Digital Object Identifier (DOl) 10.1007/s002040000110

Received:20 December 1999 / Accepted: 1 March 2000 /Published online: 24 May 2000

Springer-Verlag 2000

D-5 1368 Leverkusen, Germany

Sir,

MuBhoff et al. (1999) published results from patch clamp investigations in oocytes, which showed that 2-phenoxyethanol had an influence on the NMDA (N-methyl-D-aspartate) and kainic receptor. They used the Xenopus laevi oocyte model for testing diverse glycol ethers including 2-phenoxyethanol. Their results showed that among the 17 glycol ethers investigated, only 2-phenoxyethanol caused a noteworthy reduction in NMDA-induced membrane currents. The reduction was rather weak, i.e., 8% at a concentration of 100 uM, and the effective concentration for 50% of the group (EC50) was 362 uM. They interpreted this result as an indication of "a strong neurotoxic potential" being "in line with observations on three patients, showing an immediate and delayed neurotoxic effect of this compound", referring thereby to a publication by Morton (1990). We would like to comment on these results and their interpretation as well as on some generalizing statements they made in their paper.

2-Phenoxyethanol is and has been widely used as a cooling lubricator, antibacterial and antifungal preservative in cosmetics and pharmaceuticals, fixative and anesthetic in fish hatcheries. The toxicological profile of 2-phenoxyethanol has been studied in numerous experiments with many different endpoints using a number of species (for review, see Greim 1998). In addition, a few studies in humans have been performed. In none of the animal experiments did neurotoxic properties of 2-phenoxyethanol become apparent, even at massive, sublethal dose levels administered over months. Despite the wide range of applications and long-term use with probably considerable dermal exposure in certain applications, reports of systemically toxic effects in humans are extremely rare. Due to its rather low vapor pressure (0.04 hPa at 20 °C), a significant exposure via inhalation is not to be expected under normal circumstances. After intestinal and dermal absorption, 2-phenoxyethanol is rapidly metabolized to phenoxy acetic acid and conjugates thereof. More than 90% of all the metabolites are rapidly eliminated via urine. Therefore only small amounts of 2-phenoxyethanol and its metabolites are to be found in organs and tissues, and there is no accumulation potential.

In view of these pharmacokinetic properties it is questionable whether in vitro investigations on the parent compound have much relevance to the in vivo situation.

The aforementioned report of Morton (1990), which Mubhoff et al. consider agrees with their results, describes three women in a fish hatchery who used 2-phenoxyethanol for anesthetizing small salmon. They complained about diminished strength and sensation in their hands and fingers, headaches, dizziness, tiredness, euphoria, grogginess, feeling "drunk", slurred speech, and lightheadedness. One to two years after stopping exposure, they reported problems with constant irritability, forgetfulness, and difficulties maintaining concentration, which have led to major difficulties at work and at home. Furthermore, 2/3 patients reported an alcohol intolerance. While several neuropsychological tests confirmed cognitive impairment, etc., other tests (nerve conduction, computed tomography of brain) failed to show abnormal effects. Personal discussions between Prof. Morton and one of the present authors (W.S.) revealed that examinations were done months after the cessation of exposure, that other workers in this hatchery (and in other hatcheries, as this use of 2-phenoxyethanol was not unusual) doing the same work clearly had no symptoms. Even rough estimates of the most likely dermal exposure of these individuals were not possible due to a change in the anesthetic used, the time lag between the exposure and the diagnosis, and the lack of blood and urine samples. A number of confounders, like the use of a ‘cleaning agent", nicotine, or alcohol, were not taken into account, and the pattern of appearance/disappearance of symptoms did not consistently follow periods of the most intense use of the anesthetic. Severe symptoms were recognized only 4 years after cessation of exposure. Therefore, and because this is the only report attributing neurotoxicity to 2-phenoxyethanol poisoning after decades of uses in many different applications including fish anesthetizing, there are serious doubts about whether there is a causal relationship between exposure to 2-phenoxyethanol and these effects.

