The Q–T interval and the Mellaril story: a cautionary tale

By Edward Shorter

Is a lengthening of the Q–T interval in the ECG benign or pathological in drug action?  This produced a small controversy in the 1960s that had a major impact on patient care.  In 2000, the Novartis Company cautioned physicians about further use of the antipsychotic drug Mellaril (thioridazine).  The company announced that the drug can entail dangerous cardiac complications.  This information was already known in the mid-1960s, and not only did Sandoz (one of the predecessor companies of Novartis) ignore it, they attempted to discount it at scientific meetings and disregarded the warnings of several clinical scientists.  Moreover, in various ad campaigns Sandoz showed elderly “patients” in the artwork, emphasizing that the drug was suitable for geriatric cases, precisely the population most at risk of such complications.  The story is a textbook case of ignoring scientific warnings in favor of corporate interests.

It was known early on that Sandoz’s new antipsychotic agent thioridazine (Mellaril), launched in the United States in 1959, lengthened the Q–T interval.  But was this good or bad?

There was the benign repolarization school.  In 1964, M.H. Wendkos, a cardiologist at the Veterans Administration Hospital in Coatsville, Pennsylvania, published a paper on pharmacologic studies in a hitherto unreported “benign repolarization disturbance among schizophrenics” (Wendkos 1964).  Wendkos re-stated this position in his presentation at a psychopharmacology meeting in Quebec (see below), arguing that the recorded ECG changes “represent a benign repolarization disturbance rather than an adverse cardiac effect” (Wendkos 1965).

But events were in the saddle, and galloped in a very different direction.  Some background:  It happens quite frequently that drugs are withdrawn or new warnings of their side-effects are circulated.  Yet the story of Sandoz’s antipsychotic medication Mellaril (thioridazine) represents an almost textbook case of a company marching into trouble by ignoring warnings.

On July 31, 2000, Novartis Pharmaceuticals sent a letter to all physicians and pharmacists in Canada, warning that the use of the drug Mellaril should be significantly curtailed.  The preparation should henceforth be restricted only to those schizophrenic patients “who fail to show an acceptable response . . . to other antipsychotic drugs.” The reason?  “Mellaril has been shown to prolong the QTc interval in a dose related manner, and drugs with this potential, including Mellaril, have been associated with torsade de pointes-type arrhythmias and sudden death” (Novartis Pharmaceuticals Canada, 2000).

Simultaneously, the August 18 issue of the Psychiatric News cautioned its readers that thioridazine “will include a new boxed warning regarding potentially fatal cardiovascular effects and will be restricted to second-line use.”  The reason again was that “TdP (torsades de pointes) develops spontaneously, usually without warning, and requires immediate emergency intervention.”  The note stated that the risk of sudden death was “high” (Psychiatric News 2000).

These warnings came more than thirty years too late.  Here is how the controversy unfolded:

In 1963, H.G. Kelly and coworkers in the Faculty of Medicine of Queen’s University in Kingston, Ontario, reported 28 electrocardiograms that depicted a quinidine-like effect of thioridazine on ventricular repolarization (prolongation of the QT interval) in doses as low as 200 mg. a day.  T-waves were flattened out and sometimes inverted, occasionally S–T segments became convex and new waves appeared.  In that study, two fatal cases of arrhythmia occurred (Kelly et al., 1963).

By this time the Sandoz company, of course, knew of the Queen’s University deaths, and their medical advisor, Roy Stewart, a Montreal cardiologist, brought this to the attention of Thomas Ban, chief of the clinical research service at Verdun Protestant Hospital, a psychiatric inpatient facility in the outskirts of Montreal.  It was at Stewart’s request that Ban designed a clinical study, conducted in collaboration with André St. Jean, Scientific Director at Hôpital des Laurentides in L’Annonciation, Quebec, comparing the effects of thioridazine, chlorpromazine, and trifluoperazine on the ECG.  In 1964 the investigators reported that thioridazine “modifies the terminal portion (S–T segment, T and U waves) of the human ECG.”  They  found that, whereas similar changes took place in only 1 of the 6 subjects taking trifluoperazine, and in 3 of 6 taking chlorpromazine, such changes were noted in all 6 of the 6 patients on thioridazine by the 8th day of drug administration, i.e., with 200 to 400 mg of thioridazine per day (Ban and St. Jean, 1964).

