of biological relevance (Viguera, A.C. et al., 2007).

Summary, discussion and overall conclusion
The general aim was to identify possible critical findings and issues of potential authority concerns with regards to the reproductive and developmental toxicity profile of Lithium Carbonate.
In principle, Lithium in the form as carbonate (CAS 554-13-2) and partly as chloride and hypochlorite is comprehensively investigated with regards to its reproduction/fertility and developmental toxicity profile. Both, guideline conform OECD TG 416 and TG 414 studies in rats as well as explorative screening studies, predominantly concerning prenatal developmental toxicity are available in rats, rabbits, mice, monkeys and pigs. Thus, the reproductive toxicity profile of lithium carbonate can be considered as sufficiently and appropriately examined.

With regards to developmental toxicity, the most reliable and key study is the prenatal developmental toxicity study performed in Wistar rats (LPT, 2010) according to OECD TG 414 under GLP conditions covering a sufficient dose range between 10 mg/kg bw/d – 90 mg/kg bw/d. Clear NOAELs were obtained. The NOAEL (in fact a NOEL) for maternal toxicity is 30 mg/kg bw/d and that for prenatal developmental toxicity including teratogenicity is 90 mg/kg bw/d, demonstrating the lack of developmental toxicity even at the systemic toxic dose level.

In addition, the prenatal developmental toxicity study in rabbits, performed by Gralla und McIlhenny, 1972, is considered as a sufficient study within a weight of evidence (WoE) approach. For WoE this study together with the other prenatal developmental toxicity studies, performed in mice, monkeys and rats can be used. Within all these studies, fetal effects, especially in mice, occurred exclusively at excessive high dose levels. These high dose levels induced overt signs of toxicity up to mortalities. In contrast, lower dose levels with no or only minor signs of maternal toxicity did not lead to signs of developmental toxicity. In most of the cases, clear NOAELs for developmental toxicity, generally as high or higher than the NOAELs for maternal toxicity were obtained. Taken all the studies together, the results obtained under the respective conditions of the studies should not lead to concerns regarding prenatal developmental toxicity.

Beside the above mentioned studies there are a number of additional literature data on the developmental toxicity of lithium carbonate in rats (Fritz, 1988; Marathe and Thomas, 1986) and in mice (Smithberg and Dixit, 1982; Loevy and Catchpole, 1973; Szabo, 1970 (range-finding study); Mroczka et al., 1983) available. However, these studies have to be disregarded due to their questionable reliability. All of them have several limitations either due to missing controls and/or a single dose, no appropriate exposure route (intraperitoneal, subcutaneous) and/or limited exposure periods down to single days only. Deficiencies in reporting and assessment were also evident in these studies. In most cases effects on development were only observed at maternally overt toxic and/or lethal dose levels.

This is consistent with the conclusion of an earlier review performed by Leonard et al. already in 1995 (see IUCLID section 7.12). These authors concluded at that time a considerably variation in the results of the investigated animals. Several types of abnormalities (e.g. reduced number and weight of the litter, more resorptions, 'wavy' ribs, incomplete ossification) were observed by some authors but not by others. They assumed that these discrepancies might be due to a different sensitivity of the species and strains used, the stress of daily injection and/or differences in lithium concentrations present in serum during critical periods of development. This was based on observations that pregnant mice given lithium carbonate over several days yielding serum levels comparable to those in man treated for manic-depressive disorders did not show any effect, but six times higher doses caused malformations in the offspring. Chronic exposure to lithium doses that produced serum levels of the same order as seen in patients was toxic but did not affect the entire litter nor was it teratogenic to individual embryos.

Finally, the literature data of lithium carbonate supported the conclusion that in general the developmental toxicity profile was sufficiently and appropriately examined. All of these studies supported the findings observed in the key study. Moreover, all disregarded screening studies should not lead to relevant concerns regarding pre- and/or postnatal development toxicity due to mentioned deficiencies and uncertainties.

Taken all the studies together, there is no need to investigate lithium carbonate again in a second species as sufficient information on the developmental toxicity profile is available in other species beside the rat.

Justification for classification or non-classification

Classification, Labelling, and Packaging Regulation (EC) No 1272/2008
The available experimental test data are reliable and suitable for classification purposes under Regulation (EC) No 1272/2008. Based on the results obtained from reproduction/developmental testing, the test item is not classified according to Regulation (EC) No 1272/2008 (CLP), as amended for the tenth time in Regulation (EU) No 2017/776.