Even if one ascribes the observations made in these three patients to 2-phenoxyethanol they do not fit well the proposed mechanism by Mubhoff et al. in our opinion for the following reasons: (a) a qualitatively similar reduction of the NMDA-induced membrane currents was found also with neuroprotective compounds like MK-801 or L 701.324. These compounds were developed to protect the neurons from overstimulation ("excitotoxicity") of the NMDA receptor with glutamate, as occurs in epilepsy or stroke (Anderson et al. 1987; Obrenovitch and Zilkha 1996); (b) an important property of the NMDA receptor in the brain is the regulation of long-term potentiation (LTP) and long-term depression (LTD). LTP and LTD are possible surrogates of memory (LTP) and forgetting (LTD). The intensity of a stimulus is the decision point between memory and forgetting. However, this mechanism is not only regulated exclusively by this receptor. This was shown with NMDA antagonists like 2-amino-5-phos-phonovaleriate (2-AP5) and MK-801, which did not induce LTD in contrast to compounds activating the voltage-gated Ca²⁺ channels or the gamma-aminobutyric acid (GABAA) receptor (Akhondzadeh and Stone 1995; Palmer et al. 1993). However, 2-AP5 inhibits the induction of LTD (Gean and Lin 1993). In summary, the inhibition of the NMDA receptor does not lead necessarily to an impairment of memory functions nor to other symptoms seen in those patients.

Certainly, there are no concrete hints for alcohol abuse in the three patients described by Morton. However, ethanol may well serve as an example for discussing clinical and receptor effects. High acute and/or chronic ethanol abuse cause, for example, memory loss, depression, euphoria, slurred speech, etc. Also, ethanol has an inhibitory effect on the NMDA and GABA channel, but it is not a specifically antagonistic one (Gilbert and Burdette 1995). Solvents and narcotic compounds administered at high concentrations to in vitro systems change neuronal membranes in a dramatic way and therefore lead to massive changes in the electrical signaling of the cell. These properties are mostly seen on a variety of receptors: in the case of ethanol, it is also the muscarinergic acetyicholine receptor, in addition to NMDA and GABA (Little 1991; Pick et al. 1993). Such rather non-specific receptor interactions have to be discriminated from specific ones which always occur at much lower concentrations. In our opinion, it is highly unlikely that such high concentrations (as used in vitro) of 2-phenoxyethanol can be reached in vivo under the exposure conditions described. Taken together, the data presented by Mulbhoff et al. neither indicate a "strong neurotoxic potential" nor really fit the observations of Morton.

Apart from these more scientifically oriented discussions, we are deeply concerned about the way in which language and contents are dealt with in this paper, and how some single and isolated observations are being generalized. This is exemplified by the following citation:

"Effects on the central nervous system additionally have to be expected, since the GE [glycol ethers] are also capable of rapidly crossing the blood-brain barrier (Ahmed et al. 1983). In fact, disturbances of brain functions, for example headache and persistent cognitive impairment, have been found in humans exposed to GE (Morton 1990; Ohi and Wegmann 1978; Parsons and Parsons 1938; Rowe and Wolf 1982)."

Comment

The publication of Ahmed et al. (1983), which is presented twice in the paper as proof that GE rapidly reach the target CNS, deals exclusively with acrylonitrile and is therefore totally unsuitable for the purpose of this argumentation. Of the four papers cited to indicate disturbances of brain function by GE, that of Morton has been discussed above. The other three citations deal with cases where people have been highly exposed to 2-methoxyethanol only. This compound is certainly quantitatively as well as qualitatively different from several other GE and especially 2-phenoxyethanol (Hardin 1983; Hobson et al. 1986). Therefore, we feel that a generalization from essentially one compound (with proven neurotoxic side-effects) to a whole class of compounds with quite different physicochemical and biological activities is not justified. In addition, the predictive value of the model of MuBhoff et al. is refuted by the fact that 2-methoxyethanol, which in contrast to 2-phenoxyethanol is said to be very slowly metabolized in the body (Rowe and Wolf 1982), has not shown any activity in their model up to the very high concentrations they used.

At the moment, we are witnesses to a stage in the history of toxicology in which a high degree of uncertainty is generated in the general public as well as among the politicians and authorities in charge by questionable results obtained with non-validated methods, dubious study approaches without reproducibility, inadmissible extrapolation of in vitro investigations to in vivo conditions, etc. Responsible handling of scientific findings (even if they are probably reproducible as in the present case) is therefore absolutely imperative today, especially when whole classes of industrial products of enormous importance which reach the end consumer are affected.

References

Rowe VK, Wolf MA (1982) Derivatives of glycols. In: Clayton GD, Clayton FE (eds) Patty’s industrial hygiene and toxicology, 3rd edn. Wiley, New York, pp 3911—3919