The marketing stance for the commercial rollout of Zopiclone (sold as Lunesta) in the United States is that Zopiclone is not addictive. However, there is evidence that Zopiclone and the other popular non-benzodiazepine sleep aid Zolpidem both cause tolerance (people increase their dose with regular use) and withdrawal (people report substantial difficulty sleeping or other side effects when stopping after regular use). Although many of the U.S. news stories about Lunesta fail to mention it, Zopiclone has been available in Europe and around the world since November 1989.

The new U.S. product, Lunesta, is being sold in 1, 2, and 3 milligram tablets of just the S-isomer (they have named this eszopiclone), which is considered to be the primary active isomer. The European brands (Imovane, etc.) contain racemic mixtures and are sold in 5.0 and 7.5 mg dosages. It is possible that there is a lower risk of addiction because of the difference in dose and isomers, but this remains to be seen as the general public in the U.S. is exposed to it.

A number of early news reports about the FDA's approval of Zopiclone for sale in the U.S. under the brand Lunesta included comments that suggested it was "non-addictive" and could be taken safely for long periods of time. The Jan 25, 2005 Washington Post story about the approval stated, with a hint of ironic disbelief:

It sounds like an insomniac's dream: a sleeping pill that can be taken for weeks or even months at a time, without the risk of addiction or morning-after grogginess.

In the next several weeks consumers will see splashy print and television ads touting the lyrically named Lunesta, which was approved last month by the U.S. Food and Drug Administration (FDA). Unlike other sleeping pills, including market leader Ambien, which are not supposed to be taken for longer than 10 days at a time, Lunesta has no FDA-recommended time limit.

A news story in the L.A. Times also reported that Lunesta could be taken for long periods without any tolerance or withdrawal:

The studies of Lunesta found that patients taking it for the long haul did not build up a tolerance for drug the a concern that worries many doctors who prescribe sleep aids for long periods. And when its use was stopped "cold turkey," according to Dr. Andrew Krystal, who ran one of the trials at Duke University, researchers found few to none of the symptoms withdrawal including delirium and a worsening of symptoms insomnia that patients sometimes experience when discontinuing other prescription sleep aids. [Source: LA Times, 17 January 2005]

However, the official literature produced by the pharmaceutical company and the FDA both include mentiones of tolerance and withdrawal, although they are clearly presented as negligible. A 25-page article on the research leading to the FDA approval from the official Lunesta website can be found:
http://www.lunesta.com/PostedApprovedLabelingText.pdf

and a similar, but not identical, document on the FDA's website:
http://www.fda.gov/cder/foi/label/2004/021476 lbl.pdf

These both include several mentions of withdrawal:

The clinical trial experience with LUNESTA revealed no evidence of a serious withdrawal syndrome. Nevertheless, the following adverse events included in DSM-IV criteria for uncomplicated sedative/hypnotic withdrawal were reported during clinical trials following placebo substitution occurring within 48 hours following the last LUNESTA treatment: anxiety, abnormal dreams, nausea, and upset stomach. These reported adverse events occurred at an incidence of 2% or less....

No development of tolerance to any parameter of sleep measurement was observed over six months.
[NDA 21-476 Approved Labeling Text - Dated December 15, 2004]

There are concerns that the problems associated with frequent use and addiction are underreported. The MedSafe Editors state:

Withdrawal effects & dependence may occur & are being under-reported At June 1997, the WHO database held 46 reports of dependence and 42 of withdrawal syndrome with zopiclone. From March 1994 to June 1997, Rhône-Poulenc Rorer received 17 spontaneous reports of dependence and 13 of withdrawal syndrome or symptoms with zopiclone.5 These figures suggest a very low rate of occurrence of these problems. However, during a 2.5 year period, Tranx Services, Auckland (an organisation that assists people to withdraw from addiction to minor tranquillisers) saw 24 clients seeking help for dependence who were taking zopiclone.6 As only a small proportion of dependent people seek assistance from addiction services, extrapolating this figure would suggest that there is significant under-reporting worldwide.
[medsafe.govt.nz : Dependence with Zopiclone, July 1998 : http://www.medsafe.govt.nz/Profs/PUarticles/3.htm ]

There are medical case reports of addiction to zopliclone, including the following case:
http://www.smw.ch/pdf200x/2002/35/smw-10074.PDF