The study had been completed in 1963, but before it was published, the following incident occurred at Hôpital des Laurentides:  A patient who had been receiving high (1500 mg per day) doses of thioridazine over a period of ten weeks, suddenly became unconscious and passed into a state of shock.  It happened that there were two physicians in the room, one of them a cardiologist.  An ECG demonstrated ventricular tachycardia.  It was noted that a prior ECG of the patient, six weeks after the initiation of thioridazine therapy, had shown bradycardia and prolongation of the QT (Desautels et al., 1964).

These findings led Ban and co-workers to conduct a survey to determine the incidence of cardiac conductance changes with thioridazine.  It was clear that such complications existed, but what was the size of the problem?  Ban presented the results later in 1964 at the fourth congress of the Collegium Internationale Neuro-Psychopharmacologicum (CINP) in Birmingham, England.  Of the 92 patients receiving drugs other than thioridazine, 12, or 13 percent, displayed an abnormal ECG.  Seventeen, or 77.3 percent of all patients receiving thioridazine, manifested abnormal ECG's (Ban et al., 1965).

In 1964 or 1965 Ban travelled to Basel to report these findings to Sandoz and met with the president and head of pharmacology of the firm (Ban TA, personal communication to E Shorter, 11 Mar 2013).

On June 4, 1965, the Quebec Psychopharmacological Research Association organized a special symposium at the Hôpital des Laurentides on ECG changes with psychoactive drugs.  Ban and coworkers reviewed the aforementioned studies as well as some findings based on a further series of four studies which indicated that “the lowest dose (of thioridazine) which brought about changes  was 150 mg per day” (St Jean et al., 1965).  At the same meeting Edward Kingstone, in his review of the literature on “neuroleptic drugs and the ECG,” pointed out that in 1964 Graupner and Murphree also described ECG changes associated with the use of thioridazine (Kingstone, 1965).  Of the 55 patients they studied, 44 percent developed abnormal electrocardiograms.  Most of the changes were concerned with the T-wave.  They appeared at all dose levels from 150 to 900 mg per day (Graupner and Murphree, 1964).

In organizing the symposium, Ban wanted to ensure that a fair picture of Mellaril was offered.  He had mentioned the meeting to Sandoz, and the company paid the travel cost for Wendkos to attend (Ban, 2011).

Here is where events took over.  Other investigators began learning of the cardiac dangers of thioridazine.  In the mid-1960s, Louis Gottschalk, then at the Cincinnati General Hospital, warned Sandoz privately that Mellaril was dangerously increasing the QT interval.  Gottschalk later said in an interview, “We got the idea to find out whether there are any differences in the psychoactive drug metabolites in people that get these cardiac irregularities.  And lo and behold, we did discover that a metabolite that is not  psychoactive, sulforidazine, does have an adverse cardiovascular effect . . . and  [we] tried to get the drug companies to provide further financial support so we could study the biochemical basis. . . .  But they were doing so well marketing their drugs, that they would not fund it” (Gottschalk, 2011a).  Gottschalk, who in the meantime had moved to the Irvine campus of the University of California, reported with co-workers the existence of this previously unknown metabolite of mesoridazine and thioridazine in 1974 (Dinovo et al., 1974); details of a GLC analysis followed in 1976 (Dinovo et al., 1976).

Did Sandoz then become interested?  Not really.  Gottschalk later said, “Everybody told me that the metabolite was not pharmacologically active.  I asked the head of the organic chemistry department at UCI whether she could manufacture it for me because I wanted to test the effects of the metabolite on cardiovascular function in dog experiments.  She could do it for a certain amount of money, but I never was able to obtain the necessary funds.  In general, pharmaceutical companies are not very interested in trying to discover what triggers the adverse side effects of drugs” (Gottschalk, 2011b)..

Gottschalk was not the only researcher to be brushed off by Sandoz.  In 1974 Donald Gallant and co-workers at Tulane University reported a double-blind ECG comparison of thioridazine and thiothixene (Dillenkoffer et al, 1974).  “Only one of the 13 thiothixene patients had prolongation of the Q–T,” said Gallant later in an interview, “but 13 out of 13 patients on 800 milligrams a day of thioridazine, and 7 of 13 on 400 milligrams a day had prolongation of the Q–T interval.  We published that.  In fact, my cardiology fellow that read the EKGs could identify thioridazine, blind. . . .  After we published, somebody from Sandoz called and started yelling on the phone at me, criticizing me, saying I was unethical for publishing the data.  This was 1972 [1974], and I was shocked that someone from a pharmaceutical firm would start telling me I’m unethical for publishing these findings. . . .  It was solid, solid data and Sandoz Company never made any mention about it” (Gallant, 2011).

These early warnings did not prevent Sandoz from further marketing the preparation. Indeed, to go by the visual content of the company’s advertisements for Mellaril, the drug was pitched to physicians as especially suitable for geriatric use, a population at risk of cardiac complications.  And in 1978 George Simpson and co-workers at Rockland State Hospital found that it was precisely in the elderly that thioridazine prolonged QT intervals (Branchey et al, 1978).  “I stopped using thioridazine at that time,” Simpson later said in an interview (Simpson, 2011).

An analysis of images depicted in Mellaril advertisements in Diseases of the Nervous System (after 1989 the Journal of Clinical Psychiatry) showed that Sandoz launched four major ad campaigns featuring elderly “patients.”  For example in three ads which appeared between May and July 1983, a clearly elderly woman was shown and the text stated that Mellaril “helps keep the disturbed geriatric at home” (Dis Nerv Syst, 1983).  An ad featuring an older male golfer (“effective control of psychotic symptoms”) ran 14 times (Dis Nerv Syst, 1979–80).  Ban in his Psychopharmacology for the Aged, published in 1980, noted that “thioridazine has become one of the most extensively employed psychotropic drugs in the aged” (Ban, 1980).

While the Ban studies showed that cardiac conductance changes appeared at daily dosages above 150 mg., the above-mentioned ads indicated that dosages below 300 mg were relatively safe. (“Daily doses in excess of 300 mg should be used only in severe neuro- psychiatric  conditions.”)  (Dis Nerv Syst, 1979–80).

For Sandoz – and its successor organization Novartis – it was irresponsible not to say reckless to have ignored such warnings for more than thirty years, putting the lives of many patients at risk.  The entire story of shortsightedly placing corporate interests ahead of science could be found in an MBA curriculum on how not to market a pharmaceutical preparation.



Ban TA.  Psychopharmacology for the Aged. Basel: Karger; 1980. p 66–7.

Ban TA interviewed by Bunney WE, Jr, in An Oral History of Neuropsychopharmacology – The First Fifty Years: Peer Interviews (Thomas A Ban, editor), Volume 9 – “Update” (Barry Blackwell, volume editor). Nashville: American College of Neuropsychopharmacology; 2011. p. 19–35, 26–27.

Ban TA, St. Jean A, Desautels S. The effects of phenothiazines on the human electrocardiogram, in Neuropsychopharmacology (D Bente, PB Bradley, editors), Volume 4. Amsterdam: Elsevier; 1965, p.467–70.

Ban TA, St. Jean A. The effect of phenothiazines on the electrocardiogram. J Can Med Assoc 1964; 91:537–40.

Branchey MH, Lee JH, Amin R, Simpson GM.  High- and low-potency neuroleptics in elderly psychiatric patients.  J Am Med Assoc 1978; 239: 1860–2.