Here are some excerpts, emphasis added:

We report a case of excessive use of zopiclone and its withdrawal. A 67-year-old man used almost 340 mg/d to treat his insomnia... His insomnia started after a surgery and his first divorce. The patient started using sedatives in 1967. In the beginning he took flunitrazepam - prescribed during his hospitalisation up to 3 mg per day for about 20 years (1 to 4 mg on a daily basis), later he was put on zolpidem (approximately for half a year, using 20 mg/d) and zopiclone (starting with 7,5 mg/d). His second wife was diagnosed with cardiac insufficiency in 1999. From this time on our patient augmented his use of zopiclone up to 337.5 mg daily. ... In normal volunteers changes in sleep pattern and rebound anxiety appeared upon discontinuation of the drug (7,5 mg/d for 21 days). Withdrawal reactions like headache, anxiety or agitation have been found only in 0.05% (N = 7) of 13177 subjects of a prescription-event monitoring study. Those patients had been taking up to 225 mg/d for up to 3 months. Our patient used alcohol and sedatives to cope with insomnia and psycho-social stress. He could easily stop drinking 1.5 litres beer each day but had severe problems in reducing zopiclone. It is noteworthy that our patient could stop taking diazepam completely.

The British Medical Journal also reported several cases of Zopiclone addiction in 1998:
http://bmj.bmjjournals.com/cgi/content/full/316/7125/117

... Claims that zopiclone does not cause rebound, dependence, or withdrawal phenomena may have led to it being considered a safe option for treating insomnia. We present a series showing evidence of zopiclone dependence and problems caused by withdrawal.

Case 1 : A 29 year old man who had had a pneumothorax was prescribed zopiclone 7.5 mg nightly. As he was anxious about a recurrence he increased the dose to 22.5 mg supplemented with tablets from other sources. After eight months he realised he was misusing the drug and stopped taking it suddenly. This resulted in severe anxiety with tachycardia, tremor, sweating, and rebound insomnia. He was treated with amitriptyline 25 mg three times daily for several weeks and made a full recovery.

Case 2 : A 26 year old man was prescribed zopiclone 7.5 mg nightly for insomnia. The dose was eventually increased to 7.5 mg four times daily, depending on availability. If he stopped taking zopiclone he experienced anxiety, tremors, sweats, flushes, palpitations, and derealisation. He was stabilised with 7.5 mg zopiclone nightly and 3.75 mg daily and monitored closely. He described a strong craving for the drug, and attempts to reduce the dose further were firmly resisted.

Case 3 : A 49 year old woman received zopiclone 7.5 mg for insomnia during inpatient treatment for depression. She was taking this dose at discharge but subsequently increased it to 15 mg and then 22.5 mg. She remained on 22.5 mg one year later. She had tried to stop taking zopiclone twice but experienced severe rebound insomnia and anxiety. She was therefore reluctant to reduce the dose.

Case 4 : A 36 year old woman with bipolar affective disorder was prescribed zopiclone 7.5 mg nightly for insomnia. The following year she was taking 7.5 mg four times daily, obtaining supplies by changing doctors. If she reduced her intake suddenly she experienced sweating, palpitations, tremor, and anxiety. Her intake was monitored and the dose was gradually reduced. She had been dependent on benzodiazepines although these were stopped before zopiclone was prescribed. She had no history of alcohol dependence.

All these patients increased their intake of zopiclone above the dose initially prescribed and their withdrawal symptoms included craving, anxiety, and insomnia. Although it has been suggested that zopiclone is not associated with dependence or withdrawal phenomena, rebound insomnia was experienced by normal volunteers after taking the drug for only two weeks. ...

Studies of up to four weeks of zopiclone use have not demonstrated evidence of dependence or withdrawal problems. Our cases reveal problems with use over a considerably longer period which may more closely resemble clinical practice.
[Jones IR, Sullivan G "Physical Dependence on Zopiclone: case reports". BMJ 1998; 316(117) Jan 1998.]

The Feb 14, 2005 DEA scheduling announcement that came out of the FDA approval states "Clinical trials indicate that withdrawal effects from eszopiclone are similar to those of benzodiazepines."
http://www.deadiversion.usdoj.gov/fed_regs/rules/2005/fr0214.htm