Desautels S, Filteau C, St.Jean A. Ventricular tachycardia associated with administration of thioridazine hydrochloride (Mellaril). J Can Med Assoc 1964; 90: 1030–1.

Dillenkoffer RL, George RB, Bishop MP, Gallant DM. Electrocardiographic evaluation of thiothixene: a double-blind comparison with thioridazine. Adv Biochem Psychopharmacol 1974; 9: 487–95.

Dinovo EC, Gottschalk LA, Noble EP,  Biener R. Isolation of a possible new metabolite of thioridazine and mesoridazine from human plasma.  Res Commun Chem Pathol Pharmacol 1974; 7: 489–96.

Dinovo EC, Gottschalk LA, Nandi BR, Geddes PG.  GLC analysis of thioridazine, mesoridazine, and their metabolites.  J Pharma Sci 1976; 65: 667–9.

Dis Nerv Syst 1979;40 (8) through 1980; 41 (11), Mellaril advertisement (14 instances).

Dis Nerv Syst 1983; 44 (5–7), Mellaril advertisement (3 instances).

Gallant DM interviewed by Ban TA, in An Oral History of Neuropsychopharmacology – The First Fifty Years: Peer Interviews (Thomas A. Ban, editor), Volume 4 – “Psychopharmacology” (Jerome Levine, volume editor).  Nashville: American College of Neuropsychopharmacology; 2011. p. 75–124, 88.

Gottschalk LA interviewed by Bunney WE. Jr, in An Oral History of Neuropsychopharmacology – The First Fifty Years: Peer Interviews (Thomas A. Ban, editor), Volume 1 – “Starting Up” (Edward Shorter, volume editor). Nashville: American College of Neuropsychopharmacology; 2011.  p. 265–75, 270.

Gottschalk LA interviewed by Ban TA, in An Oral History of Neuropsychopharmacology – The First Fifty Years: Peer Interviews (Thomas A. Ban, editor), Volume 9 – “Update” (Barry Blackwell, volume editor). Nashville: American College of Neuropsychopharmacology; 2011. p. 105–34, 119–20.

Graupner KL, Murphree OD. Electrocardiographic changes associated with the use of thioridazine. J Can Med Assoc 1964; 91: 537–40.

Kelly HG, Fay JE, Laverty SG.  Thioridazine hydrochloride (Mellaril): its effect on the electrocardiogram and a report of two fatalities with electrocardiographic abnormalities. J Can Med Assoc 1963; 89:546–54.

Kingstone E. Review of the literature – neuroleptic drugs and the ECG, in Toxicity and Adverse Reaction Studies with Neuroleptics and Antidepressants ( HE Lehmann, TA Ban, editors). Montreal: Quebec Psychopharmacological Research  Association; 1965, p. 121–8.

Novartis Pharmaceuticals Canada. Important Drug Warning, Sumegi B, Rousseau G to all physicians and pharmacists, July 31, 2000 (“Dear Doctor or Pharmacist”).

Psychiatric News, August 18, 2000, p. 28 (“FDA Approves New Warnings for Thioridazine, Valproate”).

Simpson GM interviewed by Ban TA, in An Oral History of Neuropsychopharmacology – The First Fifty Years: Peer Interviews (Thomas A. Ban, editor), Volume 9 – “Update” (Barry Blackwell, volume editor).  Nashville: American College of Neuropsychopharmacology; 285–312, 298.

St. Jean A, Desautels S, Ballon J, Ban TA. Studies with thioridazine, in  Toxicity and Adverse Reaction Studies with Neuroleptics and Antidepressants (HE Lehmann, TA Ban, editors). Montreal: Quebec Psychopharmacological Research Association; 1965 137–42.

Wendkos MH. Experiments with thioridazine, in Toxicity and Adverse Reaction Studies with Neuroleptics and Antidepressants (HE Lehmann, TA Ban, editors). Montreal: Quebec Psychopharmacological Research Association; 1965, p. 143–55.

Wendkos MH. The significance of electrocardiographic changes produced by thioridazine. J New Drugs 1964; 4:322–32.